tailings and mine waste 2010

ANDMINE

WASTE‘10

Tailings and Mine Waste ’10 contains the contributions from the thefourteenth annual Tailings and Mine Waste Conference held byColorado State University of Fort Collins, Colorado in conjunctionwith the University of Alberta and the University of British Columbia.The purpose of these conferences is to provide a forum for discussionand establishment of dialogue among all people in the mining industryand environmental community regarding tailings and mine waste.

Tailings and Mine Waste ’10 includes over 40 papers which presentstate-of-the-art papers on mine and mill tailings and mine waste, aswell as current and future issues facing the mining and environmentalcommunities. This includes matters dealing with technical capabilitiesand developments, regulations, and environmental concerns.

TAILINGS

TA

ILIN

GS A

ND

MIN

E W

AST

E ‘10

an informa business

cover_hinshaw10_LG.qxd 13-09-2010 15:33 Pagina 1

TAILINGS AND MINE WASTE ’10

PROCEEDINGS OF THE 14TH INTERNATIONAL CONFERENCE ON TAILINGS AND MINEWASTE, VAIL, COLORADO, USA, 17–20 OCTOBER 2010

Tailings and Mine Waste ’10

CRC PressTaylor & Francis Group6000 Broken Sound Parkway NW, Suite 300Boca Raton, FL 33487-2742

© 2011 by Taylor & Francis Group, LLCCRC Press is an imprint of Taylor & Francis Group, an Informa business

No claim to original U.S. Government worksVersion Date: 20121218

International Standard Book Number-13: 978-0-203-83088-8 (eBook - PDF)

This book contains information obtained from authentic and highly regarded sources. Reasonable efforts have been made to publish reliable data and information, but the author and publisher cannot assume responsibility for the valid-ity of all materials or the consequences of their use. The authors and publishers have attempted to trace the copyright holders of all material reproduced in this publication and apologize to copyright holders if permission to publish in this form has not been obtained. If any copyright material has not been acknowledged please write and let us know so we may rectify in any future reprint.

Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced, transmitted, or uti-lized in any form by any electronic, mechanical, or other means, now known or hereafter invented, including photocopy-ing, microfilming, and recording, or in any information storage or retrieval system, without written permission from the publishers.

For permission to photocopy or use material electronically from this work, please access www.copyright.com (http://www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC), 222 Rosewood Drive, Danvers, MA 01923, 978-750-8400. CCC is a not-for-profit organization that provides licenses and registration for a variety of users. For organizations that have been granted a photocopy license by the CCC, a separate system of payment has been arranged.

Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe.

Visit the Taylor & Francis Web site athttp://www.taylorandfrancis.com

and the CRC Press Web site athttp://www.crcpress.com

Tailings and Mine Waste ’10© 2011 Taylor & Francis Group, London, ISBN 978-0-415-61455-9

Table of Contents

Preface IX

Organization XI

Keynote papers

Improving the safety of mine waste impoundments 3N.R. Morgenstern

History and developments in the treatment of oil sands fine tailings 11J.C. Sobkowicz

Mill tailings

Tailings impoundment failures, black swans, incident avoidance, and checklists 33J. Caldwell & L. Charlebois

New directions in tailings management 41C. Strachan & J. Caldwell

Overview: Tailings disposal and dam construction practices in the 21st century 49A.J. Breitenbach

A history of South African slimes dams engineers 59J. Caldwell & G. McPhail

Unique geosynthetic liner system for uranium mill tailings disposal 65G.T. Corcoran & H.R. Roberts

Optimizing tailings deposition concentration at Minera Yanacocha, Peru 71M. Keevy & R. Cooke

Geotechnical considerations

Peak and critical-state shear strength of mine waste rock 79Z. Fox & J.A.H. Carraro

Ore geotechnical testing for heap leach pad design 91J. Lupo & A. Dolezal

Critical state liquefaction assessment of an upstream constructed tailings sand dam 101C.D. Anderson & T.L. Eldridge

Heap leach pad cover design analyses Salmon, Idaho 113I. Hutchison, A. Whitman, J. Juliani & T. Hadj-Hamou

The effect of tailings characteristics on cover system success 121J. Keller, M. Milczarek, T.M. Yao & M. Buchanan

V

Water management and water treatment

Dewatered tailings practice – trends and observations 133M. Davies, J. Lupo, T. Martin, E. McRoberts, M. Musse & D. Ritchie

Groundwater modeling at the Panna Maria uranium facility in supportof an ACL application 143M. Gard, J. Warner, L. Cope & K. Raabe

A priori and posterior probabilities in operational water balancesfor tailing storage facilities 157S.F. Truby, V. Lishnevsky & J.R. Kunkel

Single process arsenic and antimony removal using coagulation andmicrofiltration 165J.R. Tamburini, H.C. Liang & S.J. Billin

Mitigating impacts from acid-producing rock in Tennessee roadconstruction projects 171J.J. Gusek, V. Bateman, J. Ozment, L. Oliver, D. Kathman, J. Waples,T. Rutkowski, H. Moore, W. Bowden & A. Reither

20-day design build to save $50 million worth of equipment 187S.J. Tamburini & S.J. Billin

The simultaneous removal of arsenic and manganese at a gold mine in Nevada 195H.C. Liang, S.J. Billin & J.R. Tamburini

Geochemistry

The impact of short-term variations of weather conditions on the chemismof rain water runoff from flotation wastes of Mississippi Valley-type Zn-Pb ores(southern Poland) 203A. Bauerek

The effect of weathering on the acid-producing potential of the GoathillNorth Rock Pile, Questa mine, NM 213V.T. McLemore, N. Dunbar, S. Tachie-Menson & K. Donahue

Effect of reservoir pool changes on metals release frommining-contaminated sediment 229T. Moyer, B. Striggow, J. Eldridge & C. Zeller

Neutral mine drainage water-quality impacts from a former taconite mine 241B. Hanna

Benefits of timely and valid geochemical characterization of mine wastefor life of mine and closure planning: A case study of Newmont Boddington GoldMine in Western Australia 253N. Amoah, R. Haymont & G. Campbell

Containment systems

Disposal of coal mine slurry waste using geotextile containers at theNorth River Mine, Chevron Mining Inc. 265M. Watts & E. Trainer

VI

Identification, management and disposal of PCB-containing equipmentused in mines 275D.W. Bench

Waste management practices at Alaska’s large mines 285J. DiMarchi & J. Vohden

East Mission Flats Repository design—challenges and case history 295D.K. Vernon, Jr. & A. Mork

Physical properties of mill tailings as foundation material for wasterepositories, Bunker Hill Superfund Site 305J.S. Woolston

Dry Stack/Paste

Dry stack tailings design for the Rosemont Copper project 315L. Newman, K. Arnold & D. Wittwer

Dry stack tailings – design considerations 327J. Lupo & J. Hall

Reprocessing of tailings of Chador-Malu iron ore, Iran 335H. Nematollahi

Oil Sands

Suncor Pond 5 coke cap – The story of its conception, testing,and advance to full-scale construction 341P.S. Wells, J. Caldwell & J. Fournier

Treatment of fluid fine tailings with silica 347R.H. Moffett

Filtration tests on PVD filter jackets in fine oil sands tailings 355Y. Yao, A.F. van Tol, B. Everts & A. Mulder

Suncor oil sands tailings pond capping project 367G. Pollock, X. Liu, E. McRoberts, K. Williams, P.S. Wells & J. Fournier

Review of oil sands tailings technology options 381C.B. Powter, K.W. Biggar, M.J. Silva, G.T. McKenna & E.B. Scordo

Case study: Sand capping of weak tailings at Suncor’s Pond 1 393E. Olauson, R. Dawson & P.S. Wells

The use of geosynthetics in the reclamation of an oil sands tailings pond 401C. Athanassopoulos, P.S. Wells, S. Trinca & W. Urchik

A new approach to oil sand tailings management 409L. Lawrence & Z. Ali

Environmental issues

A landscape design approach for the sustainable reclamation activities of a post-miningarea in Cartagena, SE Spain 419S. Kabas, Á. Faz, D.M. Carmona, S. Martínez-Martínez, R. Zornoza & J.A. Acosta

Priority setting in Idaho’s Coeur d’Alene Basin 427B. Adams & D.R. Pitzler

VII

Acid mine drainage as a sustainable solution to eliminate risk and reduce costs 439J. Cormier

Chemical compound forms of cadmium in uranium tailings of Schneckenstein 451T. Naamoun & B. Merkel

Uranium residue impacts on ground and surface water resources at theSchneckenstein site in East Germany 457T. Naamoun & B. Merkel

Author index 471

VIII


Preface

This marks the fourteenth annual Tailings and Mine Waste Conference. The purpose of theseconferences is to provide a forum for discussion and establishment of dialogue among peopleinvolved in the mining industry and environmental community regarding tailings and mine waste.Previous conferences have been successful in providing opportunities for formal and informaldiscussion, exhibits by equipment and instrumentation companies, technical exhibits, and generalsocial interaction.

This year’s conference includes over 40 papers. These papers address the important issues facedby the mining industry today. These proceedings will provide a record of the discussions at theconference that will remain of value for many years.

IX


Organization

Organized by the Department of Civil and Environmental Engineering, Colorado State University,Fort Collins, Colorado in conjunction with the University of Alberta, Edmonton, Alberta and theUniversity of British Columbia, Vancouver, British Columbia.

ORGANIZING COMMITTEE

Daniel Overton Engineering Analytics, Inc., Fort Collins, Colorado(Committee Chair)

Peter Mundy Alfa Laval, Inc., Calgary, AlbertaMichael Smith AMEC, Englewood, ColoradoLoel Renshaw Ausenco PSI, Concord, CaliforniaAntonio Carraro Colorado State University, Fort Collins, ColoradoShawn Steiner ConeTec, Inc., Salt Lake City, UtahBill Thompson Golder Associates, Inc., Lakewood, ColoradoBryan Ulrich Knight Piésold Consulting, Elko, NevadaClint Strachan MWH, Fort Collins, ColoradoRobert Cooke Paterson & Cooke, Denver, ColoradoAndrew Robertson Robertson GeoConsultants, Inc., Vancouver, British ColumbiaLarry Cope SRK Consulting, Inc., Fort Collins, ColoradoMike Henderson Tetra Tech, Inc., Golden, ColoradoKirk Palicki URS Corporation, Denver, Colorado

XI

Keynote papers


Improving the safety of mine waste impoundments

N.R. MorgensternUniversity of Alberta, Edmonton, Canada

ABSTRACT: Although failures of tailings storage facilities persist, there have been numerousimprovements in practice over the last decade that enhance the safety of mine waste impoundments.Examples are provided of improvements in corporate and regulatory responsibility related to thisissue. The dam safety system applied to the Alberta oil sands industry is put forward as a successfulmodel. The role of Independent Tailings Dam Review Boards is discussed and emphasized as avaluable component in the safety system applied to all tailings storage facilities.

1 INTRODUCTION

In 1996 and 1999 the writer published two presentations that summarized the then state of practicewith respect to tailings and other related mine waste management (Morgenstern, 1996; 1999).The assessments were case history based, with a focus on modern structures as opposed to legacyfacilities. The recommendations arising from these studies were as follows:

“The mining industry must take action to reduce risk associated with waste management by:

• Improving quality control• Documenting construction and quality control by more use of as-built records• Improving construction procedures consistent with recommendations from well-qualified

geotechnical engineers familiar with the mining industry• Utilizing more third party reviews• Ensuring that there is no conflict between short term profitability and integrity of containment• Ensuring that the responsibility for failure of waste containment structures is understood at the

highest corporate levels and that the standard of care is set by senior mine management.”

The intent of this presentation is to survey how far the industry has come in responding to theserecommendations. The role of third party reviews and highlights from the experience of the Writerwith such reviews will be singled out for more detailed discussion. It is the Writer’s hope that thisupdate will encourage on-going assessments of the safety management systems that are developingin different jurisdictions.

2 RECENT HISTORY OF MAJOR FINDINGS

In order to illustrate the current state of practice, Table 1 has been created that lists publicly knowntailings dam failures over the past decade (2001–2010). Waste dump, pit wall and heap leach failureincidents are not included.

This list is not intended to be definitive. Davies and Martin (2009) summarize a comprehensivedata base, augmented by their personal files, that indicate that there is no substantive reduction inthe temporal pattern of failure incidents.

Intriguingly, and not unreasonably, they suggest that there is a pattern with the periods ofcommodity price peaks that indicates that safety may be compromised by market forces. A numberof reasons are put forward to account for this correlation that challenge designers, operators andregulators alike. Reference to some of these issues will appear later in this presentation.

3

Table 1. Tailings dam failures.

Name (Place) Country Year

Sebastião dos Äquas Claras Brazil 2001San Marcelino Philippines 2002Cerro Negro Chile 2003Malvési France 2004Partizansk Russia 2004Riverview USA 2004Pinchi Lake Canada 2004Bangs Lake USA 2005Miliang China 2006Nchanga Zambia 2006Taoshi China 2008Kingston USA 2008Huayuan China 2009Karamken Russia 2009Las Palmas Chile 2010

Source: www.wise-uranium.org/mdaf.htmlwww.geerassociation.org

Table 1 and other unpublished data confirm conclusions made by the Writer in the past; namely:

• the failures reflect the current state-of-practice internationally in the industry• there is no socio-economic pattern among the cases, with regulatory environments ranging from

weak to strong• in no case, to the knowledge of the Writer, was there systematic third party review.

3 IMPROVEMENTS IN CORPORATE RESPONSIBILITY

In response to the international failure incidents in the 1990’s, the Mining Association of Canadaestablished a task force in 1996 to promote safe and environmentally responsible managementof tailings and mine waste. The Task Force concluded that the main priority should focus onimprovement of tailings management. This resulted in the establishment of the MAC TailingsWorking Group with broad industrial representation.

The need for a Tailings Management System was regarded as necessary to support industry’scommitment to continual improvement in health, safety and environmental stewardship. The firstproduct of this effort was the document “MAC Guide to the Management of Tailings Facilities”which presents:

• a framework of management principles, policies and objectives• checklists for implementing the framework through the life cycle of a tailings facility• lists of technical considerations.

It is in a format that is adaptable to specific site and corporate considerations.Following implementation of this advance it was recognized that there was a need for further

guidance to outline site specific procedures for the safe operation, maintenance and surveillance(OMS) of facilities. This resulted in the publication “Developing OMS Manuals for Tailings andWater Management Facilities”.

It is common in mining practice to verify conformance through technical audits. The need forguidance in this regard led to the production of a third guide, “A Guide to Audit and Assessmentof Tailings Facility Management”.

A history of the development of these guides has been presented by Gardiner and Gladwin(2009) and the guides themselves are freely available from the Mining Association of Canada(www.mining.ca).

4

In the experience of the Writer who has conducted occasional corporate wide audits, if corporatemanagement is found deficient, it is usually a result of lack of compliance with MAC guidelines.The MAC guidelines are readily adaptable to non-Canadian jurisdictions and site conditions ofany kind. They represent best available technology in the mining industry to-day and the Writerrecommends that all operators commit to compliance with them, subject to their adaptation to localcircumstances.

4 IMPROVEMENTS IN REGULATORY RESPONSIBILITY

It is the view of the Writer that the practice of the mining industry is too variable, as a whole, forit to be self-regulating. A well-supported, technically skilled, transparent, regulatory system is anintegral part of sustainable mining and safe management of mine waste.

On October 11, 2000, near Inez, Kentucky, a breakthrough occurred in which a 72 acre surfaceimpoundment of waste materials of the Martin County Coal Corporation released approximately250 million gallons of slurry into a nearby underground coal mine and subsequently into nearbycreeks and rivers. While there was no loss of life, the environmental damage was significant.This incident prompted the U.S. Congress to request the National Research Council to examineways to reduce the potential for similar accidents in the future. To conduct this study, the NationalResearch Council appointed the Committee on Coal Waste Impoundments, which included theWriter. Following a number of meetings, the Committee issued its report in 2002 (NRC, 2002).

Arising from the many observations and conclusions made by Committee, recommendationswere made to improve the regulatory process as practiced by the Mine Safety and Health Admin-istration (MSHA) and the Office of Surface Mining (OSM). Recommendations were made bothwith regard to technical and process-related considerations with a strong emphasis on dam safetyconsiderations.

The Writer was pleased to see the recent update and revision of the MSHA Engineering andDesign Manual – Coal Refuse Disposal Facilities which, together with other publications, is respon-sive to the needs felt by the NRC Committee. A detailed discussion of the advances made withinMSHA related to the safety of mining industry dams is presented by Fredlund (2009).

It is not possible to ensure that the regulatory process will be constructive in all environmentsand jurisdictions encountered by the mining industry. The industry has to be particularly diligenton its own when regulatory review does not contribute effectively to technical assessment. As isabundantly clear, there is much more to ensuring tailings dam safety than getting a permit.

In circumstances where regulatory review is technically weak, the Owner and Engineer shouldexercise considerable caution to ensure that appropriate, as opposed to minimal, safety standardsare being met.

5 DAM SAFETY AND THE ALBERTA OIL SANDS

“If any of those tailings ponds were ever to breach and discharge into the river, the world wouldforever forget about the Exxon Valdez”.

David Schindler, internationally respected water ecologist.

The Alberta Oil Sands contain an estimated bitumen resource of about 2 trillion barrels which,if recoverable, could satisfy North American oil demands for several generations. Currently, thereserves, which are recoverable with current technology, are estimated at about 177 billion barrels.Of the total reserves about 20% are recoverable by current mining methods while the rest requirecurrent and developing in-situ techniques to extract the bitumen. At this time production is 1.5million barrels per day, split almost evenly between mining and in-situ methods. Based on currentplans, it is projected to double by about 2020 and will produce about 1 billion barrels per year. Ifoil averaged about $100/barrel over this period, the gross revenue in 2020 would be $100 billionper year. Current and projected investment in the Alberta oil sands is a vast undertaking by anystandard.

The bitumen is extracted from mined oil sands by water-based processes. No other techniqueshave been demonstrated to be commercial. Tailings are a necessary outcome of current methodsof extraction and tailings ponds have been used extensively to manage the tailings. At this time,

5

tailings ponds are estimated to have a surface area of 130 square kilometers (50 square miles) witha volume of 720 billion litres (190 billion gallons). The footprint of the ponds is clearly visible oncurrent satellite imagery. A common strategy in mine development is to begin operations with anout-of-pit pond and to subsequently deposit tailings in-pit. Hence, not all ponds are contained bydams. Nevertheless, there are numerous large dams and ensuring their integrity is of paramountimportance. As is evident from the opening quotation, considerable attention is focused on theirbehavior.

The first tailings structure began construction in June, 1968 under the auspices of Great CanadianOil Sands, now Suncor. It was initially conceived as a dyke, 12 m high, to contain fluids from tailingsdeposited from the top of the escarpment of the Athabasca River. The tailings were expected to takea slope of 8% and release water sufficiently clarified that it could be re-cycled to the process. Thisproved not to be the case and considerable innovation was required in tailings dam construction.The ultimate dyke constructed was 92 m high, partly on muskeg and normally consolidated clay.This was a considerable achievement in its time. Morsey et al., 1995 report on several aspects ofthe foundation behavior.

The second stage of commercial development was initiated by Syncrude Canada Ltd. The out-of-pit tailings pond required to support this project was about 18 kms in perimeter and is likely thelargest earth structure in the world in terms of volume of engineered fill. Parts of the foundationof this structure is comprised of high plasticity clay shales whose strength has been reduced to theresidual state by means of glacial drag process. The observational method was employed throughoutthe construction of this structure in order to bring it to final design. Significant deformationsdeveloped in parts of the foundation, even though the overall factor of safety was about 1·3. Themechanics of this mechanism are now well understood and are discussed by d’Alencar, 1994.

Both soft and pre-sheared clayey foundations have been encountered in a number of the tailingsdams. In other instances it has been necessary to construct tailings ponds over sand channels. Herea primary focus is on hydrogeological considerations with relief wells required to depressurize thefoundations of the dam and extensive cutoffs and pumped well collection systems utilized to containthe process-affected water within the lease. The industry currently operates under conditions ofzero release of process-affected water.

To date, all tailings containment structures in the oil sands industry have been managed in a safemanner and it is of interest to describe and understand the dam safety system that has arisen.

It is the view of the Writer that the dam safety system applied to the Alberta oil sands industryis the best in the world. It relies on responsibilities of a number of stake holders and it is based onan intimate understanding and application of the observational method (Peck, 1969).

It has the following components:

• each owner is cognizant of its responsibilities to provide a tailings management system consistentwith the MAC guidelines.

• each owner has staff qualified in the management of tailings dams.• owners retain consulting engineers for design and construction supervision who are well-known

for their expertise in tailings dam design with special reference to the circumstances associatedwith the oil sands industry; the designer acts as the Engineer-of-Record; senior internal reviewof design submissions is expected.

• designs are compliant with at least CDA (Canadian Dam Association) Guidelines.• designs rely on the detailed application of the observational method for risk management.• designs are reviewed by theAlberta Dam Safety Branch, the regulator, who have staff well-versed

in dam design and construction.• an annual report is submitted each year to the regulator by the owner, supported by the Engineer-

of-Record, that the dam is behaving as intended; if not actions that have been or need to be takenare indicated.

• in accordance with CDA Guidelines, approximately every five years the owner retains anengineer, other than the Engineer-of-Record, to undertake an independent assessment of damsafety.

• each owner retains an Independent Geotechnical Review Board comprised of senior specialists,to provide on-going third party review of geotechnical issues of significance to the operation.One of the major responsibilities of such Boards is to review all aspects related to safety oftailings dams over the life cycle from design, construction, operation and closure.

6

The success of the dam safety system applied to the Alberta oil sands industry relies on responsi-bilities shared by the owner, the Engineer-of-Record, the regulator and various levels of independentreview. The Writer is aware that in many jurisdictions, not all of these components will be mature.Under these circumstances, the remainder of the safety management team should exercise addi-tional caution to compensate for regional limitations. As many case histories continue to remindus, a permit to operate is not a guarantee against failure.

6 TAILINGS DAM REVIEW BOARDS

The appointment of an Independent Tailings Dam Review Board (ITRB) to provide third partyadvice on design, construction, operation and closure of tailings dams has become increasinglycommon and is recognized to provide value.

World Bank and other Lenders Groups are requiring formation of an ITRB. International FinanceCorporation/World Bank guidance and operating principles OP4.01 and OP4.37 establish therequirement to review the development of tailings dam design, construction and initial dam filling.Maintaining an ITRB through operations and closure will depend upon the scale of the facility.Often a single Board will be formed during operations to provide advice on all geotechnically sen-sitive matters, including slope stability, waste dumps and tailings management to closure. Somelarge corporations retain a third party review board for on-going advice on tailings operations tocomplement their internal technical audit systems.

Senior review is often invoked by designers and regulators. However for an ITRB to fulfillits role in an effective manner, it should be retained by the owner. Clear terms of referenceshould be established. The process of organizing an ITRB is now sufficiently well-known thatthe process of establishing terms of reference is not difficult. A general requirement is to coverat least the stages from design to first filling and to evoke an international standard of care in itsassessment and review. This does not preclude regionally tested experience provided that the out-comes are consistent with the appropriate standard of care. Hoek (2001) discuss geotechnicalreview boards in mining. The discussion is presented in the context of a review board con-cerned with mine slope stability issues, but the general guidance is equally applicable to tailingsdams.

Independent Review Boards have a long history in design and construction of water dams forpower, irrigation and water supply. Legislation in a number or countries such as France and theUnited Kingdom formalized third party reviews. After several catastrophic dam failures in the1960’s and early 1970’s regulation of dam safety was strengthened in North America and the roleof Independent Review Boards grew accordingly.

The Writer’s first experience with a review board in the mining industry was as an early memberof the Geotechnical Review Board (GRB) established by Syncrude Canada Ltd. in 1972. The GRBwas originally retained as a Board of Consultants to advise on the choice of mining method forthis large oil sand mining venture. The choice was to either follow previous experience of utilizingbucket wheel excavators or take advantage of the very large draglines that were coming on stream.Ultimately draglines were selected and the early years of the GRB were dominated by considerationsof safe and productive mining practice under very challenging conditions. As the mine matured andafter draglines were replaced by truck/shovel operations, the agenda became increasingly focusedon tailings related issues. McKenna (1998) has vividly described the operation of Syncrude’sGRB over its first twenty-five years (1972-1997). The GRB is still active. It meets twice yearlyand tailings management issues, including reclamation concerns, are major areas of discussion.All other oil sand mining operations have followed the Syncrude model in establishing reviewboards.

While the Writer had been involved in a number of aspects of dam design, including tailingsdams, in the early 1970’s, his first participation in a comprehensive assessment of dam safety was asa member of the Tar Island Tailings Dyke Design Review Panel, appointed byAlberta Environment,in 1975. Alberta Environment, the regulator, had at that time strengthened its dam safety regulationsand one of its first actions was to appoint this review panel. While it was commissioned by theregulator, representatives of the owner and Engineer-of-Record participated in the review (AlbertaEnvironment, 1977). Matich (1986) has provided a summary of review board practice to the mid1980’s.

7

In the experience of theWriter, geo-environmental concerns related to water quality and reclaima-bility were not considered in the design reviews of the 1970’s and for much of the 1980’s. It appearsthat the first recognition to improve tailings management in this regard emerged in the UnitedStates with legislation in 1978 leading to control of uranium tailings. ARD and metal-relatedawareness began about a decade later, as did the increased control of process-affected dischargesto the environment.

Geo-environmental considerations are now an integral part of the agenda for ITRB activities,even if closure is perceived to be far off. It is increasingly recognized that mining, consistent withsustainability objectives, requires ongoing interaction between the mine plan, the processing plan,the waste management plan and the reclamation plan. Neglect of the interface between these plansis a common source of unwelcome surprises. The intent of this presentation is to focus on damsafety and not to address further the geo-environmental design issues. It would be of interest todocument the historical emergence of their recognition.

7 EXPERIENCE

In the following, the Writer offers some observations based on his experience:

i) Perspective – The ITRB should reflect corporate values and international standards of care.It is not uncommon for the well-meaning Engineer and his client, the well-meaning ProjectManager, both sensitive to the demands of schedule and budget, to make recommendations thatexceed the corporate appetite for risk. The ITRB provides one check against this possibility. Itis also not uncommon for regional practice to lag international practice and accepting regionalpractice may also incur extra risk. The ITRB also can assist the Owner in assessing risk whenit is perceived that regulatory requirements are unreasonable. The concept of “a permit at anyprice” carries with it risks that require evaluation.

ii) Phasing – An ITRB is best appointed at the conceptual design phase of a Project. Currentexperience indicates that one is usually necessary at the “bankable feasibility study” phase andthereafter if the project proceeds. Following agreement on feasibility design, the next milestoneis final design and documents, issued for construction (IFC).

The Writer finds it increasingly productive for the ITRB to be intimately involved at theIFC stage and immediately thereafter to ensure that the specifications, QA and QC programare clear and executable by all parties; the Owner, the Engineer and the Contractor, includingEPCM Contractor, if appropriate.

The Writer finds that additional effort for construction compliance is sometimes neededwhen the Owner acts as Contractor. The obligations of the Owner as “Contractor-of-Record”should not differ from those of a third party “Contractor-of-Record”.

The Writer places a great emphasis on the Construction Report which is intended to doc-ument that construction proceeded as intended. It is more than just a summary of as builtsdrawings, compaction test data, and membrane test data. The Report is the fundamental ref-erence document for subsequent evaluations of dam safety. In the view of the Writer, it isbest co-ordinated and prepared by the Engineer-of-Record, with inputs as appropriate from thevarious Contractors. The significance of these Construction Reports is under-estimated in theindustry.

The ITRB should continue through construction and start-up. Whether it, or some equivalent,continues through operations and closure will depend upon the practice, scale and needs of theOwner.

iii) Board Practice –The shelf life of ITRB reports during design, construction, and commissioningis limited. It is essential for the Board to debrief site and senior management on their findingsat the end of the meeting. Submission of written reports should follow shortly thereafter.

iv) Recurrent Technical Issues – The Writer finds that the following design/QC issues occursufficiently frequently that they merit recording here:• Although there is an adequate consensus on how to establish the seismic hazard at dam-sites,

there is a tendance to underestimate the consequences of failure during operations and adoptearthquake loading that is too low; this tendancy is accentuated in areas of minor to moderateseismicity.

8

• There is variable practice in assessing Probable Maximum Precipitation and Floods fordesign against extreme events. Corporate risk taking should be cognizant of consequenceassessments and recommendations provided by modern design guides.

• Notwithstanding the understanding and guidance in the literature regarding the need toaddress the potential undrained failure of saturated contractant tailings in terms of undrainedresistance, the issue remains poorly understood in much practice (Martin and McRoberts,1999; Fourie, 2008; Veillette et al., 2008). This is particularly the case where dams are con-structed by the upstream method, with assistance by sub-aerial drying. Even under theseconditions, it is possible to develop zones of saturated loose tailings in the deposit.

8 CONCLUDING REMARKS

Over the last decade or so there has been substantial progress in responding to the issues identifiedby the Writer in 1996. Progress has been made in strengthening corporate understanding andresponsibility for tailings management, improving technical tools for the design, enhancing thecapability of regulators and increasing oversight by utilizing more third party reviews. Howeverfailures still occur.

The price for improved safety is consistent review of the safe management protocols applied toany project to assess whether they are adequately in place, coupled with continued vigilance.

ACKNOWLEDGEMENT

The writer appreciates discussions with M.A.J. (Fred) Matich, P.Eng., in the preparation of thispaper.

REFERENCES

Alberta Environment, 1977. Report on Great Canadian Oil Sands Tar Island Tailings Dyke, Design ReviewPanel.

d’Alencar, J., Morgenstern, N.R. and Chan, D.H., 1994. Analysis of foundation deformations beneath theSyncrude tailings dyke. Canadian Geotechnical Journal, Vol. 31, p. 868–884.

Davies, M. and Martin, T., 2009. Mining market cycles and tailings dam incidents. Tailings and Mine Waste ’09,Proceedings 13th International Conference on Tailings and Mine Waste, p. 3–14, University of AlbertaGeotechnical Centre, Edmonton, Canada.

Fourie, A.B., 2008. Future tailings management strategies – High time we took the high road. Tailings andMine Waste ’08. Proceedings 12th International Conference on Tailings and Mine Waste, p. 3–15, CRCPress/Balkema, The Netherlands.

Fredland, J.W., 2008. Developments in the safety and security of mining industry dams. Tailings and MineWaste ’08. Proceedings 12th International Conference on Tailings and Mine Waste, p. 345–354, CRCPress/Balkema, The Netherlands.

Gardiner, E. and Gladwin, D., 2008. Working for responsible management of tailings facilities. Tailings andMine Waste ’08. Proceedings 12th International Conference on Tailings and Mine Waste, p. 337–344, CRCPress/Balkema, The Netherlands.

Hoek, E., 2001. Geotechnical review boards in mining. Geotechnical News, March, p. 43–45.Martin, T.E. and McRoberts, E.C., 1999. Some considerations in the stability analysis of upstream tailings

dams. Tailings and Mine Waste ’09, Proceedings 13th International Conference on Tailings and Mine Waste,p. 303–313, University of Alberta Geotechnical Centre, Edmonton, Canada.

Matich, M.A.J., 1986. Design and review boards. Alberta Dam Safety Seminar, Alberta Environment, 11pages.

McKenna, G., 1998. Celebrating 25 years: Syncrude’s Review Board. Geotechnical News, September,p. 34–41.

Morgenstern, Norbert R., 1999. Geotechnics and mine waste management – an update. Proceedings of Work-shop on Risk Assessment and Contingency Planning in Tailings Management Systems, Buenos Aires,International Council on Metals in the Environment, p. 171–175.

9

Morgenstern, Norbert, R., 1996. Geotechnics and mine waste management. Proceedings, International Sympo-sium on Seismic and Environmental Aspects of Dam Design: Earth, Concrete, and Tailings Dams. Santiago,Chile, Vol. 2, p. 5–26, Souidad Chilena de Geotecnica.

Morsey, M., Morgenstern, N.R. and Chan, D.H., 1995. Simulation of creep deformation in the foundation ofTar Island Dyke. Canadian Geotechnical Journal, 1995, Vol. 32, p. 1002–1023.

National Research Council, 2002. Coal Waste Impoundments: Risks, Responses and Alternatives. NationalAcademy Press.

Peck, R.B., 1969. Advantage and limitations of the Observational Method in applied soil mechanics.Geotechnique, Vol. 19, p. 171–187.

Velillette, M.F., Martin, T.E. and Larreta, S.A., 2008. Stabilized upstream tailings dam and converted into afiltered tailings facility. Tailings and Mine Waste ’08. Proceedings 12th International Conference onTailingsand Mine Waste, p. 437–448, CRC Press/Balkema, The Netherlands.

10


History and developments in the treatment of oil sands fine tailings

J.C. SobkowiczThurber Engineering Ltd., Calgary, Canada

ABSTRACT: Oil sand mine operators have encountered many challenges in storing and treatingfine tailings since commercial mining started in 1967. A remarkably broad research and develop-ment effort has addressed these challenges, finding solutions such as enhanced capture of finesin beaches, non-segregating tailings and thickened tailings – all intended to improve fines captureand reduce fluid fine tailings inventories. Recently developed treatment methods, such as in-linethickening of tailings combined with thin-lift dewatering or with centrifuging, or freeze-thaw con-solidation of tailings, are in advanced stages of pilot testing. Industry efforts to find effective andeconomically responsible solutions to the fine tailings challenge have been focused by the releaseof the ERCB Tailings Directive 074 in 2009. This paper reviews the history of oil sands tailingsresearch, presents recent developments, discusses the value and weaknesses of Directive 074, andcomments on where treatment of fine tailings is likely headed.

1 INTRODUCTION

Operators have been coping with the vagaries of oil sand tailings since the very beginning ofcommercial mining of oil sands, at the start-up of the Great Canadian Oil Sands (GCOS; nowSuncor) mine in 1967. The original scheme envisaged the discharge of tailings off the AthabascaRiver escarpment, with the sand and fines settling out in a long beach (with an 8% slope), and cleanwater being collected behind a 12 m high toe dyke and recycled to the bitumen extraction plant.The first (of many) painful industry “lessons learned” was that the fines portion of the tailings doesnot settle nor consolidate quickly, and thus water and fines must be stored in ponds for a significantperiod of time. For GCOS, this required the rapid design and eventual construction of a 92 m highdyke (the Tar Island Dyke) and a large tailings pond, to continue operations. The dyke design andconstruction encountered many geotechnical challenges on its path to completion.

Some newcomers to the industry (young engineers and scientists, staff of new oil sand miningcompanies, new regulators, and newly-aware environmentalists), seem to form the opinion thata) little effort has been made to solve the oil sands fine tailings challenges, b) potentially appli-cable technologies from mines in other parts of the world have been ignored, and c) progress onclosure and reclamation of tailings ponds is too slow. Use of intentionally inflammatory and mis-leading statements by some stakeholders does not help matters, but rather perpetuates confusionand misunderstanding of the real and complex issues.

The author’s opinion, based on 35 years of personal involvement, is that the truth is largelycontrary to the above-stated opinions, and specifically that:

1. A large group of engineers and scientists, working for different companies and research insti-tutes, has carried out a remarkably broad research and development effort, with the intent ofunderstanding and solving the many oil sands tailings issues that have arisen since 1967. Thishas led to the adoption of helpful technologies, such as the enhanced capture of fines in tail-ings beaches, non-segregating tailings, and thickened tailings. Research and development efforthas continually increased over the past 43 years and now is at its most intense (with the resultdescribed at the end of point 3).

2. Tailings treatment technologies from existing mines around the world have been considered, pilottested, assessed economically, and adapted where possible to oil sands tailings. Considerable

11

effort has also been expended in universities and private research facilities to develop newtechnologies or progress existing technologies for tailings treatment.

3. The number of tailings ponds has increased in relationship to an increase in the rate of miningand with the need to temporarily store fine tailings while effective and economically responsiblesolutions are found for their permanent storage in a closure landscape. At times, “dead ends” havebeen encountered in finding solutions, but overall there has been impressive progress. Thanksmainly to their own efforts and partly to encouragement by regulators, industry operators areclose to achieving this objective, with the first tailings ponds at Suncor now capped and reclaimed,several tailings ponds at Suncor and Syncrude in the process of being capped and reclaimed, andcommercial operations underway at Suncor to curtail the growth of MFT inventories. Develop-ment work in the final stages of field demonstration will bring similar and additional, affordabletechnologies for decreasing MFT inventories into production at the other operating mines withina few years.

It takes time to find, absorb and appreciate the wealth of existing information on oil sands tailingstreatment methods and technologies – there is a large body of information extending back over40 years! Any uninformed person would be wise to make that investment, and not to settle for theabundant misinformation easily available on the Internet. As a guide to both the newcomer andpracticing engineers and scientists, this paper provides a brief (and admittedly incomplete) historyof the research and development work on oil sands tailings treatment, an opinion on where tailingstreatment technologies are headed, and a discussion of the role of the regulator in the overall process.

As background information, a detailed discussion of oil sand tailings challenges, current tailingsdisposal practices, and a full suite of tailings treatment methodologies may be found in Sobkowiczand Morgenstern (2009) and in Hyndman and Sobkowicz (2010). A comprehensive bibliography oftechnical papers on these topics will be available soon on the Oil Sands Research and InformationNetwork (OSRIN) web site – http://www.see.ualberta.ca/OSRIN.cfm. An excellent list of researchcarried out over the past 10 years and currently underway at the Oil Sands Technology and ResearchFacility (OSTRF) can be found at www.ostrf.com/research.

2 HISTORICAL TAILINGS MANAGEMENT

2.1 Tailings and water storage imperatives

“Bitumen is extracted from mined oil sands by water-based processes. No other techniques havebeen demonstrated to be commercial. Tailings are a necessary outcome of current methods ofextraction and tailings ponds have been used extensively to manage the tailings. At this time,tailings ponds are estimated to have a surface area of 130 square kilometers (50 square miles) witha volume of 720 billion litres (720 million m3 or 190 billion gallons). The footprint of the pondsis clearly visible on current satellite imagery. A common strategy in mine development is to beginoperations with an out-of-pit pond and to subsequently deposit tailings in-pit.” (Morgenstern, 2010).

The following points, adapted from Sobkowicz and Morgenstern (2009), highlight some of thetailings disposal challenges encountered in oil sands mining, and explain why oil sand tailingsponds are so large and take a long time to bring to closure:

– The mine pits are 100 m or more deep. To develop in-pit storage for tailings, mining mustproceed to the base of ore, expand an area for the pond itself, and then open up a sufficientlylarge footprint at the base of the mine to accommodate any in-pit dykes required for containment.Due to poor foundation conditions, these in-pit dykes can have very flat slopes and thus theyplace a high demand on providing footprint at the base of the mine, before dyke constructioncommences. Thus, development of the mine pit and construction of containment dykes to providein-pit tailings storage can take from 5 to 10 years.

– While in-pit tailings storage space is being developed, the tailings must be stored in a relativelylarge, out-of-pit, above ground facility. As mentioned above, this out-of-pit facility must providesufficient volume to store 5 to 10 years’ worth of tailings production. Once tailings disposalmoves in-pit, the out-of-pit pond is usually still needed for clarification and recycle of water tothe extraction plant, which further delays the closure and reclamation of the out-of-pit pond (thecontainment dykes can be reclaimed during this time).

12

– When tailings disposal does move in-pit, the storage areas are deep and often laterally constrained,which can result in the accumulation of very thick deposits of relatively soft tailings in a shortperiod of time.

– Fine tailings settle and consolidate very slowly, and the rate of consolidation is slowest inquickly accumulated, thick deposits. In the past, ponds containing fine tailings were not readyfor reclamation and closure for tens of years after cessation of operations, depending on theclosure method selected.

– Due to the abundance of ore grade deposits, many of the mine leases are highly constrained inavailable surface area on which to dispose of waste. Overburden dumps, tailings ponds, DDAs,thin lift dewatering areas, and other waste disposal facilities all compete for limited out-of-pitand in-pit space. This fact complicates tailings operations and often prevents early closure oftailings ponds.

– The industry is constrained during operation to zero discharge of process-affected water. There-fore tailings management is intimately related to the site-wide water balance and the provision ofreclaim water to the extraction plant. Efficient water management also delays closure of tailingsponds.

2.2 Out-of-pit tailings ponds

All operating mines started with an out-of-pit tailings pond (Suncor – Tar Island Dyke / Pond 1 attheir base mine and Pond 8A/8B at the Millennium Mine; Syncrude – the Mildred Lake SettlingBasin [MLSB] for the Base Mine and the Aurora Settling Basin [ASB] for the Aurora North Mine;Shell – the External Tailings Facility [ETF] for the Muskeg River Mine; CNRL – Pond 1 for theHorizon Mine).

As discussed in the opening paragraph, experience with Pond 1 at GCOS (Suncor) revealed theslow settlement behaviour of fine tailings and the need for large external ponds. Syncrude sizedthe MLSB accordingly, but after a few years of operation found that their predicted MFT make ratewas a bit high. They accordingly reduced the size of the MLSB footprint and moved the location ofthe west dyke on the topographically high side of the pond to the east, prior to its construction. Finetailings storage issues during simultaneous operation of Suncor’s three mines (Base, Steepbank andMillennium) forced construction of an additional external pond – the South Tailings Pond (STP).

There was only a limited ability to optimize fines capture in these external ponds, as there wasa continual pressure to meet basic tailings storage needs, and the construction of the pond dykes(from either overburden or hydraulically placed sand) was often just ahead of projected pond levelrise (plus freeboard and contingency storage requirements). However, it was recognized that normalbeaching practices often resulted in significant fines capture in the beaches placed sub-aqueouslyand sub-MFT (of from 60% to 75% of the fines in the ore and in the WT line).

2.3 In-pit tailings ponds

Early in-pit tailings ponds at Syncrude and Suncor were operated in a similar manner to out-of-pit ponds, and thus experienced some of the same operational constraints and challenges. The oneadvantage of an in-pit pond is that any fluid tailings can be contained below the surrounding groundsurface, and thus poses a much smaller risk of release off-lease.

3 DEVELOPMENTS IN THE TREATMENT OF OIL SAND FINE TAILINGS

3.1 General comments

This section provides a brief discussion of the various research and development efforts made bymany researchers in numerous organizations. The intent has been to emphasize the work done andits chronology, and to acknowledge the responsible oil sand companies and research institutes. Dueto space and time limitations, credit is not given to the individuals who did the work. The authorapologizes to those involved for this necessary omission.

A chronology of the various research programs and field trials was compiled on a 24 column by30 row table (time running vertically down the table and each column referring to specific tailings

13

Table 1. Chronology of mine start-up and related events.

Year Mine applications and start-up Other events

1967 GCOS mine start-up (later became Suncor)1978 Syncrude – Mildred Lake Mine start-up1989 Fine Tailings Consortium formed199019911992 Syncrude revised its Base Mine plan, looking

at the potential for storing non-segregatingtailings in the mined-out pit.

1993 Fine Tailings Symposium sponsored by theConsortium

1994 CONRAD formed1995 Book “Advances in Oil Sands Tailings Research”

published19961997 Syncrude Mildred Lake – North Mine start-up.

Shell Muskeg River Mine application approved.1998 Suncor Millenium Mine application approved1999 UofA seminar on paste and thickened tailings2000 Syncrude Aurora North Mine start-up CONRAD oil sands tailings seminar;

UofA seminar on paste and thickened tailings.20012002 Shell – Muskeg River Mine start-up and Jackpine Book “Paste and Thickened Tailings – A Guide”

Mine application; True North Fort Hills Mineapplication; CNRL Horizon Mine application

20032004 Suncor North Steepbank Mine application CONRAD oil sands tailings seminar;

OSTRF becomes operational.2005 Total buys Joslyn Lease from Deer Creek Energy200620072008 June – ERCB releases draft Tailings Directive;

OSTRF – First International Oil Sands TailingsConference in Edmonton

2009 CNRL – Horizon Mine start-up February – ERCB issues Tailings Directive 074(industry to submit plans to comply bySeptember 2009); Tailings and Mine Wasteconference in Banff, (focus on Oil Sands)

2010 Shell – Jackpine Mine projected start-up April – ERCB approves Fort Hills and SyncrudeTailings Plans; June – ERCB approves SuncorTailings Plans

2012 Imperial – Kearl Lake Mine – projected startup

treatment methods or related events). It was impossible to fit the full version of the master tableinto this paper, so it was sub-divided into the 10 tables included herein, each one dealing with twotreatment methods. A full version of the table can be obtained by contacting the author (e-mailaddress given on first page).

3.2 Chronology of mine applications and start-ups, and related events

As a backdrop for the research and development efforts described in the remaining tables, Table 1contains a listing of associated major events, such as oil sand Mine Applications and start-ups,early seminars, publication of books, and events related to the ERCB Directive 074.

14

Table 2. Chronology of in-pit and out-of-pit fine tailings storage R&D.

Year Out-of-pit storage of fine tailings In-pit storage of fine tailings

1967 GCOS discovers that fluid fine tailingsdoes not settle quickly in Tar Island Pond

Late 1960’s1970’s1978 Syncrude planning basis is to permanently

store MFT in mined-out pits, below OGSEarly 1980’s Syncrude learns more about consolidation

of MFT and “MFT make rate”, and reducessize (footprint) of MLSB (1981–82)

Late 1980’s1990 Trials at Syncrude on mixing lime and

acid-lime with various tailings products,with the objective of developing higherstrength, sufficient for reclamation.

199119921993 Syncrude hearings; discussion of the

concept of end-pit lakes with water-cappedMFT

1994 ERCB gives Syncrude approval to build awater-capped in-pit MFT pond (to store150 Mm3 of MFT).

1995 Transfer of MFT from MLSB to WIPat Syncrude

19962005 Start of Suncor South Tailings Pond

construction200620072008200920102012 Projected end of MFT placement in WIP at

Syncrude; start of BML trial.

3.3 Chronology of R&D on in-pit and out-of-pit storage of fine tailings

Table 2 contains some general comments on work directly related to sizing and use of in-pit andout-of-pit tailings ponds. The lessons learned from out-of-pit ponds have been discussed previouslyin Sections 1 and 2.2.

The main research effort on in-pit ponds is associated with the concept of storing some MFTbelow original ground level in water-capped “end-pit” lakes. The water in these lakes would cap thesettling MFT and bioremediate the naphthenic acid coming from it (or from off-spec CT) to provideacceptable water for discharge into the environment. This would also allow flow through of surfacewater in the closure landscape. The ERCB gave approval for a full-scale trial of this concept in 1994,and Syncrude has since been building the trial facility (referred to as the “Base Mine Lake”) andfilling it with MFT and water. The Base Mine Lake is expected to be complete and the trial to start in2012. A minimum of about 10 years will be needed to demonstrate the viability of the concept, andcertainly lessons will be learned about how the end-pit lake needs to be managed, in its early stages.

3.4 Chronology of R&D on filtration and of spiking of MFT into whole tailings

Table 3 lists research and development of filtration to increase fine (and in one case coarse) tailingsdensity, and of injecting MFT into WT lines (referred to as “spiking”) to directly increase finescapture in beaches.

15

Table 3. Chronology of R&D on filtration and MFT spiking to increase fines capture.

Year Filtration Spiking MFT into WT

Late 1960’s GCOS obtains patent for flocculationand drum filtration of TFT and MFT (1968)

1970’s Extensive study of filtration of WT atSyncrude in the late 1970’s; determinedto be too slow at fines contents >12%,to require too much filtration area, and tobe costly for disposal.

1978Early 1980’s Syncrude (owners) in 1980 obtains a

patent for flocculation and filtrationof oil sands tailings.

Late 1980’s OSLO (now Aurora South) proposedspiking of MFT into WT

1990 Syncrude studied the mechanisms of finescapture on beaches, including spiking WTwith MFT

1991 Syncrude conducts a field demonstrationof MFT spiking as “proof of concept”

1992 Syncrude application to ERCB includes20% volume reduction in MFT by “slurrydensification” and MFT spiking

1993 Prototype scale test of MFT spiking intoWT at Syncrude MLSB

1994 In 1994–95, filtration tests were run on Large scale tests of MFT spiking into WTthe whole tailings stream produced by the on Syncrude SWSS beaches (produced soft“Bitmin” process (a non-CWHE process). deposits), followed by limited commercial

operations. Operations suspended due toissues with “soft” beaches.

19951996199720042005 Total tests filtration of coarse and fine

tailings at bench scale (2005–2007?)2006 Fort Hills project – pilot of belt filtration

of bitmin whole tailings, with disposalby dry stacking.

20072008 OSTRF research on cross-flow filtration

of various oil sands tailings products(flow through porous pipe)

200920102012

Filtration of tailings has been practiced at other mines for many years and had early consider-ation by oil sand operators. Both GCOS and Syncrude obtained patents on filtration before fullyunderstanding how it would be implemented. Syncrude conducted an extensive study of filtrationin the late 1970’s, but found the then existing technology to be inadequate to the task (due to theimpact of fines and residual bitumen) and too costly. Later studies by other operators came toessentially the same conclusions.

More recent research at OSTRF on cross-flow filtration of tailings (as they flow through a porouspipe) show promising results, but have only been demonstrated at a lab scale and need to be scaledup to higher flows using the variable tailings typical of an operating environment.

16

Syncrude conducted extensive field testing of MFT spiking of WT in the early 1990’s, which wassufficiently promising that it was included in a 1992 application to the ERCB and led to prototypeand then limited commercial operation in the mid-1990’s. While successful in capturing additionalfines, issues with safe operation on soft beaches brought an end to the program. Some researcherswere of the opinion that this decision was premature, and fines spiking is still under considerationfor future application at several tailings facilities.

3.5 Chronology of R&D on non-segregating tailings and of the implementation of CT

One of the most intense research and development efforts in oil sand tailings has been on makinga tailings product that captures an optimal amount of fines and does not segregate upon dischargeinto a storage facility. The numerous efforts in this regard are listed in Table 4. Early work focusedon making “consolidated tailings” or CT by mixing MFT with WT, and achieving a stable mix byadding various chemicals, such as acid, lime, gypsum or CO2. As it became evident that highersolids contents were also necessary to make a non-segregating mix, the WT were run through acyclone and CT was produced by mixing MFT with CUT. Later, as thickeners became part ofthe extraction process, non-segregating tailings were made by mixing TT with CUT, which in theindustry was referred to as NST.

The laboratory and pilot scale testing of CT showed a viable technology for increasing finescapture, and the industry thus adopted it in commercial operations, in about 1996 at Suncor and2000 at Syncrude. Production issues were still encountered. The first was in producing a CT thatmet a prescribed, robust recipe – variable WT feed density, bitumen content and rock content,and unreliable supply of chemical additive, all posed challenges. It took operators several years toidentify and implement process changes and controls, so that they could reliably produce a CT thatmet specifications, but this is now “standard practice” at both Suncor and Syncrude.

The second challenge was to provide a sufficiently low energy environment when CT (or NST)is discharged to a DDA, so that it does not segregate under high shear stresses. Partial segregationof CT has been encountered in existing CT ponds. The issue is understood but does not have asimple solution, due to several challenges associated with winter tailings operations and optimaltremie operation from a floating platform (not always possible or desired). Oil sand operators areactively researching and developing discharge techniques to solve this problem.

3.6 Chronology of R&D on TT and on biogenic methods of increasing MFT density

In the mid-1990’s, oil sand operators became interested in using thickeners as part of the bitumenextraction process – primarily as a means of recovering clean water and heat in the extraction plant.But they also realized that a thickened tailings product, with a sand to fines ratio of around 1 and asolids content (by total mass) of 45% to 50%, provided opportunities to achieve higher fines capturein tailings deposits. From 1995 through to 2002, a considerable amount of field and pilot testingwas carried out as a joint industry effort under the CONRAD umbrella, (with the involvement ofSyncrude, Shell, Suncor and CANMET), as listed in Table 5.

The Shell Muskeg River Mine was the first to implement conventional thickeners in their tailingsprocess. The thickeners proved effective in meeting their primary objectives (of clean water andheat recovery) but the actual thickened tailings was quite variable in solids content due to variablefeed input from the extraction plant PSV’s. Shell intensely researched the use of high rate and pastethickeners at their pilot facility at the Muskeg River Mine from 2007 to 2010. As a result, theyhave reliably produced several higher solids content TT and NST deposits and have a more robustthickener design for their Jackpine Mine. The formation of DDA’s containing high solids contentTT (proposed for Shell’s Jackpine Mine), or NST made from high solids content TT, is a distinctpossibility arising from this research.

Research on various biogenic methods of densifying MFT is also given in Table 5. Two possiblescenarios have been investigated. One is associated with the apparent impact of methane gasformation, as observed at the MLSB in the mid-1990’s, on increased MFT consolidation rates.While progress has been made on the science, it is not clear that the “technique” is actuallyan effective or viable one. The other scenario is associated with the application of plant rootdewatering on MFT, which has significant implications for establishing accessible crusts on soft

17

Table 4. Chronology of R&D on CT and of implementation at Syncrude & Suncor.

Year CT at Suncor CT at Syncrude

Early 1980’s McGill University researches treating Experimental work at SyncrudeSuncor tailings with various chemicals to indicates that WT can be made“stabilize” the tailings. non-segregating by adding lime

Late 1980’s Research in the late 1980’s and very early 1990’s at the UofA on producing CT (using variousadditives to prevent segregation).

1990 Small field test of CT on MLSB toeberm. Lime was added to WT toinvestigate segregation behaviour.

19911992 Samples from 1991 field program

were tested at the UofA1993 Pilot test of both gypsum-CT and Syncrude conducted comprehensive

acid-lime-CT (UofA involvement) laboratory program on non-segregatingtailings at the UofA

1994 Mine planning and economic screeningstudies on non-segregating tailings.

1995 Commercial trials of CT in Pond 5 Medium scale demonstration of CT at(1995–96; using 30% of extraction tailings) SW corner of MLSB (100K m3).

Small add-on program looking atdischarging CT below MFT.

1996 Development and installation of verticaltremie barges for CT deposition (1996–97)

1997 Start of commercial operation of CT, 5 M m3 CT Prototype test at Syncrude,discharging into ponds in the NW corner of MLSB

199819992000 Syncrude starts commercial production

of CT in EIP200120022003 Tests on improving CT density by use of

underdrainage at Syncrude2004 Development of concept to sub-areally

beach CT20052006 Block model techniques developed for

assessment of CT in Ponds2007 Process/control improvements to produce CT to a more reliable and robust recipe at both

Suncor and Syncrude. Studies salso underway to assess best methods for dischargingCT in a low energy/low shear environment (these are still ongoing in 2010).

20082009 Suncor starts production of CT into Pond 7 Syncrude starts CT production into SWIP2010 Syncrude essentially completes CT placement

at EIP20122014 Projected start-up of CT production at

Syncrude’s Aurora North Mine

tailings deposits, and could have application in DDA’s if the ERCB’s Directive 074 is revised toallow capping of soft tailings (see Section 4.4).

3.7 Chronology of R&D on in-line flocculation of WT and of centrifuging MFT

Table 6 lists research and development work on in-line flocculation of WT and centrifuging ofMFT.

18

Table 5. Chronology of R&D on TT and on Biogenic methods of increasing MFT density.

Year TT Biogenic activity

Early 1980’sLate 1980’s Rumours of gas production and

biogenic activity on Suncor’sPond 1

19931994 UofA PhD thesis on plant root

dewatering of fine tailings.1995 Starting in 1995, research on thickening technology has1996 proceeded under the CONRAD umbrella. The late Significant methane gas bubbles1997 1990’s work included the following Syncrude/ noted in MFT in Syncrude MLSB1998 CANMET pilot tests:

Phase 1 bench tests 1995/1996.Phase 2 (2 tph unit at Syncrude Research using CHWE

and LEE process) December 1996 to April 1997.Phase 3 (Mildred Lake Site) Stream 73 (floatation line)

Nov, Dec 1997; CT prototype overflowAug, Sept 1997.Phase 4 (2 tph unit at Syncrude Research using LEE

process) Nov, Dec 1998.

Field trials on plant root dewateringof fine tailings at Syncrude.

1999 CONRAD – 10 m diameter conventional thickener builtat Aurora North mine

Syncrude noticed significantdecrease in MFT make(1999–2004) – one possiblecause was gas-enhanced drainageand consolidation of MFT

2000 TT tests at Aurora facility, first on tailings produced fromthe LEE process; later in 2001 on tailings producedfrom the CHWE process.

2001 Field investigation of gas bubblesin Syncrude MLSB by UofA

2002 Addition of hydro-cyclones; further tests of TT at Aurorafacility. Start-up of Shell’s Muskeg River mine usingconventional thickeners.

2003 Initial work on paste thickener at Syncrude UofA – significant research efforton impact of gas bubbles on MFTconsolidation.

2004 Work on a deep bed thickener at Syncrude200520062007200820092010

Ongoing research and development at Shell (MRM) onhigh rate and paste thickeners, on producing NST fromhigher solids contentTT and CUT, and on the behaviourof TT and NST deposits.

In-line flocculation of WT was trialed at both Syncrude and Suncor at up to a commercial scale(one full tailings line). If successful, this method would capture the fines carried in the WT stream,reducing or preventing segregation on discharge, and thus ultimately increasing fines capture in thesand beaches. While small scale tests showed some promise, large-scale tests were not successfulin preventing segregation upon discharge over the variable range of WT solids and fines contents.

Centrifuging of fine tailings, particularly MFT, is a viable tailings treatment method that has beenwell researched at CANMET and Syncrude, starting with bench scale tests in 2006, progressing tosmall and medium scale tests in ensuing years. The 2010 pilot is focused on field testing severaldifferent 600 mm diameter centrifuges, and on various methods of transporting and discharging theend product to a DDA (including trucking, positive displacement pumping and conveyor stacking).The 2008 medium scale tests were successful in achieving moderately high solids contents (∼60%),which would allow deposition without segregation. Post-deposition consolidation or environmental

19

Table 6. Chronology of R&D on in-line flocc and centrifuging to increase MFT density.

Year In-line flocculation of WT Centrifuging MFT

1967Late 1960’s GCOS obtains patent for flocculation and

centrifuging of oil sand tailings fines (1969).20012002 Syncrude test of In-line flocculation of

“middlings” (fine tailings), with“in-ground” thickener. High dilution of feedbefore addition of polymer for thickening.

200320042005 In-line flocculation testing of WT at

Syncrude – EIP. Small in-lineflocculation field test of WT atSuncor – Pond 6.

2006 Bench scale testing of centrifuging fine tailingsat CANMET

2007 Large in-line flocculation field test ofWT at Suncor – South Tailings Pond

Small scale field test at Syncrude, usingcentrifuge for drilling mud

2008 Medium scale field pilot at Syncrude (MLSB)using larger centrifuges

20092010 Medium scale, longer term test at Syncrude

MLSB. Using three × 600 mm diametercentrifuges. Deposition by trucking, PDpumping and conveyor stacking.

treatment to further increase solids content and develop sufficient strength for reclamation is stillbeing studied.

3.8 Chronology of R&D on MFT drying and on thin-lift dewatering to increase MFT density

Research on combined methods for in-line flocculation of MFT followed by thin-lift dewatering(or “drying”) has been seriously pursued starting in 2003 with bench scale tests at Suncor (Wellsand Riley, 2007; Wells, 2010). As indicated on Table 7, Suncor has consistently pursued this work,first with inorganic additives (2003 to 2006), and then with organic polymers starting in 2007.Successful piloting of this technique in 2008 led to a scale up of operations to commercial levelsin 2009, and then expansion of treatment areas and formal adoption of the treatment techniqueas Suncor’s Tailings Reduction Operations (TRO) in 2010. The TRO technology was accepted asmeeting the intent of Directive 074 by the ERCB in June of 2010.

Syncrude ran some initial trials of a similar technology in 2009, and has scaled up to a largerpilot test in 2010 at the MLSB, in conjunction with Total. Similar sized trials are also being run in2010 at the CNRL Horizon mine and Shell Muskeg River mine sites.

3.9 Chronology of freeze-thaw consolidation of MFT and of Rim ditching

Considerable work has been undertaken on freeze-thaw consolidation of MFT since the late 1980’s.This is not surprising, given the lengthy, cold winters in Fort McMurray, which the oil sand operatorswould hope to use to advantage in dewatering fine tailings. Fundamental testing was conductedthroughout the 1990’s at the University of Alberta. Small (0.3 hectare) to medium (1–2 hectare)sized field scale tests were carried out at both Syncrude and Suncor in the mid-1990’s, withadditional testing at Suncor (on Pond 1 beaches) in 2006. This work demonstrated the significantdewatering of MFT that can be accomplished (from a solids content of 30% to a solids content of45% to 55% after one freeze-thaw cycle) and the beneficial effects of freeze-thaw on improving

20

Table 7. Chronology of R&D on MFT drying and TLD to densify and store MFT.

Year MFT drying and TLD at Syncrude MFT drying and TLD at Suncor

19941995 Trials of wind-blown sand control at

Syncrude’s MLSB provides valuableinformation on MFT evaporative drying

19962003 First bench scale trials of thin lift MFT drying2004 Field trials on Suncor’s Pond 1 north beach – focus

on use of hydrated lime and gypsum additives.200520062007 Published 2004–6 work at 2007 paste conference.

Started lab trials on addition of polymerflocculant to MFT.

2008 Field pilot of in-line flocculated, thin liftdewatering of MFT at Pond 1.

2009 Initial trials of TLD at Syncrude MLSB Increasing commercialization of in-lineflocculated, thin-lift drying of MFT at variouslocations on Suncor lease. Adoption oftechnology by Suncor as TRO. Approval ofTRO technology by ERCB in June 2010.

2010 Major trials of TLD at Syncrude (partneredwith Total), CNRL, and Shell mine sites

Table 8. Chronology of R&D on Rim Ditching and freeze-thaw consolidation of MFT.

Year Rim Ditching Freeze-thaw

Early 1980’sLate 1980’s Research by AEC on freeze-thaw dewatering of

Syncrude (MLSB) MFT, from 1988–1993.199019911992 Initial lab work on freeze-thaw of Suncor fine tailings

by UofA1993 Freeze-thaw trial at Suncor (two × 1 hectare areas)1994 Freeze-thaw trial of Syncrude MFT (with additives)

by UofA – multiple layers in a 0.3 hectare area1995 Large scale freeze-thaw trial of MFT at Syncrude

BML (230K m3 of MFT placed partially on ice andpartially on shoreline).

19961997 Ongoing research of freeze-thaw behaviour of fine

tailings at UofA199820052006 Freeze-thaw tests on Pond 1 beach at Suncor – added

gypsum (later polymer); run on a sloped surface.2009 Rim ditching of in-line flocculated

MFT at Syncrude MLSB2010 Continuation of Rim Ditching trial

at Syncrude

21

Table 9. Chronolgy of R&D on capping of soft tailings deposits.

Year Sand capping of soft tailings deposits andReclamation

Coke capping of soft tailings deposits

1998 Sand capping of CT protype area in NW corner ofMLSB at Syncrude

1999 Reclamation of CT prototype area in NW cornerof MLSB at Syncrude

2000 “Bearing Capacity” field trials at Syncrude forsand capping of CT

2002 Trials at Suncor to sand cap CT2003 Hydraulic placement of coke cap on MFT at

Syncrude’s MLSB, starting about 2003 andcontinuing until current time

2004 Pilot for hydraulic capping of CT with cokeat Suncor

2006 Sand capping of CT at Syncrude’s NEIP Pilot of capping CT using various techniquesat Suncor Pond 1

20072008 Field trials of Coke capping on Suncor Pond 5

by mechanical placement20092010

At Syncrude: Fen construction in northernportion of NEIP. Landform construction(hummocks, swales, marshes, etc.) in southernportion of NEIP. At Suncor: capping of soft(Plant 4) tailings on Pond 1.

Commerical scale coke capping by mechanicalplacement on Suncor Pond 5. Haul roadsover soft deposits supporting fully loaded777’s.

MFT permeability and thus rate of consolidation. Freeze-thaw shows promise as a complementarytailings treatment method, but scale-up effects and area requirements have yet to be assessed.

Rim ditching of thick deposits of in-line flocculated MFT have recently been instigated atSyncrude (2009). While results so far have been very positive, research on this treatment methodis still in its early stages.

3.10 Chronology of R&D on capping of soft tailings deposits

Table 9 lists research and development on methods for capping soft tailings deposits, which has beenactive since the late 1990’s when CT became a viable tailings treatment process. Work has focusedon two types of capping material – sand and coke. The sand or coke can be placed hydraulicallyor mechanically, depending on the strength of the material to be capped and the economics ofplacement. Both Syncrude and Suncor have demonstrated that a coke cap can be placed on verysoft materials (Syncrude on MFT and Suncor on off-spec CT), with strengths much less than thevalue of 5 kPa specified in Directive 074. Once the cap is in place, the deposit can be accessed byheavy equipment (e.g. loaded 777 haul trucks) for further reclamation activities. Both companieshave also demonstrated construction of closure landscapes (hummocks, swales, etc.) by hydraulicplacement of sand caps.

3.11 Chronology of R&D on other tailings treatment techniques

Other tailings treatment techniques have been researched, as listed in Table 10. Two of these showsome promise and are still being actively investigated. Both promote drainage and consolidationof fine tailings – one is the use of vertical drains (such as wick drains, installed from the top ofa capped deposit) and the other is electrophoresis (for which significant technical advances havebeen made in recent years, resulting in significantly reduced power requirements).

Table 10 also mentions the timing of fundamental studies on the geotechnical behaviour of finetailings at the University of Alberta, and other studies focused on reducing the amount of clayparticle dispersion resulting from the bitumen extraction process, which is one of the primarycauses of the slow rate of settlement/consolidation of oil sands fine tailings.

22

Table 10. Chronology of R&D on other tailings treatment techniques and of reclamation and closure effortson tailings deposits.

Year Other treatment techniques and relatedmethods

Reclamation and closure of fine tailingsdeposits

1970’s GCOS obtains a patent to capture MFT in the bottomof a pond by dispersing a tailings stream across thesurface of the pond and “raining” sand down ontothe MFT layer.

1978Early1980’s

Early investigations on electrophoresis andelectrokinetics in the early 1980’s at Syncrude

19921993

1994

1995

Fundamental geotechnical studies on the settlementand consolidation behaviour of fluid fine tailingsand the UofA.

Considerable research on CHWE and OHWEprocesses, as well as other tailings technologies,in an effort to find an extraction method that had alower dispersion of fines and thus a lowergeneration of TFT and MFT. A lot of this researchled to the LEE process initially tried at Syncrude’sAurora Mine. Work continued until 2000.

1996 Formal closure planning for soft tailingsdeposits starts at Syncrude

19971998 Large, lab scale work at Syncrude to study the

effectiveness of wick drains in consolidating MFTFirst closure plan completed at Syncrude

19992000 Lab scale work on electrophoresis at Syncrude200120022003 Syncrude constructed a cyclo-stacker prototype,

placing several cones of sand to a height of 20 mover a 1 month period

2004

2005

Second phase cyclo-stacker trial at Syncrude, placingcones up to 40 m high, with evaluation ofgeotechnical properties

200620072008

Development of methods to displace, recover andpump high density MFT from Suncor Pond 1

2009 Small field scale electrophoresis tests on SyncrudeMFT (reduced power requirements). Initial trials atSuncor on enhanced consolidation of MFT and CTusing wick drains.

Work at Suncor to cap Pond 1 and atSyncrude to construct closurelandscapes (fens, marshes, hummocks,swales) on the NEIP deposit.

2010 Suncor Pond 1 capping (including cappingof soft, fine tailings) and reclamationcompleted

3.12 Reclamation and closure of tailings deposits

The final word of this section must be given to the oil sand operators who have successfullyreclaimed large areas of tailings deposits. These include:

– Tens of square kilometers of sand beaches and tailings sand slopes on all active mining leases.All reclaimed areas have healthy vegetation cover and some have been reclaimed for over 20years. While these are not “fine tailings”, they are nevertheless challenging to reclaim in a lastingmanner.

23

– 10 hectares of capped CT at the Syncrude prototype – this was reclaimed in 1999/2000, and iswell vegetated and treed. The vegetation and environmental conditions have been monitored onan ongoing basis at this site since 2000.

– Other small areas of soft tailings that have been capped and reclaimed (MLSB west and east toeberms; Syncrude Coke Cell 5).

– Syncrude has capped an area of about 5 km2 (500 hectares) on their NEIP pond (containingCT) and are in the process of constructing a) several fens at the northern end of the NEIP, andb) hummocks, swales and marshes in the southern part of the NEIP. Final reclamation work isexpected to be complete in 2012 (the fens taking the longest time to establish).

– Suncor has capped about 2 km2 (200 hectares) on their Pond 1 (Tar Island Pond), built a closurelandscape (with all associated surface water drainage), and planted trees and other vegetation.This brings to final reclamation the first full tailings pond in the industry.

4 ERCB TAILINGS DIRECTIVE 074

4.1 Contents

The ERCB issued a draft Tailings Directive in June of 2008, which after a brief discussion withindustry participants was issued in final form in February of 2009 (ERCB, 2009). The Directivewas developed in response to direction given to ERCB staff in July of 2004.

The Directive requires operators to reduce fluid tailings through fines captured in dedicateddisposal areas (DDAs), and to form and manage DDAs. The operators are further required tosubmit tailings plans, pond status reports, DDA plans, and annual compliance reports for DDAs,which will allow the ERCB to assess overall compliance with the directive.

The Directive was developed as the “. . . first component of a larger initiative to regulate tailingsmanagement. . .” (ERCB, 2009), and contains a list of long-term objectives which presumably willgovern the development of further regulations on tailings management.

Of relevance to this paper are several “technical” requirements contained in the Directive:

– Certain targets are given for fines capture in DDA’s, expressed as a percentage of the mass ofdry fines in the oil sands feed (this target is in addition to the fines captured in hydraulicallyplaced dykes and beaches). The target is 20% for mid-2010 to mid-2011, 30% for mid-2011 tomid-2012, and 50% thereafter.

– “DDAs must be formed in a manner that ensures trafficable deposits.” The stated criteria to meetthis objective is a) to achieve a minimum undrained shear strength of 5 kPa for material depositedin the previous year, b) to be ready for reclamation within five years of end of deposition, and c)when ready for reclamation, the trafficable surface layer must have a minimum undrained shearstrength of 10 kPa.

– Once the trafficable surface layer has been achieved, the operator is to file an application toabandon the DDA, (what “abandonment” means is not explicitly stated in the Directive, nor isthe need to establish a closure landscape addressed).

4.2 Impact on industry

All oil sand mine operators were required to submit, by September 2009, a plan explaining howtheir tailings management would comply with Directive 074. The ERCB reviewed these plans andhas to date approved those for Fort Hills, Syncrude and Suncor (Table 1).

In the author’s opinion, Directive 074 has (so far) had several impacts on the industry players,some beneficial and some onerous:

– One of the beneficial impacts is that the industry has had to focus their tailings research anddevelopment efforts to find specific tailings treatment solutions that can be implemented in arelatively short time period (consistent with their Directive submissions, as approved by theERCB). This was no doubt one of the ERCB objectives. However, without the large amount ofresearch and development work that had already been carried out by the operators (and associatedinstitutes), the task of implementing tailings technologies that met the Directive would have takenmuch longer (ten years or more). This is a fact that many industry detractors chose to ignore.

24

– An associated, less desirable outcome is that, with the rush to meet the Directive timeline,less than optimal tailings treatment technologies are being chosen. One might argue that opti-mization of technology and reduction of cost can occur as experience is gained implementingproposed solutions. However, the ability to do so is limited by the fact that significant (andat some facilities drastic) changes are being made to mine and tailings plans. As a result, theability to implement future changes (to a different tailings treatment technology) will be moreconstrained.

– Another beneficial impact of the Directive is that operating companies have renewed a level oftechnical cooperation in their research and development efforts that existed in the mid to late1990’s, again conducting joint field trials on new technology. This has increased the rate at whichimportant research information is shared and at which new technologies can be implemented.

– One of the onerous consequences is that the Directive requires operators to build facilities andtake actions that are incompatible with aspects of approved mine and tailings plans. For somemines, the amount of incompatibility is small; for others it is large. This is a complicated andserious issue, which unfortunately cannot be properly addressed here. The “short story” is thatthere is a considerable investment in developing a plant and mine, and similarly a considerableamount of time and money required to implement major technology changes. The ERCB isattempting to be flexible in their application of the Directive at each mine site, taking intoconsideration “. . . particular mining and tailings plans, facilities, and the status of a project. . .”(ERCB, 2009). However, one wonders how well ERCB staff actually understand this issue. Timewill tell how flexible the ERCB will be and what perhaps unnecessary cost will be imposed onthe operators. The amount of time that is being required for ERCB approval of company tailingssubmissions suggests that they are having a difficult time assessing all the issues and balancingall the competing objectives and priorities.

– An issue that is being given less attention at the moment, by both the operators and the ERCB,in the rush to progress tailings treatment methods already under development to the commercialscale, is the potential water quality impacts of the various tailings additives being proposed. Thisissue is not being ignored, and it will be examined fully in due course, but perhaps not in asorderly a manner as if the research had progressed at its own rational pace. The risk is that theERCB will endorse or approve a particular tailings treatment technology before all of the waterquality issues are fully understood.

– One of the main complaints of the operators is that the Directive, with good intentions, isimposing some conditions that are artificial and unnecessary in meeting stated ERCB objectives,and that will prevent operators from implementing equally effective but less costly solutions thanthose mandated by the Directive. This is discussed further in Section 4.4.

4.3 Roles and responsibilities of the regulator and the operators

It is clear that the Regulator has had and now has an important role to play in overseeing industryinitiatives and efforts to find effective tailings storage solutions. One would like to believe that theRegulator’s level of involvement has been about right and that the operators have, by and large, beenresponsibly pursuing these solutions. The truth may not be quite so rosy – perhaps the industry wasa little under-regulated prior to the mid-2000 time period, and perhaps the operators were being alittle less than fully diligent in pursuing their research and development efforts (or at a minimum,not entirely coordinated in their efforts).

It is difficult to make a fair assessment of both the Regulator’s and the operator’s roles in thepast, because the technical issues were complex and not easily solved, there were difficult economicpressures in the industry, and there were difficult challenges in scaling up and integrating apparentsolutions for tailings treatment to a commercial operation. Unforeseen challenges and costs aroseduring the implementation of the CT and TT technologies, and these were not always recognizedby the Regulator, nor responded to in a timely manner. However, operators were persistent inimproving problematic technologies and in pursuing new ones – credit for this should be givenboth to their own sense of corporate and social responsibility and to the regulatory environment inwhich they functioned.

Similar challenges will undoubtedly occur with the newer technologies for dewatering MFT, asthey are implemented, and the Regulator now has in place monitoring and reporting mechanismsthat will allow it (and stakeholders) to see what is happening.

25

It is the author’s opinion that Directive 074 is a useful regulation, responsibly conceived by theERCB. It is not perfect, but it is timely and it will have positive impacts on the industry. It alsoprovides a level of transparency to industry operations that will be welcomed by all shareholders.There is opportunity and need for the Directive to be improved, so that its objectives can bemet without unnecessary technical (and related cost) constraints, which will be discussed in thefollowing section.

The challenge for the ERCB will be to revise the Directive in a reasonable manner and to developfuture tailings management regulations that maintain the right balance of encouragement and “roomto move”, that is, to neither under-regulate nor over-regulate the industry. Careful thought will haveto be given to the impacts of economic cycles on the ability of the industry to maintain higher costtailings management practices. Current opinion is that the old tailings management practices wereinsufficient, but will the new ones be sustainable? And if stakeholders are to take a larger role inproviding useful feedback on industry practices, they need to be more meaningfully engaged in theprocess – less rhetoric and more well-informed, rational thinking and discussion is called for.

4.4 Can the directive be improved?

Directive 074 can be improved, particularly if one focuses on its objectives and removes its technicalimpediments. The following is a suggested set of objectives for DDAs, both during their operatinglife, and then at and after closure. They are taken from Hyndman and Sobkowicz (2010), and are,in the author’s opinion, equivalent to or compatible with the tailings management objectives givenin Directive 074.

Goals for the operating period of the mine preceding mine closure include, (to the greatest extentpractical):

– Reclaim tailings as mining proceeds, avoiding excessive accumulations of contained fluid finesthat must be remediated at or near the end of mine life.

– Limit the required containment volumes of MFT (in particular, in out-of-pit dam structures) tothat required for effective tailings management.

– Without compromising the essential elements of a closure landform design, conduct as muchremediation as is practical during the active mine life, when there is operating revenue to coverthe costs and while the mine organization and operating infrastructure are in place to efficientlyconduct the activity.

The following are general objectives for returning mine site lands to the public without ongoingliabilities:

– Avoid DDAs in the reclaimed landscape that require ongoing maintenance for decades followingactive mining.

– Attain landforms with geotechnical stability that are resistant to natural processes and are self-healing after natural erosion, with self-sustaining, native vegetation cover.

– Design productive, self-sustaining land and water features that are integrated into the naturalecosystem without adverse consequences to downstream watercourses.

Specific to the immediate reclamation of a DDA, tailings should with time meet several importantobjectives:

– They should develop strength at a rate sufficient to allow timely capping, in order to meetreclamation and closure requirements.

– They should develop a low compressibility so as to minimize post-closure settlement and notdisrupt the closure landscape.

The author notes that while most of the tailings management objectives given in Directive 074are reasonable and achievable, some of them require tempering against operation requirementsand the overall objective of maximizing bitumen recovery. One is the objective on maximizingintermediate process water recycling, which may be impacted by as yet poorly understood waterchemistry impacts on bitumen extraction efficiency. Another is the objective to eliminate or reducecontainment of fluid tailings in out-of-pit ponds during operations, which is incompatible withsound tailings planning and water management practices, particularly during the first 10 to 15 years

26

of a new mine’s life. These comments regarding the objectives could be cleared up by relativelyminor wording changes that recognize practical mining constraints.

4.4.1 An alternative to the 5 kPa strength criteriaIn the author’s opinion, the strength criteria contained in Directive 074 are too restrictive, giventhe number of technical options being considered and under development by oil sands operators tomeet the Directive objectives. A better approach would be to replace the existing strength criteriaby a detailed consideration of the following (Sobkowicz and Morgenstern, 2009):

– What are the anticipated properties of the tailings in the DDA?– How will they be capped?– What will be the schedule for capping?– What are the time-strength trajectory and the associated subsidence of the tailings deposit, and

how is this addressed in the reclamation plan?

This approach is less “formulaic” than the one advocated in Directive 074, and has the flexibilityto adapt tailings management plans to the specifics of each DDA. Application of the approachrequires a proper understanding of the different types of oil sands tailings materials and the appro-priate reclamation strategy that can be used for each. This is discussed further in Morgenstern andSobkowicz, (2010).

In adopting the approach advocated above, one must recognize that trajectory and demonstrationof behaviour is more important than meeting artificial goal posts. For example, what better meetsreclamation objectives – a DDA that is capped soon after completion and demonstratively improves(following a predicted trajectory) over a period of say 10 or 15 years, to a point where a closurelandscape can be constructed, or a DDA that meets all of the Directive 074 criteria but neverthelesstakes 50 years or 100 years to reach the same point? The latter scenario is quite possible and reflectsthe dangers associated with setting the wrong goal posts. Recognizing as well that “consolidationover time” is usually less costly than forcing accelerated dewatering of soft tailings, one shouldquestion the economic values associated with too rigid a set of tailings management criteria.

4.4.2 Improved performance measuresThe ERCB has chosen as their primary measurement of DDA performance the undrained shearstrength of the deposit in various locations and at various times after deposition. This might seemlike a preferable approach as it directly measures the property that is perceived to be the mostrelevant to the performance of the deposit. However, there are a number of issues associated withthis approach (Morgenstern and Sobkowicz, 2010):

– Inaccuracy and unreliability of measuring low strength values, (which would be further exac-erbated in the case where one accepts the use of reclamation strategies at very low surfacestrengths).

– Applicability of undrained shear strength measurements in granular materials.– Difficult access for, and high cost of, the moderately large equipment needed to measure strength.– Needless restriction of reclamation to terrestrial-based methods (other methods do not require

the same surface conditions).

There are other types of measurements that can be used to monitor the progress of consolidation/densification in the various DDA deposits, to predict the performance of the deposits at any time,and to assess the readiness of the deposit to accept any particular reclamation strategy. One type ofmeasurement that avoids the disadvantages described above is sampling of the deposit combinedwith the measurement of solids content and related material index tests, (such as Atterberg Limits).This method allows:

– Easier access (for lighter equipment) and less costly monitoring campaigns.– Accurate results.– Clear indication of improving material state (consolidation in the fines-dominated materials and

densification in the granular materials).– Direct measurement of, or correlation to, desired geotechnical characteristics of the deposit.– A basis for projecting long term behaviour, e.g., to forecast the subsidence of the tailings deposit.

27

5 WHERE ARE WE HEADED?

5.1 Promising tailings treatment technologies

The tailings treatment technologies described in Section 3, which historically have been, or currentlyare, under consideration and development by industry operators, run the full gamut of identifiedpossible technologies (Sobkowicz and Morgenstern, 2009). The ones that hold the greatest promisefor meeting Directive 074 objectives may be grouped as follows:

– Methods that are currently in commercial operation (although also still being improved). Thiswould include CT, TT, and at Suncor, in-line flocculation and thin-lift dewatering of MFT.

– Methods that have seen intensive research and development, and are almost ready for commercialimplementation. This would include in-line flocculation and thin-lift dewatering at all mine sitesexcept Suncor, centrifugation (combined with some form of environmental assist or acceleratedconsolidation), and MFT spiking of WT.

– Methods that have had comprehensive research and development, and while not quite ready forcommercial implementation, show great promise. This would include freeze-thaw consolidationof fine tailings.

It is likely that any particular operating company will have to adopt several tailings treatmentmethods to address all of the fines tailings issues on its lease, as discussed in Sobkowicz andMorgenstern (2009).

In addition to the treatment methods listed above, there is active research directed at improvingexisting oil sand capping technologies, adapting capping technologies for soft deposits from otherindustries, and developing new capping technologies.

5.2 The shift in storage/treatment focus from tailings to water

One important (but somewhat subtle) point to note is that the more effective operators become withdewatering oil sand tailings, the more water will be released from those tailings. That will providegreater opportunities for water recycle, with perhaps attendant water chemistry issues (e.g. impacton bitumen extraction). However, it will also shift waste storage challenges from tailings to water,and in some cases force the need for water treatment and release. This is a separate, complex issuethat only bears mentioning in this paper.

5.3 What is on the horizon?

There are other tailings treatment technologies that have received research attention but are not yetsufficiently advanced, in the author’s opinion, to be considered “promising”. This does not detractfrom the interest shown in them or from the importance of continuing their research efforts, butis only a statement of how close they might be to implementation. At present, they are “legitimatepossibilities”. These include, listed in order of least to most advanced:

– Electrophoresis.– The use of plants and plant growth to form stabilized crusts on soft tailings deposits.– Rim ditching.– Some kind of tailings filtration (the current favourite being cross-flow filtration).– Wick drains (to enhance soft tailings consolidation).– End-of-pit lakes.

The oil sand industry has great hopes for end-of-pit lakes because they will have a tremendousimpact on the overall cost of tailings storage. They are included here as a possible rather thanpromising technology because they have not yet been demonstrated. Even if the science proves thatthey are technically sound and effective, there are still public perception hurdles to be crossed withvarious stakeholders, before they can be implemented on the scale currently envisaged.

Other tailings treatment methods have been proposed, several of them involving the use ofproprietary chemical additives, but insufficient research has been done (to date) to give a sense oftheir potential. No doubt these and other yet to be conceived methods will soon make it onto the“possible” and then the “promising” list.

28

5.4 Is there a perfect tailings management practice?

As a last comment on future tailings management practices, the author states his opinion that themost cost-effective and technically effective solutions will involve the separation in a pragmaticway of coarse tailings from fine tailings. With this approach, the higher cost of dealing with thefine tailings is offset by the lower cost of storing the coarse tailings, and the resulting products arethe most efficient in terms of total required storage space and early release of water.

Past methods of constructing tailings ponds, which efficiently stored sand and captured fines,and which also efficiently collect MFT, are a robust and sensible tailings management practice. Asreasonable methods of dewatering MFT are demonstrated at the commercial scale (Section 5.1),the combination of conventional tailings ponds and MFT dewatering may prove to be superior tomethods that rely on combining sand and fines (such as CT and TT). If this proves to be the case,there are also implications for how long specific tailings ponds are operated. In any case, the juryis still out, and your crystal ball may be much better than mine!

6 CONCLUSIONS

Oil sand mine operators have encountered many challenges in storing and treating fine tailingssince commercial mining started in 1967. A remarkably broad research and development effortover many years has addressed these challenges, finding solutions such as enhanced capture offines in beaches, non-segregating tailings and thickened tailings – all intended to improve finescapture and reduce fluid fine tailings inventories. Recently developed treatment methods, such asin-line thickening of tailings combined with thin-lift dewatering or with centrifuging, or freeze-thawconsolidation of tailings, are in advanced stages of pilot testing.

Industry efforts to find effective and economically responsible solutions to the fine tailingschallenge have been focused by the release of the ERCB Tailings Directive 074 in 2009. This is auseful and timely regulation that has had major impacts on the oil sand industry (some beneficial,some not so). Suggestions are given herein for improvements to the Directive, to make it moresuited to the realities of oil sand mining and tailings operations.

A number of “promising” and “possible” tailings treatment methods have been discussed. Pre-dictions are given for which ones will “win” and which will be incorporated into a better tailingsmanagement practice. Time will tell how clear our crystal ball is.

7 ACRONYMS

The following acronyms have been used for technical terms in this paper. Other acronyms fororganizations and locations have been defined where they were first used.

BT – Beached TailingsBAW – Tailings beached sub-aeriallyBBW – Tailings beached sub-aqueouslyCHWE – Clark Hot Water Extraction ProcessCT – “Consolidated” TailingsCUT – Cyclone Underflow TailingsCOT – Cyclone Overflow TailingsDDA – Dedicated Disposal AreaLEE – Low Energy Extraction ProcessMFT – Mature Fine TailingsNST – Non-segregated TailingsOHWE – Oslo Hot Water Extraction ProcessPD – Positive Displacement PumpingPSV – Primary Separation VesselTFT – Thin Fine TailingsTT – Thickened TailingsWT – Whole Tailings (from extraction plant)

29

ACKNOWLEDGMENTS

The author expresses sincere thanks to the following “old-timers” who provided valuable informa-tion and insight when compiling the history of tailings research and development: Al Hyndman,Jonathan Matthews, Nordie Morgenstern and Sean Wells. Bill Shaw, who has been at the centerof most of the research and development work at Syncrude, expended effort far beyond the callof duty, and deserves special thanks. Gord McKenna, on short notice, provided a very insightfulreview of the paper; his comments and thoughts are greatly appreciated.

REFERENCES

Energy Resources Conservation Board 2009. Directive 074: Tailings Performance Criteria and Requirementsfor Oil Sands Mining Schemes. Province of Alberta.

Hyndman, A. & Sobkowicz, J.C. Oil SandTailings: Reclamation Goals & the State of Technology. In Geo2010;Proc. 63rd Canadian Geotechnical Conf., Calgary, Canada, 12–16 September 2010. Canadian GeotechnicalSociety.

Morgenstern, N.R. 2010. Improving the Safety of Mine Waste Impoundments. In Tailings and Mine Waste ‘10;Proc. Intern. Conf., Vail, Colorado, 17–20 October 2010. Colorado State University.

Morgenstern, N.R. & Sobkowicz, J.C. 2010. Reclamation and Closure of an Oil Sand Tailings Facility. InInternational Oil Sands Tailings Conference 2010; Proc. 2nd Intern. Conf., Edmonton, Canada, 5–8December 2010. OSTRF/CONRAD, University of Alberta.

Sobkowicz, J.C. & Morgenstern, N.R. 2009. A Geotechnical Perspective on Oil Sand Tailings. In Sego,Alostaz & Beir (eds.), Tailings and Mine Waste ‘09; Proc. Intern. Conf., Banff, Canada, 1–4 November2009. University of Alberta Geotechnical Center/OSTRF.

Wells, P.S. & Riley, D.A. 2007. MFT Drying – Case Study in the Use of Rheological Modification andDewatering of Fine Tailings Through Thin Lift Deposition in the Oil Sands of Alberta. In Fourier & Jewell(eds.), Paste 2007; Proc. 10th Intern. Seminar on Paste and Thickened Tailings, Perth, Australia, 2007.ACG, University of Western Australia.

Wells, P.S. 2010. Oil Sands Pond Closure – Sand, Sun and Soft Tailings. In Mine Closure 2010; Proc. Intern.Conf., Vina del Mar, Chile, 23–26 November 2010. ACG, University of Western Australia.

Wells, P.S., Caldwell, J. & Fournier, R. 2010. Suncor Pond 5 Cap: The Story of its Conception, Testing, andAdvance to Full-Scale Construction. In Tailings and Mine Waste ‘10; Proc. Intern. Conf., Vail, Colorado,17–20 October 2010. Colorado State University.

30

Mill tailings


Tailings impoundment failures, black swans, incident avoidance,and checklists

J. Caldwell & L. CharleboisRobertson GeoConsultants, Vancouver, BC, Canada

ABSTRACT: The thesis of this paper is that tailings impoundments fail as a result of a string ofincidents, each of which is trivial and within the bounds of normal events, but which, taken together,constitute an event so unusual that it lies outside of the bound of normal occurrence and experience.The string of incidents leading to the failure of a tailings impoundment may be understood andevaluated in the light of the theory of the Black Swan, an event that nobody could have foreseen,that results in extreme consequences, and which can be explained after its occurrence by all onthe basis of standard knowledge. In this paper we examine current theories and hence methods foravoiding failure of tailings impoundments. We find them all lacking, and so we proceed to set outproposed approaches based on incident control, checklists, and Black Swan avoidance to limit andhopefully eliminate the possibility of failure of tailings dams and the consequent loss of life andproperty.

1 INTRODUCTION

This paper is about the philosophy of slimes dam failures.You can call them tailings impoundments,processed material containment facilities, storage locations, or mine geowaste areas. The factremains they are dams that contain slime—thus we prefer the name we grew up with and which westill think is accurate and descriptive.

You may blame the senior author for the opinions in this paper. You must thank the junior authorfor checking the facts and having the courage to be associated with our ideas.

We seek to get to the bottom of an every-pressing issue: why do slimes dams fail? Someattribute failure to engineering issues; some attribute failure to institutional practices; some blamethe designer or the mine. A favorite reason is Acts-of-God, most often extreme precipitation. Thebest theory we know of attributes accidents to a failure to control incidents; the idea is that tenunattended incidents equal one accident; ten accidents equal one fatality. We believe this is the rootcause; and if not the cause, at least the best way to proceed to eliminate the failure of slimes damsand the attendant deaths.

We are fascinated by the theory of the so-called Black Swan; the idea that some things are sounusual that nobody can foresee their occurrence, although after the event everybody can explainwhy the event occurred. Thus we examine the role of Black Swans in causing slimes dam failuresand suggest practical ways to hunt and kill Black Swans before they come to kill you by failingyour slimes dams.

2 THE BLACK SWAN

After the economic collapse in late 2008, many sought the fundamental reasons for the economiccollapse. Before the collapse, nobody saw it coming—although some may claim they did, wehave seen no convincing evidence of anybody predicting its occurrence. The event had an extremeimpact; economic edifices came crashing down and many lost their houses, their investments, andtheir jobs.

33

After the collapse, we find it very simple to explain why it happened: the collapse is the resultof the granting of mortgages to people who did not have the ability to pay. We all know somebodywho kept buying a bigger and bigger house costing many multiples of their annual salary. We allknow people who kept taking bigger and bigger loans against their house for foreign travel, a newcar, more clothes, or just more plastic or exercise equipment to clutter the garage.

Some ascribe the collapse to the imprudence of banks and finance houses who bundled badmortgages into packages to be sold to distant investors rendered incautious by near-dishonestratings firms. Behind it all there appears to have been a battalion of overly trained mathematiciansarmed with flawed theories of statistics; specifically statistics that said that there was so vanishinglyslim a chance of the mortgages going sour that reasonable people need not be concerned. This wastopped by inattentive managers who seemingly did not know or understand what their underlingswere doing

In the book, The Black Swan – The Impact of the Highly Improbable by Nassim Nicholas Taleb,the author notes three characteristics of a Black Swan:

• It is an outlier, as it lies outside the realm of regular expectation, because nothing in the past canconvincingly point to its possibility.

• It carries an extreme impact.• In spite of its outlier status, human nature makes us concoct explanations for its occurrence after

the fact, making it explainable and predictable.

Let us explore the thesis that tailings impoundment failures are generally Black Swan events inthat nothing in the past convincingly points to their occurrence, that they have an extreme impact,and that after the failure, human natures makes the even explainable and predictable.

3 BAFOKENG

We have previously written about the failure in 1974 of the Impala Mine, Bafokeng slimes damthat killed thirteen and cost millions to clean up. This is what we wrote:

The mine, like all mines in the area, was perpetually short of water, so they stored as muchwater as possible on the top of the impoundment. The day of the failure, the pool was very close toand some say lapping up against the outer dike thrown up to make a place for the discharge pipesand the next lift of tailings discharge. Then it rained. The bulldozer driver was sent to shore up avulnerable-looking part of the outer dike. Who knows: maybe he vibrated the wet tailings and theyliquefied; maybe he dug too deep or too inexpertly with his bucket as he struggled in the rain todo something unfamiliar and he just took away the freeboard; maybe some profound geotechnicaloccurrence happened deep in the tailings. Regardless, the water and liquid tailings flowed out,flowed far, and killed miners.

Geoff Blight in his new book writes in detail about possible failure causes. He is forced toconclude: “It appears at first sight that the dyke did not fail by conventional overtopping. Eyewitnessaccounts all point to a failure by piping erosion. However, a satisfactory explanation of how theinitial hole formed in the wall was never reached.”

Professor Jennings, for whom I worked to collect the data to evaluate the failure, was convincedthat piping began between two layers of low permeability slime bordering a zone of higher per-meability sand tailings. Such zones existed, as I saw too often climbing over the failure zone tomeasure things.

Personally I think it is easy: to pinpoint the root cause of the failure of the Bafokeng dam. Nobodycould or did foresee the Black Swan. Those who ran the mine and who operated the impoundmentsaw no reason for concern. They had no concept that the dam could or would fail—nothing in thepast pointed to such a failure, and nothing pointed convincingly to the possibility of such a failure.But the failure occurred, and the failure had great impact.

Now it is clear: the pool was too close to the perimeter dikes, it was raining hard, there wasseepage flow in saturated sand layers between clay layers, and the bulldozer operator inducedliquefaction in the confined sand layer. Thirteen died. Now we know we must avoid all thesefactors.

In the years following the failure, the senior author evaluated two more platinum slimes damsthat failed. The causes of failure, as at Bafokeng, included a pool too close to the perimeter dikes

34

and high pore pressures in the outer shell. In addition, it was clear that there was slope failure causedby sliding along the weak clays that are found throughout the mining district. If you consider thatat that time, there were no more than about ten platinum tailings impoundments in the area, theprobability of failure works out to almost 33 percent.

Some changes were made to the standards of practice in the wake of these three failures, butthey were insufficient to preclude the Merrispruit failure which we now consider.

4 MERRIESPRUIT

Geoff Blight has this to say about the failure of the Merrispruit slimes dam in 1994;“On the night of February 22, 1994, a 31-m high dyke upslope of the village of Merrispruit,

South Africa, failed with disastrous consequences. The dyke breached a few hours after 30 to55 mm of rain fell in approximately 30 minutes during a late afternoon thunder storm. The failureresulted in some 600,000 cubic meters of liquefied tailings flowing through the town causing thedeath of 17 people.”

Wikipedia records this—and it sound like the person writing knew what actually happened;“In March 1993 an inspection noticed seepage along the north wall and it was agreed to stop

deposition into compartment 4A. According to the contractor, the freeboard at this time was anacceptable 1.0+ m. The division of compartments 4A and 4B was breached some time before thedisaster, resulting in drainage from 4B to 4A. The extra drainage led to a freeboard of 300 mm.Despite the termination of daywall construction, excess plant water containing tailings continuedto be deposited, with the water decanted by the penstock and the remaining tailings using up theremaining freeboard.”

One of the authors has chatted informally with some of those who were involved in examiningthe area after the failure. They all acknowledged that the mining company had cut to the bone toreduce costs and as a result the dam was neglected, competent people were not involved, and thecontractor’s staff were overly confident (to the point of negligence) of their abilities. Undoubtedlythere was institutional inertia and over-confidence based on past successes—or at least an absenceof past failure of similar structures in the area. Regulations were in place, but they were inadequate,not followed, and not enforced. Everybody involved, from top to bottom of the chain of commandand responsibility, failed to foresee the event which we must conclude lay outside the realm ofregular expectation, because nothing in the past convincingly pointed to its possibility.

The failure had an extreme impact. In spite of its outlier status, soon after the event almosteverybody involved was able to concoct explanations for its occurrence and render the failureexplainable and predictable. As Blight points out: “Everybody was to blame and everybody blamedsomebody else and nobody was held responsible. And the man on the ground at the slimes dam didnot conceive of another big storm and what it could, and did do.”

5 INCIDENT CONTROL

On the basis of but these two spectacular slimes dam failures, we submit that they were indeedmanifestations of the arrival of a Black Swan. We believe the Black Swan was able to come andwreak havoc because of the failure by too many people to control the little incidents that if attendedto would have blocked the path of an oncoming bird of doom.

We need hardly expound on the theory of incident control here. It is well documented in theliterature and in countless sites on the web. Suffice to say that the essence of incident control isthat if you control the little things, the big things do not happen. As we said in the introduction, tenunattended incidents equal one accident; ten accidents equal one fatality.

One story to illustrate. Many years ago the senior author was digging and profiling test pits inthe cover of a steep (1.4 horizontal to 1.0 vertical) and high (100 m) side of a landfill perched abovethe I60 Freeway leading east from Los Angeles. A clod from the pile of soil dug to make the testpit rolled down the slope, jumped the fence, and hit a passing car, causing a big mud splat on thedoor. The driver continued home—we never heard from him.

The company for which the senior author worked had in place an Incident Control Program. Sohe reported the incident. The result was a long and high metal and plywood safety fence along the

35

perimeter of the landfill. We proceeded to strip the cover and replace it with a geogrid reinforcedcover. We did over $100 million of work and there were no serious injuries and no fatalities. Ibelieve the Incident Control Program is to be credited with this success.

We are not aware of any mine that has in place a Tailings Storage Facility Incident ControlProgram. No wonder failures occur and will occur.

6 FAILURE CAUSES

The literature and the web are replete with records of slimes dam failures. Common postulatedtechnical causes of failure include: earthquakes, foundation and/or slope failure, overtopping duringheavy rains, washout pursuant to pipe failures, and piping resulting from the pool being too closeto the perimeter dikes.

Some proceed to blame non-technical factors for slimes dam failures. They say root causesinclude:

• Institutional, including cultural and stakeholder attitudes and practices that fail to be aware orconcerned about the presence and/or condition of the tailings impoundment.

• Management, including the absence of knowledgeable managers and/or the failure of those withthe power and responsibility to do so to act.

• Regulatory, including an absence of or inadequate regulations, and/or a failure of those chargedwith doing so to enforce existing regulations.

We submit these are but sources of incidents. We concur that, as at Merrispruit and Bafokeng,there were indeed serious and systematic institutional, management, and regulatory lapses. Eachand every one gave rise to an “incident” that if eliminated would have blocked the path of the deadlyBlack Swan.

7 OTHER POSTULATED REASONS FOR FAILURES

An early paper by Edwin Smith (1973) notes that “according to mining folklore, no tailings dam hasever been completed without at least one failure occurring during the deposition of the particles.”He lists these causes of failure: foundation failure; slope failure; overtopping by flood waters;erosion of face; piping; collapse of dewatering conduit; and liquefaction. He recommends theapplication of the Observational Method as the best way to build tailings impoundments and keepthem safe.

Over the years, papers in previous Tailings and Mine Waste Conferences have sought to pinpointthe cause of failure. The most dramatic assignation of responsibility is in the 2003 proceedingswhere Allen Gipson blames “owners, engineers, designers, and operators [who] are not performingtheir work in accordance with the standards of practice that should be followed.” There is not muchmore to say in the face of such an assertion.

Steven G. Vick (2002) also examines reasons for failure of mine geowaste facilities. In hismagnificent book he explores the role of subjective probability and engineering judgment. Hisbook is so profound that we do no more than note it here and maintain that anybody charged withkeeping mine geowaste facilities safe must read it.

Michael Davies and Todd Martin (2009) present a tantalizing institutional reason for the failureof tailings dams. They examined the timing of cycles of boom and bust in the mining industry andthe timing of failures. They conclude that between one and two years after a period of depressedmining activity (poor economic conditions), there is likely to be a slimes dam failure. They ascribethis to the possibility that during poor economic times, mine management cuts back on the costsof managing and operating the tailings facilities and inevitably this leads to a failure. They providesome plausible reasons why the tailings dam could fail after the economic boom:

• Permit haste• Fast track investigation, design, and construction• Cost cutting after the boom• Inexperienced but overconfident designers

36

• Lack of independent third party peer review• Rapid turnover of mine personnel• Disconnect between design expectations and operational reality• Development of deposits that have been left undeveloped for good reason• “Cookie cutter” designs.

We are not aware of any evidence that the managers ultimately responsible for the failure of theBafokeng dam had cut back in any way preceding the failure. In fact, the construction companyoperating the dam on behalf of the mine had, some time before the failure, employed their firstqualified civil engineer in an attempt to improve the standard of their service to the mining industry.

In the case of the Merrispruit failure, there are hints that management had cut back on provisionof services related to operation and oversight of the dam. But the slimes dam had always beenoperated that way and there was no valid reason to believe it would act any differently in the future.

8 HOW TO AVOID FAILURES OR AVOIDING THE BLACK SWAN

Let us proceed to examine what can be done to stop the Black Swan from coming to induce failureof slimes dams. Of course the easy answer is: control incidents. We recognize, however, that thisimplies positive action. Thus we take a more detailed look at what individuals, companies, andsociety as a whole can do to augment a good Incident Control Program.

Taleb says this about incorporating Black Swan thinking into your life: “I am very aggressivewhen I can gain exposure to positive Black Swans—when a failure would be of small moment—andvery conservative when I am under threat from a negative Black Swan. I am aggressive when anerror in a model can benefit me, and paranoid when an error can hurt. This may not be interestingexcept that it is exactly what other people do not do. In finance, for instance, people use flimsytheories to manage their risks, and put wild ideas under “rational” scrutiny.”

Clearly the individual charged with some aspect of a slimes dam can hunt the Black Swan byacting as Taleb does. But this takes a bold, confident, and knowledgeable professional. Too often,cultural, societal, and even professional practice precludes prudent, individual action that amountsto Black Swan hunting. In the case of a heap leach pad failure of which we are aware, professionalindividuals were pressured to perform fast, were greedy to profit by fast performance, and fell intothe trap of relying on statements by others instead of undertaking the evaluations themselves. Theywere too proud to consult with peers or submit to review. Now the lawyers are on their tail and thecountry is saddled with a mining mess.

Many, but by no means all, mining companies have taken many steps to prevent a recurrence ofpast tailings dam failures. These include:

• Employ experienced professional staff. This is a problem in most times as the perpetual callingby head-hunters attests.

• Engage reputable consultants. Most consultants are reputable, but most are also susceptible tolapses if improperly managed and controlled.

• Demand conservative engineering. This is good to do, but nearly impossible if the accountantshave any sway and the project manager is profit and bonus motivated.

• Inspections by regulators. As we write, the US eastern coal mining industry is under a cloud asa result of a mine accident that killed nearly thirty miners. It appears as thought the regulationswere in place, the inspections were being made by the regulators, and the technology existed topreclude the accident. The sad fact, however, is that the mine in question had been inspectedthe very morning of the accident. The inspector, the miners, and management failed to conceiveof the accident that did occur. It was outside of their field of imagining, although once all thecauses are established, it will seem obvious that they should have known.

• More conservative engineering designs. We note our perennial favorite, namely the UMTRAProject where we had to design in accordance with U.S. Federal law for a period of stability of1,000 years. Thus we designed for the maximum probable precipitation, the maximum credibleearthquake, and so on. It can be done as a way to avoid both short-term and long-term BlackSwans, but how many societies have the courage or check-book to demand this of their mines?

37

9 SUGGESTED APPROACH

We suggest the following as the set of actions that should be undertaken if you are an individual, aconsultant, or a mining company charged with the safety of a tailings impoundment and you wishto go Black Swan hunting.

• Use the Wisdom of the Group. There are many ways to capture the wisdom of the group—asgood a way as we know of to identify potential Black Swans. Some people call the process riskassessment, but there are many ways of facilitating wise thinking, including: Risk Assessment;Failure Modes and Effect Analysis; Value Engineering; and Multi Accounts Analysis. The seniorauthor personally prefers the FMEA approach although he has participated in many other sessionsadopting variants, and they mostly worked.

• Prepare a Risk Assessment Report and Monitoring Plan. The overall objective of using the wis-dom of the group is to compile a RiskAssessment that is used as the basis of a Monitoring Plan. Inother words for each identified risk of malfunction or failure, put in place an observation routine(visual and/or instruments) so that you get early warning of potential impending malfunctionand/or failure. Then compile, as we note below, an Observational Method Plan that providespredetermined courses of action as practical responses to observed (monitored) performance ordeviation from anticipated performance.

• Compile an Observational Method Plan. The Observational Method is well-known in geotech-nical engineering, so we say no more about the method here other than that it forces you toestablish logical monitoring and observation routines, to identify what may go wrong, and toestablish before you start what you are going to do if things start to go wrong. The ObservationalMethod, correctly applied, is no more than the construction of look-out forts, their consistentmanning, keeping a look out for an oncoming Black Swan, and the preparation of an arsenal ofweapons to slay the swan as it glides to your project and discredit.

• Implement an Incident Control Program. We know of no tailings facility that is part of an IncidentControl Program. It is a nuisance and sometimes offends the powers that be. Yet we submit thata comprehensive Incident Control Program will nip in the bud most things that have ultimatelylead to slimes dam failures.

• Compile Checklists of what to do when designing, constructing and operating a tailings facility,and focus on what to do when things start going wrong. Note the book by Atul Gawande whoproposes that proper use of checklists can improve the practice of medicine and the safety ofairplanes. There are many checklists for tailings impoundments built into the many documentsput out by national organizations. One that violates all the recommendations for a good checklistis from the Mining Association of Canada (1998). We suspect none of them takes a “kill theBlack Swan” perspective. We submit that all are probably too unconservative. In particular, avoidany that have sustainable in the title, for by definition this means they avoid the truth, or havebeen written by consultants trying to make it easier for cash-strapped clients. They obviouslyneed to be re-written to be more outlier-event averse, but that is no good reason not to start now.

• Ensure Regular Peer Review. Peer reviewers are easy to fool and mislead. Peer reviewers areas susceptible as any group to herd-thinking, and blindly following the lead of one dominantindividual who does not fall asleep in peer review meetings. Nevertheless, if carefully chosenthey are at least independent, force the designers and operators to prepare presentations andargue their case (probably the greatest benefit of the whole process), and they may just noticesomething management is too busy to attend to.

10 CONCLUDING THOUGHTS

In this paper, we reject simplistic explanations of slimes dam failures predicated on economic cycles,heavy rains, piping, management inattention, regulator incompetence, and so on. We submit thatevery slimes dam failure can be traced to a string of events (positive incidents or failure-to-actincidents), that in concert or in sequence are the cause of the failure and the attendant deaths.

We submit that the immediate causes of most slimes dam failures are such that the failurecould have been predicted and/or precluded by application of standard engineering practices andtechniques of the time. It would be fair to say that at the time of the Bafokeng failure, the engineering

38

knowledge and techniques were in place to predict failure, if an engineering examination had beenundertaken. Without doubt, at the time of the Merrispruit failure, the engineering knowledgeexisted; had it had been brought to bear, failure would not have occurred.

Thus it is not technology or a failure of technology that kills. Lives are lost when a string ofincidents occurs, each in itself relatively trivial, but in concert, deadly. The answer is simple: controlthe incidents, technical, engineering, and institutional. It is the graffiti theory: control the graffitiand you reduce crime.

In this paper, we recommend that for every slimes dam/tailings facility/processed materialcontainment area there should be:

• Regular facilitated wide person deliberations to think things through.• Regular peer review to put the spotlight on practices, good and bad.• A comprehensive Risk Assessment to establish the things that could go wrong and to form the

basis of the instrumentation and monitoring plan, to be implemented in conjunction with anObservational Method Plan.

• An Observational Method Plan, to guide the monitoring and associated actions.• An Incident Management Plan, to enable you to deal proactively with the little things that if left

unattended will combine to enable the Black Swan to triumph.• Simple Checklists on what to do when things start to go wrong.

REFERENCES

Blight, G. “Geotechnical Engineering for Mine Waste Storage Facilities” CRC, 2010.Blight. G. and Fourie, A. “A review of catastrophic flow failure of deposits of mine waste and municipal refuse.

At this link: http://ww.unina2.it/flows2003/flows2003/articoli/G.E.%20BLIGHT%20&%20A.B.%20FOURIE.pdf

FMEA Info Centre. At this link http://www.fmeainfocentre.com/Gawande, Atul “The Checklist Manifesto.” At this link http://gawande.com/the-checklist-manifestoGipson, A. H. (2003) “Tailings dam failures – the human factor.” Tailings and Mine Waste ’03, page 451.

Balkema.Michael Davis, Todd Martin, and Peter Lighthall. “Mine Tailings Dams; When Things Go Wrong.” At this link

http://www.infomine.com/publications/docs/Davies2002d.pdf, The Mining Association of Canada (1998)“A Guide to the Management of Tailngs Facilities.” www.mining.ca

Smith, E.S. (1973) “Tailings Disposal—Failure and Lessons.” Tailings Disposal Today Miller FreemanPublications. Page 356.

Tailings.info at this link http://www.tailings.info/accidents.htmTaleb, Nassim Nicholas “The Black Swan – The Impact of the Highly Improbable” Penguin Books 2007.Vick, Steven G. (2002) “Degrees of Belief – Subjective Probability and Engineering Judgment.” ASCE Press.

39


New directions in tailings management

Clint StrachanMWHAmericas Inc., Fort Collins, Colorado, USA

Jack CaldwellRobertson GeoConsultants, Vancouver, BC, Canada

ABSTRACT: Two MillionTonnes a Day –A MiningWaste Primer issued by MiningWatch Canadain December 2009 sets out conceivable ills associated with current and possible new tailings disposalmethods. In this paper, we set out to examine current and proposed methods of tailings disposaland management.

While there have been failures of tailings impoundments, it is possible to design, operate, andclose these facilities safely, and in ways that protect the environment. This paper will establish thatthe issues raised in MiningWatch and similar documents posted on other web sites that attack miningand tailings disposal practices are founded on a lack of technical knowledge; failure to collect,collate, and understand the facts; and a desire to make statements derived from pre-establishedprejudices and perspectives.

This paper summarizes the body of knowledge, new technologies, and practical experience, alongwith case histories that substantiate the fact that tailings disposal can be managed in compliance withinternational guidelines and standards; local regulations and requirements; in an environmentallyresponsible manner; and in a manner that provides employment.

1 BRIEF HISTORY OF TAILINGS DISPOSAL

1.1 Initial operations

Tailings disposal was initially a technique of trial and error. In the early 1900s in SouthAfrica, FraserF Alexander pioneered practical and inexpensive methods to construct tailings impoundmentsentirely with tailings (slimes dams). He and those who succeeded him, have a long history ofsuccessful impoundment construction and operation. However, this method in South Africa hashad failures. For example, failures of the Bafokeng and Merrispruit slimes dams led to awarenessof the consequences of failures, and changes in the laws.

Similar development in the knowledge and techniques for tailings disposal and impoundmentmonitoring occurred in Canada and the United States. Construction of the first embankments withtailings was by trial and error. The 1960s and 1970s included the beginning of the use of soilmechanics to assess tailings behavior and tailings impoundment stability. In the mid 1970s we findthe first technical papers by civil and geotechnical engineers on tailings impoundments in NorthAmerica. It is interesting to return to these papers and see the use of fundamental principles indesign and operation of tailings impoundments, and to see the first words on vegetated covers aspart of facility closure.

1.2 Use of filtration

In the early 1980s both authors were involved in the Greens Creek Mine in Alaska. This was oneof the first mines to filter their tailings, truck them to a disposal facility, and compact the tailingsas a solid material. This approach was dictated by the wet climate and seismic risk of the area, aswell as limitations in impoundment area. But it showed that filtered tailings disposal is practicalunder certain conditions, and is still used at the mine. Filtered and other dewatering treatments of

41

tailings have been in adopted elsewhere in recent years where impoundment space is limited orwater reuse must be optimized.

1.3 Earth and rockfill tailings embankments

In the 1980s, the authors were involved with the Cannon Mine in Washington, which consisted ofa rockfill embankment for containment of tailings, and the McLaughlin Mine in California, whichconsisted of an earthfill dam for containment of tailings. Both embankments were designed andconstructed to applicable standards of dam design practice. Other rockfill and earth embankmentshave been constructed in a similar fashion.As demonstrated by the successful closure and productiveongoing use of both sites, mines close to communities and upstream from key water resources canbe operated and closed successfully.

1.4 Impoundment closure

The authors were also involved with the closure of uranium mill tailings impoundments across theUnited States. As established by over twenty years of observation, these reclaimed impoundmentsare performing as expected (in terms of impoundment stability and cover performance). The per-formance criteria for these facilities are to be stable for the long-term period of performance (200to 1,000 years).

Due to the volumes of materials involved and the cost and time associated with earthmovingand water management, tailings impoundment closure is not a simple or inexpensive exercise. Thelesson learned from the uranium mill tailings impoundment closures is that the technology andpractice exists to achieve safe and stable long-term closure. New tailings impoundments must beassessed for feasibility and designed and constructed with closure (and its associated time andcost requirements) as part of the evaluation. Although “design for closure” has been discussed intailings impoundment design for over 20 years, this concept still needs to be stressed.

Some mines are using or planning to filter their tailings and so dispose of them as solid or “drystack” materials. Mines are also thickening and polymer-amending their tailings in order to creatematerials that can be used for mine backfill or enhance the settling characteristics of dischargedtailings to hasten water recycling and tailings consolidation.

The case histories outlined above demonstrate that safe and effective tailings management isfeasible and practical – if the value of the ore body can accommodate the associated tailingsmanagement costs.

2 IMPOUNDMENT FAILURES AND LESSONS

2.1 Information on failures

Although there are significant amounts of publicity and documents on tailings impoundmentfailures, the important factors for engineers are the underlying causes of failures and how thesecan be prevented. Documents that have examined reported tailings impoundment incidents (fromevents requiring repair or mitigation to major failures) include USCOLD (1994), UNEP (1996), andICOLD (2001). The major factors that can be drawn from this body of information are listed below.

1. The majority of reported incidents were during the period when larger impoundments werebeing constructed and soil mechanics theory was starting to be applied to tailings. The body ofinformation does not cover unreported incidents prior to this period.

2. The majority of reported incidents were for smaller impoundments (with embankment heightsless than 30 meters).

3. The majority of reported incidents were associated with embankments constructed with or overtailings using the upstream method (described in Vick, 1983; ANCOLD, 1999).

4. Most of reported incidents were associated with improper water management (overtopping,seepage and erosion) or seismic effects (liquefaction or excessive embankment deformation).

This information indicates that these incidents could have been prevented with proper design forsite conditions, proper construction and operation, and effective operation and monitoring. While

42

the body of information described above tries to assign one specific cause to an incident, the failureof a tailings impoundment is often the result of a string of small events that combine in a uniqueway to bring about an unanticipated failure.

2.2 Lessons from failures and unanticipated results

As an industry, we need to turn our attention and practice to the failure methods effects analysis(FMEA) and other systems analysis approaches that have long been used in the nuclear powergenerating and other industries. As long as the industry fails to take broad-based, comprehensivelooks at the systems that constitute a tailings impoundment, failures will continue.

We can also simplify our approaches and emulate the way in which the airline industry achievesan enviable safety record by relying on simple checklists and duplicate oversight of every act. Asimilar approach (behavior-based safety procedures) has been successfully used in oil refineriesand other industrial settings.

In addition, we must be bold in recognizing that there are some places where it simply is notpractical to mine and construct impoundments that will be stable and endure in perpetuity. Thereare places where mining may not be practical, or the costs of site development, mining, and closuredo not justify the development. In certain areas, the ore body must be rich enough to afford theexpensive engineering works that are required to produce stable structures, maintain perpetualwater treatment, and provide containment and erosion resistance under long-term closure.

3 CODES AND GUIDELINES

3.1 Summary of guidelines and regulations

As the brief history above illustrates, it is possible to design, operate, and close tailings impound-ments that protect human health and the environment. What is takes is the mandate, effort, andcapital outlay for effective design, operation, monitoring, and closure. The mandate includesapplicable codes, guidelines, and regulations. These range from general guidelines to specificregulations.

Guidelines outline the accepted methods of tailings impoundment design, construction, oper-ation, and closure in general terms (such as ICOLD, 1982; ANCOLD, 1999). General guidelinesfor embankment stability have been outlined in Wilson and Marsal (1979), U.S. Army Corp ofEngineers (1982), ICOLD (1987), and ICOLD (1996). Guidelines for design storm events andembankment freeboard depend on the risk classification of the structure, as outlined in FEMA(2001), ICOLD (1987), and ICOLD (1992). Projects that include International Finance Corporationfinancing require compliance with their guidelines (IFC, 2007).

Land management agencies in the US have guidelines and regulations on mining. Some stateshave regulations affecting embankment stability administered by dam safety agencies. Other stateshave regulations on embankment stability and impoundment containment administered undergroundwater protection regulations (such as ADEQ, 1998; NDEP, 1989). Nevada and other statesrequire a closure plan and bond before construction of mine facilities.

Projects outside the US have varied regulatory requirements, ranging from specific regulationson facilities (typically in countries where there has been a mining history), to water protectionstandards or water use laws. Where there are not clear regulations, most international miningcompanies adopt North American or corporate standards or policies.

The guidelines and regulations are effective as long as there is an appropriate regulatory agencyto administer the regulations in a fair manner (without political or external pressures) or a reviewboard to check compliance with guidelines and follow-up of recommendations. These guidelinesand regulations are useful only if they are followed to achieve the ultimate goal: design, operate, andclose the impoundment so that it performs as a solid material and becomes a stable geochemicaland geomorphic form in the environment.

3.2 Variations in regulations

There is a significant variation in specificity and practicality of regulations. The uranium milltailings reclamation work mentioned above was conducted under the Uranium Mill Tailings

43

Remediation Control Act (UMTRCA) of 1977. The regulations were structured as performancecriteria (Appendix A of 10 CFR 40), stating that tailings will be isolated and impoundments willbe stable to the extent practical for 1,000 years and at any rate for at least 200 years. How thisperformance criterion was achieved left room for creative analyses and engineering.

Compare this to the recent Directive 74 from the Alberta Energy Resource Conservation Board,that states that the tailings shall have a strength of 5 kPa one year after deposition. There is noindication of how to measure the strength or what kind of strength this is, and no guideline as towhat is supposed to be achieved by this requirement. It is not clear whether this is a requirementupon discharge or a requirement for trafficability (5 kPa is a bearing capacity that is not sufficientfor foot traffic or vehicle traffic).

The engineer is thus faced with translating these kinds of objectives and goals into practicalengineering criteria: a thousand-year design life translates to design for the probable maximumprecipitation and the maximum credible earthquake.

4 RECOMMENDATIONS

4.1 Variations in site conditions

If there is a solution to these issues, we submit it rests in recognition that a set of detailed regulationsdoes not apply world-wide. Because mineralized deposits occur in all parts of the world, minedmaterials vary enormously from country to country, from region to region, and from climate toclimate. Standard practice for the construction and operation of tailings impoundments differs fromplace to place for many reasons, including these:

• Ore Host Rock: The host rock in which gold occurs in South Africa is different from the sandsfrom which oil is extracted at an oil sand mine in Alberta.

• Processing: The crushing, grinding, and milling that may be necessary to make it possible toextract platinum in the Bushveld is different to what needs be done to liberate diamonds fromkimberlite in the Canadian Northwest Territories.

• Chemicals: The chemicals added to liberate the ores impacts the waste disposal facilities.Cyanide added to a Nevada heap leach pad imposes vastly different constraints than sulphuricacid added to liberate copper or the lixiviant in Namibia to liberate uranium.

• Topography: In the steep valleys of British Columbia you have to design and operate the tailingsimpoundment and waste rock dump in a completely different way to what you may do in the flatdeserts of Australia.

• Climate (precipitation): If it rains eight meters a year, as is the case in Papua New Guinea, wastefacilities will be different than those at a mine near Tucson, Arizona. Too much water is an issuein the first case; too little water may be an issue in the second case.

• Climate (temperature): In Northern Canada, planning for snow, ice, winter freeze and springthaw is necessary in the design or closure of a mine waste facility. Conversely in the heat ofNorthern Chile, sun drying and evaporation may lead to a limited water management approachor a very different closure cover.

• Laws and Regulations. While there are international guidelines and codes to abide by, the ultimatereality is the law of the country where the facility is located. In California mine pit backfillingis required, with the idea being that the mine waste will be used to backfill the pit. In Canada,it may be allowed to put mine waste in a lake and plan for a long-term water cover.

• Historical Practice and Precedent. In South Africa, Fraser F Alexander is the name of the leadingcontractor building and operating slimes dams. .In practice, Fraser F Alexander (who started asa foreman on a mine’s tailings impoundment, and started his own company when he realizedhe could make a profit operating tailings impoundments) succeeded and established precedentsfor operation that prevail today. Similar people in all countries have had decisive impacts onstandard practice—many of which have found their way into laws and regulations.

The only common factors from the variables listed above are the principles and practice ofscience and engineering, specifically geotechnical and civil engineering. To ensure adherence to

44

these principles, implementation of standard practices include the following steps made in thestages of the project.

4.2 Initial studies and documents

At the start of the project, information on the site is collected, with information included in thefollowing documents, as they pertain to the facility:

• Site Selection Report• Alternatives Evaluation Report• Conceptual Design Report• Preliminary Closure Plan

These reports go by many other different names. But the purpose is the same, regardless of thename: characterize the area, identify and compare potential waste facility sites, compile plans andcost estimates to build, operate, and close the facilities. In addition, reports on site and facilityinformation should be produced that include the following information:

• Regional and site geology and geohydrology• Regional and site seismicity• Regional and site climate (precipitation, evaporation, temperature, wind)• Relevant aspects of air and water quality in the region and at the site• Site characterization information (surface and subsurface)• Construction materials identification and properties• Tailings and mine waste geotechnical and geochemical characteristics

This information is used to realistically evaluate the feasibility of the project. If the feasibi-lity of the project is favorable to proceed, permitting and planning activities proceed. Additionaldocuments typically produced at this stage include:

• Design Criteria Document• Regulatory Compliance Plan• Risk Assessment Evaluation• Peer Review Reports

4.3 Design studies and documents

With these reports (or the local variants or equivalents) in place, detailed design may proceed,typically with the following elements:

• Design drawings• Technical specifications• Design report, including supporting design calculations and analyses• Engineer’s cost estimate

Design calculations and analyses are a very important part of the design process, and requireclear documentation for thorough checking. Analyses and calculations that should be preparedinclude:

• Foundation seepage and stability analyses• Slope stability analyses (with deformation analyses as necessary)• Water and chemical mass balance calculations• Settlement and consolidation analyses of tailings and other structures• Hydraulic analyses of channels and other hydraulic structures• Hydraulic analyses of impoundments for appropriate embankment freeboard or spillway capacity• Cover performance analyses on final surfaces

45

4.4 Construction and operation documents

After the financing is arranged, the permits are in place, the mining team is assembled, andcontractors are selected, construction and operation may begin. The key documents at this stageinclude:

• Management and Operating Manual• Instrumentation and Monitoring Plan.• Emergency Response Plan

The Management and Operating Manual should set out in clear and specific language how thefacility will be operated and managed. This includes the personnel involved and their roles andresponsibilities—also their contact details and how often they should be involved and consulted.The manual should also describe (by way of simple diagrams) the components of the facility andhow they tie together. The level of detail should be such that the field personnel can use the diagramand text to understand how it all comes together and should work. The manual should include briefchecklists of how to operate each component. Common items include:

• Delivery pipes and valves• Penstocks and return water barges• Sediment ponds and runoff control facilities• Dikes and embankments

A recommended section in the manual is an Incident Management Plan. This consists of docu-mentation of incidents (or near misses in safety nomenclature), with root causes and mitigationmeasures. We believe that ten incidents make for one accident, and ten accidents make for oneserious accident—or worse, a catastrophic failure of the facility and a significant environmentalimpact.

The Instrumentation and Monitoring Plan outlines the instrumentation to be installed in thefacility, how it is operated and maintained, how often the readings are taken, and how often the datais downloaded or collected. There should be precise instructions on frequency of documentationand sending the results of observations and monitoring up the chain of command to responsibleengineers and managers.

We hardly need go into the contents of an Emergency Response Plan. Suffice it to say it listsall personnel who need to be informed, called in, or set to work to deal with an emergency. Thechallenge to the geotechnical engineer is to make sure all emergencies that may occur are identified.The most basic items are listed below:

• Embankment slope movement or failure• Overtoppping of embankment• Release of tailings or process water• Fire

An additional document is a plan outlining response to the ever-changing mining conditions(Observational Method Implementation Plan) that typically evolve as the mine develops. This plan isdifferent from an emergency response plans, and is a long-range planning document. Recommendedprovisions responding to changing conditions include the following:

• Foundation material characteristics and performance• Construction material properties• Tailings, waste rock, and/or heap leach material characteristics• Water balance performance (precipitation, runoff, and too much process water or too little

makeup water)• Erosion and sediment buildup• Capacity for tailings, waste rock, process water and other materials

46

Another recommended document during operations is a community relations plan. Many miningoperations have a community relations person or department to address local and regional concernsand issues. These issues change during mine operation and as the facility transitions into closure.

4.5 Closure documents

As the mine’s end of life approaches, ideally the existing closure plan has been updated duringoperations, so that actual closure is consistent with existing plans. Key documents associated withclosure include the following:

• Updated Closure Plan• Closure Construction Design Documents (drawings, specifications, design calculations and

analyses)• Post-Closure Monitoring Plan• Post-Closure Emergency Management Plan

The focus of closure and post-closure becomes water management, including surface watermanagement and erosion protection, residual process water evaporation or treatment, and tailingsporewater management.

5 CONCLUSIONS

Mining and associated tailings disposal involve relatively large areas and volumes, and consumesignificant quantities of water. The changes in topography from mining and disposal of tailingscreate landforms that will remain for centuries. This makes mining a visible target for anti-mininggroups and NGOs, even without failures and unanticipated incidents. The efforts of these groups areseen in pressures placed on legislatures and regulatory agencies, and in challenges in the permittingprocess and after permits have been issued. These efforts are despite the fact that development ofnatural resources is a key component of a productive society.

There is a credible body of information about tailings impoundments, including failures and theircauses, as well as guidelines for proper design, construction, operation, and closure. Land man-agement agencies have requirements for mining and tailings disposal on their lands, and individualstates have regulations for embankment safety and tailings containment. Both land managementand state agencies have surety bonding requirements to cover closure and reclamation costs. Theserequirements and regulations are effective if the regulatory agencies have support and enforcementmandates, and the regulated mining companies are serious about quality operations and communityrelations.

In addition to these guidelines and regulations, procedures for safety management and documen-tation used in other high-risk industries should be adopted for operation of tailings impoundments.Review of operation and closure of impoundments by knowledgeable and experienced regulatoryagency personnel or (if not available) third-party reviewers should be used.

The MiningWatch document, referenced in the abstract of this paper, outlines the failures oftailings impoundments and problems with mining activity. This is done to stress the point oflimiting where and how tailings can be disposed and recommending recycling of metals to reducethe need to mine. The MiningWatch document is one of many written products that are founded ontruth, but are created to make a specific point, without independent review or cross examination.

An egregious example of prejudice in engineering analysis that the authors were involved withwas an accusation that a mine was the cause of cracks in houses in the vicinity of the mine. Thelogic for this accusation was that there was no other simple explanation for the damage to thehouses. This false engineering conclusion has since been refuted with sound data collection andinterpretation from experts in seismic analyses. But it will take time to undo the publicity createdfrom the initial false accusation.

This example is not an isolated incident. It may be argued that such people, engineers included,serve a purpose in prompting the mining industry to act to examine the truth and act proactivelyto negate the potential for slander from such people. But, as long as the mining industry does notoperate properly and to the highest standards, mine facilities are fertile stalking ground for badpublicity and misinformation.

47

REFERENCES

Arizona Department of Environmental Quality (ADEQ), 1998. Arizona Mining BADCT Guidance Manual,Aquifer Protectiojn Program, Water Quality Division, September.

Australian National Committee on Large Dams (ANCOLD), 1999. Guidelines on Tailings Dam Design,Construction, and Operation, October.

California Mining Association (CMA), 1992. Mine Waste Management, Hutchison and Ellison Eds., LewisPublishers.

Federal Emergency ManagementAgency (FEMA), 2001. “The National Dam Safety Program, Research NeedsWorkshop: Hydrologic Issues for Dams.”

International Commission on Large Dams, Committee on Mine and Industrial Tailings Dams (ICOLD), 1982.“Manual on Tailings Dams and Dumps,” ICOLD Bulletin 45, ICOLD, Paris.

International Commission on Large Dams, Committee on Mine and Industrial Tailings Dams (ICOLD), 1989.“Tailings Dam Safety,” ICOLD Bulletin 74, ICOLD, Paris.

International Commission on Large Dams (ICOLD), 1987. “Dam Safety – Guidelines,” ICOLD Bulletin 59,ICOLD, Paris.

International Commission on Large Dams (ICOLD), 1992. “Selection of Design Flood, Current Methods,”Bulletin 82.

International Commission on Large Dams, Committee on Mine and Industrial Tailings Dams (ICOLD), 1996.“A Guide to Tailings Dams and Impoundments,” ICOLD Bulletin 106, ICOLD, Paris.

International Commission on Large Dams, Committee on Tailings Dams and Waste Lagoons (ICOLD), 2001.“Tailings Dams – Risk of Dangerous Occurrences,” ICOLD Bulletin 121, ICOLD, Paris.

International Cyanide Management Institute (ICMI), 2005. “Implementation Guidance for the InternationalCyanide Management Code,” November.

International Cyanide Management Institute (ICMI), 2008. “The International Cyanide Management Code,”August.

International Finance Corporation (IFC), 2007. “Environmental, Health and Safety Guidelines for Mining,”December 10.

Nevada Division of Environmental Protection (NDEP), 1989. “Regulations Governing Design, Construction,Operation and Closure of Mining Operations.”

United Nations Environment Programme (UNEP), 1996. “Environmental and Safety Incidents ConcerningTailings Dams at Mines,” based on survey conducted by Mining Journal Research Services for UNEP.

U.S. Army Corps of Engineers, 1982. “Engineering and Dam Stability for Earth and Rockfill Dams,”EM-1110-2-1920, U.S. Government Printing Office.

U.S. Committee on Large Dams, Committee on Tailings Dams (USCOLD), 1994. Tailings Dam Incidents,USCOLD.

Vick, S.G., 1983. Planning, Design, and Analysis of Tailings Dams, John Wiley and Sons, NewYork; reprinted1990, BiTech Publishers Ltd., Vancouver.

Wilson S.D, and R.J. Marsal, 1979. “Current Trends in Design and Construction of Embankment Dams.”American Society of Civil Engineers.

48


Overview: Tailings disposal and dam construction practicesin the 21st century

A.J. BreitenbachAusenco Vector, Denver, Colorado, USA

ABSTRACT: Tailings disposal practices in the mining industry have been changing in the last10 years from high water content conventional slurry tailings disposal to more water efficientthickened tailings, paste tailings and dry stack tailings disposal in recent time. These changes areoccurring in part due to more scarce fresh water resources to sustain the mill production operations,and in part due to more efficient thickener tank designs for removal of excess tailings slurry waterfor plant reuse. Tailings disposal storage practices are also changing to co-disposal of tailings inmine waste piles with the potential for end of operations reprocessing of tailings as backfill disposalinto depleted mine excavations at closure.

This paper discusses the recent changes in tailings disposal and storage practices and their impacton tailings dam construction in the 21st Century.

1 INTRODUCTION

Mine operations for tailings disposal have been evolving in recent time due to three primary factorsrelated to water supply at mine sites: 1) the improvement in thickener tank technology within the last10 years in reducing the solids to water ratio in slurry tailings for less mill plant water demand, 2)the scarcity of fresh water supplies in dry climate areas of the world, and 3) the general awarenessin the government and private sector to maintain and conserve existing fresh water sources forsustainability going into the 21st century.

The overall trend in modern day mining is to decrease the water content in tailings disposaloperations and switch from upstream method dam construction to centerline and downstreammethod dam construction. In some cases such as paste or dry filter tailings disposal, tailingsco-disposal in mine waste fill, or tailings backfill in depleted mine excavations, the trend goeseven further to limited or no dam construction. This discussion presents an overview of the typesof tailings disposal and storage operations and their impact on tailings dam construction.

2 TAILINGS DISPOSAL AND STORAGE DEVELOPMENT

2.1 General

The four basic types of tailings transport and disposal operations listed in the general order ofdecreasing tailings moisture content include conventional, thickened, paste and dry filter tailings.The 4 categories of tailings disposal storage facilities include co-disposal in mine waste piles,backfill in depleted underground or open pit mine excavations, backfill in natural water featuresincluding rivers, lakes and deep sea disposal, and the more common practice of land based tailingsdam containment. The different types of tailings disposal operations and storage facilities arediscussed in this section. Photo 1 shows the larger diameter conventional tailings thickeners beingreplaced with smaller sized thickeners for thickened or paste tailings at lower water content.

2.2 Conventional tailings disposal

Conventional tailings refer to the most common type of tailings mill waste slurry used in the 20th

century having a relatively low solids to water ratio ranging from about 35 to 50 percent by total

49

Photo 1. Large conventional thickener to smaller high rate (thickened tailings) to high wall and high angle(paste tailings) thickener, courtesy of Outotec.

weight, depending on various tailings rheology factors such as the whole tailings gradation, specificgravity, viscosity and hydraulic flow friction loss. The conventional tailings slurry is generallycharacterized by turbulent flow and segregation of coarse and fine tailings during transport and atthe final disposal points. The high conventional tailings slurry water content allowed “desanding”of the more coarse copper tailings grind in classifier boxes since the 1910’s, and enhanced cyclonecentrifugal separation of coarse underflow sands from the more fine overflow “slimes” for buildingperimeter tailings sand dams since the 1950’s.

The conventional tailings slurry could be transported long distances away from the mill by gravityflow in open launder chutes or enclosed pipelines to the impoundment. Energy dissipation dropboxes or centrifugal pumping were added as needed for discharge from single or multiple spigotpoints around the tailings impoundment. The settled tailings density for conventional tailings wouldbe relatively low with the more coarse tailings forming a flat alluvial fan beach surface at eachdisposal point. The finer tailings slimes with low sand content flowed further into the interiortoward the water pool. Rotation of the slurry disposal points from active to inactive areas aroundthe impoundment perimeter allowed thin layers of the low density beach material to drain and dryfor densification and increased strength. Clear water from the water pool surface would be decantedby gravity flow in towers and pipelines extending beneath the tailings dam or by lower risk floatingbarge pumps in modern times for water return to the plant.

Visually the conventional tailings slurry would appear to have excess water and a low settleddensity at the discharge point, although exceptions include non-segregating fine clayey gold andsilver tailings, phosphatic clays and oil sand sludge-like slurries that have their own unique defi-nition with very low solids to water ratios. The conventional tailings beach slope would generallybe in the range of 0.5 to 1 percent grades and flatter toward the water pool. The cycloned sandfrom conventional tailings with a high sand content (typically copper, molybdenum, lead and zinctailings) can develop underflow sand pile slopes beneath the cyclone towers steeper than 50 percentgrades (2 horizontal to 1 vertical slope).

2.3 Thickened tailings disposal

Thickened tailings refer to a higher solids to water ratio slurry typically in the range of 50 to 65percent and characterized by non-segregating laminar slurry flow in the open launder chute orenclosed pipeline. The more dewatered thickened tailings are developed in high rate and high wallthickeners at the plant. The high rate “thickener within a thickener” is a recent development in thelast 10 years that can reduce the thickener time and reduce the size of the tank diameter compared to

50

conventional thickeners. The hydraulic friction loss increases with reduction in the tailings slurrywater content, however the friction loss remains in the acceptable range of centrifugal pumps. Thesettled density at the discharge point increases with less entrapped water content in the tailingsbeach and the water pool return pumping is less to the plant.

Visually the thickened tailings slurry would appear to have a “soupy” non-turbulent dischargeflow with less segregation of coarse and fine tailings in the tailings beach surface materials. Thethickened tailings beach slope surface generally increases with the degree of thickener dewateringto between 1 to 2 percent grades. As noted earlier for the conventional tailings, there are alwaysthickened tailings slurry “exceptions to the rule” for beach slopes, including flatter beach slopesduring startup high rate of rise conditions or steeper tailings beach slopes with submerged disposalinto a startup water pool.

The end result of thickened tailings impoundment disposal is a more uniform tailings gradationfrom beach to water pool at a higher settled density and whole tailings strength with less waterdemand at the plant compared to conventional tailings. There are pros and cons to the use ofthickened tailings, like allowing for a better seal in unlined basins (less disposal water contentminimizes seepage and segregation of tailings sand from the finer slimes materials for a lowerpermeability perimeter tailings beach seal), and the opposite of this affect is that cyclone separationof tailings sands from slimes becomes more difficult with less water content in the thickened tailings.

2.4 Paste tailings disposal

Paste tailings refer to an even higher solids-to-water ratio at 65 to 75 percent typical for the non-segregating mud-like slurry compared to the more fluid conventional and thickened tailings slurry.The mud like paste tailings are developed with high wall and high angle thickeners at a slower plantproduction rate. The paste tailings have high hydraulic friction losses for pumping and require aswitch from centrifugal to positive displacement pumping for pipeline transport. The paste tailingssettled density is significantly higher and the tailings beach slope increases between 2 to 10 percenttypical for impoundment disposal. Visually the paste tailings discharge slurry would appear to bea mud flow with a low amount of “bleed” water seeping from the beach slope.

Paste tailings have been used in the backfilling of underground mines since the 1980’s andrarely used in tailings disposal in above ground impoundments due to the high cost in the plantthickener, pumping transport and disposal of mud-like tailings (high capital and operating costs).In recent times a few mine sites have used above ground paste tailings disposal, where water supplyis scarce. The impoundment cost savings in having less tailings dam containment and highersettled density are offset by high disposal point maintenance costs and loss of interior storagecapacity due to steep beach grades from the interior towers. Other restrictions to adequate pastetailings containment include liquefaction issues in high seismicity areas and sediment transportcontainment from sloping mud tailings surfaces in high rainfall areas.

2.5 Dry filter tailings disposal

Dry filter tailings refer to dewatered tailings that can no longer be pumped and require mechanicaltransport by trucks or conveyors for disposal in the tailings impoundment. The dry filter tailingssolids to water ratio at 75 to 85 percent typical allows the material to be stacked at a relativelyhigh density and strength compared to the pipeline transported tailing slurry. The dry filter tailingscapital and operating costs are higher than pipeline transport; however, the higher costs are offsetby the low risk of tailings spills (no water pool and retention dams), order of magnitude reductionin the tailings area with stacked tailings fill lifts, and the lowest water demand at the plant.

A few gold and silver operations have used dry stack tailings since the 1980’s and a secondarybenefit of the dry filter water recovery at the plant includes capturing the majority of the cyanidewater for reuse in operations in addition to lower cyanide destruction costs. The stacked tailingspiles also allow for early reclamation of the final exterior slopes during operations for less end ofmine closure costs.

2.6 Co-disposal in mine waste fill

Co-disposal of tailings in mine waste fill started in the 1990’s with construction of lined andunlined impoundments contained by perimeter mine waste piles. The importance of engineered

51

Photo 2. Depleted mine pit lined and backfilled with conventional or thickened tailings disposal.

compaction and filter control containment in mine dump fills was demonstrated by the loss of thetailings impoundment water pool through a clay core soil liner in 1996 at the Omai gold project inGuyana. Mixing of tailings within mine waste rock fill is a new development and likely will notreplace the more practical concept of using perimeter mine waste piles with interior filter drainagecontrol, and in particular the use of geomembrane liners where the water pool may be located nearthe waste dump loose lift fill slopes.

2.7 Depleted mine excavation disposal

Numerous underground mines have been backfilled with paste or treated thickened tailings backfillsince the 1980’s for safety or mine closure reasons. Tailings backfill in completed underground mineworkings included paste or thickened tailings materials mixed with cement and other stabilizingadditives, which reduce the required amount of tailings to be stored in above ground impoundmentfacilities.

Open pit mines have historically been left in an open condition during operations to closure,unless unstable wall conditions warranted partial backfilling with mine waste fill to complete thepit ore excavations. Most open pit walls are constructed to a safety factor of 1 to extract as muchore from the ground with the least amount of stripping to expose the ore body.

The backfilling of depleted mine pits with lined landfills, tailings impoundments and heap leachpads is a recent development at several mines (Breitenbach 2008). Lined or unlined tailings backfillin underground and open pit mine excavations, where practical, would significantly reduce the minedisturbance area and related reclamation closure costs, in addition to minimizing construction ofabove ground tailings containment dams. Photo 2 shows a lined tailings impoundment at startupwithin a depleted mine pit in 2004 at the El Valle mine in Northern Spain.

Re-milling of existing tailings piles for removal of residual ore metals started in the late 1980’sand is becoming more economic in recent time due to the increase in metal prices and the desirefor sustainability of the mine. Re-slurrying of existing tailings piles back through the mill plant atthe end of mine operations provides an excellent opportunity to backfill the mine pit excavationswith reprocessed tailings to reduce the above ground tailings pile and any pit lake issues at closurewith the recovered metals paying for the end of mine life low cost operations.

2.8 Rivers, lakes, and deep-sea disposal

Disposal of tailings into the natural rivers, lakes and the deep sea environment is a practice thatoccurs to the present day. The environmental impact of direct tailings disposal into rivers and lakesis significant with destruction of habitat life and fresh water sources, and therefore on-going mining

52

Photo 3. Deep water disposal of copper tailings on Vancouver Island courtesy of Amazon web site.

operations like in Papua New Guinea are becoming a less acceptable practice in the 21st century.The environmental impact of deep sea disposal can be argued to have less of an impact comparedto containment on land, particularly for acid generating tailings and waste rock. Photo 3 shows aVancouver Island copper mine in 1999 with nearby deep water disposal before the depleted minepit was flooded with sea water at the end of operations.

2.9 Land based tailings impoundment disposal

The most common type of tailings disposal involves construction of a dam or perimeter dikesystem for lined or unlined containment of conventional, thickened or paste tailings. More minesare adopting the use of compacted earth and rock fill dams with geomembrane liner systems inmodern times for tailings disposal with long term containment stability and improved protectionof baseline groundwater conditions. The primary purpose of the tailings dam containment is tominimize environmental risks with safe disposal of the tailings without spills to closure. One of theearliest geomembrane lined tailings impoundments was the Sweetwater Dam in 1976 for a uraniummine in Wyoming, USA.

3 TAILINGS DAM CONSTRUCTION

3.1 General

Tailings dam construction generally involves a starter dam for initial low capital cost tailingsdisposal containment and for development of a stable tailings beach sloping toward an interiorwater pool. A gravity flow pipeline or barge pump decant water return system would route theimpoundment water back to the plant for reuse in operations. Raises above the starter dam crestlevel included three basic types of dam construction including upstream method, centerline, anddownstream method construction discussed in this section. Photo 4 shows an upstream methoddam raise in 1971 using cycloned tailings towers elevated onto the existing tailings beach fill inNew Mexico, USA. Photo 5 shows a centerline method dam raise in 2001 using vertical cyclonetowers to develop tailings sand dam materials for compacted fill placement at the Quebrada HondaDam in Southern Peru. Photo 6 shows a downstream method dam raise using earth and rock fillmaterials for compacted fill containment at the Marlin Dam in Guatemala.

Conventional tailings disposal is associated with the cycloned upstream and centerline methodtailings dams for producing a clean sand perimeter for drainage and high strength containment.Thickened and paste tailings disposal would generally apply to non-cycloned tailings dam con-struction and co-disposal filter control in waste rock piles. Dry filter tailings stacks, depleted pit

53

Photo 4. Upstream method raise with cyclone towers on the tailings beach surface

Photo 5. Centerline method raise with vertical hydro-cyclone towers and compacted lifts.

backfill, and deep sea disposal would eliminate the need for land based containment dams. Freshwater lake and river disposal hopefully will end world wide as a tailings disposal option.

3.2 Upstream method dam raises

Upstream method tailings dams have been used extensively in the mining industry in the 20th

century with less frequent use in the 21st century due to the higher risk of instability compared toother types of dams. The single most important factor in upstream method dam stability is adequatetailings beach drainage, which requires the ability to deposit settled tailings above the impoundmentwater pool level (prevent submerged tailings disposal with related low density, strength, and poordrainage issues). Therefore the lowest risk upstream method dams have the water pool located awayfrom the dam as much as practical after startup operations. In addition, the perimeter tailings beach

54

Photo 6. Downstream method raise with conventional compacted earth and rock fill.

surface should be allowed to dry by rotation of active disposal areas for densification and reductionof pore pressures in controlled and relatively thin hydraulic fill layers.

The upstream method tailings dam construction typically consists of building a berm raiseabove a starter dam on the existing tailings surface in an upstream fashion. This can be donewith several construction techniques including hydraulic slurry disposal at multiple spigot loca-tions from incremental berm raises, placement, flooding, and drainage of multiple paddock cellwalls, or hydro-cyclone peripheral deposition of tailings sand beach material to develop an exteriorembankment berm fill with drained cyclone underflow sand strength sloping to an interior waterpool.

The most predominant type of upstream method raise construction included excavating thedried inactive perimeter tailings beach fill surface with an excavator and dozer to create a low levelperimeter berm raise in segments around the impoundment. Construction of the perimeter bermswith earth and rock fill materials added additional strength and early reclamation of the exteriorslopes during operations. Tailings cells or multiple impoundments were generally created to allowa portion of the impoundment to remain inactive for the segmental berm raises. As each segmentof berm raise is completed, the tailings delivery pipeline is reset on the new berm crest level forcontinued disposal operations. The paddock wall construction is labor intensive and requires moretime to construct in small raises. The flooded walls are allowed to drain and dry for the next wallconstruction and animal traffic was sometimes used for compaction within each wall cell.

The cyclone sand dam construction with conventional tailings disposal has been more commonfor copper tailings dams, due to the more coarse mill grind sand size and content available in thecopper tailings compared to finer grind gold and silver tailings to form a perimeter sand dam fill.The cyclones provide a more efficient way to separate the larger sand sized particles from the finersized slurry sands, silts and clays by the use of centrifugal force. The tailings slurry is pumpedinto the cyclones under low pressures to minimize pump, pipeline and cyclone wear maintenance.The spinning slurry motion in the cyclones forces the larger sand particles to spiral to the outsidetoward the apex end as underflow sand, while the finer slurry materials are forced to the center ofthe cyclone into an attached overflow pipeline to the impoundment.

The upstream method tailings dams are the most economic to construct for dam raises, unlesslarge quantities of excavated mine overburden waste rock materials are locally available for othertypes of dam construction. The upstream method dam construction was common into the 1980’sand 1990’s, however studies of world-wide dam failures indicate hydraulic fills are susceptible toseismic (earthquake) liquefaction, overtopping, and tailings delivery or water return gravity decantpipe break instability compared to all other types of dam construction (USCOLD 1994). A fastrate of rise in hydraulic fill tailings disposal operations can also increase the potential risk of staticliquefaction failure (Martin and McRoberts 1999).

55

Therefore upstream method tailings dams have seen less frequent use in recent times in highseismic zones of the world, as well as any areas where the tailings dam can be classified as ahigh hazard structure. Earthquake prone countries like Peru and Chile have discontinued the useof upstream method dam construction by law since the 1990’s.

Compacted rock fill shells have been used to improve the stability of existing upstream methodhydraulic fill tailings dams in high seismic zones since the 1970’s, starting with the CodelcoBarahona copper tailings dams in central Chile. Essentially the upstream method dams can beconverted to a more stable structure with the use of compacted mine waste rock fills in the exteriorshell of the dam.

3.3 Centerline method dam raises

Centerline method dams rarely have been used in past construction for various reasons mainlyrelated to the popularity of upstream method dams into the 1980’s, as well as the need for a widetwo way traffic downstream fill zone with each raise to accommodate large haul trucks loaded withlow cost mine waste rock fill. However, more centerline raises are being constructed in moderntimes due to cyclone “sand dam” construction primarily in the copper industry. The concept oftailings sand beach centerline raise construction above the starter dam crest began in the 1910’swith the use of sand declassifier boxes that appear to be first used by Kennecott on the Barahonacopper tailings dams at the El Teniente mine in Chile. The declassifier boxes allowed the heavierand larger particle sizes of wet underflow sands to drop down from the towers in the downhilldirection, while the overflow finer tailings slurry particles were routed in open channel launderchutes and deposited toward the interior impoundment limits. The concept of better drained andhigher quality cycloned tailings sand for more efficient upstream and centerline method dam raiseconstruction started in the 1950’s and continues to the present day. Some modern day sand damsare using elevated tailings delivery pipelines along a series of vertical steel pipeline towers, whichallows the tailings slurry pipeline disposal to continue without stopping and relocating the pipelinefor each centerline dam raise.

The centerline dam raise essentially maintains a vertical crest raise above the starter dam crestto the ultimate dam height utilizing the peripheral tailings beach fill as the buttressing support tothe upstream section of the dam. The disposal of slurry tailings in thin beach fill lifts stabilizes theupstream slope of the crest berm at 4 to 6 m typical berm raises, while the crest and downstreamsections are constructed with earth and rock fill or cyclone sand fill. Controlled lift placementand compaction are added in the downstream dam zone for stable conditions in high seismicitylocations. Earthen dams may include vertical drains and filter control in the crest fill raises withconnection to the starter dam drain system for the option of operating water pools or designstorm storage near the dam limits. Water pools next to sand dams would have a higher risk ofinstability.

Some earlier centerline method dams typically included earth and rock fill materials at thecrest and in downstream zones of the dam. A variation to the earth and rock fill centerline raiseincludes steep upstream slopes with compacted rock fill for subsequent buttressing by the hydrauli-cally placed tailings beach fill. An example of this type of dam includes the Cannon Mine Damin Washington, USA with a steepened upstream slope raise of 20 m that allowed the existingdam crest to be raised another 5 m vertically to the ultimate dam height of 146 m above thedownstream toe.

Examples of mobile centerline method cyclone sand dam raises include the Doe Run Dam inMissouri, USA). The mobile cyclone unit deposits a new layer of sand fill along the crest with theunderflow sands from the cyclone forming an approximate 3H:1V downstream slope assisted by adozer. The solids to water ratio, coarse tailings sand underflow to fine tailings slime overflow ratioand slurry pressure to the cyclone are controlled with each pass of the mobile unit for continuouscenterline dam crest raises.

An example of a stationary centerline method cyclone sand dam raise with vertical hydro cyclonesincludes the Quebrada Honda Dam in Southern Peru shown in Photo 5. The dam crest is raised ver-tically by mega cyclone stations or multiple cyclone clusters along the inside crest, with underflowsands excavated by draglines and placed in controlled lifts by dozer for compaction. The dam wasapproximately 85 m high at a crest length of 3.9 km, when it was subjected to the June 2001 8.4 Mearthquake. The tailings beach slope liquefied according to mine personnel and minor cracks were

56

observed along the upstream dam crest (GEES web site 2008). The sand dam crest and downstreamslope remained stable.

3.4 Downstream method dam raises

Downstream method dams have the best historic record for stability and are commonly used forwater storage dam raises with adequate filter and drain control. A bench is provided at the upstreamslope for continued tailings disposal while the downstream section of the dam is raised in phasesto the ultimate dam height. Drain and filter systems in the starter dam can be extended in thedownstream raises as needed, as well as transition zones from lower to higher strength downstreamshell materials. The downstream raises allow geomembrane liners to be placed on the upstreamslope for lined impoundment facilities. An example of a downstream raise to an existing tailingsdam with the water pool located in the dam area is shown in Photo 6.

4 CONCLUSIONS

An overview of the tailings disposal operations into the 21st century indicate a transition fromconventional high water content tailings disposal to lower water content thickened tailings dis-posal. Where fresh water supply is scarce, paste or dry filter tailings disposal have been used at asignificantly higher tailings disposal costs for sustainability of the fresh water supply sources.

Tailings disposal facilities include co-disposal in mine waste fill, backfill disposal in depletedmine excavations, disposal in lakes, rivers and deep sea environments, and disposal in land basedlined and unlined tailings impoundments. More mines are adopting the use of compacted earth androck fill dams with geomembrane liner systems in modern times for tailings disposal with longterm containment stability and improved protection of baseline groundwater conditions.

A transition from upstream method dam construction to centerline and downstream methodconstruction has occurred since the 1980’s related to the higher risk of upstream method daminstability, especially in seismic (earthquake) active areas of the world.

A transition from tailings dams to other options for lesser dam storage or “no dams” will likelyoccur more often in the 21st century including small containment berms for paste and dry filtertailings containment, and co-disposal in waste rock piles or tailings backfill in depleted mine pitexcavations. Re-mining of existing above ground tailings piles at the end of mine operations forresidual metal recovery and pit backfill may become the best sustainable option for mines at closurethat benefit the owner (post-mining operation pays for itself), the dam engineers (less long termcontainment risks) and the environment (less above ground dam and backfilled pit issues).

REFERENCES

Breitenbach, A.J. (2008), “Backfilling Depleted Open Pit Mines with Lined Landfills, Tailings Impound-ments and Ore Heap Leach Pads for Reduced Closure Costs”, GeoAmericas 2008 Conference Proceedings,Geosynthetics in Mining Session, IFAI, Cancun, Mexico.

GEES.usc.edu/GEER/peru_earthquake (August 2008), “Geotechnical Aspects of Mine Facility Performance”,Quebrada Honda Tailings Dam.

Martin, T.E. and McRoberts, E.C. (1999), “Some Considerations in the Stability Analysis of Upstream MethodTailings Dams”, Tailings and Mine Waste Conference, Colorado State University, Ft. Collins, Colorado.

USCOLD (1994), “Tailings Dam Incidents”, United States Committee on Large Dams, p. 82.

57


A history of South African slimes dams engineers

J. CaldwellRobertson GeoConsultants, Vancouver, BC, Canada

G. McPhailMetago Environmental Engineers (Australia) Pty Ltd, Perth, Australia

ABSTRACT: Geoffrey Blight, retired Professor at the University of the Witwatersrand in Johan-nesburg, South Africa has just published a new book Geotechnical Engineering for Mine WasteStorage Facilities. In this book he comprehensively describes the history and current state-of-the-artof South African tailings disposal. Go to almost any other country where there are mines, and youfind the offices of consulting practices that started in South Africa and that have expanded globallyon the strength of their tailings expertise: SRK and Knight Piesold are but two examples. This paperis a short history of tailings disposal design and construction starting in South Africa with Fraser F.Alexander and the ring dike system and thin lift subaerial deposition; then progressing through theyears of emigration from South Africa of engineers and companies to most other mining countries.We tell the story of how South African engineers and consultants have become so integral a part ofcurrent international tailings practice by adapting South African practice to international needs andby adopting and integrating practices pioneered by the great North American engineers faced withearthquakes, cold climates, dry deserts, and social & environmental concerns not initially part ofearly South African practice.

1 INTRODUCTION

Geoff Blight’s new book on tailings impoundments, or mine waste storage facilities as he calls them,is destined to become a classic. While we disagree with his term, storage facilities, we acknowledgethat he has written a great book that documents the technical aspects of SouthAfrican “slimes dams”and waste dumps, as we used to call them—harkening back to the days before political correctnessforced us into verbal acrobatics.

Geoff’s book tells nothing of the people of the South African slimes dam history and achieve-ments. Thus we write this paper to record what we remember of them. We tell their tales frommemory and personal recollections; we make no attempt at documented research. All we hope isthat this paper stimulates others to more closely research and document a fascinating history thatis in danger of disappearing into the mists as memories fade and colleagues die.

Some may say this paper is advertising for ourselves and the friends and colleagues of whomwe write. That is simply wrong. This is a personal history based on recollections. If it helps thoseof whom we write, so be it. If it hurts, that is what we recall. We hope only that others take up theirpens to document the history of the people of tailings impoundments worldwide—a story too bigfor one paper.

2 FRASER F ALEXANDER

Gary Rae who managed Fraser Alexander, or Frasers as we called them, told that the company hadbeen started by Fraser F Alexander near the turn of the last century.

The story goes thus: Fraser F Alexander was a foreman on a mine. He was in charge of buildingthe slimes dam. He did this mostly by experience and native skill. He was bedeviled by the fact that

59

the slimes never came at a consistent rate of consistent water content. Sometimes he could buildwall with the slimes; but too often the slimes were nearly pure water and he could not build wallswith the slimes. One day he went to the mine manager to complain. The response was predictable;he was fired. He went home to sit on the stoep and relax.

But things at the slimes dam went from bad to worse. Those left behind could not build the damand fell behind and threatened to shut down the mill as they could not keep up with the slimes. Indesperation, the mine manager set a delegation to Fraser F Alexander to ask him to return to hisold job. He declined. Things got worse and worse at the slimes dam, so the delegation returned.

He offered to return, not as an employee, but as a contractor and he demanded he be paid a pennya ton of tailings. His offer was accepted, he returned, and the slimes dam was back in operation.He never complained again.

Many mines bought into his concept for he had a simple way of building dams: dig a trench;deposit some slime; let it dry; and then using the abundant labor of the times, build a small dike;place more slime and proceed. The process as it developed and became a subject of academicinterest is well described by Blight.

Gary told me that the company struggled as there were always failures and fights with minemanagement. Then one day, the Daly brothers took control of the company. Gary married JuneDaly, and we remember her as a beautiful and vital woman, who loved luxury but was so nice thatyou could not but love her and enjoy every minute in her company.

Fred Daly, June’s father, owned Fraser F Alexander when we worked on slimes dams in SouthAfrica. He was small and round. He had skin cancer and had been told to avoid the sun. He wouldscurry from a car with darkened windows to an office of drawn blinds. He was tough and demandedresults of Gary, who was an ebullient and energetic fellow with a constant smile and an almostinstinctive understanding of slimes dams.

3 JERE JENNINGS

Jere Jennings was the father of soil mechanics in South Africa. The majority of the engineers listedin this paper went through undergraduate and many of them, postgraduate, teaching by “the Prof”as he was affectionately known at the University of the Witwatersrand where he was for many yearsthe head of the Department of Civil Engineering. Jere was the first engineer in South Africa toput a filter drain into a slimes dam. This was at St Helena Gold Mine in the Welkom Goldfieldsof South Africa. The second author of this paper had the privilege of not only studying under “theProf” but also of carrying out a seepage and stability assessment of the St Helena slimes dam in thelate 1970’s approximately 10 years after the drain had been installed. This facility is still operatingwell and keeping the phreatic surface under control.

When the now infamous Bafokeng dam failed in 1974 it was Jere Jennings that the miningcompany turned to for help in establishing the cause of the failure in which 3 million cubic metersof liquefied tailings flowed down a shaft and underground, killing 13 miners, and, at surfacecontinued to flow 40km to the Vaalkop dam where the flow stream was arrested. His explanation ofa “flukey” piping failure caused by layering of fine and coarse tailings is still oft disputed by rivalsand doubters but the fact remains that he convinced a Judge of the unforeseeability of the failure.

4 OSKAR STEFFEN

When it came to time to fix up the failed dam and bring it back into operation Jere turned to OskarSteffen saying to him words to the effect “you need to get involved in these blasted things becauseif you don’t who knows what will eventuate”. So it was that Jere brought geotechnical thinking andanalysis to slimes dams in South Africa and Oskar took it all to the next level.

Oskar oversaw the re-commissioning of the Bafokeng facility and brought on board a num-ber of the names mentioned in this paper to carry out what Jere had charged him with – to getinvolved. Even today Oskar is still involved with the tailings facilities at Bafokeng where the currentoperational facility will be taken to a height of 150m as an upstream-constructed structure.

The second author has fond memories of being the gofer for Oskar as a young engineer andaccompanying him on the “Quarterly Inspections” at Bafokeng. These were comprehensive affairs

60

involving aerial inspections using a helicopter followed by a presentation of the quarterly report bythis author to a group of some 20 people who included everyone from the General Manager downto the Fraser Alexander slimes dam operator. After the presentation and discussions there was theground inspection of issues identified while in the air, such as wet spots and gullys and any issuesfrom the meeting. At the meeting, after the welcome, the first item on the agenda was “the reporton the aerial inspection by Doctor Steffen”. Oskar would have them spell bound with his uncannyand astute observations and insights on what he had observed. For a young engineer, this was thestuff of heroes and a jolly good show for the day.

5 FRITZ WAGENER

Just around the time that Jere Jennings was twisting Oskar Steffen’s arm about taking on the Bafo-keng work Fritz Wagener of the firm Jones and Wagener began taking an interest in slimes dams.Fritz went on to do very important and challenging work in the West Rand Goldfields where thegeology comprises dolomites compartmentalized by dolerite dykes at 3 to 5km intervals. The dykesenabled individual mines to dewater the dolomites within a compartment so that deep undergroundmining could proceed. This dewatering of a compartment inevitably led to the development ofsinkholes some large enough to engulf entire gold processing plants or large segments of slimesdams. Fritz went on to become the pre-eminent authority on the construction of slimes dams overthe dolomites and ultimately completed his PhD on this subject. Today Fritz is still called upon toprovide invaluable specialist input on the issue of sinkholes, finding developing cavities and thendealing with them.

6 JOHN WATES

John was a protégé of Fritz in his earlier years and became proficient in slimes dam design andmanagement in the goldfields of South Africa. John went on to develop his own consulting practiceand was instrumental not only in furthering the engineering of slimes dams but also of incorporatingenvironmental engineering considerations into slimes dam planning, design, operation and closure.Today John has moved into the contracting arena and heads up the Strategic Projects group of FraserAlexander.

7 MIKE SMITH AND JOHN ROBBERTZE

Gary Rae started calling Oscar Steffen, one of the founders of what is now called SRK, for adviceon the design and operation of slimes dams. More-so after the failure of the Bafokeng slimes dam,one of the many that Frasers constructed. Oscar could not do all the work and so he passed the taskon to Mike Smith and John Robbertze, at that time juniors in SRK. They grew wise in the ways ofslimes dam design and operation, and so they left SRK, founded their own slimes dam design andoperating company and grew rich. Along the way they fought many battles with Frasers, but theyovercame. Sadly John passed away a few years ago. Mike sold out to Stefanuti Construction butstill works each day with humor and laughter.

With Mike is Dave Jansen, one of the originals from Fraser FAlexander. He too has an instinctiveunderstanding of the way slimes flow and how best to build a dam of them. The first author of thispaper first met him when designing the Richards Bay gypsum slimes dam. He taught us much thatwe have carried with us in our daily practice. When we last met him he was silver gray and still ascalm and humane as always, with a profound insight into the practicalities of slimes dams.

8 MIKE GOWAN

Gary Rae decided that Frasers needed a civil engineer. Somehow Gary persuaded Mike Gowan,an old friend from the University, to join Frasers. Mike loved the fast cars the company providedhim, and delighted in driving around the Witwatersrand, and then calling me on his car-phone (a

61

rarity in those days) as he drove home. We would spend an hour or two discussing what he hadseen that day and what he wanted me to do the next day: usually run a stability analysis; revise alayout drawing; or compose a memo to set things aright.

Today Mike is in Brisbane working with Golder Associates and traveling the world consultingon tailings impoundments. He is the specialist in co-disposal of waste rock and tailings and anexpert in the reworking of slimes dams, which are so often reprocessed to recover yet more goldand uranium.

9 WLPU

The dominant consulting company in the 1960s and 1970s in Johannesburg was Watermeyer,Legge, Piesold, and Uhlman, more commonly referred to as WLPU. The company started as F. E.Kanthack and Partners. Dr Kanthack had worked for the Department of Water Affairs. He foundedthe company and when he retired, his partners put their names to the letterhead.

Time passed and WLPU changed their name again to Knight Piesold. They moved into the designof tailings dams. Rob Williamson and Ronnie Schurenberg have, for many years spearheaded theirwork in tailings within South Africa with a strong focus on Palabora Copper Mine in the westernpart of the country.

WLPU expanded their operations to North America. Don East came to start the Denver opera-tions. I met him at a conference in Fort Collins and he impressed me with his verve and drive. Hetold me that he was determined to succeed, and determined to outdo Andy Robertson and SRK.

Don secured a tailings dam project in Peru. He expanded the Knight Piesold operations toSouth America at a time when we only dreamed of those horizons and was later joined by RonnieSchurenberg. Don has since left the company, and I am told he is married to a Peruvian womanand happy in Lima.

Today Knight Piesold says this about their practice; note the change from the company’s initialfocus on Dr. Kanthack’s water practices, through the power stations, to freeway construction:

“Our largest area of business is providing geotechnical and environmental services to the globalmining industry. We are committed to sustainable mining that recognizes social, environmental andeconomic responsibilities. Our specialized expertise related to tailings and mine waste management,waste characterization, heap leach pads, rock mechanics, water management and environmentalservices is directly relevant to upholding these principles.”

10 SRK

The story of SRK is large. Here we can only pause to recall that SRK was the breeding ground ofmuch slimes dam engineering with Oskar Steffen as the godfather of the group.

11 THE POLITICAL CLIMATE

In the early days, the clients for consultants providing slimes dam design services were the Johannes-burg mining companies. The five major South African mining companies were housed in buildingsthat clustered in the south-west part of the city; the second suite of SRK offices were but a shortwalk from all. It was thrill to walk fast as could be to keep up with Oscar Steffen as we marchedoff to yet another meeting with the mining group’s head-office chief consulting engineer.

But before you could enter, you had to pass through security. Those were the days of extremesecurity fear in South Africa and the mining houses were well equipped to control security. Namesand affiliations taken slowly and solemnly recorded in a black book; drawings would be unrolledfor inspection; we never bothered with brief cases, for they would be inspected via a completeunpacking; and then the fateful call to the person you had come to see. They had to come down tothe foyer to collect you and march you up to the allotted conference room.

This pervasive atmosphere of fear and insecurity borne of politics and segregationism inducedmany of South Africa’s slimes dam’s engineers to emigrate and spread their skill through the world.Here follows sketches of some of them.

62

12 ANDY ROBERTSON

Andy Robertson was also one of the founders of SRK. His expertise was rock slope stability andfoundations. When he decided, at the insistence of his wife, to go to Vancouver, he decided tobecome an expert in slimes dams, shrewdly judging that he could use that discipline as an entre tothe NorthAmerican market. He arranged a Saturday flight for Jack Caldwell and Professor Jennings.We flew for hour after hour over the slimes dams of the Witwatersrand as Andy questioned and ProfJennings and Jack answered, explained, elucidated, and in a day taught Andy all we jointly knewof slimes dams. That we taught well is attested to by the fact that today Andy is never at home—healways somewhere else sitting on yet another peer review group looking at yet another slimes damfrom Canada to South America.

13 METAGO

Metago is now a successful consulting company with offices in Perth and Johannesburg. Its guidinglight is Gordon McPhail, one of the authors of this paper. Gordon joined SRK at the urging of JackCaldwell, also an author of this paper. We worked together until Jack left to go to Tucson at AndyRobertson behest. Gordon took over and flourished on the platinum and gypsum slimes dams thatwere a mainstay of SRK work. Growing professionally powerful, he left SRK and started Metagowith partners. But he too grew weary of the South African political scene and, when businessopportunities that enabled the expansion of Metago into Australia emerged, he relocated to Perth toget those operations going. Now he travels far and wide and most often to Namibia to the uraniumslimes dams that he has looked after for the last 25 years.

14 IAN HUTCHINSON

Today Ian Hutchinson runs Strategic Engineering and Science (SES) from luxurious offices inIrvine, California. Ian got his start in South Africa dealing with hydrology. But soon he left forToronto. Andy Robertson pulled him from Acres, an international consulting company, to run theDenver offices of SRK. It was there that he and Jack Caldwell got involved designing a new tailingsimpoundment for the McLoughlin mine in California. We walked the golden hills and reveled inthe blue lake. Ian fell in love with California and moved to Laguna Niguel where he wrote anotherof those great works on mine waste disposal facilities, namely Mine Waste Management. It is stillon our shelves and we still consult it, particularly when we work with Ian on the closure of theRoyal Mountain King mine and its waste disposal facilities in central California.

15 ROB DOREY

Rob Dorey now runs Dorey Associates. He started in SRK in Johannesburg as a young and brilliantyoung engineer from Imperial College. He lectured us on evaporation from the surface of tailingssurfaces. Then he moved to Vancouver, and hence to Denver where for many years, he managedSRK. The call of independence took him to his own company and he still does great work.

16 TONY CREWS

Tony Crews started in South Africa in slimes dams as a young man straight out of the university.He grew bored and went sailing the world for so many years we lost track of him and forgot abouthim. One day out of the blue he arrived at the door of our house in Levenworth, Washington,where we (the first author of this paper) lived as we designed-as-we-built the Cannon Mine tailingsimpoundment. Tony needed a job, for all he had was two changes of clothing. He took a job, workedfor two years, and took off sailing again. But he returned, married, divorced, and now runs his ownconsulting company from Reno, Nevada, servicing mining clients in the United States and SouthAmerica,

63

17 BRIAN WRENCH

Also in Australia is Brian Wrench. He started in South African tailings when he took over theRichards Bay gypsum dam for SRK. He completed a PhD thesis on the properties of those tailingsand how to close the dam for possible use as an industrial area. This has never been done, althoughwe remain convinced it can be using the technologies developed worldwide. Time and economicswill find a way yet.

18 THURBERS

Thurbers is one of the old, respected geotechnical consulting companies. In its office in Calgaryis Jeremy Boswell. He learnt the trade of tailings with Fraser F Alexander in South Africa. Nowhe is consulting to the oil sands industry helping them solve the myriad problems that characterizetailings deposition in cold climates. He relies on the skills and knowledge gained in the hot, dryclimes of South Africa.

19 GEOFF BLIGHT

Finally we must pay our respects to Geoff Blight whose book started this paper. He supervisedGordon McPhail and his PhD thesis on beach formation on tailings dams, and encouraged himwhen the going was bleak. He helped Jack Caldwell on projects ranging from control of erosionon slimes dams around Johannesburg to the design of a 1,000-ft high embankment in northeastWashington, never built because the market for molybdenum crashed. He has been a guide andhelp and inspiration to both of us. We revere him and know him as the most brilliant of all SouthAfrican engineers who have turned their minds and attention to slimes dams. He has led the way,and we are proud to own his new book.

20 CONCLUSIONS

We write of but a few of the many South African people who have done the work that is distilledinto technical and engineering advances in South African slimes dams, tailings impoundments, ormine waste storage facilities, as you may choose to term them. We have written of only those weknow and recall. But we acknowledge them all. Theirs was and is a singular challenge and success.There have been spectacular failures. There are spectacular successes. They have changed slimesdam design and operation from the instinctive genius of Fraser Alexander to a discipline based onthe strictest scientific principles. We wish only that we could celebrate all of them and all theirstruggles and achievements. We cannot. So we encourage other to follow this stumbling start andfully document their histories.

64


Unique geosynthetic liner system for uranium mill tailings disposal

G.T. CorcoranGeosyntec Consultants, San Diego, California, USA

H.R. RobertsDenison Mines (USA) Corp., Denver, Colorado, USA

ABSTRACT: The design and construction of the Cell 4A and 4B tailings disposal cells at theDenison Mines (USA) Corp. (DMC) White Mesa Mill facility in Blanding, Utah includes the useof a double liner system to contain process liquids and ultimately uranium/vanadium mill tailingswaste materials. The double liner system is comprised of two high density polyethylene (HDPE)geomembranes and a Geosynthetic Clay Liner (GCL). The effectiveness of the composite linersystem was demonstrated through laboratory testing of the hydraulic conductivity properties ofthe GCL materials when permeated with very low pH liquids. In addition, a geosynthetic stripdrainage material was used for the slimes drain installed overlying the primary geomembrane toprovide drainage of the hydraulically placed tailings.

This paper provides a description of the waste materials, design of the liner and slimes drainsystems, and construction of the Cell containment system elements.

1 INTRODUCTION

The White Mesa Mill is the only actively operating uranium and vanadium mill in the UnitedStates and has been maintained in an operational state for over 30 years. The mill processes oresmined in Utah, Colorado, and Arizona using an acid leach process to extract uranium, and in somecases vanadium, from the ore. Lined cells at the facility accept process liquids, waste tailings, andby-products associated with the processing operations. Process liquids are typically acidic with a pHgenerally between 1 and 2. Waste tailings are comprised of ore that is ground to a maximum grainsize of approximately 28 Mesh (US #30 Sieve) (0.6 millimeters (mm) (0.023 inches)), resulting ina fine sand and silt material.

Cells 4A and 4B are approximately 16 hectares (40 acres) with a maximum depth of 12 meters(40 feet) with side slopes as steep as 2H:1V (Horizontal to Vertical). Cells are initially utilizedfor storage and evaporation of process liquids. Waste tailings are hydraulically placed in the cells,below process liquid levels, until a beach is developed, at which time the placement piping is movedto a new location within the cell. Once the cell is full of waste solids (i.e. no higher than the exteriorberm height), the free solution is evaporated or pumped to another cell, and the slimes drain systemis actively pumped to remove free liquids from the tailings solids prior to reclamation.

2 LINER SYSTEM DESIGN

The liner system for Cells 4A and 4B was designed to provide a Cell for disposal of by-productsfrom the onsite processing operations while protecting the groundwater beneath the site. The linersystem was designed to meet the BestAvailableTechnology requirements of the UAC R317-6, whichrequires that the facility be designed to achieve the maximum reduction of a pollutant achievableby available processes and methods taking into account energy, public health, environmental andeconomic impacts, and other costs. The liner system includes the following primary components,from top to bottom:

• Slimes drain system;• Primary geomembrane liner;

65

Figure 1. Slimes drain header section.

• Leak detection system;• Secondary geomembrane liner; and• Geosynthetic clay liner.

2.1 Liner system

A double liner system was designed and constructed for Cell 4A and 4B. The liner system, for boththe bottom area and side slopes, consists of (from top to bottom):

• Slimes drain system (Cell bottom only);• 1.5 mm (60 mil) thick smooth high density polyethylene (HDPE) geomembrane (Primary Liner);• 7.5 mm (300 mil) thick geonet drainage layer (Leak Detection System);• 1.5 mm (60 mil) thick smooth HDPE geomembrane (Secondary Composite Liner);• Geosynthetic clay liner (GCL) (Secondary Composite Liner); and• Prepared subgrade.

2.2 Slimes drain system

A slimes drain system was constructed on top of the primary geomembrane liner in the bottom ofthe cell to facilitate dewatering of the tailings prior to final reclamation of the cell. The slimes drainsystem is not a continuous blanket drainage system, rather a system of discreet collection headersand laterals. The slimes drain system consists of a single perforated 102 mm (4 inch) diameterschedule 40 polyvinyl chloride (PVC) header pipe, drainage aggregate, cushion geotextile, andfilter geotextile as shown in Figure 1.

In addition, a strip composite consisting of a 25 mm (1 inch) thick by 305 mm (12 inch) wideHDPE core and polypropylene geotextile filter wrap was used as slimes drain laterals. Since thecell will initially be filled with process liquids, the strip composite is continuously covered withconcrete sand filled sand bags as ballast and to act as additional filtration between the waste tailingsmaterial and the strip composite core so as to minimize the potential for clogging. Given the acidicnature of the process liquids that will contact the slimes drain components, aggregate and sandmaterials are comprised of natural materials exhibiting a carbonate content loss of no more than10 percent by weight, and manufactured materials are comprised of polymers that will withstandthe acidic environment. The laterals convey liquids to the header pipe and on to the sump forremoval from the cell. The laterals are spaced at 15 m (50 ft) intervals and are a unique use of stripcomposite, more commonly known as edge drain, that is more cost effective than the use of moretraditional pipe, aggregate, and filter geotextile components. The strip composite is provided in50 m (165 ft) long rolls that are easily installed and joined to create the slimes drain laterals. Theslimes drain lateral design is shown in Figure 2.

66

Figure 2. Slimes drain lateral section.

The Slimes Drain sump includes a side slope riser pipe to allow installation of a submersiblepump for removal of liquids in the sump.

2.3 Primary liner

The primary liner consists of an exposed smooth 1.5 mm (60-mil) thick HDPE geomembrane witha white surface that limits geomembrane movement and the creation of wrinkles due to thermalexpansion and contraction from temperature variations. HDPE geomembrane was selected due toits high resistance to chemical degradation and ability to retain durability in an acidic environment.

Splash pads were installed at several locations that allow for operations personnel to placetemporary pipes for filling of the cells. Splash pads consist of a separate textured 1.5 mm (60-mil)thick HDPE geomembrane, black side up, installed on top of the primary geomembrane liner atspecific locations.

2.4 Leak detection system

The leak detection system (LDS) underlies the primary liner and is designed to collect potentialleakage through the primary liner and convey the liquid to the sump for manual detection throughmonitoring of sump levels. The LDS consists of a 7.5 mm (300-mil) thick geonet and a network ofgravel trenches throughout the bottom of Cell 4B. The trenches contain a 102 mm (4-inch) diameterperforated schedule 40 PVC pipe, drainage aggregate, and a cushion geotextile, which drain to asump for detection and removal.

The Action Leakage Rate (ALR) was calculated for the LDS in accordance with Part 254.302of the USEPA Code of Federal Regulations. Based on the ALR calculation, the liquid head on thesecondary liner does not exceed 0.15 mm (0.006 inches), well below the required maximum limitof 305 mm (12 inches).

The LDS sump includes a side slope riser pipe and submersible pump to allow for removal ofliquids in the LDS sump. Flows into the LDS are monitored on a continuous, real time basis, andrecorded for regulatory compliance purposes.

2.5 Secondary composite liner system

The primary purpose of the secondary liner is to provide a flow barrier so that potential leakagethrough the primary liner will collect on top of the secondary liner, then flow through the LDS to

67

the LDS sump for removal. The secondary liner also provides an added hydraulic barrier againstleakage to the subsurface soils and groundwater. The secondary liner consists of a composite linerthat is comprised of a 1.5 mm (60-mil) thick HDPE geomembrane overlying a GCL.

2.5.1 Secondary geomembrane linerThe geomembrane component of the secondary composite liner system consists of a smooth 1.5 mm(60-mil) thick HDPE geomembrane that meets the same criteria as the primary liner geomembrane.

2.5.2 Secondary GCL linerThe GCL component of the secondary composite liner system consists of bentonite sandwichedbetween two geotextile layers that are subsequently needle-punched together to form a singlecomposite hydraulic barrier material. Although the GCL is used as an element of the secondarycomposite liner system and is not expected to be in contact with process liquids (i.e., the processliquids have to migrate through defects, if any, in the primary liner and then build up enough head todrive the process liquids through the secondary liner into the GCL), a testing program was devised todemonstrate that the GCL would exhibit low hydraulic conductivity (permeability) when permeatedwith a low pH liquid similar to the process liquids anticipated to be contained by the liner system.

Testing of the GCL consisted of permeating GCL samples with varying degrees of initial moisturecontent. Moisture content was established in the laboratory using deionized water to achieve 50%,75%, 100%, and 140% moisture content. Each specimen was then immediately permeated with aliquid with a pH of 1.0 (pH established using hydrochloric acid) under a normal stress of 34.5 kPa(5psi). Testing was performed in accordance with ASTM D 6766, Scenario 1.

The results of the hydraulic conductivity testing are as follows:

Approximate Approximate ApproximateApproximate permeability permeability permeability

Percent initial after one half after one after twohydration of permeability pore volume pore volume pore volumesGCL sample (cm/sec) (cm/sec) (cm/sec) (cm/sec)

50% 1.0 × 10−9 2.0 × 10−9 1.2 × 10−8 3.0 × 10−8

75% 6.0 × 10−10 3.0 × 10−9 9.0 × 10−9 2.5 × 10−8

100% 1.2 × 10−9 4.5 × 10−9 1.0 × 10−8 3.5 × 10−8

140% 8.0 × 10−10 4.0 × 10−9 1.2 × 10−8 4.5 × 10−8

Based on the test results and the ALR calculation of the head on the secondary liner, pore volumetravel time through a GCL pre-hydrated to a moisture content of 50% was estimated to be morethan 150 years for the first pore volume of permeant, which is well beyond the time when the cellwill be drained of free liquids. Based on this analysis, the regulatory agency agreed that a minimummoisture content of 50% should be achieved in the GCL installed for this project.

Initially, a test pad was constructed to demonstrate that the GCL would hydroscopically adsorbwater from the underlying subgrade, which exhibited a moisture content of 12.3%. The test padwas comprised of a 3 m (10 ft) by 4.6 m (15 ft) GCL panel overlain by a single layer of 1.5 mmHDPE geomembrane, white side facing up, anchored on all four sides with a 150 mm (6 inch)deep anchor trench. The initial GCL moisture content was tested and found to be approximately14.0% (very low relative humidity and high temperatures at the site reduced the as-delivered GCLmoisture content while stored on-site). After one week of installation in the test pad the GCLmoisture content increased to 22.8% moisture content, and after two weeks the GCL moisturecontent had increased to 26.3%. Literature suggests that in most cases, soils adjacent to GCLswill readily give up moisture to the stronger suction characteristics of the bentonite component ofthe GCL, resulting in more than 100% moisture content in approximately 2 weeks. In this case,the test pad size and site weather conditions likely contributed to much lower moisture contentdevelopment in the GCL. Based on the failure of the demonstration, field hydration of the GCLprior to installing the overlying secondary liner was determined to be the best approach to achievingthe project design goals.

68

Figure 3. Water cannons being used to hydrate GCL.

3 CONSTRUCTION

Construction of Cells 4A and 4B involved standard construction methods for earthwork, includingsoil and rock excavation, engineered fill, and subgrade preparation. Subgrade preparation includedmoisture conditioning of the soil surface in addition to compaction and elimination of protrusionsover 12 mm (1/2 inch).

Geosynthetic material installation was complicated by the need to hydrate the GCL prior toinstallation of the overlying secondary liner. Hydration was performed using water cannons andhand held hoses to distribute water on the surface of the GCL (Figure 3). Initial hydration activ-ities were closely monitored, and frequent field testing of moisture and application rates wereperformed to ensure that the minimum moisture content of 50% was being achieved in the GCLprior to deployment of the secondary liner. Shallow pans were used to capture spray application ofwater, which was then compared to calculated values of water theoretically needed to achieve theappropriate moisture content from the installed dry moisture content. In addition, moisture contentwas monitored in the field in accordance with ASTM D 4643 (Microwave method of determiningmoisture content) to quickly evaluate the moisture condition of the GCL prior to installation of thesecondary liner. Field moisture conditioning of the GCL typically took over 2 hours to allow thewater to adsorb, and in the case of side slopes, two applications of water were often required toattain the minimum moisture content.

Once the moisture content in the GCL exceeded 50%, secondary geomembrane was deployed andseamed using dual track fusion welding equipment. It was found that standard HDPE geomembraneseaming methods could be employed to achieve acceptable seams, even with the underlying hydratedGCL. Seams were tested at a frequency of one per 152 m (500 ft) with no failures (Figure 4).Installation of the leak detection system geonet and primary geomembrane followed standardindustry practice.

Slimes drain installation involved the placement of over 30,000 sand bags overlying the stripcomposite in each cell. Slimes drain header pipe and aggregate were installed using conventionalmethods; however, getting the construction materials out to the header location involved use of lowground pressure equipment operating on top of the primary geomembrane, which resulted in theneed for some minor repairs.

4 CONCLUSIONS

The unique application of geosynthetic materials for the design and construction of the White MesaMill Cell 4A and 4B liner and slimes drain systems resulted in an effective and protective liner

69

Figure 4. Secondary geomembrane destructive and GCL hydration sampling.

Figure 5. Cell 4A in use. Waste tailings beach in foreground, process liquids in background. Black patches onfar slope are splash pads comprised of a sacrificial HDPE geomembrane installed black side up for protectionof the primary liner and to allow operational access.

system that complies with the regulatory requirements while being very cost effective. Since thestart of filling of Cell 4A, the liner system has performed well with minimal leakage (significantlybelow the ALR) detected in the leak detection system. Figure 5 shows the completed Cell 4A inuse with a waste tailings beach above the level of the process liquids. Cell 4B is currently underconstruction and is expected to be in service during the first quarter of 2011.

70


Optimizing tailings deposition concentration at MineraYanacocha, Peru

M. Keevy & R. CookePaterson & Cooke, Denver, USA

ABSTRACT: Minera Yanacocha commissioned a new gold mill and processing plant in 2008.The new facilities included a tailings pumping system from the plant to a tailings storage facility(TSF) located within a heap leach facility. The location of the TSF resulted in a requirement forreducing the amount of water sent to the facility, to reduce the infiltration of water from the tailingsinto the leach pad.

Preliminary designs for the tailings pumping system included piston diaphragm pumps fordelivering high solids concentration tailings to the facility. Paterson & Cooke were contracted tocarry out the detailed design of the pumping system. The initial step was the evaluation of thedesign concentration and pumping requirements, with respect to the requirements at the TSF, toevaluate the optimum pumping concentration.

This paper presents the trade-off analysis carried out in the design of the Yanacocha pumpingsystem, as well as providing details of the subsequent design and implementation.

1 INTRODUCTION

Minera Yanacocha, located in Peru, is one of the world’s largest gold mines. In 2008 a new goldmill and processing plant was commissioned for the treatment of high grade ores. Prior to this theoperation was a heap leach operation. Part of this new plant is the tailings pumping system thattransfers processed slurry from the plant to the deposition site.

The deposition site is an impoundment facility constructed within an active heap leach pad. Tomaintain the integrity of the facility it is important that the deposit does not significantly increasethe saturated zones within the heap leach pad. The design was therefore focused initially focusedon the deposition of thickened tailings at the storage facility.

Preliminary designs for the system were based on operating the tailings stream at a concentrationof 75%m. This system would require piston diaphragm pumps to accommodate the high pressuresand have a high power consumption.

The impoundment facility design was carried out by Knight Piésold. During the design ofthe impoundment facility, Knight Piésold determined that the facility could be operated at lowerconcentrations than the original 75% without impacting the leach pad stability, due to the waterremoval systems incorporated into the design. Paterson & Cooke’s aim was therefore to determinethe optimum operating concentration for the system.

The geotechnical aspects of this design will not be discussed in this paper, but can be found inKerr et al., 2007.

2 DESIGN BASIS

2.1 Test work results

The tailings particle size distribution is illustrated in Figure 1.Rheology test work was carried out on a sample of the mine tailings. As shown in Figure 2, the

results indicated that the tailings rheology starts to increase at a concentration of approximately60%m and exhibits a sharp upswing between 70%m and 72%m.

71

Figure 1. Tailings particle size distribution.

Figure 2. Slurry rheology.

The material properties were used to estimate the settling velocity and pipeline transporta-tion requirements. The slurry remains a heterogeneous settling mixture even at concentra-tions above 70%m. If the material does settle it forms a compact bed that is difficult tore-suspend.

72

Figure 3. Absorbed power for varying slurry concentration.

2.2 Pump types

The preliminary design made use of piston diaphragm pumps due to the high pumping pressurerequirements. Due to the project schedule there was a desire to use centrifugal pumps instead toreduce the equipment lead time and avoid potentially delaying the plant start-up date. The lowercapital cost of a centrifugal pump system was also attractive.

The desire to use centrifugal pumps further increased the drive to find an optimum concentrationwith potentially lower pumping pressures, as typical centrifugal pumps are limited to a maximumdischarge pressure of 4 MPa.

3 SOLIDS CONCENTRATION ANALYSIS

To facilitate the selection of an appropriate solids concentration, the system absorbed power wasselected as the comparison basis. The pump station capital, energy and maintenance costs arerelated to the installed power.

3.1 Theoretical analysis

The minimum pressure and hence the power absorbed in any slurry system can be calculated ifthe rheological and transport properties are known. To achieve the lowest power consumption thesystem must operate at the lowest velocity that satisfies the transportation velocity requirements.

To determine the shape of the curve the calculations can be performed with an ideal combinationof pipeline size and concentration (i.e. the concentration is selected and a non standard pipe internaldiameter calculated to meet the transportation velocity requirements).

The typical shape of the absorbed power graph is shown in Figure 3. As the concentrationincreases the power requirements decrease. In this low concentration range the increase in rheologydoes not significantly affect the friction losses, but the increase in concentration reduces the flowrate and hence the absorbed power.

As the concentration increases further it reaches the point where the impact of changing rheologybalances the reduction in flow rate and the power consumption starts to increase at an acceleratingrate due to the exponential relationship between concentration and rheological properties.

73

Figure 4. Yanacocha absorbed power for varying slurry concentration.

The shape of this curve is dependent on the tonnage, slurry rheology and the pipelineconfiguration for the system.

3.2 Detailed analysis

The analysis was carried out for Yanacocha using the basis described above. In addition, standardpipe sizes and wall thicknesses where selected to relate the calculations to a real world system. Theresults are shown Figure 4. The use of standard pipe sizes results in a saw tooth pattern, due to thechanging velocity from one pipe size to another, but the overall trend is still visible.

The analysis indicates that the lowest power consumption can be achieved using either a 350 NB(14′′) pipeline at a concentration of 60%m, or a 300 NB (12′′) pipeline at a concentration of 69%m.Additionally operating at 69%m results in 72 m3/h of additional water reporting to the impoundmentfacility compared to 75%m, while operating at 60%m results in 207 m3/h of additional water.

3.3 Pump selection

A similar analysis was carried out to determine the pump discharge pressure for the system. Theresults are shown in Figure 5. In addition to the nominal tonnage used in determining the normalpower consumption and velocity, the pressure calculation also includes the peak design tonnagefor the system. There is a wide operating range for the system, resulting in a significant increasein pressure from the nominal to the peak duty.

Centrifugal pumps can be used up to a concentration of 71%m for the nominal case, but oncethe peak tonnage is considered it can be seen that operation should be limited to 69%m; beyondthat piston diaphragm pumps will be required.

3.4 Discussion

For this project a design concentration of 69% was selected. The additional volume of water sentto the impoundment facility was discussed with Knight Piésold and found to have no significantimpact on the impoundment stability.

The selection of this design concentration also enables the use of centrifugal pumps.

74

Figure 5. Yanacocha pump discharge pressure for varying slurry concentration.

The drive for high concentrations for this system initially lead to high pumping system capitaland operating costs. Further analysis of the impoundment facility indicated that additional watercould be tolerated. This provided to opportunity to reduce costs without negatively impactingperformance.

4 IMPLEMENTATION

The system has been operating successfully since commissioning during 2008.

4.1 Pump selection

Two trains of four Warman 8/6 AHP pumps were installed for the system, each with a 250 kWmotor with a variable frequency drive. More detailed on the specifics of the pump selection havebeen presented in Keevy & Hackney, 2007.

4.2 Density control

To ensure reliable operation within the concentration range the system includes two density controlloops. The thickener underflow fed to the system feed tank is measured and dilution water added toachieve a concentration of 69%m in the tank. However, this system does not give direct adjustmentto the pump station discharge as it monitors the feed and adjusting the water into the tank will alsohave a slow response time.

A second density measurement is taken at the pump station discharge. If this density is too highdilution water is injected into the suction of the pump train to correct it. This system providesimmediate correction if the density in the tank is too high and also alerts the operator to check thecalibration of the density meters.

Finally, a flushing system is included with a water booster pump feeding the inlet of the firststage pump. Flush is required after each shut down to prevent a bed forming in the pipeline, butcan also be used to pump slurry out the pipeline in sections, initially to a drain area located at thepipeline low point, and then to the first discharge point on the storage facility.

75

Figure 6. Yanacocha pump station.

5 CONCLUSIONS

By investigating the absorbed power of the system P&C were able to optimize the operatingconcentration for the system, resulting in capital and operating cost savings. The system deliversthickened tailings to achieve the reduced water delivery required, while avoiding excessive costpremiums for achieving this.

The pumping system is designed for the optimized concentration and the design includes controlmeasures to ensure that these limits are not exceeded.

REFERENCES

Keevy, M.B. & Hackney, K. 2007. Pump station drive selection case study. Hydrotransport 17, The 17thInternational Conference on the Hydraulic Transport of Solids, Cape Town, 7–11 May 2007. Johannesburg:The Southern African Institute of Mining and Metallurgy.

Kerr, T.F., Duryea, P.D., Grobbelaar, W. & Hackney, K. 2007. Design of a thickened mill sands managementsystem within a heap leach pad. Paste 2007, Tenth International Seminar on Paste and Thickened Tailings,Perth, 13–15 March 2007, Australian Centre for Geomechanics.

76

Geotechnical considerations


Peak and critical-state shear strength of mine waste rock

Z. Fox & J. Antonio H. CarraroDepartment of Civil and Environmental Engineering, Colorado State University,Fort Collins, Colorado, USA

ABSTRACT: Proper characterization of the peak and critical-state shear strength parametersrequired for geotechnical analyses involving geomaterials with large particle sizes such as minewaste rock, rockfill and coarse aggregates is challenging. This paper summarizes the results ofprevious studies related to large-scale triaxial testing and analysis of such geomaterials. Relevantexperimental issues such as appropriate sample-size ratio selection, membrane penetration andparticle breakage are discussed. The main steps required to perform a rigorous analysis of triaxialtest results are outlined that avoid usual misinterpretations associated with conventional analysis oftriaxial tests. A methodology integrating a robust experimental protocol with a rigorous theoreticalframework is presented. The proposed methodology can be used to predict the shear strength ofmine waste rock, rockfill, coarse aggregates and other types of uncemented geomaterials withlarge particle sizes under combinations of density and mean effective stress that are relevant tomost geotechnical and mining applications.

1 INTRODUCTION

1.1 Sample-size ratio

The triaxial apparatus is one of the most widely used devices to evaluate the shear strength andstiffness of geomaterials. However, geomaterials used in a wide range of geotechnical and miningapplications have particle sizes much larger than the maximum particle size (Dmax) tested in aconventional triaxial apparatus with specimen diameter (d) ranging from 50–70 mm. In triaxialtesting, the sample-size ratio can be defined as the ratio (d/Dmax) of specimen diameter to maximumparticle size (Vallerga et al., 1957, Marachi 1969, Indraratna 1993). The use of sample-size ratiossmaller than five has been shown to introduce testing errors due to particle size effects (Marsal1969, Leslie 1969, Nitchiporovitch et al., 1969) especially when more than 30% of the sample isretained on the largest sieve size (Marachi 1969). Use of a minimum sample-size ratio of six, asrecommended by ASTM D4767, leads to a maximum particle size of 12 mm for a 70-mm-diameterspecimen. A maximum particle size of 12 mm may represent a small portion of geomaterialswith large particle sizes such as mine waste rock, rockfill, and coarse aggregates. This leads touncertainties associated with the assignment of shear strength and stiffness parameters for suchmaterials during modeling and design.

Limitations of the conventional triaxial apparatus with regard to maximum particle size werefirst addressed by Holtz and Gibbs (1956) who used a large-scale triaxial (LSTX) apparatus totest gravelly soils containing particle sizes up to 75 mm. Results of 183 tests indicated that frictionangles increased with increasing particle size and angularity as well as with increasing gravelcontent up to around 50–60%. Above this threshold gravel content, the shear strength and densityof the specimens decreased with increasing gravel content.

Marachi et al. (1972) investigated particle-size effects on shear strength by testing parallel gra-dations of materials of similar mineralogical composition and geologic history with specimendiameters equal to 70, 305 and 914 mm using a constant sample-size ratio of six. As both specimendiameter and maximum particle size decreased, the measured friction angles increased and volumet-ric strains became more dilative (or less contractive) in triaxial compression (Fig. 1). All tests shownin Figure 1 were isotropically consolidated to the same mean effective stress [p′ = (σ ′

1 + 2σ ′3)/3]

79

Figure 1. Effect of specimen diameter and maximum particle size on the drained triaxial compression responseof quarried argillite materials with parallel grain size distributions used in the Oroville Dam (modified afterMarachi et al., 1972).

equal to 210 kPa, where σ ′1 and σ ′

3 are the effective major and minor principal stresses, respec-tively. The corresponding volumetric strain [εp = ε1 + 2ε3] in triaxial (or, perhaps more rigorously,axi-symmetric) compression shown in Figure 1 can then be consistently defined and associatedwith changes of the octahedral mean stress invariant p′, where ε1 and ε3 are the major and minorprincipal strains, respectively. Marachi et al. (1972) also noted relatively minor effects of specimensize on volumetric strains during isotropic compression for constant initial relative densities. Useof conventional triaxial testing equipment with a specimen diameter of 70 mm led to an overesti-mation of the peak friction angle (φp) of about 3–4◦ (6–8%), as it might be deduced from the datashown in Figure 1.

Differences in the φp of geomaterials with large particle sizes, relative to the corresponding φpof similar geomaterials with smaller particle sizes, may be pronounced (Fig. 2). These differencesmake the LSTX apparatus an important and necessary tool to characterize the shear strength andstiffness parameters of geomaterials with large particle sizes such as those used to construct damsand other geotechnical structures.

In many cases, characterization of the true field-scale shear strength parameters may not bepossible, even when using the largest triaxial apparatus available, due to the limited specimendiameters and maximum particle sizes that can be practically used during testing. In order toassess this limitation, the three different geomaterials with parallel grain-size distributions testedby Marachi et al. (1972) may be used to estimate the potential variation in φp values resultingfrom testing samples with maximum particle sizes of at least 12 mm taken from original field-scale samples with maximum particle sizes equal to 150 mm (Fig. 2). It should be noted thatparallel grain-size distributions are identical in shape, but shift along the grain size axis in aconventional grain-size distribution diagram. As shown in Figure 2, the difference in φp valuesmeasured using conventional 70-mm-diameter specimens with 12-mm maximum particle sizes or914-mm-diameter specimens with 150-mm maximum particle sizes may be as large as 5◦ (∼11%).If 150-mm-diameter specimens with 25-mm maximum particle sizes were prepared using the sameparallel gradation criterion and tested at the same p′ = 210 kPa and sample-size ratio of six usedby Marachi et al. (1972), this error would be reduced to about 2–3◦ (∼5%).

80

Figure 2. Effect of maximum particle size on the peak friction angle of three different geomaterials withparallel gradations (modified after Marachi et al., 1972).

1.2 Membrane penetration

Another aspect making LSTX testing more challenging than conventional triaxial testing involvesproperly accounting for penetration of the rubber membrane into the specimen during consolidationand shearing. Large void spaces between particles commonly exist at the lateral boundary ofspecimens containing large particle sizes, regardless of their density. Drainage of the pore wateroriginally filling these voids due to the deformation of the membrane into the specimen voidswith increasing p′ during consolidation may result in measurement of artificially large εp values.Results of numerous studies on this topic have identified particle size, specimen size, confiningpressure, and membrane characteristics as significant factors associated with membrane penetration(Sivathayalan and Vaid 1998, Ansal and Erken 1996, Nicholson et al., 1992, Choi and Ishibashi1992, Kramer et al., 1990, Dendani et al., 1988, Baldi and Nova 1982, Molenkamp and Luger1981, Frydman et al., 1973).

Failure to correct for these artificially high εp values may lead to additional errors. For example,effective axial stress (σ ′

a) errors may arise due to uncorrected εp values as the specimen crosssectional area during consolidation and drained triaxial compression is conventionally updatedbased on the current level of radial strain (εr), which, in turn, may be deduced from current valuesof εp and axial strain (εa) if local axial and radial strain transducers are not used. During drainedtriaxial compression, the effective radial stress (σ ′

r) remains constant, which helps keep the amountof membrane penetration relatively constant once consolidation is finished. However, significanterrors remain in undrained triaxial compression (when �σ ′

r may not be negligible) as pore pressuremeasurements during this stage will be affected by membrane penetration (Ansal and Erken 1996,Molenkamp and Luger 1981).

The total volume of pore water drained out of the specimen due to membrane penetration nor-malized by the initial contact area between the membrane and the lateral surface of the specimenis defined as the unit membrane penetration or unit normalized penetration (Choi and Ishibashi1992, Kramer 1989, Dendani et al. 1988, Baldi and Nova 1984, Frydman et al., 1973). Earlyexperimental methods attempting to quantify this error invoked many assumptions regarding truespecimen deformation relative to the measured volume changes in the triaxial apparatus. While themagnitude of unit membrane penetration estimated by different early studies may be pronounced

81

Figure 3. Effects of maximum particle size and mean effective stress during consolidation on unit membranepenetration for five materials with different maximum particle sizes (modified after Dendani et al., 1988).

(Choi and Ishibashi 1992), many studies observed a linear relationship between the logarithm ofp′ and unit membrane penetration (Dendani et al., 1988, Frydman et al., 1973).

Dendani et al. (1988) discussed the major effect of particle size on unit membrane penetration(Fig. 3) as well as inaccuracies associated with assumptions of linear relationships between unitmembrane penetration and logarithmic changes in p′ during consolidation with increased particlesize and/or confining pressure.

Other analytical methods have represented the lateral surface of a triaxial specimen as an arrayof spheres of varying diameters related to the grain size distribution of an actual soil (Sivathayalanand Vaid 1998, Ansal and Erken 1996, Nicholson et al., 1992, Kramer et al., 1990, Molenkamp1981, Baldi and Nova 1982). These studies showed similar observations of linear semi-logarithmicplots of unit membrane penetration with the logarithm of effective stress for sands and glass beadsalike. Various empirical relations have been presented to predict membrane penetration based onexperiments taking these more influential factors into account. However, large differences in theproposed corrections remained, especially for grain sizes larger than 3 mm, until Nicholson et al.(1993) showed the nominal particle size D20 (in mm) is the most accurate parameter to estimatethe stress-normalized unit membrane penetration (Sσ ′3) after isotropic compression (which is, inthis format, normalized by the log of σ ′

3). Based on experimental data including specimens testedin a 300-mm-diameter LSTX apparatus equipped with internal radial transducers, Nicholson et al.(1993) proposed Sσ ′3 be estimated according to:

The LSTX specimens tested by Nicholson et al. (1993) were subjected to isotropic compressionp′ levels as high as 1200 kPa to encompass the usual testing ranges for which membrane-complianceeffects may be of concern. Data used to develop Eq. 1 is shown in Figure 4.

The preferred method to evaluate membrane penetration relies upon direct measurement of thetrue radial strain in the specimen. In the absence of local strain transducers, Eq.1 is superior to allother methods presented in the literature to evaluate membrane penetration during LSTX testingof geomaterials with large particle sizes.

82

Figure 4. Effect of nominal particle size D20 on the stress-normalized unit membrane penetration (modifiedafter Nicholson et al., 1993).

1.3 Dilatancy and critical state

Reynolds (1885) coined the term dilatancy after observing the effect of density on the volumetricresponse of sands during shearing. For p′ levels associated with most typical geotechnical appli-cations (10 to 500 kPa), loose sands contract with increasing shear deformation until critical stateis reached at constant shear stress and constant volume (Schofield and Wroth 1968). On the otherhand, dense sands dilate and mobilize peak shear stress before critical state is reached at largestrains. Dilation, which is primarily affected by the soil state (density and effective stress), is themain factor responsible for the curvature of the failure envelope of uncemented geomaterials usedin most geotechnical applications.

Leps (1970) reviewed the literature regarding the shear strength of rockfill and compiled a largeamount of LSTX results to show the linear dependence of φp on the logarithm of the “normal stressacross the failure plane” for sands and rockfill materials (Fig. 5). From a more rigorous, conceptualstandpoint, φp determined from triaxial tests should be actually related to the peak mean effectivestress p′

p (Bolton 1986).For Ottawa sand, for example, additional lower and upper bounds with different slopes could be

identified and superimposed to the data shown in Figure 5 for relative density (DR) levels equal to0 and 100%, respectively, as φp depends not only on p′

p but also on density. Likewise, additionalupper and lower bounds can be defined for density states varying between the loosest and denseststates possible to be achieved for each material shown in Figure 5.

Unlike dilatancy, the critical state of a geomaterial is conventionally and uniquely related to itsintrinsic characteristics such as particle shape, mineralogy and grain size distribution (Schofieldand Wroth 1968).

1.4 Particle breakage

In the absence of particle breakage, increases in p′ suppress dilation for geomaterials with rela-tively high grain strength. However, particle breakage may also occur during triaxial testing of

83

Figure 5. Typical linear relationship between the peak friction angle and the logarithm of normal stress acrossthe failure plane for various uncemented geomaterials (modified after Leps 1970).

geomaterials, which can further influence their dilatancy response. Marsal (1967) and co-workersdesigned an LSTX apparatus capable of testing cylindrical specimens with diameter and heightequal to 1100 and 2500 mm, respectively, to characterize the material used in the construction ofthe 148-m-high “El Infiernillo” Dam in Mexico. He noted “the most important factor affecting bothshear strength and compressibility is the phenomenon of fragmentation undergone both during uni-form consolidation and during deviator load application.” Marsal also observed increasing particlebreakage with increasing uniformity of the materials. Reduced particle breakage was observed formore well-graded materials upon shearing.

Lee and Farmoohand (1967) used a 70-mm-diameter triaxial apparatus to demonstrate particlebreakage increases with increasing particle size for specimens sheared at the same initial p′ = 8 MPa(Fig. 6). The after-test particle size distributions of samples with larger particle size shown inFigure 6 approach the maximum-density particle size distribution proposed by Fuller andThompson(1907). Particle breakage also increases with increasing p′ and increasing particle angularity formaterials with the same initial particle size distribution (Fig. 7). While the p′ levels shown inFigures 6 and 7 may be too high for most geotechnical applications, these results still provideuseful insights into the effects of particle size, particle angularity and p′ on particle breakage.

Lee (1992) suggested that the reduction of the dilation component of φp with the logarithm of p′should be further normalized by grain tensile strength to take particle crushing into consideration.Using a theory of successive fractal failure of the smallest grain sizes due to the macroscopic stressapplied to the surface of the grain, McDowell et al. (1998) developed a numerical model of crush-able aggregates and compared results of the model with previously completed experimental data.Isotropic compression results show particle size distributions approaching a constant uniformitycoefficient and illustrate the effect of grain strength on the linear semi-logarithmic relationshipbetween φp and p′ normalized by grain tensile strength.

Ueng and Chen (2000) separated the components of φp for two different sands (Fulung Riverand Tamsui River sands) and a decomposed granite previously tested by Miura and O-hara (1979)into the friction angle excluding both particle breakage and dilatancy (φf ), thus equivalent to thecritical-state friction angle (φc), and the friction angle excluding dilatancy and including particle

84

Figure 6. Changes in particle size distribution due to particle breakage of various samples determined beforeand after triaxial testing (modified after Lee and Farhoomand 1967).

Figure 7. Effects of the (a) mean effective stress after consolidation and (b) particle shape on particle breakagefor isotropic and anisotropic conditions (modified after Lee and Farhoomand 1967).

breakage (φfb), to determine the actual contribution of particle breakage (i.e., φfb − φc) on φp.The relative effect of particle breakage on the φp values of the three different materials studied byUeng and Chen always increased with increasing initial p′ (after isotropic compression) used in thetests (Fulung sand results are shown in Fig. 8), whereas the magnitude of particle breakage wasinversely proportional to the grain strength of the material (Ueng and Chen 2000). Indraratna andSalim (2002) followed a procedure similar to that outlined by Ueng and Chen (2000) to evaluatethe amount of particle breakage of latite basalt with maximum particle size of 53 mm using a300-mm-diameter LSTX apparatus. The relative effect of particle breakage on the φp values of thelatite basalt with large particle sizes was of the same order of magnitude as that reported by Uengand Chen (2000) for the Fulung sand, which had strong grains. This effect may be quantified byevaluating the (φfb − φc)/φp ratio for some of the highest p′ levels used in the tests for both theFulung sand and the latite basalt (p′ ≈ 300 to 383 kPa), which yields a (φfb − φc)/φp ratio of about8–9% for both materials. This value is much lower than the typical (φfb − φc)/φp ratio of about16–21% observed for the other two materials with weak grains (Tamsui River sand and densedecomposed granite) studied by Ueng and Chen.

Results from these previous studies suggest the effect of particle breakage on φp is not significantfor typical geomaterials with strong particles tested under density and stress states associated with

85

Figure 8. Effect of mean effective stress (after isotropic compression) on the peak friction angle and particlebreakage of Fulung sand (Ueng and Chen 2000) and Latite basalt (Indraratna and Salim 2002).

typical geotechnical applications (p′ < 300–500 kPa). For geomaterials with low grain strengthand/or subjected to higher p′ levels, the procedure outlined by Ueng and Chen (2000) can be usedto systematically quantify the impact of particle breakage on their shear strength.

2 PROPOSED METHODOLOGY

2.1 Conceptual framework

A rigorous conceptual framework is needed to properly evaluate the shear strength of uncementedgeomaterials with or without large particle sizes. Such analyses should be based on careful consid-eration of both state variables and intrinsic parameters known to significantly affect the mechanicalbehavior of geomaterials. Intrinsic parameters are uniquely defined for a specific geomaterial andremain independent of its current state. On the other hand, stress, density, and fabric representtypical examples of state variables that fundamentally affect geomaterial behavior (Salgado 2008).

For axi-symmetric conditions, such as those associated with triaxial testing, the friction angleof an uncemented geomaterial may be deduced from the Mohr’s circle of stress through:

where σ ′1/σ ′

3 = N = flow number, stress obliquity, or effective principal stress ratio.A stress-dilatancy relationship for plane-strain based on minimum energy assumptions (Rowe

1962, De Josselin de Jong 1976) can be expressed for geomaterials with high grain strength as:

86

Figure 9. Schematic representation of the critical state line and various peak failure envelopes for a hypo-thetical geomaterial (with Q = 10, R = 1 and φc = 30◦) for various combinations of relative density and meaneffective stress.

where Nc = flow number at critical state; M = dilatancy number = 1 − dεp/dε1, with dε1 anddεp = major principal strain and volumetric strain increments, respectively; Nc and M can beexpressed in terms of φc and the dilatancy angle (ψ) according to:

The value of ψ approaches a maximum at the maximum dilatancy rate (Schofield and Wroth1968) and can be deduced for axisymmetric conditions from the Mohr’s circle of strain as:

where dε3 = minor principal strain increment.Dilatancy is suppressed with increasing p′ (Leps 1970, Bolton 1986), as discussed earlier (Fig. 5).

Conversely, dilatancy increases with increasing density. Bolton (1986) accounted for the effects ofstate variables (DR and p′

p) and intrinsic parameters (Q, R and φc) on the dilatancy of uncementedsands in axi-symmetric (triaxial) compression through:

87

Figure 10. New large-scale triaxial apparatus at Colorado State University for testing of geomaterials withlarge particle sizes (specimen diameter and height equal to 150 and 300 mm, respectively).

Figure 11. Area, membrane and shear failure mechanism corrections for triaxial testing (after La Rochelleet al.).

where pA = reference stress (=100 kPa, for p′p values given in kPa); Q, R and φc are intrinsic

parameters that can be determined for various geomaterials such as clean sands (Bolton 1986),nonplastic silty sands (Salgado et al., 2000), and mixtures of sands with either plastic or nonplasticfines (Carraro et al., 2009). By performing a series of LSTX tests in a systematic manner andunder controlled levels of state variables, these intrinsic parameters can also be determined for anyuncemented geomaterial with large particle sizes to allow prediction of the peak shear strength (orφp) of the material under any state. Figure 9 schematically illustrates this point for a hypotheticalgeomaterial with Q, R and φc equal to 10, 1 and 30

◦, respectively, where the axi-symmetric (or

triaxial) deviator stress invariant is defined as q = σ ′1 − σ ′

3. The critical state line (CSL) and thepeak failure envelopes were determined using this procedure, which is outlined in detail by Salgado(2008), for various combinations of DR and p′ (Fig. 9).

88

2.2 Experimental protocol

In order to properly characterize the intrinsic parameters of a geomaterial in such a way that generalpredictions of its shear strength can be made, a systematic experimental protocol must be followedas well. This protocol would systematically account for the effects of the state variables on themechanical behavior of the geomaterial by following an experimental testing program that wouldtake into account at least three different levels for each state variable (DR and p′). In the case ofgeomaterials with large particle sizes such as mine waste rock, rockfll, and coarse aggregates, theexperimental program should also be designed to address the issues of sample-size ratio, membranepenetration and particle breakage discussed above. A new large-scale triaxial testing apparatus hasbeen developed at Colorado State University to allow testing of specimens with diameter and heightequal to 150 and 300 mm, respectively (Fig. 10). Finally, systematic correction and calibrationtriaxial protocols should be followed to properly address the usual issues associated with area,membrane and shear failure mechanisms (Fig. 11), as outlined by La Rochelle et al. (1988).

3 CONCLUSIONS

The main conclusions derived from a comprehensive literature review on the large-scale triaxialtesting and shear strength of geomaterials with large particle sizes can be summarized as follows:

1) Adoption of a minimum sample size ratio of six and specimen diameters of at least 150 mm,along with the use of parallel gradations allow reasonable estimation of the peak friction angle ofmine waste rock, rockfill, coarse aggregates and other geomaterials with large particle sizes. Thedifference between the peak friction angle of the material measured in the laboratory followingthe above criteria and the peak friction angle of the actual material in the field might be expectedto be less than about 5%.

2) Critical state and dilatancy are the two most fundamental aspects associated with the properevaluation of the shear strength of geomaterials. Proper evaluation of these important aspects ofgeomaterial behavior can be carried out by systematically taking into account the effects of themain state variables (density and mean effective stress) during characterization of the intrinsicparameters (e.g., critical-state friction angle, Bolton’s correlation parameters Q and R for thepeak friction angle, and the maximum and minimum void ratios) of the material.

3) Determination of the intrinsic parameters mentioned above requires the use of an appropriatelarge-scale triaxial protocol. In turn, this more rigorous and fundamental approach would allowrobust and more comprehensive predictions of the shear strength of mine waste rock, rockfill,coarse aggregates and other geomaterials with large particle sizes to be made for the mostrelevant combinations of density and mean effective stress encountered in geotechnical andmining applications.

REFERENCES

Ansal, A.M. and Erken, A. (1996) Posttest Correction Procedure for Membrane Compliance Effects on PorePressure, Journal of Geotechnical Engineering, Vol. 122, No. 1, 27–38.

Baldi, G. and Nova, R. (1984) Membrane Penetration Effects in Triaxial Testing, Journal of GeotechnicalEngineering, Vol. 110, No. 3, 403–420.

Bolton, M.D. (1986) The strength and dilatancy of sands, Geotechnique, Vol. 36, No. 1, 65–78.Carraro, J.A.H., Prezzi, M., and Salgado, R. (2009) Shear Strength and Stiffness of Sands Containing Plastic

and Nonplastic Fines, Journal of Geotechnical and Geoenvironmental Engineering, Vol. 135, No. 9, 1167–1178.

Choi, J.W. and Ishibashi, I. (1992) An Experimental Method for Determining Membrane Penetration,Geotechnical Testing Journal, Vol. 15, No. 4, 413–417.

De Josselin de Jong, G. (1976 ) Rowe’s Stress-Dilatancy Relation Based on Friction, Geotechnique, Vol. 26,No. 3, 527–534.

Dendani, H., Flavigny, E., and Fry, J.J. (1988) Test for Embankment Dams: Interpretation and Validity,Advanced Triaxial Testing of Soils and Rock, ASTM, STP 977, 486–500.

Frydman, S., Zeitlen, J.G., and Alpan, I. (1973) The Membrane Effect in Triaxial Testing of Granular Soils,Journal of Testing and Evaluation, Vol. 1, No. 1, 37–41.

89

Fuller and Thompson (1907) The laws of proportioning concrete Journal of Transportation Engineering,ASCE, Vol. 59, 67–172.

Holtz, W.G., and Gibbs, H.J. (1956) Triaxial shear tests on pervious gravelly soils, Journal of Soil Mechanicsand Foundations Division, ASCE, Vol. 82, No. SM1, Proceedings Paper 867, 1–22.

Indraratna, B. and Salim, W. (2002) Modelling of particle breakage of coarse aggregate incorporating strengthand dilatancy Geotechnical Engineering, Vol. 155, No. 4, 243–252.

Indraratna, B., Ionescu, D., and Christie, H.D. (1998) Shear Behavior of Railway Ballast Based on Large-scaleTriaxial Tests, Journal of Geotechnical and Geoenvironmental Engineering, Vol. 124, No. 5, 439–449.

Indraratna, W. B., Wijewardena, L.S.S., and Balasubramanium, A.S. (1993) Large-scale Testing of GrewackeRockfill, Geotechnique, Vol. 43, No. 1, 37–51.

Kramer, S.L., Sivaneswaran, N., and Davis, R.O. (1990) Analysis of Membrane Penetration in Triaxial Test,Journal of Engingeering Mechanics, Vol. 116, No. 4, 773–789.

LaRochelle, P., Leroueli, S., Trak, B., Blais-Leroux, L., and Tavenas, F. (1988) Observational Approach toMembrane and Area Corrections in Triaxial Tests, Advanced Triaxial Testing of Soils and Rock, ASTM, STP977, 715–731.

Lee, D.M. (1992) The angles of friction of granular fills, Ph.D. dissertation, Cambridge University.Lee, K.L., and Farhoomand, I. (1967) Compressibility and crushing of granular soil in anisotropis triaxial

compression, Canadaian Geotechnical Journal, Vol. 4, No. 1, 68–86.Leps, T. M. (1970) Review of Shearing Strength of Rockfill, Journal of Soil Mechanics and Foundations

Division, ASCE, Vol. 96, No. SM 4, 1159–1170.Leslie, D.D. (1969) Relationships between Shear Strength, Gradation, and Index Properties on Rockfill Mate-

rials Proceedings, 7th International Conference on Soil Mechanics and Foundation Engineering, MexicoCity, 201–210.

Marachi, N.D., Chan, C.K., and Seed, H.B. (1972) Evaluation of Properties of Rockfill Materials, Journal ofSoil Mechanics and Foundation Engineering, ASCE, Vol. 98, No. SM1, 95–114.

Marachi, N.D. (1969) “Strength Characteristics of Rockfill Materials” Seventh International Conference onSoil Mechanics and Foundation Engineering, Mexico City, 217–224.

Marsal, R.J. (1969) Particle Breakage in Coarse Granular Soils, Proceedings, 7th International Conference onSoil Mechanics and Foundation Engineering, Mexico City, 155–166.

Marsal, R.J. (1969) Shear Strength of Rockfill Samples, Proceedings, 7th International Conference on SoilMechanics and Foundation Engineering, Mexico City, 225–234.

Marsal, R.J. (1967) Large Scale Testing of Rockfill Materials, Journal of Soil Mechanics and FoundationEngineering Division, ASCE, Vol. 93, SM2, 27–43.

McDowell, G.R., Bolton, M.D., and Robertson, D. (1996) The Fractal Crushing of Granular Materials, Journalof Mechanical Physics of Solids, Vol. 44, No. 12, 2079–2102.

Miura, N. and O-hara, S. (1979) Particle crushing of decomposed granite soil and shear stresses, Soils andFoundations, Vol.19, No. 4, 1–14.

Molenkamp, F. and Luger, H.J. (1981) Modeling and minimization of membrane penetration effects in testsof granular soils, Geotechnique, Vol. 31, No. 4, 471–486.

Nicholson, P.G., Seed, H.B., and Anwar, H.A. (1992) Elimination of membrane penetration compliance inundrained triaxial testing, Canadian Geotechnical Journal, Vol. 30, 727–738.

Nitchiporovitch, A.A. (1969) Shearing Strength of Coarse Shell Materials Proceedings, 7th InternationalConference on Soil Mechanics and Foundation Engineering, Mexico City, 211–216.

Reynolds, O. (1885) On the dilatancy of media composed of rigid particles in contact, with experimentalillustrations, Philosophical Magazine, Series 5, Vol. 20, 469–481.

Rowe, P.W. (1962) The Stress-Dilatancy Equation for an Assembly of Particles in Contact, Proceedings of theRoyal Society of London. Series A. Mathematical and Physical Sciences, Vol. 269, No. 1339, 500–527.

Salgado, R. (2008) The Engineering of Foundations, McGraw-Hill Book Company, New York.Salgado, R., Bandini, P., and Karim, A. (2000) Strength and Stiffness of Silty Sand, Journal of Geotechnical

and Geoenvironmental Engineering, Vol. 126, No. 5, 451–462.Schofield, A.N. and Wroth, C.P. (1968) Critical State Soil Mechanics, McGraw-Hill Book Company, New

York.Sivathayalan, S. and Vaid, V.P. (1998) Truly undrained response of granular soils with no membrane penetration

effects, Canadian Geotechnical Journal, Vol. 35, 730–739.Ueng, T.S. and Chen T.J. (2000) EnergyAspects of Particle Breakage in Drained Shear of Sands, Geotechnique,

Vol. 50, No. 1, 65–72.Vallerga, B.A., Seed, H.B., Monismith, C.L., and Cooper, R.S. (1957) “Effect of Shape, Size and Surface

Roughness of Aggregate Particles on the Shear Strength of Granular Materials” ASTM, STP 212, 63–74.

90


Ore geotechnical testing for heap leach pad design

J. Lupo & A. DolezalAMEC Earth and Environmental, Englewood, Colorado, USA

ABSTRACT: Heap leach pad design requires input from several disciplines, includinghydro-metallurgy, process engineering, unsaturated hydrology, civil engineering, geochemistry,hydraulics, geosynthetics design, and geotechnical engineering. One of the many important aspectsof heap leach pad design is characterization of the ore to define its mechanical and hydraulic behav-ior during stacking and leaching conditions. This has become more important in recent years, asthe size of leach pads has increased, both laterally and in ore height. Some leach pads in North andSouth America currently have ore heights exceeding 150 meters (m), with design ore heights over240 m. Under these conditions, ore compression, permeability, and shear strength become veryimportant factors for heap stability as well as recovery.

In addition, weak compressible ores, such as those encountered for some copper, nickel, gold,and uranium operations are being stacked and leached in pads. While these ore materials are beingstacked at low ore heights (6 to 10 m), ore compression, permeability, and shear strength are stillcritical for heap stability and recovery. To understand the behavior of ore materials under load andleaching conditions, geotechnical laboratory tests are required. These tests are used to define therange of potential behavior under anticipated conditions within the ore heap. This paper presents adiscussion on common and some unique geotechnical tests that are used to support modern heapleach pad design. The paper highlights the importance of testing and discusses how the test resultsare integrated into pad design.

1 INTRODUCTION

Geotechnical laboratory tests are required for the design of heap leach pads to develop the properdesign parameters. Laboratory and field tests are typically conducted on various components of theleach pad design, such as the liner system (e.g. underliner, overliner, and geomembrane liner) aswell as the ore materials to characterize the mechanical and hydraulic properties of the materials,which are then integrated into the leach pad design. The results from laboratory tests are also usedto guide the selection of suitable construction materials, define what type of processing may berequired for the materials, and to develop construction specifications.

Laboratory testing programs conducted for heap leach pad design primarily consist of teststhat are based on accepted standards put forth by the American Society for Testing and Materials(ASTM), while other tests do not have any recognized standard, yet are important for design. Themost common laboratory tests used in the design of leach pads include the following:

• Sieve analysis;• Hydrometer analysis;• Atterberg limits;• Natural moisture content;• Specific gravity;• Moisture/Density relationships;• Consolidated-undrained triaxial compression;• One-dimensional consolidation;• Modified one-dimensional compression;• Permeability testing;• Load-Percolation Testing;

91

Table 1. Geotechnical testing summary table.

Level of design

DetailedMaterial Geotechnical tests Scoping Prefeasibility Feasibility design

Underliner – Particle size distribution X X X XLiner System Atterberg limits X X X X

Specific gravity X XMoisture-density relationship X X XSaturated hydraulic conductivity X XDirect shear/triaxial shear strength X XInterface shear strength X X

with geomembrane

Overliner – Particle size distribution X X X XLiner System Atterberg limits X X X X

Specific gravity X XSaturated hydraulic conductivity X XAir permeability (for pads with X

internal aeration)Direct shear/triaxial shear strength X XInterface shear strength X X

with geomembrane

Ore Particle size distribution X X X XAtterberg limits X X X XSpecific gravity X XOne-dimensional compression XSaturated hydraulic conductivity X X

(under load)Load-percolation X xSoil-Water characteristic curve XDirect shear/triaxial shear strength X X

Geomembrane – Liner load test X XLiner System Interface shear testing X X

• Liner Load Testing;• Soil-Water Characteristic Curve.

The testing requirements vary depending on the type of material under consideration (underliner,overliner, ore, etc) and what stage of design is being considered (scoping, pre-feasibillity, feasibility,or detailed engineering). A summary of the typical tests conducted for various materials in a heapleach pad at different stages of design is presented in Table 1.

The focus of this paper is on the geotechnical testing on ore materials to be placed in the leachpad. The reader is referred to Lupo (2009) for the testing requirements to support leach pad linersystems.

2 ORE GEOTECNICAL TESTING

Geotechnical testing of ore materials is required to assess the mechanical and hydraulic behavior ofthe ore under the anticipated loading conditions within the leach pad. This is a critical componentto leach pad design, as inadequate or improper testing can result in instability of the leach padand/or poor recovery from the ore.

Prior to development of a testing program, it is important to recognize that the character of orematerials may change during the leaching process. For example, significant degradation of orematerials may occur when subjected to acidic leaching solutions or biological activity (Theil and

92

Figure 1. One-dimensional compression test frame.

Smith, 2004). Ore degradation can result in instability of the ore heap, increased internal inventory,decreased recovery and/or lengthened leaching time. The potential degradation of the ore can beassessed by testing leach column residues, which have already been aged by leaching solutions.

The recommended approach in developing a geotechnical testing program for ore materials isto consider the properties of the ore as it is placed in the heap (e.g. agglomeration, pre-treatment,etc), and after leaching. The as-placed ore mechanical and hydraulic properties will influence themethod of ore placement (truck haulage versus conveyor stacking), the type of leach pad to bedesigned (dedicated, valley fill, on/off, hybrid), and initial leaching rates. While the properties ofthe leached ore will affect the ore stacking configuration for stability, leach solution applicationrates of stacked ore, and solution management during operations and closure. From a recoverystand-point, the properties of the leached ore will also affect in-heap inventory and methods torecover the inventory.

As presented in Table 1, ore materials are often subject to the following laboratory tests:

– Particle size distribution– Atterberg limits– Specific gravity– Modified one-dimensional compression– Saturated hydraulic conductivity (under load)– Load-percolation– Soil-Water characteristic curve– Direct shear/triaxial shear strength

Particle size distribution, Atterberg limits, and Specific gravity are commonly used to characterizethe ore materials in terms of geotechnical analogs. For example, an ore with high fines content andplasticity (derived from Atterberg limits) would be expected to have similar attributes to a clayeysoil; while a coarse grained ore with little fines would be expected to behave in a similar fashionas a gravel or coarse sand. While these tests do not provide numerical values that are used directlyin the design of leach pads, the information they provide are useful for correlating the observedbehavior of the ore from the other tests.

2.1 Modified one-dimensional compression

The modified one-dimensional compression test is conducted by placing the ore (fresh or leached)into rigid-wall test vessel (see Figure 1). A load is applied to the ore through a steel platen andhydraulic jack. The change in the height of the ore is measured with a micrometer or other instru-mentation. The measured compression of the ore can then be used to calculate the increase in oredensity and decrease in ore porosity as a function of the applied load. The applied load can beconverted to equivalent ore height using the calculated ore density.

93

Figure 2. Ore compression test results.

One-dimensional compression tests are used to assess the response of the ore under load. Asshown in Figure 2, ore can undergo significant compression and volume reduction with increasingheight of ore on the leach pad (e.g. the lower ore lifts are compressed by the self-weight of theupper ore lifts). The decrease in volume can affect both the hydraulic and mechanical properties ofthe ore.

The durable ore sample lost approximately 50 percent of its original porosity, while the com-pressible ore material lost approximately 90 percent of its original porosity. These changes inporosity can have a profound affect on the ore mechanical and hydraulic response, as discussed inthe following sections.

2.2 Saturated hydraulic conductivity tests

Saturated hydraulic conductivity tests (often referred as permeability tests) are used to measure thehydraulic characteristics of the ore under fully saturated conditions. While ore materials are mostcommonly leached under unsaturated conditions, the saturated ore hydraulic conductivity providestwo useful measurements:

1. The saturated hydraulic conductivity of the ore represents the maximum solution application ratefor the ore. In other words, if the solution application rate exceeds the saturated ore hydraulicconductivity, not only will the ore heap become saturated, but it may become unstable due tothe high phreatic surface within the heap. Therefore, the solution application rate should bemaintained well-below the conditions that could saturate the ore; and

2. By measuring the saturated hydraulic conductivity of the ore under load, the heap leach paddesigner can assess whether the ore heap is likely to become saturated under future conditions,as more ore is stacked higher onto the leach pad. This information may be used in the design ofstacking plans to avoid or minimize saturation and enhance stability.

Saturated hydraulic conductivity tests under load can be conducted using either flexible or rigidwall permeameters. Flexible wall permeameters are used on fine-grained ore materials, while rigidwall permeameters are typically used for coarse-grained ore materials.

Figure 3 presents results from saturated hydraulic conductivity tests conducted under load. Thetests shown on Figure 3 are for a strong, durable ore and weak, compressible ore material. Asexpected, the saturated hydraulic conductivity of the ore decreases with increased ore load (i.e. oreheight). The decrease in hydraulic conductivity is less than an order of magnitude for the durableore, while the compressible ore exhibits a loss of almost three orders of magnitude, reflecting asignificant change in the ore hydraulic properties.

94

Figure 3. Saturated hydraulic conductivity under load.

Figure 4. Equivalent application rate under load.

Figure 4 presents saturated hydraulic conductivity test data for a compressible copper ore. In thisfigure, the data are presented in terms of equivalent solution application rate, in liters per hour persquare meter (L/hr/m2). By plotting the data in this form, it is possible to estimate the maximumore height under which the ore will become saturated under leaching conditions. For example, thedesign solution application rate for this leach pad is 7.8 L/hr/m2. Under this application rate, OreType #1 and #2 are anticipated to become saturated under an ore height of 5 to 8 meters, while OreType #3 and #4 will not become saturated until an ore height of 40 to 45 meters is achieved.

Presenting the data in this format provides a useful tool for the heap leach pad designer and theoperator. The data presented in Figure 4 suggests that if ore is to be stacked greater than 8 meterson the leach pad, then the following leach pad design options are available:

1. Delay stacking Ore Type #1 and #2 on the leach pad until the last 8 meters at the top of the heap;2. Blend all of the ore together, resulting in a composite ore that can be stacked higher than 8 meters.

The ore blend would have to be tested to determine the blend ratio;

95

Figure 5. Load percolation test frame schematic.

3. Design the leach pad to accommodate low ore stacking for Ore Type #1 and #2, and high orestacking for Ore Type #3 and #4; or

4. Consider using an interlift liner for Ore Type #1 and #2. The benefits and costs for an interliftliner would need to be carefully evaluated before selecting this option.

2.3 Load-percolation tests

Load-percolation tests are an alternate method for assessing the hydraulic characteristics of oreunder load, however the focus of the test is to determine what conditions are required to maintainunsaturated ore percolation (applied load and/or applied solution rate).

Load-percolation tests are typically conducted using a modified rigid wall permeameter, asshown in Figure 5. The tests are conducted by placing an ore sample within test vessel, betweenporous plates and perforated load/bearing plates. Leach solution is introduced at the top of the oresample at the design application rate and the effluent is collected at the bottom of the vessel. Asthe leach solution is applied, the load on the ore sample is increased incrementally to the desiredmaximum load. During loading, the ore compression is measured based on the change in heightof the sample. These data are used to calculate bulk density and porosity of the ore sample duringleaching. The volume difference between the applied solution and effluent can be used to estimateore moisture content under leach and moisture up-take prior to leaching. These moisture contentvalues are important as they are used directly in solution management (water balance) calculations.

If, during loading, the applied solution pools on top of the ore sample (above the perforatedloading plate), this would be an indication that the ore sample has become saturated and furtherpercolation is occurring under saturated conditions. The load at which the ore becomes saturatedcan be converted to an equivalent ore height and used to guide the design of maximum ore heightfor the heap leach pad.

For both the saturated hydraulic conductivity and load percolation tests, it is recommended touse actual leach solution, rather than water. Depending on the mineralogy of the ore, the hydraulicresponse may vary considerably on the leach solution chemistry.

2.4 Soil-water characteristic curve

In recent years, some projects have conducted tests to define the soil-water characteristic curve(SWCC) for ore materials. A SWCC relates the soil (ore) moisture content with the soil (ore)suction pressure. This relationship is unique to each ore type and can be used to relate ore moisturecontent to unsaturated and saturated hydraulic conductivity, which is used for percolation andairflow studies within ore heaps. A full discussion of SWCC’s in heap studies is outside the scopeof this paper. However, when considering tests to define the SWCC for ore materials, it is important

96

to note that most SWCC tests are generally conducted on small samples with small particle sizes(minus 25 mm), although some tests have been conducted on larger particles. In addition, theSWCC parameters are likely to vary with ore depth and with ore degradation.

2.5 Triaxial compression/direct shear tests

An integral part of leach pad design is defining the shear strength of the ore materials. Ore materialsare often tested under triaxial compression and/or direct shear tests to evaluate the shear strengthunder the anticipated loading conditions. The decision on whether to use triaxial compressionor direct shear, or both types of testing is generally based on the ore characteristics (fresh andleached) and the anticipated loading conditions (static and seismic loads, ore loading rates, and oreplacement method).

Triaxial compression tests are commonly used to evaluate the shear strength of rock and soilmaterials. There are several types of triaxial compression tests that may be conducted, dependingon the type of loading and drainage conditions that may occur within the ore heap. The generaltypes include:

– Consolidated – Undrained (CU)– Consolidated – Drained (CD)– Unconsolidated – Undrained (UU)

CU tests are conducted whereby the ore is first saturated then consolidated (under drainedconditions) to an effective mean stress that is equivalent to desired ore depth. Once the ore hasbeen consolidated under the effective mean stress, the sample is then sheared under undrainedconditions. CD tests are typically conducted on ore samples that have been prepared at/or near theleaching moisture content of the ore. The sample is consolidated to an effective mean stress that isequivalent to desired ore depth, and then sheared under fully drained conditions. It is important torun CD tests at a strain rate that is slow enough to prevent development of excess pore pressureswithin the sample. Finally, UU tests may be conducted on either saturated ore or an ore sampleat/or near the leaching moisture content. For UU tests, the sample is confined to an effective meanstress that is equivalent to the desired ore depth, and sheared under undrained conditions.

CU tests are the most common test used for ore shear strength testing. At first, testing the oreunder CU conditions may not seem compatible with the concept of leaching under unsaturatedconditions; however in the lower portions of the heap (particularly next to the liner) the ore canbecome saturated. It is these lower portions of the heap that can have a significant impact onthe overall stability of the ore heap. In addition, if excess pore pressures are measured, CU testscan provide both effective and total stress parameters, which allow stability of the ore heap to beassessed under different drainage conditions. CD tests may also be used for coarse, well-drainedore materials that are not anticipated to generate excess pore pressures under load. UU tests maybe used to assess the shear strength of fine-grained ore, with very poor drainage properties thatare to be loaded rapidly (e.g. loading under haul truck traffic). Direct shear tests can also be usedfor leach pad design. These tests should be conducted on saturated or nearly saturated ore andsheared at a strain rate that will not generate excess pore pressures (generally between 0.0025 and1 millimeter per minute).

The results from triaxial compression and/or direct shear tests can be plotted in various formats.One useful format is to plot the data in stress path space, by plotting the mean and deviatoricstresses [mean effective stress [p′] = (σ ′

1 + σ ′3)/2 and deviatoric stress [q] = (σ ′

1 − σ ′3)/2].

Plotting the data in stress path space, allows the designer to observe the behavior of ore undershear. If the ore exhibits dilatant behavior, the ore may gain shear strength under strain (e.g. strainhardening), which is beneficial to heap stability. The opposite of dilatant ore is strain softening(contractive) ore, which rapidly loses shear strength under strain. Under certain conditions, acontractive ore can lead to abrupt failure of the ore heap in a leach pad. A general stress path plotof strain hardening and strain softening ore is presented in Figure 6.

Using the results from triaxial compression tests, the designer can identify if the ore will behavein a contractive or dilatant sense, and make appropriate changes in the heap leach pad design toaccommodate for this behavior. It is important to note that the behavior of the ore under straincan change, depending on the stress level. Figure 7 presents a graph showing ore behaving as acontractive material at low stresses, while transitioning to a dilatant material at higher stresses.

97

Figure 6. Stress path plot (undrained loading).

Figure 7. Transition in ore behavior.

These types of transitions in behavior are important to identify in the design stage, before ore isbeing stacked and leached.

3 CONCLUSIONS

Geotechnical testing to define the mechanical and hydraulic characteristics of ore materials iscritical to the design of heap leach pads. Often the focus on geotechnical testing of the leach padcomponents (liner system, drainage layers, etc), but the testing of ore is also critical to the design.The hydraulic and mechanical properties of the ore will influence:

• Ore stacking (height and placement methods)• Leaching rates

98

• Stability• Recovery• Internal inventory (dissolved).

When developing the geotechnical testing program for ore materials, it is important to consider thepotential changes to the ore characteristics due to ore loading (depth of ore), stacking methods, anddegradation from leach solutions. The geotechnical testing program should be designed to capturethese potential changes so they can be incorporated into the design of the overall leach pad.

REFERENCES

Lupo, J. 2009. Liner System Design For Heap Leach Pads, Geotextiles and Geomembranes, No. 9.Theil, R. and M. Smith, 2004. State of the practice review of heap leach pad design issues, Geotextiles and

Geomembranes, 22, pp 555–568.

99


Critical state liquefaction assessment of an upstream constructedtailings sand dam

C.D. AndersonGolder Associates Ltd., Burnaby, Canada

T.L. EldridgeGolder Associates, S.A, Santiago, Chile

ABSTRACT: The stability of an upstream constructed tailings sand dam is highly dependent onthe strength of the existing tailings to form the foundation of the subsequent raise. Dams that areraised above the original design height or are constructed with a narrow zone of coarse sand onthe upstream face may have portions of the dam constructed over potentially liquefiable tailings.Whereas silts (e.g. fine tailings) have in the past been viewed as non-liquefiable, that view isnow changing such that sufficient ‘fines content’ is no longer regarded as sufficient protectionfrom liquefaction and alternative means of assessment are required. The paper will present anexample where the liquefaction potential of silt-sized tailings was assessed as part of the stabilityassessment for a raise of an existing upstream constructed tailings sand dam. The liquefactionassessment was based on the critical state concept, and used two approaches. First, the liquefactionpotential was assessed using disturbed and undisturbed samples to define the void ratio profileof the tailings with depth, which was then compared to the critical state locus (CSL) determinedfrom a suite of laboratory tests on the tailings. Second, the liquefaction potential was assessedin-situ using piezocone (CPTu) measurements and soil-specific calibrations. The estimated in-situstate parameter from the two methods was remarkably similar, giving confidence to the overallapproach. The calibrated CPTu data was then adopted for assessment of the dam as a whole. Withthe in-situ state parameter profile defined, estimation of liquefaction potential is straightforward,and independent of ‘fines content’, as described in the paper.

1 INTRODUCTION

The ability to adequately manage the risks associated with operating an upstream constructedtailings sand dam relies on the ability to understand and account for the behaviour of the materialsinvolved. Reasonable identification of potential consequences of constructing on tailings requiresthe ability to adequately assess how the soil behaviours such as soil strength and liquefaction willcontrol the dam geometry and raising sequence.

Upstream constructed tailings sand dams are higher risk structures than downstream constructedsand dams because of the high degree of operator attention that must be applied to depositing thetailings in a manner to maximize future stability, maintain the pond in a position to allow sufficientsegregation of coarse particles along the beach in the area of future raises and maintain a low phreaticsurface within the dam shell. The use of hydrocyclones to separate the coarse fraction of the tailingsfor dam construction can improve the operation by eliminating the reliance on particle segregationon the beach to produce the dam shell material. Regardless of the method of segregation of thecoarse fraction, the finest and lowest specific gravity silt particles (often called slimes) depositwithin the tailing pond. If the pond is maintained in a single location, thick layers of these softsilts can develop, and as the dam is raised, the crest may progress over these slimes, as shown onFigure 1. An understanding of the shear strength, consolidation behaviour, hydraulic conductivityand liquefaction potential of these slimes is required in order to predict the performance of the damwith reasonable confidence.

101

Figure 1. Typical configuration of upstream constructed tailings sand dam.

While the prediction of the behaviour of the coarse tailings shell is relatively straightforward, theprediction of the behaviour of the silt tailings is considerably more challenging. While sufficientlycoarse-grained materials and sufficiently fine-grained materials occupy either of the relatively wellunderstood ends of the spectrum of soil behaviour, soils with a high percentage of silt-size particlesoccupy a relatively broad band of the spectrum, where grain size is an insufficient indicator ofanticipated behaviour. Complicating matters, tailings slimes are man-made materials, a product ofa series of crushing, grinding and chemical processes that produce a “soil” that has not experiencedweathering, aging, or other natural processes to which natural soils are subjected. Where standardliquefaction evaluation techniques have been developed based on natural soils, typically clean sands(e.g. Youd et al. 2001, Bray & Sancio 2006, Boulanger & Idriss 2007), such techniques must beused with caution in evaluation of tailings slimes.

To characterize the tailings slimes recently encountered during the investigations for an upstreamraise of a tailings dam, a critical state approach was adopted, which utilizes fundamental physicalprinciples to predict the strength and liquefaction resistance of the tailings slimes (Shuttle &Cunning, 2007). Once the tailings slimes were characterized within the critical state framework, sitespecific correlations were developed for the tailings slimes in order to demonstrate the deviation ofthe behaviour of the tailings slimes from that predicted using the standard techniques. In particular,the extent of potentially liquefiable materials was significantly under-predicted using the standardtechniques, which would lead to unconservative estimates of potential liquefaction.

1.1 Background

The assessment of liquefaction susceptibility of silt tailings was required as part of an evaluationof the stability of a proposed 25 m raise of an upstream constructed sand dam. At the time of theproposed expansion the tailings impoundment had been in nearly continuous operation for about40 years, and had a maximum crest height of about 70 m. The sand dam provides containment onthree sides of the impoundment, with a total crest length of about 3500 m. The dam is raised duringoperations using lifts of about 1.5 m constructed from coarse sand reclaimed from the tailingsbeach. The proposed raise would advance the crest about 125 m upstream, resulting in a face slopeof 5 horizontal to 1 vertical.

Two types of tailings were encountered during the investigation along the final crest alignmentof the 25 m dam raise. “Coarse” tailings were found in the upper portion of the borehole and CPTusoundings, consisting of sand and silty sand to depths of up to about 25 m. Below the coarse tailingsvery soft “fine” silt tailings interlayered with coarse tailings were encountered. The thickness ofthe layers of these very soft silt tailings ranged from 1 m up to 20 m, and the interlayered sequenceextended to depths of up to 70 m. A typical CPTu sounding is shown in Figure 2. The particle sizedistribution of the bulk tailings discharged from the mill along with the coarse and fine fractionsresulting from segregation after deposition are shown in Figure 3.

The coarse tailings encountered during this investigation were generally consistent with previ-ous investigations and assessments at the site, and are generally well understood from thoroughlaboratory and in-situ testing. The coarse tailings behave as a typical loose to compact sand withrelatively high permeability.

The silt tailings that were encountered in this investigation had not been previously encountered inthe investigations carried out in this tailings dam. These silt tailings are soft, and have a relativelyhigh plasticity, as shown on the plasticity chart in Figure 4. In general, these silt tailings wereexpected to behave in an undrained manner as a soft, cohesive material with very low permeability.

102

Figure 2. Typical CPTu data showing upper zone of coarse tailings and lower zone of soft silt tailings.

Figure 3. Particle size distributions of bulk tailings, coarse fraction and fine fraction resulting fromsegregation.

103

Figure 4. Plasticity chart showing plasticity of tailings slimes ranging from low to high.

However, the susceptibility of these soft tailings to liquefaction, either through earthquake loadingor other triggering mechanisms such as rapid loading was not well understood.

The presence and significant thickness of very soft silt tailings required that the behaviour ofthese materials be characterized in some detail to determine their in-situ strength, susceptibilityto liquefaction or softening under dynamic or static loading, and subsequently their critical state(residual) strengths for use in the stability assessment of the dam raise.

2 SILT TAILINGS BEHAVIOUR

2.1 Behaviour of silt tailings

The potential for silts to liquefy has in the past been assumed to be very low, particularly for naturalsilts with some plasticity (Robertson & Wride 1998). In effect, it was assumed that either soilswith a certain fines content did not have sufficient void space to allow collapse of the soil structureduring loading, or that the plasticity was indicative of some tendency for the soil structure to holdtogether under loading.

More recently, case studies and research (Bray & Sancio 2006, Wijewickreme et al. 2005) haveshown that soils with high fines content (greater than about 40% passing the #200 sieve) can andwill liquefy under loading, as void ratios can in certain situations be very high, and the ability ofplasticity to resist soil structure collapse is limited. For silt-size tailings in particular, the relativelyyoung age of the deposit results in an unconsolidated or consolidating mass consisting of angularcrushed particles with high void ratios and potential for collapse of the soil structure. Althoughtailings are often assumed to have low to no plasticity because they result from crushing of rock,the geology and mineralogy of the ore and host rock, and particularly the alterations that are oftenassociated with the ore, coupled with segregation of the finest and lightest particles of the tailingsto the pond can result in layers of tailings with plasticity.

2.2 Correlations for classification of liquefaction potential

The existing methods for classification of liquefaction potential fall into two categories. The firstcategory of methods requires some measurable soil property to be correlated to observed lique-faction or non-liquefaction, either through observations made following natural events such asearthquakes or other failures, or through laboratory testing carried out in controlled conditions.Recent research (Bray & Sancio 2006, Boulanger & Idriss 2007) has focused on the relatively easilymeasured soil index properties of plasticity index (PI), liquid limit (LL) and natural water content

104

Figure 5. Criteria for evaluating liquefaction potential based on soil index testing.

(wc). Using the soil index tests, PI is plotted against either the ratio of wc/LL (Bray & Sancio2006), or against LL (Boulanger and Idriss 2007), with zones of “non-susceptible”, “moderatelysusceptible” and “susceptible” identified based on the observations of samples that did or did notexperience liquefaction, as shown in Figure 5. A wc/LL ratio of between 0.65 and 0.8 is identifiedas the threshold below which the soil will not liquefy.

The second category of methods are those that utilize cone penetration testing (CPTu) data tocorrelate soil behaviour during CPTu sounding to observed liquefaction, also observed from naturalevents or laboratory testing. Research in this area has focused on a modified soil behaviour typeplot, which uses state parameter ψ and soil classification index Ic limiting values to delineatevarious zones of strain softening or non-strain softening soils (Robertson 2008, Shuttle & Cunning2008), as shown in Figure 6.

2.3 Critical state concepts for behaviour of silt tailings

To provide a consistent framework for prediction of soil behaviour the critical state soil mechanics(CSSM) approach has been widely adopted. CSSM forms the basis of several methods of evaluationof liquefaction potential (Been et al. 1991, Plewes et al 1992, Boulanger 2003, Jefferies & Been2006). An overview of the CSSM theory is illustrated on Figure 7. Any soil with a mean effectivestress, p′, and void ratio, e, that plots above the critical state line (or locus), CSL, will contract duringdrained loading, or generate excess pore pressure, reducing the effective stress during undrainedloading, until it reaches the CSL. Conversely, soil with a p′ and e below the CSL will dilate duringdrained loading, or decrease pore pressure, increasing the effective stress during undrained loading,until it reaches the CSL. Once the CSL is reached a soil continues shearing with no change in e or p′.

The strength and behaviour of in situ soils depends on the state parameter, ψ, defined as thevertical difference between the in-situ void ratio and the void ratio at the critical state at the same p′(see Figure 7). Loose or normally consolidated soils have a void ratio above the CSL (positive ψ)which will subsequently exhibit contractive behaviour during shearing (leading to change in voidratio, or excess pore pressure), while dense and over-consolidated soils have void ratios below theCSL (negative ψ) and will exhibit dilative behaviour during shearing (leading to change in voidratio or decreased pore pressures). Depending on the drainage conditions present within the soil,the contractive or dilative behaviour will result in either a change in volume of the soil (drainedbehaviour), or a change in effective stress at constant volume (undrained behaviour). When thereis a substantial loss of strength resulting from the reduction in effective stress during undrainedshearing (strain softening), liquefaction is said to have occurred.

The CSL is associated with large strains (i.e. once initial fabric has been destroyed), hence theCSL may be determined using laboratory testing on reconstituted samples. This is advantageousfor silts which are difficult to sample in an undisturbed manner and differs from other laboratory

105

Figure 6. Criteria for evaluating liquefaction potential based on CPTu.

Figure 7. Critical state soil mechanics concepts.

106

Figure 8. Critical State Line determination from triaxial tests.

based evaluation procedures which require undisturbed samples. With the CSL known the in-situstate of the soils can be easily assessed using measurements of water content to calculate in-situvoid ratios and an estimate of in-situ effective stress.

3 CHARACTERIZATION OF TAILINGS

3.1 Laboratory testing results

The laboratory testing of the tailings focused on samples from the middle and lower depths con-sidered to be representative of the soft silt tailings. The objective of collecting and testing thesesamples was to measure the parameters of the silt tailings, of tailings at greater depths than havepreviously been tested, and of the softest observed samples in order to establish the degree ofvariation in tailings properties throughout the tailings impoundment.

A series of triaxial tests on reconstituted and undisturbed samples from thin-walled piston-tubesampling was undertaken to determine the strength properties of the silt tailings and to establish theCSL. Since the soil properties at the critical state are unaffected by either initial fabric or density,the choice of test conditions was based on obtaining the clearest possible determination of theCSL. Samples were consolidated to a range of confining pressures and sheared in both drained andundrained conditions. The derived CSL is shown in Figure 8.

The CSL is defined by the following equation:

where, for these silt tailings, � = 2.063 is the void ratio of the CSL at a reference pressure of 1 kPa,and λ10 = 0.541 is the slope of the CSL in the plot of e against log(p′).

3.2 In-situ testing results

The state and strength of the soft silt within the interbedded stratum was the principal focus as theselayers contain the weakest and softest soils in the profile, and are therefore possibly a controllingaspect from a stability perspective. As can be seen from the processed CPTu data in Figure 9, boththe normalized pore pressure response during penetration Bq, as defined in Equation 3, and the soilbehaviour type index Ic, as defined by Jefferies & Been (2006) in Equation 3 are about constantwithin the various soft silt interbeds, and thus the silt can be treated as essentially the same materialthroughout the interbedded unit.

107

Figure 9. Characterization of silt tailings from CPTu data.

When the CPTu data are plotted on a modified soil behaviour plot, the silt tailings plot wellbelow the demarcation line between strain hardening and strain softening, and in a similar locationas other silt tailings (Shuttle & Cunning 2008), as shown on Figure 10.

In order to calibrate the CPTu for use in the assessment of liquefaction susceptibility, the labora-tory testing results were used to provide a site-specific calibration of the CPTu behaviour using theprocedure established by Shuttle & Cunning (2007). Such site specific calibration was necessarybecause the tailings were much softer and higher in plasticity than the natural soils from which thestandard correlations were developed.

3.3 Numerical modelling for site specific correlation

Finite element simulations were carried out to determine the relationship between normalized CPTresistance and the state parameter, ψ, of the soft silts. These simulations used the soil proper-ties determined from the laboratory testing. The simulations used the methodology presented inShuttle & Cunning (2007), which was an extension of the earlier work on drained penetration byShuttle & Jefferies (1998). This finite element code has been calibrated to CPT chamber testsand has an accuracy between the calibration data and the finite element simulations of about�ψ = ±0.03 (about 5% of the range in ψ for a soil) when the shear modulus has been measured.When dealing with undrained CPT soundings, the trends are simplified into a near-unique rela-tionship if presented in terms of the parameter group Qp(1 − Bq) + 1, as discussed in Shuttle &Cunning (2007).

Figure 11 shows the computed relationship between this parameter group and ψ0 for the siltsfor a range of reasonable in-situ stiffness and NorSand fitting parameters. The effect of stresslevel has been included. There is a simple trend essentially independent of uncertainty in the inputparameters. This plot forms the basis for the calibration of the in-situ state parameter of the siltwithin the interbedded unit using the CPT data.

108

Figure 10. Soil behaviour plot showing the boundary between strain softening and strain hardening behaviour.

Figure 11. Computed relationship between normalized CPTu resistance and initial state parameter fortailings slimes.

109

Figure 12. Susceptibility to liquefaction based on PI and wc/LL.

The relationship between the state parameter, ψ, and the CPT measurements for these tailingsis thus:

where the coefficients k = 3.84 and m = 4.26 are the curve fit parameters determined during thecalibration process.

Based on Figure 10, a reasonable ‘characteristic’ normalized penetration resistance for the siltis about Qp(1 − Bq) + 1 ≈ 1.3. This value applies throughout the depth range of the deposit, asthe CPT data shows reasonably constant silt state with depth. Taking the central trend through thefinite element results in Figure 11, the corresponding range for the state parameter of the “slimes”is +0.18 < ψ0 < +0.23, which is a very good match to the measured state parameter from watercontents shown in Figure 9.

4 LIQUEFACTION ASSESSMENT

Based on the characterization of the silt tailings discussed above, the silts may potentially weakenduring and shortly after an earthquake. Any such reductions in strength are conventionally includedin the definition of ‘liquefaction’, and in the following discussion the term ‘liquefaction’ will beused to include all seismically induced reductions of strength, such as ‘cyclic mobility’.

Within the critical state framework, it has been shown in previous studies (Shuttle & Cunning,2007, 2008) that for tailings at an in-situ state less than −0.05, the response to shearing will be strainhardening (dilative), resulting in low or no susceptibility to liquefaction. For soils at in-situ state,ψ, greater than −0.05, the response to shearing will be strain-softening (contractive), resulting inpotential susceptibility to liquefaction. This demarcation line of ψ = −0.05 is shown on Figure 10,and the silt tailings plot well below this line indicating substantial susceptibility to strain softeningand subsequent liquefaction.

The second approach to liquefaction assessment discussed above, using correlations based ona threshold PI and minimum ratio of wc/LL, is illustrated in the plot of the silt tailings data inFigure 12, showing a range in anticipated susceptibility to liquefaction due to the range in PI,which is not indicated when using the CSSM approach.

When using the CPT as an investigation tool, this concept of threshold plasticity has effectivelybeen extended to the use of Ic as a proxy for PI as an indicator of susceptibility to liquefaction.Threshold Ic values for natural soils have been established by Jefferies and Been (2006), where Icvalues greater than 2.4 indicate non-susceptible to liquefaction.

Subsequently, in order to include a PI threshold in the assessment of the liquefaction susceptibilitydirectly from the CPT profile, a site specific correlation between Plasticity Index and Ic wasdeveloped. In the absence of sufficient laboratory testing to confirm the zones of liquefactionsusceptibility proposed by Bray & Sancio (2006), a conservative PI threshold of 20 was adopted.

110

Figure 13. Correlation between PI and Ic to develop threshold Ic for liquefaction susceptibility of silt tailings.

The correlation between Ic and PI for the silt tailings is shown in Figure 13, which shows a verystrong linear trend. Based on the correlation between Ic and PI the threshold Ic values to considerthe soft silt tailings as non-liquefiable was considered to be 3.3.

5 SUMMARY

While it is commonly perceived to be conservative to assume tailings are susceptible to liquefactionwith an Ic or PI below a certain threshold value, the foregoing assessment using a systematiccritical state framework and site specific correlations illustrates that this assumption may not be asconservative as previously thought.

The characterization of the tailings using laboratory testing and CPTu soundings provided thenecessary information to calibrate the CPT soundings using numerical modelling techniques, result-ing in a refined estimate of the state parameter ψ which had impressive agreement with estimatesof ψ from water content and laboratory determined CSL. This refined estimate of the state para-meter could then be applied to the CPT soundings to provide a continuous profile of liquefactionsusceptibility, rather than snapshots from specific sampling locations. Using the CPTu profileswithin the critical state framework and plotting data on a modified soil behaviour plot indicatedthat the silt tailings are expected to behave in a highly strain softening manner which is likely toresult in liquefaction of the silt tailings from an applied seismic load of sufficient magnitude.

Using the simplified approach of correlations of liquefaction susceptibility based on PI andwc/LL, and the extension to relating PI to Ic, site specific values for a liquefaction threshold Ic for thesilt tailings were found to be 3.3 and 3.45 in place of 2.4 and 2.6 for determinations of Ic by Jefferiesand Been (2006) or Robertson and Wride (1998), respectively. This higher value of a threshold Iceffectively increases the extent of tailings that would be identified as potentially liquefiable.

Subsequently, determinations of the anticipated extent of potential liquefaction using the stan-dard approaches could significantly underestimate the size of the liquefied zone. If subsequentengineering decisions are based on the anticipated extent of liquefaction, these decisions may notin fact be achieving the intended level of performance.

REFERENCES

Been, K., Jefferies, M.G., & Hachey, J. 1991. The critical state of sands. Geotechnique 41(3): 365–381.Boulanger, R.W. 2003. Relating Ka to Relative State parameter index. Journal of Geotechnical and

Geoenvironmental Engineering 129(8): 770–773.

111

Boulanger, R.W., & Idriss, I.M. 2007. Evaluation of cyclic softening in silts and clays. Journal of Geotechnicaland Geoenvironmental Engineering 133(6): 641–652.

Bray, J.D. & Sancio, R.B. 2006. Assessment of the liquefaction susceptibility of fine-grained soils. Journal ofGeotechnical and Geoenvironmental Engineering, 132(9): 1165–1177.

Canadian Dam Association (CDA) 2007. Dam Safety Guidelines & Technical Bulletins. Canadian DamAssociation, Edmonton, Alberta.

Hynes-Griffin, M.E. & Franklin, A.G. 1984. Rationalizing the seismic coefficient method Miscellaneouspaper GL-84-13 US Army Corps of Engineers Waterway Experiment Station, Vicksburg, Mississippi.

Ishihara, K. 1993. Liquefaction and flow failure during earthquakes. Géotechnique 43(3): 351–415.Jefferies, M.G. & Been, K. 2006. Soil liquefaction, a critical state approach. Taylor and Francis.Marcuson, W.F., Hynes, M.E. & Franklin, A.G. 2007. Seismic design and analysis of embankment dams: The

state of practice. 4th Civil Engineering Conference in the Asian Region, Proc., Taipei, 25–28 June 2007.Olson, S.M. & Stark, T.D. 2002. Liquefied strength ratio from liquefaction flow case histories. Canadian

Geotechnical Journal 39(3): 629–647.Plewes, H.D., Davies, M.P., & Jefferies, M.G. 1992. CPT based screening procedure for evaluating liquefaction

susceptibility. 45th Canadian Geotechnical Conference, Proc., Toronto, Ont. 26–28 October 1992.Robertson, P.K. 2008. Discussion of “Liquefaction potential of silts from CPTu”. Canadian Geotechnical

Journal 45(1): 140–141.Robertson, P.K., & Wride (Fear), C.E. 1998. Evaluating cyclic liquefaction potential using the cone penetration

test. Canadian Geotechnical Journal 35(3): 442–459.Robertson, P.K. 2010. Evaluation of flow liquefaction and liquefied strength using the cone penetration test.

Journal of Geotechnical and Geoenvironmental Engineering 136(6): 842–853.Seed, H.B., Cetin, K.O., Moss, R.E.S., Kammerer, A., Wu, J., Pestana, J., Reimer, M., Sancio, R.B., Bray, J.D.,

Kayen, R.E., & Faris, A. 2003. Recent advances in soil liquefaction engineering: A unified and consistentframework. 26th Annual ASCE Los Angeles Geotechnical Spring Seminar, Keynote presentation, LongBeach, CA 30 April 2003.

Shuttle, D.A. and Cunning, J. 2007. Liquefaction potential of silts from CPTU. Canadian Geotechnical Journal44(1): 1–19.

Shuttle, D.A. and Cunning, J. 2008. Reply to the discussion by Robertson on “Liquefaction potential of siltsfrom CPTu”. Canadian Geotechnical Journal 45(1): 142–145.

Shuttle, D.A., and Jefferies, M.G. 1998. Dimensionless and unbiased CPT interpretation in sand. InternationalJournal for Numerical and Analytical Methods in Geomechanics 22(5): 351–391.

Wijewickreme, D., Sanin, M.V., and Greenaway, G.R. 2005. Cyclic shear response of fine-grained minetailings. Canadian Geotechnical Journal 42(5): 1408–1421.

Youd, T.L., Idriss, I.M., Andrus, R., Arango, I., Castro, G., Christian, J., et al. 2001. Liquefaction resis-tance of soils: Summary report from the 1996 NCEER and 1998 NCEER/NSF workshops on evaluationof liquefaction resistance of soils. Journal of Geotechnical and Geoenvironmental Engineering 127(10):817–833.

112


Heap leach pad cover design analyses Salmon, Idaho

I. HutchisonStrategic Engineering & Science, Inc., Irvine, California, USA

A. WhitmanMeridian Beartrack Company, Reno, Nevada, USA

J. Juliani & Tarik Hadj-HamouStrategic Engineering and Science Inc., Irvine, California, USA

ABSTRACT: In accordance with Federal and State laws a former gold mine in Idaho is undergoingfinal closure. The closure criteria for the Heap Leach Pile (HLP) which contains 23 million tons(21 millions metric tons) of gold ore residues from cyanide leaching operations are the flow andquality of leachate released. Flows meeting the established chemical load criterion can be released toinfiltrate without treatment. Combinations of soil and geosynthetic covers and long term treatmentwere evaluated. Financial life cycle costs analyses including consideration of risk and long termmaintenance costs were performed to determine the least cost cover/treatment plant combinationand assess the incremental costs of the other alternatives. Decision analyses were performed toevaluate which alternatives would best meet a set of closure objectives. The results were utilizedby the Mining Company to select the cover type that best fit its financial goals, internal riskmanagement approach and criteria for minimizing liability.

1 INTRODUCTION

The Heap Leach Pad (HLP) covers approximately 119 acres (47.6 acres) and contains approximately23 million tons (21 millions metric tons) of crushed ore residues (Figure 1). The crushed ore residueis characterized as silty gravel with sand, and is generally sized less than 3 inches (75 mm) indiameter. The crushed ore residue originated from the pits that were exploited during the activelife of the mine. The crushed ore residue is relatively nutrient poor and consequently the slopes aresparsely vegetated.

The HLP was located against an existing hill side and presents two fill slope faces, referred toat the East side slope and the West side slope, respectively. The major characteristics of the slopesare as follows:

– East Side Slope: varies in height from (34 to 55 m), is approximately 1,880 ft (565 m) long,and covers approximately 25 acres (10 ha). The slope inclination ranges from about 27 to 29per cent (i.e., 3.44 horizontal to 1 vertical).

– West Side Slope: varies in height from about 50 to 170 ft (15 to 21 m), is approximately 1,570 ft(479 m) long, and covers approximately 18 acres (7.2 ha). Inclination of the slope varies from21 to 28 per cent.

The top deck covers approximately 76 acres (30.4 ha) and was closed with a geomembrane basedcover in 2007. Selection and implementation of the closure approach for the slope is discussed inthe balance of this paper.

Sources of leachate from the HLP are the remnant of the cyanide-base solution used to leach outthe gold from the ore and water from infiltrating precipitation (rain and snow melt). The volume ofleaching solution in the HP is diminishing annually as it drains out of the HLP whereas the volumeof infiltration form precipitation is continues and varies seasonally depending on the total amountof precipitation that falls on the HLP. The average annual precipitation at the site is on the order

113

Figure 1. Heap Leach Pad in background.

of 24 in. (61 cm), with the greatest amount of precipitation occurring between the months of Mayand June. Measurements of the snow-pack at several locations on the site between February 1997and March 1999 indicated that the typical late winter snow-pack has been 28.8 in. (71 cm) (i.e., 3to 5 in. (7.5 to 12.75 cm) water equivalent). The major accumulations of snow occur in Decemberthrough February. The annual average potential evaporation is about 32 in. (81 cm).

Leachate and a portion of the runoff from the HLP are collected in a lined ditch located alongthe western boundary. The lined ditch conveys the collected liquids to the Operating Pond. Recentreclamation activities have resulted in diversion of a significant quantity of the runoff from theheap into the stormwater management system so it does not report to the Operating Pond.

2 HLP CLOSURE

2.1 Objectives and alternatives

The basic design objectives for the HLP side slope cover include infiltration reduction, long termdurability, cost effectiveness, and sustainability

To achieve these objectives a series of side slope cover alternatives were considered for the slopesof the HLP.

• Alternative 1 – Composite Soil/Geosynthetics cover, consisting of a vegetated cover soil, under-lain by a drainage geocomposite and a flexible geomembrane, with either a thin soil cover (2 ft(0.61 m)) or a thick soil cover (4 ft (1.2 m)).

• Alternative 2 – A capillary break soil cover consisting of a vegetated soil cover underlain by acapillary break consisting of a drainage geocomposite or coarse layer of soil.

• Alternative 3 – Evapotranspirative soil cover, consisting of either a thin (2 ft (0.61 m)) or a thicksoil cover (4 ft (1.2 m)).

• Alternative 4 – Hybrid cover, consisting of a composite (Alternative 1) cover on a portion ofthe slopes and a capillary break soil cover (Alternative 2) or an evapotranspirative soil cover(Alternative 3) on the remaining slopes. The composite (Alternative 1) cover would be placedon those portions of the HLP that yielded the highest constituent concentrations.

Because of the location of the mine at near 7,000 ft (2,130 m) elevation in northern Idaho, theconstruction season is limited to four months, and therefore construction of the final cover willtake place over two years.

114

Table 1. Key parameters associated with closure of the HLP.

Side slope cover type

Parameters Alternative 1 Alternative 2 Alternative 3 Alternative 4

Long-Term Residual Infiltration <1 [3.8] 12 [45.6] 10 [38] 5 [19](Calculated – gpm [liters per day])

Long-Term Residual Chemical LeachateLoads (lb/year[gpm/year])

Cadmium: 0.03 [14] 0.5 [230] 0.4 [180] 0.4 [180]

Mercury: 0.10 [45] 1.9 [860] 1.6 [730] 0.2 [90]

Approximate Cover Capital Cost $8.3 $6.6 $4.4 $5.8($ Million)

Potential Leachate Treatment Costs $1.2 $2.4 $2.6 $1.8(NPV $ Million)

Period of Treatment (Years)1 5 years >30 years >30 years 10 years

Approximate Total Cost ($ Million) $9.5 $9.0 $7.0 $7.6

1 Thereafter leachate metals loads are low enough for land application.

Key parameters, including infiltration rates, leachate metal loadings and costs associated withthe alternatives are provided in Table 1. Initial analyses had indicated that a soil cover over theentire HLP would not result in enough reduction of infiltration and therefore reduction in leachateproduction to meet the NPDES requirements. Consequently it was decided to construct a compositesoil/geosynthetics cover on the top deck (Alternative 1) and evaluate the above four alternativesfor the slope cover. Even with he geosynthetic on the top deck the projected long-term leachateloadings for some of the side slope cover alternatives, may not reduce leachate mercury loading,for example low enough for discharge (See Table 1). In the pond National Pollution EliminationDischarge System (NPDES) limits for mercury have been as low as 0.1 lb/year.

2.2 Top deck

The final cover for the deck (TRC, 2007) and was constructed in 2007. The final cover for the deckconsists of the following from bottom to top:

– Foundation: graded crushed ore residue.– Geosynthetic Barrier: 60-mil (1.5 mm) double textured LLDPE geomembrane.– Geosynthetic Cushion: 16-oz/sq. yd. (540 g/m2) nonwoven geotextile.– Select Protective Soil: 1-ft (0.30 m) thick glacial till layer processed with a maximum opening

size of 4 in. (10 cm).– Vegetative Soil: 1-ft (0.30 m) thick soil layer.– Final Cover Surface: fine graded and seeded.

The deck is approximately 76 acres 930.4 ha) and therefore represents 64 percent of the watershedof the HLP. By eliminating infiltration into the HLP through the top deck the toal infiltration andhence leachate production has been significantly reduced.

2.3 Alternative covers for the slopes

2.3.1 Alternative 1 – composite soil/geosynthetics coverThe advantages of Alternative 1 include:

– Optimum performance. Best practically achievable cover performance infiltration control, henceleast long-term treatment costs.

115

Disadvantages associated with Alternative 1 include:

– Durability/Potential Long-Term Maintenance Required. Frost, seepage etc. considerations min-imized by intermediate anchor trenches, drainage outlets and benching Potential stabilityconcerns regarding the creep and large-scale movement of the side slope covers because lack ofdrainage to minimize seepage parallel to slope.

– Highest Cost: Higher Capital Cost and Construction Cost than other alternatives.– Schedule – not feasible for completion in single season.

Risks associated with Alternative 1:The most significant risk associated with Alternative 1 is potential instability of the soil

cover veneer and geosynthetic cover components on the sideslopes. To minimize the risk ofdownslope creep and mass movement of the composite cover, the CCM incorporates additionalgrading/benching, drainage outlet and anchor trench elements and costs for anticipated requiredroutine maintenance and repair costs. The need to significantly repair or replace the cover morefrequently would drive up the net present value of this alternative. On the other hand, the risk oflong term treatment and discharge of a significant flow of leachate (i.e., dependence on maintainingthe NPDES permit) is mitigated.

2.3.2 Alternative 2 – soil/capillary breach coverThe advantages of Alternative 2 include:

– Constructability – equipment can place in horizontal lifts, with little to no material processingrequired.

– Performance – Next best practically achievable cover performance infiltration control, henceleast long-term treatment costs.

– Schedule – feasible in one season.

Disadvantages associated with Alternative 2 include:

– Cost: Higher capital cost & construction cost.

Risks associated with Alternatives 2:The most significant risks associated with Alternative 2 are the possibility that the proposed

soil cover would not be as effective as anticipated in reducing infiltration and that the period oftransient drainage would be longer than assumed. If the soil cover is not as effective as assumed,a soil/geocomposite cover could easily be installed at a later stage, in effect deferring a portion ofthe capital cost of implementing Alternative 1 (with some increase due to required some additionalregrading). Slower transient drainage would require longer active treatment and hence higheroperating costs and net present value. This alternative also balances the risk of long-term treatmentand discharge by reducing the flow of leachate (i.e., mass loads would be reduced, although not tothe extent of Alternative 1, a treatment and discharge would still likely be required).

2.3.3 Alternative 3 – evapotranspirative (soil) only coverThe advantages of Alternative 3 include:

– Constructability – equipment can place in horizontal lifts, with little to no material processingrequired.

– Schedule – Construction expedited. Single construction season. Reasonable and practicalamount of construction in 2010.

– Least cost reduced capital expenditure and net present value commensurate with projectobjectives.

The disadvantages with Alternative 3 include:

– Infiltration control.– Ongoing transient drainage. The possibility that transient drainage quantities are not reduced

enough may necessitate subsequent implementation of Alternative 1 or other improvements inthe future.

The risks associated with Alternative 3 are the same as those presented for Alternative 2:

116

2.3.4 Alternative 4 – hybridThe advantages of Alternative 4 include:

– Infiltration control.

The disadvantages associated with Alternative 4 include:

– Cost.– Ongoing transient drainage. The possibility that transient drainage quantities may still not

be reduced enough may necessitate subsequent implementation of Alternative 1 or otherimprovements in the areas that were closed with a soil cover only.

The risks associated with Alternative 4 are a combination of those identified for the individualalternatives but to a reduced scale.

2.4 Evaluation parameters for slope cover

Based on costs alone (Tables 1 and 2) the soil/evapotranspiration cover (Alternative 3) ranks thehighest, even though it will require leachate treatment in the very long-term. This result in largelydue to the fact that all the alternative covers require treatment for at least 5 years and because netpresent values are used to represent these long-term costs. To further analyze the cover alterna-tives a multi-parameter analysis was completed using parameters such as performance, long termdurability, cost effectiveness, and sustainability. The criteria for each objective are detailed in thefollowing subsections.

2.4.1 Performance:– Reduction of infiltration: Cover types that maximize infiltration reduction are considered more

highly. This is a qualitative evaluation parameter which addresses the ease of leachate manage-ment. The actual approach to managing the leachate (i.e. collection, treatment and disposal) isaccommodated under the cost parameter described below.

– Constructability: This parameter considers the complexity associated with the cover constructionand the availability of local materials.

2.4.2 Cost:Net present values costs including capital and long-term leachate management.

Capital cost includes:

– Design, construction management and quality assurance– Contractor mobilization and demobilization– Existing slope grading and surface preparation– Preparation of soil cover fill material– Construction of cover– Vegetation of cover

Treatment costs include:

– Operation of leachate management system.

2.4.3 SustainabilityNeed for leachate treatment:

– The need for leachate treatment, an administrative support system is considered a negative.Therefore Alternatives requiring treatment for longer periods are rated lower than those whichmay only require it for a short period of time.

Risk of future treatment:As future NPDES discharge permit limits are unknown, the higher the long-term residual leachate

flow is, the higher the risk of requiring treatment, even if no treatment was required before.

117

Figure 2. Decision tree.

2.4.4 Long-term durabilityIn order to evaluate performance, the following sub-parameters were considered:

Healthy vegetation:

– This parameter considers the type and durability of the vegetation that can be supported by thecover alternatives and also the extent to which the cover’s performance can be adversely affectedor damaged by vegetation; e.g. reduced compaction caused by root penetration.

Susceptibility to erosion during storm conditions:

– This parameter evaluates the ability of the alternative cover types to minimize surface erosionand soil loss. As all of the alternatives will have similar soil cover and surface drainage features,this is not a distinguishing feature and is not considered further in these evaluations.

Susceptibility to damage during snowmelts conditions:

– Frozen soil will thaw from the top down creating a saturated soil mass over the still frozen soil.There is a risk for some sloughing or localized failure of this saturated soil mass.

Susceptibility to tree damage:

– Thinner covers or soil covers overlying geosynthetics are considered a higher risk to tree damage,particularly in the event trees topple over.

3 EVALUATION OF THE ALTERNATIVES

3.1 Method of analyses

To compare the four alternatives decision analysis software was utilized using the followingapproach:

– A “decision tree” was constructed to provide for a multi-parameter analyses (See Figure 2)– The relative weights of the various levels of decision parameter were determined using input

from a multi-disciplinary team and the owner.– The alternatives were scored according to the parameters.

118

Table 2. Multi-parameters decision analyses results.

Multi-parameter analysis Cost only analysis

Alternative Score Rank Score Rank

1. Soil/Geosynthetics 57.5 2 37.0 42. Soil/Capillary Break 45.7 3 48.1 33. Soil/Evapotranspiration 41.5 4 60.4 14. Hybrid (Alternatives 1 and 2 or 3) 67.8 1 48.9 2

The decision analysis software utilized calculates an overall score for each of the alternatives.For comparative purposes, the decision analysis was run for the weighted parameters shown onFigure 2, as well as for the cost only parameters.

3.2 Results

The results are provided in Table 2 and illustrated:

– For the multi-parameter analysis the hybrid alternative (No. 4) ranked first, as if combined thebenefits of the reduced metal loading from the geosynthetic cover over portion of the slopes andstill retained some of the durability and ease of construction of a soil only cover on the remain-ing slopes. The soil/geosynthetic (Alternative 1) ranked second, mainly because it minimizesinfiltration and the risks of future treatment should NPDES permit conditions change.

– For the cost-only analysis, the soil/evapotranspiration (Alternative 3) cover ranked first with thehybrid (Alternative 4) again ranking fairly high at second.

4 CONCLUSIONS

Selection of a final cover for the HLP at the mine in Salmon, Idaho was based on a multi-criteriaanalysis. Identification of the objectives and criteria that need to be considered was paramount tothe process. Often taken as the dominant criterion in the selection process of engineering option,cost was not so in this case. Issues such as sustainability and maintenance for instance were alsocritical in the selection of the final cover.

The analyses could be expanded to incorporate other parameters such as those related to soilmass wasted during exploitation of the borrow sources to compare the cost of importing soil versusexploiting that on site, the susceptibility for failure of the cover drainage systems an consequenceson the overall management of the stormwater system.

The analysis presented here has been simplified. However the overall approach followed atthis site can be implemented at other sites considering final closure especially remote sites whenaccess may be difficult at certain period of the year and where maintenance may be difficult.The systematic approach followed allows the designer to include in the decision and selectionprocess criteria sometimes overlooked. The assignment of weight allows the designer to assess theimportance of those criteria.

119


The effect of tailings characteristics on cover system success

J. KellerGeoSystems Analysis, Inc, Hood River, OR, USA

M. Milczarek, T.M. Yao & M. BuchananGeoSystems Analysis, Inc, Tucson, AZ, USA

ABSTRACT: Mine tailing properties significantly differ from other mine waste (e.g. waste rockand heap leach material) such that cover system design criteria for cover success and post-closuremonitoring should require different approaches. Tailings can generally be classified into threematerial types corresponding to location within the impoundment, with each material type pos-sessing distinct physical and hydraulic properties. Finding from tailings reclamation research andperformance monitoring at five tailings facilities in the southwestern United States indicate that:(1) alternative cover system designs based on location within the impoundment can maximize per-formance (and reduce costs), (2) tailings underlying shallow evapotranspirative cover systems playa significant role in reducing net percolation, whether they are non-acid or acid, and (3) dependingon the cover material properties and climate, monolayer covers over acid tailings may show limitedacidification and salinization. Consequently, tailings cover system design should consider potentialinteractions between the tailings, cover material and vegetation.

1 INTRODUCTION

Closure and reclamation of mine tailings facilities are guided by three general goals. The firstgoal is to develop a sustainable reclaimed land which is stabilized against wind and water erosion,revegetated, and in the long-term the reclaimed surface is a soil material that has structure andnutrients of a typical soil for the area. These components are inherently related such that non-acidic tailings stabilization can be achieved if the tailings material has a nutrient composition andhydraulic properties that will support vegetation. In cases where the tailings material is too acidicor saline to support vegetation, cover material may be utilized which provides a growth mediumand protects the tailings material from erosion.

A second goal is to minimize deep percolation and drainage from the tailings material, whichcan serve as a long-term pollution source to surface and groundwater. In arid and semi-arid envi-ronments, deep percolation can be reduced by placing an appropriately designed cover system thatacts to store water within the cover material where it is available for evaporation and transpiration(Dwyer, 2003; Albright et al., 2004; Milczarek et al., 2009). An appropriately designed store andrelease cover system will use cover material with adequate structure and nutrient composition tosupport vegetation, and with hydraulic properties that allow for sufficient soil-water storage toretain infiltrated water from rainfall or snowmelt events.

The final goal is to develop a closure and reclamation plan that optimizes performance whilelimiting capital, operation and maintenance expenses. This requires developing site-specific clo-sure and reclamation plans that account for conditions (e.g. climate, tailings properties, borrowmaterial properties, area vegetation) specific to that site. Applying a “one size fits all” closure andreclamation plan may fall short of meeting the needs or may result in unnecessary work, both ofwhich increase either short-term and long-term costs.

This paper presents a general summary of findings from over a decade of copper tailings recla-mation research and performance monitoring at five copper tailings facilities in the southwesternUnited States. These copper tailing facilities are located in the Sonoran or Chihuahuan deserts and

121

Figure 1. Example particle size distributions for beach sand and slimes tailings material.

are characterized by average annual precipitation that range from 300 to 450 mm and referenceevaporation conditions exceeding 1700 mm per year. The findings are presented within the contextof the three closure and reclamation goals.

2 IMPORTANT TAILINGS CHARACTERISTICS

Mine tailing properties differ significantly from other types of mine waste, such that reclama-tion design, the criteria for reclamation success, and post-closure monitoring require differentapproaches from the standard methods used for waste rock and heap leach material. Tailings lackorganic matter, soil microbes, soil structure, and plant nutrients which complicate reclamationactivities. In addition, tailings can have a low hydraulic conductivity and high moisture retentionsuch that drainage from saturated tailings material may take decades to centuries.

2.1 Physical characteristics

Tailings are poorly graded material primarily made up of mostly silt sized particles and lack soilstructure. Due to fluvial deposition processes, significant sorting and layering of the tailings mate-rial typically occurs within an impoundment regardless of the deposition method. In general, threetextural areas are created: (1) beach sands which represent coarser textured material that settledout first, (2) the slimes which represent finer textured material that settled out last, and (3) a mixedarea between the slimes and beach sands. Figure 1 provides an example of particle size distributionfor beach sand and slimes material.

2.2 Hydraulic characteristics

The soil water characteristic curve (SWCC) which describes the soil-water content versus pressurehead, and the hydraulic conductivity function which describes the hydraulic conductivity versussoil-water content or pressure head, varies significantly by the different impoundment texturalareas. Figure 2 shows an example SWCC and hydraulic conductivity function for beach sand andslimes material. Slimes material has greater soil water retention capacity and hence greater plant

122

Figure 2. Example moisture retention curve and hydraulic conductivity function for beach sand and slimestailing material.

available water than the sand material. However, the slimes material is less conductive than thebeach material at wetter (less negative) pressure heads. Under these pressure head conditions (i.e.during and after tailings deposition) the slimes material impede downward flow more than thebeach material and will result in significantly increased drainage times of tailings water comparedto that from the beach area material. As an example, assuming initial capillary pressures of −10 cmand a tailings impoundment thickness of 100 ft, it would take approximately 500 years for theslimes material to drain free water, whereas the beach material would only take 0.5 years. Theresult is drain down of the slimes material can take decades to centuries, albeit at very low rates(i.e. 1 gpm/acre of impoundment). Depending on the size and height of the impoundment, variablesaturation and drainage conditions can be expected within the different tailing textural areas.

2.3 Geochemical characteristics

Tailings are typically plant nutrient limited, have minor levels of organic carbon and a functioningmicrobial community, and can be saline to hyper-saline. All of these factors limit the potential fordirect revegetation of tailings material. Additionally, the ore body mineralogy can result in highacid generation potential, acidity and high plant available metals. Nonetheless, circumneutral tomoderately acid tailings have been successfully revegetated in a variety of climatic environmentsusing organic matter addition and lime amendments as needed (i.e. Brown et al., 2005; Sauer,et al., 2002; Bengson, 2000; Munshower et al., 1995). In general, the effect of tailing geochemicalcharacteristics on potential revegetation and whether a cover system is needed can be classified asshown in Figure 3.

3 REVEGETATION

General observations regarding vegetation success on reclaimed copper tailings in the southwesternUnited States are as follows. Organic amendments can be successfully used to reclaim circumneutraltailings, however, low to moderate amendment rates should be used to limit high-nutrient conditionsthat favor for undesirable non-native species. Volunteer revegetation on copper tailings has beenobserved on circumneutral tailings, though vegetation is generally limited to the slimes area andhalophyte species (Milczarek, 2006). Greenhouse and field experiments with raw tailings treatedwith biosolids and green waste showed significant vegetative cover with native species whichoutperformed untreated plots over at least eight years (Thompson et al., 2001; Milczarek et al.,

123

Figure 3. Tailings geochemical characteristics and influence on relative cover system depth potentiallyrequired to support revegetation.

Figure 4. Mean vegetation groundcover for amended and non amended plots.

2006). These tailings also showed no significant changes in geochemical weathering and nitrateleaching (Pond et al., 2005).

Other long-term experiments with organic amendments added to 30 cm and 60 cm cover depthsover acid tailings have shown that significant differences in vegetation density were sustained after10 years of reseeding (Milczarek et al., 2009). Figure 4 shows that the addition of biosolids attwo different levels resulted in significantly greater mean native and non-native vegetation groundcover, grass, and forb and shrub groundcover than in unamended plots. However, unamended plotsgenerally showed greater native species diversity, but lower overall frequency and biomass. Theinfluence of organic amendments on vegetation ground cover was observed to persist over 10 yearsafter application relative to the unamended plots.

124

Figure 5. Mean vegetation groundcover for 30 cm and 60 cm cover plots.

Figure 6. Vegetation on circumneutral (left) and moderately acidic (right) tailings plot with 15 cm covermaterial.

In this same study, there were no significant differences over ten years in observed vegetativeground cover between 30 cm and 60 cm cover depth test plots over acid tailings (Figure 5). Thiseffect may be due to the endemic presence of South African grasses in the southwestern UnitedStates. In general, the South African grasses did well in all test plots, but, greater native speciessuccess was observed on the 60 cm cover depths (Milczarek et al., 2009).

High salinity and/or acid tailings has been shown to restrict vegetation success in shallow covers(e.g. less than 15 cm) most likely due to root contact with high salinity and acidity levels. Virtuallyall semi-arid plant species are acid intolerant with soil pH levels below 5 considered to adverselyaffect vegetative growth (i.e. Borden et al., 2005; Barth, 1986; Shafer, 1979). Salt-tolerant plantscan withstand higher salinity levels, however, vegetative density and the ability to extract waterefficiently may diminish with increasing salinity. Examples of reclaimed copper tailings with a15 cm cover overlying circumneutral tailings and a neighboring area with 15 cm cover overlyingmoderately acidic (pH > 5) tailings are shown in Figure 6. Both pictures were taken ten years afterseeding and planting of trees and shrubs. Tree and shrub planting was limited to the circumneutral

125

Table 1. Rooting profile descriptions for circumneutral and moderately acidic reclaimed tailings plots.

Root Density1

Trench Cover depth Soil cover 0–10 cm below 10–20 cm belowID (cm) (above tailings contact) tailings contact tailings contact

P1 13 4 2 0P2 24 4 1 1P3 19 4 2 0P4 12 4 1 0P5 21 5 4 1P6 16 4 3 1P7 21 4 1 0P8 24 4 3 2P9 11 4 3 2P10 19 5 4 3P11 17 4 3 2P12 18 3 2 2

1Root density descriptions use a modified USDA classification system for root abundance: 0 = none, 1 = veryfew/none, 2 = few, 3 = few/common, 4 = common, 5 = common/many, 6 = many.

tailings area, otherwise seeding treatments were identical. The greater vegetative success in thecircumneutral tailings plot compared to that in the moderately acidic tailings plot can be observed.

The ability of plant roots to propagate into tailings is influenced by many components, includingcompaction, salinity, and acidity. Frequently, a combination of tailings salinity and the densenature and generally poor soil structure of deposited and consolidated tailings limits root extensionand density. Moreover, the generally low permeability of mixed and tailing slime areas limitsthe downward infiltration of moisture at depth, resulting in root concentration near the surface.At several reclaimed copper tailings in the southwestern United States, plant roots have beenobserved to actively root into circumneutral and moderately acidic (pH > 5) tailings (Milczareket al., 2006). Table 1 presents rooting profile descriptions for several reclaimed tailings areas. Rootswere observed down to 20 cm below the tailings and cover material contact, although rooting wasat much lower densities than in the soil cover material. The implications for reclamation planningare that rooting into the tailings material extends the depth of plant water extraction and makes thetailings a component of the overall cover system.

Finally, vegetation characteristics vary with location with mesic type vegetation (e.g. creosoteand salt cedar) in the slimes and xeric type vegetation (e.g. cattails) in the beach sands. This maychange over time as the slimes area dries out and if surface runoff is not available to replenishdrained and evaporated moisture.

The vegetation monitoring results indicate that an understanding of the geochemical character-istics (e.g. pH and salinity) and hydraulic characteristics (e.g. slimes or beach area) of the tailingsmaterial and their spatial distribution will allow for increased likelihood of long-term revegetationsuccess.

4 INFILTRATION AND NET PERCOLATION

Infiltration is the process of water entry into the soil (e.g. rain or snowmelt event). Infiltrated watermay return to the atmosphere through evaporation or transpiration of plants. Water that remains inthe soil profile and continues downward past the evapotranspiration zone is termed net percolationand over the long term can be considered equivalent to aquifer recharge. Cover systems act toincrease the evapotranspiration zone, water storage capacity and return of infiltrated water to theatmosphere by evapotranspiration processes.

Monitoring data collected on cover systems in the southwestern United States indicate thatshallow cover systems can effectively store and release precipitation, though episodic sequences

126

Figure 7. Pressure head measurements for vegetated plots with 30 cm and 60 cm cover and bare tailings.

of above-average precipitation can result in net percolation past the cover system (i.e. Milczareket al., 2009; Fayer and Gee, 2006; Waugh et al., 2006; Nyhan, 2005; Scanlon et al., 2005; Albrightet al., 2004; Dwyer, 2003). In the case of cover systems over tailings, the contrast in hydraulicproperties between the tailings and cover material also can significantly affect the cover systemperformance. For example, Figure 7 presents in-situ soil water pressure head data collected at180 cm below ground surface under 30 cm and 60 cm coarse-grained cover material and a no (0 cm)cover (Milczarek et al., 2009). Wetting and drying patterns shown at 180 cm below ground surfaceindicate that under conditions of normal precipitation little to no wetting of the subsurface occurs atdepth with either cover system depth. However, when above-average precipitation follows very dryperiods, equivalent or greater wetting occurs at depth below the 60 cm cover than the 30 cm cover(i.e. August 2002 and July 2006). These data indicate that after periods of drought, differences inevapotranspiration rates could be diminished and the thicker profile of higher conductivity covermaterial over low conductivity tailings may actually result in increased net percolation due to morerapid downward percolation of precipitation through the upper 60 cm. Of note, the bare-tailingsplots consistently showed drier conditions than did the covered plots at the 180 cm depths. Thisresult is due to higher runoff rates from the bare tailings surface than from the cover material.

Table 2 presents estimated total and average downward flux across tailing reclamation treatmentsusing in-situ soil water pressure measurements and the simplified two-layer flux model described inMilczarek et al. (2009). Predicted downward fluxes through the 60 cm cover systems were slightlygreater than the 30 cm cover systems. The higher estimated flux rates through the deeper coversare due to observed lower-permeability tailings layers below the 30 cm cover plots than the 60 cmcover plots. With the exception of the bare-tailings plot, the average estimated flux rates are notsignificantly different. These predictions also indicate that the underlying tailings permeabilityhave a significant affect on cover system performance in controlling net percolation.

Finally, Figure 8 presents the predicted net percolation using a calibrated unsaturated flow model(UNSAT-H Fayer, 2000) and applying a 98-year climate record for a coarse-grained cover systemlocated over tailing beach sand and slimes areas. Model predictions indicate that for the tailingsbeach and sideslope materials the cover system could be expected to limit net infiltration to betweenapproximately 4.5 to 7 mm per year, depending on cover thickness. Increasing the cover thicknessfrom 45 cm to 90 cm was predicted to only nominally decrease the net percolation by 2.5 mm peryear. However, decreasing the estimated saturated hydraulic conductivity of the underlying tailings

127

Table 2. Estimated downward flux rates for different treatments.

Flux

Treatment cm/yr Percent of precipitation

30 cm cover, low vegetation 0.37 1.330 cm cover, high vegetation 0.12 0.360 cm cover, low vegetation 0.55 1.760 cm cover, high vegetation 0.48 1.5Bare tailings 0.02 0.1

Figure 8. Predicted net percolation with different cover material depths and tailings saturated hydraulicconductivity (Ksat).

to approximate hydraulic property differences between the slimes and beach areas showed greaterpredicted reductions due to reduced wetting front depths and subsequently higher available moisturefor evapotranspiration. The unsaturated flow model results and estimated flux rates presented inTable 2 indicate that increasing cover thickness can have less influence on net percolation than theunderlying tailings characteristics.

5 TAILING SOLUTION MIGRATION INTO COVERS

Low pH and high electrical conductivity (EC) of copper mine tailings in semi-arid and arid envi-ronments raise concerns regarding potential upward migration of salinity and acidity into the covermaterials. Limited upward salinity migration from acidic tailings into reclaimed mine-spoil coversoils has been observed on time-scales up to 25 years (i.e. Dollhopf et al., 2001; Dollhopf et al.,2003; Munk et al., 2006). Salinity and acid migration has also been observed to be negligible undermoderately acidic conditions and limited to approximately 15 cm above the cover and tailingscontact (Milczarek et al., 2009; Milczarek et al., 2010).

Figure 9 displays profiles of pH and EC relative to the tailings-cover material contact (Milczareket al., 2010). pH and EC results were observed to be highly variable across the test plots, however,samples generally displayed decreased pH and higher EC values within 5 cm to 10 cm above the

128

Figure 9. pH and electrical conductivity for different treatments.

tailings/cover system contact. EC and pH returned to near background levels within 15 cm abovethe contact. Similar tests performed five years prior showed that pH and EC values were essentiallysimilar over the five year period. The observed nominal effects in pH and EC migration are believedto be in part affected by cover material neutralization potential due to the cover material beingstrongly calcareous.

EC and pH effects from acidic tailings have been observed to be greater in shallow (i.e. 30 cm)than in deeper (i.e. 60 cm) cover systems at equivalent depths above the tailings/cover system contact(Milczarek et al., 2010). Because the tailings/cover contact in shallow covers is closer to the surfacethan in deeper covers, hydraulic gradients which drive upward advective flux may be greater at thetailings/cover contact. Diffusion may also be a secondary cause of decreased pH and increased ECat depths near the tailings/cover contact. However, if diffusion were the primary cause, EC andpH levels would be expected to be generally uniform across depths. Advection of tailings solutioninto the cover system is likely to be driven by episodic rainfall events that wet the tailings and arelimited to a very shallow region above the tailings/cover contact due to rapid decreases in hydraulicconductivity with distance above the contact due to drier conditions nearer the surface. Vegetationmonitoring results from several semiarid reclaimed tailing sites in the southwestern United States(i.e. Milczarek et al., 2009; Milczarek et al., 2010; Munk, 2006) indicate that pH and EC changesabove the cover contact has not negatively affected vegetative cover, rooting dynamics, or coverperformance for a range of cover system depths and varying tailings chemistries.

6 CONCLUSIONS

General findings from over a decade of tailings reclamation research and performance monitoringat five tailings facilities in the southwestern United States indicate that circumneutral tailingscan be directly revegetated with organic amendments or using a shallow cover. The effectivedepth of a cover system in supporting vegetation and controlling net percolation can range from15 cm for circumneutral tailings to 60 cm for acidic tailings. Revegetation seed mixes shouldconsider differences between beach sand and slimes areas as well as cover depth, such that mesicspecies can be used in slimes areas and xeric species will be more successful in beach/mixed areas.Deeper covers also may promote better success of native seed mixes. Plants can actively root intocircumneutral and moderately acidic tailings, indicating that water balance modeling of the coversystem should allow for evapotranspiration at depths into the tailings. Low permeability tailingsalso serve to slow down infiltration and retain water in the cover and can have a greater effect onnet percolation than does cover depth. Finally, upward acidity and salinity migration into covers

129

appears to be limited to shallow depths above the cover-tailings contact. These finding indicate thattailings affect the performance of store and release covers and their influence and spatial variabilityshould be considered during cover design.

REFERENCES

Albright, W.H., Benson, C.H., Gee, G.W., Roesler, A.C., Abichou, T., Apiwantragoon, P., Lyles, B.F., andRock, S.A. (2004) Field Water Balance of Landfill Final Covers. J. Environ. Qual. 33:2317–2332.

Barth, R.C. (1986) Reclamation Technology For Tailings Impoundments: Part 2. Revegetation. Mineral &Energy Resources vol. 29, no. 6, pp 1–24.

Bengson, S.A. (2000) Reclamation of Copper Tailings in Arizona Utilizing Biosolids, Mining, Forest and LandRestoration Symposium and Workshop, Golden, CO, July 17–19, 2000.

Borden, R.K. and Black, R. (2005) Volunteer Revegetation of Waste Rock Surfaces at the Bingham CanyonMine, Utah, J. Environ. Qual. 34:2234–2242.

Brown, S., Sprenger, M, Maxemchuk, A., and Compton, H. (2005) An Evaluation of Ecosystem FunctionFollowing Restoration with Biosolids and Lime Addition to Alluvial Tailings Deposits in Leadville, CO. J.Environ. Qual. 34:139–148.

Dwyer, S.F. (2003) Water Balance Measurements and Computer Simulations of Landfill Covers. PhDDissertation, The University of New Mexico, Albuquerque, NM.

Fayer, M.J. (2000) UNSAT-H Version 3.0: Unsaturated Soil Water and Heat Flow Model: Theory, User Manual,and Examples. PNNL-13249, Pacific Northwest National Laboratory, Richland, WA.

Fayer, M.J. and Gee, G.W. (2006) Multiple-Year Water Balance of Soil Covers in a Semiarid Setting, J. Environ.Qual. 35:366–377.

Milczarek, M.A., Yao, T.M., Vinson, J., Word, J., Kiessling, S., Musser, B., and Mohr, R. (2004) Monitoringthe Performance of Mono-layer Evapotranspirative Covers in the Southwestern United States. Design-ing, Building, & Regulating Evapotranspiration (ET) Landfill Covers, US EPA Remediation TechnologiesDevelopment Forum, Denver, CO, March 9–10.

Milczarek, M., Grahn, H., and Watson, A. (2006) The Development of the Tailings Closure Cover Design forthe San Manuel Plant Site. Hard Rock 2006, Sustainable Modern Mining Applications, US EPA, Tucson,AZ, November 14–16.

Milczarek M.A., Buchanan, M., Keller, J., Yao, T.M., Word, W., and Steward, M. (2009) Ten Years of TailingsReclamation Experiments at the Morenci Mine. 8th International Conference on Acid Rock Drainage, June22–26, 2009, Skellefteå, Sweden.

Milczarek M.A., Steward, F.M. Jr., Word, WB., Buchanan, M.J., and Keller, J.M. (2010) Salinity/pH Inter-actions and Rooting Morphology in Monolayer Soil Covers above Copper Tailings. Mine Closure 2010,November 23–26, 2010, Viña del Mar, Chile.

Munk, L., Jaworski, M., Jojola, M., and Romig, D. (2006) Upward Migration of Constituents in Soil Coversat Semi-Arid Mine Sites, 7th International Conference on Acid Rock Drainage, March 26–29, St. Louis,Missouri.

Munshower, F.F., Neuman, D.R, Dollhopf, D.J., Jennings S.R., and Goering., J.D. (1995) Revegetation ofStreambank Tailings Along Silver Bow Creek, Montana. Proc. 12th Annual Meeting of Amer. Soc. SurfaceMining and Reclam., pp. 729–740. Gillette, WY. June 3–8, 1995.

Nyhan, J.W. (2005) A Seven-Year Water Balance Study of an Evapotranspiration Landfill Cover Varying inSlope for Semiarid Regions, Vadose Zone J. 4:466–480.

Pond, A.P., White, S.A., Milczarek, M., and Thompson, T.L. (2005) Accelerate Weathering of Biosolid-Amended Copper Mine Tailings, J. Environ. Qual. 34:1293–1301.

Sauer, H., Williams, T., and Duvall, E. (2002) Revegetation of Nine Square Miles of Copper Tailings.Reclamation NAAMLP Annual Conference.

Scanlon, B.R., Reedy, R.C., Keese, K.E., and Dwyer, S.F. (2005) Evaluation of Evapotranspiration Covers forWaste Containment in Arid and Semiarid Regions in the Southwestern USA, Vadose Zone J. 4:55–71.

Schafer, W.M. (1979) Guides for Estimating Cover-Soil Quality and Mine Soil Capability for use in Coal StripMine Reclamation in the Western United States. Reclamation Review, vol. 2, pp 67–74.

Thompson, T.L., Wald-Hopkins, M., and White, S.A. (2001) Reclamation of Copper Mine Tailings UsingBiosolids and Green Waste. National meeting of the American Society of Surface Mining Reclamation,Lexington, KY.

Waugh, W., Smith, G., Danforth, B., Gee, G., Kothari, V., and Pauling, T. (2006) Performance Evaluation of theEngineered Cover at Lakeview, Oregon, Uranium Mill Tailings Site. Proceedings of the Waste Management2007 Symposium, University of Arizona, Tucson, Arizona.

130

Water management andwater treatment


Dewatered tailings practice – trends and observations

Michael DaviesAMEC Earth & Environmental, Vancouver, BC, Canada

John LupoAMEC Earth & Environmental, Denver, CO, USA

Todd Martin & Ed McRobertsAMEC Earth & Environmental, Vancouver, BC, Canada

Marcelo MusseAMEC Earth & Environmental, Santiago, Chili

David RitchieAMEC Earth & Environmental, Toronto, ON, Canada

ABSTRACT: The traditional impoundment where a dam(s) retains pumped slurried tailings withsolids contents typically in the range of 30–40% remains the most common method for storingthe tailings from milling operations. However, there are a growing number of operations whereby thetailings are dewatered and placed in the tailings storage facility with less accompanying water. Thedegree of dewatering and the method of placement can vary significantly dependent upon designcriteria and site specific constraints.

This paper reviews the state of dewatered tailings practice in a global sense. The review providesa summary of trends in the practice as well as some valuable lessons learned from observationsthroughout the world. Like many things that are relatively new to an industry, both excessiveoptimism and pessimism about the value and outcomes from using dewatered tailings have devel-oped. The paper provides insight to both ends of this spectrum and, through actual operating anddesign experience, offers pragmatic and supportable “real case” scenarios to many of the more con-tentious issues surrounding dewatered tailings practice. The paper addresses critical issues such as“dry landscape” concepts and beach slopes misconceptions for thickened/paste tailings. Finally,the paper provides some needed guidance in terms of nomenclature.

1 DEWATERED TAILINGS – WHAT ARE THEY?

1.1 Tailings water removal

Essentially any tailings can be dewatered, by definition, if water is removed by some processbetween the point of resource extraction and their subsequent placement in the tailings managementfacility (TMF). The simplest form of explanation is in considering the traditional slurried milledtailings stream that is typically, for hard rock mining, in the 30 to 40% solids range. This slurry canhave its effective solids content increased by dewatering through increasing mechanical interventionthrough the dewatering continuum as shown in Figure 1.

However, classification of tailings in the dewatering “continuum” can be confusing. This is par-ticularly true for thickened/paste tailings where the concept of segregating versus non-segregatingbecomes included in the definitions. There is clear understanding as to the meaning of segregatingversus non-segregating tailings as separated by a segregation threshold. At solids contents higherthan the threshold, and often as augmented by coagulants or flocculants, a tailings slurry afterdischarge does not separate into sand beaches and fines (also known as slimes or sludges) and

133

Figure 1. Tailings dewatering continuum.

releases only water. In some parts of the world, the term thickened tailings is used to describe arange encompassing non-segregating tailings just dry of the segregation threshold and paste tailingsfor tailings slurries that are at even higher solids contents. Both thickened tailings and paste areideally types of non-segregating tailings though the former is not necessarily so; hence just one ofthe challenges lies in nomenclature for these materials.

1.2 Cycloning or classification

One of the more common methods of creating a dewatered product is that which comes from theunderflow from a cyclone. However, a significant proportion of “coarse” material, typically morethan 50% sand size, is needed to be efficient in a single-stage operation. When it was introducedin the 1960’s, cycloning represented the first stage in the evolution of tailings dewatering, and wasdriven by the attractive economics associated with recovery of the coarse fraction of the tailingsfor embankment dam construction. Performance depends on grain size, equipment design, andequipment operation. Challenges can come in extreme cold climates and several operations havepoorly anticipated these challenges leading to severe mass balance problems (and, in one case, afacility overtopping failure). In a typical application, underflow (coarse sand) material is obtainedat about 70% solids concentration. Transportation may require positive displacement pumping(depends on the relative position of the cyclone station and the deposit itself). Often the cyclonedsand will be re-slurried to a lower water content to facilitate conventional pumping and distributionif the sand is being used for dam construction. If deposited in managed cells, it can be compactedto provide a very stable containment material; even in extreme environments (high seismicity andrainfall).

For success with cycloning, quality of the tailings material is a key factor too. Maximum finescontent is a limitation to ensure obtaining the amount of material and, more important, to facilitateadequate drainage to allow for compaction, and to allow development of a low phreatic surface.Overflow (fines, slimes) material is typically delivered from the process at about 30% solidsconcentration and it can also be dewatered and deposited upstream the of the cycloned sand shellof the TMF, increasing the recovery of water.

1.3 Thickened and paste tailings

Given the definitions that abound in current literature, there is often only semantics to separatethickened from paste tailings. Perhaps a convenient approach is to view the origin of thickened

134

disposal on a large commercial scale. Thickened tailings are a technique that has been proposedfor about 35 years and has been implemented in a few operations. The main premise of thickeneddisposal is that tailings may be thickened to a degree that they may be discharged from one orseveral discharge points to form a possibly non-segregating tailings mass with little or no waterpond. In the most classical connotation, thickened tailings are assumed to form a conical mass withthe tailings surface sloping downwards from the centre of the cone. A thickened tailings system,if successful, should require lower retaining dykes in compare to traditional slurry, as storage isgained by raising the centre of the impoundment.

Unfortunately, in many of the first instances where thickened tailings was implemented, thick-ening technology was not capable of producing a consistent non-segregating material, so fineswould form a very flat slope and require additional dyking at the toe. As well, flatter than pro-jected slopes were experienced, and it was not possible to steepen these slopes to avoid extensiveland use impact. Moreover, promised improved water management and control of everything fromliquefaction susceptibility to sulphide oxidation simply did not result in these initial efforts. Fromthe above experiences, thickened disposal did not become widely used in temperate regions. Ithas, however, been successful in very arid regions, such as the gold mining districts of Australia.In recent years, high density thickening technology has been developed which make it useful tore-examine thickened disposal. Further, irrespective of issues with anticipated versus achievedslopes and storage capacity utilization, thickening offers significant advantages, especially in aridregions where makeup water supply can be problematic, for water recovery.

With the addition of “paste”, where the tailings are even further dewatered, there is an increasingtrend to consider thickened/paste tailings for tailings management. Paste tailings has a long suc-cessful history for underground backfill but a “spotty” record for surface impoundments due in partto a lack of appreciation for the variations in mill feed and the resulting challenges in addressingthis variation in the thickening process. However, “paste” tailings in their best form can meet anumber of the promised attributes that came with the original visions for thickened tailings (in, forexample, Robinsky, 1975).

Using “paste” as a generic term is the heart much of the discussion with some parts of the industryusing the term to describe what many might be content to call thickened tailings. There can be clearcut quantitative criteria for a specific type of tailings for which ranges in solids content can beused to define significant changes in tailings performance, rheology, strength, pumpability and soforth. Common subjective definitions of paste are varied. One is that paste cannot be pumped bycentrifugal pumps, but requires positive displacement equipment. However, recent developments inpumping technology render this definition impractical. Another discriminator is that paste does nothave a critical velocity and therefore will not settle in a pipeline. Yet another is that on deposition,paste will have little or no bleed water. Part of the issue in so characterizing paste versus thickenedtailings is non-technical, with proponents of some disposal scenarios focusing on terminology – notperformance. In reality, discrimination between thickened and paste tailings involves considerationof many factors including:

• Type of dewatering facilities• Type of process aids, on-line chemical modifications• Conveyance of tailings to deposition area: positive displacement pumps, conveyors, trucks.• Amount of bleed water• Slope angle on deposition, deformations mechanisms• Can post depositional consolidation be predicted using effective stress based models or is

consideration of factors such as gel strength, creep and zone settling required• Strength: can geotechnical constructs be used, or is a rheological based approach more

appropriate• Is tailings self supporting, or are containment facilities required.

Maintaining this discrimination between thickened tailings or paste may not in fact be that useful.Unfortunately, there has been branding of the terminology to the point where it has become blindlyproposed on projects where it not the ideal tailings management solution. Even worse, it has shownup on projects where it has impacted project viability by simply being the wrong technology for

135

that given project. Part of the problem is what has been proposed in these cases is based upon anexpectation based upon a flawed terminology understanding.

1.4 Filtered tailings

Development of large capacity vacuum and pressure belt filter technology has presented the oppor-tunity for storing tailings in a solid state, rather than as conventional slurry and/or in the “pastelike” consistency associated with thickened or paste tailings. Tailings are dewatered to moisturecontents that are no longer pumpable and that are, in fact, below a material’s liquid limit. Thefiltered tailings are transported by conveyor or truck, and placed, spread and compacted to ideallyform an unsaturated, dense and stable tailings stack (often termed a “dry stack”) requiring no damfor retention. Filtered tailings, still considered primarily for operations under 20,000 tpd but thattrend is changing quite quickly, can have attraction for projects with the following attributes:

a) arid regions, where water conservation is an important issueb) situations where economic recovery is enhanced by tailings filtrationc) where very limited space and/or very high seismicity contraindicates some forms of conventional

tailings impoundmentsd) cold regions, where water handling is very difficult for significant portions of the year.

Moreover, filtered tailings stacks have regulatory attraction, require a smaller footprint fortailings storage (lower bulking factor), are easier to reclaim, can be reclaimed in a progressivemanner, and have lower long-term liability in terms of structural integrity. One challenge withthe technology can be defaulting to the less-expensive vacuum filtration for tailings materials thatrequire pressure filtration to achieve the required moisture content (Davies and Rice, 2001).

A curious aspect of filtered tailings is that it is used on substantially more operating mines thanpaste tailings yet there are but a few publications on filtered tailings versus the extensive (andgrowing) number of papers, specialty conferences and even texts on paste tailings.

2 DEWATERED TAILINGS – NO TAILINGS PONDS?

The trend to the concept of a “dry landscape” once a mine decides to start dewatering their tailingsis indeed seductive but seldom is it accurate. The only case where “no pond” can be stated withsome confidence is for filtered tailings and then only where there is sufficient system redundancyto handle feed variations and limits to equipment availability. One culprit in the “no pond” myth isa surprising large portion of the recent literature includes statements or implications that adoptionof thickened/paste technology will eliminate tailings ponds. This is particularly curious in non-aridor cold regions where operating without some form of a tailings pond is a near crippling constraintto put on a mill. While it might be the case that no pond is required in some centrally dischargedtailings stacks in very arid climates, in many others it is not. In fact, even applications of thickenedtailings in arid environments often have bleed water ponds. In other words, unless a project is at avery unique combination of tonnage, climate and feed consistency, if the dewatering technology isnot filtered tailings requiring trucking or conveying, then planning around not having any form ofa tailings pond is contraindicated by experience. Some projects involving central stack dischargehave been promoted as dry tailings, notably projects in Canada, and have ended up with extensivepond areas to manage a range of other requirements such as recycle water clarification due to solidsfrom non-segregating or off-specification episodes, concerns about sulphide oxidation and rainfallor snowmelt runoff.

To that end, there is also a prevailing theme by some in the literature that tailings ponds themselvesare inherently “bad” whereas the opposite is often the case during the operating phase of a mine.A tailings pond allows:

• Storing the pore fluids expressed through bleed and consolidation processes.• Saturation which limits oxidation of sulphide bearing constituents, as experience to date with

thickened/paste tailings demonstrates oxidation of these materials will occur.• A facility to store and provide water during drier periods.

136

• A location to allow clarification of recycle water; often process requirements result in a need forsignificant areas to meet those requirements. Note however that optimal process water recoveryin many environments is achieved via some level of dewatering.

• Provide a mixing zone to manage ionic levels.• Water management buffering – storage of storm water, direct and runoff, that becomes tailings

contact water and cannot be directly released to the environment.• Geothermal factors in cold climates: a large pond can be a large heat source. A small pond with

very limited water has much less of a heat reservoir. When coupled with long beaches, pondtemperatures can readily approach the freezing point and can significantly complicate winterpond operations, potentially including the inability to obtain reclaim water due to excessive iceand inadequate water pond depths. However, dewatered tailings can provide decent heat recoverwhere the tailings are processed at elevated temperatures relative to the environment.

3 THICKENED AND PASTE TAILINGS BEACH SLOPES

Another important trend in dewatered tailings practice is the better understanding of the actualmechanisms involved with beach slopes in thickened/paste tailings. The design of thickened orpaste tailings stacks, or central discharge operations using some form of non-segregating tailingsin conjunction with desiccation or drying is used in perhaps 20 to 30 operations (Fitton, 2007).For tailings intended to be self supporting in a large cone the specification of the design angle ishighly critical, and several stacks have been laid out with design slopes that were far steeper thanthose ultimately achieved in the real operating conditions. The common theme of many such slopedesigns (i.e., for thickened tailings or paste tailings) is the reliance on small scale flume or “fish-tank tests” which while universally appealing given their common use, are in fact fundamentallywrong.

Non-segregating tailings flow on beaches does not occur via simple sheet flow observed in verylow discharge rate experiments. It can be observed that achieved slope angle can be correlated to theheight or length of a deposit. The concept of “sheet flow” provides a simple model that supposedlyjustifies the use of small scale flume tests. The equation for sheet or skin flow, where during thedeposition of the slurry lift, the rheological properties of the slurry in a zone settling mode and thelift thickness determine the slope angle achieved is as follows:

where:θ = slope angle,Cu = undrained strength of the slurry (as determined from rheological considerations of yield

strength and strain rate),γ = unit weight of the slurry, andd = lift thickness measured perpendicular to the slope.

Inspection of Equation 1 indicates that three parameters must be manipulated to achieve a desiredslope angle. Consider a slurry discharged at a constant continuous rate of volume Q, and at a fixedsolids content or constant operative strength, Cu. Consider also a conical shape of Height h, raisedby continuous discharge of tailings Q at the tip of a cone (Figure 2).

Of interest is the incremental thickness of tailings on the cone surface with time, in this instancerepresented by an increment in volume placed which is assumed placed evenly around the conegenerated by the surface at a given height.

From this is can be seen that the “dz/dv” is function of the cone slope and the current height. Ifthe slope angle is fixed then the only way to get sheet flow to all parts of the cone is to continuallyadjust either Q or Cu. This conundrum is well summarized by Pirouz et al. (2000).

137

Figure 2. Idealized tailings cone.

Pirouz et al. (2000) state:

If a very large area of a tailings stack is eventually to be developed it can be intuitivelyunderstood that this will not occur over the full 360◦ and the full radial distancesimultaneously. This would require a progressively thinner and thinner sheet eachday, and/or a progressively steeper and steeper slope in accordance with sheet flowequations. Furthermore the radial velocity would have to also increase since the timefor zone settling within a layer is fixed, and the flow must reach the outermost perimeterof the area before this occurs.

The second problem with a simple sheet flow model is that the sheet flow can only move so farbefore it stops or self arrests due to dewatering. When a fresh non segregating layer is deposited,dewatering occurs due to several mechanisms. Firstly, the solids content of the tailings increases dueto zone-settling via a mechanism first presented by Kynch (1952), and first discussed in the contextof soil settling by McRoberts and Nixon (1976), and the theory expanded by Bartholomeeusen(2003). The procession is as follows:

• The zone settling process is relatively rapid especially when fine tailings are highly flocculatedby the addition of process aids.

• As the fine tailings participates in zone settling, they also move downslope, either by sheet flow(commonly assumed) or by flow channelization (commonly observed). As the tailings movedownslope, water loss occurs downwards into the previously desiccated layer causing the finetailings to consolidate.

• The combination of zone settling and downwards flow causes fine tailings to eventually “freeze”in place or self arrest; however, as discussed further below, flow channelization complicates thisprocess.

• Once downslope movement ceases, water loss occurs upwards and downwards. Depending on thetailings characteristics and surface slope, desiccation may initially be inhibited by consolidationwater arriving at the surface at a flux rate greater than the evaporative rate.

• Eventually, desiccation is initiated and consolidation is enhanced by the imposition of tension inthe pore water within the desiccation zone drawing water upwards. As the tailings consolidates,the rate at which water can be driven out is reduced because the hydraulic conductivity of thetailings reduces.

138

A key factor in understanding non-segregating fine tailings behavior stems from the work ofM.P.A. Williams who has identified the observed phenomenon of flow channelization as distinctfrom sheet flow. This is clearly stated by Williams (1992) as follows:

It is necessary to understand the mechanics of beach formation which, contrary tointuitive expectations does not occur as a result of ever expanding radial sheets ofslurry flow. Rather the flow concentrates into a self formed channel. From time to timethe direction of flow from the plunge pool below the point of discharge changes andbreaks out over a new area of beach…”

Flow channelization has been observed in essentially every thickened or paste tailings depositdeveloped. If flow is maintained by channelization, then such flow can be turbulent and zone settlingin laminar flow will not occur. While channelization will move slurry out over longer distancesfrom the discharge point, it is thought that the time for second water release is extended. Moreover,turbulent flow may affect the integrity of flocculation, as it is know that shear-thinning can affectthe rheological properties of flocculated slurries. That is to say, segregation thresholds obtainedfrom column tests or low energy flumes are not valid under shear. Fitton (2007), following on thework of Williams, (1992) provides a lucid explanation of the phenomenon of flow channelization.

In summary, experience and theory indicate that it is invalid to assume deposition slopes andprocesses that are independent of slope length, deposit height, and are derived on the basis of flumetests that due to scale inevitably exclude many of the key processes, including post-depositionaldewatering, driving performance at a field scale. There can be plenty of useful information gainedfrom bench scale tests with dewatered tailings but providing an indication of what the operatingslopes will be is not one of them.

4 CURRENT GLOBAL TRENDS

Initiatives to limit the amount of water sent to tailings management facilities is increasing in allmajor mining districts in a global sense. The reasons vary in each case but the clear objective in eachcase is to reduce water consumption/water losses related to tailings management. Some specifictrends that have been noticeable are described below.

In Chile, there are a high number of large tonnage productions with commensurate high costsof equipment and pumping, large containment embankments and design concerns with the impactof large seismic and infrequent, but large, rainfall events. Important potential savings of waterexist at most of the major mines in Chile and dewatered tailings are being considered on multipleprojects and being implement on several. Examples of actual dewatered tailings projects currentlyin operation include:

• La Coipa (filtered)• Mantos Blancos (filtered)• El Peñón (filtered).

The El Indio Mine (currently closed) was also a dewatered tailings operation having had a seriesof conventional and dry (filtered) tailings facilities.

Projects that will soon be in operation include the Las Cenizas Project (small scale, paste) and theEsperanza Project which will be Chile’s first large scale paste tailings project. Other projects haverun multiple studies at different engineering levels, but results have not been sufficient compelling(technically and/or economically) to make the change from conventional tailings though the use ofcycloned classification of tailings remains a very common practice in Chile. Of interest, Chileanregulators have already included different types of dewatered tailings in governance literature.

In Canada, there is a slow, but steady, increasing use of dewatered tailings but on large scaleprojects, like in the Alberta Oil Sands, commitment to full-scale production has been slow due tochallenges at large scale. Lower tonnage in the northern regions of Canada (and Alaska), have morereadily embraced thickened/paste and filtered tailings and some of the best global examples of thesetechnologies exist in this region. The Pogo Project in Alaska, for example, has been operating a

139

Figure 3. Trends in use of dewatered tailings in mining.

very efficient filtered dry stack since 2006, Raglan Mine in Northern Quebec several years longer.However, some of the least successful thickened tailings facilities also exist in Canada where theseprojects have provided some valuable lessons learned to the industry as a whole but significantnegative impact to the mines that utilized this technology.

In Australia, filtered tailings have not been embraced quite as quickly as in other regions butthe region is by far the leader in thickened tailings applications. The pioneering work by M.P.A.Williams and his colleagues has resulted in some excellent project examples in the area of thickenedtailings for use in arid environments at modest tonnage. This same expertise has been applied atmuch higher throughputs in Middle Eastern countries such as Iran where scarce water resourceshas driven the need for dewatered tailings.

Besides regional trends, another important trend is the size of the facilities using dewateringtechnologies. For the most part, the current operations using thickened/paste tailings are mostlyunder 30,000 tpd whereas most filtered operations are generally under 10,000 tpd. There areexceptions in each case with operations of greater throughputs currently existing. Moreover, theauthors are aware of proposed mines in advanced project stages where thickened tailings in excessof 100,000 tpd and filtered tailings in the range 70,000 tpd. This significant increase in potentialthroughput has come from a combination of increased economies of scale with advances in thedewatering equipment and increased “drivers” for having dewatered tailings as the base case formany new projects.

Figure 3 provides a summary of the relative (and somewhat actual) number of dewatered facilitieson a global scale. While this paper does not specifically address co-disposal (fine tailings mixedwith a coarser material, often waste rock), it is a dewatered tailings that is used in such applicationsand for that reason the trend in these projects was included in Figure 3.

5 DEWATERED TAILINGS – THE FUTURE

Using cyclones to classify tailings, and effectively dewater the underflow, started in the late 1960son an appreciable scale, driven by the desire to recover the coarse fraction of the tailings as aconstruction material for embankment dams. The first thickened tailings facilities were being con-ceptualized and trialed in the mid 1970s, filtered tailings roughly 10 years later and paste facilitiesstarting about 1990. Through misapplication and other considerations, none of the dewateringmethods, other than cycloning, gained much initial momentum. However, since about 2000, allmethods of dewatering noted and some newer concepts such as centrifuging, have become moreroutine in design consideration.

140

Tabl

e1.

Dew

ater

edTa

iling

sC

lass

ific

atio

n.

Thi

cken

edco

agul

ants

/flo

ccul

ants

Past

eC

onve

ntio

nals

lurr

ycy

clon

es/th

icke

ners

deep

bed/

cone

thic

kene

rsC

ake

Dew

ater

ing

faci

litie

s1no

ne(c

yclo

nes

poss

ible

)on

-lin

ech

emic

alm

odif

icat

ions

addi

tiona

lche

mic

als

filte

rs(v

acuu

mor

pres

sure

)

Typi

calS

lurr

yD

ensi

ty2

30–4

0%so

lids

45–6

5%so

lids

65–7

0%so

lids

80–8

5%so

lids

(uns

atur

ated

)Ty

pica

lcon

veya

nce

Cen

trif

ugal

pum

psPo

sitiv

edi

spla

cem

entp

umps

Tru

cks

orco

nvey

ers

Vis

cosi

tyre

duce

rsB

leed

Wat

erFi

nes

topo

ndw

ithC

onsi

dera

ble

wat

erto

man

age

Litt

leto

nobl

eed

wat

erto

man

age

Con

solid

atio

non

ly–

can

sign

ific

antw

ater

tom

anag

ebe

negl

igib

leD

epos

itC

hara

cter

Segr

egat

ing

Poss

ibly

segr

egat

ing

Non

-seg

rega

ting

Soil-

like

(non

-seg

rega

ting)

Phys

ical

stre

ngth

mod

elFl

uid

rheo

logy

Flui

drh

eolo

gyT

rans

ition

toso

ilm

echa

nics

Soil

mec

hani

csB

each

Slop

esFl

atw

ithco

ncav

ityL

esse

rco

ncav

ity.S

light

lyA

ppro

achi

ngsh

eetf

low

N/A

stee

per

than

conv

entio

nals

lurr

y.bu

tstil

lhas

chan

neliz

atio

n.B

each

esfo

rmth

roug

hB

each

ing

can

bein

hibi

ted

chan

neliz

edfl

owby

stre

ngth

Post

Dep

ositi

onal

Cre

epm

ayst

illef

fect

non-

linea

rfi

nite

stra

inpr

edic

tions

App

roac

hing

trad

ition

alT

radi

tiona

lsoi

lmec

hani

csm

odel

sSe

ttlem

ent(

Con

solid

atio

n)so

ilm

echa

nics

mod

els

Con

tain

men

tofT

ailin

gsR

eten

tion

stru

ctur

esfo

ral

lC

anbe

self

-sup

port

ing

onC

anbe

self

-sup

port

ing

onSe

lf-s

uppo

rtin

gat

high

suba

eria

ldep

ositi

onve

rylo

wan

gle

slop

es.

mod

ests

lope

s.C

anpo

tent

ially

angl

esl

opes

with

out

Mod

estr

eten

tion

stru

ctur

esbe

deve

lope

dw

ithne

edfo

rre

tent

ion

stru

ctur

esst

illre

quir

edm

inim

alre

tent

ion

stru

ctur

es

1C

entr

ifug

es,n

otlis

ted,

can

deve

lop

taili

ngs

prod

ucti

nter

med

iate

topa

ste-

cake

.2

Typi

cald

ensi

ties

are

for

min

eral

taili

ngs

from

met

alm

inin

g.V

aria

tions

inbe

havi

orw

ithot

her

min

edm

ater

ials

can

beex

pect

ed.

The capital and per ton operating cost of the various dewatering options continue to improveas their scale of use increases. This economy of scale is being met by the increasing “cost” ofwater use (real and political) to the point where for many projects, dewatered tailings is not onlythe best technical option it is also the most logical option from an overall economics perspective.Moreover, as life-cycle costs (including closure) are more comprehensively factored into feasibilitystudies and project economics, the perceived capital costs associated with dewatering become morepalatable in many cases.

The collective experience of the authors indicates that dewatered tailings have become a viablealternative for mining projects of essentially any tonnage with the main limitation being in ensuringthe right technology is selected based upon the nature of the tailings, not some misconception takenfrom an alarming amount of misinformation available in current literature. With the increased useof these methods, more case records will be developed and better information will flow to operators,designers, regulators and educators and improved decisions will result.

Finally, a classification guide provided as Table 1 is provided to assist with the nomenclaturechallenge and to assist those not as familiar with the dewatering options available to the currentmining industry.

REFERENCES

Bartholomeeusen, G., 2003. Compound Shock Waves and Creep Behaviour in Sediment Beds. A thesissubmitted for the degree of Doctor of Philosophy, University of Oxford.

Davies, M.P. and Rice, S., 2001. An alternative to conventional tailing management – “dry stack” filteredtailings. Tailings and Mine Waste 2001.

Fitton, T., 2007. Tailings Beach Slope Prediction. A thesis submitted in fulfillment of the requirements for thedegree of Doctor of Philosophy, RMIT University

Jewell, R.J. and Fourie, A.B., 2006. Paste and Thickened Tailings – A Guide (Second Edition). Australia Centrefor Geomechanics. 257 p.

Kynch, G.J., 1952. A Theory of Sedimentation. Transactions Faraday Society, 48, 166–176.McRoberts, E.C. and Nixon, J.F., 1976, A Theory of Soil Sedimentation. Canadian Geotechnical Journal.Pirouz B., Kavianpour M.R. and Williams M.P.A. 2000. Thickened Tailings Beach Deposition, Field

Observations and Full-Scale Flume testing. Paste 2000 Santiago, Chile.Pirouz, B. and Kavianpour, M.R., 2005. Thickened Tailings Beach Deposition. Field Observations and Full-

Scale Flume Testing. Paste 2005, Santiago, Chile.Robinsky, EI 1975. ‘Thickened discharge – A new approach to tailings disposal’, CIM Bulletin, vol. 68,

pp. 47–53.Robinsky, E.I., 2000. Thickened Tailings in the Mining Industry.Williams, M.P.A., 1992. Australian Experience with the Central Thickened Discharge Method for Tailings.

Environmental Issues and Waste Management in Energy and Minerals Production, Singhal et al. (eds)Williams, M.P.A. and Seddon, K.D., 1999. Thickened Tailings Discharge: A Review of Australian Experience.

Tailings and Mine Waste 1999.

142


Groundwater modeling at the Panna Maria uranium facilityin support of an ACL application

M. Gard & J. WarnerGard Water Consulting, Loveland, Colorado, USA

L. CopeSRK Consultants, Fort Collins, Colorado, USA

K. RaabeRio Grande Resources Corporation, Hobson, Texas, USA

ABSTRACT: Alternative Concentration Limits (ACLs) are a viable option for groundwater com-pliance at mining properties. The numerical groundwater flow and transport mode is a critical toolto establishACLs at a point of compliance (POC) that can be shown to be protective at downgradientpoints of exposure (POE). A three-dimensional variably-saturated multilayer groundwater modelwas developed to support an ACL Application for a uranium mill tailings facility. The objective ofthe model was to simulate and predict seepage from the Panna Maria Tailings Impoundment andthe long term net effect this seepage has on the groundwater in the various aquifers in the vicinityof the site and ultimately on the San Antonio River. The model domain incorporated regional andlocal groundwater divides and simulated unsaturated and saturated flow and contaminant transport.It was developed using MODFLOW-SURFACT to simulate flow and transport through variablysaturated clay units that lie between the tailings and underlying near surface and deeper, regionalsand units. The resulting calibrated model demonstrates the concentrations to be defined at POCmonitoring wells that will be defined as ACLs that are protective of human health and the environ-ment at downgradient POE locations. The ACLs will be used to establish groundwater protectioncriteria that are protective of human health and the environment from potential releases from thefacility. The proper calibration and application of this model will provide a reliable and defensiblemodel and a pathway for prompt regulatory approval of the ACLs.

1 INTRODUCTION

The Panna Maria Uranium Operation Facility (Facility) is located in Karnes County, Texas, near thetown of Hobson, Texas. The Facility was operated as an open pit uranium mine and conventionalmilling operation. After the completion of mining, the mining pit and mill site were closed andreclaimed. The tailings impoundment was reclaimed using a multilayer soil cover. The cover wasdesigned to limit the release of radon gas from the tailings, to manage the runoff from the reclaimedimpoundment, and to limit precipitation infiltration. The multilayered cover consists of randomfill that was placed as an interim cover, and an infiltration barrier of clayey soil having a hydraulicconductivity of no greater than 1 × 10-8 cm/sec. Above the infiltration barrier, a soil cover wasplaced to enhance vegetative rooting and further reduce the emission of radon gas. The surface ofthe cover consists of topsoil.

The objective of the groundwater flow and transport model was to establish alternate uraniumand sulfate concentration limits at the points of compliance. The model accurately simulates andpredicts seepage from the Panna Maria Tailings Impoundment and the corresponding long termnet effect this seepage has on the groundwater in the Alluvial Sand, A-Sand, slough sediments,B1-Sand and ultimately the San Antonio River. The model is expected to accelerate regulatoryapproval and reduce the overall time and costs of the approval process.

143

The model was developed using MODFLOW-SURFACT Version 3.0, (Modflow-Surfact, 1996),a variably saturated, fully integrated flow and transport code, based on the U.S. Geological Survey(USGS) groundwater modeling software, MODFLOW (McDonald and Harbaugh, 1988). Thismodel was implemented to address the variably saturated clays in the alluvium and between thealluvium and the underlying Catahoula sandstones. The groundwater model was calibrated togroundwater elevations and chemical analytical results collected from 1992 to 2008. The calibratedmodel then simulated the 1,000 year predicted fate and transport of uranium and sulfate.

2 SITE DESCRIPTION

The Facility is located approximately six miles north of Karnes City, Texas and is located approxi-mately one half to one mile northeast of the San Antonio River. The Facility was operated as anopen pit uranium mining site with onsite uranium milling and onsite waste disposal ponds andtailings impoundment. Approximately 6.8 million tons of uranium ore was processed at the Facilityduring active onsite mining that occurred from 1977 until 1992. Reclamation of the Facility wasinitiated in 1992 and was completed in 2000.

2.1 Geologic settings

The site is underlain by southeast dippingTertiary sediments, consisting primarily of poorly consoli-dated siltstones, claystones, claystones and fine grained sandstones. In the site area, these sedimentsinclude the lower sequence of the Catahoula Formation, underlain by the Tordilla Member of theWitsett Formation of the Jackson Group.

The geologic formations of primary focus beneath the site are the Catahoula Formation andoverlying alluvial sediments. The Catahoula Formation consists of poorly consolidated claystonesand alternating sandstones. This formation contains distinctive fluvial channel-fill, crevasse splay,flood-plain, and lacustrine facies, which tend to persist vertically through the section. The CatahoulaClay composition reflects alteration to montmorillonite and kaolinite of large volumes of depositedvolcanic ash.

Situated above the Catahoula Formation are alluvial and fluvial sediments that consist ofunconsolidated clays, silty clays, silts and silty sands. The deposition and subsequent erosionof these sediments began penecontemporaneously (Miocene) with the deposition of the CatahoulaFormation and continued into the Quaternary.

The upper portion of the Catahoula Formation and the overlying unconsolidated sediments hasbeen subdivided into eight hydrostratigraphic units (HSUs) for the purpose of numerical modeling.More than one hundred well logs were evaluated to create the model’s geologic layering. Thegeologic data were interpreted using the Environmental Visualization Software package (EVS). Allof the available well logs were collected and analyzed to determine the various geologic units at eachlog’s specific spatial location. These data were recorded on a log by log basis and entered into anEVS database. EVS implemented three-dimensional kriging, a geostatistical gridding algorithm, togenerate the three-dimensional geologic structure. The EVS-generated geologic layering allowedfor a three-dimensional visual analysis of each geologic layer. EVS also enabled analysis of crosssections to maintain the integrity of the data within the geologic conceptual model.

The HSUs are summarized below from oldest to youngest.

2.1.1 B1-SandThe B1-Sand is the lowest HSU included in the flow and transport models. The B1 Sand consistspredominantly of a silty to clayey sand. The average thickness of the B1-Sand is twenty to twenty-five feet. The B1-Sand can be as thick as fifty feet in the center of the model and thins to zerothickness in the northwest where it has been eroded.

2.1.2 B1-ClayThe B1-Clay (Catahoula) is situated above the B1-Sand. It is the first HSU that is contiguous thoughout the model domain. The B1-Clay consists predominantly of stiff hard sandy silt, with regions ofhard silty sandy clay. The B1-Clay thickens substantially along the dip slope to the southeast to amaximum thickness of about one hundred feet.

144

2.1.3 A-SandThe A-Sand deposits include the first encountered sand unit beneath the eastern portion of thetailings impoundment. The A-Sand deposits also extend beyond the tailings impoundment to thenorth, east and northeast to some extent. The A-Sand consists primarily of fine grained, silty sand,ranging from sandy gravel to very dense silty or clayey sand and layers of gravel. In addition, theA-Sand has scattered caliche deposits.

TheA-Sand consists of two main units. The lowerA-Sand unit consists of pedogenically modifieddeposits of Quaternary valley fill and is most likely older than the Alluvial valley fill sand. TheA-Sand also includes a younger overlying tributary alluvium. Typically at the interface betweenthe two main A-Sand units there are intermittent layers of cemented silty or clayey sand (WWL,1993).

The A-Sand deposit is absent in much of the model domain. It occurs primarily beneath andto the northeast of the tailings impoundment. Thickness of this unit ranges from zero to as muchas twenty feet beneath the tailings impoundment. Some A-Sand deposits have been interpreted toexist east and north of the tailings impoundment. The A-Sand is truncated immediately south ofthe tailings impoundment and is interpreted to have been truncated by subsequent erosion. TheA-Sand has been eroded by and is in direct contact with the Alluvial Sand on portions of the southside of the tailings impoundment. The margin of this contact is characterized by the presence oflow permeability clay representing bank and overbank deposits.

Additionally, the A-Sand is present near the surface of the Manka Slough. This previouslyunnamed slough is located east of the tailings impoundment and the Manka Ponds are locatedwithin this slough. The A-Sand deposits have been eroded by the Manka Slough drainage system.Groundwater discharges to the Manka Slough sediments in this area. The A-Sand is also presentnear the surface east of the Manka Slough. The A-Sand subcrop is important hydrogeologicallybecause it may provide a direct pathway to the subsurface waters of the slough sediments.

2.1.4 A-ClayThe A-Clay is a silt, silty clay or clay deposit situated near the ground surface. The clay hasbeen eroded by the Miocene-aged streams that deposited the alluvium. The modern alluvium,where present, has typically dissected or truncated the A-Clay. The A-Clay is generally dry butgroundwater has been encountered in isolated locations. This unit is discontinuous and pinchesin and out within the model domain. Geologic logs describe the A-Clay as predominantly stiff ordense sandy silt to sandy clay. The A-Clay is not contiguous throughout the model domain. TheA-Clay varies in thickness from zero to a maximum of about forty feet within the model domain.

2.1.5 Alluvial sandTheAlluvial Sand generally describes the sands and clayey sands that are first encountered under thesouthwest portion of the tailings impoundment and that extend primarily to the south. The AlluvialSand is an important hydrogeologic unit in the area. Distinct from Quaternary Alluvium, they areOligocene in age, deposited penecontemporaneously with the clay deposits of the Catahoula. TheAlluvial Sand is discontinuous within the model domain and as discovered in the 1992 investigation,is separated by a horseshoe shaped, primarily fine grained silty clay ridge to the south of the tailingsimpoundment (WWL, 1993).

The Alluvial Sand is not vertically continuous and is often divided by silty clay to clayey siltlenses. This was defined during the 1992 investigation and was determined to be vertically separatedin places by a muddy floodplain or paleosol deposit. The Alluvial Sand unit is generally saturatedto various degrees and is typically unconfined.

2.1.6 SloughsIn addition to the HSU layering at the site, the geologic conceptual model contains several surficialand near surface alluvial deposits in sloughs that are interpreted to represent surficial expressions ofolder Miocene-aged Catahoula streams. These streams may have devolved into the existing sloughsdue to fine grained sediments deposited during flood events. The existing Lake and Manka Sloughsfound on the aerial photos and USGS topographic maps are thought to be the remnant expressionsof these paleo-streams.

Subsequent to the paleo-stream sediment deposition, the sloughs in the flood plain were coveredthrough time with fine-grained flood deposits. Because of their coarse texture and corresponding

145

high relative hydraulic conductivity, the sloughs were considered preferential transport pathwaysfor potential off-site migration of potentially impacted groundwater. The slough sediments consistof poorly sorted sands and silty or clayey sands. Some locations contain silty clay or gravel lensesand carbonate nodules are sometimes present.

2.1.7 Surficial soilsThe surficial unit consists of surficial deposits including the topsoil and tailings. The topsoil consistsof a variety of materials, ranging from loose organic rich silty sand with gravel to an organic richsandy loam to a stiff sandy silt or clay. Surficial soils in the floodplain typically contain smallgypsum crystals. The tailings consist of sandy clay or silt with some interbedded deposits ofchemical precipitates or slimes.

2.2 Waste disposal facilities

The waste disposal facilities at the Facility consisted of a mine and mill drainage pond, an industrialwaste pond, a molybdenum storage pond, decantation ponds and a major tailings impoundment.The Facility placed the tailings and effluent from the milling process into the tailings impoundment.An acid leach procedure utilizing sulfuric acid was used in the milling process at the Facility; hencethe water in the tailings impoundment has a very low pH, with a historical range between 2 and 6standard units. During reclamation the fluid and wastes that accumulated in these auxiliary pondswere transferred to the main tailings impoundment and the ponds were backfilled with clean soil.

The tailings impoundment was constructed using dikes that contained a central core of lowpermeable clay. The tailings impoundment dikes consisted of a zoned earthen embankment con-sisting of a central core flanked by shell zones with a blanket drain system on the downstreamside (IECO, 1977). The central core was designed to be a minimum of ten feet wide at the top andtwenty feet wide at the bottom where it was keyed through the Alluvial Sand or A Sand and at leasttwo feet deep into the underlying Catahoula Clay. The central core was designed to be relativelyimpervious and was constructed of select mine overburden material with an average permeabilityof 1 × 10−6 cm/sec. The inner portion of the tailings impoundment was not excavated down to theCatahoula Clay; therefore the tailings were placed directly on top of the existing Alluvial Sand orA-Sand units. The tailings impoundment covered approximately 150 acres and had a capacity of10 million tons of tailings (Shepherd Miller, 2002). In 1996 construction began on a reclamationcover for the tailings impoundment, which was constructed to minimize infiltration of water intothe tailings.

2.3 Hydrogeologic conceptual model

Groundwater movement follows four primary transport pathways from the closed tailings impound-ment to discharge points into the San Antonio River. The pathways are illustrated on the figurebelow. The easternmost pathway consists of the Manka Slough sediments. Fluids from the tailings

146

impoundment seep through or beneath the clay core along the eastern margin of the tailingsimpoundment and are intercepted by the Manka Slough sediments. The tailings fluid is comin-gled with groundwater from up gradient and east of the tailings impoundment. The Manka Pond issituated just east of the tailings impoundment and groundwater can be observed rising to the surfacedown gradient of the constructed dam as a result of under flow and dam seepage. These fluids arethen transported down the slough sediments onto the floodplain of the San Antonio River. Thefluids in the Manka Slough sediments are then captured by the paleo-San Antonio River sediments.The groundwater then migrates through the San Antonio Paleochannel to the current San AntonioRiver.

The second and third transport pathways begin on the southwest corner of the tailings impound-ment. The tailings fluids seep through or beneath the clay core at this location and are intercepted bythe Alluvial Sand sediments. Just south of Farm to Market (FM) Road 81, an A-Clay ridge bisectsthe Alluvial Sand and forms two separate transport pathways. The eastern limb of the AlluvialSand transports seeped tailings fluids downgradient until intercepted by the paleochannel of theSan Antonio River.

The western limb of the Alluvial Sand is a significant transport pathway for the tailings fluid.Transport within the western limb of the Alluvial Sand is intercepted by the sediments of the LakeSlough along the western margin of the Alluvial Sand deposit. The Lake Slough sediments thenprovide a pathway to the floodplain sediments where they are intercepted by the paleochannel ofthe San Antonio River. Transport continues until the paleochannel is intercepted by the currentriver.

The fourth transport pathway is downward through the B1-Clay into the B1-Sand. Only a minoramount of the tailings fluids are found along this pathway.

3 GROUNDWATER MODEL

The model’s geologic layering structure was based on the conceptual geologic model discussedpreviously and incorporated the hydrogeologic conceptual model. The groundwater flow system atthe Facility was simulated based on the post-mining conditions. Initially, a steady-state model wasdeveloped for the summer of 1992, a time period that coincided with an extensive hydrogeologicfield investigation. The 1992 investigation yielded key spatially-distributed groundwater level datafor theAlluvial Sand andA-Sand units that was otherwise lacking. The 1992 groundwater elevations,coupled with the groundwater data collected from the quarterly monitoring program, produced asignificant amount of spatially distributed groundwater elevation data. The calibrated steady-stategroundwater elevations were used as initial heads for the transient flow model.

Subsequently, a transient groundwater model was developed to simulate the conditions collectedbeginning July of 1992 until the end of December, 2007. The transient model was calibrated tomeasured quarterly groundwater elevations, the precipitation record, and well extractions.

3.1 Model domain

The model grid extends to natural flow boundaries in the region that include the San Antonio Riverand natural groundwater flow divides. The total model grid extends 4.27 miles by 3.83 miles andconsists of 1,344,168 total cells; with 224,028 cells per layer. The active grid consists of 1,170,510total cells, with 195,085 active cells per layer and the active model domain is illustrated in theaccompanying figure. A very fine grid spacing of 25 feet by 25 feet was used in the vicinity ofthe Panna Maria tailings impoundment and extends south, encompassing the southern portion ofthe Lake Slough. The fine grid, relative to the size of the model domain was required to accuratelysimulate the constituent transport through the hydrogeologic units of concern, due to high ground-water gradients between the tailings impoundment and the adjacent sediments. Outside this area ofthe fine grid, grid spacing was gradually increased to a maximum spacing of 500 feet by 500 feet.

3.2 Perimeter boundary conditions

Existing natural boundary conditions were used to define the extents of the Panna Maria modeldomain. The top hydrogeologic unit of the Catahoula Formation is the B1-Sand. This unit outcrops

147

to the north of the site and was used as reference points for the northern extent of the active modeldomain. As the development of the model progressed, the Rincon Lake was originally includedin the model. The Rincon Lake is the remnants of the westernmost mine pit converted to a lake.During construction of Rincon Lake, over one hundred feet of low hydraulic conductivity clay wasused to line the old mine pit. This low hydraulic conductivity liner material acted in effect like ano-flow boundary and created unnecessary numerical instability in the model. The abandoned minepits to the east of the Rincon Lake were also backfilled with low permeability clay. The backfillmaterial serves to isolate the sediments south of the pits from the hydrogeologic regime north ofthe pits.

The model cells in Rincon Lake and backfilled mine pits area were converted to no-flow cellsand the corresponding northern extents of the active model domain were reduced to the southernedge of Rincon Lake and abandoned mine pits. This elimination of the thin, low conductivity cellsin this area increased numerical stability of the model.

The San Antonio River acts as a natural western and southern hydrologic boundary to the modeldomain. The SanAntonio River was input into the model as river boundary cells and the river surfaceelevation was digitized and utilized as the river stage for each river model cell. River boundaryconditions were placed in cells corresponding to the defined sand units including the B1-Sand.

The eastern model boundary condition consists of a series of no-flow cells located at a naturaltopographic ridgeline. The alignment of the topographic ridgeline is to the southeast, the samegeneral direction of the corresponding subsurface groundwater flow, making the ridge a reasonableno-flow boundary for the relatively shallow groundwater flow system. Precipitation that occurseast of this topographic ridgeline flows east and recharges potential alluvial outside of the area ofinterest for this modeling effort. Therefore, no-flow boundary cells were used along this topographicridgeline/groundwater divide.

3.3 Internal boundary conditions

Internal boundary conditions were utilized to represent naturally occurring surficial sloughsand associated ponds, the central core of the tailings impoundment dikes and the drain systemincorporated with the tailings impoundment.

3.3.1 Sloughs and pondsThe naturally occurring sloughs in the vicinity of the tailings impoundment include the Lake Sloughand associated tributary in the western and southern portions of the model domain and the MankaSlough in the mid-eastern portion of the model domain, which ultimately is a tributary to the LakeSlough. The sloughs originate in the vicinity of the Alluvial and A sands.

Miocene-aged streams that drained the Catahoula deposits carried a stream load of silty sandsand fine to medium grained sands. The streams sediments are, in general, approximately 100

148

feet wide and between five and twenty feet thick. In the upland, near the Facility, the sedimentsare hydraulically connected to either the Alluvial Sand or the A-Sand. The hydraulic connectiontypically occurs near the surface. As the sloughs extend southward into the floodplain the sedimentshave been buried by the deposition of flood deposits of the San Antonio River to a depth ofapproximately forty feet. The floodplain sediments consist of primarily clay and silty clay that aregenerally dry.

The sloughs are important for two reasons. The surficial expression of the sloughs providessurface drainage from the uplands to the San Antonio River. The subsurface slough sedimentsprovide a pathway for groundwater to the San Antonio River.

The MODFLOW-SURFACT drain package was used to simulate the surficial sloughs based onthe concept that the sloughs do not significantly recharge the alluvium and primarily act as a conduitfor periodic surface water flow to the San Antonio River. The slough sediments are represented byhigh conductivity lenses in the low conductivity floodplain deposits.

Ponds and lakes on and near the Facility were simulated using constant head boundary conditions.The constant head at each boundary condition was set at the average head observed in the existingwater body.

3.3.2 Central core of tailings impoundment embankmentConstruction of the tailings impoundment dikes entailed a zoned earthen embankment consisting ofa central core flanked by shell zones with a blanket drain system on the downstream side. The centralcore of the impoundment dikes were designed to be relatively impervious. The average permeabilityof the select mine overburden material used to construct the central core was 1 × 10−6 cm/sec. Thewidth of the central core was constructed to be a minimum of twenty feet wide at its base where itwas keyed into the underlying Catahoula Clay.

Horizontal flow barrier boundary conditions were employed to simulate the central core on thetailings impoundment dikes. A preliminary width of twenty feet and a corresponding permeabi-lity of 1 × 10−6 cm/sec were used to define the total conductance of the horizontal flow barriermaterial. This preliminary conductance was modified during the flow calibration process, ensuringadequate model seepage values were maintained while concurrently matching the observed tailingsimpoundment dewatering curve.

3.3.3 Tailings impoundment drainsThe tailings impoundment contains a drain system on the downgradient side of the zoned earthenembankment. The drain system was designed to remove any runoff from precipitation occurringon the dikes as well as any seepage that infiltrated through the dikes. MODFLOW-SURFACTboundary condition drain cells were used in the model to simulate the drains.

149

3.3.4 Tailings impoundment headsThe site is unique in that there are three groundwater monitoring wells installed within the tailings.These wells provide both groundwater elevation and analytical data not available for most tailingssites. In the steady-state model constant head cells were used to represent the fluid elevation inthe tailings impoundment. These constant head cells were implemented to ensure correct startinghead elevations for the fluid in the tailings impoundment. The tailings impoundment’s steady stateconstant heads were set at an elevation of 360.12 feet amsl, based on the fluid elevation in welldata measured in the summer of 1992. The tailings impoundment constant heads were only usedin the steady-state model and subsequently removed for the transient calibrations.

The head in the tailings impoundment in the transient flow model were initially set at the 1992observed levels. Model parameters representing the central clay core of the tailings impoundmentdikes and the underlying clay were adjusted until the drainage curve was approximated.

3.3.5 Tailings uranium and sulfate concentrationsThe tailings source concentration of uranium and sulfate was simulated using constant concentrationboundary conditions. The tailings impoundment was subdivided into twelve zones. The zones wereestablished to account for the special variability in tailings composition.

4 MODEL CALIBRATION

The Facility was modeled implementing three separate, but inter-related models. A steady-statemodel was developed to simulate the summer of 1992, when much of the groundwater elevationsand analytical data was collected. This model provided the initial conditions for subsequent transientmodels. A transient calibration model was used to simulate the data collected between 1992 and2008, when the temporary piezometers were installed. The transient model included variationsin precipitation, well pumping, and analytical results. When parameters were changed during thecalibration of either the steady-state or the transient models, those parameters were updated in allmodels. The third model was used to predict the long term impacts of the tailings impoundmenton the surrounding groundwater and the San Antonio River.

The parameter estimation program PEST was used during both the steady-state and the transientcalibration. PEST was allowed to adjust key parameters within an established range of variability toreduce the objective function of the model and thus improve the model calibration. The parametervariability was established using available data, literature values and professional experience.

4.1 Flow model calibration

The steady-state flow model calibration statistics are summarized in the table below. Prior to modelcalibration, a target standard deviation (SD)/Range value of less than 0.05 (5 percent) indicated agood calibration, greater than 0.05 but less than 0.10 (5 to 10 percent) was considered a satisfactorycalibration and greater that 0.10 (10 percent) was considered a poor calibration. The calculatedSD/Range value of 0.07 indicates a satisfactory steady-state calibration.

The results of the steady-state calibration were used as the initial conditions for the afore-mentioned transient calibration. The transient flow model calibration was somewhat better thatthe steady-state calibration. Using the previously defined calibration criteria, the transient modelcalculated SD/Range value of 0.036 represented a good calibration.

4.2 Transport model calibration

After the steady-state and transient flow models were calibrated, both the sulfate and uranium trans-port models were calibrated. The initial 1992 uranium and sulfate plume extents were estimatedusing the data available from the 1992 investigation. Transport parameters including porosity anddispersion were calibrated using sulfate since it was expected to act as a conservative constituent,i.e. no transport retardation was expected. Adjustments were made to the initial plume configu-ration, porosity values and dispersion values until sulfate transport calibration was achieved. Thefinal calibration SD/Range statistics were calculated to be 0.046 and the sulfate transport modelcalibration classification was considered good.

150

The initial uranium plume distribution was estimated using the 1992 investigation data. Uraniumcalibration consisted of making minor adjustments to the initial plume distribution and to thedistribution coefficient, Kd. The resulting calibrated Kd for the slough sediments and the A-Sandwas 0.09 ml/g. The observed breakthrough curves at an off-site well for uranium and sulfate werecoincident, indicating that no retardation was affecting the uranium transport. Kd for the AlluvialSand was set to zero. The resulting calculated calibration statistics (SD/Range = 0.059) indicatethat the uranium transport model is satisfactorily calibrated.

5 PREDICTIVE FLOW AND TRANSPORT MODEL

Predictive modeling entailed simulating the calibrated flow and sulfate transport model and thecalibrated flow and uranium transport model for a 1,000 year timeframe. The calibrated physi-cal parameters were directly incorporated into the flow and transport models for the 1,000 year

151

simulations. The average precipitation of record, 30.9 inches per year, was utilized as the long termaverage precipitation for the 1,000 year simulations.

The groundwater elevation in the tailings impoundment was allowed to naturally decline basedon the calibrated tailings impoundment dewatering curve, which is primarily a function of thelimited recharge through the overlying infiltration barrier cap, the conductance of the horizontalflow barrier representing the central core of the embankment and the vertical hydraulic conductivityof the underlying B1-Clay.

During the 1,000 year simulations the remaining model boundary conditions were unchanged,including the constant concentration zones representing source constituent concentrations withinthe tailings impoundment. A mass balance was performed on the approximate quantity of uraniumin the tailings impoundment to determine if the uranium source term was anticipated to be depletedwithin the 1,000 year predictive time frame. The calculated depletion was considered negligible.

6 FLOW AND TRANSPORT MODEL RESULTS

The results of the flow and transport predictive modeling indicate that neither sulfate nor uraniumwill reach the San Antonio River at concentrations exceeding the respective Maximum Contami-nant Levels (MCLs). The primary transport pathway is through the Alluvial Sand sediments andconnecting Lake Slough sediments.

Tailings seepage is expected to change over time. Seepage was estimated to be 24 gallons perminute from 1992 until 1996, when the tailings were capped. In 1997, the seepage estimatedwas reduced to 17 gallons per minute. Current tailings seepage estimates are about 8.3 gallonsper minute. Ultimately, tailings seepage estimates will approach a steady-state condition when itapproaches the recharge through the barrier cap, of approximately 4.6 gallons per minute.

Tailings seepage has a significant impact on the groundwater flow and site groundwater qualityat the Facility. The tailings fluid is the source of the constituents at and downgradient of the site.Increased seepage from the tailings impoundment results in higher constituent concentrations.Addi-tionally, the tailings seepage increases the groundwater gradients near the tailings impoundmentand the Facility. Higher gradients result in faster flow velocities and transport rates. Conversely, asthe tailings impoundment drains, the gradients and resulting constituent load decline.

Tailings fluids are discharged through the clay core of the impoundment into both the A-Sandand theAlluvial Sand. TheA-Sand is hydraulically connected to the sediments of the Manka Sloughand ultimately to the Lake Slough sediments. The Alluvial Sand is hydraulically connected to theLake Slough sediments. The Lake Slough, in the flood plain, is interpreted to be a paleochannelof the San Antonio River and discharges to the San Antonio River. The final transport pathwayof constituents of concern is through the unsaturated B1-Clay into the underlying B1-Sand. TheSan Antonio River is considered a groundwater divide within the model domain and is expected toaffect flow in the B1-Sand. It is assumed that the B1-Sand discharges to the San Antonio River. Thisassumption provides a conservative estimate of the constituent load to the river from all potentialsources.

Neither the A-Sand nor the Alluvial Sand is currently used for domestic water supply nor are theyexpected to be capable of sustaining enough groundwater to be used as a drinking water source.Due to their limited extent, the slough sediments likewise are not capable of supplying sufficientquantities of groundwater to be used as a drinking water source.

7 CONCLUSIONS

The modeling code MODFLOW-SURFACT was able to simulate the constituent transport throughthe variably saturated clays. The calibrated model adequately simulates the transport of the con-stituents of concern through the sediments present on and near the Facility. The modeling effort forthe Panna Maria ACL Application indicates that there is a significant area that will require insti-tutional controls to transfer the Facility to the Department of Energy, especially down the westernAlluvial Sand/Lake Slough transport pathway. The proper application and calibration of the robustmodel, supported by sufficient site-specific geologic, hydrologic and water quality data, provides

153

a reliable and defensible basis for the establishment of the ACLs. The model will facilitate the ACLApplication acceptance and accelerate Facility closure.

It was discovered that the coincident breakthrough of uranium and sulfate at an off-site well waspossibly the result of well casing failure that allowed shallow groundwater already impacted byuranium and sulfate to reach the well screen. This observation negates our conservative assumptionthat retardation was not a factor for uranium transport along theAlluvial Sand/Lake Slough transportpathway. This result is significant as it will affect the timing of the uranium arrival at all locationswithin and down gradient of the Alluvial Sand. The extent of the modeled plume will be reducedas the result of increasing the distribution coefficient (Kd). Currently additional work is proposedto adjust this parameter in the model and estimates the resulting uranium extents. This will likelyreduce the area where institutional controls are required.

Point of Compliance (POC) alternate concentration limits have not yet been estimated for theACL application. This effort has been delayed until regulatory approval of the calibrated model hasbeen indicated. The ACL application is anticipated to be complete by mid-year, 2010.

REFERENCES

American Society for Testing and Materials (ASTM), 2008a. Standard Guide for Conducting a SensitivityAnalysis for a Ground-Water Flow Model Application. Designation: D 5611-94 (Reapproved 2008). ASTMInternational, West Conshohocken, Pennsylvania.

American Society for Testing and Materials (ASTM), 2008b. Standard Guide for Comparing Ground-WaterFlow Model Simulations to Site Specific Information. Designation: D 5490-93 (Reapproved 2008). ASTMInternational, West Conshohocken, Pennsylvania.

American Society forTesting and Materials (ASTM), 2008c. Standard Guide for Defining Boundary Conditionsin Ground-Water Flow Modeling. Designation: D 5609-94 (Reapproved 2008). ASTM International, WestConshohocken, Pennsylvania.

American Society for Testing and Materials (ASTM), 2008d. Standard Guide for Defining Initial Conditionsin Ground-Water Flow Modeling. Designation: D 5610-94 (Reapproved 2008). ASTM International, WestConshohocken, Pennsylvania.

American Society for Testing and Materials (ASTM), 2008e. Standard Guide for Conceptualization and Char-acterization of Ground-Water Systems. Designation: D 5979-96 (Reapproved 2008). ASTM International,West Conshohocken, Pennsylvania.

American Society for Testing and Materials (ASTM), 2008f. Standard Guide for Calibrating a Ground-Water Flow Model Application. Designation: D 5981-96 (Reapproved 2008). ASTM International, WestConshohocken, Pennsylvania.

American Society for Testing and Materials (ASTM), 2008g. Standard Guide for Developing Conceptual SiteModels for Contaminated Sites. Designation: E 1689-95 (Reapproved 2008). ASTM International, WestConshohocken, Pennsylvania.

American Society for Testing and Materials (ASTM), 2006a. Standard Guide for Subsurface Flow and Trans-port Modeling. Designation: D 5880-95 (Reapproved 2006). ASTM International, West Conshohocken,Pennsylvania.

American Society for Testing and Materials (ASTM), 2006b. Standard Guide for Documenting a Ground-Water Flow Model Application. Designation: D 5718-95 (Reapproved 2006). ASTM International, WestConshohocken, Pennsylvania.

American Society for Testing and Materials (ASTM), 2004. Standard Guide for Application of a Ground-WaterFlow Model to a Site-Specific Problem Designation: D 5447-04. ASTM International, West Conshohocken,Pennsylvania.

Anderson, M.P. and W.W. Woessner, 1992. Applied Groundwater Modeling: Simulation of Flow and AdvectiveTransport. New York, New York, Academic Press Inc.

Doherty, J., 2004. Model-Independent Parameter Estimation Users Manual: 5th Edition. Watermark NumericalComputing.

Gard Water Consultants, Inc. and SRK Engineering, Inc., 2009. Supplemental Data Collection and ModelUpdate Report. Prepared for Rio Grande Resources, Hobson, Texas.

Gard Water Consultants, Inc. and SRK Engineering, Inc., 2009. Predictive & Historical Modeling of Ground-water Flow and Transport in the Vicinity of the Decommissioned Panna Maria Uranium Operations Facility,Texas. Prepared for Rio Grande Resources, Hobson, Texas.

International Engineering Company, Inc. (IECO), 1977. Design Report, Panna Maria Uranium Mine WasteDisposal Facilities, Volume I and 2, Karnes City, Texas. Prepared for Chevron Resources Company.

154

Langmuir, D.L., 1997. Aqueous Environmental Geochemistry. Prentice Hall, New Jersey.Mace, R.E. Davidson, C.E. Angle, and W.F. Mullican, 2006. Aquifers of the Gulf Coast of Texas, Texas Water

Development Board, Report 365.MODFLOW-SURFACT, 1996. MODFLOW-SURFACT Software (Version 3.0) Overview: Installation,

Registration, and Running Procedures. Hydrogeologic, Inc.McDonald, M.G. and A.W. Harbaugh, 1988. A Modular Three-dimensional Finite-difference Groundwater

Flow Model. U.S. Geological Survey Professional Paper.Rumbaugh, J., and D. Rumbaugh, 2007. Groundwater Vistas, Version 5, developed by Environmental

Simulations, Inc. Reinhold, Pennsylvania.Shepherd Miller, 2002. As-Built Report for the Closure of the Panna Maria Tailings Pond. Prepared for Rio

Grande Resources Corporation, Hobson, Texas.Simons, L.H. and Taggart, Jr., M.S., 1953. Clay Mineral Content of Gulf Coast Outcrop Samples. Humble Oil

& Refining Co., Houston Texas.SRK Consulting, Inc., 2000. Responses to Texas Department of Health Comments on Application for Alterna-

tive Concentration Limits. Rio Grande Resources, Panna Maria Tailings Impoundment (TDH License No.L02042). Prepared for Rio Grande Resources, Hobson, Texas.

Strachan, C.L., and Raabe, K.L., 2009. “Reclamation of the Panna Maria Uranium Mill Site and TailingsImpoundment: A 2008 update.” Tailings and Mine Waste ’08. Taylor and Francis Group, London. 381–358.

U. S. Geological Survey (USGS) and U. S. Department of the Interior, 2006. Hydrologic and Water-QualityData, Honey Creek State Natural Area, Comal County Texas, August 2001-September 2003. Data Series200. In cooperation with the U.S. Department of Agriculture, Natural Resources Conservation Service, andthe San Antonio Water System.

VTN Environmental Services, 1977. Panna Maria Uranium Project Environmental Baseline Report. Preparedfor Chevron Resources Company, Hobson, Texas.

Water, Waste & Land, Inc. (WWL), 1991a. Groundwater Compliance Plan, Panna Maria Project. Prepared forChevron Resources Company, Hobson, Texas.

Water, Waste & Land, Inc. (WWL), 1991b. Response to TDH Comments, Groundwater Compliance Plan.Prepared for Chevron Resources Company, Hobson, Texas.

Water, Waste & Land, Inc. (WWL), 1993. Hydrogeologic Investigation and Preferred GroundwaterRemediation Plan. Prepared for Panna Maria Uranium Operations, Hobson, Texas.

Water, Waste & Land, Inc. (WWL), 1994a. Modification of Ground Water Monitoring Plan, Panna MariaUranium Operations. Prepared for Panna Maria Uranium Operations, Hobson, Texas.

Water, Waste & Land, Inc. (WWL), 1994b. Summary of Approach Alternate Concentration Limit Application.Prepared for Panna Maria Uranium Operations, Hobson, Texas.

Water, Waste & Land, Inc. (WWL), 1994c. Summary of Field Activities, Panna Maria Uranium Operations.Prepared for Panna Maria Uranium Operations, Hobson, Texas.

Water, Waste & Land, Inc. (WWL), 1996. Predictive Modeling of Ground Water Flow and Contaminant Trans-port in the Vicinity of Panna Maria Uranium Operations. Prepared for Rio Grande Resources Corporation,Hobson, Texas.

Westec, 1997. Application for Alternate Concentration Limits for Uranium, Selenium and Gross Alpha, PannaMaria Uranium Operations, Hobson, Texas. Prepared for Chevron Resources Company.

Xu, M., and Y. Eckstein. 1995. “Use of Weighted Least Squares Method in Evaluation of the RelationshipBetween Dispersion and Field Scale.” Groundwater, Vol. 33, No. 6: 905–908.

Zheng, C., and G.D. Bennett. 1995. Applied Contaminant Transport Modeling: Theory and Practice. JohnWiley & Sons, New York.

Zheng, C., and P. Wang, 1999. MT3DMS: A Modular Three-Dimensional Multispecies Transport Model forSimulation of Advection Dispersion and Chemical Reactions of Contaminants in Groundwater Systems;Documentation and User’s Guide. Washington D.C.

Zheng, C., 1990. A Modular Three-dimensional Transport Model for Simulation of Advection, Dispersion andChemical Reactions of Contaminants in Groundwater Systems. U.S. Environmental Protection Agency.

155


A priori and posterior probabilities in operational water balances fortailing storage facilities

Steven F. Truby, Victor Lishnevsky & James R. KunkelKnight Piésold and Co., Denver, Colorado, USA

ABSTRACT: An a priori probability distribution for a Tailings Management Facility (TMF)water balance variable (pond volume) is one in which you can see that it is true just lying on yourcouch. You don’t have to get up off your couch and go outside and examine the way things are inthe physical world. You don’t have to do any science. In a posteriori probability distribution for aTMF water balance variable, knowledge or justification is dependent on water balance outcomes orempirical evidence. This paper will examine the difference between modern water balance modelswhere the selected input parameters (precipitation and evaporation) are assigned a priori probabilityfunctions and a large number of water balance output realizations for a given variable (pond volume)are generated, versus examining all possible combinations of input parameters and operationalparameters (TMF evolution) and then fitting a posteriori probability function to the outcomes.

1 INTRODUCTION

The procedure used in the operational water balance to perform TMF pond design is the robustwater balance model previously described by Kunkel (2001) and Kunkel and Lishnevsky (2002).The traditional approach to an operational water balance is to use the systematic climatological timeseries data as average monthly values and once-through operation. This approach uses only a 12-“season” model with each season represented by a month. An alternative approach for operationalwater balance modeling is to use the complete monthly time series for as many years of data as areavailable, or to use a stochastically generated monthly time series. Modern stochastic climate datagenerating techniques are easily applied to obtain data “in-fill” and/or extension of the historicalclimatological data. Typical stochastic data generation models include WGEN (Richardson andWright, 1984) and ClimGen (Stöckle and others, 1998; 1999). Similar stochastic models also areavailable within the EPIC computer program (Sharpley and Williams, 1990) which is extensivelyused by Knight Piésold and Co., as well as GoldSim™.

The robust operational water balance approach (Kunkel and Lishnevsky, 2002) uses the sys-tematic climatological record or a synthetically generated monthly time series. Figure 1 shows ahypothetical monthly climatological time series of 33 years and a mine life of 6 years for a Perúmine site, which indicates how the robust operational water balance operates. The first operationalwater balance model run is placed at the start of the mine life in the first year of the 33 yearsof precipitation data, while the second model run is placed in the second year of the data, andso on. Each model run is considered to be independent of the previous run. Therefore, each ofthe 33 runs produces 33 equally-likely, independent outcomes which can be analyzed and fit to aprobability distribution to calculate the probability of occurrence of given output variables from theruns. The reason for performing 33 equally likely and independent water balance runs is because itis unknown, a priori, when the mine project will come on line or what the climate conditions willbe when the project comes on line.

2 WATER BALANCE MODEL COMPONENTS AND MODELING GOALS

A simple operational water balance problem including a TMF, a processing plant, and an exter-nal makeup water source was modeled. The water balance input components included the

157

Figure 1. Hypothetical operational water balance procedure for 6-year mine life and 33 years of climatologicaldata.

following: (1) TMF pond (elevation, area, capacity data), (2) inflows to the TMF from internaland external areas, (3) seepage from the TMF, (4) tailings production over time (water, solids,tailing pore water storage), (5) precipitation, (6) evaporation (from tailings and non-tailings areas),and (7) process plant water requirements. Calculated water balance outputs included the following:(1) TMF pond volume, (2) TMF water supplied to the process, (3) TMF water discharged to treat-ment and/or the environment, and (4) required external makeup water for the process. A schematicof these components and their interactions is shown on Figure 2.

Using a 14-year set of monthly climatological data and the robust water balance techniquementioned above with the Knight Piésold Minder™ program, a water balance was generatedand probabilities calculated for operating tailing pond volumes and operational makeup waterrequirements for the process as the two example modeled variables. Monthly outcomes from theKnight Piésold Minder™ program were fit to an Extreme Value Type I (Gumbel) probabilitydistribution (Kite, 1977); although a two-parameter Log Normal or a Weibull distribution wouldgive essentially the same conclusions. This same approach was used deterministically in GoldSim™in order to have a common starting point for the a priori analysis. Monthly outcomes from theGoldSim™ program also were fit to an Extreme Value Type I probability distribution.

For comparison, the same 14-year monthly climatological time series data were fit to a two-parameter Log Normal probability distribution and the moments for these data then used inGoldSim™. The goal of the modeling was to assess differences, if any, between assigning aprobability function to the input climate data (a priori) to obtain water balance outcomes, ver-sus, calculation of the probabilities of the water balance outcomes (posteriori) from a series ofindependent, identically distributed random variables.

GoldSim™ provides two stochastic distribution models that would be suitable for modelingprecipitation data, namely the Log-Normal and Weibull distributions. The log normal distributionwas selected for the modeling as it has been shown by the Food and Agriculture Organization ofthe United Nations (FAO, 1999) to provide reasonable approximations of monthly precipitationvalues, and the input parameters required for the log-normal distribution are easy to generate from

158

Figure 2. Simple TMF water balance model components and their interactions.

historical data. Separate log-normal distribution functions were generated for each month of theyear, with the appropriate function being selected by the model during runtime, depending on themonth being modeled.

Functions were generated to relate precipitation to evaporation from water surfaces, wet tailings,dry tailings and natural ground. A separate function was generated for each component of evapora-tion, and for each month of the year, resulting in a total of 48 evaporation functions. The appropriatefunction was selected by the model during run time depending on the month of the year, and thecomponent of evaporation being modeled. The functions were based on monthly precipitation andevaporation data generated by Knight Piésold for the project. Similarly, a function was generatedrelating seepage into existing ground surfaces to precipitation, with a separate function being gen-erated for each month of the year. The outcomes presented below comparing the two programsutilized to model the water balance for the simple TMF system are considered to be preliminaryand did not show favorable comparisons. The reason(s) for the sometimes large differences in theoutcomes using the two programs is yet to be completely resolved.

3 WATER BALANCE OUTCOMES AND COMPARISONS

3.1 Deterministic model outcomes

Water balance outcomes for operational TMF pond volumes (in m3) and operational makeup waterrequirements (in m3/hr) for the process were modeled deterministically using both the KnightPiésold Minder™ and the GoldSim™ programs. Monthly results for average operational TMFpond volumes in cubic meters (m3) are shown graphically on Figure 3 for the deterministic cases.

Analysis of Figure 3 showing the average operational TMF pond volumes indicates that both theKnight Piésold Minder™ and the GoldSim™ programs generally predict the same maximum TMFpond storage which would be used for design of the TMF. However, the GoldSim™ program mod-eled significantly lower (by approximately 50 percent) minimum TMF storage volumes. The reasonfor this is not well understood but may be due to the evaporation component of the water balance,as the two models were independently programmed and even though the individual programmerscommunicated with each other, there were small differences in the codes which may have resultedin a systematic increase in evaporation during the southern hemisphere summer (rainy season forthe site utilized).

159

Figure 3. Deterministic TMF operational pond volumes.

Figure 4. Deterministic makeup water requirements.

Deterministic outcomes for makeup water requirements using the two programs are shown onFigure 4. The GoldSim™ program shows much larger makeup water requirements at the end ofthe dry season (October of each year). The reason for this is not well understood but may be dueto overprediction of the evaporation even though the monthly evaporation input functions were thesame for both programs.

3.2 Probabilistic model outcomes

A Gumbel probability distribution was fit to the deterministic outcomes for TMF operational pondvolumes and makeup water requirements based on the outcomes of the deterministic modeling

160

Figure 5. Posteriori 100-Year TMF operational pond volumes.

Figure 6. Posteriori 100-Year makeup water requirements.

using both the Knight Piésold Minder™ and the GoldSim™ programs. Results of this exerciseare shown respectively in Figures 5 and 6 for the operational pond volumes and makeup waterrequirements.

Analyses of these two figures indicates that the 100-Year return period using both programsgiving the monthly 1-percent chance exceedance volumes are generally the same and match wellfor the dry season months (April through October), but differ by over 50 percent during the wetseason (November through March). This difference indicates that for some reason, even though theinput data were the same, GoldSim™ program underpredicts the TMF pond volume and, as can beseen in Figure 6, over predicts the makeup water requirements.

161

Figure 7. A priori vs. Posteriori 100-Year TMF operational pond volume.

Figure 8. A priori vs. Posteriori 100-Year makeup water requirements.

3.3 Probabilistic model outcome comparisons

Figures 7 and 8 compare the a priori and posterior 100-Year (1-percent chance) monthly outcomesfor, respectively, TMF operational pond volumes and makeup water requirements using the KnightPiésold Minder™ and the GoldSim™ programs. In this case the probabilistic outcomes usingthe GoldSim™ program were the result of 2,000 realizations using a Log-normal probabilitydistribution fit to the monthly data for 14 years, and evaporation functions based on the sameevaporation data used in the Knight Piésold Minder™ program.

162

Figure 9. GoldSim™ A priori and Posteriori 100-Year makeup water requirements.

The presumption was that the a priori Log-normal precipitation inputs to the GoldSim™ programwould result in generally Log-normally distributed outcomes for both the TMF operational pondvolumes and the makeup water requirements. Comparison of the a priori (GoldSim™ program 1%stochastic) and posteriori (Knight Piésold Minder™ program 1% deterministic) outcomes shownon Figures 7 and 8 indicate that the results vary greatly during some portions of the year. Mostlikely the underestimation of TMF storage volume during the dry season is due to overestimationof evaporation by the GoldSim™ program. Makeup water requirements are over-estimated by thestochastic GoldSim™ program model during the rainy season, primarily as a result of precipitationbeing under-estimated over this period by the Log-normal rainfall generation routine in the program.

Given that there may be differences in the user programming in the Knight Piésold Minder™and the GoldSim™ programs, a comparison was done for the a priori and posteriori 100-yearmakeup water requirements from only the GoldSim™ program as shown on Figure 9. As would beexpected the stochastically generated values tend not to preserve the extremes which are typicallyexhibited by hydrologic data and actual operational variables. Whereas the GoldSim™ 1 percentdeterministic (posteriori) results do show some months in which makeup water requirements arezero (rainy season), the GoldSim™ 1-percent stochastic (a priori) results never show zero makeupwater. This could cause operational issues if water storage reservoirs are smaller than necessary,and/or result in water shortages during some months of the year.

4 CONCLUSIONS

A comparison of two water balance programs, Knight Piésold Minder™ and the GoldSim™, haveshown that even with consistent programming, the outcomes from these programs can differ by50 percent or more during some months of the year. Additionally, the comparison of probabilisticoutcomes for TMF operational pond volumes and makeup water requirements show substantial dif-ferences if the analyses are performed deterministically and then probabilities fit to these outcomes,versus stochastic inputs with presumed probabilistic outcomes.

The stochastic and deterministic models generated very similar results over many of the months ofthe year. Makeup water requirements are, however, overestimated by the stochastic model during

163

the rainy season, primarily as a result of precipitation being under-estimated over this period.This, along with non-realistic wet season makeup water requirements indicates that care must beutilized in interpreting the outcomes of even simple water balance models. More complex waterbalance models which have, for example, multiple heap leach pads, multiple liquid storage ponds,and complex piping and pumping may be best analyzed utilizing the robust water balance modelproposed by Kunkel (2001) and Kunkel and Lishnevsky (2002).

The outcomes presented herein comparing the two programs utilized to model the water balancefor the simple TMF system are considered to be preliminary and did not show favorable compar-isons. The reason(s) for the sometimes large differences in the outcomes using the two programsis yet to be completely resolved.

REFERENCES

Food and Agricultural Organization of the United Nations (FAO). 1999. A Statistical Manual for ForestryResearch. Regional Office for Asia and the Pacific. Bankok. 231p.

Kite, G.W. 1977. Frequency and Risk Analysis in Hydrology. Fort Collins: Water Resources Publications.224 pp. http://www.wrpllc.com/links.html

Kunkel, J.R. & Lishnevsky, V. 2002. A Robust Water-Balance Method for Sizing Heap Leach Solution Ponds.Proceedings of the SME Annual Meeting, Phoenix, Arizona. February 25–27, 2002 (only available oncompact disk). Preprint 02-049. 5p.

Kunkel, J.R. 2001. A Robust Water-Balance Method for Sizing Heap Leach Solution Ponds and Reservoirs.Proceedings of the XXV Convention of Peruvian Mining Engineers, Are-quipa, Peru, September 10–14,2001 (only available on compact disk). 11p.

Richardson, C.W. & Wright, D.A. 1984. WGEN: A Model for Generating Daily Weather Variable.U.S. Department of Agriculture, Agricultural Research Service. ARS-8, August. http://soilphysics.okstate.edu/software/cmls/WGEN.pdf

Sharpley, A.N. & Williams, J.R. 1990. EPIC-Erosion/Productivity Impact Calculator: 1. Model Documenta-tion. US Department of Agriculture Technical Bulletin No. 1768. 235 p., http://www.epa.gov/nrmrl/pubs/600r05149/600r05149epic.pdf

Stöckle, C.O., Steduto, P. &Allen, R.G. 1998. Estimating Daily and Daytime Mean VPD from Daily MaximumVPD. 5th Congress of the European Society of Agronomy, Nitra. The Slovak Republic.

Stöckle, C.O., Campbell, G.S. & Nelson R. 1999. ClimGen Manual. Biological Systems EngineeringDepartment. Washington State University. Pullman, WA. 28 p., http://bsyse.wsu.edu/climgen/

164


Single process arsenic and antimony removal using coagulationand microfiltration

Joseph R. Tamburini & H.C. LiangTetra Tech, Denver, CO, USA

Samuel J. BillinTetra Tech, Elko, NV, USA

ABSTRACT: Several arsenic removal strategies have been approved by the EPA as Best Avail-able Technologies; and with the promulgation of the Arsenic Rule in 2001, these strategies are wellunderstood and widely implemented. However, strategies for antimony removal are not as widelyimplemented. Bench scale testing at a remote water treatment site in Nome,Alaska showed that a sin-gle treatment process including pH adjustment and coagulation followed by low pressure filtrationcould be used to remove both contaminants at 90 percent removal efficiency. The treatment processhighlighted here is advantageous over other treatment methods such as high pressure reverse osmo-sis filters because it requires less power consumption, creates less waste, and is lower in capital cost.

1 BACKGROUND

Arsenic has been found to impair biological metabolic pathways as well as cause cancer (specificallylung and skin cancer) in humans in chronic doses (CDC 2000). Due to the prevalence of arsenic inmany ground water sources scattered throughout the United States, the Environmental ProtectionAgency (EPA) issued an arsenic standard on all drinking water sources, promulgated as the ArsenicRule in 2001. Several arsenic removal strategies have been approved by the EPA as Best AvailableTechnologies including: ion exchange, activated alumina, reverse osmosis, coagulation followed bymicrofiltration, modified coagulation/filtration, modified lime softening, electrodialysis reversal,and oxidation/filtration (EPA 2001). Although many of these technologies have been implementedfor some time, after the promulgation of the Arsenic Rule, these strategies became even betterunderstood and widely implemented. The discussion in the paper will focus on coagulation followedby microfiltration (C/MF) for treating both arsenic and antimony.

1.1 Arsenic chemistry

Under natural conditions, arsenic (As) is present in two different oxidation states: arsenous, orAs(III) and arsenic, or As(V). When present in water, arsenic forms the dissolved species presentedin Figure 1, dependant on the pH and oxidation reduction potential (ORP) of the solution. Asillustrated in Figure 1, under typical groundwater conditions, As(III) occurs as H3AsO0

3 and As(V)occurs either as H2AsO−

4 or H2AsO2−4 . This is significant when trying to remove arsenic using

C/MF because As(III) occurs in solution as a neutral species while As(V) occurs as a mono- ordi-anion, depending on the pH.

In order to remove dissolved arsenic from water using coagulation followed by microfiltration,a coagulant must be added to adsorb the arsenic and form a floc particle that can be removedby the microfiltration unit. Typically, an iron-based or aluminum-based chemical is added as thecoagulant. For the purposes of this paper, an iron-based coagulant will be used since that is thechemical available for Alaska Gold Company (AGC) at their mine site. Ferrous sulfate was usedin the bench scale tests for arsenic removal.

165

Figure 1. Arsenic Pourbaix Diagram (Liang, 2008 and Garrels, 1965).

When ferrous sulfate (FeSO4) is added to water, it forms ferric hydroxide that precipitates outof solution. At neutral and low pH, a floc with the simplified formula of Fe(OH)+2 can form fromferric hydroxide, Fe(OH)3, according to Equation 1.

There is a net positive charge on ferric hydroxide as long as the pH of the solution is less than8 (Chwirka et al., 2004), and the net positive charge increases as the pH decreases. The negativelycharged form of dissolved arsenic (H2AsO−

4 ) is attracted to the positively charged ferric hydroxidefloc particles. Once the dissolved arsenic is adsorbed by the ferric hydroxide floc, the arsenic can beremoved from solution by removing the floc particles with a microfilter. However, if the oxidationstate and the pH of the solution is below the red line in Figure 1, the dissolved arsenic is in the formof As(III), which has a neutral charge, meaning that it has little attraction to the ferric coagulant.This shows that it is essential to have the correct pH and oxidation state balance for optimal arsenicremoval.

1.2 Antimony chemistry and removal

While the coagulation chemistry for arsenic is well understood and widely implemented, the coagu-lation chemistry of antimony is not as well understood. Antimony resides in the same group withinthe periodic table as arsenic, indicating that its chemistry should be similar to that of arsenic. Infact, antimony (Sb) is generally present in two oxidation states: antimonous, Sb(III) and antimonic,Sb(V). Pourbiax et al.(Pourbiax 1966) developed a Pourbaix diagram for antimony similar to thatshown in Figure 1 for arsenic. This diagram suggests that a negatively charged dissolved gaseousspecies (SbO−

3 ) is present under highly oxidized conditions, similar to that of arsenic. Therefore,similar coagulation chemistry should occur for the removal of antimony as for the removal of arsenic.

2 METHODOLOGY

A pilot study was conducted in October 2006 at the Alaska Gold Company’s (AGC) Rock Creekmine site in Nome, Alaska. Water from three different dewatering wells was collected and the rawwater tested for water quality parameters including arsenic and antimony and used throughout thepilot study. The water quality from each of the three wells is summarized in Table 1 in comparisonto the maximum contaminant limits (MCL). Groundwater pumped from AGC Rock Creek mine pit

166

Table 1. Raw water quality summary.

Parameter Well 1 Well 2 Well 3 MCL

Alkalinity, mg/L 200 262 172 N/ApH, s.u. 7.72 7.59 7.90 N/ATotal Dissolved Solids, mg/L 324 410 232 N/AAntimony, mg/L 0.000 0.171 0.006 0.006Arsenic, mg/L 0.079 0.205 0.155 0.010

area must be treated to meet, or exceed, the more stringent of either the drinking water requirements(or MCLs) or the aquatic life standards. As illustrated in Table 1 the three wells are between 8 and20 times above the MCL limit with respect to arsenic. Also, Well 2 and Well 3 are at or above theMCL limit with respect to antimony. Both contaminants need to be removed from the raw water ina water treatment plant.

The goal of the pilot study was to determine the optimum pH and coagulant dose for both arsenicand antimony removal via low pressure filtration. The mine utilizes ferrous sulfate (FeSO4) forother coagulation processes in the processing of ore; therefore, the water treatment plant is requiredto utilize this same chemical. The study included adding ferrous sulfate coagulant at various dosesranging from 14 to 160 mg/L as FeSO4 (5.1 to 58.8 mg/L as Fe2+). The pH was also adjusted tovarious pH values ranging from approximately 4.0 to neutral pH using sulfuric acid to decreasethe pH. Sodium hypochlorite (NaOCl) was also added to the raw water to oxidize arsenic (III) toarsenic (V).

Each set of tests were conducted with four 1000 mL jars all with the same oxidant and coagulantdose added to the jars and mixed thoroughly. A 4-jar gang mixer was used to mix all samplessimultaneously. Next, the pH was adjusted using sulfuric acid at varying pH values in each of thefour jars starting at a pH of approximately 4.0 and increasing one pH unit in each jar to 5.0 and 6.0.The fourth jar was kept at a neutral pH, which varied depending on the coagulant concentrationadded. The jars were then mixed again at which time floc formation was evident. The samplesfrom each jar were then filtered through a 1.0 micron filter using a vacuum filter apparatus. Theresulting filtrate for each sample was express-shipped to a laboratory to be tested for arsenic andantimony under Method E200.8.

One set of tests was performed for three different coagulant doses with each of the three wellwater sources. A summary of test conditions can be found in Table 2.

3 RESULTS

Water quality reports for each of the samples were received in October 2006. The concentration ofantimony in the raw water of Well 1 was non-detect (less than 0.001 mg/L), so all results from Well1 showed non-detect levels of antimony. The concentration of antimony remaining in the filtrate ofthe bench scale tests for Well 2 ranged from 0.002 mg/L to 0.135 mg/L as illustrated in Figure 2.Most of the Well 3 antimony concentration values were below the detection limit; however, fourvalues were above that detection limit as seen in Figure 3.

Although the raw water from each of the three wells showed elevated levels of arsenic as presentedin Table 1; interestingly, all filtrate samples from the C/MF bench scale testing showed non-detectlevels of arsenic. Regardless of coagulant or oxidant dose, and regardless of the water source, allsamples were below the arsenic detection limit of 0.001 mg/L. In order to confirm that this wasnot testing error, six of the samples were analyzed for a second time using Method A3114 B, andagain showed non-detect levels of arsenic.

4 DISCUSSION AND CONCLUSIONS

Results from this bench scale test indicate that both arsenic and antimony removal is possible usingcoagulation followed by microfiltration. Both contaminants can be effectively removed in one passthrough the system assuming the pH and the coagulant dose are adequate. Arsenic removal wasacceptable at all pH values and coagulant doses tested. Antimony removal varied significantly based

167

Table 2. Sample testing conditions.

Water pH FeSO4 Dose as NaOClSample Source (s.u.) Dose (mg/L) Fe2+ (mg/L) Dose (mg/L)

102 Well 1 4.0 14 5.1 20103 Well 1 4.9 14 5.1 20104 Well 1 6.1 14 5.1 20105 Well 1 7.9 14 5.1 20106 Well 1 3.9 30 11.0 20107 Well 1 5.1 30 11.0 20108 Well 1 6.1 30 11.0 20109 Well 1 7.9 30 11.0 20110 Well 1 4.0 60 22.1 20111 Well 1 4.9 60 22.1 20112 Well 1 6.4 60 22.1 20113 Well 1 7.6 60 22.1 20114 Well 2 3.9 80 29.4 20115 Well 2 5.1 80 29.4 20116 Well 2 6.1 80 29.4 20117 Well 2 7.7 80 29.4 20118 Well 2 4.1 120 44.1 20119 Well 2 5.1 120 44.1 20120 Well 2 6.1 120 44.1 20121 Well 2 7.4 120 44.1 20122 Well 2 4.0 160 58.8 20123 Well 2 5.1 160 58.8 20124 Well 2 6.1 160 58.8 20125 Well 2 6.8 160 58.8 20127 Well 3 3.9 80 29.4 20128 Well 3 5.0 80 29.4 20129 Well 3 6.1 80 29.4 20130 Well 3 7.2 80 29.4 20131 Well 3 3.8 120 29.4 20132 Well 3 5.0 120 29.4 20133 Well 3 6.0 120 29.4 20134 Well 3 6.9 120 29.4 20136 Well 3 3.9 160 29.4 20137 Well 3 5.0 160 29.4 20138 Well 3 6.0 160 29.4 20139 Well 3 6.5 160 29.4 20140 Well 2 4.1 120 44.1 10141 Well 2 5.0 120 44.1 10142 Well 2 6.0 120 44.1 10143 Well 2 6.7 120 44.1 10144 Well 2 4.1 120 44.1 0145 Well 2 5.1 120 44.1 0146 Well 2 6.1 120 44.1 0147 Well 2 6.5 120 44.1 0

on pH and coagulant concentration; therefore, the conditions leading to cost effective antimonyremoval will dictate at the full scale treatment system. Antimony removal efficiency for the Well 2bench scale tests is presented in Figure 4.

The conditions required for adequate antimony removal are summarized in Table 3.Figure 4 shows antimony removal at the various pH values, coagulant doses and oxidant doses.

This chart shows that the most cost effective antimony removal would likely occur at a pH ofapproximately 5.0, ferrous sulfate dose of 120 mg/L and a sodium hypochlorite dose of 10 mg/L.Under these conditions antimony removal was seen to be above 80 percent, which is more thansufficient to meet the required MCL outlined in Table 1. Increasing the dose and/or lowering the pH

168

Figure 2. Filtrate Antimony Concentration for Well 2 Bench Scale Tests.

Figure 3. Filtrate Antimony Concentration for Well 3 Bench Scale Tests.

would increase the removal efficiency as high as possibly 92 percent removal, as seen in Figure 4,but the incremental increase is small compared with the cost of additional chemical.

The relatively low pH needed for effective antimony removal compared to the higher pH valuesunder which arsenic can be removed using ferrous or ferric salts can be rationalized by the overallcharges of the antimony and arsenic species involved and their differing affinity for the ferrichydroxide floc. For example, while the predominant arsenic species removed is the di-anionicspecies HAsO2−

4 near neutral pH and high ORP values, the antimony species removed is the mono-anionic SbO−

3 . Based on coulombic interactions, it can be speculated that near neutral pH, wherethere is a mixture between cationic ferric hydroxide Fe(OH)+2 and neutral Fe(OH)3 floc, adsorptionand subsequent removal of the higher charged HAsO2−

4 should be more efficient than that of thesingly-charged SbO−

3 . It can further be argued that as the pH decreases and there is a higher

169

Figure 4. Antimony Removal Efficiency.

Table 3. Optimum operating conditions.

Parameter Value

Coagulant (Ferrous Sulfate), mg/L 120Oxidant (Sodium Hypochlorite), mg/L 10pH, s.u. 5.0

percentage of positively-charged Fe(OH)+2 , the binding and removal of antimony becomes betterdue to more affinity between SbO−

3 and the ferric hydroxide floc.

4.1 Full scale design considerations

The full scale treatment system has been designed and is currently under construction. The designincludes pH adjustment with sulfuric acid and ferrous sulfate coagulant addition. Due to the highdose of coagulant required, a lamella-type plate settler is installed after coagulation, ahead ofmicrofiltration, to protect the membrane modules from excessive calcium sulfate buildup, whichcan result from coagulant addition. The pH is then adjusted back to neutral with lime slurry andthen filtered through a low pressure microfiltration membrane.

The results of this pilot testing show that antimony can be removed in the same process along witharsenic using low pressure microfilters. This is advantageous over high pressure reverse osmosisfilters because it requires less capital cost, much lower operating pressures, and produces only2 to 5 percent waste compared to 20 to 25 percent waste from a high pressure membrane system.

REFERENCES

U.S. Environmental Protection Agency (2001), Arsenic Rule 2001, 40 CFR 141, Washington, D.C.Center for Disease Control (2000), Case Studies in Environmental Medicine: Arsenic Toxicity, Atlanta, GA.Chwirka Joseph D. et al. (2004), Arsenic Removal from Drinking Water Using the Coagulation/Microfiltration

Process, Journal AWWA, 96 (3), 106–114.Liang, H. C., et al. (2008), The Simultaneous Removal of Arsenic and Manganese for water in Northern

Nevada, Nevada Water and Environment Association 2008 Conference.Garrels, Robert M.; Christ, Charles L. (1965) Solutions, Minerals, and Equilibria; Freeman, Cooper &

Company.Pourbaix, Marcel et al. (1966), Atlas of Electrochemical Equilibria in Aqueous Solutions, Translated from

French by James A. Franklin, Pergamon Press, Oxford, New York.

170


Mitigating impacts from acid-producing rock in Tennessee roadconstruction projects

J.J. GusekGolder Associates, Lakewood, Colorado, USA

V. Bateman, J. Ozment, L. Oliver & D. KathmanTennessee Department of Transportation, Nashville & Knoxville, TN, USA

J. Waples, T. Rutkowski, H. Moore, W. Bowden & A. ReitherGolder Associates, Lakewood, Colorado, and Atlanta, Georgia, USA

ABSTRACT: There is a potential for runoff to become polluted with sulfuric acid and metals(mostly iron) when the pyrite/sulfide rock weathers in road projects where Chattanooga shale andother pyrite-bearing or sulfide-bearing rock formations are exposed. As a part of surface waterpollution management, the Tennessee Department of Transportation updated its standard operatingprocedure to create a new guidance document. A team of geologists and GIS experts developeda database to quickly identify projects that need to follow the new guideline to avoid impacts byhighlighting zones of geologic formations known to contain pyrite or acidic pH-neutralizing rockssuch as carbonates.

Despite the implementation of best management practices, some residual acidic/metal runoffmay occur. For these situations, BMPs from mining industry experience are applied, and may havereverse application in mine waste situations. TDOT’s new guidelines are the most comprehensiveconstruction-related acidic rock drainage BMPs of any state DOT.

1 INTRODUCTION

The Tennessee Department of Transportation (TDOT) has in recent years been involved in thedetection, testing, and mitigation of rock material containing minerals that, under certain conditions,are capable of producing acidic runoff. In late 2006, a focused effort began to replace an earlierstandard operating procedure (SOP) regarding this issue. The new guideline was based on existingliterature and published practices by others faced with the challenges of encountering acid producingrock (APR) which can lead to acid rock drainage (ARD). The new document, entitled Guidelinefor Acid Producing Rock Investigation, Testing, Monitoring, and Mitigation (TDOT 2007), andcalled the “APR Guideline,” was designed to provide consistent guiding principles rather thanstrict analytical/procedural protocols, to be applied to TDOT projects for investigation, prevention,and mitigation of potential ARD. Thus, it considers professional judgment as acceptable inputin decision-making. Notably, it was produced in cooperation with the Tennessee Department ofEnvironment and Conservation (TDEC).

While the primary focus of the new guideline was ARD prevention, it also included a secondaryfocus on ARD treatment not addressed by the original SOP (TDOT 1990).

The first phase, project screening, will be conducted using geographic information system (GIS)-based data, TDOT personnel professional experience, and other available geological literature andmaps. The goal of this phase is to determine if a project, or a project’s components, is located inmedium- or high-risk APR potential zones. Figure 2 shows the various risk zones as developed bythe APR guideline team. These GIS data are based on a bedrock geology map (1:250,000 scale)produced by the Tennessee Division of Geology (Hardeman 1966) and were deemed sufficientlycomprehensive for use as a general guide for site geology and potential risk. Project components in

171

Figure 1. Phases of project investigations and Activities (TDOT 2007).

Figure 2. Acid Producing Rock Risk Map for Tennessee (TDOT 2007).

low-risk APR zones could likely be exempted from additional phases such as sampling and testing.However, all project sites would require an initial site visit and/or knowledge from previous visitsto determine that potential APR materials are not present. While the primary purpose of the sitevisit is to verify the accuracy of the GIS mapping data and expected geology of the site, the visitmay be combined with other tasks related to geotechnical data needs.

Projects with components located in medium- or high-risk APR zones are to follow samplingand testing guidelines during the life of the project and monitoring at the conclusion of the project.Data generated would be examined using guidance provided in the “Triggers and Thresholds”sections of the APR Guideline to identify if further sampling or mitigation measures are warranted.In addition, if potential APR materials are identified at any point during the project, the APRGuideline provides direction for appropriate APR mitigation design approaches. Mitigation isdivided into two methodologies: prevention and treatment.

172

Figure 3. Relative costs within the realm of APR mitigation strategies.

The realm of prevention and treatment technologies has two logical endpoints as shown graphi-cally in Figure 3. At one end, an APR situation might be completely mitigated by implementing a“walk-away” prevention design remedy that is nearly permanent, requiring little or no maintenancewith just cursory post-construction monitoring. The upfront costs of implementing this approachmay be much more than Tennessee taxpayers are willing to spend for a new transportation project.At the other end, it may not be practical to implement APR prevention measures in which case acommitment to perpetual treatment of acidic drainage and monitoring will be required in the eventthat acidic drainage actually forms. The long-term costs and problems of this approach may beequally unacceptable. Some projects may have components encompassing both endpoints, and thevast number of combinations in between. The proportioning of prevention and treatment risk is tobe resolved by the professional judgment of qualified engineers and/or geologists based on project-and site-specific circumstances.

Guideline-based recommendations may vary within a given project depending on the currentproject phase and with changes in geology, site conditions, and disturbance area. Pre-, mid-,and post-construction activities may require different levels of sampling and testing. Also, dueto the linear nature of highway construction projects, guideline applicability may vary with mile-post/stationing as a function of the geology combined with the depth of construction and other siteconditions. Lastly, the type of project might influence APR assessment and response procedures.These include:

– Building a new-alignment road in undisturbed terrain– Widening or modifying an existing road segment, and– Implementing ARD mitigation at a previously-constructed project.

The principles and general direction included in the new APR Guideline were derived fromexisting literature, previous TDOT professional experience, past practices, and experience reportedby others, such as the US EPA, other states, and the Federal Highway Administration. GermaneTDOT experience was obtained from recent TDOT highway projects involving acid producing rockmaterial in Blount, Carter, Sevier, and Unicoi counties.

Based on literature searches as well as direct contact with the transportation departments in otherstates, it appears that no other state transportation agency has a guideline document for dealingwith APR at this time, though several other states are aware of these issues and are researchingthem as well. The US EPA, many state agencies, and mining companies are confronted with APRsituations related to existing and abandoned mines; therefore, it was appropriate to consider someof this experience in identifying and characterizing APR and in developing mitigation guidelinesfor potential APR from TDOT projects.

173

While the new APR Guideline attempts to provide up-to-date and state-of–the-art practices forAPR and roadway construction, new tests, standards, or mitigation technologies may be developedin the future. Factors affecting APR generation include mineralogy, weathering rates, climate,material size and surface area, mineral occlusion or exposure, exposure of the material to air andwater, hydrologic regime, and material placement method and location (EPA 1994; Nordstrom &Alpers 1998); these factors and their complex interactions are being continuously studied andresearched in a variety of settings. Therefore, it is anticipated that the new APR Guideline will bereviewed and updated periodically to account for new developments, including findings developedin-house by TDOT based on site-specific observations at Tennessee road projects.

Those observations may include the assessments of mitigation strategies thatTDOT implementedat the outset of dealing with APR issues over a decade ago. There is no better gauge of a mitigationdesign’s effectiveness than the test of time. The protocols developed in the new APR Guidelineshould facilitate this ongoing process into the future.

2 APR RISK MAP GENERATION

The GIS/APR map (GIS dataset) was developed by researching the geology of the State of Tennesseeto identify known geologic units that have the potential to be sources of APR. In addition, theAPR Guideline team also identified geologic units that contain neutralizing materials for APR. Asnoted earlier, the GIS/APR Map was based on the Geologic Map of Tennessee (Hardeman 1966).Geologic units shown on the map legend were researched as well as a general internet search forAPR and pyrite-containing formations in the State. While pyrite is the most common component inAPR, other sulfide-bearing minerals can also be present. In addition, the team researched availablehard-copy publications to complete the effort.

The research identified individual geologic formations and groups of formations that containknown APR sources, potential APR sources or sources of neutralizing materials. The team definedseven categories for the GIS/APR map (five with APR potential to varying degrees and two withAPR neutralizing potential) and color coded them as follows:

– Red–individual Formations which are known sources of APR– Light Red–groups and supergroups that include formations which are known sources of APR– Orange–formation that may contain potentially APR– Yellow–formations that are potential sources of APR– Navy Blue–Fort Payne and Chattanooga Shale (specific, historically problematic, high APR

potential rock formations)– Green–limestone (material with comparatively high neutralizing ability)– Light Green–dolomite (material with comparatively lower neutralizing ability).

The details supporting these categorizations would constitute a separate technical paper and arenot discussed here. An example of the individual formation data is provided in Table 1 below whichis an excerpt of the geological data that was inserted in the GIS metadata table (TDOT 2007).

The GIS/APR Map includes a number of layers that contain political or geographical informationfor orientation and to make the map more useful. These layers include the following:

– TDOT Regions– TDOT jurisdictional roads– County Names, boundaries, and county seats– Waterways– 303(d)/305(b) waters impaired by pH and/or metals– 7.5 Minute (1:24,000 scale) USGS quadrangle map names and boundaries– State Plane 1983 Coordinate System:

FT Zone 4100 Tennessee with a North American Datum (NAD) 1983 projection– Latitude and longitude

The GIS/APR Map is intended to be an evolving “living” tool which can be updated and refinedwith more detailed information which can be incorporated into the GIS database to supersede theexisting database. Several directions for additional effort or further refinement of the GIS database

174

Table 1. Typical geological formation data from the GIS metadata table.

Geo. Formation ID Geological formation(s) APR Category

Pcm Cross Mtn Formation Includes formations that may containacid producing rock

Pco Crab Orchard Mountains Group: Includes formations that containContains Whitwell Shale acid producing rock

Pcg Crab Orchard Mountains and Gizzard Includes formations that containGroups: Contains Whitwell Shale acid producing rock

p€o Ocoee Supergroup Includes formations that containacid producing rock

p€w Walden Creek Group: Includes formations that containContains Sandsuck Formation acid producing rock

p€ss Sandsuck Formation Formation that containsacid producing rock

p€rb Rich Butt Sandstonep€g Great Smoky Group Includes formations that contain

acid producing rockp€s Snowbird Group Includes formations that contain

acid producing rockp€m Mount Rogers Groupp€r Roan Gniess

were identified during the research that may be of particular value. For example, a significantamount of published geologic mapping exists that could be evaluated, digitized and incorporatedinto the GIS/APR database to provide more detail and precision. Several areas of 303(d)/305(b)impaired waters are covered by 1:24,000 scale geologic mapping which could provide additionaldetail in these critical locations. In addition, site-specific geologic mapping of APR data could beincorporated into the existing database. Another opportunity for refinement would be to incorporatethe Soil Survey Geographic (SSURGO) Database developed by the National Resource Conserva-tion Service (NRCS). This soils mapping is complete and available in digital format for much ofthe state. The SURRGO soils mapping provides soil properties based on shallow (60-inch deep)soils borings and laboratory testing which includes classification testing, basic soils mechanicsproperties, erosion characteristics, permeability, and soil pH.

3 PROJECT SCREENING AND SITE ASSESSMENT

The APR Guideline prescribed three preliminary phases to be conducted as a part of a potentialAPR evaluation. The first phase is project screening. Project screening should identify whether aproject or project components are located in areas of low-, medium-, or high-risk APR zones. Thesezones are based upon the geology of Tennessee and professional knowledge of Tennessee geologicformations with respect to APR. Tennessee formations have been classified as those with knownpotential (high risk zones), likely potential (medium risk zones), or minimal to rare potential (lowrisk zones). Locations of these zones are determined using a GIS database, published geologicalliterature and maps, as well as internal institutional or professional knowledge.

The second phase includes a dedicated site visit and/or assessment of observations noted inprevious site visits, referred to here as a visual and geographic assessment. The purpose of thevisual and geographic assessment is to confirm that site conditions match those predicted by thedatabase or other existing information, to assist with development of a sampling plan (SP), and toidentify areas or zones that should be targeted for future sampling and testing.

The third phase is the development of a SP. For projects with components containing medium-or high-risk APR zones, a SP, or multiple SPs if necessary, should be developed. The SP(s) shouldbe prepared at the conclusion of the project screening and visual and geographic assessment using

175

Figure 4. Summary of water, rock, and geophysical sampling programs (TDOT 2007).

project site-specific information and information collected as a part of the screening. The SP, orSPs, should incorporate recommendations presented in the APR Guideline.

4 SAMPLING AND TESTING

The “Sampling and Testing” section of the TDOT APR Guideline provides details for pre-construction and construction phase planning and sampling if the project is located in medium- orhigh-risk APR zones. Sampling of water and rock is required for those areas of the project that arelocated in medium- or high-risk APR zones, or in areas identified by the visual and geographicassessment. Sampling may vary throughout the project or in different areas, depending on theproject type, phase of the project, and results from earlier phases. Figure 4 provides a summary ofthe recommended water, rock, and geophysical sampling programs.

For completeness, theAPR Guidance document provides recommendations for specific samplingmethods and guidance for analytical testing methods. Results from the sampling and testing wouldbe assessed using the information contained in the triggers and thresholds discussion to determineif additional actions are required.

5 GUIDELINES FOR MATERIAL CHARACTERIZATION AND MITIGATIONTHRESHOLDS

The APR Guideline provides direction for examination and use of data collected during the sam-pling and testing phases of a project, as well as for the initial screening and monitoring phases.Numerical thresholds are provided for each of the testing methods, or a combination of the test-ing methods. If these thresholds are exceeded, additional effort, such as sampling or mitigationdesigns and appropriate material handling during construction, are therefore “triggered.” How-ever, these numerical thresholds must be considered with the site-specific conditions and past orknown behavior of the materials. Actual known field behavior of materials may be considered morereliable than laboratory testing performed in a sterile environment. To facilitate understanding and

176

communication regarding this complex issue, theAPR Guideline provided figures that summarizedrecommendations if thresholds are exceeded during the initial screening as well as flow charts fordecision-making based upon water and rock sampling results.

5.1 Visual and Geographic assessment thresholds

The Visual and Geographic Assessment can provide an excellent indication that potential-APR(P-APR) or APR materials are present or information about the field behavior of these materials.Site thresholds include the following that are associated with site geologic conditions:

– Waters of distinctive colors, such as iron/red, yellow, white, or black stained streambeds, oriron/red staining with large amounts of algae

– Staining of rocks or surface materials, particularly on hillsides, streambeds, road cuts, roadwaysor sidewalks, or other surfaces

– Low pH values (<5) or elevated conductivity values [>2,000 microsiemens per centimeter(µS/cm) depending on background] or

– Kill zones, or areas devoid of vegetation– Cementation crusts or areas of mineral precipitation from evaporating water– Geologic formations at the site, as outcrops or on geologic maps—of particular interest are

those known to be rich in sulfides (e.g. pyrite), have a history of APR impacts, or are carbonatematerials (e.g. limestone)

– Proximal P-APR sites, such as coal mines or, road cuts, and– Proximal road fills and any seeps emanating from the fills.

5.2 Rock thresholds

Laboratory test results drive the following thresholds and categories for rock materials. Based oninitial laboratory acid-base accounting (ABA) testing, including paste pH and pyritic sulfur values,materials will fall into one of four categories, as listed below.

– APR-Neutralizing Materials– Non-APR Materials– Potential APR Materials– APR Materials

Flow charts for identifying materials falling within these categories were developed and providethe foundation of the APR thresholds; one two sets of guidelines may be applicable depending onthe TDOT’s experience at a given site or geological situation.

The primary rock characterization guideline is based on existing institutional knowledge. TDOThas been actively and progressively working with P-APR and APR materials for many years andtheir practices to date have not resulted in significant ARD problems. Therefore, a primary set ofguidelines has been provided based on practices to date. These guidelines may be more appropriatefor sites and materials for which TDOT has previous experience where previous material handlingand placement procedures have not resulted in ARD.

A second set of guidelines provides thresholds that represent state-of-the-art practices withrespect to ARD evaluations applicable to the geologic setting of Tennessee but where institutionalknowledge may be lacking. These thresholds are necessarily conservative in order to account for thewide variety of factors that can influence ARD development. An appropriate future course of actionfor TDOT may be to collect and analyze historical and current data on handling and placement ofnon-APR, P-APR and APR materials to date in order to formally calibrate the thresholds proposed.

Selection of the particular set of guidelines should be made by a qualified engineer or geologistbased on site-specific and material-specific information based on the previous experience with asite or material.

5.2.1 APR characterization overviewThe characterization of a particular geologic horizon falls with a continuum ranging from APR-neutralizing materials to APR materials. The behavior of a geologic horizon is dependent on anumber factors, such as its mineralogy, weathering rates, material size and surface area, mineral

177

occlusion or exposure, exposure of the material to air and water. Therefore, characterization of amaterial relies upon several tests or aspects of the material to classify it as APR-neutralizing, APR,or somewhere in between. Paste pH, net neutralization potential (NNP) or neutralization potentialratio (NPR) values, and sulfur values are all considered in the APR Guideline to determine whethera given material needs to be fully or partially encapsulated or blended. In general, avoidance ofconstruction in APR horizons should be the preferred action.

5.2.2 Water thresholdsThere are several water chemistry indicators for the presence ofAPR.As described by Skousen et al.(1987), water affected by APR in the Appalachian region (Alabama, Indiana, Illinois, Kentucky,Maryland, Ohio, Pennsylvania, Tennessee, Virginia, and West Virginia) generally has pH valuesless than 5.0 or a combination of the following:

– total iron greater than 7 mg/L– total manganese greater than 4.0 mg/L– other dissolved metals greater than EPA MCLs– elevated acidity– elevated conductivity (>2,000 µS/cm, depending on background), and– elevated sulfate concentrations.

If these conditions are observed, thenAPR conditions may have developed. It is worth noting thatnot all of these water geochemistry indicators may be present to indicate that APR conditions aredeveloping; professional judgment and understanding of site geology should be used to determineif all or some of these conditions present indicate the development of ARD. Additional samplingshould be performed in anticipation of development of APR mitigation. If the above thresholds areobserved in surface water or groundwater, this should trigger periodic measurement of flow rates,which are necessary for design of mitigation systems.

In addition, trends in water chemistry through time are just as important as the stated valuesabove. Coupled with visual assessment clues (e.g. fresh iron staining), a professionally-judgedincrease in metals, sulfate, or acidity concentrations, or a coincidental decrease in alkalinity or pHvalues with time may be an indication that ARD is occurring. Increasing sulfate and decreasingalkalinity of the water, without increasing metals concentrations, may indicate that oxidation ofsulfides and subsequent consumption of neutralizing potential (NP) is occurring. If the NP becomesfully depleted then ARD conditions may occur. Therefore, if these trends are observed, increasedmonitoring should be performed and APR mitigation designed if ARD conditions have occurred.

6 MITIGATION MEASURES

Mitigation techniques are needed for two general situations: excavated material and cut slopes.Several mitigation techniques, referred to here as best management practices (BMPs) are providedfor both situations. For excavated materials, techniques range from blending to full encapsulation,with an intermediate of partial encapsulation. The techniques may be viewed as distinct methodsor as a continuum that may be adjusted to fit site specific conditions or materials.

6.1 Mitigation of excavated material

Techniques for the mitigation of APR excavated material have been proposed by the Federal High-way Authority (Byerly 1990) and TDOT (2005). TDOT has had significant experience with APRmitigation and has published research on updated mitigation methods (Moore 1992). The currentAPR Guideline expands or furthers these publications and experience. Techniques or BMPs ofseveral phases of road construction are provided below.

6.1.1 Design phase best management practices (BMPs)If pre-construction sampling and analysis indicates the presence of P-APR/APR, theAPR Guidelineindicates that:

– Excavation of P-APR/APR should be avoided where possible and always minimized.– The expected quantity of P-APR/APR should be estimated from construction drawings.

178

Figure 5. Partial encapsulation cross section view (TDOT 2007).

– Sites for disposal of all anticipated P-APR/APR should be identified.– On-site borrow areas from which adequate quantities of cover material for burial of the APR

should be identified.– Logistics for hauling P-APR/APR, the lime and limestone, and cover material to the disposal

sites during construction should be developed to eliminate, if possible, temporary storage of theP-APR/APR.

– Drainage should be diverted away from all excavations and encapsulating embankments ifpossible.

– Drainage ditches or other water conveyances along excavated and encapsulated APR should belined with geomembrane or other impervious material such as clay.

– Underdrains, pipe culverts, and storm drains in areas of excavated and encapsulated APR shouldbe constructed of inert plastic.

6.1.2 Blasting BMPsIf blast hole sampling and testing indicate the presence of P-APR/APR, blast designs may beadjusted to minimize the production of “fine-grained” P-APR/APR. This BMP is implementedonly if it results in blasted fragments that may be safely and cost-effectively loaded into haulagevehicles or moved into encapsulation zones.

6.1.3 Construction phase BMPsThree different construction phase BMPs are described in this section, including blending, partialencapsulation, and full encapsulation. These three methods are appropriate for different thresholds;however, variations or modifications to or between the methods may be appropriate given site-specific conditions or site-specific materials. These BMPs should be selected in consultation withTDEC. Four major BMPs were developed in the APR Guideline:

– Blending of P-APR and APR with APR-neutralizing material [i.e. limestone, calcareous shale,or rock material with a net neutralizing potential (NNP) value greater than 50 Tons of calciumcarbonate (CaCO3) per kiloton (kT) of rock]. Grain sizes and mixing recommendations areprovided.

– Partial Encapsulation (See Figure 5).– Full Encapsulation (See Figures 6 and 7) which may occur within the roadway or at a dedicated

waste site repository.

The full encapsulation BMP conceptual design includes both clay and geomembrane liners;Figure 6 shows the geomembrane liner option for a roadway embankment location of the wasterepository.

179

Figure 6. Roadway embankment full encapsulation cross section view (TDOT 2007).

Figure 7. Waste-Site repository encapsulation cross section view (TDOT 2007).

6.1.4 Cut slopes–ARD preventionCut slope ARD prevention BMPs include designing the slopes to be as steep as possible withingeotechnical stability constraints and public safety. Pre-split blasting to minimize rock face over-break is a BMP that limits exposure of APR to water and oxidizing conditions. If near-verticalslopes are not recommended, the slopes would be flattened to allow placement of nonAPR andplant growth medium. Bactericides, which are considered a temporary BMP, may be used in thiseffort to suppress pyrite oxidation as the plant community matures.

Other cut slope BMPs include: attention to bench designs, stabilizing friable rock slope covers,and rapid revegetation protocols. Post-construction BMPs include the placement of oxic limestonechannels and mixing of limestone into native soils/plant growth medium prior to revegetation.

7 WATER TREATMENT

While the goal of the guideline is to avoid generation of ARD, and if proper planning and mitigationBMPs have been followed, the likelihood of generating ARD should be minimized. However,treatment of ARD would be necessary if other implemented prevention measures have not achievedthe level of control required. Water treatment is costly, and in some cases, must be continued inperpetuity. In addition, this may not have occurred at some older sites that pre-date effectivemitigation methods.

The spectrum of ARD treatment ranges from active to passive and includes a “semi-passive”category. Active treatment processes typically require mixing and settling tanks, pumps, electricity,

180

chemical addition, and some level of filtration in addition to the labor required to operate andmaintain these systems. Active treatment plants also generate sludge which requires disposal on aregular schedule. Because of these permanent infrastructure requirements, active treatment systemsare deemed inappropriate for TDOT projects.

Passive treatment, on the other hand, consists of oxic limestone channels, free water surface wet-lands, and bioreactors that treat water without electricity, day-to-day labor, or chemical addition.Passive treatment systems (PTS) require occasional maintenance and must be refurbished, depend-ing on the type of system, every 10 to 20 years. The primary limitation of the PTS technology isthat large areas may be required to treat high flow rates and/or high metal concentrations. Sometypes of PTS may require National Pollutant Discharge Elimination System (NPDES) Permits.

Semi-passive treatment is an off-the-shelf technology that uses water-powered chemical feedersto add reagents either continuously or intermittently to ARD. The reservoirs of chemical reagentsrequire refilling perhaps on a monthly to bi-monthly schedule, depending on the ARD treatmentsituation.

The new APR Guideline was not intended to be a PTS design manual but instead to offerdirection for situations in which PTSs are appropriate. If a site requires water treatment, a qualifiedprofessional engineer should evaluate the site water, and design the appropriate PTS. Public-domainsoftware, AMD Treat©, is available from the internet to assist the project engineer in sizing anddesigning a PTS and/or a semi-passive treatment system in typical situations.

7.1 Water treatment implementation triggers

Water treatment should be initiated based on the following triggers:

– The source of the ARD cannot be eliminated or remediated, or– Water leaving the site is in violation of TDEC water quality criteria for Fish and Aquatic Life, or– Water leaving the site has a pH of less than 5 (site dependent).

The decision to treat water at a particular site will be based on a variety of site factors includingbackground water quality, flow rate, land ownership, historic land use, and future land use. In thecase of background water quality, it is possible for streams to have naturally-occurring pH valuesless than 5. In this situation, TDOT and TDEC could waive the water treatment requirement. Thisdocument attempts to provide generalized guidance for initiation of water treatment at potentialARD sites. The final decision to treat water at any particular site should be made based on thetriggers listed above and TDOT and TDEC recommendations.

7.2 AMD Treat© public domain software

AMD Treat© is a computer application for estimating remediation costs for mine drainage orgeneric ARD. Version 4.0 of AMD Treat© can be downloaded from the internet from the Officeof Surface Mining website (http://amd.osmre.gov/amdtreat.asp); the website also offers an on-line tutorial in learning how to use the software. The software can be used to estimate constructionquantities and costs (capital and operating) for a variety of passive and chemical treatment methods,including:

– vertical flow ponds – oxic limestone channels– anoxic limestone drains – caustic soda– anaerobic wetlands – hydrated lime– aerobic wetlands – pebble quicklime– bio reactors – ammonia– manganese removal beds – oxidation chemicals, and– limestone beds – soda ash treatment systems.– settling ponds

The treatment estimating modules in bold above have been identified as preferred treatmentmethodologies at TDOT sites. However, these preferences are not necessarily all-inclusive andother methodologies may be appropriate.

181

7.3 Water treatment methods

7.3.1 Short-term semi-passive treatmentIf ARD is discovered during construction, immediate capture and semi-passive treatment of thewater should begin to prevent off-site impacts. This short-term treatment method will be employeduntil a permanent system is designed and built. Short-term semi-passive treatment measures follow.

– Retention Pond Sizing–The flow rate of the ARD should be measured. If the ARD flow isthe result of precipitation events, a qualified hydrologist/engineer should estimate the 10-yr,24-hr runoff volume. A geomembrane-lined retention pond with a 24-hr retention time shouldbe constructed to capture the ARD. See the following modules in AMD Treat©: Ponds, FlowCalculation Tools, and Acidity Calculator. Periodic sediment and/or sludge removal will berequired for the retention pond. Clean stormwater should be diverted from the retention pond.

– Aquafix™ Treatment–Aquafix™ units are water-wheel powered pebble lime-dosing machines.Aquafix™ systems require neither electricity nor constant monitoring but function better undercontinuous flow conditions. If the ARD flows are intermittent but can be stored and released asa continuous feed, an Aquafix™ unit may be appropriate. Contact and ordering information forAquafix™units can be found at http://www.aquafix.com/. See the following modules in AMDTreat©: Ponds, Pebble Lime, Flow Calculation Tools, and Acidity Calculator.

– Wheel-treaterTM Treatment–Wheel-treater™ units are water-wheel powered caustic soda(sodium hydroxide solution)-dosing machines. These units require neither electricity norconstant monitoring. They function well under both continuous and intermittent flowconditions. Contact and ordering information for Wheel-treater™ units can be found athttp://www.chemstream.com/. See the following modules in AMD Treat©: Ponds, Caustic SodaFlow Calculation Tools, and Acidity Calculator.

– Other Semi-Passive Units–Vendors offering semi-passive units that feed limestone or otheracid-neutralizing reagents should be investigated on a case-by-case basis.

– Water Treatment Sampling Program–A water quality sampling program should be initiatedas soon as the retention pond receives water. Pond influent and pond water samples shouldbe collected and analyzed for the parameters listed on the advance sampling suite. The pondinfluent sample should be collected upstream of the pond and the semi-passive unit. The pondwater sample should be collected from the surface of the pond near the pond discharge point.If the ARD flow is driven by precipitation events, samples should be collected after significantprecipitation events (rainfall > 1 inch in 24 hours). A sampling quality control plan should bedeveloped in accordance with TDEC regulations to ensure a successful sampling program.

– Semi-Passive Reagent Feed Rate Adjustment–The target pH for pond water should be 8 or less,depending on the pH of the receiving stream. Increasing the pH to this level should remove asignificant portion of metals. Based on the pH levels measured in the pond water, the lime feedof the Aquafix™ unit or the caustic soda feed of the Wheel-treater unit should be adjusted toprovide the target pH level.

– Constituents of Concern and Reporting–Sampling results should be reported toTDOT andTDECon a quarterly basis and after the completion of construction. Based on the sampling results, alist of contaminants of concern should be developed upon which to base future sampling efforts.

7.3.2 Long-term passive treatment implementationAfter the short-term semi-passive treatment system is in place, the long-term PT implementationphase begins and consists of the design and construction of a suitable PT system to address mitiga-tion of ARD at the site. After an appropriate PT system has been constructed and commissioned,the operation of the semi-passive unit can be suspended. However, retaining the semi-passive uniton site in standby status is recommended for at least six months. The APR Guideline providesdecision criteria for three different types of PT systems as listed below. Some types of PTS mayrequire National Pollutant Discharge Elimination System (NPDES) Permits. Long-term passivetreatment measures addressed in the APR Guideline follow.

– Analyze water quality data from the short-term semi-passive treatment phase.– If the site water has a pH < 5 or if any metals concentrations exceed the TDEC water quality

criteria, long-term PT will be required. The site conditions and water quality will dictate whichPT system (PTS) is appropriate among the options listed below.

182

Figure 8. Decision tree for selection of long term passive treatment system (PTS).

1) PTS I–Settling Pond, Open Limestone Channel (OLC)2) PTS II–Settling Pond, Surface Flow Wetland (SFW)3) PTS III–OLC, Setting Pond, two Sulfate-Reducing Bioreactors (SRBRs), and SFW.

Detailed descriptions of these systems, sizing criteria, and installation guidance are provided inthe APR Guidance document. A decision tree diagram for choosing the most appropriate PTS isshown on Figure 8. Sulfate reducing bioreactors are discussed in more detail in Gusek (2002).

8 POST-CONSTRUCTION MONITORING

If P-APR/APR materials are identified during the course of the project, then the guideline indicatesthat post-construction monitoring should be performed for a minimum of two years followingconstruction to ensure that mitigation and design measures are working effectively. If a PT systemis constructed, monitoring should be performed as long as the system is in operation. If adverseimpacts from APR disturbance/exposure develop, they would most likely be detected in surfacewater, runoff, or groundwater associated with the project. Sampling of rock in the post-constructionphase is impractical relative to water sampling.

8.1 Monitoring locations

The guideline recommends that any area of construction that contains P-APR/APR materials shouldbe monitored. Monitored areas include, but are not limited to:

– road cuts fill zones, constructed or exposed embankments, and blended fill areas associatedwith APR

– structures designed for encapsulation, mitigation, or remediation of P-APR/APR– and PT systems.

183

In order to monitor these areas, designated sampling points should be established to capturegroundwater, seepage, and runoff from these areas. Surface water sampling points should includeprovisions for flow rate measurement, if this data requirement is triggered. Monitoring locationsshould be established in a site-specific monitoring plan to monitor areas associated with APRmaterials.

The sample locations should be accounted for during the pre-construction design phase to ensurethat the sampling sites will provide representative samples of water leaving the site. If impacts arenoted down-gradient, appropriate up-gradient samples should be collected.

8.2 Monitoring period

The monitoring period should be established in a site-specific monitoring plan that accounts forthe specifics of each project. It is recommended that water sampling should be performed on aquarterly basis for the first year following construction, or in accordance with permitting, andsemi-annually until one year after vegetation is established on cut faces, graded areas, slopes, andembankments; however, this frequency may be varied based on site conditions and professionaljudgment. If no indication of ARD generation is shown in this time, sampling may be discontinued.Background groundwater should be sampled on the same frequency as down-gradient waters.

If no indication of ARD generation is observed during these monitoring periods, sampling maybe discontinued. If indications of ARD are observed, sampling should be increased to bi-monthly inorder to evaluate the ARD generation. PT systems should be monitored on a quarterly basis for thefirst year following construction and on a semiannual basis thereafter. Treatment systems shouldbe monitored as long as they are in operation. If a PT system is regulated by a NPDES permit, thepermit will specify the monitoring frequency.

8.3 Monitoring suite

The analysis suites for post-construction monitoring are the same as those presented in the “WaterTesting Methods” section of the APR Guideline. Two sampling suites are specified there: 1) if ARDis not present, the analysis suite should include an “abbreviated” set of parameters; 2) if ARD isknown to exist or if a PT system is in operation, an extended sampling suite is recommended. Thesampling suites can be modified based on site conditions and professional judgment.

9 CLOSING REMARKS

Roads are relatively narrow, linear design features when compared to typical mining disturbancesthat may or may not expose APR. As such, it is not appropriate to apply the TDOT APR Guidelineto mining situations. There are many handbooks and guidelines available for mitigating acidicand neutral/alkaline mining influenced water (MIW), including the free, on-line GARD Guide(www.gardguide.com) compiled by a mining industry consortium, the International Network forAcid Prevention [INAP], and recent publications developed by the Acid Drainage TechnologyInitiative’s (ADTI’s) metal mining and coal mining sectors. See Gusek and Figueroa (2009) for anexample of one of six ADTI Metal Mining Sector guide books available. Collectively, these aregeneric guideline publications whose recommendations can be embraced just about anywhere onthe planet.

In contrast, TDOT’sAPR Guideline is a site-specific guideline, applicable toTennessee’s geologicconditions, that builds on the general knowledge related to MIW that has been compiled by theINAP and ADTI practitioners. As such, it has prescriptive controls and triggers incorporated intoit that are intended to provide consistent APR management decisions which are jointly accepted byroad designers and the state environmental agency that approves those designs. One of its betterfeatures is planned flexibility: it is considered a living document (particularly the GIS dataset) thatshould improve over time provided that the findings of others can also be incorporated to update itwhen needed. It also allows professional judgment to override prescriptive controls. Some of thosecontrols, or BMPs, might find reverse application in certain mining situations if properly tailoredto the individual mining site.

184

REFERENCES

Byerly, D.W. 1990. Guidelines for Handling Excavated Acid-Producing Materials. (FHWA-FL-90-007).[Washington, DC]: US Federal Highway Administration. 81 p.

EPA. 1994. Acid Mine Drainage Prediction; Technical Document. (EPA-530-R-94-036; available NTIS PB94-201829). Washington, DC: US Environmental Protection Agency, Office of Solid Waste.

Gusek, J. J. 2002. Sulfate-Reducing Bioreactor Design and Operating Issues: Is This the Passive TreatmentTechnology for Your Mine Drainage?, presented at the National Association of Abandoned Mine LandPrograms, Park City, Utah, September 15–18, 2002.

Gusek, J. and L. Figueroa (eds.) 2009. Mitigation of Metal Mining Influenced Water. Littleton, CO. Society forMining, Metallurgy, and Exploration, Inc. for ADTI Metal Mining Sector.

Hardeman, W. D., 1966. State [Tennessee] Geologic Map, scale 1:250,000 (1 inch = 4 miles), in 4 sheets.Nashville, TN: Tennessee Department of Environment and Conservation

Moore, H. 1992. The Use of Geomembranes for Mitigation of Pyritic Rock. In 43rd Annual Highway GeologySymposium, Fayetteville, AK, August 1992. Asheville, NC: The Symposium.

Nordstrom D.K & Alpers, C.N. 1998. Geochemistry of Acid Mine Waters. In G.S. Plumlee & M.J. Logsdon,(eds.) The Environmental Geochemistry of Mineral Deposits, Part A: Processes, Techniques, and HealthIssues. (Reviews in Economic Geology Volume 6A). Littleton, CO: Society of Economic Geologists, Inc.

Skousen J.G., Sencindiver, J.C., & Smith, R.M. 1987. A Review of Procedures for Surface Mining and Reclama-tion in Areas with Acid-Producing Materials, in cooperation with the West Virginia Surface Mine DrainageTask Force, the West Virginia University Energy and Water Research Center, and the West Virginia Miningand Reclamation Association. Morgantown, WV: The Center.

TDOT. 1990. State ofTennessee Special Provision RegardingAcid Producing Materials, revised May 30, 2003.Nashville, TN: Tennessee Department of Transportation.

TDOT. 2005. Standard Operating Procedure for Acid Producing Rock; Investigation, Testing, Monitoring,and Mitigation, revised July 2005. Nashville, TN: Tennessee Department of Transportation, GeotechnicalEngineering Section.

TDOT. 2007. Guideline for Acid Producing Rock Investigation, Testing, Monitoring and Mitigation, preparedby Golder Associates, Inc., Lakewood, CO. October, 2007.

185


20-day design build to save $50 million worth of equipment

S.J. TamburiniTetra Tech Inc., Denver, CO, USA

S.J. BillinTetra Tech Inc., Elko, NV, USA

ABSTRACT: To prevent the flooding and loss of $50 million worth of equipment in an under-ground nickel mine, a new Effluent Water Treatment Plant (EWTP) was design, procured, built, andcommissioned in 20-days. The mine’s interim tailings storage facility which was storing the dewa-tering water was full, and the mine was unable to meet its discharge permit without the constructionof the EWTP.The EWTP targeted the removal of total suspended solids, radium, ammonia, and totaldissolved solids from the water stored in the interim tailings storage facility. It was designed as anintegrated membrane filtration plant with microfiltration and reverse osmosis membrane systems.Through great team work from the mine staff, the design engineer, contractors, and equipmentsuppliers, the EWTP was completed and met permit limits within the extremely tight timeline.

1 BACKGROUND

Crowflight Minerals Incorporated owns and operates the Bucko Lake Mine (Mine) near Wabowden,Manitoba. The Mine is underground with onsite milling facilities for nickel production. The Mine’sdewatering pumps keep the mine shaft dry by pumping at a rate between 300 to 400 gpm. When themine failed to meet its discharge permit regarding total suspended solids (TSS) and trout toxicity,they diverted the dewatering flow to their interim tailings storage facility (ITSF) until a solutioncould be found, and the discharge could meet permit limits. The mine hired Tetra Tech Inc. to helpthem develop and implement a strategy to bring the mine into compliance with their dischargepermit. When Tetra Tech was hired, there was 20-days of storage in the ITSF before it exceededits design capacity. If the ITSF reached its capacity the mine would be faced with the difficultdiscussion to either discharge water that knowingly violated their permit (which was illegal), orturn off the dewatering pumps which would cause the mine shaft to flood, destroying $50 millionworth of equipment.

The first step to solve the problem was to develop a treatment strategy that could reliably meetall permit requirements. Once the conceptual treatment scheme was developed, the EWTP neededto be designed in such a way that the contractors could begin construction without waiting forthe design to be completely finished. One of the biggest challenges was to procure the necessaryequipment that closely met the design requirements which could be delivered to the site within therequired timeline. Once construction was complete, the EWTP needed to be commissioned, andeffluent quality needed to meet the discharge requirements in order resume discharging.

2 WATER QUALITY

To develop a treatment strategy, the contaminants of concern were first identified. The Mine collectseffluent water samples on a regular basis which are analyzed for total metals, dissolved metals,inorganic compounds, specific organic compounds, and other important water quality parameters.By comparing the water quality data to the discharge permit parameters, the contaminants ofconcern were identified. Table 1 summarizes the Mine’s permit parameters and Table 2 summarizes

187

Table 1. Discharge permit limits.

Parameter Units Grab Sample Maximum 30-day Maximum Average

pH S.U. 6.0–9.5 6.0–9.5Arsenic mg/L 1 0.5Copper mg/L 0.6 0.3Lead mg/L 0.4 0.2Nickel mg/L 1 0.5Zinc mg/L 1 0.5TSS mg/L 30 15Radium 226 Bq/L 1.11 0.37Toxicity Trout LCD50 NA >100%Toxicity Daphnia LCD50 Report Report

Table 2. Untreated discharge water quality summary.

Parameter Units Average Value Maximum Value Number of Exceedances*

pH S.U. 7.8 7.9 0Arsenic mg/L 0.0006 0.0008 0Copper mg/L 0.003 0.006 0Lead mg/L 0.0004 0.0007 0Nickel mg/L 0.38 0.49 0Zinc mg/L 0.017 0.040 0TSS mg/L 30.7 72.0 3Radium 226 Bq/L 0.20 0.50 1Toxicity Trout LCD50 35% 60% 2Toxicity Daphnia LCD50 NA NA NA

* Number of exceedances is the number of individual samples that exceeded the 30-day maximum averagevalue and does not necessarily indicate a permit violation.

the average and maximum discharge values along with the number of samples greater than the30-day maximum average.

Of the permitted water quality parameters, the main contaminants of concern include TSS,radium 226, and trout toxicity. The water quality data presented shows that all three of theseparameters exceeded their allowed discharge value at least once. Each of these parameters wasviewed individually in order to determine a reliable overall treatment approach. It is difficult topin point a treatment process for trout toxicity that will be “guaranteed” to work because the exactparameter responsible for killing the fish in the toxicity test cannot be determined. To determinean appropriate treatment technology to limit toxicity, the water quality as a whole was evaluated.

3 ENGINEERING

To develop the complete treatment process, each of the three parameters of main concern (TSS,radium 226, and trout toxicity) were evaluated separately to determine the recommended individualprocess for that parameter. Each one of these parameters can be treated using various treatmentsystems; however, to create an overall efficient treatment scheme, processes that can removemultiple contaminants were identified and were given preference over processes that target singlecontaminants.

3.1 TSS

Two main options that could be implemented within the given timeline capable of removing the TSSfrom the effluent water were identified. These options were chemical addition with sedimentation,

188

and microfiltration (MF). Chemical addition followed by sedimentation would include the additionof a metal salt (such as alum or ferric chloride) and/or addition of a polymer to improve settleabilityof the solids suspended in solution. The chemicals would be added to the process water in the millbefore it is sent to the existing settling ponds where the solids would be allowed to settle out ofsolution. The main advantage of this process was that it did not require much additional equipment,and it was easy to mobilize to the site. The main disadvantage of this process was that it is difficultto control or ensure the facility will meet the discharge limit. This treatment method also adds tothe total dissolved solids which could become a future permit issue.

The MF option consisted of leasing or purchasing a used membrane microfiltration system thatcould be used to physically separate the solids from the water. Prior to visiting the site, the engineercalled manufacturers check the availability of already manufactured microfiltration systems thatwere available. It was found that a model AP-6 manufactured by Pall Corporation was currentlyavailable. This system consists of a skid mounted module with racks of membranes, a control skid,a feed tank, and a clean-inplace (CIP) solution tank. An AP-6 is rated for a flow rate of 750 gpm andis approximately 98% efficient (meaning that 2% of the water processed through the membraneis used for cleaning sediment off the membranes). The main benefit of the MF system was that itgives the Mine a positive barrier which can be controlled to deal with TSS removal. MF is ideal forremoval of suspended solids, and the process should remove TSS to approximately 1.0 mg/L whichis significantly lower than the permitted value of 15 mg/L. Chemical addition and sedimentationcannot achieve TSS values in this range and it is significantly less reliable. Using the MF system, itwould take multiple catastrophic events before the TSS would be violated. The main disadvantageof this process was that it requires construction of an EWTP and a structure to house the equipment.

3.2 Trout toxicity

For trout toxicity, it is difficult to determine what treatment techniques need to be applied becausethis parameter does not target a specific contaminant. To assure compliance with this parameter thewater quality was viewed as a whole. The main parameters of concern for toxicity are ammonia,nitrite, strontium, and possibly total dissolved solids (TDS). The average ammonia concentrationin the water of 23 mg/L can be toxic to fish and required removal. While the nitrite concentration israther low with an average concentration of 0.97 mg/L, the nitrate concentration is relatively highwith an average concentration of 20 mg/L. Nitrite is extremely toxic to fish while nitrate is not;however, a high nitrate concentration is still concerning, because, under certain conditions nitratecould be reduced to nitrite. The average strontium concentration of 19 mg/L is also higher than thesecondary acute value which could have negative impacts to fish.

3.2.1 AmmoniaThe main contaminant of concern with regard to trout toxicity is ammonia. There are severaltreatment techniques that remove ammonia that could be implemented within the timeframe. Thesealternatives include reverse osmosis membranes, break point chlorination, and ammonia stripping.Reverse osmosis (RO) uses a membrane with very small pore sizes that allows water to passthrough based on osmotic pressure. RO will also remove other metals and contaminants, suchas nitrate and strontium, at the same time. The main disadvantage of RO is that it requires veryhigh operating pressure (approximately 300 psi) which results in high energy consumption. This isnormally a serious concern for RO systems but less of a concern in Manitoba where energy costsare significantly lower than average. This process would also require additional equipment leasesor purchases thus increasing the cost of treatment system.

Break point chlorination consists of adding a source of chlorine (usually sodium hypochlorite)to oxidize the ammonia to nitrate. Ammonia stripping requires the pH to be raised to approximately11.5 which will convert all ammonia to a gaseous form which can then be stripped from solution.Ammonia stripping was not considered further, because it requires chemical addition for raisingand then lowering the pH, plus it requires stripping towers that have a high capital cost. Long-termuse of treatment options that rely on significant chemical addition for pH adjustment or oxidationare not unfavored as they will gradually raise the TDS in the ITSF reclaim as the waste streams arerecycled back to the tailings circuit.

Break point chlorination is an effective way to remove ammonia to reduce the toxicity of theeffluent. This process includes adding sodium hypochlorite (NaOCl) at an approximate dose of

189

10 mg NaOCl per mg ammonia removed. The hypochlorite will oxidize the toxic ammonia tonitrate which is less toxic. Excess hypochlorite is also toxic thus requiring the installation ofdechlorination prior to discharge to the environment. Dechlorination can be accomplished throughthe addition of sodium thiosulfate or other dechlorination chemicals. The main advantage of thisprocess is that it does not require much equipment, and it can be implemented very quickly. Themain disadvantage of this process is that it requires close operational control to assure enoughchlorination is provided and enough dechlorination is provided.

Special considerations were considered to readily remove ammonia using RO membranes. Bothmolecule size and molecular ionic strength contribute to the removal of contaminants using ROmembranes. The pore size of the RO membranes is designed to allow molecules the size of water topass through while larger molecules cannot. The size of an ammonia molecule is similar to water,and ionic strength of ammonia is similar to water, indicating that ammonia cannot be removed byRO membranes. While this is true, the acid and base chemistry of ammonia indicates that the pHconditions of the water at the Mine convert nearly all of the ammonia (NH3) to the ammonium ion(NH+

4 ) form. The ionic strength of ammonium is greater than that of water; therefore, ammoniumis removed by the RO membrane. If the pH of the mine water increased, ammonium would beconverted to ammonia and the removal efficiency would decrease. It is important to maintain a pHless than 8.2 to remove ammonium.

3.3 Radium

For radium treatment, the Mine was historically adding barium chloride (BaCl2) to the dewateringwater before the final settling pond as well as in the underground shaft. The BaCl2 reacts withsulfate in the water to form a barium sulfate precipitate which settles out of the water. While thebarium sulfate settles, the flocs attract and entrap the radium in the water, thus also removing itfrom solution. This treatment technique was usually successful yet occasionally unreliable.

There are several other ways to remove radium from solution. These methods include ionexchange, adsorption, and hydroxyl manganese oxide (HMO). These three treatment alternativesall operate on the same premise; they use electrostatic attraction to pull the radium from the wateronto the surface of a media. For ion exchange and HMO, the main disadvantage is that once themedia is saturated with radium and other like cations, the media needs to be regenerated. The wastefrom the regeneration process would require disposal in some manner. In this case the waste fromthe regeneration process would have to be either recycled to the tailings pond or treated further toproduce a solid. Both of these disposal options have significant negative ramifications. Recyclingthe waste to the tailings pond would cause a buildup of cations in the system creating a cycle ofmore frequent regeneration cycles. In the long term this operation would not be feasible; how-ever, as an interim treatment, further study would be needed to determine whether the buildup ofcations would cause the process to fail within the interim treatment timeframe. Adsorption has asimilar negative impact. When the adsorption media is saturated, it requires removal and disposal.Considering radium is a radionuclide, disposal of the media may be a permitting issue.

Another treatment option for radium removal is RO membranes. The radium particles are toolarge to pass through the RO membranes and the radium remains in the RO reject. Considering therecommendation to deal with trout toxicity includes installation of these membranes, the radiumremoved by the RO membranes was accounted for. Without installation of a larger RO systemthe influent radium concentration is high enough to require a higher blend of RO treated water tomeet the discharge limit. Based on the treatment alternatives available, it was recommend the Minecontinue to use BaCl2 addition and sedimentation.

3.4 Design criteria

The recommended treatment strategy included installation of MF followed by RO. The MF systemwill remove TSS and will provide adequate pretreatment for the RO membranes. RO will removealmost all contaminants in the water including ammonium, nitrite, nitrate, strontium, and all othermetals in the water. The RO system was installed after the MF system so that the TSS does not clogthe small pore spaces of the RO membranes. The RO system was designed so that only a portionof the water from the MF would be sent to the RO system to limit the size of the RO treatment

190

system. The portion of the water that is only treated by the MF system will be blended with the ROpermeate so that the blended water meets discharge limits and complies with toxicity tests.

The EWTP needed to be designed to treat the maximum dewatering flow rate of 400 gpm. Whenthe dewatering pumps can keep up using a lower flow rate, the Mine wanted the ability to reduce thevolume in the ITSF by treating the water. The design RO to MF blend was 60% meaning that 60%of the water treated by the MF was also treated by the RO system. The MF system acquired hadexcess capacity and could treat 750 gpm however, while only 400 gpm flow rate was required. TheRO system had a input capacity of 230 gpm with a 75% recovery or a permeate flow of 175 gpm.The wastewater from the EWTP was sent back to the ITSF.

4 PROCUREMENT

Procurement and delivery of the major pieces of equipment was the most difficult part of the projectand required the most timing coordination. When determining the possible treatment alternatives forthe project, the engineer contacted Pall Corporation to see if they had any pilot scale treatment skidsavailable for immediate delivery to the mine. Pall happened to have one 750 gpm treatment skid insouthern California ready for shipment back to Pall’s factory for refurbishing. The mine signed atemporary lease and the MF skid was shipped directly to the mine site. Considering the equipmentwas being shipped internationally, the equipment had to clear customs at the US and Canada border.To assure the equipment cleared customs without being delayed, the Mine hired a shipping broker.

RO membrane manufacturers also needed to be contacted to determine if they had any systemsavailable for immediate delivery. While there were several pilot units identified, none of themcould be shipped within the timeline; however a newly manufactured RO system was availablefrom Applied Membranes. The fastest this RO skid could be delivered to the site was three daysbefore the dewatering pumps would have to be shutoff. This meant all equipment and piping hadto be installed and ready for operation when the RO skid arrived, so it could be connected, wiredand started by the deadline. This equipment was manufactured in California, and it also requiredthe help of a shipping broker to minimize possible delays.

Booster and transfer pumps that met the design criteria were also required. Local suppliers outof Winnipeg were contacted to determine what pumps were in stock that could be shipped in timethat could also meet the flow and head conditions required. While the EWTP was designed tohave a standby redundant pump, the suppliers only had one pump in stock. The Mine purchasedthe stocked pump and ordered a second pump for installation at a later date. The combination ofCanadian and American equipment created another problem. The Mine’s mill and the Canadianpumps used a 600V power supply while the American equipment used 480V motors. In order totransform the 600V power supply available to the 480V required by the MF and RO systems atransformer was needed. A used transformer was acquired; although when it arrived onsite, it wasnot functional and needed to be repaired.

The Mine is remotely located; therefore, there are not construction supply companies readilyavailable. To ensure the Mine can continue operation and not experience significant down time dueto equipment or mechanical failures, the Mine maintains an onsite warehouse. In the warehouse,pipe, pipe fittings, concrete, power cable, control panels, motor starters, etc are stored; therefore,construction did not rely on shipment of construction materials. Once the design phase of theproject was nearly complete the, Mine had to make one order for additional materials that were notstored in the warehouse.

5 DESIGN BUILD AND COMMISSIONING

5.1 Design-build

The construction of the EWTP began immediately after the conceptual design was complete. Tocomplete the project in the timeframe allowed, it required good teamwork between the engineer,contractors, mine staff, and equipment manufacturers. Typically the progression of a design projectdoes not follow exactly the sequence of construction; however, in order to complete the projectwithin the timeline, the design process was modified so that the engineer’s design was one step

191

Table 3. Treated discharge water quality summary.

Parameter Units Average Value Maximum Value

pH S.U. 8.0 8.1Arsenic mg/L 0.0004 0.0004Copper mg/L 0.0005 0.0008Lead mg/L 0.00007 0.0002Nickel mg/L 0.12 0.14Zinc mg/L 0.006 0.011TSS mg/L 4.3 9.0Radium 226 Bq/L 0.07 0.12Ammonia mg/L 6.7 8.1Toxicity Trout LCD50 0% 0%

ahead of the contractor. For example, the first step in the construction of the EWTP was to installthe concrete slab. The engineer needed to figure out the size of the pad and the approximate weightwhich it needed to support before the complete equipment layout was complete. The progression ofthe design portion of the project was driven by what the next task was in the construction sequence.While the contractors were waiting for the design of the next construction item at the EWTP, theyhad to construct the piping within the Mill to transport the tailings water to the EWTP and transportwastewater from the EWTP back to the tailings pond. This process required input from the engineersand mine staff to determine how to incorporate the EWTP without affecting the milling process.

Several problems arose during the construction. First, when the 600V to 480V transformer arrivedonsite, it was not operational and some of the internal wiring had to be replaced. This delayed thestartup of equipment and minimized the amount of time for troubleshooting and commissioning.Another problem encountered during construction was the air compressor for the MF system wasnot functional. The MF equipment was scheduled to be refurbished at the factory before being sentto the next project, so the MF manufacturer had to refurbish the equipment onsite. In order to getthe equipment running as soon as possible, a temporary compressed air line was installed from themill compressed air system to the MF unit. This effort was led by the MF system manufacturerand the Mine staff, demonstrating the importance of having a good relationship with all partiesinvolved in the project.

5.2 Commissioning

Due to finalizing electrical power issues, the equipment could not be started until there was less than2 days before the dewatering pumps had to be turned off. The majority of problems that occurredduring the commissioning required electrical troubleshooting to make sure the power and controlwiring were correctly installed at the control panels. By the end of the first day of commissioning,the MF system was operable; however the RO system would only run for a few minutes before itwould shutdown due to low pressure alarms. To correct this problem some last minute changes wererequired. The controls between the RO booster pumps which were fed by the MF system requiredsignificant changes to the timing which could have jeopardized the commissioning. To change thiscontrol loop so that the EWTP could be operated in a manual mode, a small tank was installed. TheMF system effluent filled the tank which was connected to the suction of the RO booster pumps.This last minute change in the design allowed the EWTP to be started in manual operation.

The effluent quality of the treatment system met the discharge limits and the design criteria imme-diately upon manual startup. The effluent flow rate was set at 400 gpm to offset the dewatering flowrate. Sixty percent of the water was treated by MF and RO which was blended with the remainder ofthe water that was only treated by the MF system. Once the Mine began to discharge, they collectedfour consecutive days of water samples. The water quality results of the treated water are shown inTable 3. Once the results of the effluent samples were received and it was noted that the effluentquality was not close to a violation for any parameter, the percent RO/MF split was decreased from60% to 50% to increase the overall throughput of the EWTP. After this operational change, theeffluent water quality concentrations increased slightly; however, they were below the permit limits.

192

After the plant was operational in manual mode, the EWTP was automated in the following daysto minimize the required operations time. After automation was complete, the Mine’s existing milloperations staff was able to incorporate the operations of the EWTP into their normal operations.Considering the MF system was oversized and the Mine wanted to quickly decrease the water levelin the ITSF, they leased a second RO system to increase the through put of the plant.

6 FUTURE CONSIDERATIONS

The project was constructed in the summer months which allowed the EWTP to be constructed andoperated without an enclosure. After the EWTP was operational, the Mine ordered a pre-engineeredstructure to house the treatment plant to prevent freezing during the winter months.

Due to the tight timeline of the project, disposal of the RO reject was not considered. The ROreject was recycled back to the ITSF where it was mixed with rest of the tailings water and iseventually retreated at the EWTP. This continuous cycle causes a buildup of TDS in the tailingspond, and over time, it would affect the overall treatment and milling process. An increase in TDSin the tailings water will also lead to decreased performance of the RO membranes resulting ina lower permeate flow rate. If the TDS increase goes unchecked, the permeate flow rate coulddecrease to the point where the Mine could not get the right blend of RO treated water and a permitviolation could result.

To prevent the buildup of TDS in the system, a RO reject handling system must be designed. ROreject treatment system alternatives include passive evaporation, mechanically enhanced evapora-tion, crystallization, or deep well injection. Manitoba receives more rain than there is evaporation;therefore passive evaporation is not an option. The remaining possible options for the RO rejectdisposal have high capital and operations costs associated with them and therefore require carefulconsideration.

7 RESULTS

In 20 days, the mine staff, contractors, and engineering team was able to design, procure, deliver,build, and commission a MF and RO treatment system. Developing a treatment strategy involvedidentifying contaminants of concern, then evaluating what treatment options could be implementedgiven the timeline available. By maintaining good relationships with manufacturers and workingas a team, the equipment was procured and delivered. The design approach was modified so thatthe design was one step ahead of the construction sequence. The EWTP was able to meet dischargelimits as soon as all equipment was functional. By allowing the mine to resume discharging, themine was able to save over $50 million worth of underground mining equipment. In addition tosaving equipment, the treatment plant allowed the mill to resume the production of nickel. In thefuture the Mine plans to install calcite filters to restabilize the RO permeate so that the water canbe used for potable use. The mine currently hauls all potable water for the site by truck at the costof $14,000 per month.

193


The simultaneous removal of arsenic and manganese at a goldmine in Nevada

H.C. Liang, S.J. Billin & J.R. TamburiniTetra Tech, Inc. Denver, Colorado, USA

ABSTRACT: Arsenic, which occurs naturally in many parts of the earth, is commonly releasedinto the environment from mining activities. Not only is arsenic acutely toxic towards human healthat higher levels, low level chronic exposure and ingestion of arsenic is also a major human healthconcern and can lead to various diseases, including diabetes and cancer. The latest research alsoshows that arsenic can act as a potent endocrine disrupter and adversely impact hormone functionsin humans. The paper highlights a case study at a gold mine in Nevada where water with arseniclevels up to 4 mg/L were treated down to below 50 µg/L, and manganese levels were treated down tobelow 100 µg/L, using existing infrastructure at a water treatment plant by changing the treatmentprocess configurations. Both bench scale testing results and full-scale water treatment plant dataand analyses are presented, and the rationale for the refinement and changes to the arsenic removaltreatment processes are discussed. Studies of the water chemistry and the adjustments of both thetreatment chemicals used and the treatment procedures, such as the chemical addition points, arealso presented, and the challenges for the simultaneous removal of manganese and arsenic as wellas the relevant water chemistry behind the treatment processes are also discussed.

1 INTRODUCTION

Arsenic and metals removal are commonly required for mining-impacted waters due to the propen-sity of mining activities to mobilize these contaminants (Liang & Thomson, 2009; Liang &Thomson, 2008). Arsenic toxicology, both as an acute and chronic toxicant, is well-known (Hall,2002; Gebel, 2000). Arsenic can cause death when ingested at higher levels, while long-term con-sumption of lower levels of arsenic can lead to a host of diseases such as cancer, diabetes, andblackfoot disease (CDC, 2000). Many arsenic removal technologies that have been deemed BestAvailable Technologies (BAT) by the U.S. Environmental Protection Agency (U.S. EPA), such asactivated alumina adsorption, anion exchange, greensand filtration, coagulation/filtration, limesoftening, or reverse osmosis (RO) are commonly used for arsenic removal (EPA 2001). Althoughmanganese is considerably less toxic than arsenic (Howe et al., 2004), its removal from mine watersprior to discharge is also important due to the potential adverse effects of manganese on humansand aquatic life (Howe et al., 2004). The case study and research described in this paper focus onthe use of oxidation, ferric-based coagulation, clarification, and pH adjustment to achieve effectivearsenic and manganese removal at a mine water treatment plant (WTP). Because of the very differentwater chemistry parameters that are required for removing arsenic using ferric-based coagulation(vide infra) compared to the parameters required for manganese precipitation and removal (videinfra) and the inherent challenges in removing both contaminants in the same treatment process(Chang et al., 2006), the successful implementation of the water treatment process required carefulmodifications in chemical addition and doses at the existing WTP at the mine site in northernNevada.

2 BACKGROUND

In December of 2008, the authors were contacted by representatives at a gold mine in northernNevada. The WTP at the mine site had been treating the combined dewatering flows from two nearby

195

Figure 1. Original treatment process and flow diagram.

mines to remove high influent arsenic levels (up to 4 mg/L) and manganese for discharge, withtheir NPDES (National Pollutant Discharge Elimination System) permits stipulating that the arseniclevels need to be below 50 µg/L and that the manganese need to be below 100 µg/L. At that time,however, their treated water consistently contained arsenic and manganese concentrations higherthan 700 µg/L and 200 µg/L, respectively, and the treated water could not be discharged and hadto be diverted to a tailings pond. Analyses of the existing WTP and the original treatment processshowed that waters from two mines were pumped into a thickener feedbox, where the oxidantpotassium permanganate (KMnO4) and the coagulant ferric sulfate (Fe2(SO4)3) were added to theinfluent waters, which then flowed to a solids thickener for solids removal and then to a reactiontank followed by two parallel settling ponds for polishing before the treated water effluent wasdischarged to an infiltration basin (Figure 1).

In November of 2008, when effluent manganese concentrations started consistently exceedingthe discharge limit of 100 µg/L, the WTP staff ceased adding KMnO4 to the treatment processto decrease the overall manganese input into the treatment system. Thereafter, the treated waterarsenic levels started to climb dramatically, sometimes up to greater than 700 µg/L. The WTP staffresponded to the high treated water arsenic levels by increasing the dosage of ferric sulfate, whichnot only did not lower the arsenic levels but instead exacerbated the manganese problems in thetreated water, leading to treated water arsenic and manganese concentrations higher than 700 µg/Land 200 µg/L, respectively.

3 RESULTS AND DISCUSSION

Initial Data and Analyses. Although no arsenic speciation data were available when the authorswere first called to resolve the operational issues at the mine WTP in northern Nevada, it was clearfrom analyzing the data that most of the influent arsenic was in the form of reduced arsenic(III).Voluminous data and research exist which show that ferric coagulation is ineffective for removingarsenic(III), and that arsenic(III) would need to be oxidized first to arsenic(V) prior to coagulationfor effective removal (Hering et al., 1997). The vast contrast in removal efficiencies betweenarsenic(III) and arsenic(V) by ferric coagulation can be attributed primarily to the differencesin the overall charges between As3+ vs. As5+ compounds over the pH ranges of water that arecommonly encountered at WTPs. For example, from examining the Pourbaix Diagram (Pourbaix,1966) of arsenic in water (Figure 2), it can be seen that while arsenic(III) exists as the unchargedH3AsO3 arsenous acid over a wide pH range, arsenic(V) exists primarily as the oxyanions H2AsO−

4

and HAsO2−4 .

One of the consequences for the difference in overall charge between arsenic(III) and arsenic(V)compounds in water is that ferric hydroxide floc formed from the addition of a ferric coagulantwould more effectively adsorb and remove arsenic(V) due to the higher attraction between thenegatively charged arsenic(V) species and the positively charged ferric floc, especially at lower pHlevels (Figure 3). As seen from Figure 3, at lower pH, the average formula (not a discrete molec-ular formula) for ferric hydroxide changes from the neutral-charged “Fe(OH)3” to the positivelycharged “Fe(OH)+2 ,” which would more effectively adsorb negatively-charged arsenic(V) speciesdue to stronger coulombic attractions. Because of the ineffective removal of arsenic(III) by ferric

196

Figure 2. Pourbaix Diagram of arsenic in water.

Figure 3. Distribution diagram for ferric floc speciation vs. pH (Adapted from Chwirka et al., 2004).

coagulation, cessation of potassium permanganate addition at the WTP led to skyrocketing effluentarsenic levels.

When the WTP staff attempted to mitigate the resulting high arsenic effluent levels by increasingferric sulfate dosage, the effluent manganese levels increased. This can be explained by examiningthe speciation chemistry of manganese in water and from analyzing the effects of ferric coagulationon the pH. The hydrolysis of the ferric ions in water results in an increase in hydronium (H3O+)ion concentrations in water and decreases the pH of the solution (Equation 1):

197

Figure 4. Pourbaix Diagram of manganese in water.

Consequently, as the ferric sulfate dosage was increased at the WTP, the pH was decreasedfurther. As can be seen from the Pourbaix Diagram of manganese (Figure 4), manganese becomesmore soluble at lower pH levels as the reduction of oxidized, insoluble manganese oxide speciesto the soluble Mn2+ species becomes more facile while the desired reverse reaction of manganeseoxidation and precipitation becomes more thermodynamically unfavorable. Because of this, as thepH became further depressed from higher ferric sulfate dosages at the WTP, the effluent manganeselevels climbed even higher, both due to ineffective precipitation of dissolved manganese as well asthe re-dissolution of precipitated manganese oxides in the settling ponds.

Process Improvements. After analyzing the data, it was determined that the WTP needed tore-introduce an oxidant to the treatment process. Although permanganate is a highly effectivereagent for oxidizing arsenic, and the manganese content from permanganate could be removedas an insoluble manganese oxide if the process were finely tuned, because the process controlsat the WTP were not well-developed, the authors felt that switching to another oxidant such assodium hypochlorite (NaOCl) would be better to insure that no additional manganese sources wereintroduced into the WTP.

Bench scale testing was conducted which led to optimization of the NaOCl dosage to effectivelytreat the influent waters for both arsenic and manganese. Furthermore, it was discovered in thecourse of the investigations that the ferric sulfate that was used at the WTP contained approximately600 mg/kg of manganese. Therefore, usage of the batch of ferric sulfate was discontinued andreplaced with polyferric sulfate (PFS) which was assayed and determined to contain negligiblelevels of manganese impurities. Besides helping to decrease the amount of manganese added intothe WTP, PFS was also chosen because it had been shown to be more effective at arsenic removalthan ferric sulfate (Fan et al., 308).

Another process improvement, based on bench scale testing results, was to divide the ferricaddition to two steps, where it was added at both the thickener feedbox and after the solids thickener(Figure 5). The two-step PFS addition led to higher arsenic removal efficiency and lower overallPFS usage compared to single-step PFS addition. Not only did the two-step PFS addition lead tolower PFS dosage requirements and chemical cost savings on PFS, it also helped to decrease theamount of pH depression from ferric addition and improve manganese removal. Sodium hydroxide(caustic soda) was also added prior to the settling ponds to increase the pH above 8 to furtherfacilitate manganese precipitation and removal in the settling ponds. The process flow diagram ofthe improved process is shown in Figure 5.

After implementing the process modifications and improvements described above, the WTPconsistently treated the mine waters to below 50 µg/L arsenic and 100 µg/L manganese, most ofthe time with arsenic levels below 10 µg/L.

198

Figure 5. Improved treatment process and flow diagram.

4 CONCLUSIONS

Comprehensive knowledge of water and inorganic chemistry is critical in effectively treatingmining-impacted waters for inorganic contaminants such as arsenic and manganese. Although thewater parameter requirements for arsenic and manganese removal are vastly different, by under-standing the speciation chemistry of both contaminants, and by conducting bench scale testing todevise the optimal treatment schemes and implementing process changes based on the bench scaletesting results, the treatment process at a mine dewatering WTP in northern Nevada was improved,and the effluent arsenic and manganese levels were brought back to compliance.

ACKNOWLEDGMENTS

HCL is grateful to W. Brinson Willis and Sarvin S. Tabatabaei for their kind help on other aspectsof this project which were not discussed in this paper.

REFERENCES

Centers for Disease Control (CDC). 2000. Case Studies in Environmental Medicine: Arsenic Toxicity.Atlanta, GA.

Chang, Y. J.; Black, B. D.; Chang, D.; Gehling, D. 2006. Advanced Processes for Simultaneous Arsenic andManganese Removal. AWWA Research Foundation, Denver, CO.

Chwirka, J. D.; Colvin, C.; Gomez, J. D.; Mueller, P. A. 2004. Arsenic removal from drinking water using thecoagulation/microfiltration process. J. AWWA 96(3): 106–114.

Fan. M.; Brown, R. C.; Sung, S. W.; Huang, C.-P.; Ong, S. K.; van Leeuwen, J. H. 2003. Comparisons ofpolymeric and conventional coagulants in arsenic(V) removal. Water Env. Res. 75(4): 308–313.

Gebel, T. 2000. Confounding variables in the environmental toxicology of arsenic. Toxicology 144: 155–162.Hall, A. H. 2002. Chronic arsenic poisoning. Toxicol. Lett. 128(1–3): 69–72.Hering, J. G.; Chen, P.-Y.; Wilkie, J. A.; Elimelech, M. 1997. Arsenic removal from drinking water during

coagulation. J. Environ. Engin. 123(8): 800–807.Howe, P.; Malcolm, H.; Dobson, S. 2004. Manganese and its compounds: Environmental aspects. World

Health Organization, United Nations Environment Programme, International Lab.Liang, H. C.; Thomson, B. M. 2008. Minerals and Mine Drainage. Water Environ. Res. 80: 1481–1509.Liang, H. C.; Thomson, B. M. 2009. Minerals and Mine Drainage. Water Environ. Res. 81: 1615–1663.Pourbaix, M. 1966. Atlas of electrochemical equilibria in aqueous solutions, Translated from French by James

A. Franklin, Pergamon Press, Oxford, New York.U.S. Environmental Protection Agency (2001), Arsenic Rule 2001, 40 CFR 141, Washington, D.C.

199

Geochemistry


The impact of short-term variations of weather conditions on thechemism of rain water runoff from flotation wastes of MississippiValley-type Zn-Pb ores (southern Poland)

A. BauerekCentral Mining Institute, Katowice, Poland

ABSTRACT: The article presents the results of investigations of runoff water generated by heavystorms on a slope of tailings pond made from Zn-Pb flotation wastes of the Mississippi Valley-type ore formation (Silesia-Cracow ore district). For selected indicators of contaminations (SO2−

4 ,Ca2+, Zn, Cd, and Pb) leached from wastes through water runoff main statistical parameters arepresented. The impact of short-term variations of weather conditions on the chemistry of runoffwater is clarified. It has been pointed out that SO2−

4 concentrations in rain water runoff havehighly negative correlation with the sum of rainfalls and relative air humidity during days pre-ceding runoff occurrences. Among heavy metals only cadmium concentrations in waters point outan some dependence on weather conditions, occurring before runoff events. For sulfates, as themain contamination indicator, regression models (linear, polynomial and exponential) of depen-dences of their concentrations in waters on precipitation and air humidity were tested. Moreover, anon-standard technique of runoff water sampling and a system of information about approachingrainfalls, based on meteorological radar data, have been presented.

1 INTRODUCTION

According to estimations, zinc-lead (Zn-Pb) deposits of the Mississippi Valley-type (MVT) for-mation contain about 25% of world resources of metals mentioned above (Paradis et al. 2005).As a result of exploitation and processing of MVT zinc and lead ores, considerable quantities offine-grained wastes, rich in heavy metals (Zn, Pb, As, Cd), arise. Most wastes of this type are dis-posed in the central part of the United States, in deposit areas Tri-State and Old Lead Belt MiningDistrict (Missouri, Oklahoma, and Kansas). However, they are systematically managed. Accordingto assessments, in these regions from previously disposed 750 million tons of wastes only about200 million tons remained unmanaged (Kring et al. 2007). In Poland flotation wastes of Zn-Pbores occur in three regions of the Silesia-Cracow ore district (southern Poland): Bytom – about30 million tons, Olkusz – about 60 million tons and Chrzanow – about 33 million tons (Girczyset al. 2002). The thickest fractions of currently produced wastes are used for the reclamation ofabandoned open cast mines of Zn-Pb ores (Eckes et al. 1998).

Investigations carried out in Poland have pointed out that flotation wastes from MVT Zn-Pb oreprocessing play an essential role in the migration of heavy metals (HM) and sulfates (SO2−

4 ) intothe environment. The environmental impact of these wastes results mainly in soil contaminationwith fine-grained fractions rich in heavy metals spread by the wind (Krzaklewski et al. 1990,Cabała et al. 2006) and infiltration of leachates containing SO2−

4 into ground waters (Adamczyket al. 1994, Górecka et al. 1994). Girczys & Sobik-Szołtysek (1999) indicate that waters flowingdown from the embankment surfaces, having a short time contact with the buffering environmentof wastes rich in Ca and Mg carbonates, are able to leach contaminants, including heavy metals.

The initial results of investigations into contamination leaching through rain water runoff (RWR)from flotation wastes and mineralogical transformation accompanying the phenomenon were pre-sented by Bauerek et al. (2009). These investigations have pointed out that in the waters of thesurface runoff are dominated by sulfate (SO2−

4 ) and calcium (Ca2+) and accompanied by zinc (Zn)and cadmium (Cd). The results of later investigations (Bauerek et al. 2010) indicate the seasonal

203

Figure 1. Location of study area.

concentration variability of main contamination indicators in RWR and are suggesting a possiblesignificance of local meteorological conditions for the environmental impact.

In the United Kingdom the recognition of the phenomenon of contamination leaching in RWR,conducted in the 1990s, concerned environments with low carbonate content, generating acidmine drainage (AMD). The results from post-mining areas Wemyss in western Wales indicatethat leaching of heavy metals during torrential rainfall should be taken into consideration in theevaluation of contaminant transfer into the environment (Merrington et al. 1994). Episodic heavyrainfalls causing waste erosion, can release Pb, Zn, and Cd in hundreds of kilograms into theenvironment (Gao et al. 1995). The results of investigations carried out in the area of the abandonedin 1954 Zn-Pb Parc Mine (northern Wales) indicate that HM are released into the environment asa consequence of erosion and leaching of fine-grained processing wastes.

The investigations carried out in the United States indicate that intensive rainfalls followingdirectly after a dry period can generate acid runoff especially enriched in HM and SO2−

4 . In thezone of temperate climate essential significance for RWR enrichment in metals has the dissolutionof secondary sulfates containing HM precipitated as a result of evaporative crystallisation (Keithet al. 2001, Hammarstrom et al. 2005). One of few examples of occurrence of basic surface runoffwaters containing high concentrations of sulfates as well as Pb and Ni is the area of the TrojanNickel mine in Zimbabwe (Lupankwa et al. 2006). The investigations carried out there indicatethat even RWR with pH between 7.0 and 8.5 can contain considerable concentrations of sulfatesand metals leached from wastes.

The aim of this investigation was to link sulfate (SO2−4 ), calcium (Ca2+), zinc (Zn), lead (Pb), and

cadmium (Cd) concentrations in surface runoff waters from the tailings pond of flotation wastesof MVT ores to selected weather variables. In the analysis selected regression models to describethe variability of SO2−

4 concentrations in RWR were used. As independent variables the averagerelative air humidity and sum of rainfalls in days preceding runoff episodes were also determined.

2 SITE DESCRIPTION

The investigations at a site of flotation wastes coming from processing of Zn-Pb ore of MVTformation, were carried out. These wastes build the embankments of the tailings pond located nearOlkusz (southern Poland). The tailings pond, where wastes have been disposed since 1957, hasa surface area of 110 ha and height up to 42 m above the terrain’s surface (Fig. 1). Currently 60million tons of wastes are disposed there.

3 GEOLOGICAL SETTING OF DEPOSITS AND MINERAL COMPOSITION OF WASTES

The MVT deposits occurring in the area of the Silesia-Cracow ore district are of epigenetic origin.They were formed as a result of metasomatic transformation of carbonate formations of the middle

204

Figure 2. Average mineral composition of flotation wastes from tailings ponds in Olkusz (Cabała et al. 2006).

Triassic system under the influence of low-temperature (75–200◦C) hydrothermal solutions (Kon-stantynowicz 1979, Leach et al. 1993, Leach et al. 2001). Ore bodies, containing mainly simplemetal sulfides (sphalerite, galena, marcasite or pyrite) occur mostly in secondary dolomites as thefilling of karst voids, brecciated zones or as ore mineralization in bedrock. Shallow occurrence ofore bodies favours the development of weathering features in their roof zones (Cabała 2001).

In the mineral composition of flotation wastes carbonates (about 73%) represented by dolomite,ankerite, and calcite (Fig. 2) prevail. Sulfates and minerals Zn, Pb, and Fe oxides constitute 20%by weight of wastes (Górecka et al. 1994).

Among sulfide minerals iron sulfides (marcasite and pyrite) prevail; their share in wastes dis-posed in various time periods changes from 11 to 17% by weight. Zn and Pb sulfides and carbonatesconstitute 1.9% by weight and 0.8% by weight, respectively. Clayey minerals such as illite, mont-morillonite and kaolinite constitute on average about 7% by weight of wastes (Cabała et al. 2006).The total contents of Ag, As, Ba, Cd, Co, Cr, Cu, Ni, Sr and Tl in wastes do not exceed 0.1% byweight (Górecka et al. 1994, Cabała 2000).

4 MATERIALS AND METHODS

4.1 Sampling and laboratory tests

Water samples were collected during all runoff episodes that took place on the embankment of thetailings pond in 2008 and 2009. In the period from June to October 2008 five runoff events werenoted, in which total 24 samples were collected. The next 20 samples represent four episodes fromMay, June, and July 2009. For water sampling the surface area on the outside slope of the tailingspond was selected, with an inclination of 30◦, covered with fine-grained wastes. In order to samplethe waters of the surface runoff at the base of the slope, five 2-metre gutters were set up, creatingfive plots with a total surface area of 60 m2 (Fig. 1). The gutters were designed in such a way thatthe waters originating from the runoff should not mix with rain water (Fig. 3).

Water samples were collected to polyethylene bottles of 0.5 l. Measurements of pH and electricconductivity (EC) were performed in the field using an integrated meter WTW MultiLine P4. Ionchromatography (IC) tests (Dionex, ICS-2500) were used to determine SO2−

4 (uncertainty ±10%).Inductively coupled plasma atomic emission spectrometry (ICP-AES) (PerkinElmer, OptimaTM

3000 DV) was used to determine Ca, Cd, Pb, and Zn (uncertainty ±5%). The analyses wereperformed in the laboratories of the Department of Environmental Monitoring at the Central MiningInstitute in Katowice.

4.2 Meteorological data and information system about storms

The data concerning the relative air humidity and rainfall quantity were collected at an automaticmeteorological station belonging to the network of stations of the Institute of Meteorology andWater

205

Figure 3. Construction of gutters for sampling of rain water runoff.

Figure 4. Scheme of meteorological radar data obtaining system.

Management (IMWM). The station is located 6 km from the tailings pond. The measurements ofrain quantity were carried out using the rain-gauge Hellmann D200 with resolution 0.1 mm, whereasair humidity was measured by means of a hygrometer of Hassmann type with a resolution of 0.1%.

In cooperation with IMWM a system of heavy rains (Fig. 4) monitoring based on the data fromthe meteorological radar was created. The system activates itself automatically in the case of arainfall with intensity above 5 mm/h at the distance up to 10 km from the tailings pond (1). Theradar data are transmitted to the headquarters of IMWM in Warsaw and after processing they aredirected every 10 minutes as maps and data tables to the server of the Central Mining Institute(2). At the same time short message services (SMS), containing current data on rain intensity, aregenerated on mobile phone. The system provides the possibility to reach the testing site before therainfall appearance (3).


5.1 Chemistry of rain water runoff

The results of surface water runoff analyses from the embankments of the flotation tailings pondof MVT zinc and lead ores indicate an essential significance of 5 indicators. These are SO2−

4 andCa2+ as main ions, indicating the sulfate-calcium (SO4-Ca) hydrochemical type of water runoffaccording to the Monition classification (Macioszczyk 1987), and heavy metals Zn, Cd, and Pb.

206

Table 1. Statistical parameters of the selected indicators in RWR samples collected in 2008 and 2009.

Parametersmg/l pH SO2−

4 Ca2+ Zn Cd Pb

2008

Frequency 24 24 24 24 24 24Minimum 6.8 579.0 240.5 1.1 0.030 0.010Maximum 7.7 939.0 388.8 2.4 0.075 0.028Range 0.9 360.0 148.3 1.3 0.045 0.018Mean 7.3 756.3 318.2 1.6 0.051 0.016Median 7.3 736.0 315.6 1.4 0.052 0.016Standard deviation 0.2 84.5 37.1 0.4 0.012 0.004

2009

Frequency 20 20 20 20 20 20Minimum 7.1 179.0 80.6 0.6 0.020 0.013Maximum 8.0 917.0 382.8 2.7 0.068 0.038Range 0.9 738.0 302.2 2.1 0.048 0.025Mean 7.5 499.6 214.3 1.6 0.038 0.025Median 7.5 476.0 208.4 1.4 0.037 0.025Standard deviation 0.3 215.9 87.4 0.6 0.013 0.006

Probability p 0.79* 0.00006** 0.00003** 0.92*** 0.0007* 0.000008*for α = 0,05

The following statistical tests for the comparison of means were applied: ∗) – F tests, ∗∗) – Cox tests withseparate variance estimation, ∗∗∗) – U Mann-Whitney test.

The of statistical parameters of the indicators show that only pH and Zn concentrations in runoffwaters do not differ considerably during 2 years period of the study. However, the differencesbetween SO2−

4 , Ca2+, Cd and Pb concentrations, measured in 2008 and 2009, are statisticallysignificant (Tab. 1).

The ranges of results of SO2−4 and Ca2+ as well as HM concentrations obtained in individual

seasons, described by minimum and maximum values and range, are considerable. However, therelatively similar values of mean and median, for the results related to individual seasons suggestthat these sets are homogeneous (Tab. 1).

5.2 The impact of weather conditions on SO2−4 , Ca2+, Zn, Cd, and Pb concentrations in RWR

The weather parameters used to classify the variability of SO2−4 and Ca2+ concentrations in surface

runoff waters, tested in the summer months of 2008 and 2009, were the average relative air humidityand sum of rainfalls within the period of five days preceding the surface runoffs. The appliedweather-related data are presented in Table 2.

The investigation results indicate that SO2−4 and Ca2+ concentrations in RWR representing

episodes of surface runoff with numbers 1–6 and 9 (Fig. 5) are high (from 525 to 939 mg ofSO2−

4 /l and from 226 to 387 mg of Ca2+/l). These runoff events are generated by intensive storms,which followed after dry and hot periods lasting several days. The values of relative air humidity(from 41.6 to 65.8%) and sum of rainfalls (from 0.0 to 10.8 mm) in 5-day periods preceding runoffepisodes were relatively low (Tab. 2). Such conditions have favoured the evaporative crystallisationof easy leachable sulphate minerals, of which the greatest significance has gypsum (CaSO4 . 2H2O)or bassanite (CaSO4 . 0.5H2O) (Bauerek et al. 2009). Thus every intensive rainfall, causing runoff,resulted in dissolution of these salts and release of considerable amounts of SO2−

4 and Ca2+.The relatively low concentrations of main ions (from 179 to 492 mg of SO2−

4 /l and from 81 to208 mg of Ca2+/l in water from rainfalls with numbers 7 and 8, noted in 2009, are an example of theimpact of unstable weather and long wet periods on the reduction of contaminant concentrationsleached by RWR (Fig. 5). The relative air humidity reaching 69.4 and 84.0% and high sum of

207

Table 2. Meteorological data.

Number of runoff Average relative Sum of 5-dayepisode air humidity [%] precipitation [mm]

2008

1 52.0 0.02 62.0 10.83 47.4 2.64 41.6 0.05 65.8 2.3

2009

6 49.4 11.37 69.8 50.68 84.0 26.19 53.6 2.3

Figure 5. Distribution of SO2−4 and Ca2+ concentrations in waters and values of average relative air humidity

and sum of rainfalls.

rainfalls reaching 50.6 and 26.1 mm have not favoured sulfate minerals precipitation on the wastesurface. This caused a gradual depletion of the surface waste layer in easily leachable components.

Thus, the described dependence confirms the the supposition about an important role of localweather conditions (Bauerek et al. 2010). However, according to US Geological Survey (Sealet al. 2002), not only the temperately warm climate with considerable rainfall quantity is thecondition intensifying contamination leaching. The time-related rainfall distribution and atmo-spheric conditions occurring in periods separating heavy storms that cause surface runoffs are alsoimportant.

The distribution of Cd concentrations in RWR fits the best the data concerning main ions leaching.The lowest Cd concentrations were noted in runoff water originating from the episodes 7 and 8,which followed after wet days with high rainfalls (Fig. 6). However, the analysis of data in thediagram indicates that cadmium concentrations have also other minima, which are difficult tointerpret (runoff episodes No. 3 and 5).

The distribution of lead (Fig. 6) and zinc (Fig. 7) concentrations in runoff waters are even moredifficult to interpret, indicating the possibility of impact of other factors that were not taken intoconsideration. Though zinc concentrations, similarly as main ions and Cd concentrations, also reachthe minimum during the runoff No. 8 (Fig. 7). Lead concentrations, however, have a distribution

208

Figure 6. Distribution of Cd and Pb concentrations in waters and values of average relative air humidity andsum of rainfalls.

Figure 7. Distribution of Zn concentrations in waters and values of average relative air humidity and sum ofrainfalls.

approximately balanced during the study period and do not show dependences on meteorologicalconditions (Fig. 6).

5.3 Regression of SO2−4 concentrations in relation to air humidity and sum of rainfalls

Among the tested contamination indicators (SO2−4 , Ca2+, Zn, Cd and Pb), SO2−

4 concentrations inrunoff waters and their dependence on weather variables were selected for regression analysis. Theselection of SO2−

4 is based on two reasons:

• The SO2−4 ion concentrations are characterized by a strong, negative correlation with selected

meteorological factors:– −r = −0.73 for the dependence of SO2−

4 concentrations on the 5-day sum of rainfalls,– −r = −0.72 for the dependence of SO2−

4 concentrations on average air humidity from fivedays,

• The SO2−4 ions are the main indicator affecting water quality in the area of disposal of flotation

wastes from MVT ore processing (Adamczyk et al. 1994, Górecka et al. 1994, Bauerek et al.2010).

209

Table 3. The results of regression analysis of SO2−4 concentrations in relation to weather variables.

Index ofModels Equations adjustment

Relation: concentration of SO2−4 – sum of rainfall

Linear (A) SO2−4 = 754.3808 − 9.2109*rainfall R2 = 0.515

Polynomial (B) SO2−4 = 796.1808 − 20.1259*rainfall + 0.2253*rainfallˆ2 R2 = 0.567

Exponential (C) SO2−4 = 773.547*(exp(−0.0183* rainfall)) R2 = 0.560

Relation: concentration of SO2−4 – average relative air humidity

Linear (D) SO2−4 = 1322.4503−11.5981*hum. R2 = 0.511

Polynomial (E) SO2−4 = 308.1379+22.3791*hum. − 0.2719*hum.ˆ2 R2 = 0.550

Exponential (F) SO2−4 = 2481.4737*(exp(−0.0242*hum.)) R2 = 0.427

hum. – humidity

Figure 8. Diagram of dispersion of SO2−4 concentrations and rainfall parameters approximated by regression

lines.

The tested regression models (linear polynomial of second grade and exponential) explain from42.7 to 56.7% the variabilities of SO2−

4 concentrations in relation to air humidity and sum ofrainfalls (Tab. 3). The adjustment of lines, described by equations shown in Table 3, to basic dataare presented in dispersion diagrams (Fig. 8–9). The rightness of the models mentioned above wasconfirmed by the Shapiro-Wilk tests (S-W) on distribution normality of remainders and assessmentof variance homogeneity of the random component of remainders.

The remaining, not clarified part of SO2−4 concentration variabilities is connected to non-tested

factors, of which the highest significance may have the quantity of water running off from theembankment during heavy storms.

For practical application, linear models are proposed for estimation of SO2−4 concentrations on

the basis of the analysed weather variables. Their equations are the most simple and their indicesof adjustment to variables (R2) do not considerably depart from those counted for other modelswith better fits. Weather variables, i.e. the sum of rainfall and air humidity are characterisedby relatively high, mutual correlation (r = 0.65). Thus, the proposed regression models can be

210

Figure 9. Diagram of dispersion of SO2−4 concentrations and air humidity approximated by regression lines.

used interchangeably when estimating SO2−4 concentrations, dependent on the availability of data

concerning air humidity or rainfall quantity.

6 CONCLUSIONS

In runoff waters flowing from tailings pond embankments, comprising carbonate–sulfide flotationwastes, SO2−

4 and Ca2+ ions and heavy metals Zn, Cd, and Pb are dominant. The weather variables,such as the average relative air humidity and the sum of rainfalls for 5-days periods preceding runoffepisodes do not modify the qualitative composition of runoff waters. The weather conditions havean impact, however, on the main ion concentrations and, to a lower extent, on Cd concentrations,while Zn and Pb concentrations in waters remain independent of air humidity and sum of rainfalls.

The variability of SO2−4 concentrations, as the main contamination indicator in tested rain water

runoff, can be explained by short-term fluctuations of weather conditions. The performed regressionmodels (linear, polynomial and exponential) indicate that independent weather variables explainabout 50% of the variability in the SO2−

4 concentrations data.

ACKNOWLEDGEMENTS

The author is grateful to The Mining and Metallurgy Plant ZGH “Boleslaw” in Bukowno forrendering the testing ground accessible and for the permission to carry out research procedures.

The research work was financed from means for science within 2010-2011 and as a researchproject No 5682/B/T02/2010/38.

The anonymous reviewer is thanked for helpful suggestion and comments.

REFERENCES

Adamczyk, A. & Haładus, A. 1994. The influence of large sources of contamination on groundwater in theintensively drained basin (S part of GZWP 454 Olkusz-Zawiercie). In A. Kleczkowski (eds), Methodicalprinciples of groundwater protection. KBN Research Project No 9 0615 91 01: 133–153. Kraków: AkademiaGórniczo-Hutnicza (in Polish with English summary).

211

Bauerek, A., Cabała, J. & Smieja-Król, B. 2009. Mineralogical alterations of Zn–Pb flotation wastes of theMississippi Valley Type ores (Southern Poland) and their impact on contamination of rain water runoff.Polish Journal of Environmental Studies 18(5): 781–788.

Bauerek, A. & Łaczny, M. J. 2010. Contaminated water runoff from dam slopes of tailings pond of flotationwastes of Zn-Pb Mississippi Valley-type ores at Bolesław near Olkusz. Przeglad Geologiczny 58(1): 54–59(in Polish with English summary).

Cabała, J. 2000. Prospects for Zn-Pb ore mining in Poland with regard to ore quality in discovered deposits.In G. N. Panagniotou & T. N. Michalakopoulos (eds), Mine planning and equipment selection: 177–182.Rotterdam: Balkema.

Cabała, J. 2001. Development of oxidation in Zn-Pb deposits in Olkusz area. In A. Piestrzycski, et al. (Eds.),Mineral deposits at the beginning of the 21st century 121–124. Lisse: Balkema.

Cabała, J. & Teper, L. 2006. Metalliferous Constituents of Rhizosphere Soils Contaminated by Zn-Pb Miningin Southern Poland. Water Air Soil Pollut 178: 351–362.

Eckes, T., Gołda, T., Gruszczynski, S. & Trafas, M. 1998. The possibilities of utilization the post-flotationwastes from zinc and lead ores treatment for the reclamation of post-mining terrains. Archiwum OchronySrodowiska 24(2): 95–117 (in Polish with English summary).

Gao,Y. & Bradshaw, A.D. 1995. The containment of toxic wastes: II. Metal movement in leachate and drainageat Parc lead-zinc mine, north Wales. Environmental Pollution 90: 379–382.

Girczys, J. & Sobik-Szołtysek, J. 1999. Release and elimination of heavy metals in tailings pond filled ofblende flotation wastes. Fizykochemiczne Problemy Mineralurgii 33: 33–44 (in Polish).

Girczys, J. & Sobik-Szołtysek, J. 2002. The wastes of zinc-lead industry. Czestochowa: WydawnictwoPolitechniki Czestochowskiej (in Polish with English summary).

Górecka, E., Bellok, A., Socha, J., Wnuk, R. & Kibitlewski, S. 1994. Variation of metals contents in flotationwastes of Zn-Pb ores (ZGH Bolesław, Olkusz area). Przeglad Geologiczny 42: 834-841 (in Polish withEnglish summary).

Hammarstrom, J.M., Seal, R.R., Meier, A.L. & Kornfeld, J.M. 2005. Secondary sulfate minerals associatedwith acid drainage in the eastern US: recycling of metals and acidity in surficial environments. J. ChemicalGeology 215: 407–431.

Keith, D.C., Runnells, D.D., Esposito, K.J., Chermak, J.A., Levy, D.B., Hannula, S.R., Watts, M. & Hall, L.2001. Geochemical models of the impact of acidic groundwater and evaporative sulfate salts on BoulderCreek at Iron Mountain, California. Appl. Geochem. 16: 947–961.

Konstantynowicz, E. 1979. Geology of mineral resources, T.2 Deposits of metal ores. Katowice: 256–270:Uniwersytet Slaski (in Polish).

Kring, D. & Gene, G. 2007. Mine Waste. Fact sheet. U.S. Environmental Protection Agency (EPA),www.epa.gov/region7/factsheets/2007/fs_mine_waste0707.htm.

Krzaklewski, W. & Wójcik, J. 1990. An influence of industrial pollution of the atmospheric air on the selectedcomponents of forests in the Olkusz region. In M. Trafas & K.P. Zajac (eds), Zeszyty Naukowe AkademiiGórniczo-Hutniczej, Sozologia i Sozotechnika z.32: 201–216. Kraków: Wydawnictwo AGH (in Polish withEnglish summary).

Leach, D.L., Bradley, D., Lewchuk, M.T., Symons, D.T.A., De Marsily, G. & Brannon, J. 2001. MississippiValley-type lead-zinc deposits through geological time: implications from recent age-dating. MineraliumDeposita 36: 711–740.

Leach, D.L. & Sangster, D.F. 1993. Mississippi Valley-type lead-zinc deposits. In R.V. Kirkham, W.D. Sinclair,R.I. Thorpe & J.M. Duke (eds), Mineral Deposit Modeling: 289–314. Toronto: Geological Association ofCanada Sp. Paper.

Lupankwa, Keretia, Love, David, Mapani, Benjamin, Mseka, Stephen & Meck, Maideyi 2006. Influence ofthe Trojan Nickel Mine on surface water quality, Mazowe valley, Zimbabwe: Runoff chemistry and acidgeneration potential of waste rock. Physics and Chemistry of the Earth 31: 789–796.

Macioszczyk, A. 1987. Hydrogeochemistry. Warszawa: Wydawnictwa Geologiczne (in Polish).Merrington, G. & Alloway, B.J. 1994. The transfer and fate of Cd, Cu, Pb and Zn from two historic metaliferous

mine sites in the U.K. Applied Geochemistry 9: 677–687.Paradis, S., Dewing, K. & Hannigan, P. 2005. Mineral Deposits of Canada. Mississippi Valley-type Lead-Zinc

deposits (MVT). Natural Resources Canada.Seal, R.R. & Foley, N.K. (ed.) 2002. Progress on Geoenvironmental Models for Selected Mineral Deposits

TypesU.S. Geological Survey Open-File Raport 02-195.

212


The effect of weathering on the acid-producing potential of the GoathillNorth Rock Pile, Questa mine, NM

Virginia T. McLemore & Nelia DunbarNew Mexico Bureau of Geology and Mineral Resources, NM Inst. of Mining and Tech., Socorro, NM, USA

Samuel Tachie-MensonFreeport McMoRan, Morenci, AZ, USA

Kelly DonahueTelesto Solutions Inc., Fort Collins, CO, USA

ABSTRACT: The Goathill North (GHN) rock pile was constructed in 1964–74 and re-graded in2004–05. GHN samples were subjected to static, petrographic, and chemical tests to characterizeacidity and future acid-producing potential (AP). Samples that have higher concentrations of pyriteare more likely to have a higher AP. Static tests indicated that, although the rock-pile materialcontains acid consuming minerals, acid generation has occurred. Samples with rhyolite rock frag-ments have lower NP (neutralization potential) then samples with andesite rock fragments. Sampleswith rhyolite rock fragments had undergone hydrothermal quartz-sericite-pyrite (QSP) alteration,whereas samples with andesite rock fragments were propylitically altered; these hydrothermalalterations occurred after the molybdenum mineralization, but prior to mining. No single compo-nent controls the ABA (acid-base accounting) and NAG (net acid generation) tests. This lack ofcorrelation between ABA and NAG tests, mineralogy, and chemistry is a result of 1) weatheringreactions in the soil matrix producing precipitation of coatings surrounding the rock fragments andpreventing further weathering of the rock fragments, especially around pyrite crystals, 2) a lackof the water available that is required for weathering in the 25–40 years since the formation ofGHN rock pile, 3) non-uniform weathering of pyrite within the rock pile due to heterogeneous airflow, and 4) little or no weathering of the rock fragments. Also, the AP capacity of the rock-pile isdetermined by the combination of K-feldspar, calcite, smectite, illite, and pyrite, which are con-trolled in part by the lithology, pre-mining hydrothermal alteration, and post-mining weathering.The effects of pre-mining hydrothermal alteration and post-mining weathering both affect the statictests, emphasizing the need to perform detailed petrographic and mineralogic investigations alongwith the static tests to determine the AP of any mine waste material.

1 INTRODUCTION

1.1 Purpose

Rarely do waste-rock-pile characterization studies allow for detailed petrographic, mineralogical,and geochemical characterization of the undisturbed interior of large rock piles in situ. During theperiod of open-pit mining (1969–82) at the Questa molybdenum mine, NM (Chevron Mining Inc.,formerly Molycorp, Inc.), approximately 317.5 million metric tons of overburden rock was removedand deposited onto mountain slopes and into tributary valleys, forming 9 rock piles surrounding theQuesta open pit. After the rock piles were emplaced, a foundation failure occurred at the GoathillNorth (GHN) rock pile that resulted in sliding of the rock pile (Norwest Corporation, 2004). There-grading and stabilization of GHN rock pile provided an opportunity to examine, map, sample,and develop a conceptual model of the interior of a large rock pile in situ by means of trenches cut intothe rock pile. GHN has since been reclaimed and the gravitational sliding of the material has beenstopped. The purpose of this paper is to describe the effects of mineralogy, lithology, hydrothermal

213

Figure 1. Questa rock piles and other mine features, including location of trenches constructed in GHN.

alteration, and weathering of samples from throughout the rock pile on the static tests used todetermine the acid potential (AP) and neutralization potential (NP) of the rock-pile material fromGHN. Tachie-Menson (2006) reported preliminary results of the effects of mineralogy, lithology,hydrothermal alteration, and weathering of samples on AP and NP; this paper significantly updatesthat work.

1.2 Location of site

The Questa mine is on the western slope of the Taos Range of the Sangre de Cristo Mountains innorth-central NM (Fig. 1) and is on southward facing slopes at elevations of 2290 to 3280 m.

1.3 Definitions

Rock piles, the preferred term by many in the metal mining industry today, refer to the man-madestructures consisting of piles of non-ore material that had to be removed in order to extract ore. Thismaterial, referred to in older literature as mine waste, mine soils, overburden, subore, or proto-ore,does not include the tailings material, which consists of non-ore material remaining after milling.

Alteration is a term describing the changes in mineralogy, texture, and chemistry of a rockas a result of a change in the physical, thermal, and chemical environment in the presence ofwater, steam, or gas (Henley and Ellis, 1983; Reed, 1997; Neuendorf et al., 2005). Alterationincludes the effects produced by hypogene or hydrothermal (primary) and supergene (secondary)alteration and weathering. Hydrothermal alteration is the change in original composition of rockby hydrothermal (warm to hot) solutions during or after the mineralization. Hypogene alterationoccurred during the formation of the ore body by upwelling (ascending) hydrothermal or warm tohot fluids. Supergene alteration is the natural weathering of the ore body, at low temperatures at andnear the Earth’s surface by descending fluids. In this study, hydrothermal alteration refers to pre-mining processes. Weathering is the set of physical and chemical changes, including disintegration,of rock by physical, chemical, and/or biological processes occurring at or near the earth’s surface thatresult in reductions of grain size, changes in cohesion or cementation, and changes in mineralogicalcomposition (Neuendorf et al., 2005). In this study, weathering occurred after the material wasemplaced in the rock pile.

1.4 Acknowledgements

This project was funded by Chevron Mining Inc. (formerly Molycorp, Inc.) and the New MexicoBureau of Geology and Mineral Resources, a division of New Mexico Institute of Mining andTechnology (NMIMT). Special thanks go to the numerous students of NMIMT, who did much ofthe hard work from sampling to data compilation.

214

2 SITE DESCRIPTION

2.1 Climate

The climate at Questa is alpine and semi-arid, with cold snowy winters and moderate warm sum-mers with monsoons during July and August. The annual average temperature is 4◦C and theannual average precipitation and snowfall are approximately 50 and 371 cm, respectively. Dailytemperatures generally fluctuate by 18◦C throughout the year.

2.2 Geology

The geology and mining history of the area is complex and is described by others (Lipmanand Reed, 1989; Roberts et al., 1990; Meyer, 1991; Robertson GeoConsultants, Inc. 2000a, b;McLemore, 2009). Lithologies also are diverse, ranging from metamorphic to volcanic rocks,granites, shales, limestones, and sandstones (McLemore et al., 2009b). The Questa deposit is aClimax-type porphyry Mo (±W) deposit, which is a large, low-grade (0.1–0.2% Mo) deposit thatcontains disseminated and stockwork veinlets of Mo sulfides and is associated with Si- and F-richporphyritic granitic intrusions (Ludington et al., 2005). The Questa ore deposit contains quartz,molybdenite, pyrite, fluorite, calcite and other minerals. Climax-type deposits produce concen-tric zones of hydrothermal alteration (Ludington et al., 2005). There are seven major hypogenealteration types at Questa: 1) propylitic (chlorite, epidote, albite, calcite), 2) argillic and advancedargillic, 3) potassic (biotite, potassium feldspar, quartz, fluorite, molybdenite), 4) quartz-sericite-pyrite (QSP, also called phyllic, sericitic and silicic), 5) magnetite veining, 6) silicification, and7) post-mineral carbonate-fluorite veining, which are described by Meyer (1991), Ludington et al.(2005), among others. Supergene alteration is commonly superimposed on the hypogene alteration.Natural alteration scars occur in the Questa area that also are of hypogene and supergene origin.

The Questa rock piles were constructed using standard mining practices, primarily by haul-truckend-dumping in high, single lifts, which involved the dumping of rock over the edge of the hillslopes and resulting pile crests (Fig. 3; URS Corporation, 2003). Multiple areas of the open pit weremined at the same time. Records of the quantity, lithology, and rock-pile location of individualoverburden material were not maintained during construction of the rock piles. An estimate ofthe construction history of the rock piles was determined by examination of aerial photographs,as summarized by URS Corporation (2003). The upper portion of the rock pile tends to be moresoil-like (matrix-supported), whereas the lower portion tends to be rock-like (cobble-supported).The base of the rock pile is coarse rock and cobble supported, and is referred to as a boulder rubblezone. The resulting layers are locally at or near the angle of repose and subparallel to the originalslope angle. More details are in McLemore et al. (2009a).

The GHN rock pile contained approximately 4.2 million m3 (14.5 million metric tons) of over-burden material with slopes similar to the original topography, approximately at an angle of reposeof 38◦ (Fig. 2). GHN rock pile was approximately 192 m high and 61 m thick (URS Corporation,2003; Norwest Corporation, 2004), and was constructed during 1964–74 when material was enddumped in an alteration scar area, which is a natural, actively eroding landslide area caused byacidic weathering (Norwest Corporation, 2004). GHN rock pile is stratified consisting of locallyalternating layers and lenses of coarse- and fine-grained material that increase in grain-size downslope (Fig. 2). GHN was divided into 2 areas: a stable and an unstable area (Fig. 3). The unstableportion of the rock pile was the active land slide area, involving 1.9 million m3 of material (NorwestCorporation, 2004). Molycorp stabilized this rock pile by removing material off the top portion ofboth areas to the bottom of the pile (Norwest Corporation, 2004). This re-grading has decreasedthe slope, reduced the load, and created a buttress to prevent movement of the rock pile. This reportdescribes only the stable portion of GHN.

3 PROCEDURES

3.1 Mapping procedures

Remote sensing techniques, ground penetrating radar surveys (van Dam et al., 2005) and thermalcamera imaging surveys (Shannon et al., 2005) were used to select the location of trenches

215

Figure 2. Conceptual geological cross section of the stable portion GHN rock pile as interpreted from surfacemapping. Locations of trenches and drill holes are shown. Geologic units described in McLemore (2009a).

Figure 3. GHN before re-grading, looking east. Solid line indicates approximate location of trenches con-structed in summer-fall 2004 and the line of the cross section in Figure 2. Dashed line indicates the boundarybetween the stable and unstable portions of the rock pile.

within GHN during reclamation. Standard geologic mapping techniques were used (Lahee, 1961;McLemore et al., 2009a). Each unit on the surface and in the subsurface of GHN was examined andmapped, and were differentiated mostly on the basis of color, grain size, lithologic composition,texture, stratigraphic position, dip, thickness, and other soil properties (McLemore et al., 2009a).Longitudinal sections were made of each bench in the GHN rock pile and geologic maps weremade for each trench. McLemore et al. (2009a) described the mapping and field procedures andincludes specific locations of samples and construction data for each trench.

216

3.2 Sampling procedures

Sampling procedures, descriptions, and analytical methods typically used for soil profiles wereemployed, since the rock-pile material is similar to mine soils (URS Corporation, 2003; Smith andBeckie, 2003; Haering et al., 2004; Stormont and Farfan, 2005). Most samples were channel com-posites collected along approximate 1.5-m-long horizontal slots. Some samples were compositescollected along specific layers that were less than 1.5 m thick. Sample locations are in Figure 2 andMcLemore et al. (2009a). Only the upper third of the GHN rock pile was trenched, mapped, andsampled in detail using trenches. Data from 3 drill holes that were drilled into the rock pile (Fig. 2)also were used. Samples from the surface of the toe of GHN were used to define the toe region.These data were extrapolated for the entire rock pile.

3.3 Laboratory procedures

The laboratory analyses, summarized in Figure 4, were performed at NMIMT using standardlaboratory procedures (SOPs). Petrographic analyses (mineralogy, lithology, hydrothermal andweathering alteration) were performed using a binocular microscope and supplemented by thinsection petrography, microprobe, X-ray diffraction (XRD) analyses, and whole-rock chemicalanalyses. Clay mineralogy, in terms of the major clay mineral groups, was determined usingstandard clay separation techniques and XRD analyses of the clay mineral separates on orientedglass slides (Moore and Reynolds, 1989; Hall, 2004). This method does not liberate or measure theamount of clay minerals within the rock fragments, just within the soil matrix. The concentrationsof major and trace elements, except for S, SO4, LOI (loss on ignition), and F, were determinedby X-ray fluorescence spectroscopy at the New Mexico State University and Washington StateUniversity laboratories. F concentrations were determined by fusion and single-element electrodeand LOI concentrations were determined by gravimetric methods at NMIMT. S and SO4 weredetermined by ALS Chemex Laboratory. Acid base accounting tests were performed on selectedsamples at NMIMT (Tachie-Menson, 2006).

The acid-base accounting (ABA) test comprises two separate procedures: the acid potential (AP)test and the neutralization potential (NP) test. The AP test measures the potential for a sample togenerate acid, while the NP test determines the potential for a sample to neutralize acid. From thesetwo tests, values are obtained for NP and AP, both expressed in kg of CaCO3 per metric ton ofmaterial (i.e. parts per thousand). A net neutralization potential (NNP) is then calculated as NP–AP. The NNP also is referred to as the acid-base account of a sample. Factors that affect the NNPare the concentrations and types of acid-producing minerals such as pyrite, and acid-consumingminerals such as calcite. The genesis and development of these test procedures and their variationsare published by many researchers (Smith et al., 1974; Sobeck et al., 1978; Cruywagen et al., 2003;Tachie-Mensen, 2006). The hydrogen peroxide procedure was used for the AP test and the StandardSobek Method was used for the NP test.

3.4 Mineral abundances

Mineralogical data is obtained by different techniques, including: 1) petrographic analysis of abulk grab subsample using a binocular microscope, 2) petrographic analysis of thin sections of therock fragments using a petrographic microscope (including both transmitted- and reflected-lightmicroscopy), 3) electron microprobe analysis of both the fine-grained soil matrix and the rockfragments, 4) clay mineral determination of a bulk sample split using clay separation techniquesand XRD analysis (Moore and Reynolds, 1989; Hall, 2004), 5) Rietveld analysis of heavy mineralseparates (Oerter et al., 2007), and 6) other methods of determining mineralogy (spectral analysis,XRD, fizz test).

Petrographic analysis of rocks has been traditionally performed with optical microscopy usingthin sections and point counting to provide a modal mineralogy. However, this method typicallydoes not provide accurate and reproducible mineral proportions for sedimentary, volcanic, and soil-like material, such as that found in the Questa rock piles, because rock fragments, groundmass, andmatrix are typically identified as separate phases, not as specific minerals. In addition, groundmassphases in volcanic rocks and soils can be too fine-grained to identify individually by petrographicmicroscopes. Quantitative mineral abundance determination by whole rock XRD analysis also is

217

Figure 4. Flow chart showing characterization analyses of selected samples. Not all analyses are performedon every sample. Bucket, metal tin, and bags refers to size of sample collected. XRF = X-ray fluorescence anal-yses, XRD = X-ray diffraction analysis, ICP = Induced-coupled plasma spectrographic analysis, NAG = netacid producing tests, ABA = acid base accounting tests. Specific details of sample preparation are describedin the project reports and are available upon request.

difficult, because of factors such as mineral crystallinity, preferential orientation in the samplemount, differential absorption of X-rays, and overlapping peaks by different minerals affect thediffractogram patterns and makes their interpretation difficult. Some minerals, such as hematite andother iron oxides, have poor crystallinity and are not always easily detected from the backgroundby XRD. Another approach that can be used to determine quantitative mineralogy is to calculatea normative mineralogy from the whole-rock chemical composition. A normative mineralogy is aset of idealized minerals that are calculated from a whole-rock chemical analysis (Neuendorf et al.,2005), but not all calculated minerals in the normative approach are always actually present in thesample.

For this project, the mineral abundances were determined by the modified ModAn technique(Paktunc, 1998, 2001; McLemore et al., 2009c), which provides a quantitative bulk mineralogythat is consistent with the petrographic observations, identified minerals, electron microprobeanalysis, clay mineral analysis, and the whole-rock chemistry of the sample. Unlike most normativemineral analyses, all of the minerals calculated for the quantitative mineralogy are in the actualsample analysis using the modified ModAn technique. ModAn is a normative calculation thatestimates modes “. . . by applying Gaussian elimination and multiple linear regression techniques tosimultaneous mass balance equations” (Paktunc, 1998, 2001) and allows location-specific mineralcompositions to be used. Representative mineral compositions for minerals in the Questa sampleswere determined from electron microprobe analysis and used in ModAn for this study (McLemoreet al., 2009c).

3.5 Precision and accuracy of data

Precision and accuracy are measured differently for each field and laboratory analysis (i.e. para-meter), and are explained in the project reports and summarized by McLemore and Frey (2008).

218

Figure 5. Sample GHN-KMD-0048, backscattered electron image showing original igneous texture shownby presence of large, altered phenocrysts. Phenocrysts are patchily replaced by clay minerals (generally illite).Other secondary minerals include epidote and Fe oxide. These are hydrothermal minerals. This sample exhibitsno weathering textures.

The chemical analyses are accurate to within ±5%. The mineralogical analyses are estimated to beaccurate to within ±10% of the reported value (McLemore et al., 2009c).

4 DESCRIPTION OF GHN ROCK PILE

4.1 Composition of original material removed from the open pit (i.e. overburden)

Chemically, the volcanic rocks in the Questa-Red River area are calc-alkaline, metaluminous toperaluminous igneous rocks (McLemore, 2009). The GHN samples are a mixture of 2 or morebasic rock types that were hydrothermally altered before mining, typically rhyolite (Amalia Tuff)and andesite. The rhyolite has more quartz and little to no epidote and chlorite compared to theandesite. The rhyolite typically has higher SiO2, K2O, Rb, Nb, less TiO2, Al2O3, Fe2O3T, MgO,CaO, P2O5, and Sr than the andesite. Little, if any, unaltered rocks went into the Questa rockpiles, which resulted in large variations in mineralogical and chemical composition. Amphiboles,pyroxenes, and feldspars were replaced by biotite and albite during alteration. Biotite, hornblende,and pyroxenes were hydrothermally altered to chlorite, sericite, smectite, illite, and mixed layerclays (prophyllitic alteration). Feldspars show varying degrees of hydrothermal alteration to illite,kaolinite, smectite, quartz, mixed layer clays (QSP overprinting prophyllitic alteration). Many rockswere silicified. Pyrite occurs as fine disseminated crystals in the host-rock matrix and as stockworkveins up to 15 cm thick. Other sulfide minerals are rare. A typical texture of hydrothermal-alteredandesite is in Figure 5.

4.2 Geologic units of GHN rock pile

The geologic units mapped in GHN typically consist of numerous elongate to lobate, wedge-shapedlenses and layers of a few centimeters to a meter in thickness and were differentiated mostly basedon similar color, grain-size, lithologic composition, texture, stratigraphic position, and other soilproperties (Fig. 2; McLemore et al., 2009a). Individual layers pinched and swelled or gradedvertically down slope or laterally across the width into other lenses. Very few individual layers arecontinuous through the entire length of the rock pile, but many of the geologic units do appear

219

to be continuous until cut off by the coarse rubble zone forming the toe of the rock pile. Rockfragment lithology is generally consistent within mapped geologic units and correlates well withmineralogy and chemistry. The units in GHN are generally youngest to oldest on the basis ofstratigraphic position because the relative time of deposition of the units in GHN increases fromwest to east. Unit boundaries ranged from horizontal to vertical, but most dipped between 20◦ and40◦ westward to northwestward. A specific geologic unit probably represents a combination of 1)similar lithologic composition of overburden material mined from the upper portion of the open pitand dumped by individual truck loads, 2) aqueous movement in finer-grained material down slopeand vertically through the rock pile material by rain-fall events in between individual truck loads, 3)differences in hydrothermal alteration and 4) subsequent weathering of the rock-pile material. Notethat all rock piles are different in terms of their construction, composition of overburden materials(including hydrothermal alteration), and weathering, therefore this model only represents the stableportion of GHN rock pile and similar constructed rock piles.

The Questa rock-pile materials are a mixture of different lithologies and hydrothermal alterationmineral assemblages before being emplaced in the rock piles, therefore changes of mineralogy andchemistry between the outer, oxidized zone and the interior, unoxidized zones of the rock pile area result of differences due to pre-mining composition as well as post-mining chemical weathering.These differences can be difficult to distinguish and the changes due to hydrothermal alteration aremore pronounced than those due to weathering.

4.3 Weathering of the rock-pile material

Typical chemical weathering is based upon the acidity derived from the CO2 system, where thedissolution of feldspar to form clays is the most important chemical reaction (Drever, 1997; Priceand Velbel, 2003). However, in the Questa rock piles, unlike most natural residual soil weatheringprofiles, dissolution of pyrite produces H2SO4 as the dominant weathering acid, with subsequentdissolution of calcite, and to a lesser extent chlorite, illite, and other silicate minerals. Thesereactions result in 1) elevated dissolved solutes in water seeping from the rock piles and 2) theprecipitation of gypsum, jarosite, soluble efflorescent salts, and Fe oxide/hydroxide minerals.These reactions can occur within years to hundreds of years, until the source of S is consumed.Weathering or oxidation of pyrite and other sulfide minerals generally requires four components:water, S (sulfide), air (oxygen) and bacteria (McLemore, 2008) and the result is sulfuric acid, locallycalled acid drainage (AD), acid mine drainage (AMD), or acid rock drainage (ARD). The resultingsulfuric acid does not entirely escape the rock pile, but resides as pore fluids, which can dissolveminerals within and at the surface of the rock pile. Water and oxygen appear to be the rate limitingfactors in the oxidation of sulfide minerals, especially in arid and semi-arid environments (Leónet al., 2004). Recent experimental studies by Jerz and Rimstidt (2004) shows pyrite oxidizes fasterin moist air than under saturated conditions, thereby accelerating the weathering of the rock piles,at least locally. Specific factors that affect pyrite oxidation are oxygen concentration, temperature,pH, pyrite surface area, concentration of ferric iron (Fe+3), the presence of bacteria or other livingorganisms, and water. Both water and air flow through the Questa rock piles (McLemore et al.,2009a; Reiter, 2009).

It is difficult, but possible to distinguish between pre-mining hydrothermal alteration and post-mining weathering in the rock piles, because of both the fine-grained texture of the soil-like matrixmaterial and the extensive pre-mining hydrothermal alteration. Detailed field observations andpetrographic analysis (especially using electron microprobe analyses, McLemore et al. 2009d) areused to define the paragenesis (sequence of events). Some of the rock in the rock piles also hadbeen weathered in the natural supergene environment before open-pit mining began (Campbell andLueth, 2008). The field observations and petrographic analyses are important to understand in orderto properly evaluate the mineralogy and chemistry of the rock piles and the effect on predictionsof AP and NP. The evidence for weathering in the Questa rock piles includes (McLemore et al.,2009a, d):

• Change in color from darker brown and gray in less weathered samples (original color of igneousrocks) to yellow to white to light gray in the weathered samples. However, the fast-reacting (lessthan 40 yrs) weathering of pyrite produces precipitates of secondary reaction minerals on thesurface of existing rock fragments and within the soil matrix, which result in the yellow to orange

220

Figure 6. BSE images showing a) delicate gypsum blades with intergrown jarosite cement (bright phase) andb) feathery blades in adhered soil matrix. These are typical weathering textures. Note the lack of weatheringof the rock fragments.

color, reflecting the role of ferric iron in the mineral structures. The secondary precipitates formcoatings on exterior surfaces, rims, and fill macro- and micro-fractures. Thus, the macro-scaleimpression of weathering, caused by the discoloration from the observed secondary precipitatesis exaggerated (McLemore et al., 2009d).

• Thin yellow to orange, “burnt” layers within the interior of GHN, where water and/or air flowedand oxidized the rock-pile material.

• Paste pH, in general, is low in oxidized, weathered samples and paste pH is higher in lessweathered samples.

• Increase in abundance of jarosite, gypsum, Fe oxide minerals and soluble efflorescent salts(locally as cementing minerals) (Fig. 6), and low abundance to absence of calcite, pyrite, andepidote in weathered samples.

• Textures of gypsum crystals indicate in-situ formation with supporting evidence from S and Oisotopes (Campbell and Lueth, 2008).

• Tarnish or coatings of pyrite surfaces, as well as more complete oxidation or dissolution of pyritewithin weathered samples.

• Dissolution textures of minerals (skeletal, boxwork, honeycomb, increase in pore spaces, frac-tures, change in mineral shape, accordion-like structures, loss of interlocking textures, pits,etching) within weathered samples (McLemore et al., 2009d).

• Change in bulk texture of the rock-pile material as compared to the original mined material,including increase in soil:rock ratio, piping or stoping within the rock pile, and decrease in grainsize due to physical weathering.

• Chemical classification as potential acid-forming materials using ABA methods (Tachie-Menson, 2006).

• The chemical composition of waters from the Questa rock piles (i.e., seeps and runoff watersfrom the rock piles, chemistry of leachate waters obtained by laboratory leaching of rock pilesamples) imply that silicate dissolution is occurring within the rock piles in a similar manneras that suggested by surface and ground water documented in the alteration scars by Nordstromet al. (2005).

• The chemical analyses of these water samples are characterized by acidic, high sulfate, high TDS(total dissolved solids), and high metal concentrations. Sulfate is the predominant anion and Al,Mg, Fe, and Mn are the predominant cations. The chemical analyses of these waters reflect thedissolution of calcite, pyrite, gypsum, and soluble sulfate and hydroxide minerals. The high Mgand Al in the waters possibly are related to chlorite, epidote, and other clay minerals, and arelikely a result of incongruent dissolution (i.e. selective removal of cations) by the acidic waters.The high Mn is possibly related to Mn-bearing carbonate (calcite, dolomite, rhodochrosite),chlorite, epidote, and smectite. The high F is possibly related to dissolution of fluorite, illite,beryl, apatite, and smectite, which contain anomalous fluorine concentrations (McLemoreet al., 2009a). The high Fe and S are possibly related to the dissolution of pyrite and othersulfide minerals.

221

Figure 7. Net NP (neutralizing potential) versus paste pH for samples in GHN. Geologic units shown inFigure 2 and described in McLemore et al. (2009a).

Not all weathered samples exhibit all of these features. It is common for the fine-grained soilmatrix to be weathered, but most rock fragments within the rock-pile material exhibit little or noweathering, except on their surface (McLemore et al., 2009d). As the pyrite oxidizes, calcite dis-solves, and secondary gypsum and Fe oxides precipitate; the mineralogical changes can be observedmicroscopically and, locally, macroscopically. Extensive evidence of mineral transformations duespecifically to weathering in the rock pile is not observed (i.e. feldspar to kaolinite).

Four different zones of weathering can be distinguished at GHN (Fig. 2; McLemore et al.,2009a): 1) outer oxidized zone (includes the surface and geologic units C, I), 2) intermediate zone(includes unit J, N), 3) inner, less oxidized, weathered zone (includes units K-W), and 4) basaloxidized zone (includes geologic units R and rubble zone). In all benches and drill holes sampled,the interior, less oxidized units (east, units K-W, excluding unit N) of the piles are uniformly darkto light brown or gray with visible pyrite that are interbedded with occasional yellow to gray zonesof oxidation associated with little or no pyrite. The inner, less oxidized zone typically containsabundant calcite, chlorite, and clay minerals and accordingly, has high paste pH values and lowerAP than the outer units. The outer surface-atmosphere interface represents a zone with the mostactive geochemical processes noted in the rock piles. The outer, oxidized units consist of highlyleached and oxidized rock comprising mainly quartz and secondary iron sulfates, with smectite andmixed layer illite-smectite and some pyrite. This zone is characterized by low paste pH, low NP,and high AP. Extensive interchange of water and oxygen occurs in this zone, which enhances pyriteoxidation. Inside the leached zone (J) is a zone of clay accumulation. The clays are predominantlyillite and smectite with increasing chlorite toward the center of the pile. This unit is typically greento orange with moderate to low paste pH. Inward from the zone of clay accumulation is a zone ofsulfate mineral accumulation. Jarosite and gypsum become more abundant and the zone is typicallyorange. In between the outer, oxidized and interior, unoxidized zone is an intermediate zone (UnitN) of light to dark brown material that is well cemented by clay. It contains local zones of brightorange to yellow oxidized sandy clay. Clays are dominated by illite, smectite, and chlorite.

5 STATIC TESTS

ABA and NAG tests indicate that the GHN material ranges from potentially acid-generating to non-acid generating samples (Fig. 7). Generally, AP depends on the amount of pyrite and other sulfide

222

Table 1. Summary of ABA results for the stable portion of GHN rock pile (Tachie-Menson, 2006). PastepH2 = paste pH on powdered samples; NP = neutralization potential; AP = acid potential; NNP = NP – AP;no = number of samples.

Paste pH2 NNP (kg CaCO3/t)

Hole/Trench No Min. Max. Avg. Std dev Min. Max. Avg. Std dev.

LFG-005 3 3.98 7.75 5.37 2.07 −1.51 41.03 12.96 24.32LFG-006 30 3.14 8.08 5.58 1.74 −12.92 47.78 9.23 14.71LFG-007 3 5.47 7.81 6.47 1.21 0.32 13.33 6.55 6.52LFG-008 13 4.14 7.69 5.77 1.16 −10.66 28.44 4.38 13.25LFG-009 16 3.05 6.78 4.46 1.08 −12.81 51.74 2.82 14.02Surface 2 3.72 4.09 3.91 0.26 −5.59 −5.49 −5.53 0.05TH-GN-01 32 3.11 8.08 5.23 1.51 −18.8 29.15 3.48 10.64Overall 99 3.05 8.08 5.30 1.53 −18.8 51.74 5.43 13.25

Table 2. Summary of NAG test results for the stable portion of the GHN rock pile (Tachie-Menson, 2006).No. = number of samples.

NAG pH2 NAG4.5 (kg CaCO3/t)

Hole/Trench No. Min. Max. Avg. Std dev Min. Max. Avg. Std dev.

LFG-005 3 2.96 8.99 5.38 3.18 0.00 1.27 0.42 0.73LFG-006 29 2.42 9.29 6.32 2.06 0.00 29.74 2.02 7.14LFG-007 3 6.26 8.51 7.27 1.14 0.00 0.00 0.00 0.00LFG-008 22 2.43 8.62 5.10 1.96 0.00 14.77 1.58 3.70LFG-009 16 2.03 8.49 4.55 2.10 0.00 25.89 4.99 9.54Surface 2 2.84 3.00 2.92 0.11 1.66 3.98 2.82 1.64TH-GN-01 32 1.37 8.06 4.39 2.16 0.00 31.18 3.33 6.44Overall 107 1.37 9.29 5.16 2.22 0.00 31.18 2.68 6.55

minerals and NP depends upon the amount of calcite and other acid-neutralizing minerals. Samplesthat have higher concentrations of pyrite are more likely to have a higher acid generation capacity.Samples with rhyolite rock fragments in GHN have lower NP then samples with andesite rockfragments. Samples with rhyolite rock fragments were hydrothermally altered to QSP alteration,whereas samples with andesite rock fragments were propylitically altered; these hydrothermalalterations occurred after the molybdenum mineralization, but prior to mining. The samples fromthe interior of the GHN rock pile (units L, K, O, R, S, U, V, M, T) are less weathered since itsemplacement and have less acid-generating capacity. However, the majority of GHN samples donot show any strong correlation between paste pH, AP, NP, mineralogy, or chemistry (Tables 1, 2,3). No single component controls the ABA and NAG tests.

This lack of correlation between ABA and NAG tests, mineralogy, and chemistry may be con-trolled by several factors. First, as weathering progresses, we observe that reactions in the soilmatrix producing precipitation of coatings on mineral grains in rock fragments. This is particularlytrue for pyrite, which is coated by goethite. This coating process may slow the rate at which pyriteweathers and produces acid, although this process is not completely halted because some fullyaltered relict pyrite is observed (McLemore et al., 2009d). Second, the local climate in the Questaarea is relatively arid, and therefore much of the GHN rock pile has remained relatively dry inthe 25–40 years since its construction. Water is a key component of the acid-generating weather-ing process, and in this case we suggest that uniform weathering was hampered by the relativelysmall amount of available water. Third, the air flow through the GHN rock pile has been shownto be heterogeneous (Reiter, 2009), leading to a heterogeneous weathering process. Fourth, forsome of the reasons mentioned above, there has been little or no weathering of the interior of the

223

Table 3. Pierson correlation coefficients for various parameters. Note the lack of any significant correlations.

Paste pH NAG pH NAG value AP NP Net NP NPAP

SiO2 −0.40 −0.36 0.07 0.03 −0.34 −0.21 −0.24TiO2 0.45 0.45 −0.10 −0.07 0.37 0.24 0.38Al2O3 0.39 0.32 −3.51E-04 0.06 0.34 0.15 0.26Fe2O3T 0.31 0.33 −0.07 −0.07 0.25 0.15 0.21MnO 0.51 0.52 −0.37 −0.322 0.35 0.36 0.30MgO 0.56 0.61 −0.29 −0.25 0.59 0.46 0.59CaO 0.61 0.64 −0.32 −0.31 0.56 0.50 0.59Na2O 0.42 0.49 −0.36 −0.35 0.30 0.37 0.34K2O −0.37 −0.29 0.09 −0.009 −0.36 −0.31 −0.22P2O5 0.44 0.39 −0.11 −0.12 0.34 0.25 0.34S −0.22 −0.34 0.87 0.85 −0.16 −0.20 −0.24SO4 −0.23 −0.37 0.25 0.24 −0.28 −0.24 −0.30C 0.25 0.25 −0.14 −0.09 0.30 0.20 0.15LOI −0.14 −0.25 0.25 0.29 −0.13 −0.23 −0.25Ba 0.34 0.32 −0.06 −0.04 0.18 0.11 0.21Paste pH 1 0.71 −0.31 −0.27 0.56 0.49 0.47NAG pH 0.70 1 −0.48 −0.41 0.62 0.52 0.56NAG value −0.31 −0.48 1 0.74 −0.24 −0.31 −0.23AP −0.27 −0.41 0.74 1 −0.20 −0.37 −0.30NP 0.56 0.63 −0.24 −0.20 1 0.74 0.68Ne tNP 0.49 0.52 −0.31 −0.37 0.74 1 0.55NPAP 0.47 0.56 −0.23 −0.30 0.68 0.56 1quartz −0.52 −0.59 0.25 0.23 −0.43 −0.36 −0.41K-feldspar 0.10 0.18 −0.24 −0.34 0.05 0.20 0.20plagioclase 0.41 0.41 −0.28 −0.28 0.26 0.34 0.29epidote 0.57 0.51 −0.22 −0.26 0.43 0.38 0.49calcite 0.32 0.41 −0.18 −0.13 0.45 0.33 0.23pyrite −0.22 −0.38 0.78 0.80 −0.17 −0.44 −0.16Fe-Mn oxides −0.04 −0.09 0.09 −0.002 −0.10 −0.06 −0.12gypsum −0.04 −0.002 −0.09 −0.09 −0.12 −0.07 −0.15biotite −0.12 −0.14 0.81 0.18 −0.09 −0.12 −0.07fluorite −0.09 −0.16 −0.07 0.02 0.29 0.09 0.14apatite 0.23 0.41 −0.22 −0.23 0.38 0.34 0.340Total clay −0.16 −0.23 0.28 0.34 −0.10 −0.28 −0.23kaolinite 0.14 0.17 −0.13 −0.08 0.06 0.05 −2.73E-04chlorite 0.56 0.61 −0.29 −0.25 0.60 0.46 0.50illite −0.32 −0.38 0.34 0.39 −0.20 −0.34 −0.31smectite 0.24 0.36 −0.16 −0.15 0.14 0.15 0.15mixed-layered clays −0.14 −0.03 −0.06 −0.23 −0.13 −0.10 −0.15

rock fragments, which comprise much of the rock-pile material (Fig. 6). Also, the acid producingcapacity of the rock-pile is determined by the combination of K-feldspar, calcite, smectite, illite,and pyrite, which are controlled in part by the lithology, pre-mining hydrothermal alteration, andpost-mining weathering. The rock-pile material forming the interior of GHN has not experiencedsignificant weathering since emplacement except where water and air flowed (McLemore et al.,2009a, d). A similar lack of correlation between weathering and geotechnical properties (such asslake durability, point load indices, and shear strength) is observed at GHN (Gutierrez et al., 2008;McLemore et al., 2009a).

6 CONCLUSIONS

In the Questa rock-pile materials, dissolution of pyrite, calcite, and to a lesser extent some combina-tion of chlorite, illite, feldspars, smectite, and other silicate minerals are the predominant chemical

224

weathering reactions that results in 1) elevated dissolved solute concentrations in water seepingfrom the rock pile and 2) the precipitation of gypsum, jarosite, soluble efflorescent salts, and Feoxide/hydroxide minerals. The Questa rock-pile materials are a mixture of different lithologies andhydrothermal alteration mineral assemblages before being emplaced in the rock piles, thereforechanges of mineralogy and chemistry between the outer, oxidized zone and the interior, unoxidizedzones of the rock pile are a result of differences due to pre-mining composition as well as chem-ical weathering. No single component controls the ABA and NAG tests. This lack of correlationbetween ABA and NAG tests, mineralogy, and chemistry is a result of 1) weathering reactions inthe soil matrix producing precipitation of coatings surrounding the rock fragments and preventingfurther weathering of the rock fragments, especially around pyrite crystals, 2) a lack of the wateravailable that is required for weathering in the 25–40 years since the formation of GHN rock pile,3) non-uniform weathering of pyrite within the rock pile due to heterogeneous air flow, and 4)little or no weathering of the rock fragments. Also, the AP capacity of the rock-pile is determinedby the combination of K-feldspar, calcite, smectite, illite, and pyrite, which are controlled in partby the lithology, pre-mining hydrothermal alteration, and post-mining weathering. Samples thathave higher concentrations of pyrite are more likely to have a higher AP. The effects of pre-mininghydrothermal alteration and post-mining weathering both affect the static tests, emphasizing theneed to perform detailed petrographic and mineralogic investigations along with the static tests todetermine the AP of any mine waste material.

REFERENCES

Campbell, A.R, and Lueth, V.W., 2008, Isotopic and textural discrimination between hypogene, ancientsupergene, and modern sulfates at the Questa mine, NM: Applied Geochemistry, v. 23, p. 308–319.

Cruywagen, L.M., Usher, B.H., Hodgson, F.D.I., and de Necker, E., 2003, Towards a standardized static testingmethodology for opencast collieries in South Africa: 6th International Conference on Acid Rock Drainage:Cairns, Queensland, Australia, p. 203–210.

Drever, J.I., 1997, The Geochemistry of Natural Waters: Surface and Ground water Environments (3rd edition),Prentice Hall, New Jersey, 436 p.

Gutierrez, L.A.F., Viterbo, V.C., McLemore, V.T., and Aimone-Martin, C.T., 2008, Geotechnical and geome-chanical characterisation of the Goathill North Rock Pile at the Questa molybdenum mine, New Mexico,USA; in Fourie, A., ed., First International Seminar on the Management of Rock Dumps, Stockpiles andHeap Leach Pads: The Australian Centre for Geomechanics, University of Western Australia, p. 19–32.

Haering, K.C., Daniels, W.L., and Galbraith, J.M., 2004, Appalachian mine soil morphology and properties:effects of weathering and mining method: Soil Science Society America Journal, v. 68, p. 1315–1325.

Hall, J.S., 2004, New Mexico Bureau of Mines and Mineral Resource’s Clay Laboratory Manual: UnpublishedNew Mexico Bureau of Geology and Mineral Resources report.

Henley, R.W. and Ellis, A.J., 1983, Geothermal systems, ancient and modern: a geochemical review: EarthScience Reviews, v. 19, p. 1–50.

Jerz, J.K. and Rimstidt, J.D., 2004, Pyrite oxidation in humid air: Geochimica Cosmochimica Acta, v. 68, p.701–714.

Lahee, F.H., 1961, Field geology: McGraw-Hill Book Company, New York, 926 p.León, E.A., Rate, A.W., Hinz, C., and Campbell, G.D., 2004, Weathering of sulphide minerals at circum-

neutral-pH in semi-arid/arid environments: influence of water content: SuperSoil 2004, 3rd Australian NewZealand Soils Conference, University of Sydney, Australia, 7 p., http://www.regional.org.au/au/asssi/.

Lipman, P.W. and Reed, J.C., Jr., 1989, Geologic map of the Latir volcanic field and adjacent areas, northernNew Mexico: U.S. Geological Survey, Miscellaneous Investigations Map I-1907.

Ludington, S., Plumlee, G., Caine, J., Bove, D., Holloway, J., and Livo, E., 2005, Questa baseline andpre-mining ground-water quality Investigation, 10. Geologic influences on ground and surface waters inthe lower Red River watershed, New Mexico: U.S. Geological Survey, Scientific Investigations Report2004–5245, 46 p.

McLemore, V.T., ed., 2008, Management Technologies for Metal Mining Influenced Water, Volume 1: Basicsof Metal Mining Influenced Water: Society for Mining, Metallurgy, and Exploration, Inc., Littleton, CO,102 p.

McLemore, V.T., 2009, Geologic Setting and Mining History of the Questa mine, Taos County, NewMexico: New Mexico Bureau of Geology and Mineral Resources, Open-file Report 515, 29 p.,http://geoinfo.nmt.edu/publications/openfile/details.cfml?Volume=515

225

McLemore, V.T., Dickens, A., Boakye, K., Campbell, A., Donahue, K., Dunbar, N., Gutierrez, L.,Heizler, L., Lynn, R., Lueth, V., Osantowski, E., Phillips, E., Shannon, H., Smith, M., Tachie-Menson, S., van Dam, R., Viterbo, V.C., Walsh, P., and Wilson, G.W., 2009a, Characterization ofGoathill North Rock Pile: New Mexico Bureau of Geology and Mineral Resources, Open-file Report523, http://geoinfo.nmt.edu/publications/openfile/details.cfml?Volume=523

McLemore, V.T. and Frey, B.A., 2009, Appendix 8—Quality control and quality assurance report (TaskB1); in McLemore, V.T., Dickens, A., Boakye, K., Campbell, A., Donahue, K., Dunbar, N., Gutierrez,L., Heizler, L., Lynn, R., Lueth, V., Osantowski, E., Phillips, E., Shannon, H., Smith, M., Tachie-Menson, S., van Dam, R., Viterbo, V.C., Walsh, P., and Wilson, G.W., Characterization of GoathillNorth Rock Pile: New Mexico Bureau of Geology and Mineral Resources, Open-file Report 523,http://geoinfo.nmt.edu/publications/openfile/details.cfml?Volume=523.

McLemore, V., Heizler, L., Donahue, K., and Dunbar, N., 2009d, Characterization ofWeathering of Mine Rock Piles: Example from the Questa Mine, New Mexico, USA: Secur-ing the Future and 8th ICARD, June 23–26 2009, Skelleftea, Sweden, conference proceedings,10 p., http://www.proceedings-stfandicard-2009.com/ pdfer/Virginia_McLemore_P_T2_Characterization-of-weathering-of-mine-rock-piles-example-from-the-Questa-mine-New-Mexico-USA.pdf

McLemore, V.T., Sweeney, D., and Donahue, K., 2009b, Lithologic atlas: New Mexico Bureau of Geology andMineral Resources, Open-file Report 516, 73 p., http://geoinfo.nmt.edu/publications/openfile/details.cfml?Volume=516

McLemore, V., Sweeney, D., Dunbar, N., Heizler, L. and Phillips, E., 2009c, Determining quantitative min-eralogy using a combination of petrographic techniques, whole rock chemistry, and MODAN: Society ofMining, Metallurgy and Exploration Annual Convention, preprint 09–20, 19 p.

Meyer, J.W., 1991, Volcanic, plutonic, tectonic and hydrothermal history of the southern Questa Caldera, NewMexico: University Microfilms, Ph.D. dissertation, 348 p.

Moore, O.M. and Reynolds, R.O., Jr., 1989, X-ray diffraction and the identification and analyses of clayminerals: Oxford University Press, New York, 378 p.

Neuendorf, K.K.E., Mehl, J.P., Jr., and Jackson, J.A., 2005, Glossary of Geology: American GeologicalInstitute, 5th ed., Alexandria, Virginia, 779 p.

Nordstrom, D.K., McCleskey, R.B., Hunt, A.G., and Naus, C.A., 2005, Questa Baseline and Pre-MiningGround-Water Quality Investigation. 14. Interpretation of ground-water geochemistry in catchments otherthan the Straight Creek catchment, Red RiverValley, Taos County, New Mexico, 2002-2003: U.S. GeologicalSurvey, Scientific Investigations Report 2005-5050.

Norwest Corporation, 2004, Goathill North Slide Investigation, Evaluation and Mitigation Report: unpublishedreport to Molycorp Inc., v. 3, 99 p.

Oerter, E., Brimhall, G.H., Jr., Redmond, J., and Walker, B., 2007, A method for quantitative pyrite abun-dance in mine rock piles by powder X-ray diffraction and Rietveld method: Applied Geochemistry, v. 22,p. 2907–2925.

Paktunc, A.D., 1998, MODAN: An interactive computer program for estimating mineral quantities based onbulk composition: Computers and Geoscience, v. 24 (5), p. 425–431.

Paktunc, A.D., 2001, MODAN—A Computer program for estimating mineral quantities based on bulkcomposition: Windows version. Computers and Geosciences, v. 27, p. 883–886.

Price, J.R. and Velbel, M.A., 2003, Chemical weathering indices applied to weathering profiles developed onheterogeneous felsic metamorphic parent rocks: Chemical Geology, v. 102, no. 3–4, p. 397–416.

Reed, M.H., 1997, Hydrothermal alteration and its relationship to ore fluid composition; in Barnes,H.L., ed., Geochemistry of hydrothermal ore deposits: 3rd ed., John Wiley and Sons, New York,p. 303–365.

Reiter, M., 2009, Fluid flow estimates in molybdenum mine rock piles using borehole temperature logs:Environmental and Engineering Geoscience, v. 15, no. 3, p. 175–195.

Roberts, T.T., Parkison, G.A., and McLemore, V.T., 1990, Geology of the Red River district, Taos County,New Mexico: New Mexico Geological Society, Guidebook 41, p. 375–380.

Robertson GeoConsultants Inc, 2000a, Interim Background Characterization Study, Questa Mine, NewMexico: RGC Report 052008/6.

Robertson GeoConsultants, Inc., 2000b, Background study data report, Questa Mine, New Mexico: RobertsonGeoConsultants, Inc. Report 052008/12, January.

Shannon, H., Sigda, J., van Dam, R., Hendrickx, J., and McLemore, V.T., 2005, Thermal Camera Imagingof Rock Piles at the Questa Molybdenum Mine, Questa, New Mexico: National Meeting of the AmericanSociety of Mining and Reclamation, Breckenridge, Colo, June, CD-ROM.

Smith, L. and Beckie, R., 2003, Hydrologic and geochemical transport processes in mine waste rock:Mineralogical Association of Canada, Short Course Series, v. 21, p. 51–72.

226

Smith, R.M., Grube, W.E., Arkle, T.A., and Sobek, A.A., 1974, Mine Spoil Potentials for Soil and WaterQuality: U.S. Environmental Protection Agency, EPA-670/2-74-070.

Sobek, A.A., Schuller, W.A., Freeman, J.R., and Smith, R.M., 1978, Field and Laboratory Methods Applicableto Overburdens and Minesoils: US Environmental Protection Agency, EPA-600/2-78-054.

Stormont, J.C. and Farfan, E., 2005, Stability evaluation of a mine waste pile: Environmental and EngineeringGeosciences, v. XI, no. 1, February, p. 43–52.

Tachie-Menson, S., 2006, Characterization of the acid-Producing potential and investigation of its effecton weathering of the Goathill North Rock Pile at the Questa Molybdenum Mine, New Mexico: M.S.thesis, New Mexico Institute of Mining and Technology, Socorro, NM, 209 p., http://geoinfo.nmt.edu/staff/mclemore/Molycorppapers.html.

URS Corporation, 2003, Mine rock pile erosion and stability evaluations, Questa mine: Unpublished Reportto Molycorp, Inc., 4 volumes.

van Dam, R.L., Gutierrez, L.A., McLemore, V.T., Wilson, G.W., Hendrickx, J.M.H., and Walker, B.M., 2005,Near surface geophysics for the structural analysis of a mine rock pile, northern New Mexico: 2005 NationalMeeting of the American Society of Mining and Reclamation, Breckenridge, Colorado, June, CD-ROM.

227


Effect of reservoir pool changes on metals release frommining-contaminated sediment

Thomas MoyerBlack & Veatch, Federal Services Division, Arvada, CO, USA

Brian StriggowEPA Region 4 Science and Ecosystems Support Division, Athens, GA, USA

James EldridgeBlack & Veatch, Federal Services Division, Woodinville, WA, USA

Craig ZellerEPA Region 4, Division of Superfund, Atlanta, GA, USA

ABSTRACT: Eastern Parksville Reservoir in Tennessee contains a large sediment delta contam-inated with copper slag, sulfide minerals and other mine wastes. Repeated exposure of the deltaduring winter drawdowns has oxidized the upper few feet of the delta, which is flooded duringsummer months. Seasonal inundation creates pore water in the oxidized zone with low pH (<4)and high concentrations of metals and sulfate. Water in the reduced zone has near-neutral pH, lowconcentrations of most metals, and variable iron. Surface water collected within inches of the sedi-ment interface often exceeds chronic aquatic life criteria for copper and zinc, suggesting sedimentinfluence. Winter drawdown permits poor quality water to drain to the reservoir perhaps throughpreferential pathways. Investigation of the delta illustrates that understanding complex mechanismsof contaminant release and cycling from sediment is critical to assessing potential long-term riskssuch as those associated with reservoir management practices.

1 INTRODUCTION

The Ocoee River in southeasternTennessee historically received acid mine drainage, mining wastes,and soil eroded from the Copper Basin mining district. Much of this material was depositedapproximately 32 km downstream in eastern Parksville Reservoir where a sediment delta wasformed (Fig. 1). Studies of the reservoir, conducted as part of a remedial investigation (RI) of theOcoee River (Black & Veatch, 2008), revealed adverse chronic effects to the aquatic ecosystemcaused primarily by metals and pH. This prompted a more detailed study of the delta in order tobetter understand potential contaminant release mechanisms.

The upper Ocoee River drainage (known as the Toccoa River in Georgia) is underlain by Precam-brian metamorphic rocks, some of which contain disseminated iron sulfide minerals. The absenceof calcareous strata in this region creates surface water with low hardness (commonly <20 mg/L)and little alkalinity. The Copper Basin mining district occurs within the Great Smoky Group and ischaracterized by massive sulfide deposits containing pyrite, pyrrhotite, chalcopyrite, and sphalerite(Slater, 1982). Ore from surface and underground mines was beneficiated using flotation separa-tion, roasting, and smelting to produce iron calcine, copper metal, zinc concentrate, and sulfuricacid and other sulfur products (EPA, 2010). In the late 1800s, local hardwood was cut and used asfuel to roast ore in open sheds prior to smelting. The deforestation, combined with the effects ofsulfur dioxide gas partly to wholly devegetated a 130 square kilometer area, leading to substantialsoil erosion. Granulated copper slag, iron calcine, tailings, and other mining and processing wastesalso were eroded and conveyed to the Ocoee River. Beginning in the 1970s, revegetation effortsbegan to stem the widespread erosion, although tributary streams continued to convey water with

229

Figure 1. Location map showing the location of Parksville delta at the eastern end of Parksville Reservoir insoutheastern Tennessee.

low pH and high concentrations of aluminum, copper, iron, manganese, zinc, sulfate, and acidityto the river. In 2001, the mining district and Ocoee River downstream of the site were designated asa Superfund Alternative megasite (EPA, 2010). Actions taken since that time have addressed creekdischarges through selective removals and water treatment using lime neutralization.

Sediment and mine contaminants discharged from the creeks moved downstream through theriver, forming lateral bars and mid-stream islands before being deposited in the still water behindParksville (Ocoee No. 1) dam, which was built in 1911. Two additional dams (Ocoee No. 2 and No.3) were constructed in 1913 and 1942, respectively to provide water for power generation (both) andlimit downstream sediment migration (No. 3). Reservoir pool elevation within Parksville Reservoiris managed by the Tennessee Valley Authority (TVA) primarily for power generation. From Maythrough October, pool level is kept at an elevation that floods the delta with 0.6 m or more of water.In winter months, pool elevation is lowered by about 2.7 m, which exposes approximately 1.05 km2

of sediment.

2 STUDY APPROACH

Data collected from the delta as part of the Ocoee River RI included grab samples of oxidized andreduced sediment, sediment cores collected from a transect along the delta axis, shallow groundwater from temporary locations in the oxidized sediment and wells installed in reduced sediment,surface water from the reservoir, and field parameters collected from the water column at thedelta toe as pool level was lowered in the fall. To support a feasibility study, the RI data weresupplemented with measurements of ground water concentrations though a full reservoir cycle andbench tests to investigate the effect of amending sediment with lime.

2.1 Delta ground water investigation

Data on ground water concentrations were gathered from three sets of nested wells screened atdifferent depths in the sediment column, including both the oxidized and reduced sediment zones(Fig. 1). The wells were installed in February 2008 prior to the rise in reservoir level to summer pooland monitored until water level was lowered in December. Samples were collected approximatelymonthly and analyzed for selected metals, sulfate, acidity, and field parameters; ground water eleva-tion also was measured at each location. In addition, samples of surface water were collected at eachlocation and from the river upstream of the delta. The results of the well investigation were combinedwith results previously obtained during the RI to provide a more complete picture of the delta.

Wells were sampled using low-flow techniques with water purged until field parameters werestable. Water samples for metals were field-filtered through certified 0.45 µm filters, preservedto pH <2 using nitric acid, then placed on ice. Samples for sulfate and total acidity were not

230

filtered and were placed on ice immediately following collection. Field parameter measurementsof temperature, pH, specific conductance, dissolved oxygen, redox potential, and turbidity weremade at the time of sample collection using a flow-through cell and Eureka Manta multi-parametersondes that were calibrated daily.

Water samples were analyzed by GEL Laboratory in Charleston, SC following accepted proto-cols. Aluminum, calcium, cobalt, copper, iron, magnesium, manganese, and zinc were analyzedusing EPA Method 6020A (inductively coupled plasma-mass spectrometry; EPA, 1986). Total sul-fate was analyzed following EPA Method 300.0 (ion chromatography; Pfaff, 1993). Total aciditywas analyzed following Standard Method 2010B (titration; Eaton, et al., 2005).

2.2 Bench testing

Oxidized delta sediment was collected from the upstream end of the delta for use in bench testing.This location was chosen for its sandy composition to minimize potential problems with percolationof water through the material. The sediment was homogenized and aliquots withdrawn for use inthe bench tests. Test cells were constructed using 5 gallon (19 L) buckets fitted with pre-packedsample screens installed horizontally near the base of the cell to permit sampling of pore water. Thecells were filled with about 16 kg of delta sediment and saturated with deionized water; the cellswere maintained in a laboratory at EPA’s Science and Ecosystem Support Division. Test protocolsincluded static tests to mimic permanent inundation of the delta and dynamic tests to imitate thecurrent reservoir management practices. Sediment remained inundated in the static cells for the 8month test duration, whereas sediment in the dynamic cells was inundated for 80 days, drained andexposed to the atmosphere for 100 days, then flooded for an additional 66 days. Agricultural limeamendment obtained from local suppliers was incorporated into the upper portion of the sedimentcolumn in both static and dynamic tests.

Water collected from the test cells was analyzed following the protocols listed above. Two splitsof the sediment were analyzed by SVL, Inc. in Kellogg, ID using standard protocols. Aluminum,calcium, cobalt, copper, iron, magnesium, manganese, and zinc were analyzed using EPA Method6010C (inductively coupled plasma-emission spectrometry; EPA, 1986). Neutralizing potentialand sulfur forms (total, sulfide, sulfate, and non-extractable) were analyzed following Sobek, et al.(1978). Saturated pH was measured following McLean (1982).

3 PARKSVILLE DELTA SEDIMENT

Delta sediment was investigated through grab sampling, hand augering, and coring the upperlayers of the sediment column. Sediment is moderately to poorly sorted and varies from silty clayto coarse sand. Sand fractions are dominated by quartz and other metamorphic minerals, with minormetamorphic rock fragments, granulated copper slag, and trace amounts of sulfide minerals. Siltand clay is typically layered on intervals of less than a few millimeters and leaf litter and organicdetritus commonly defines bedding tops/bottoms throughout the sediment. Sediment layers at thesurface of the emergent delta are oxidized and have an orange-brown color distinct from the darkgray of underlying reduced material (Fig. 2). The change from oxidized to reduced sediment variesfrom sharp to gradational over approximately 15 cm.

A boat-mounted vibracore core rig was used to collect cores from the upper 3 m of the deltasediment; the six cores defined a transect along the strike of the delta axis (Fig. 1). Becauserecovered core length was typically 50 to 75 percent of the drilled depth, there is uncertainty in thedepositional history of the recovered material. Coring showed that the zone of oxidized sedimentthinned from nearly 0.6 m near the reservoir inlet to zero where sediment remains permanentlyinundated and that silty clay layers are interbedded with sand units across the delta. A well-sortedslag-bearing sand layer identified at 1 m below the surface near the inlet is thought to correlateto a normally graded sand layer identified in other cores. That this layer was found in coresseparated by about 1.5 km along strike illustrates that some sedimentary units are widely distributedacross the delta and represent significant depositional events. Such layers may form preferentialpathways for water flow through the delta. Testing of samples collected from the delta suggest thathydraulic conductivity varies widely (at least 3 orders of magnitude), with maximum measuredvalues exceeding 0.75 m/day.

231

Figure 2. Photo of Ocoee River bank cutting through Parksville delta, showing orange-brown oxidizedsediment overlying dark gray reduced sediment. Thickness of oxidized zone is about 20 cm.

Table 1. Median and range of concentrations in Parksville delta sediment.

Oxidized Reduced

Median Minimum Maximum n Median Minimum Maximum n

Aluminum 19,000 3400 40,800 25 21,000 8200 35,300 21Calcium 770 160 9000 25 2800 820 7600 21Copper 630 110 1600 25 800 290 2000 21Iron 56,000 20,000 92,900 25 59,000 31,000 98,400 21Lead 220 68 980 25 270 130 1080 21Magnesium 3400 730 4800 25 3600 2000 5900 21Manganese 630 68 3300 25 400 250 1200 21Zinc 850 230 2000 25 1300 680 2540 21Paste pH 4.75 3.69 6.34 21 6.16 3.84 7.00 11Oxidizable S 0.07 0.04 0.41 11 0.97 0.64 1.28 11Neutral. Pot. <0.3 <0.3 <0.3 11 16.6 2.17 20 11

n = number of samplesmetals in mg/kg; pH in standard units; oxidizable sulfur in %; neutralizing potential in tons CaCO3equiv/kiloton.

Sediment samples collected from across the exposed portion of the delta had variable con-centrations of metals (Table 1; see Black & Veatch, 2008 for additional details). Most exceededsite-specific and literature-based risk levels for copper, iron, lead, and zinc in sediment and a fewsamples exceeded toxicity reference values for cadmium (TRV; the concentration at which toxiceffects to receptors is expected). Samples also had low to moderate concentrations of total phos-phorous and organic carbon. Paste pH varied from 3.7 to 7.0; of the eleven samples having pH lessthan 5.0, ten were collected from the upper 15 cm of the soil column or from material describedas oxidized. Most samples had low concentrations of oxidizable sulfur (total sulfur minus sul-fate sulfur) and neutralizing potential (Table 1). Median metals concentrations in oxidized andreduced sediment samples were generally similar for the two groups; however, oxidized sedimenthad consistently lower zinc and paste pH. Moreover, oxidized sediment had little to no oxidizablesulfur or neutralizing potential, which contrasted with higher values for both in reduced sedimentsamples. These results suggest that oxidation of delta sediment used up most available neutralizingpotential, resulting in lower sediment pH.The oxidation also appears to have mobilized considerable

232

Table 2. Well construction data.

Screen Top Screen BottomLocation Well ID Sediment Screened (m bss) (m bss)

PR200 PR200R reduced 1.43 1.59PR200X2 oxidized 0.30 0.46PR200X1 oxidized 0.15 0.30

PR201 PR201R reduced 1.13 1.28PR201T transition 0.43 0.58PR201X oxidized 0.15 0.30

PR202 PR202R reduced 0.88 1.04PR202X2 oxidized 0.30 0.46PR202X1 oxidized 0.15 0.34

m bss = meters below sediment surface

zinc from the sediment, but had less effect on copper, lead, and most other metals. Grain size alsowas found to influence sediment chemistry. Analysis of sieved splits of a few sediment samplesshowed that silt and finer material had significantly higher concentrations of metals and oxidizablesulfur than sandy material.

4 PARKSVILLE DELTA GROUND WATER

Shallow ground water contained within the oxidized and reduced sediment of Parksville deltawas investigated by installing three clusters of shallow wells whose locations are illustrated inFigure 1. Wells in reduced sediment were screened just below the permanent water table defined bywinter low pool in the reservoir; wells in the oxidized zone were dry at the time of installation andcontained water only when reservoir water flooded their locations. Reservoir information obtainedfrom TVA showed that pool elevation began to rise from winter low pool around April 1, 2008and had reached full summer pool by the end of the month (at summer pool, the delta is covered by0.6 to 1.5 m of water). A sharp drawdown to winter pool occurred in early to mid-May (exposingthe delta) and water elevation did not regain typical summer pool until the last week of May.From June through November, pool elevation remained stable, with fluctuations less than 0.3 m.Winter drawdown began the first week of November and low pool was achieved by early December.Pressure transducers installed at each well location showed that ground water elevation at stationsPR200 and PR202 responded quickly to changes in lake elevation, whereas ground water elevationat PR201 responded more slowly, lagging by up to three weeks. The lag suggests local anisotropyin water flow through the delta.

Table 2 summarizes information for the wells at each cluster. Sand filter packs were installedaround each well screen and the screens were isolated with bentonite plugs above and below thesand packs. In addition to the wells shown in the table, each cluster contained a fourth well screenedbelow the permanent water table and fitted with a pressure transducer and data logger to recordground water level. Samples were collected from the wells on seven occasions. The initial sampleswere collected when reservoir pool had risen to about 70 percent of summer pool (as measured fromwinter low pool). The second set of samples was taken during the May drawdown when reservoirpool stood at about 25 percent of summer elevation. The subsequent three sets were collected atfull summer pool, while the remaining two sample sets were collected when reservoir pool hadfallen from summer elevation by approximately 10 percent and 75 percent. Analytical data for allsamples can be found in Black & Veatch (2008).

A composite sediment sample was collected from the upper 1.9 m of the delta in 2002 andsubmitted for leach testing using the State of Nevada Meteoric Water Mobility (MWMP) procedure(NDEP, 1996). The sample included reduced sediment with subordinate oxidized material. The testreleased numerous metals from the sediment in concentrations generally similar to those of groundwater collected from reduced sediment near the base of the sampled interval. This indicated thatthe composition of ground water in the delta sediment could be expected to be a function of thematerials in which it occurs.

233

Table 3. Median and range of concentrations of ground water in Parksville delta sediment.

Water in Oxidized Sediment Water in Reduced Sediment

Median Minimum Maximum n Median Minimum Maximum n

Aluminum 3610 26.2 16,800 33 9.7 <5 228 21Calcium 11,700 7490 27,400 33 56,200 10,800 161,000 21Copper 2560 49.8 11,500 33 1.7 0.27 14.5 21Iron 6290 <50 66,000 33 21,100 7540 829,000 21Magnesium 5110 2260 9530 33 12,400 10,200 79,200 21Manganese 878 62.4 3040 33 586 157 14,000 21Zinc 9280 3000 24,100 33 6.8 <2.6 659 21Sulfate 237 101 665 33 164 <5 1990 21pH 4.6 3.5 7.1 33 6.8 6.0 7.8 21ORP 284 68 380 33 −54 −153 78 21Conductivity 508 281 700 33 552 261 2908 21

n = number of samplesmetals in µg/L; sulfate in mg/L, pH in standard units; ORP in mV; conductivity in µS/cm.

Figure 3. Ground water contained within Parksville delta sediment. Filled symbols collected from oxi-dized sediment; open symbols collected from reduced sediment. A. Dissolved zinc vs. dissolved copper. B.Oxidation-reduction potential vs. pH.

4.1 Water in the oxidized and reduced sediment zones

Table 3 summarizes concentrations found in ground water collected from oxidized and reduceddelta sediment during the 2008 study. The results show that water sampled from oxidized sedimenthas higher concentrations of numerous dissolved (0.45 µm filtered) metals including aluminum,cobalt, copper, manganese, and zinc. For most of these metals, median concentrations are two ormore orders of magnitude higher than the median concentrations in samples collected from reducedsediment (Fig. 3A). Reduced water had median concentrations of calcium, iron, and magnesiumhigher than those of oxidized water by factors of 5 or less. Water contained within oxidized sedimentalso had higher acidity and sulfate, lower median pH (by 2 log units), and positive redox potential(Fig. 3B). Importantly, pH and the concentrations of metals in most samples of ground watercollected from oxidized sediment exceeded applicableTennessee State chronic surface water qualitycriteria for the protection of aquatic life in Parksville Reservoir (Cu = 2.0 µg/L; Zn = 26 µg/L; pH6.0 to 9.0) and the recommended Federal criterion for aluminum in surface water (87 µg/L). Incontrast, only one sample of water from the reduced zone exceeded the surface water criterion foraluminum, while ten samples exceeded the criterion for copper (7 of these from location PR201),and seven samples exceeded the criterion for zinc (all from location PR201).

4.2 Local variation in ground water concentrations

Abrupt changes with depth were observed in the composition of ground water within the delta ateach of the three well clusters. The magnitude of these changes varied significantly from one well

234

Figure 4. Dissolved metal concentrations in interstitial pore water from the Parksville delta at location PR201.A. Dissolved iron vs. dissolved aluminum in ground water collected from oxidized, transitional, and reducedsediment and surface water from the same location. B. Dissolved zinc vs. dissolved copper.

cluster to another. Ground water contained in oxidized and reduced sediment layers at locationPR201 provides an example of the large variations that are locally observed, in this case within adepth range of about 1 m. Cluster PR201 contained a well screened fully within oxidized sediment,a well screened fully within reduced sediment, and a well screened across the sharp transitionbetween the two. These wells were screened in relatively fine-grained sediment; the oxidized sed-iment well was dry during the May and December sampling events and the transitional sedimentwell was dry in May when the reservoir level was down. Although aluminum, cobalt, and copperoccurred in low concentrations in the reduced sediment well, the reduced water contained elevatedconcentrations of dissolved iron, manganese, zinc, sulfate, and hardness that contrasted with waterin reduced sediment in other locations.

Figure 4 illustrates the wide variation in ground water concentrations present at location PR201and shows that water from the well screened across the oxidation boundary has a compositionintermediate to the oxidized and reduced endmembers. Moreover, the concentrations of metals insurface water collected a few cm above the sediment surface at this location show that the compo-sitional differences in ground water cannot be explained simply as a function of dilution by surfacewater (Fig. 4A).

4.3 Changes in ground water composition over a reservoir cycle

Ground water composition was observed to change throughout the course of the reservoir cycle.Figure 5 provides an illustration of the types of changes observed at well cluster PR200. Obser-vations of water extracted from reduced sediment at the three well clusters showed that dissolvedaluminum concentrations more than doubled from April through September during rising and fullpool conditions, then dropped by nearly as much as pool level fell to winter conditions. Dissolvedzinc initially dropped in two of the three reduced sediment wells then began to rebound as the waterlevel was lowered. Lowering of the reservoir pool also caused redox potential, which remainedrelatively constant at high pool, to increase in these samples, in some cases to positive values.Other constituents, such as pH, iron, and copper, displayed inconsistent behavior, increasing insome wells and decreasing in others.

The concentrations of copper and zinc decreased by about half in samples collected from oxidizedsediment during rising and high reservoir pool, then increased by a lesser amount as reservoir poolwas lowered. In contrast, iron concentrations increased throughout the reservoir cycle, with thegreatest increase occurring during rising and high pool. Field pH generally decreased throughoutthe cycle whereas redox potential increased during rising and full pool until remaining essentiallyunchanged as water level declined in the fall.

4.4 Field measurements of reservoir water during drawdown

An effort was made to investigate potential water quality impacts to the reservoir caused by therelease of water from the delta as the reservoir level was drawn down to winter pool in 2005. Three

235

Figure 5. Dissolved metal concentrations in ground water from the Parksville delta at location PR200. A.Ground water in reduced sediment at 1.43 to 1.59 m below the top of sediment (note log scale for concentration).B. Ground water in oxidized sediment at 0.30 to 0.46 m below top of sediment.

Figure 6. Profiles of specific conductance in Parksville reservoir with time as reservoir level was drawndownin the fall of 2005. Delta toe station located 30 to 60 cm above delta sediment at mid-point of toe; reservoirstation suspended adjacent to toe, 3 m above sediment. A. Mid-November, 2005 shows abrupt increase inspecific conductivity near the end of the monitoring period at the delta toe station. B. Early December, 2005,shows continued higher conductivity with delta toe water significantly more elevated than the reservoir as awhole. This departure is consistent with high conductivity water draining from the delta toe and graduallymixing into the reservoir as drawdown continues.

stations were located in the reservoir along the western toe of the delta and sondes were suspendedin the water column at these stations for multiple days on three occasions. Each sonde was equippedwith probes to measure temperature, pH, specific conductance, and dissolved oxygen. Two stationswere established at the mid-point and bottom of the delta toe within 30 to 60 cm of the sediment–water interface, while the third station was located within the water column approximately 3 mabove the sediment. The stations were occupied for several days in mid-October 2005 prior todrawdown, in mid-November as the reservoir level was dropping, and in early December as thereservoir was nearing low pool.

Figure 6 shows the records for the stations located at the mid-point of the delta toe and at 3 m abovethe sediment for the November and December events. The conductivity data showed that all stationshad similar values prior to and during the early stages of drawdown. As drawdown progressed,however, higher conductivity was noted in the delta toe station (Fig. 6A) and this eventually wasmatched by higher conductivity higher in the water column of the reservoir (Fig. 6B). The increasein conductivity is consistent with high conductivity water draining from the delta into the reservoir.

5 BENCH TESTING OF PARKSVILLE SEDIMENT

Sediment used in the bench test was collected adjacent to well cluster PR202 (Fig. 1) and con-sisted of medium to coarse, orange-brown micaceous sand with small gravel-sized fragments of

236

metamorphic rocks. The sand was quartz-rich, with the quartz typically stained yellow and itincluded 1 to 3 percent black, glassy granulated slag by visual estimate. Biotite mica was oxidizedand appeared brassy; sulfide minerals were not observed with a hand lens. The sediment wasthoroughly homogenized in a single pile from which aliquots were taken for testing.

Analysis of the delta sediment used in the bench tests found that the test sample had metalconcentrations lower than the median oxidized sediment shown in Table 1. Most significant werealuminum and copper which were lower by about 70 percent, iron which was lower by 37 percent,and zinc which was lower by 22 percent. In addition, the test sample had paste pH that was higher by0.6 log units, oxidizable sulfur that was lower by 43 percent, and measurable neutralizing potential(3.5 tons/kiloton CaCO3 equivalent).

5.1 Static tests

Five cells were used to conduct the static tests. These tests were intended to mimic long-terminundation of the delta sediment that occurs as reservoir elevation is maintained at summer pool. Thetests used about 16 kg of delta sediment saturated with 9 to 10.5 L of deionized water. Agriculturallime from two different suppliers was incorporated into the upper 3 to 4 cm of the sediment in four ofthe cells; the fifth cell was unamended to serve as a control. Water fully saturated the sediment anda pool of water 5 to 7 cm deep was maintained above it. Lids were placed on the cells to minimizeevaporative loss and the dissolved oxygen content of the surface pool was monitored weekly toensure that reducing conditions did not ensue. Three cells received 40 g of lime whereas the fourthcell received 80 g of lime. These amounts mimicked application rates of 3 and 6 tons/acre, whichis typical for amending acidic soils in the mining district.

The static tests began operating on May 15, 2008 and continued through January 15, 2009. Sixpore water samples were collected from each cell during that interval. The cell amended with 6tons/acre was inadvertently drained immediately after completion; the amount of water flushedthrough the cell is uncertain but less than the 9.5 L of water that were used to initially saturate it.The cell was refilled with water; no sediment or amendment was added.

Figure 7 illustrates that water produced in the unamended control cell was generally similarin composition to actual ground water collected from the delta, with the exception that the testcell water contained higher concentrations of cobalt and manganese; the cause of these elevatedconcentrations is unknown. Similar to ground water in the oxidized delta sediment, water in theunamended test cell exhibited trends of decreasing concentrations with time for most metals exceptiron, which progressively increased by two orders of magnitude.

Significant changes in pore water quality were observed in all cells over the course of thestatic test. Aluminum, calcium, copper, and zinc decreased with time while iron increased. Forsome metals, changes were substantial: copper in the control cell decreased by a factor of 12and iron increased by a factor of 124 from the first to the last sampling event. Redox potentialalso decreased throughout the test and pH typically increased by less than 1 unit. These changescomplicate interpretations of the effect, if any, of lime amendment on pore water quality. Althoughmost metals had similar concentrations in the control and amended cells, calcium in the amendedcells was nearly double that of the control cell. Moreover, relative to the control cell, amendedcells exhibited greater decreases in aluminum, copper, and zinc over time and a less substantialincrease in iron. The differences were greatest for the cell receiving the most lime. This suggeststhat amending sediment with lime could potentially produce positive effects on pore water quality.

5.2 Dynamic tests

Three cells were used to conduct the dynamic tests. These tests were intended to mimic the periodicflushing of reservoir sediment that occurs annually when pool level is drawn down from summerto winter pool. The dynamic cells were set up similarly to those of the static tests, using the sameamounts of sediment, lime amendment, and reagent-grade water. One dynamic cell was unamendedto serve as an experimental control; the other two cells were amended with two different types ofagricultural lime at an application rate of 3 tons/acre.

The cells were subjected to two flushing events. During the first event, the cells were filled withwater on May 15, 2008 and a pool of water a few cm deep was maintained above the sediment.Samples of pore water and the surface water pool were collected after 80 days after which the cells

237

Figure 7. Comparison of water produced in the unamended control test cell to ground water collected fromthe deep oxidized well at location PR202 on the delta. Concentrations for dissolved metals are in µg/L; thosefor hardness, acidity, and sulfate in mg/L.

were permitted to drain. The cells were then rested and the sediment allowed to dry and oxidizefor approximately 100 days. At that point (November 10), the cells were refilled with water anda surface pool was again maintained atop the sediment. Samples of pore water and surface waterwere collected after 66 days at which point the cells were drained and the experiment terminated.

Results from the unamended control cell showed that constituent concentrations in the pore waterand surface pool were similar for the first and second tests, with concentrations slightly lower in thefirst flush test samples. During each test, concentrations were typically higher in pore water than insurface water and constituents that were elevated in pore water also were found to be elevated in thesurface water sample. The surface water had significant concentrations of many constituents (e.g.,Cu 534 and 863 µg/L; Zn 4060 and 5860 µg/L in the first and second tests) and acidic pH (4.5and 4.1 s.u). Since the surface pool was not disturbed after the cell was filled, constituents clearlyhad diffused out of the sediment and into the deionized water used to fill the cells, illustrating thatdelta sediment can negatively affect surface water quality.

Cells that were amended with lime obtained from two different sources responded similarly toone another. As observed in the control cell, pore water and surface water had similar concentrationsin the two flushing tests, with the exception of pH, which dropped by up to log unit in surface waterin the second test. In these tests, however, there was a marked difference between surface water andpore water composition during each flushing event, with pore water having slightly lower calcium,magnesium, and hardness, and concentrations of cobalt, copper, manganese, and zinc that were upto 2 orders of magnitude higher.

Figure 8 compares constituent concentrations in pore water and surface water generated duringthe second flushing event in the unamended test cell and in a test cell amended with lime. Porewater in the amended cell had pH and constituent concentrations that were similar to or slightlylower than those in the unamended cell (Fig. 8A). In contrast, surface water in the amended cell hadsignificantly higher pH, slightly higher concentrations of calcium, magnesium, iron, and hardness,and substantially lower concentrations of aluminum, cobalt, copper, manganese, and zinc (Fig.8B). The dynamic tests clearly indicate that amending the sediment with lime can limit diffusiveexchange between surface water and sediment pore water.

6 DISCUSSION

The assessment of ecological risk to the reservoir conducted as part of the Ocoee River RI (Black &Veatch, 2008) found that water quality in the reservoir was generally good. Exceptions includedwater adjacent to the sediment delta collected during reservoir drawdown and during winter monthsand deep water (hypolimnion) in the reservoir collected during the summer. These areas suffered

238

Figure 8. Comparison of concentrations in pore water and surface water in amended and unamended dynamictest cells. Concentrations for hardness, acidity, and sulfate in mg/L; all others in µg/L. A. Pore water from thesecond flush test. B. Surface water from the second flush test.

chronic risk, primarily from zinc. Data collected as part of the RI pointed to the sediment delta atthe eastern end of the reservoir as a primary source of the chronic effects and this was substantiatedduring the more detailed study of the delta described herein.

For 4 to 5 months each winter, drawdown of the Parksville Reservoir exposes about 1 km2 ofa sediment delta contaminated with slag and other mining-related wastes eroded from the CopperBasin mining district. As a result, the upper surface of the delta, extending locally to a depth ofabout 0.6 m has become oxidized. Delta sediment contains concentrations of copper, iron, lead,and zinc that exceed site-specific and literature-based risk levels, indicating that toxic effects tobenthic macroinvertebrates are likely. Sediment analysis showed that the oxidation process createdsediment with saturated pH values less than 5 and low concentrations of oxidizable sulfur andneutralizing potential that contrast with reduced sediment (saturated pH of about 6; 1 percentoxidizable sulfur; 16 tons/kiloton CaCO3 equivalent neutralizing potential). Moreover, oxidizedsediment was found to have lower concentrations of a few metals, most notably zinc.

Samples collected from wells installed in the delta showed that despite the typically low concen-trations of oxidizable sulfur, oxidized sediment produced acidic ground water with concentrationsof aluminum, cobalt, copper, manganese, and zinc up to two orders of magnitude higher than thatof reduced sediment. At any given location, the composition of ground water within the delta wasfound to vary with depth and as a function of sediment oxidation. Dilution by the overlying surfacewater column does not appear to play a significant role in controlling water composition within thedelta sediment although it is expected to be the source of water in the oxidized sediment zone.

The sediment and ground water chemistry is consistent with the oxidation of iron sulfide mineralsto produce acidity and release metals (Stumm and Morgan, 1996). Over time, neutralization ofthis acidity is thought to have used up available neutralizing potential within the oxidized zoneultimately creating sediment with acidic pH. Continued exposure to oxidizing conditions may becreating water soluble acid salts as intermediate products that remain in the sediment pore spaceuntil dissolved into pore water as reservoir pool is raised each year.

Historic aerial photos (EPA, 2000) show that the majority of the delta was formed prior to1942 when Ocoee No. 3 dam was completed. Coring of the upper 3 m of the delta found thatthe delta consists of interbedded sandy, silty, and clayey deposits, some of which are widespread.The hydraulic conductivities of these materials were found to vary over at least three orders ofmagnitude. Depending on the stratigraphic architecture created during deposition, preferentialpathways for ground water flow are expected to be present within the delta and may take the formof widespread layers, channel-fills, or both.

Reservoir drawdown permits water in the oxidized sediment layer to drain from the delta aswater level is lowered below the top of the sediment surface. The pathways taken by this water havenot been documented but are likely to include flow into the river channel that bisects the delta atlow pool and flow through preferential pathways that lead to the delta toe. Evidence for the latterwas provided by sondes suspended above the delta toe that recorded increasing water conductivityimmediately above the delta toe as drawdown proceeded.

239

The Parksville delta does not support a viable benthic community, a fish spawning area, or ajuvenile rearing area when inundated with water. It also does not support the germination, estab-lishment, and growth of vegetation (either aquatic or terrestrial) due to the water level fluctuationthroughout the year. Bench testing of oxidized sediment demonstrated that inundating the deltacreates poor quality ground water that undergoes diffusive exchange with overlying surface water.Because water flows continuously through the reservoir, the extent of diffusive effects are expectedto be small due to the short time for surface water to equilibrate with the sediment; surface watersamples collected immediately above the sediment at each well cluster were not clearly differentthan surface water in the river upstream of the delta. Nevertheless, dynamic bench tests demon-strated that diffusive effects to surface water are significant given sufficient contact time; theseeffects were greatly reduced by incorporating lime into the upper surface of the oxidized sedimenteven though the lime did not significantly improve pore water quality. Although sediment amend-ment significantly raised surface water pH, the pH of the surface pool dropped from the first flushtest to the second, suggesting that the positive effects of sediment amendment eventually may beexhausted after several reservoir cycles.

Once released from the delta sediment, metals entering the reservoir may undergo complexcycling between sediment and water in response to changing geochemical conditions. This isexemplified by the elevated concentrations of zinc that are found in samples from the hypolimniononly during summer months when water pH is commonly less than 6. That zinc concentrations aremuch lower in winter when pH is above 7 suggests that zinc cycling may be controlled by sorption-desorption reactions. These processes and their importance to risk have not yet been investigatedin detail.

REFERENCES

Black & Veatch, 2008. Remedial Investigation Report for the Ocoee River, Copper Basin Site, Operable Unit5, Polk County, Tennessee, CERCLIS ID TN0001890839. Report prepared by Black & Veatch SpecialProjects Corp. for EPA Region 4.

Eaton, Andrew D., Clesceri, Lenore, S., Rice, Eugene W., Greenberg, Arnold E., and Franson, MaryAnn H.,editors, 2005. Standard Methods for the Examination of Water and Wastewater: Centennial Edition, 21stedition. American Public Health Association, American Waterworks Association, and Water EnvironmentFederation, Washington, D.C.

McLean, E.O., 1982. Soil pH and Lime Requirement. In: A.L. Page, et al. (eds), Methods of Soil Analysis,Part 2, 2nd Edition, Agronomy Monograph 9, ASA and SSSA, Madison, WI, pp. 199–224.

Nevada Department of Environmental Protection (NDEP), 1996. MeteoricWater Mobility Procedure (MWMP),Standardized Column Percolation Test Procedure. Bureau of Mining Regulation and Reclamation.

Pfaff, John D., 1993. Method 300.0, Determination of Inorganic Anions by Ion Chromatography, Revision2.1. U.S. Environmental Protection Agency, Environmental Monitoring Systems Laboratory, Office ofResearch & Development, Cincinnati, OH.

Slater, R., 1982. Massive Sulfide Deposits of the Ducktown Mining District, Tennessee. Proceedings,Exploration for Metallic Resources in the Southeast, Athens, GA.

Sobek, A.A., Schuller, W.A., Freeman, J.R., and Smith, R.M., 1978. Field and Laboratory Methods Applicableto Overburden and Minesoils. U.S. Environmental Protection Agency, Report EPA-600/2-78-054, 204 pp.

Stumm, W. and Morgan, J.J., 1996. Aquatic Chemistry: Chemical Equilibria and Rates in Natural Waters,Third Edition. John Wiley & Sons, New York, 1022 pp.

U.S. Environmental Protection Agency (EPA), 1986. Test Methods for Evaluating Solid Waste, Physi-cal/Chemical Methods, 3rd edition. Office of Solid Waste and Emergency Response, Report SW-846,November 1986 with revisions to January 2008.

U.S. Environmental Protection Agency (EPA), 2000. Aerial Photographic Analysis, Copper Basin MiningDistrict Study Area, Ducktown and Copperhill, Tennessee, vols. 1 and 2. TS-PIC-99004449S/20004449S,Prepared by Environmental Services Division, National Exposure Research Laboratory, U.S. EPA Officeof Research and Development, Las Vegas.

U.S. Environmental Protection Agency (EPA), 2010. Copper Basin Mining District. EPA Region 4 Superfundsite summary, http://www.epa.gov/region4/waste/copper, viewed 6/2/2010.

240


Neutral mine drainage water-quality impacts from a formertaconite mine

B. HannaItasca Denver, Inc., Denver, CO, USA

ABSTRACT: Surface waters at the site of a former Minnesota taconite mine were reported tohave solute concentrations elevated with respect to water-quality standards. Waste rock and oregenerated from past mining were primarily from open pit mining of the Biwabik Iron Formation(BIF). The BIF is a variably bedded iron formation composed of inter-bedded cherty and slaty ironsilicate and iron carbonate rich beds. A geochemical characterization was conducted to identifypotential constituents of interest (COI), facilitate understanding of mechanisms controlling theirenvironmental behavior at the site, and guide future site activities. Primary COI were determinedto be SO4, hardness (predominantly from Mg), alkalinity, Fe, Mn, and Al. BIF waste rock fromthe Lower Slaty member appears to be the primary source for the identified COI. Mechanisms ofrelease are primarily attributed to pyrite oxidation and subsequent neutralization by dissolution ofmixed-composition (Ca-, Mg-, Fe-, and Mn-bearing) siderite and ankerite.

1 INTRODUCTION

Open-pit mining of taconite ore has led to the formation of several mine pit lakes at a former taconitemine on the Mesabi Iron Range, in northeastern Minnesota. Water quality from the site, was eval-uated during preliminary studies associated with redevelopment of the property and found to haveelevated solute concentrations with respect to water-quality standards or regional background con-centrations (e.g. for SO4, hardness, and alkalinity). A geochemical characterization was initiated atthat time to assess the sources of the elevated concentrations and guide environmental planning. Aspart of this characterization, a variety of geochemical tests have been conducted on waste rock andwaste rock leachate to help identify potential constituents of interest (COI) and facilitate under-standing of the mechanisms controlling their release and subsequent behavior in the environment.This paper summarizes the methods, results, and conclusions from the geochemical characteriza-tion and presents the basis of a geochemical conceptual model that can be used to guide futuremanagement and development of the site.

2 SITE GEOLOGY

The site is located at the eastern end of the Biwabik Iron Formation (BIF). The BIF dips gently to thesoutheast, although the geology of the site is complicated by local folding and faulting. The BIF isclassified into two types of iron formation: cherty iron formation, which is granular, massive, andtypically rich in quartz and iron oxides, and slaty materials, which are generally finely laminated,fine-grained, and contain iron silicates and/or iron carbonates.

Wolff (1917) divided the BIF into four informal members: the Lower Cherty, Lower Slaty, UpperCherty, and Upper Slaty. The Upper Cherty and Lower Cherty members comprise the primary orezones due to their higher concentrations of magnetite and cherty silicate taconite, whereas theLower Slaty and the Upper Slaty members are largely waste rock. The Lower Slaty member isapproximately 75 feet to 85 feet thick. The P submember comprises the upper 60 feet to 75 feet andis a fine-grained, finely-laminated, dark gray to black slate containing some disseminated pyrite. Atthe base of the P submember is approximately 10 feet to 15 feet of the Intermediate Slate (locally

241

referred to as the Q submember). The Q submember is typically a black to dark-gray graphiticsilicate- and quartz bearing ash-fall tuff containing up to 5% disseminated ultra fine-grained pyrite(Morey, 1993). Due to an erosional unconformity, the Upper Slaty member is absent from mostof the site, and therefore the Lower Slaty member comprises the majority of the approximately2 × 1011 kg of waste rock on site, particularly in areas of notable water quality.

3 METHODS

Samples of waste rock were recovered from stockpile drilling and selected for testing. Most sampleswere subjected to modified acid-base accounting, Synthetic Precipitation Leaching Procedure(SPLP) testing, aqua regia digestion, humidity cell testing, analysis of pore-water chemistry (forsamples collected from the saturated zone), and (for a subset) X-ray diffraction analysis (XRD).Samples of diamond drill core were subjected to modified acid-base accounting, SPLP, aqua regiadigestion, and (for a subset of samples) humidity cell testing and petrographic analysis. Detaileddescriptions of the methods utilized in the collection and analysis of these samples are providedbelow.

3.1 Sample collection

3.1.1 Waste rockThree boreholes were advanced in each of three stockpiles that together comprise more than 35%of the Lower Slaty waste rock on site. Sonic drilling methods were used to recover waste rockmaterial from the stockpiles without the use of water. A 4-inch diameter core barrel was advanced20 feet ahead of the 6-inch diameter casing while drilling above the level of the pit water, and wasreduced to 10-foot runs below the water level. Samples were shaken out of the core barrel into6-mil plastic sample bag sleeves. Samples were generally collected from three positions within theborehole: the surface directly below any cover materials present, just above the water level, and justabove the bedrock. Samples were crushed to less than 1/4-inch diameter using a jaw crusher, afterbeing allowed to air-dry overnight where necessary. The crusher was cleaned using brushes andjetted air between samples. After crushing, the samples were homogenized and split using the coneand quarter technique (Pitard, 1993). Samples were shipped overnight, under chain of custody, toNortheast Technical Services (NTS) in Virginia, MN and CANTEST, Ltd. in Burnaby, BC, fortesting and chemical analyses.

Continuous diamond drill core through the Upper Slaty, Lower Slaty, and the Lower Chertysubmembers immediately underlying the Lower Slaty was obtained from standard NQ diamonddrilling core from three boreholes from mining activities. Samples for chemical analyses werecollected from the Upper Slaty, Lower Slaty member of the Biwabik Iron Formation: submembersP and Q and the 20-foot interval underlying the Q submember (identified as the R submember ofthe Lower Cherty member). Samples taken for chemical analyses within these submembers weredetermined based on detailed geologic logging. Sample interval lengths were modified as necessaryto ensure that geologic contacts were honored, with only one geologic stratum comprising anysample. Samples were shipped overnight, under chain of custody, to CANTEST, Ltd. in Burnaby,BC and to ACZ for testing and chemical analyses.

3.1.2 Existing water qualitySite water-quality monitoring has been conducted as part of an ongoing monitoring program.Groundwater samples to be analyzed for dissolved metals were filtered in the field using an in-line45-µm disposable filter, following the low flow purging procedure (EPA, 1996). Surface-watersamples collected for dissolved metals were collected in unpreserved containers and were filteredand preserved (within 48 hours) upon receipt at the laboratory. Each sample container was labeledwith a unique sample identification number, placed in a cooler with ice, and submitted to thelaboratory for analysis. At each surface-water sampling site, sample bottles were filled using aclean sample bottle. Field blanks and field duplicate samples were collected at a ratio of one persampling event. Samples were sent under chain of custody to NTS for chemical analyses for generalparameters and metals (total and dissolved) by standard EPA methods (EPA, 2007).

242

Pore water was collected from waste rock stockpile boreholes using either a disposable baileror a submersible pump. Several gallons were purged from the borehole prior to sampling with thedisposable bailer. Between 20 and 30 gallons were purged prior to sampling with the pump. A totalof 15 pore-water samples were analyzed for general parameters and metals (total and dissolved)by standard EPA methods. Samples were filtered in the field with a 0.45-micron vacuum filterand preserved with HNO3 (for metals) and H2SO4 (for cations) in laboratory-provided samplingcontainers.

3.2 Mineralogy

The mineralogy of waste rock was evaluated using a combination of XRD, scanning electronmicroscopy (SEM), energy dispersive spectroscopy (EDS), and optical microscopy. The XRDanalyses included standard qualitative analysis to identify the minerals present as well as quantitativephase analysis by Rietveld refinements for select samples to identify not only the minerals present,but also their relative abundance. The SEM and optical microscopy work was performed to identifythe modes of occurrence of the carbonate and sulfide minerals, including their texture, grain size,morphology, and mineral association. The EDS analyses were conducted to determine the chemicalcomposition of the carbonate and sulfide minerals.

3.2.1 XRDSix samples of sonic drill core of the waste rock stockpiles were submitted for quantitative XRDanalysis at The University of British Columbia Dept. of Earth and Ocean Sciences. Sampleswere ground at CANTEST, Ltd. to <10 µm under ethanol in a vibratory McCrone MicronisingMill for seven minutes. Step-scan X-ray powder-diffraction data were collected over a range of3◦ to 80◦ 2θ with CoKα radiation on a standard Siemens D5000 Bragg-Brentano diffractometer.The resulting X-ray diffractograms were analyzed using Search-Match software, refined with theReitveld program Topas 3. Fifteen samples collected from diamond drill core were submitted to theMichiganTechnological University Department of Material Science and Engineering for qualitativeXRD analysis. Samples were ground, and step-scan X-ray powder-diffraction data collected overa range of 5◦ to 45◦ 2θ with a Scintag XDS 2000 Diffractometer. Background was subtracted andcrystalline phases were identified using Scintag Diffraction Management System for NT (DMSNT)software.

3.2.2 SEM and EDSSEM was used to investigate the physical characteristics of the minerals present in fifteen samplescollected from diamond drill core. In addition to the physical characteristics, the chemistry of thecarbonate minerals was analyzed using EDS in conjunction with the SEM.

Samples representing the range of modes of occurrence of sulfides and carbonates, as well as thestratigraphic units encountered, were selected from the core at the time of detailed logging. Thesesamples were polished, mounted, and carbon-coated and then examined using a JEOL 840-JXASEM equipped with a Kevex Sigma EDS system. EDS spectra were collected at an acceleratingvoltage of 20 kV for 120 seconds, maintaining a dead time of approximately 25%.

3.3 Acid base accounting

Acid-base accounting (ABA) provides a measure of the balance between the acid-producing capa-bility and acid-neutralizing capacity of mine wastes. Modified ABA was conducted on 25 samplesof waste rock (CANTEST, Ltd.) and 21 samples of diamond drill core (ACZ). Modified ABAtesting utilized partial-decomposition by wet-chemical-leach speciation methods for identificationof sulfur and carbon forms by combustion-infrared spectrophotometer, including:

– HCl-extractable sulfur (total sulfur from untreated sample minus total sulfur after extractionwith 40% HCl), which is attributed to SO4-sulfur,

– HNO3-extractable sulfur (total sulfur after HCl extraction minus total sulfur after extraction with14% HNO3), which is attributed to pyritic sulfide-sulfur,

– Non-extractable sulfur (total sulfur from untreated sample minus total sulfur after HNO3extraction), which is attributed to organic sulfur,

243

– Total carbon after extraction with 15% HCl, which is attributed to organic carbon, and– Total inorganic carbon (carbonate), which is calculated as the difference in total carbon and total

carbon after HCl extraction.

NP was measured with the method of Lawrence as described in MEND (1991), which is summarizedbelow.

Two grams of sample were pulverized to < 250 µm (minus-60 mesh) and digested at roomtemperature with 0.1 N HCl over a 24-hour period under mechanical agitation to achieve a pH ofapproximately 2. Subsequently, samples were titrated with 0.1 N NaOH to reach the endpoint pH,typically 8.3. However, in this case additional “endpoint” pH values of 4.5, 6.0, and 10 were alsorecorded.

3.4 SPLP

Twenty-five samples of waste rock and 21 samples of diamond drill core were submitted to CAN-TEST, Ltd. and ACZ, respectively, for SPLP extraction and analysis (EPA, 2007). Both labs utilizedextraction fluid number 3, which is deionized reagent water with no added acidity. The solid phasewas extracted with an amount of extraction fluid equal to 20 times the weight of the solid phase.Samples were agitated in an end-over-end rotary agitator with the extraction fluid for a period of18 hours.

3.5 Aqua regia digestion

Twenty-five samples of waste rock and 21 samples of diamond drill core were submitted to CAN-TEST, Ltd. and ACZ, respectively, for aqua regia digestion and analysis to evaluate the solid-phasechemical composition. Digestate was analyzed by inductively coupled plasma-atomic emissionspectrometry (ICP), inductively coupled plasma mass spectrometry (ICP-MS), and cold vaporatomic absorption (CVAA), to determine solid-phase elemental concentrations of select metals,metalloids, and non-metals. Analyzed constituents included Ag, Al, As, Au, B, Ba, Bi , Ca, Cd, Co,Cr, Cu, Fe, Ga, Hg, K, La, Mg, Mn, Mo, Na, Ni, P, Pb, Sb, Sc, Se, Sr, Th, Ti, Tl, U, V, W, and Zn.

3.6 Humidity cells

Humidity cells are designed to enhance the release of acidity/alkalinity, metals, and other con-stituents from solid materials by providing conditions conducive to sample oxidation and thenleaching with a fixed volume of water on a weekly basis. Samples of waste rock and diamonddrill core were submitted to CANTEST, Ltd. for humidity cell testing following the ASTM methodD5744-07, Option A (ASTM, 2007). Humidity cell tests were conducted for 25 samples of wasterock and for 12 samples of diamond drill core. For each humidity cell, one kg of rock was placedin a clear acrylic cell and was initially flushed with 750 mL of deionized water. The waste rockstockpile sample cells had an inner diameter of 11 cm and a 20-cm height from the base. Diamonddrill core samples had an inner diameter of 20 cm and height from the base of 11 cm. Subse-quently, cells were subjected to a weekly cycle composed of three days of dry air followed by threedays of water-saturated air and then a 500 mL flush with deionized water in a climate-controlledfacility. Humidity cell effluent was collected and analyzed for pH, electrical conductivity, oxida-tion/reduction potential, SO4, acidity and/or alkalinity, and dissolved metals by ICP and ICP-MS.Analytical frequency was initially weekly, but was later reduced following leachate concentrationstabilization.


4.1 Borehole observations

In general, waste rock encountered in the stockpile boring profiles was black, fine-grained, slatyor cherty iron formation, which was variously magnetic and consisted of coarse-grained gravelto cobble-sized clasts, with various amounts of sand and silt, some of which was a result of

244

Table 1. Summary of average water chemistry by site.

pH TDS Alkalinity Hardness SO4 Ca Mg Fe Mn Al As Co Cu Sb Se

Site mg/L mg/L CaCO3 mg/L µg/L

GW1 6.8 450 320 440 120 60 70 1.5 0.62 51.3 0.9 1.0 1.4 0.02 0.4GW3 7.2 760 300 640 320 110 90 20 0.65 109 5.2 0.4 0.5 0.02 0.3GW4 6.4 780 230 610 340 100 90 30 3.7 474 2.6 0.8 1.2 0.1 0.4GW5 6.8 2110 480 1700 1220 150 320 10 1.0 343 1.6 1.0 1.9 0.03 1.6GW6 7.4 370 140 290 150 60 30 3.1 0.61 1169 0.8 1.6 3.3 0.2 0.3Stream 8.0 676 290 650 330 40 130 0.80 0.16 23.4 2.0 0.4 0.4 0.1 0.6Pit 1 8.1 860 370 780 390 40 160 0.03 1.2 12.5 1.1 0.9 0.6 0.03 0.6Pit 2 8.5 440 320 400 110 30 80 0.03 0.01 12.5 4.1 0.3 0.6 0.1 0.5Pit 3 7.9 2040 570 1660 1150 50 370 0.04 2.3 12.5 2.0 4.1 1.0 0.1 2.4Pit 4 8.3 760 320 710 350 30 150 0.03 0.03 12.5 1.1 0.4 0.5 0.04 0.8

pulverization of the rock during drilling. Disseminated pyrite or similar sulfide minerals wereoccasionally noted in the samples after being crushed and dried. Pyrite also occurred as thinpartings or vein-fillings in some samples.

Diamond drill core from boreholes associated with mining operations was evaluated from threeboreholes to obtain additional information on the materials comprising the waste rock at the site. Ingeneral the holes encountered 40–50 feet of overburden, 55–65 feet of the Upper Cherty member,66–71 feet of the P submember of the Lower Slaty member, 11–13 feet of the Q submember of theLower Slaty member, and 140–145 feet of the Lower Cherty member. One hole also intersectedthe Upper Slaty member. In general, pyrite was commonly visible in the P and Q submembers.Dominant modes of pyrite included disseminated and veinlet occurrences, with veinlets generallyoccurring either parallel to or cross-cutting bedding.

4.2 Site water chemistry

Site water chemistry (summarized in Table 1) provides an indication of potential water-qualityconcerns and is useful for evaluating the potential for mobilization of constituents under variousenvironmental conditions. Average concentrations of Al, Fe, and Mn exceed drinking water-qualitystandards in all groundwater monitoring wells. Background concentrations for these constituents areelevated regionally (MPCA, 1999). GW03, GW04, and GW05 also exceed drinking water standardsfor TDS and SO4. GW01, GW03, and GW05 also exceed discharge standards for alkalinity.

Concentrations of Al, Fe, and Mn are generally lower in surface waters than in the groundwatermonitoring wells. Average concentrations ofAl and Fe are below water-quality standards for each ofthe pit lakes on site. Pit lakes exceed the discharge standard of 250 mg/L for alkalinity, the dischargestandard of 500 mg/L for hardness, and the discharge and secondary drinking water standardsfor TDS. SO4 concentrations are elevated in the pit lakes 1, 3, and 4 (390 mg/L, 1150 mg/L, and350 mg/L, respectively) compared to the secondary (aesthetic) drinking water standard of 250 mg/L.

The tendency for Al, Fe, and (to a lesser extent) Mn to precipitate from solution in circumneutralto alkaline oxygenated waters is consistent with their concentration trends at the site; they areelevated in groundwater, but have much lower concentrations in surface waters. Al does not needto be oxidized prior to hydrolysis and precipitation from solution, and iron oxidation at the mildlyalkaline pH of site waters is rapid (Eary and Schramke, 1990). However, slower oxidation kineticsfor Mn2+, particularly in the presence of elevated SO4 and bicarbonate concentrations can allow itto accumulate in solution (Hem, 1963).

The elevated concentrations of solutes in the pit lakes are consistent with the mechanisms dis-cussed previously. The (Ca+Mg):SO4 and (Ca+Mg):HCO3- ratios from waters on and near thesite are also consistent with the pyrite oxidation and neutralization mechanisms and relative ratesobserved in the humidity cells.

Pore-water chemistry is notably consistent, with pH from 7.4 to 8.0 and SO4 concentrationsranging from 1040 to 1420 mg/L for all but one sample. That sample was collected at the phreatic

245

Table 2. Quantitative XRD results in wt%.

Mineral Ideal Formula Sample 1 Sample 2 Sample 3 Sample 4 Sample 5 Average

Quartz SiO2 40.4 46.8 37.6 33.9 36.5 39Siderite Fe2+CO3 31.6 17.5 26.2 30.2 22.5 26Clinochlore (Mg,Fe2+)5Al(Si3Al)O10 15.3 0 20.6 18.1 0 11

(OH)8Stilpnomelane K(Fe,Mg)8(Si,Al)12 2.4 2.2 1.2 2.5 13.2 4.3

(O,OH)27Magnetite Fe3O4 4.3 6.5 6.2 3.5 0 4.1Minnesotaite (Fe2+,Mg)3Si4O10(OH)2 3.7 1.8 4.2 8 0 3.5Goethite α-Fe3+O(OH) 0 6.1 0 3.2 8 3.5Greenalite (Fe2+,Fe3+)2−3Si2O5(OH)4 0 4.7 0 0 11 3.1Hematite α -Fe2O3 0 5.6 0 0 4.2 2.0Ankerite Ca(Fe2+,Mg,Mn)(CO3)2 0 5.1 0.7 0.6 1.3 1.5Pyrite FeS2 2.4 0 3.3 0 0 1.1Talc Mg3Si4O10(OH)2 0 3.3 0 0 0 0.7Plagioclase NaAlSi3O8–CaAl2Si2O8 0 0 0 0 3.3 0.7Calcite CaCO3 0 0.4 0 0 0 0.1

surface and had a SO4 concentration of 2270 mg/L. The quality of pore water clearly does notrepresent either drinking water or discharge water quality. Nonetheless, comparison of pore-waterconcentrations with water-quality standards can provide an indication of constituents that meritfurther consideration when evaluating COI. Comparison of the dissolved fraction of pore-watersamples to surface-water discharge and drinking water standards indicates that alkalinity, hardness,TDS, SO4, and Mn were greater than standards in all samples, and that Co and Fe exceeded waterquality standards in one or more samples. Dissolved (Ca+Mg):SO4 ratios in pore water ranged from1.2 to 1.6, and (Ca+Mg):HCO−

3 ratios ranged from 1.4 to 2.6, consistent with partial degassing ofCO2 occurring during neutralization of acidity from sulfide oxidation.

4.3 Mineralogy

Results of the quantitative XRD analyses are provided in Table 2 and are consistent with thequalitative XRD analyses. Rocks of the Biwabik Iron Formation evaluated as part of this studygenerally contain trace to moderate pyrite (< 0.02 to 5 wt% as pyrite) and abundant Mg-rich siderite(average of 24 wt%). The minerals stipnomelane, greenalite, clinochlore, and, tentatively, magnesitewere also identified in the Lower Slaty P and Q submembers. Ankerite, calcite, and chamosite werealso identified in the Upper Cherty sample. Quartz, minnesotaite, talc, stipnomelane, greenalite,magnetite, and ankerite were also identified in the Lower Cherty.

Two general modes of occurrence were identified for pyrite: disseminated and veinlets (Fig. 1).Disseminated pyrite is generally present as euhedral grains (with an average size of approximately5 µm), as generally < 20 µm individual grains comprising aggregates, or as sieve-textured grains(small idioblasts lying within larger xenoblasts) up to approximately 300 µm. Pyrite veinlets wereobserved in samples from the Lower Slaty (P and Q submembers), generally as aggregates ofindividual euhedral grains ranging from 20 µm to 300 µm. Two orientations were common for thepyrite veinlets: parallel to bedding and cross-cutting. A tendency toward separation along thesepyritiferous veins and bedding was noted. Both modes of occurrence are generally concurrent,with the disseminated variety comprising the majority of the pyrite in all cases. Pyrite in the LowerCherty and Upper Cherty was less abundant than in the Lower Slaty and, where present, wasdisseminated. One grain of arsenopyrite was observed in a sample from the P submember of theLower Slaty (identified with SEM/EDS, but below XRD detection limits).

Siderite was the primary carbonate mineral identified in this study, although ankerite and sub-ordinate calcite were observed in the Upper Slaty and in trace amounts in the Lower Chertyand waste rock stockpiles. Based on EDS, the siderite has a consistent chemical composition

246

Figure 1. Photomicrographs of a) decussate greenalite, idioblastic carbonate, and xenoblastic sieve texturedpyrite, b) idioblastic to subidioblastic sieve textured pyrite in a siderite-rich matrix, c) atoll textured sideritewith disseminated pyrite, and d) disseminated pyrite in a granoblastic siderite and magnetite matrix.

of approximately Ca0.03Mg0.20Fe0.73Mn0.04CO3. Siderite in the P submember has a disseminatedidioblastic to xenoblastic mode of occurrence (Fig. 1). Siderite in the Q submember is alsodisseminated, but with a smaller grain size (approximately 25 µm). Siderite in the Lower Chertyis generally present as a granular aggregate with a grain size ranging from <25 µm to >300 µm.

4.4 Acid base accounting

Sulfide content of all samples ranges from 0.01 to 2.73 wt% S, with average values of 0.24 and0.66 wt% for the waste rock and diamond drill core samples, respectively. Carbonate contentranges from 0.11 wt% C to 8.3 wt% C, with average values of 2.5 and 3.1 wt% for the waste rockand diamond drill core, respectively. If all carbonate is assumed to be the Mg-rich siderite describedpreviously, this would correspond to an average content of 24 wt% siderite, which is in excellentagreement with the quantitative XRD results (26% siderite, Table 1) and the qualitative XRD, whichconsistently identified siderite as a major constituent.

The difference in composition of waste rock and diamond core is attributed to the sampling biastoward sulfidic zones (e.g. the Q submember, relative to the P submember and Lower Cherty) in thediamond core samples, and to the partial oxidation of the waste rock samples. Sulfide concentrationsare substantially higher in rocks from the Lower Slaty Q submember (average 2.68 wt%) than fromany other rock types tested. Sulfides are present at much lower concentrations in rocks from theLower Slaty P submember (average 0.14 wt%), and the Lower Cherty R submember (average0.16 wt%). Sulfide concentrations in the P and R submembers appear to increase with proximityto the Q submember. Sulfide generally comprises >95% of total sulfur for the diamond coreand >75% of total sulfur for the waste rock. The lower relative sulfide fraction observed in the

247

Table 3. Acid base accounting results.

Net neutralization potential*kg CaCO3 per tonne NP to AP ratio**

Sample type n Minimum Maximum Average Minimum Maximum Average

Waste rock 25 −11 58 22 0.59 166 21Lower Slaty (P) 8 23 91 51 5.6 38 20Lower Slaty (Q) 4 −43 −12 −31 0.51 0.87 0.63Lower Cherty 6 15 68 42 5.6 46 17Upper Slaty 1 – – 648 – – 84

*Net neutralization potential calculated as the difference in the modified NP to pH 8.3 and acid generationpotential calculated based on sulfide plus non-extractable sulfur.**Ratio of modified neutralization potential to pH 8.3 and acid generation potential calculated based on sulfideplus non-extractable sulfur.

waste samples is due to the presence of non-extractable (acid-insoluble) sulfur. For the purpose ofthis characterization, the acid-insoluble fraction has conservatively been included with the sulfidefraction in the calculation of acid-generating potential.

For the purpose of this characterization, the NP values to an endpoint of 8.3 were used forABA determinations. ABA results are summarized in Table 3. Net neutralization potentials (NNP)across all samples range from −43 to 91 kg CaCO3 per tonne with an average NNP of 22 kg CaCO3per tonne for waste rock. One sample of waste rock and five samples of diamond drill core haveNNP < 0, indicating acid-generating potential. Ten samples of waste rock and two samples ofdiamond drill core have 0 < NNP < 20, indicating potential uncertainty regarding classification ofthese samples as non acid-generating, depending on the criteria applied.

The NP to AP ratio for all samples ranges from 0.51 to 166 with an average of 19 and a median of6.8. One sample of waste rock and four samples of diamond drill core (all from the Q submemberof the Lower Slaty) had NP to AP ratios <1, indicating the potential for acid generation forthose samples. Seven samples of waste rock and one of diamond drill core have 1 < NP:AP < 4,indicating potential uncertainty regarding their acid-generating capacity classification, dependingon the criteria applied. The approach of determining a site-specific appropriate NP to AP ratio fordesignation of potentially acid-generating materials (Ferguson and Morin, 1991; Morin and Hutt,1994) has been applied to the site. As is presented below, observed (Ca + Mg):SO4 ratios for theimpacted areas of the site are approximately 1.4. Thus, an NP to AP ratio of 2 provides an additionalfactor of safety when applied to classification of site materials as potentially acid-generating.

To assess the applicability of the NP values from the Modified ABA, empirical NP values werecompared to theoretical mineralogic NP based on observed siderite composition. Comparisonwith theoretical mineralogic NP (Fig. 2) indicates that the Modified ABA to an endpoint pH of8.3 provided a generally conservative representation of the NP, consistent with published resultsobtained for similar siderite samples using the SobPer method (Jambor et al., 2003).

A final consideration relating to ABA is that of mineral availability. Two of the primary factorscontrolling mineral availability are the particle size distribution and the armoring or encapsulationof mineral grains. Both of these factors can be evaluated based on consideration of the mineralogicalevaluation. For example, because the surface area of a particle increases exponentially as its particlesize decreases, the actual reactivity will be a function of both the abundance of acid-generatingand acid-neutralizing materials and their respective particle size. Similarly, the presence of mineralcoatings (such as ferric hydroxides on calcite or sulfates on pyrite) can preclude these mineralsfrom participating in acid neutralization or generation (EPA, 2003).

As described above, grain sizes of pyrite and siderite range from <5 µm to approximately 300 µmand from <25 µm to approximately 300 µm, respectively. Furthermore, the general tendency is forthe larger carbonate grains to occur conjointly with the larger pyrite grains and, likewise, for thesmaller carbonate and pyrite grains to occur conjointly (e.g. see Fig. 1). None of the mineralogic orpetrographic analyses provided any indication of localized occurrences of pyrite in the absence ofcarbonate minerals. Because of the semi-consolidated to poorly-consolidated sedimentary nature

248

Figure 2. Comparison of mineralogic NP with NP measured using the Modified NP to pH 8.3.

of the deposits, the mineral grain sizes are small regardless of the size of the aggregate particles andthe porosity and friability are high relative to typical hard rock mines, particularly for the slaty rocktypes. Hence, the overall abundance of carbonate and sulfide is interpreted to provide a reasonableindication of the extent of their respective surface areas. Furthermore, none of the petrologicor mineralogic evaluations of waste rock provided any indication of encapsulation, armoring, oralteration rims that could potentially inhibit the reactivity of the sulfide or carbonate surfaces.

4.5 SPLP

SPLP leachate solute concentrations were generally low and pH was circumneutral to mildly alka-line (pH 7.3 to 9.0) for all samples. Specific conductance ranged from 28 to 236 µS/cm andSO4 from 1 to 87 mg/L in the waste rock samples and from 47 to 112 µS/cm and <1 to 9 mg/L,respectively, in the diamond core samples. Comparison of extract concentrations can provide anindication of constituents that merit further consideration when evaluating COI. Comparison ofSPLP extract pH and solute concentrations with various Minnesota water-quality standards forsurface-water discharge and drinking water indicates that the extracts from one or more samplesexceeded water-quality standards for Al, Cu, Fe, Mn, and Se.

4.6 Aqua regia digestion

Aqua regia digestion is a strong acid digestion that will dissolve almost all elements that couldbecome environmentally available. However, it does not necessarily indicate that constituents willbecome environmentally available (EPA, 2007). Constituents released at the highest concentrationswere Fe, Mg, Al, Mn, and Ca. Average concentrations exceeded the crustal average abundances(Price, 1997) for Mo, Ag, Mn, Fe, As, Cd, Sb, Bi, W, Hg and Se and were in excess of ten timesaverage crustal abundances for Mn, As, Bi, and Se. However, average extract concentrations forSb and Hg were below the average concentration for shale (Price, 1997). Only Ag was present atan average concentration greater than the average concentration for shale.

249

4.7 Humidity Cells

Humidity cell effluent pH ranged from 6.7 to 9.5 for the waste rock samples and from 6.9 to 8.3for the drill core samples. After effluent concentrations stabilized, SO4 concentrations ranged from<1 to approximately 200 mg/L, well below saturation with respect to the common SO4 minerals,which indicates that observed SO4 release rates are not affected by precipitation of secondaryminerals and are likely a good indicator of the sulfide oxidation rates occurring in the cells. SO4release rates from the waste rock samples calculated using weeks 15 to 26 range from 3.2 × 10−9 to4.2 × 10−8 kg of SO4 per kg of sulfide per second, with an average value of 1.5 × 10−8. Althoughthese calculated rates vary by a factor of 13, most variability is from cells with low SO4 production.Excluding cells producing less than or equal to 5 mg/L in week 26, the remaining (average weekly)rates still average 1.5 × 10−8, but are all within a factor of 2.9. Similarly, SO4 release rates fromdiamond drill core samples averaged 7.1 × 10−9.

Humidity cell effluent is not a direct measure of ultimate water quality. However, constituentswith humidity cell effluent concentrations greater than water-quality standards merit further con-sideration when evaluating COI. Humidity cell effluent from one or more samples of waste rockstockpiles has exceeded drinking water standards for Mn, Fe, Al, SO4, As (from one cell), and Cu(from one cell). Discharge standards were exceeded in effluent for hardness, and for Cu and Se fromone cell each. Humidity cell effluent from samples of diamond drill core exceeded water-qualitystandards for Mn, Fe, Al, SO4, Cu, and Se. Additionally, effluents from diamond drill core samplesfrom the Q submember of the Lower Slaty have also exceeded the drinking water-quality standardfor Sb and effluent from the R submember of the Lower Cherty and the Virginia Formation haveexceeded the discharge water-quality standard for Co.

Humidity cell leachate chemistry was evaluated using (Ca+Mg):SO4 and (Ca+Mg): HCO−3

ratios to identify relative acid-generation and acid-neutralization rates occurring in the humiditycells. The aggregate overall behavior was assessed by summing the Ca, Mg, SO4, and bicarbonateconcentrations for all cells through 26 weeks. The resulting overall (Ca+Mg):SO4 ratios were 1.5and 1.3 and the overall (Ca+Mg):HCO−

3 ratios were 0.80 and 0.70 for humidity cells on the wasterock and diamond drill core samples, respectively. The overall ratios from the humidity cell testsare similar to field averages for the pit lakes of (Ca+Mg):SO4 = 1.3 and (Ca+Mg):HCO−

3 = 1.5.The overall behavior of acid neutralization by siderite can be represented by Equation 1, whichrepresents approximately 80% CO2 degassing.

The acid-neutralizing consumption rate appears to be approximately 1.4 times faster than theacid-generating rate, both in the humidity cells and in the field.

4.8 Geochemical conceptual model

Primary COI include: SO4, hardness (Ca and Mg), alkalinity, Fe, Mn, and Al. These constituentshave been consistently observed at elevated concentrations with respect to water-quality standardsin site water and geochemical test effluent. Secondary COI include Co, Cu, As, and Se. Theseconstituents have been sporadically detected at elevated concentrations with respect to water-quality standards in site water and/or geochemical test effluent, but are not readily mobilized underprevailing conditions at the site. However, these constituents are present and can potentially bemobilized under certain conditions (such as reducing conditions or elevated pH).

Primary constituents of interest (COI) are associated with pyrite oxidation and subsequent acidneutralization, primarily by siderite and ankerite. Calcite equilibrium has been shown to controlthe relationship between pH and pCO2 in many carbonate aquifers (e.g. Langmuir, 1971; Plummer,1976). Similarly, the relationship between pCO2 and pH in the site pit lakes and waste rock stock-piles appears to be controlled by dissolution of siderite (and ankerite) in response to the amount ofsulfide oxidation occurring. In order to evaluate the conceptual model, these geochemical controls,including equilibration with atmospheric oxygen, progressive degassing of CO2, precipitation ofcalcite, precipitation of ferrihydrite, dissolution of pyrite, dissolution of siderite (with the aforemen-tioned chemical composition) to the extent necessary to yield observed site Ca+Mg:SO4 ratios,

250

Figure 3. Simulated relationship between pH and pCO2 defined by calcite equilibrium (solid line) fromLangmuir (1971) and Plummer (1976) and defined by dissolution of site siderite in response to pyrite oxidationwith subsequent precipitation of calcite and ferrihydrite.

and the pH resulting from thermodynamic equilibrium with the aforementioned controls) weresimulated in Phreeqc (Parkhurst, 1999). Comparison of the simulated results with pore-water andlake-water chemistry is shown on Figure 3. The model provides excellent agreement with measuredconditions at the site, indicating that the conceptual model provides a consistent explanation of thebehavior of primary COI at the site.

5 CONCLUSIONS

Waste rock from the Lower Slaty member of the BIF has impacted water quality at the site as aresult of pyrite oxidation and subsequent neutralization by Mg-rich siderite and ankerite, resultingprimarily in elevated SO4, alkalinity, and hardness. Sulfide concentrations are roughly an orderof magnitude higher in rocks from the Lower Slaty Q submember (average 2.7 wt%) than fromany other rock types tested. The Q submember of the Lower Slaty is potentially acid-generating,whereas the P submember of the Lower Slaty, the R submember of the Lower Cherty, and the topof the Upper Slaty are non acid-generating. However, the aggregated waste rock present in sitestockpiles appears to be net neutralizing as long as the relative rates of release of acid generatingand acid neutralizing minerals are maintained over the long term. Select humidity cells are beingcontinued for long-term testing to provide additional information regarding these relative rates ofrelease. Alternatives for management of waste rock to limit oxidation and infiltration are beinginvestigated as a means to ameliorate site water-quality.

REFERENCES

ASTM. 2007. Standard Test Method for Laboratory Weathering of Solid Materials Using a Humidity Cell. D5744-07.

Eary, L.E., & Schramke, J.A. 1990. Rates of Inorganic Oxidation Reactions Involving Dissolved Oxygen. InMelchior, D.C. & Bassett, R.L. (eds.), Chemical Modeling of Aqueous Systems II. ACS Symposium Series416, Washington, D.C.: American Chemical Society.

251

EPA. 1996. Low Stress (low flow) Purging and Sampling Procedure for the Collection of Ground WaterSamples from Monitoring Wells, Revision 2, US Environmental Protections Agency.

EPA. 2003. Characterization of Ore Waste Rock and Tailings. In EPA and Hardrock Mining: A Source Bookfor Industry in the Northwest and Alaska. US Environmental Protection Agency.

EPA. 2007. Test Methods for Evaluating Solid Waste, Physical/Chemical Methods, SW-846. US EnvironmentalProtection Agency.

Ferguson, K.D. & Morin, K.A. 1991. The Prediction of Acid Rock Drainage – Lessons from the Database.In Proceedings of the Second International Conference on the Abatement of Acidic Drainage, Montreal,Quebec, September 16–18, 1991.3:85–106.

Hem, J.D. 1963. Chemical Equilibria and Rates of Manganese Oxidation. Geological Survey Water-SupplyPaper 1667-A: A1-A64.

Jambor, J.L., Dutrizac, J.E., Raudsepp, M., & Groat, L.A. 2003. Effect of Peroxide on Neutralization-PotentialValues of Siderite and Other Carbonate Minerals. J. Environ. Qual. 32: 2373–2378.

Langmuir, D. 1971. The geochemistry of some carbonated groundwaters in central Pennsylvania. Geochimicaet Cosmochimica Acta 35:1023–1045.

MEND. 1991. Acid Rock Drainage Prediction Manual. Mine Environment Neutral Drainage Program.Morey, G.B. 1993. Geology of the Mesabi Range: Field trip guidebook (Trip 1): 39th Annual Institute on Lake

Superior Geology, Eveleth, MN .39(2): 1–18.Morin, K.A. & Hutt, N.M. 1994. Observed Preferential Depletion of Neutralization Potential Over Sulfide

Minerals in Kinetic Tests: Site-Specific Criteria for Safe NP/AP Ratios. Presented at the International LandReclamation and Mine Drainage Conference and the Third International Conference on the Abatement ofAcidic Drainage, Pittsburg, PA, USA, April 24–29.

MPCA. 1999. Baseline Water Quality of Minnesota’s Principal Aquifers – Region 1, Northeastern Minnesota.St. Paul: Minnesota Pollution Control Agency.

Parkhurst, D.L. & Appelo, C.A.J. 1999. User’s guide to PHREEQC (Version 2)—A computer program for spe-ciation, batch-reaction, one-dimensional transport, and inverse geochemical calculations: U.S. GeologicalSurvey Water-Resources Investigations Report 99–4259.

Pitard, F.F. 1993. Pierre Gy’s sampling theory and sampling practice. Heterogeneity, sampling correctness andstatistical process control. 2nd ed. CRC Press

Plummer, L.N. & Wigley, T.M.L. 1976. The dissolution of calcite in CO2-saturated solutions at 25◦C and 1atmosphere total pressure. Geochimica et Cosmochimica Acta 40: 191–202.

Price, W.A. 1997. Draft Guidelines and Recommended Methods for the Prediction of Metal Leaching andAcid Rock Drainage at Minesites in British Columbia. Reclamation Section, Energy and Minerals Division,Ministry of Employment and Investment, BC, V0J2N0.

Wolff, J.F., 1917, Recent geologic developments on the Mesabi iron range, Minnesota. Am. Inst. Mining Metall.Petroleum Engineers Trans., 56: 142–169.

252


Benefits of timely and valid geochemical characterization of minewaste for life of mine and closure planning: A case study of NewmontBoddington Gold Mine in Western Australia

Nelson Amoah & Rory HaymontNewmont Asia Pacific, Perth, Western Australia

Graeme CampbellGraeme Campbell and Associates, Bridgetown, Western Australia

ABSTRACT: The extraction of minerals often involves the removal and storage of large quantitiesof waste in the form of waste rock and tailings. These waste materials may have the potential for acidor metalliferous drainage. Many of the long-term environmental impacts associated with miningarise from the geochemical characteristics of these wastes. A significant proportion of mining costsand liabilities is linked to the storage of these wastes during the life cycle of the mine, especiallyat the closure and post-closure stages.

While geochemical characterization is usually conducted in the early stages of mine planningand continues during operations, in many cases the process is driven by the need to assess relevantmineralogical composition that has economic value, with less emphasis on the key environmentalelements that affect the long-term storage and impacts of waste materials during the life of mine,closure and post-closure.

Evidence abounds in the mining industry for the misclassification of potentially acid formingand non-acid forming materials, having significant adverse operational and environmental impli-cations during the life of mine. Environmental aspects of the geochemical characteristics of waste,particularly in terms of acid generation and other metalliferous drainage potential, are also veryimportant during closure and post closure and need to be given sufficient attention early in projectdesign and development, and as the process progresses.

Acceptable environmental performance at mine closure and post-closure is critical for the con-tinuation of mining companies’social licence to operate, and regulatory agencies in many countriesnow require a detailed closure plan that includes potential post-closure performance indicators, orclosure criteria, prior to mining approval and development. Closure and post-closure costs and per-formance go beyond the physical life of a mine and in many cases must be dealt with in perpetuity.Unfortunately, most long-term environmental issues associated with mining arise from geochem-ical characteristics of the waste in storage, which may not have been given sufficient attentionduring the early stages of mine planning and development.

At Newmont, an important element of closure planning now involves detailed assessment ofgeochemical parameters of mine waste that affects closure performance. This paper discusses thesignificant benefits of designing and implementing a geochemical characterisation program for lifeof mine, closure and post-closure planning. To illustrate this, a case study of a Western Australianmine site that has the potential to store over one billion tonnes of waste is discussed. Key risksand opportunities for understanding the specific and subtle issues critical for making managementdecisions are also discussed.

1 INTRODUCTION

Newmont Boddington Gold (NBG) operates a gold mine, located 17 km northwest of the town ofBoddington, and 100 km southeast of Perth, in Western Australia. Open pit mining of an oxide goldresource has been in operation since 1987, along with small underground mining. In 2008, NBG

253

undertook an expansion program that will take gold production to one million ounces per annumand will become the largest gold mine inAustralia by end of 2010. The project was commissioned inearly 2010 and production is at the ramp up stage. Associated with an expansion program involvinglarge open cut pits, is the requirement to store the large amount of waste rock and tailings materialsin an environmentally sound and sustainable way. Numerous investigations have been conducted atvarious stages since the feasibility stage to ensure the economic and environmental sustainabilityof the operation.

During the early operational phase, NBG is undertaking further studies to refine and update thefeasibility investigation findings and also to develop new innovative approaches for investigatingand managing the significant quantities of waste generated from the operation. Central to theseongoing studies is the need to develop an integrated life of mine and closure plan that will ensureefficient and cost effective operation of the mine both in the short-term and long-term. In achievingthis goal, the geochemical characteristics of the mine waste materials were considered critical, as itis one major factor that influences the storage performance of waste rock and tailings facilities, aswell as groundwater and surface water quality. Furthermore, Newmont recognizes that a significantproportion of mining costs and closure liabilities are linked to the storage of these wastes duringthe life cycle of the mine, and at closure and post-closure stages. Globally, Newmont’s closureliabilities amount to several hundred million dollars (Dowd, 2002; Dowd and Slight, 2009). Thisproportion of closure liability is also within the range for almost all mining companies of its sizeand above (ICMM, 2005).

To mitigate the closure liabilities, Newmont’s internal closure guideline has strict requirementsfor the development and review of closure plans at all stages, from feasibility as well as periodicreviews of closure plans during operations to ensure consistency with life of mine planning, changesin operations, stakeholder expectation, regulatory requirements and their impact of costs. The aim isto develop a company-wide system, where closure planning is embedded in the day to day activitiesthroughout the life cycle of all mining operations.

In the case of NBG, the size of the operation and the potential for future expansion requiresa thorough reassessment of the key closure assumptions and approaches at this early stage ofoperations in order to:

• Identify key components of the current closure plan that can have significant impact on futureliability.

• Ensure that current and future life of mine activities and operational decision are consistent withprogressive closure requirements and closure criteria.

• Provide the necessary input for the development of a realistic closure liability/cost estimates.• Understand and manage potential environmental impacts of the mine expansion that are likely

to require attention at the closure stage.

2 GENERAL SITE DESCRIPTION

Newmont Boddington Gold site is located about 100 km south of Perth in Western Australia(Figure 1). The site experiences a Mediterranean type climate, characterized by mild to hot drysummers with occasional storms between November and April and cool wet winter months (May toOctober). Maximum summer temperatures are typically in the 30’s Celsius, while overnight wintertemperatures are typically less than 10 degrees Celsius. Average annual (1984–2009) precipitationfor the site is 765 mm, of which 75% occurs in winter. Total annual evaporation is always higher thanrainfall and has an annual average of 1,400 mm, with 80% of evaporation (1,046 mm) occurring inthe period from the start of October to the end of March. Although the site experiences annual netpositive evaporation, average precipitation exceeds average evaporation in the winter months.

The NBG site is located within the Saddleback greenstone belt at the southwestern borderof the Yilgarn craton. From a geochemical point of view, the most common sulphides across thedeposits are chalcopyrite, pyrhotite and pyrite. Less common sulphide occurrences are molybdenite,arsenopyrite and sphalerite. The sulfur distribution with the deposit varies over a wide range. At thesouthern part (of Wandoo North Pit) total S/Cu ratios are low and coincident with chalcopyrite. Inthe northeast direction, the ratio increases and most sulfur in these zones is inferred to be related topyrite and arsenopyrite. A comparable trend is evident in arsenic distributions, with higher arsenicconcentrations coinciding with higher S/Cu ratios in the northeast.

254

Figure 1. Boddington Mine location map.

In summary, where sulphides occur in the andesite and diorite waste rocks, they are invariablyassociated with sulfur values less than 0.5%, and CaCO3 contents less than 1%. Weathering ofthese types of waste rocks therefore corresponds to the oxidation of trace amounts of sulphides ina groundmass where circum-neutral buffering rests chiefly with the fast-reacting calcites initially,followed by the slower-reacting primary silicates.

3 GEOCHEMICAL CHARACTERIZATION OF MINE MATERIALS

Geochemical characterization of mine materials usually forms an integral part of mine developmentfrom exploration through feasibility and mine planning and continues during operations stages.While environmentally related testing is conducted as part of this process, in many cases the process

255

Figure 2. Geological characteristics of the project.

is driven by the need to assess relevant mineralogical composition that has economic value and thekey environmental elements that affect the long-term waste storage are given less importance thanthey deserve. The environmental aspects of the geochemical characteristics of waste, particularlyin terms of acid generation and other metalliferous drainage potential are very important duringclosure and post closure performance and require sufficient attention early in project design anddevelopment and throughout operations stages.

3.1 Static tests

The most common approach for assessing the acid-forming tendencies of mine waste materialsemploys static tests based on the conventionalAcid BaseAccounting (ABA) methodology. Common

256

indices developed from this approach to quickly classify materials include Maximum PotentialAcidity (MPA), Acid Neutralization Capacity (ANC), Net Acid Producing Potential (NAPP), andNetAcid Generation (NAG).ABA classifications for mine-waste samples are generally divided intothree categories including Potentially Acid Forming (PAF), Non Acid Forming (NAF) or Uncertain(UC). Some of the empirical relations used in interpretation include the ANC/MPA ratio, and theNAPP (which is determined by MPA – ANC). If the ANC/MPA ratio is greater than 2, then thetested sample is generally considered to have low potential to acidify, and so may be classified asNAF. This ANC/MPA ratio of 2 applies especially when the ANC is dominated by the dissolutionof carbonates like calcite.

The attractiveness of this approach is that static tests are relatively quick, and inexpensive toperform. In terms of management decision making, the interpretation of the results from static testsis generally straight forward. However, this is not the case for the BGM waste rocks, because ofthe contribution made to circum-neutral buffering by primary silicates which is not fully accountedfor in the conventional suite of static tests employed in ABA assessments. This is particularlyimportant where the rates of sulphide oxidation are slow, since there is, in fact, a sizeable reserve ofalkalinity forms, arising from the hydrolysis and dissolution of primary silicates, which is effectivein maintaining circum-neutral-pH. The net outcome from static testing may then be that a minewaste sample (containing trace sulphides) is classified as PAF when it will actually never acidifyas its weathering proceeds to the full decomposition of the sulphides

Misclassification of PAF and NAF varieties of mines wastes is not uncommon, and can lead tosignificant adverse operational and environmental implications during the life of mine and beyond.Generally speaking, it is more often the case that a given lithotype may be classified initially asNAF, but which is actually PAF, and so becomes a source of acidic drainage. A common situationhere is where the ineffectiveness of strongly ferroan forms of carbonates (e.g. siderites) may resultin ANC values being significantly overestimated, so that an invalid NAF classification ensues.Given the significant quantities of waste rock materials at NBG, it is important that mine wasteclassification is valid for effective life of mine planning.

Although a conservative approach to geochemical classification of mine wastes is environmen-tally sound, this should not be taken to the extreme. Misclassification through over application ofthe precautionary principle may lead to significant storage costs that could be avoided (for examplemine wastes may be incorrectly classified as PAF, despite being NAF). In the latter example, insteadof there being a large quantity of PAF materials requiring isolation, there would be a large quantityof NAF materials suitable for use in other, cost effective operational and progressive closure works.This is a key challenge facing mine waste management at NBG, and reflects the trace occurrencesof sulphides in the andesite and diorite waste rocks which contain only trace amounts of calcites.

3.2 Kinetic tests

To ensure the full benefits of geochemical classification, testing is often extended to include kinetictests. By subjecting samples to simulated ambient weathering conditions, the rates of sulfideoxidation, carbonate depletion, acid generation and metal leaching may be determined to assessthe dynamics of weathering in a way that cannot be predicted from static tests alone (Bradhum andCaruccio, 1995). Results of NBG kinetic testing are discussed in the following sections.

4 GEOCHEMICAL CHARACTERISTICS OF NBG WASTE MATERIALS

The need to ensure environmental compliance of the significant volumes of waste material wasrecognized at the early stage of the NBG expansion project. Hence, detailed geochemical charac-terization of the various lithologies was conducted early in the feasibility stages. Comprehensiveinformation on static geochemical data which defines acid generation potential for all of the majormaterials was developed using drill core samples. In general, NAPP data derived from Total-Sulfur values are more common than ANC or NAG data. The ANC and NAG data indicate thatthe content of fast-reacting alkalinity forms (such as carbonates) is typically low, and reflects thelogging of only trace amounts of calcites (at most) in the andesite and diorite waste rocks. This lowbuffering capacity, due to reactive carbonates, was highlighted by a program of kinetic test workconducted over a two year period (McNeil and Nichols 2003). The kinetic results also highlighted

257

Figure 3. Newmont Boddington Gold total sulfur distribution in waste (parts per million).

several minor elements which have the potential to report to waste rock facility seepage at elevatedconcentrations.

Singularly, the key finding from the earlier investigations on the acid forming characteristics ofNBG waste rocks was the inability to validly account for circum-neutral buffering from primarysilicates which thereby biased the ABA assessment overall. Based on the static testing results,the majority of samples were classified as UC (uncertain) in terms of acid formation potential.The kinetic test work by McNeil and Nichols (2003) showed that NAF-Composite samples couldproduce acidic leachates. Accordingly, the earlier investigatory work resulted in almost all of thewaste bedrocks (those below the regolith profile) at NBG being classified as PAF. This equatesto approximately 60% of the total mine waste volume (this includes the sum of waste regolithand waste bedrock volumes). The majority of the so-classified PAF waste bedrocks would becharacterized by Total-S values less than 0.3% (or 3,000 mg/kg), as shown by the histogram ofTotal Sulfur distribution at NBG (Figure 3).

4.1 Operational approaches for dealing with geochemical uncertainties

The feasibility era classification led to a mine waste management plan which necessitates thecareful control of material segregation throughout operation. Static testing is undertaken duringthe blast hole sampling process and this information is used to carefully schedule and direct truckmovements to control material placement. With over 60% of the entire mine waste inventory for theproject classified as PAF, this is a significant undertaking. Although appropriate for environmentalprotection, such requirements impose significant limitations and inflexibilities on mine planning asareas planned for storage can become quickly consumed with the consequent cost impost. Hence,it was prudent to ensure that the classifications are appropriate. The geochemical uncertaintiesresulting from historical testing pose a dilemma for operational management at NBG, since theunilateral classification of all of the basement waste rock as PAF places significant constraints onwaste storage options, and ensuring implications for project costs.

5 CRITICAL APPRAISAL OF HISTORIC GEOCHEMICAL INVESTIGATIONS ANDCHARACTERIZATION

Coincident with the closure plan review, Newmont commissioned a suite of investigations to morefully characterize the geochemical nature of the waste bedrocks at NBG. Identifying the sources ofuncertainties in the historic mine waste classification, and their underlying likely causes, was central

258

Figure 4. Anomalous kinetic test results for non-acid-forming wastes from South Pit (McNeil and Nichols,2003).

to the follow-up work. A program of database interrogation and material testing was commissionedand several consultants with expertise in various technical disciplines were engaged. The geologicalresource database consisting of tens of thousands of sulfur and other elemental assays that werecompiled by the mine planning department was examined.

The investigation also included a reinterpretation of earlier kinetic testing in the light of thedesign of the kinetic testing program (similar to that described in the AMIRA (2002) document),and the slow rates of the hydrolysis and dissolution reactions of primary silicates at circum-neutral-pH. Since calcite only occurs as a trace component, its dissolution kinetics may also be suppressed,due to surface-chemical interactions arising from intimate associations between calcite and silicategrains.

An example of conflicting outcomes between predictions and observations is shown in Figure 4for the kinetic testing of a NAF-Composite sample of waste rock from the South Pit (McNeil andNichols, 2003). Although this sample had a Total-S value of approximately 0.1%, and classified asNAF, variable leachate-pH values were recorded during the two years of kinetic testing, and couldbe as low as 3.8 (see results for Col 4 S NAF below), but only during the summer months.

In a recent review of these kinetic results, the sulphide-oxidation rates (SORs) were estimated,and their variations plotted as a function of weathering history, together with the leachate-pH values(Figure 5). An inverse, and seasonal, trend is evident between the SORs and leachate-pH values,and is interpreted as reflecting both the use of a non-constant temperature room for the kinetictesting, and the particle size range (less than 9 mm with minimal fine earth fraction of less than2 mm) of the drill core derived samples. Use of flood lamps (AMIRA, 2002) during the daytimeto dewater the columns for five days each week meant that the diurnal temperature regime of thecolumns varied seasonally during the two years of kinetic testing. The coarseness of the waste rocksamples meant that the columns drained almost to completion within a matter of minutes followingcommencement of the flushing step (every four weeks) using deionized water. At the slower winterSORs, the short residence times of the flushing step was sufficient for the leachate-pH values tobe greater than 5. Leachate-pH values of 6+ would likely have been recorded if the residencetimes had been several hours (c.f. minutes), thereby allowing the opportunity for circum-neutralbuffering by the traces of calcites, and the primary silicates. By similar reasoning, at the fastersummer SORs, leachate-pH values near 4 were recorded.

It is therefore proposed that the observed acidic summer leachates are not representative of thetrue weathering pH regime, since the residence times of only minutes in the flushing steps were

259

Figure 5. Variation in sulphide-oxidation rates and leachate-pH values for non-acid-forming sample (Col 4S NAF).

too short. This reinterpretation of the kinetic results reported by McNeil and Nichols (2003) is tobe confirmed, or refuted, in the follow-up program of kinetic testing currently being planned. Itshould be noted that, from a pragmatic modeling viewpoint, the use of a non-constant-temperatureroom for kinetic testing has the advantage that seasonal variations in ambient temperature allowdirect assessment of the temperature dependence of SORs, and so is useful for modeling purposes,as required.

The study tasks currently in hand at NBG include:

• Comprehensive review of existing geochemistry data, including analysis of the very considerablestatic testing database generated since operations commenced;

• Detailed review of the key findings which underpin the earlier geochemical classification,specifically the outputs of a suite of kinetic testing undertaken since 2001;

• Preliminary testing to further understand the nature and behaviour of the material from currentas-mined materials, rather than drill core samples tested previously; and,

• Design of comprehensive set of medium to long-term tests to measure a wide range ofgeochemical signatures/attributes specific to the materials on this site.

The current geochemical investigations focus on the nature of the sulphide forms, andgroundmass-buffering properties of the NBG waste rocks (chiefly diorites and andesites), andthereby provide a better understanding of the long term weathering and leaching behaviour. Bothconventional testing methods and novel approaches relevant to this low sulphide and low bufferingcapacity materials, are being employed. The findings from the laboratory testing (both static andkinetic testing) will ultimately be extended by moving to large scale field trials on dumped wasterock materials in situ.

260

6 BENEFITS OF CRITICAL REVIEW AND FOLLOW-UP GEOCHEMICALCHARACTERIZATION

The discussions above show that the benefits from conducting careful and detailed geochemi-cal characterization cannot be overemphasized. With the limitation of storage space, the need tominimize waste rock dump sizes for short and long term cost implications, the difficulties andtime involved in obtaining approval for extensions to lease boundaries to accommodate the specialrequirements for storing PAF materials, coupled with the potential need for expansion in the nearfuture, it is clear that refinements that validly allow reclassifications are significant for both life ofmine and closure planning. This reclassification process currently being undertaken is estimatedto result in less than 40% of total waste materials being classified as PAF. When dealing with amine the size NBG where an estimated 1 billion tons of waste rock material will be stored over thelife of the mine, such a reduction in PAF materials is highly significant in all respects.

The benign component of waste rock material that has been misclassified could be used for bothprogressive and final reclamation purposes and can also be placed in areas allocated for generaldumping which will significantly mitigate the cost impost of the various management constraintsrequired for the hostile (PAF) material. The cost benefits during mine planning and the reductionin closure liabilities is in the order of tens of millions of dollars during the life cycle of the mine.

The timing of such investigation to reclassify the materials is very important for the operationof the mine, the life of which is currently around twenty years with potential for future extension.With such reclassification, a significant proportion can now be managed in a cost effective andenvironmentally responsible manner while mitigating what could have posed significant operationaland closure liability constraints.

7 CONCLUSIONS

A significant proportion of mining costs and liabilities is linked to the storage of these wastesduring the life cycle of the mine, especially at the closure and post-closure stages. This storagecost arises from the geochemical characteristics that imposes environmental restrictions on storagemethods; therefore, giving the required attention to the geochemical characteristics at an early stageis extremely important for the operations and closure of the mine. This paper has outlined, througha case study, the significant benefits of conducting a detailed, accurate and timely assessment ofgeochemical parameters of mine waste that affects life of mine closure performance.

The findings from this paper confirm the fact that, the assessment of geochemical characteristicsshould not be limited to a single approach such as static tests and its interpretation based of ABAmethodology. Because of the complexities involved in assessing and understanding the geochemicalweathering processes and the inter relationship between, weathering history and specific chemicalweathering processes such as sulphide-oxidation rates, tests should necessarily include kinetictests of various materials and should be done at the early stage of mine planning and continuedthroughout operations. Such approach is the only sure way of capturing the subtleties involved inthe acid forming characteristics of mine waste materials and their potential environmental impacts.

Furthermore, mine waste material classifications should be constantly reviewed and new findingsused in adjustment of LOM and closure planning. Such an integrated approach will save miningoperations significant cost, facilitates life of mine planning process, reduction in environmentalharm and minimize closure liabilities.

REFERENCES

AMIRA International Ltd, 2002, ARD Test Handbook, Prepared by Ian Wark Research Institute, andEnvironmental Geochemistry International Pty Ltd.

Bradham, W.S. and F.T. Caruccio (1995) Sensitivity Analysis of Laboratory Based Mine Over Burden, Analyt-ical Techniques for the Prediction of Acidic Mine Drainage, US Department of Interior/OSM, Pittsburgh,Pennsylvania Volume/Pages: p. 267.

Australian Government, Department of Industry, Tourism and Resources (2007), Leading Practice SustainableDevelopment for the Mining Industry.

261

Dowd, P. and M. Slight (2009). Business Case for Effective Mine Closure. Proceedings of the First InternationalSeminar on Mine Closure, 13–15 September. Perth Western Australia. (Ed. Andy Fourie and Mark Tibbett).Australian Centre for Geomechanics.

Dowd, P.J, 2005 The business case for Preventing Acid Drainage (Ed. C. Bell), Proceedings of the FifthAustralian Workshop on Acid Drainage, Fremantle, Western Australia. Centre for Minerals Extension andResearch, Brisbane.

Fitzgerald, P.T, (2005). Mine Closure and Exit Strategies. Risk financing Solutions for Environmental Lia-bilities and Sustainability. Canadian Institute of Mining, Technical Paper. Planning for Integrated MineClosure, Toolkit. International Council for Mining & Metals.

262

Containment systems


Disposal of coal mine slurry waste using geotextile containersat the North River Mine, Chevron Mining Inc.

Mike WattsConsultant

Ed TrainerTencate Geotube

ABSTRACT: The processing of raw coal to a saleable clean coal requires many mine operators towash the run of mine product using a processing plant. The raw coal contains impurities composedof rock and fire clay. Two waste streams are created by this process. Coarse rock and fine rockparticles. At the North River Mine the coarse rock is transported to a refuse disposal area byconveyor. The fine rock particles leave the processing plant suspended in water to form slurry.Slurry is normally disposed of via surface impoundments or injected into abandoned undergroundmine workings. The volume of this waste stream is very significant and expensive to dispose of.In this case approximately 1000 gallons per minute is created on a twenty four hour basis at theNorth River Mine in Alabama.

With a possible interruption of the primary disposal methods due to available area and construc-tion scheduling, a third method of slurry handling was sought for the interim. The mine neededto continue processing coal for shipment to be able to meet customer commitments. This requiredslurry disposal. Utilizing Geotube® Containers for dewatering the slurry waste from the processingplant solved the problem. After a successful test was conducted, and permits obtained, the minebegan using the geotextile containers to dewater and contain the solids from the waste stream.Chevron Mining Inc. solicited the help of the Alabama Surface Mine Commission, the Office ofSurface Mining, TenCate Geosynthetics, J.F. Brennan Co., Inc., Whittemore Farms Excavation,and PERC Engineering Co., Inc. to develop a unique and successful method to solve the problem.

1 BACKGROUND INFORMATION

North River Mine is an underground coal mine producing over 7 million tons of raw coal per year.The raw coal is processed to yield 3.5 million tons of clean saleable coal. Raw coal is processed atthe mine preparation plant at approximately 1000 tons per hour. This yields about 550 tons of cleancoal per hour. Refuse is therefore 450 tons per hour of which coarse rock is the primary by-product.At normal operating levels fireclay and fine rock particles suspended in water at a rate of 70 tonsper hour in dry weight is also a waste by-product. The slurry waste stream reports to a three milliongallon concrete thickener tank adjacent to the preparation plant. Solids in the thickener underfloware increased with the addition of polymers to produce a waste stream of about 1000 gallons perminute. Solids in this pump discharge vary between 25 and 35%. A total of approximately 1.5million gallons of slurry is produced per day. Particle size analysis of the fireclay and rock slurryreveal that 80% are 400 mesh or smaller. The ultra fine particles tend to stay suspended in the water.

North River Mine began seeking an alternative method of disposal of this slurry waste streambeyond the conventional methods of surface impoundment or injection into abandoned undergroundmine workings. Since containers made from geosynthetic materials have been used for dewateringvarious types of sludge wastes, it was thought that they might be able to do the same with coal minewaste slurry. One type available is the Tencate Geotube® Container. The material used for thesecontainers is fabricated from a specially engineered dewatering textile fabricated from high tenacitypolypropylene multifilament and monofilament yarns which are woven into a stable network such

265

that the yarns retain their relative position. This provides a sieve to hold the particles and allowthe water to run out. The result is a reduction in the retained water and consolidation of the solids.Containers are constructed with PVC fill ports for the attachment of pipes from the pump or dredgethrough a manifold that allows the filling of several bags at once.

2 PRELIMINARY TESTING

InAugust of 2007, a test was conducted at the mine to determine if the geosynthetic fabric containerswould successfully dewater the slurry waste sufficiently to become a viable option for disposal.(Figure 1) Two one hundred foot long geosynthetic test bags were placed on a pad that had beengraded to a 1% slope. In order to facilitate the capturing of the material in the containers and prevent‘blinding’ of the fabric, chemical injection was required.

A chemical treatment pump, tanks, and pipe manifold were assembled. (Photo 1)The treatment plant utilized an anionic flocculation polymer and a cationic coagulant polymer

to treat the slurry before it was pumped into the bags. (Photos 2 and 3)For the test, slurry was pumped directly to the bags from the preparation plant underflow. The

bags were filled on August 21st and 22nd. Polymer injection was adjusted when necessary as

266

the changing solids percentage required. As the bags were being filled with the slurry mixture,clear water flowed from them draining into the waste water sediment pond below. The effluentwas almost totally clear. For the test, a volume of about 500 gallons per minute was processedalternating between the two containers. After about two days water ceased to flow from the bagsand the bags were stable. The test was successful and determined that the system would work on alarger scale.

Once all the needed data was collected from the containers, they were split, the material loadedout with a front end loader into trucks and hauled to the coarse refuse disposal area. The materialhad a consistency of fine, wet sand.

267

3 IMPLEMENTATION

In order for the system to be employed on a larger scale the mine had to design the system andseek a permit from the Alabama Surface Mine Commission. Due to the volume of slurry to beprocessed the mine sought a safe and efficient plan that would allow the containers to be reclaimedin place instead of opening them and transporting the material to the coarse refuse disposal area.The bags are designed to contain, dewater, and consolidate the solid material. A plan was developedto construct bag fields upstream from the existing South Slurry Pond on the mine property. The

268

effluent water would drain to the pond. This pond provided an environmentally safe vessel for theeffluent from the process. The Office of Surface Mining provided a technical review of the designand permit as well. This was the first time that the use of geosynthetic bags had been utilized fordisposal of slurry waste from a mine washing facility on a large scale. (Figure 2)

Construction of the pads or bag fields included the following steps:Removal of vegetation and storage of topsoilExcavation of earth to construct the pad on grade with 1% slope back to pondCovering the bag field with 6” inches of drain material (a blend of sandstone 1/3 # 4, 1/3 # 57,

and 1/3 # 89).Covering that with 3” crushed limestone (# 57)Construction of rock drains and safety berms around perimeter (Photo 4)In order to maximize utilization of bag field areas, bags can be stacked on top of each other

in a pyramid fashion once the water runs out and the bags have stabilized. Containers were to be

269

placed on top of each other overlapping the line between them. The first field (Bag Field # 1) wasapproximately two acres in size. The design called for a total of four levels of bags to be filled.Containers were ordered for the field as follows:

11 bags 186-ft long for Level A9 bags 172-ft long for Level B8 bags 157-ft long for Level C7 bags 143-ft long for Level D

270

The bags used for the field were 60 feet in circumference and could be filled to a height of sevenfeet. (Figure 3).

Polymer tanks, pumps, and pipe manifolds were installed on the pad. Other equipment involvedincluded lift truck necessary to unload, maneuver, and stacking of the tubes, light plants, nerators,and a field office. Level A bags were then placed on Bag Field # 1 according to the design plan.(Photos 5 and 6).

The original test was conducted by pumping slurry directly from the preparation plant. Theoperation plan for the project was modified by installing a dredge in the 250 acre foot slurryimpoundment to pump to the containers. Removing the material from the impoundment wouldcreate capacity in the pond faster than the plant was producing slurry. A Dredge Supply swinging

271

ladder 8 inch dredge was placed into the existing slurry pond to pump the slurry from the pond tothe bags. This machine was capable of pumping 1750 gallons per minute to the bags. The bags onLevel A were systematically filled from one side to the other allowing Level B bags to be placedon top of Level A safely. This process was then continued to Level C and Level D thus utilizing allof the space designed in Bag Field #1. (Photos 7 and 8)

The pipe manifold allowed the operation to switch from one bag to another, using valves. Oncea bag was filled, flow was directed to another bag while dewatering occurred. The containers usedwere designed for a maximum filled height of 7 feet. The bags were immediately de-wateringthe slurry and becoming stable quickly. Once the material reached a moisture content of 35%other layers could be added. While filling Bag Field # 1, excavation of Bag Field # 2 began. Thesecond bag field was constructed prior to completion of the first providing two workable layoutareas for continuous operation thus eliminating downtime for the dredge and providing time forthe containers to drain. Using this method a total of three bag fields were constructed and utilized.The first bag field was modified to allow a second tier of four levels to be added as well.

Productivity was gauged by measuring the height of the filled and retired bags each day. At aboutthe midpoint of the project a physical survey was conducted in the slurry pond to verify the volumepumped. The dredge was initially operated on a twenty four hour basis with two 12 hour shifts perday. A crew of 5 to 6 employees was required to run the dredge, install the piping, maintain thepolymer station, position the bags, and monitor the filling operation. Later in the project only one12 hour shift per day was used. This allowed even more dewatering time for the bags being filled.

Hourly productivity for the project varied slightly with bag field distance from the dredge as wellas solids in the slurry. Productivity averaged 1750 cubic yards per twenty-four hour day. A total of200,000 cubic yards were pumped and disposed of during the project. This yardage was producedfrom January – August 2008. Since the dredge operation was independent from the preparationplant, the project operated on a different schedule from that of the mine. The completed projectutilized 240 bags with a combined total length of 42,000 linear feet. The circumference of the bagsranged from 60 to 70 feet. The containers held an average of 5 cubic yards of material per linearfoot of bag. (Photo 9)

272

Once all the bags in the field were full and dewatered, the site was then ready for reclamation.First a layer of sand was used to cover the sides of the bags to provide a filter medium and then alayer of limestone for drainage. The bags were then covered with earth and finally topsoil. A lowground pressure bulldozer was used to cover the containers. All the bag field sites were very stableand no problems were encountered during the covering operation with the equipment. (Photo 10)

Once this was complete the entire field could then be mulched and seeded using a hydro seeder.The entire project was completed without an accident or environmental incident. (Photo 11)

4 CONCLUSION

Although a relatively new application of this technology for the mining industry, the utilization ofGeotube® Containers worked extremely well and provided an alternate method of coal mine slurrywaste disposal for North River Mine.

ACKNOWLEDGEMENTS

Tim LeBlancJF Brennan Co., Inc. – LaCrosse, WIMike WindleChevron Mining Inc. – Berry, AL

273


Identification, management and disposal of PCB-containingequipment used in mines

D.W. BenchUS Environmental Protection Agency

ABSTRACT: Polychlorinated biphenyls (PCBs) are highly stable toxic organic compounds thatremain a persistent environmental threat for decades. The molecules, once valued for chemicalstability and fire resistance, made their first industrial debut in 1929, being manufactured andprocessed primarily for use as insulating fluids and coolants in electrical equipment. In 1979, theUSEPA issued final regulations banning the manufacture of PCBs and phasing out unenclosedPCB uses based on bioaccumulation and toxicity data. However, many PCB-containing pieces ofelectrical equipment remain in abandoned surface and deep mining operations worldwide, releasingthe toxic compounds into groundwater and eventually to the oceans where they bioconcentrate inphytoplankton, the basis of the ocean food chain and producer of ∼50 percent of atmosphericoxygen, as equipment deteriorates. This paper addresses five arenas of PCBs: environmentalhazards, identification information, hidden sources to look for, potential liabilities and what todo when you find PCBs.

1 PCB PROPERTIES

There is no longer any doubt that PCBs present threats to human health and the environment.They can contribute to local groundwater contamination and disseminate worldwide throughoutthe ocean, which is considered to be the final sink for PCBs (Goddard et al., 2003).

PCBs are one of the 12 chemicals targeted by the global Stockholm Convention on PersistentOrganic Pollutants (POPs). POPs are chemicals that remain intact in the environment for long peri-ods, become widely distributed geographically, accumulate in the fatty tissue of living organisms,and are toxic to humans and wildlife. POPs circulate globally and can cause damage wherever theytravel (Stockholm Convention. 2008). By implementing the Convention, governments are takingmeasures to eliminate or reduce the release of POPs into the environment. There are 152 signatoriesto the convention, including the United States.

The physical and chemical properties that make PCBs valuable commercially also make themenvironmentally detrimental. PCBs are very stable and resist breakdown at high temperaturesand from aging. Once in the environment, PCBs can easily cycle between air, water, and soil(USDHSS.2000a). Because of their solubility in water and fats, PCBs can readily enter the foodchain. The major dietary sources of PCBs are fish (especially sportfish that are caught in contami-nated lakes or rivers), meat, and dairy products (USDHSS.2000b). Human health issues/symptomsassociated with PCB exposures were one of the factors that lead to early suspicion about thechemical’s safety, which motivated the passage of theToxic Substances ControlAct (TSCA) in 1976.

The mining industry, including abandoned surface and underground mines, presents a uniqueset of challenges for identification, management and proper disposal of PCBs the release of whichcan have important large-scale negative social and ecological impacts. PCBs have been used asdielectrics in electrical equipment and as coolants in motors in continuous miners and loaders inthe mining industry. A common misperception is that because the USEPA regulations banned themanufacture of PCBs in 1979, PCB-containing electrical equipment is no longer in use. This is notthe case. The regulations authorize the use of intact and non-leaking PCB transformers, capacitors,and fluorescent light ballasts for the useful life of the equipment. Today, the mining industry stilluses PCB-containing electrical equipment and some of it continues to be abandoned undergroundas documented by USEPA Region 8 mine inspections.

275

But abandonment may translate into an unwanted corporate responsibility for PCB contaminationof humans, plants, animals, and the environment for the future. Abandoned equipment is nolonger being maintained; the next step is deterioration with the inevitable release of its chemicalcomponents, including PCBs. Most mining operations communicate very closely with groundwaterwhich can eventually reach the ocean.

In the ocean, PCBs can cause a number of potentially serious problems: PCBs and other POPswill bioconcentrate in phytoplankton, the unicellular algae in the surface layers of fresh waterlakes, rivers and the ocean that play a key role in regulating atmospheric carbon dioxide concentra-tions (Smetacek.2008). Phytoplankton is the primary food source, directly or indirectly, of all seaorganisms. Data show that PCBs affect the productivity of phytoplankton and the composition ofphytoplankton communities (Eisler.1986 and USEPA.1997). Not only do phytoplankton play a keyrole in regulating atmospheric carbon dioxide concentrations, but about 50 percent of the world’soxygen is produced by phytoplankton in the oceans (NASA.2008)

The median bioconcentration factors (BCFs) for accumulation [of PCBs] from water by phyto-plankton range from 1 × 104 to 1 × 106 (USDHSS.2000c). Given that PCB concentrations rangefrom 0.24 to 5.7 × 10−12 g/L in the North and South Atlantic oceans (Gioria.2008), PCBs canbioconcentrate in phytoplankton and then bioaccumulate in the food chain through phytoplanktonbeing eaten by zooplankton which are in turn are eaten by small fish and then larger fish that enterthe human food chain. The Federal Food and Drug Administration (FDA) has been compelled toissue the temporary PCB tolerance of 2 ppm for human consumption of the edible portion of fish(FDA.2010).

It follows that the abandonment of PCBs is not only a hazard to human health and the environmentit also presents potential pollution-related liabilities to the mining industry. This industry can playa significant role in preventing further increases of PCB contamination in the ocean by eliminatingor controlling its PCBs.

There may be no solution for the long-term water pollution that can be caused by PCBabandonment underground.

2 PCBS IN THE UNITED STATES

PCBs were manufactured in the US under the trade nameAroclor before manufacture was prohibitedby the regulations in 1979. Aroclors, which are waxes or oils, were liquefied using technical gradetri- and tetrachlorobenzenes, which give high-concentration PCB dielectrics their characteristicodor. Aroclors are odorless; different aroclors have different percent chlorine weight fractionsof PCBs that result in different properties. A typical aroclor is designated 1254. Twelve denotesthe number of carbons in the PCB molecule and 54 denotes the percent chlorine weight fraction.Aroclor 1016, commonly used in capacitors, is an exception to this nomenclature. Mixtures ofPCBs and solvents were sold under the trade names that appear on the manufacturer nameplatesof PCB electrical equipment. Some of the more common PCB dielectric trade names are: Pyranol,Interteen, Elemex, and Chlorextol. There are many others; by content they average ∼60 percentPCBs. The generic name for these fluids is Askarel.

PCBs are not the only chemicals used in mines. Underground repair facilities have used chlori-nated solvents such as trichloroethane, tetrachloroethene, and methylene chloride for cleaning anddegreasing equipment. The release of these solvents, in addition to constituting their own threatsof ground water contamination, can mobilize PCBs, facilitating transport into ground and surfacewaters. Some mines maintain their own landfills which contain improperly disposed of PCBs andsolvents.

3 IDENTIFYING PCB-CONTAINING ELECTRICAL EQUIPMENT

The regulations require transformers and capacitors containing three pounds or more of dielectricbe identified by PCB Marks (Fig. 1) placed on the equipment by the owner or user if they contain≥500 ppm (0.05%) PCBs. They are designated PCB transformers or PCB capacitors. Mineraloil transformers containing ≥50 ppm (0.005 percent) PCBs are designated PCB-contaminatedtransformers. All of the above are regulated for use and disposal. This does not necessarily mean

276

Figure 1. PCB Marks may have a yellow or white background.

Figure 2. Two 76-gallon PCB (Pyranol) transformers in the Eagle Mine at Gilman, Colorado.

that PCBs of lower concentrations or PCBs in small capacitors are not hazardous to human healthand the environment. For purposes of this paper “PCB-containing” means dielectrics containingany detectable quantity of PCBs.

4 PCB TRANSFORMERS

PCB transformers (Fig. 2) may carry a PCB trade name on the manufacturer nameplate. Theregulations require the assumption that a transformer manufactured prior to July 2, 1979, containing

277

Figure 3. PCB-marked capacitor on the left.

fluid other than mineral oil and whose concentration is not established is a PCB transformer andcontains ≥500 ppm PCBs. If the date of manufacture or the type of dielectric fluid is unknown, thetransformer must be assumed to be a PCB transformer. PCB concentrations must be established bychemical analysis or testing using SW 846-Method 8082 and other methods (USEPA.2010), (seeASTM D-4059), to document compliance with the regulations. Rebuilt transformers may carryreplacement nameplates that do not correctly identify the dielectrics.

5 MINERAL OIL TRANSFORMERS AND VOLTAGE REGULATORS

Transformers with “oil” or “mineral oil” on the manufacturer’s nameplate originally contained onlymineral oil dielectrics but may have been contaminated with PCBs. Sometimes there is no dielectriclisted. This is common when a transformer has been rebuilt. Testing is the only way to determinewhether or not a mineral oil transformer has been contaminated.

Dielectrics of mineral oil transformers and PCB transformers were often mixed during servicingresulting in transformers with mineral oil on the nameplates that may contain PCBs as a contam-inant. In writing the regulations, USEPA concluded the hazards of PCBs warranted regulationin transformers at ≥50 and ≥500 ppm PCB. This is why mineral oil transformers contaminatedwith ≥500 ppm PCBs are regulated as if they are trade name PCB transformers that typically con-tain about 600,000 ppm (60%) PCBs. Voltage regulators and substation transformers can becomeregulated PCB articles if internal small PCB starting capacitors leak (also see Section 12).

6 PCB CAPACITORS

By 1976, 95 percent of the capacitors produced in the United States were filled with PCBs(USDHSS.2000d) (Fig. 3). PCB capacitors contain the pure aroclors 1242 or 1016. Manufacturebefore July 2, 1979, or a PCB trade name on the nameplate is a good indicator of high concentrationPCBs. The regulations require the assumption that a capacitor manufactured prior to July 2, 1979,and whose PCB concentration is not established, contains ≥500 ppm PCBs and is a PCB capacitor.If the date of manufacture is unknown, the capacitor must be assumed to be a PCB capacitor andassumed to contain ≥500 ppm PCBs. Because most capacitors are sealed units, that testing willrequire penetration of the casing that may destroy their usefulness and result in leaks which must

278

Figure 4. Fluorescent light ballast.

be eliminated or contained. A helpful aid in identifying capacitors that do not contain PCBs is thatthe regulations require capacitors manufactured between1978 and 1998 be marked “No PCBs” bythe manufacturer.

7 FLUORESCENT LIGHT BALLASTS

Typical fluorescent light ballasts (Fig. 4) manufactured before May 31, 1979 contain a small capaci-tor buried in the tar or asphalt potting material filling the ballast that functions as an insulator. Thesecapacitors hold about an ounce of aroclor. Asphalt material in fluorescent light ballasts manufac-tured before 1978 has been found to have a better than 50 percent chance of containing regulatedlevels of PCBs (63 FR 35384, 35403). As with capacitors, a means of identifying fluorescent lightballasts that do not contain PCBs is a manufacturer-emplaced “No PCBs” mark required between1978 and 1998.

8 ELECTRICAL CABLE

Electrical cable (Fig. 5) can contain PCBs. In some cases, the cable is enclosed in a lead jacketwhich makes it difficult to handle. If electrical cable contains liquids or damp insulation, PCBsshould be suspected.

9 PCB REPLACEMENT

Replacement of PCB dielectrics can be a good investment considering the potential costs of cleanupfrom PCB spills or fires. In the United States, PCB contamination can result in liabilities underthe Comprehensive Environmental Response Compensation and Liability Act (CERCLA) at anyconcentration. Fires and explosions involving PCB-containing capacitors, transformers or anyother PCB-containing electrical equipment, including transformers with contaminated mineral oildielectrics, can create polychlorinated dibenzo-p-dioxins and polychlorinated dibenzofurans someof which may be more toxic than PCBs and not be easily cleaned up and can permanently shutdown facilities.

Mines and related facilities need not face these risks today because there are good alternatives toPCBs on the market. Transformers, capacitors, fluorescent light ballasts, and cable manufactured

279

Figure 5. Lead jacket electrical cable.

in the United States after July 2, 1979 should not contain PCBs. Mineral oil dielectric fluids intransformers are common but can present fire and explosion hazards as can vegetable oil dielectricswhen subject to high temperature faults within the transformer. A number of major United Statespower distribution companies have programs to eliminate PCBs and are adopting biodegradablevegetable oil dielectrics. Company representatives report biodegradability is an important con-sideration. Many coal and hardrock mining companies have adopted dry-type transformers forunderground use. Dry-type transformers are transformers in which the core and coils are in gaseous(usually air) or dry compound insulating mediums. In explosive atmospheres dry-type transform-ers must be in properly ventilated non-combustible containers. Associated switch gear, or sparkproducing items, such as fuses, circuit breakers, and relays must be in explosion proof enclosures(MR.2010).

There are a number of PCB-free capacitor dielectrics on the market. However, some of thesedielectrics can present their own hazards to human health and the environment and constitutefinancial risks if released into the environment.

10 MINES AND PCBS

Underground and surface mines and the attendant crushing, milling, and smelting facilities mayuse PCB-containing electrical equipment. PCBs are most likely to be in transformers, capacitors,fluorescent light ballasts, cable, and drums of unidentified oils. Underground inspections haverevealed transformers grouped in permanent substations, located singly, or mounted on mine carsthat can be transported throughout the mine. Capacitors are found in locations similar to thoseof transformers. PCB capacitors have been found in electric locomotives. In coal mines, PCBcapacitors have been found in wheel or skid-mounted power centers (Fig. 6).

Inspections have revealed PCB electrical equipment in just about every major electrically pow-ered mining activity including in draglines used in open-pit coal mines and in power shovels atopen-pit metal mines; inspections also found PCBs in discarded equipment at a variety of typesof mines. PCBs have been found in underground substations, pump stations, mine power centers,and electric locomotives. PCBs have also been found in surface facilities including hoist facilities,mills, smelters, metal refineries, breaker houses, and transfer facilities. Experts in the miningindustry believe that substantial quantities of electrical equipment with PCB-containing dielectricshad been abandoned underground before the advent of the PCB regulations (Personal communi-cation 1978–94). Even after promulgation of the PCB regulations, PCBs have been abandonedunderground, especially in situations where it was not cost effective to remove them.

280

Figure 6. Coal mine power center containing PCB capacitors.

11 PCB MANAGEMENT AND DISPOSAL

The following is a summary of ways to manage PCBs; this information should be supplementedby consulting PCB regulations.

Identify and inventory PCB-containing electrical equipment. Equipment should be markedso it is easily recognizable and not disposed of inadvertently. Written records identifying PCBequipment and their locations are essential. Records should include a serial number from themanufacturer nameplate or a company assigned number that has been placed on the equipment,dielectric identification from the nameplate, dielectric quantity and PCB concentrations from labanalyses.

Examine the equipment for leaks and clean up any leaks or spills potentially containing PCBs.A common concept of a leak is that it is a drip, sometimes referred to as a weep or seep. However,a leak is defined in the regulations as any instance in which a PCB article (including a transformer,voltage regulator, capacitor, lead cable, or fluorescent light ballast) or container that has any PCBson any portion of its external surface. This means a leak can be an oily film or oily dirt nearany port or opening in the equipment. Leaks should be cleaned up and the equipment repaired orthe leaking equipment containerized, moved to safe storage and replaced. PCBs that have run offthe equipment onto the concrete or soil below should be cleaned up and stored pending disposal.Personal protective equipment and cleanup materials may be contaminated with PCBs and requireproper disposal.

It is a good idea to replace and properly dispose of PCB-containing equipment that is no longerneeded, keeping in mind this can be a good long-term investment to avoid liability. Storage shouldbe in a building with an adequate roof and walls that is in a location selected to protect the PCBsfrom the possibility of release. Storage facilities should not be in a flood plain. Leaking equipmentshould be stored in metal drums with lids. Containment should prevent escape of PCBs into theenvironment through volatilization and containers should carry PCB marks. PCBs will penetratemost plastics.

12 BE AWARE OF HIDDEN SOURCES OF PCBS

The largest single hidden PCB source resulting in improper disposal is transformer bushings.The dielectrics in bushings have no fluid connections with the dielectrics in the transformers towhich they are attached so analysis of the transformer dielectric will not reveal anything aboutPCBs in the bushing. Bushings can contain anything from mineral oil to pure aroclor to tar-likecompounds containing very high concentrations of PCBs. Pot heads – cable termination apparatusthat connects transformers to incoming power sources – can be filled with a tar-like material,

281

which can contain very high concentration PCBs. Any tar-like or asphalt-like material used as aninsulator or dielectric should be suspected of containing PCBs. Small motors often require startingcapacitors that can contain PCBs. Voltage regulators and substation transformers can contain loadtap changers operated by small motors that contain PCB starting capacitors. Small motor capacitorscan leak contaminating the dielectric fluid. Asphalt material in fluorescent light ballasts, along withlubricants and caulks are other potential sources. Air compressors have been serviced with PCBcontaining lubricants. Oil-filled switches, circuit breakers, and enclosures should also be suspect.

It is of utmost importance to keep in mind the dangers and persistence of PCBs in the environmentwhen deciding on storage locations and disposal. PCB-containing dielectrics require specializeddisposal techniques that destroy the PCB molecule. Incineration of PCB-containing dielectrics is thepreferred destruction method. However, the PCB regulations require 99.9999 percent destructionof the PCBs (USEPA.2010) and incinerators can burn PCBs only if they obtain an EPA PCBdisposal approval which includes successful completion of a trial burn to demonstrate this level ofdestruction. Inefficient incinerators or open burning can vaporize and disperse PCBs and convertthem to even more hazardous dioxins. Disposal of PCB-containing dielectrics in landfills is notpermitted because of the potential for ground water contamination. If there are no adequate disposalfacilities, long-term storage will be your only option.

The extent and complexity of underground mines present opportunities for abandonment orillegal disposal of hazardous wastes. The presence of hazardous wastes may not be evident untilthey are found in the local ground water. PCBs released underground from abandoned electricalequipment can cause water pollution in mining districts which will eventually introduce PCBs intothe environment leading to contamination of the ocean and the human food chain regardless of thelocation of their release.

Abandoned underground electrical equipment may remain intact and not release PCBs for a verylong time. Testing waters issuing from abandoned mines may not indicate whether or not PCBs arepresent in intact electrical equipment.

13 CONCLUSION

PCBs are hazardous not only to human health and the environment but also to the mining industrybecause of potential worker exposure and improper disposal liabilities. Risk management should bean important part of PCB management planning. PCB managers should take into account not onlythe regulatory requirements for use and disposal, but also financial risks that can result if PCBsare inadvertently released into the environment. Compliance with the regulations will not protectagainst liabilities under other federal laws while minimal compliance can result in unanticipatedrisks. For example, the regulations permit the disposal of fluorescent light ballasts containing smallcapacitors with about an ounce of pure aroclor as municipal solid waste. Nevertheless, the preambleto the regulations states: However, disposers of fluorescent light ballasts that contain a small PCBcapacitor should be aware that they could be subject to Comprehensive Environmental ResponseCompensation and Liability Act (CERCLA) liability if the municipal solid waste landfill becomesa Superfund site (63 FR 35384, 35409). This small thimble-sized capacitor contains enough aroclorto contaminate 166 gallons of mineral oil at 50 ppm.

In the United States, PCB contamination can result in liabilities under CERCLA at any con-centration if there is an actual or threatened release of PCBs to the environment which presentsan imminent and substantial endangerment to public health or the environment. Because PCBsare considered hazardous substances under CERCLA, USEPA can also act whenever there is anactual or threatened release even in cases where there is no imminent and substantial endangerment(CERCLA.1980).

Replacement and proper disposal of PCB-containing dielectrics can protect you from futureliabilities.

Dan Bench, a mining engineer, is the U.S. Environmental ProtectionAgency Region 8 PCB Coor-dinator (E-mail: [email protected] or 303-312-7090, the Mining Hotline). For the regulationsand more information on PCBs, check www.epa.gov/pcb.

The views in this article express the opinions of the author and do not necessarily reflect EPApolicies.

282

REFERENCES

CERCLA. 1980 Section 104(a), 42 USC. Sec. 9604(a).Eisler, R. 1986. Polychlorinated biphenyl hazards to fish, wildlife, and invertebrates: a synoptic review. In

Biologic Report 85; Patuxent Wildlife Research Center, US Fish and Wildlife Service. Laurel, MD: pp. 7,14.Gioia, R. 2008. Polychlorinated biphenyls (PCBs) in air and seawater of the Atlantic Ocean: sources, trends

and process. Environmental Science and Technology. Vol 42 No 5: pp. 1416–22.Goddard, C. et al. 2003. Preliminary report on the sperm whale data collected during the voyage of the Odyssey,

Ocean Alliance, Lincoln, MA and Woods Hole Oceanographic Institute, Woods Hole, MA.http://earthobservatory.nasa/gov/Newsroom/NewImages/images.php3?img id=17405. 2008. Nine years of

ocean chlorophyll.Personal communications from company officials during inspections, 1978–1994.Smetacek, V. & Cloern, J.E. 2008. On phytoplankton trends. Science. March 7 Vol. 319: pp. 1346–48.Stockholm Convention on Persistent Organic Chemical (POPs) website. 2008.US Department of Health and Human Services, Public Health Service, Agency for Toxic Substances and

Disease Registry. 2000a. Toxicological profile for polychlorinated biphenyls: p. 2.US Department of Health and Human Services, Public Health Service, Agency for Toxic Substances and

Disease Registry. 2000b. Toxicological profile for polychlorinated biphenyls: p. 4.US Department of Health and Human Services, Public Health Service, Agency for Toxic Substances and

Disease Registry. 2000c. Toxicological profile for polychlorinated biphenyls: p. 493.US Department of Health and Human Services, Public Health Service, Agency for Toxic Substances and

Disease Registry. 2000d. Toxicological profile for polychlorinated biphenyls: p. 469.Federal Register, June 29.1998. 63 FR 35384,35403.Federal Register, June 29.1998. 63 FR 35384,35404.Food and Drugs. 2010. 21 CFR Part 109.30. Tolerances for polychlorinated biphenyls (PCBs).Mineral Resources. 2010. 30 CFR . Secs. 75.2 Permissible, and 75.340 Underground electrical installation.Protection of Environment.2010. 40 CFR Part 761.60. Testing procedures (g)(1)(iii).Protection of Environment.2010. 40 CFR Part 761.70. Nonliquid PCBs (b)(1).US Environmental Protection Agency, Office of Pollution Prevention and Toxics.1997. Management of

polychlorinated biphenyls in the United States. In Health and environmental effects of PCBs.Sec. 1.4.

283


Waste management practices at Alaska’s large mines

J. DiMarchi & J. VohdenState of Alaska, Department of Natural Resources, Fairbanks, Alaska, USA

ABSTRACT: Alaska has four large mines that include the world’s largest zinc mine, secondlargest silver mine, and two gold mines with a combined annual production of 689,000 troy ouncesgold. Two of these mines are open pit mines and two are underground mines. These mines contributeto a cumulative total of 22 Mtonnes of waste rock and 18 Mtonnes of tailings each year that haveto be managed during life-of-mine and eventually reclaimed as part of mine closure.

Each mine employs waste management practices suited to their physical environment, wasterock geochemistry and ARD/ML potential, and the approved closure plans. The constituents ofconcern and the ARD/ML potential vary widely between the mines as a function of the differencesin the geochemistry of the ore deposits. Mine reclamation and closure plans vary between minesto address the unique site, tailings, waste rock and water quality conditions in ways that will attainand maintain the physical, chemical, and biological stability of each site for the long term, and caninclude the need for water treatment in perpetuity.

1 POGO MINE

The Pogo mine is located in InteriorAlaska, 90 km southeast of Fairbanks (Figure 1). It is accessibleby road and is connected to the power grid. It receives approximately 35 cm of precipitation perannum including 127 cm as snow. The mine site is underlain by discontinuous permafrost which isgenerally restricted to wetlands and north-facing slopes. The deposit is a gold-bearing quartz veinwith an original resource of approximately 9 Mtonnes at a grade of 16 grams per tonne (gpt) Au. Itis an underground mine and ore is mined at a rate of 2300 tonnes per day. The mine opened in 2006and is scheduled to close in 2016. However, ongoing exploration is likely to extend the life of themine. It produced 389,808 troy ounces in 2009. The mine, processing facilities, access road andpower line rights-of-way are situated on State of Alaska lands. The mine is owned and operated bySumitomo Metal Mining Pogo LLC, which is owned by Sumitomo Metal Mining Co., Ltd. (85%)and Sumitomo Corporation (15%).

Waste rock at Pogo includes blast rock removed from underground as development tunnelsare advanced for access to ore stopes and to lengthen haulage ways. The mine has a waste rocksegregation procedure that relies on mine lab x-ray fluorescence (XRF) analyses of blasthole drillcuttings that determines the arsenic and sulfur content of each round (a round is the volume ofbroken rock generated with each blast – approximately 181 tonnes). Waste rock is classified as“mineralized” if it contains more than 600 ppm As or 0.5% S. Mineralized waste rock must bepermanently placed on the dry stack or back underground. Nonmineralized waste rock can be usedfor constructing roads or other general purposes at the mine. In 2009, 1149 rounds of waste weretaken and sampled in accordance with the rock segregation procedure.

Of these, 375 rounds (33%) exceeded either the As or S thresholds and were placed internallyin the drystack. Average concentrations for mineralized waste rock placed in the dry stack in 2009were 322 ppm As, 0.225% S, 4178 ppm Fe with ratio of neutralizing potential to acid producingpotential (NP:AP) of 4. The mine estimates that it will generate 756,000 tonnes of mineralizedwaste rock over the life of the mine (Sumitomo Metal Mining Pogo LLC, 2010).

The gold ore is conveyed from the underground ore bin to the mill on the surface. The millcrushes and grinds the ore to 80% passing 50µ before subjecting it to gravity concentration andflotation. The flotation concentrate comprises 10% of the ore by weight, and is reground to 80%

285

Figure 1. Location Map of Alaska’s large mines.

passing 10µ. The flotation concentrate is then subjected to vat-cyanide leaching to extract the goldthat was not recovered in the gravity circuit. Waste streams from the mill include flotation tails, andCIP tails which are the flotation concentrate after it has been subjected to cyanidation. The flotationtails consist primarily of gangue (mostly quartz). They are not exposed to cyanide and 60% of thesetails are dried using a filter press to approximately 15% moisture before being trucked, dumpedand compacted in the dry stack tailings impoundment. The remainder of these tails is used to makepaste backfill. Average concentrations for flotation tails solids placed in the dry stack in 2009 were1116 ppm As, 0.125% S, 2440 ppm Fe with ratio NP:AP of 9.47 (Sumitomo Metal Mining PogoLLC, 2010). The CIP tails represent another waste stream from the mill and account for 10% of theore by weight. These tails consist of the flotation concentrate that has been subjected to cyanidationfollowed by a SO2/air cyanide destruction process. This waste is pumped to the paste fill plantwhere it is combined with approximately 40% of the flotation tails and approximately 7% cementto make paste slurry. This slurry is pumped underground and used to fill the exhausted stopes. Thepaste cures over a 28-day period into a semi-competent solid mass. CIP tails are sampled dailyfor WAD CN to ensure that they meet State permit standards. Ninety percent of samples mustcontain <10 ppm and 100% <20 ppm WAD CN. In 2009, 100% of samples contained <10 ppmWAD CN. The procedures have been formulated to meet the specific metallurgical characteristicsof the ore but also meet the following environmental objectives: 1) minimizing the amount ofsulfides and arsenic mineralization in the surface drystack facility, 2) minimizing the volumeof cyanide-contacted material, 3) ensuring that cyanide solutions are detoxified to the greatestpractical extent, and 4) ensuring that all cyanide-contacted material is permanently entombed inpaste fill, underground.

The mine closure plan is designed to return the site to a stable condition including stabilization toprevent erosion, encourage revegetation and prevent any chemical degradation of the surface wateror groundwater. The closure is subdivided into 5 stages that include reclamation of constructiondisturbances, concurrent reclamation, final closure and reclamation, post closure reclamation (10year duration) and finally, post closure monitoring (20 year duration).

At closure the drystack will be covered with an engineered soil cover. The engineered soil coverwill consist of 0.3 m of non-mineralized waste rock applied over the surface of the crowned tailings,followed by a 15 cm sand and gravel layer to provide support for an additional 15 cm of growth

286

media (TeckCominco, 2003, Pogo Project Reclamation & Closure Plan Update). This soil coverhas been designed in view of the relatively modest annual rainfall at the site, the low hydraulicconductivity of the compacted drystack tailings, and the lack of acid rock drainage (ARD) potential.After the soil cover is in place, a system of groundwater monitoring wells will be installed near thedownstream toe of the drystack. These wells, together with 10 years of monitoring of the surfacewater flow between the drystack and the RTP, will provide the information required to assess theeffectiveness of the closure design to protect water quality for the long term.

The only waste rock stockpile at Pogo was developed during the advanced exploration programin 2000. At closure, that waste rock will be entombed in the drystack, below under the engineeredcover. The stockpile site will be reclaimed to original topography, covered with growth media andrevegetated.

The underground mine stopes are routinely filled with paste backfill. At closure, the main haulageways will be compartmentalized to minimize hydraulic conductivity. The mine will be flooded andall three adits will be permanently stabilized and sealed using a combination of select paste backfillplacement and concrete plugs to prevent access and drainage.

After all the facilities that are not required for long term post closure activities are removed, andthe site is in a stable condition, the project will enter into the 10 years of Phase IV post closurereclamation. During this phase the RTP and water treatment plant will remain in place as long asneeded to treat the drystack runoff and seepage consisting of seasonal water treatment of the RTPwater. By phase V, all surface disturbances will be stabilized and water quality will be acceptable.Post-closure monitoring of groundwater, stormwater, and surface water will continue for a 20-yearperiod, which together with the 10 year monitoring period under Phase IV, results in an estimated30 year monitoring period after closure of the mine.

The Pogo mine has provided the State with a bond in the amount of $27.6M to cover the costs ofmine reclamation and monitoring. That amount is presently under review, and expected to increase,as part of the State’s renewal process for the mine’s Reclamation Plan Approval.

2 FORT KNOX MINE

The Fort Knox Mine is located in Interior Alaska, 30 km northeast of Fairbanks (Figure 1). It isaccessible by public road and is connected to the power grid. The area receives approximately 36 cmof precipitation per annum including 127 cm as snow. The mine site is underlain by discontinuouspermafrost which is generally restricted to wetlands and north-facing slopes. The gold depositconsists of a network of structurally controlled quartz veins hosted within granite. The mine is anopen-pit operation and ore is trucked from the pit at a rate of approximately 82,000 tonnes perday; half of which is milled and the remainder is subjected to heap leaching. The mine produced263,260 ounces in 2009. Pre-mining global reserves were 332 Mtonnes of which approximately105 Mtonnes remain to be mined. The average grade of proven and probable reserves is 0.45 gptAu. The mine opened in 1986 and pit operations are scheduled to cease in 2016, although heapleach operations will continue until 2021. The mine is located on State of Alaska and mine-ownedprivate land, and is operated by Fairbanks Gold Mining Inc. which is 100% owned by KinrossCorporation.

Waste rock from the mine includes overburden known as Fairbanks schist, and granite from thepit that is barren or contains sub-economic concentrations of gold. Approximately 50,000 tonnesof waste are mined and placed on the dumps per day. This waste is stored in four dumps adjacent tothe pit. The waste rock is sampled quarterly for acid-base accounting (ABA). Typical ABA resultsfor the waste rock consistently show high NP:AP ratios due to the AP values of <1 and NP valuesas high as 69 (Fairbanks Gold Mining Inc, 2007b). Nonetheless, runoff from the dumps reports tothe tailings storage facility which is a zero discharge facility. At closure the dumps will containapproximately 181 Mtonnes of waste rock.

The Fort Knox mill crushes approximately 40,000 tpd of ore and uses a vat-cyanide leach processto extract the gold. Waste streams from the mill include coarse reject from the crushing circuit andtailings. The coarse reject is +3 inch barren granite that is conveyed to stockpiles and used forconstruction at site. The tails are pumped via slurry to the tailings storage facility (TSF). TheTSF is a dammed facility. The dam is a cross-valley, earth- and rock-fill structure with seal andfilter/transition zones. The dam has been raised three times and is in the process of being raised an

287

additional 16 m to a height of 115 m with a crest elevation of 473 m above mean sea level (MSL).The pH for tailings ranges from 9 to 10 and the permit limitations are 6 to 11. The tails contain CNand the waste management permit stipulates monthly average effluent limits of 10 mg/L WAD CNand an instantaneous limit of 25 mg/L. Typical tailing solution water reporting to the TSF facilitycontains between 8 and 9 mg/L WAD CN. Quarterly ABA analyses for the tailings solids showsthat NP:AP ratios are very high (30–50) and not likely to produce acid (Fairbanks Gold MiningInc, 2007). The TSF is a “zero discharge” facility. Dam seepage is intercepted by 8 interceptorwells situated at the toe of the TSF dam which also maintain a cone of depression in the watertable at the toe of the dam. The current pump back rate is approximately 0.11 m3/s. The chemistryof the pump back water is monitored quarterly. Eight monitoring wells situated down gradientare sampled quarterly to ensure that the interceptor system is functioning properly and there is noseepage into the surface water or groundwater regimes below the TSF dam.

In 2009, Fort Knox completed construction of Stage I of the Walter Creek heap leach facility.The facility is designed to allow the extraction of gold from low grade run-of-mine ore. The minecommissioned the heap facility in October, 2009. The entire facility (Stages I–V) will be linedwith 80 mil HDPE. Stage I of the facility will impound the solution pond and is double lined andincludes a leak collection and recovery system (LCRS). Cyanide solution is applied to the heapat a rate of approximately 0.50 m3/s. The pregnant solution is pumped from the in-heap pond at asimilar rate to the carbon-in-column circuit in the mill where the gold is removed from solutionand the CN is recycled back into the leach process. Approximately 11 Mtonnes of low grade ore isplaced on the heap annually and approximately 163 Mtonnes of low grade ore will eventually betreated on the heap pad through 2021, prior to closure.

The reclamation and closure plan for the mine is designed for long term stability of the siteincluding the waste dumps, tailings and heap leach facility, and the pit lake. The State is currentlyreviewing the closure and reclamation plan including the bond amount as required by the recurring 5yr permit renewal process. Reclamation of the waste rock dumps will include contouring, placementof 30 cm of growth media, and revegetation.

During the initial closure period, the TSF seepage interception system will continue to operate,and the seepage water will be pumped to the pit as the tailings decant pool is dewatered. During thistime sufficient storage capacity will be available to contain the 100-year, 24-hour storm event plusspring runoff, while maintaining the required freeboard. Currently, the TSF seepage does not meetdrinking water quality standards with regard to nitrate and sulfate. At closure, the remaining decantpond will be pumped to the open pit. The surface of the tails will be reshaped and covered withclean soil and revegetated. A pond and or stream will be reconstructed to restore some jurisdictionalwetlands, and a permanent spillway will be constructed in the dam to allow water to pass throughthe facility, seasonally. Based on the current closure plan it would take 12 years for the water pondbehind the dam to fill naturally to an elevation where seasonal discharges through the spillwaywould occur. The downstream face of the tailings embankment is constructed of durable rock, andis resilient to erosion and will not be capped with growth media.

After closure, the pit lake will begin to fill naturally as the dewatering wells around the pitare decommissioned. It will take approximately 80 years for the pit lake to recharge naturallyto its maximum water elevation of 448 m MSL. It is anticipated that the pit lake will meet waterquality standards, except for manganese and iron which are naturally high, once it attains maximumelevation. The closure plan calls for an annual pit lake evaluation to: 1) summarize site conditions,2) update the water balance, 3) validate the reclamation approach from a water quality perspective,and 4) validate pit lake water quality predictions based on TSF decant volume and pit wall runoffwater quality and identify an appropriate water treatment/management if required.

The Fort Knox heap leach facility will be active for at least 5 years following closure of themill. This is because the heap leach process generally continues to yield gold for several yearsfollowing the cessation of placement of ore on the heap pad. Reclamation of the heap leach facilitywill include the following: 1) residual leaching until the treatment gold recovery diminishes to thepoint that it becomes uneconomic, 2) solution recirculation/rinsing to destroy cyanide and meetcompliance standards, 3) release of drain down to the tailing impoundment, 4) release of minorlong-term seepage to the tailing impoundment, and 5) regrading and cover.

The facility will be regraded to an overall 3H:1V slope and covered with growth media. Theregrading design will include erosion control measures as necessary to avoid loss of growth media.After closure, monitoring systems for process components will remain in place until each specific

288

facility has been chemically stabilized to the satisfaction of the State. This is anticipated to be onthe order of 10 years. The long term monitoring will occur down gradient of the facility at thesurface water and groundwater monitoring points established as part of the tailing impoundmentclosure plan (Fairbanks Gold Mining Inc, 2007a).

The Fort Knox mine has provided the State with a bond in the amount of $37.6M to cover the costsof mine reclamation and monitoring. That amount is increased annually to reflect the Anchorageinflation index. In addition the closure plan will be reviewed by the State of Alaska large minepermit team during 2010 and the bond amount will likely be amended as part of that review as well.

3 RED DOG MINE

The Red Dog Mine is located in northwest Alaska on the south flank of the Delong Mountains, asub range of the better known Brooks Range (Figure 1). The mine is situated at latitude 68◦4′9′′Nin the Arctic and north of tree line and is underlain by continuous permafrost. The average annualtemperature is −6◦C and average precipitation is 23 cm including 122 cm as snow. The mineproduces Pb, Zn and Ag from a one of the largest accumulations of these metals on the planetand is the largest producer of zinc in the world with production of 516,200 tonnes Zn and 122,600tonnes Pb and 7.5M ounces Ag in 2008 (Szumigala et al., 2008). The mineralization consists ofstrataform galena and sphalerite in two deposits of about 54 Mtonnes each. The Main depositis nearly exhausted and the company is stripping overburden from the second deposit, known asAqqaluk. Aqqaluk will be mined for approximately 20 years at the current milling rate of 9800tonnes per day. The mine opened in 1989 and is scheduled to close in 2031. The mine is situatedon NANA Corporation Native lands and NANA also owns the mineral estate. Teck Corporationoperates the mine through an agreement with NANA.

Waste rock from the mine includes relatively barren overburden that is stockpiled and can beused for construction of roads, the tailings dam, and other uses at the mine. Most of this waste isstockpiled for future use in constructing covers at closure over the more mineralized waste rockdumps. The mine has stockpiled approximately 9.9 Mtonnes of overburden. The majority of wasterock contains anomalous but sub-economic concentrations of Pb, Zn and Ag plus other metalsincluding Cd. This waste is placed on the main waste stockpile for permanent storage. The mainwaste stockpile covers approximately 77 ha and contains approximately 32 Mtonnes of waste andwill contain approximately 60 Mtonnes and cover an area of approximately 111 ha at closure in2031 (Teck, 2009). The majority of the waste rock in the main waste stockpile weathers rapidlyand either already generates acid or has the potential to generate acid. As a result, water runoffand seepage from the waste dumps are collected in a series of ditches and sumps, respectively anddirected to the TSF, or to the water treatment plant and discharged into Red Dog Creek. The ditchesand sumps collect approximately 600,000 m3 per annum from the main waste sumps, overburdenstockpile and the low grade ore stockpile.

The Red Dog mill crushes and grinds the ore to sub-65µ particle size before subjecting it to aflotation treatment to produce separate Zn and Pb-Ag concentrates. Mill throughput is approxi-mately 9800 tonnes per day producing 1.25 million tonnes of concentrate and 2.3 million tonnesof waste as tailings, per annum (Teck, 2010). The tailings are pumped as a slurry to the TSF. TheTSF is a dammed facility consisting of a rock-fill dam with engineered seal and filter zones. Thecrest of the dam has been raised 4 times utilizing downstream raises. The upstream face of thedam is lined with a 100 mil HDPE geomembrane. The tailings pond holds an average of 3.0 Mm3 of free water which covers the tailings. The volume fluctuates seasonally with highest inflowsin the spring when about 3.8 M m3 enters the facility. The mine treats and discharges up to 7.6 Mm3 of water per year under the National Pollution Discharge Elimination System (NPDES) permitwith concurrent Clean Water Act (CWA) Section 401 certification by the State. The TSF water haspH of approximately 5. The average total dissolved solids (TDS) concentration is approximately4200 ppm, predominantly as sulfate. Zinc concentrations in the TSF water average 350 ppm whileFe concentrations average less than 10 ppm. The tailings solids contain approximately 4% Zn, 2%Pb, 8% Fe and 12% total S. The NP:AP ratio is low and the tailings are strongly acid generating(Teck, 2009). Maintaining the water pond over the tailings in the TSF is an effective means ofminimizing oxidation and acid generation (Teck, 2009).

289

The TSF is a “zero discharge” facility and a pump back system, consisting of a collection galleryand pumps at the toe of the dam and a well situated 76 m downstream from the toe. The pump backsystem collects and pumps seepage back to the TSF at a rate of approximately 0.05 m3/s (Teck,2010). A series of monitoring wells situated down gradient from the pump back system are sampledquarterly to confirm there is no TSF seepage advancing beyond the pump back system

The principal focus of the closure plan is the long term stability of the site. The three majorcomponents of the plan are: 1) encapsulating the mineralized waste rock dumps, 2) maintainingthe wet cover on the TSF, and 3) performing long term water treatment. The mine is on private(Native) land and the closure plan was developed over an 8 year period working closely with thelandowners and regulatory agencies to formulate a plan that met the requirements of the agenciesand embraced the landowner’s long term plans for the surface use.

The mine will exhaust ore in the Main pit in 2011 as it shifts its production to the adjacentAqqaluk deposit. As part of the mine plan, waste rock from Aqqaluk will be used to fill the Mainpit and create a waste dump atop the pit. Eventually 60 million tonnes of waste will be generatedfrom Aqqaluk. The reclamation plan calls for the waste rock dumps to be covered with engineeredsoil covers after the dumps are reshaped to a 3H:1V slope. Reclamation will be concurrent withongoing mining to the extent possible. The soil covers will consist of a lower compacted layer ofweathered Okpikruak and Kivalina shale approximately 46 cm thick and an upper uncompactedlayer of the same thickness. Test work supports the concept that the engineered cover will reducesurface water infiltration into the waste rock by up to 70% (Teck, 2009). The soil covers will beseeded with a grass mixture for short-term stabilization while native species become established.The mine is situated north of the tree line so native cover consists of an assemblage of grasses andforbs. The State of Alaska performance standard for revegetation is 40% cover in 5 years. Fieldtrials initiated in 2007 will also generate useful information that will be used to modify the coverand revegetation plans as warranted.

The closure plan for the TSF is a “wet” closure. A number of streams that are currently divertedaway from the TSF will be rechanneled into the TSF and the water pond will be maintained overthe tails in perpetuity. In the long term, water quality in the TSF is expected to improve owing tothe seasonal inflows of clean surface runoff and the cessation of the current practice of pumpingimpacted mine water (from sumps and mine surface runoff) into the TSF. A 183 m wide beach willbe constructed on the upstream side of the tailings dam prior to closure to minimize seepage oftailings water through the dam. The beach will be covered with non-mineralized soil cover andrevegetated. The beach should be effective at reducing seepage through the dam by 75%, therebyreducing the total volume of water that will require water treatment after mine closure (Teck, 2009).

Long term water treatment will be required for TSF water and impacted water accumulating inthe Aqqaluk pit. Currently all process and impacted surface and ground water is pumped to theTSF or the water treatment plant. At closure all impacted water will be pumped to the Aqqaluk pitwhich will serve as the primary water storage facility for the long term. Water from the Aqqaluk pitwill be treated and discharged seasonally as necessary. The mine reclamation plan estimates that5.3 M m3 of water will require treatment and discharge seasonally.

The mine has posted a bond in the amount of $305M to cover the cost of reclamation and closureand long term water treatment.

4 GREENS CREEK MINE

The Greens Creek Mine is located on Admiralty Island in Southeast Alaska, 29 km southwest ofJuneau (Figure 1). The mine is operated by Hecla Greens Creek Mining Company, which is ownedby Hecla Mining Company. It is accessible by boat and float plane although there are 13 milesof road on-site. The site itself includes the underground mine workings, ore concentrating mill,tailings impoundment, camp facility, ship loading facility and a ferry dock at tidewater. Employeesat the minesite commute daily via ferry from Juneau. The area receives an average of 135 cm ofprecipitation per year, primarily in the form of rain. (US Forest Service, 2003). Average annual airtemperature is approximately 6◦C, ranging from −13◦C to 21◦C degrees.

The orebody was discovered in 1975. Full-scale development was initiated in 1987 and the firstconcentrate was shipped in 1989. Low metal prices forced the closure of the operation in 1993but it was reopened in 1996 and remains open today. The mine will could close in 2015, except

290

ongoing exploration is likely to extend the mine life. The orebody contains Ag, Zn, Au and Pb. Thedeposit is a polymetallic, stratiform, massive sulfide. Mineralization occurs discontinuously alongthe contact between a structural hanging wall of quartz mica carbonate phyllites and a structuralfootwall of graphitic and calcareous argillite. Major sulfide minerals are pyrite, sphalerite, galenaand tetrahedrite. The mine is an underground operation utilizing cut and fill as well as longholestoping mining methods, which produces approximately 1900 tonnes of ore per day. The millproduces doré of Ag and Au as well as concentrates of Pb and Zn.

The Greens Creek property is situated on patented mining claims, US Forest Service land andland within Admiralty National Monument. The patented claims will revert to monument statusafter the mine is closed and reclaimed.

Waste rock at Greens Creek includes segregated production rock from underground operationsand is disposed of at Production Rock Sites 23 and D.Acid base accounting shows that portions of theproduction rock in both sites have the potential to generate acid. However, carbonate minerals in theproduction rock provide substantial buffering and a resulting long lag time to acid generation. Thisallows for the construction of a cover system which will limit oxygen and precipitation infiltrationat closure. The cover system is comprised of segregated glacial till, colluvium and alluvium whichare not acid-generating.

Limited covers have been constructed as part of concurrent reclamation. Metal concentrationsin the production waste rock are elevated and lead to metal leaching into contact water. Contactwater is collected in sumps, treated and discharged. Surface and groundwater monitoring show alack of elevated metals and establish the effectiveness of the waste management practices at themine.

Segregation of waste rock is carried out on a visual basis by a mine geologist. Visual characteris-tics have been shown to accurately reflect the geochemical properties of the waste rocks includingacid generation potential. Four classes of waste rock are recognized based on net neutralizingpotential (NNP):

Class 1: NNP greater than 100 tonnes CaCO3/1000 tonnesClass 2: NNP between +100 and −100 tonnes CaCO3/1000 tonnesClass 3: NNP between −100 and −300 tonnes CaCO3/1000 tonnesClass 4: NNP less than −300 tonnes CaCO3/1000 tonnes

Class 1 rock has no use restrictions or special handling requirements. Class 2 and 3 must go onthe waste stockpile or underground, with Class 3 being placed underground to the greatest extentpractical. Disposal of Class 4 production rock must be placed underground. (Greens Creek MiningCo., 2004)

Tailings generated by the milling process are dewatered to approximately 12.5% water usinga filter press and transported by truck to a drystack tailings facility. Fifty percent of the tailingsare placed underground as backfill after being mixed with cement; the remainders are placed inthe drystack. The tailings have an abundance of sulfide sulfur, generally greater than 10%, butalso have a high neutralizing component. Paste pH values are usually above 6.5 with the majoritybetween 7.5 and 8.5. The average NP:AP for the tailings is 0.74 (Hecla Greens Creek Mining Co.,2009). Tailings characterization and quarterly monitoring are ongoing. Approximately 447,000tonnes of tailings are placed in the drystack annually. There are currently 5 Mtonnes of tailings inthe drystack (Kennecott Greens Creek Mining Company, 2004).

ARD potential is minimized in the drystack by grading and compacting the tails to promoterunoff and discourage infiltration. Water management at the tailings facility consists of a complexnetwork of drains under the tailings, bentonite slurry walls around the perimeter of the site, andditches to divert upslope water and collect surface runoff. The site is underlain by a low permeabilitytill and other glacial/marine deposits. Contact water is collected, treated and discharged to the oceanunder an NPDES permit.

A sulfate reduction monitoring program was initiated in 2004 to determine the feasibility ofusing a carbon amendment within the drystack to promote microbial reduction of sulfates and theprecipitation of metal-sulfide minerals. This treatment could decrease the sulfate and metal loadingwithin the tailings pore water and would result in a shorter closure period during which water wouldrequire treatment prior to marine discharge. Ongoing trials with amendments such as peat, spentbrewing grain and municipal biosolids continue (Hecla Greens Creek Mining Company, 2010).

291

The reclamation plan for the mine is designed to return the site to a near-natural condition and tominimize the effects of mining. The plan is in compliance with all relevant US Forest Service, the USArmy Corps of Engineers, the City and Borough of Juneau, and the State of Alaska requirements.The tailings drystack will be contoured and an engineered cover will be constructed. This willinclude an 20–30 cm capillary break, a 61 cm compacted barrier layer, a second capillary layerwith filter fabric on top, and a 61 cm growth medium layer, which will be seeded and planted (USForest Service, 2003). It is anticipated that it will take approximately 7 years for the runoff from thereclaimed facilities will meet surface water quality standards. Once they meet these standards therunoff will be permitted to flow along its natural courses into the ocean. The site will be monitoredfor up to thirty years on a declining schedule (Kennecott Greens Creek Mining Company, 2000).The current financial assurance is a surety bond issued to the US Forest Service in the amount of$26.2M. The closure plan and related bond are under review at this time (US Forest Service, 2007).

5 CONCLUSIONS

Alaska’s four largest mines operate under State and Federal permits. The waste management prac-tices were developed to address the specific characteristics of the mine waste streams in order tomeet those permit stipulations. The identification and characterization of the waste streams typi-cally begins years in advance of mine permitting, as part of the mine’s environmental baseline andpre-feasibility studies. The waste management practices are codified in the mines Plan of Opera-tions which are approved by state agencies prior to the initiation of mine construction. In addition,management practices for waste streams are monitored, and modified where necessary, throughoutthe life of these mines. The waste management strategy has the objectives of both minimizing impactto the natural environment during operations and supporting an effective mine closure. Designingmines for closure and performing concurrent reclamation are considered standard industry prac-tices in Alaska. Alaska’s largest mines are long-lived, and this affords the operators and regulatorswith the opportunity to refine the closure and reclamation plan prior to mine closure. Recurringpermit renewals on 5-yr cycles combined with third party compulsory environmental audits alsoprovide timely opportunities to revaluate the adequacy of the waste management practices, and theclosure plans including the financial assurances associated with the closure plans.

REFERENCES

Fairbanks Gold Mining Inc, 2007a, Fort Knox Gold Mine Reclamation & Closure Plan, Unpublished CompanyReport, 128 p.

Fairbanks Gold Mining Inc, 2007b, Fort Knox Gold Mine Solid Waste Permit #0031-BA008 Quarterly Report,Second Quarter 2007, Unpublished Company Report, 71 p.

Fairbanks Gold Mining Inc, 2010, Fort Knox Mine 2010 Annual Activities Report, Unpublished CompanyReport, 27 p.

Greens Creek Mining Co., 2004, Site23-D Hydrogeology and Geochemistry Analysis, prepared byEnvironmental Design Engineering, Unpublished Company Report, 91 p.

Greens Creek Mining Co., 2009, Environmental Audit of the Greens Creek Mine-Final Report. Submittedto Sate of Alaska Department of Environmental Conservation, State of Alaska Department of NaturalResources, the United States Department of Agriculture Forest Service, Hecla Greens Creek MiningCompany, Unpublished Company Report, 225 p.

Hecla Greens Creek Mining Company, 2010, Tailings and Production Rock Site, 2009 Annual Report,Unpublished Company Report, 169 p.

Kennecott Greens Creek Mining Company, 2004, General Plan of Operations, Appendix 3: TailingsImpoundment, Unpublished Company Report, 43 p.

Kennecott Greens Creek Mining Company, 2000, General Plan of Operations, Appendix 14: ReclamationPlan, Unpublished Company Report, 39 p.

Sumitomo Metal Mining Pogo LLC, 2010, Pogo Mine 2009 Annual Activity and Monitoring Report,Unpublished Company Report, 32 p.

Szumigala, D.J., Hughes, R.A., Harbo, L.A., 2008, Alaska’s Mineral Industry 2008, State of Alaska Divisionof Geological & Geophysical Surveys Special Report 63, 90 p.

TeckCominco, 2003, Pogo Project Reclamation & Closure Plan Update, Unpublished Company Report, 48 p.

292

Teck, 2010, Red Dog Mine 1st Quarter Report 2010 for State of Alaska Waste Management Permit No.0132-BA002 Reclamation Plan Approval F20099958, Unpublished Company Report, 23 p.

Teck, 2009, Red Dog Mine Closure and Reclamation Plan, Prepared by SRK Consultants, UnpublishedCompany Report, 66 p.

US Forest Service, 2003, Greens Creek Tailings Disposal Final Environmental Impact Statement, Report#R10-MB-482a, 403 p.

US Forest Service, 2007, FS Agreement No. 07MU-11100500-059. Memorandum of understanding betweenthe United States Department of Agriculture Forest Service, The State of Alaska Departments of Environ-mental Conservation and Natural Resources and Kennecott Greens Creek Mining Company ConcerningReclamation/Closure Bonding and Periodic Cooperative Audits for the Greens Creek Mine, 10 p.

293


East Mission Flats Repository design—challenges and case history

Donald K. Vernon, Jr.TerraGraphics Environmental Engineering, Inc., Boise, Idaho, USA

Andy MorkIdaho Department of Environmental Quality, Boise, Idaho, USA

ABSTRACT: The East Mission Flats Repository was designed to dispose of metals contaminatedsoils and debris wastes. Several of the engineering design challenges included repository confi-guration to reduce visual impacts; access analysis to reduce impacts to local roads; stormwatermanagement integrated into the phased construction to retain sediment; slope stability for wasteconfiguration and placement; evapotranspiration cover to reduce precipitation infiltration throughthe waste; and groundwater quality protection.

1 NEED FOR EMF REPOSITORY

United States Environmental Protection Agency (USEPA) and IDEQ have developed a Basin-wide Waste Management Strategy to guide waste repository siting and design to safely containcontaminated soils from the Superfund cleanup. Waste sources and quantities have been forecastbased on the interim Operable Unit 3 Record of Decision requirements and the Basin ICP. ThisWaste Management Strategy identified a Lower Basin repository as a high priority for the near-term.There is not enough room in the existing repositories to dispose of the wastes. In addition, there isno repository to serve the Lower Basin, where the Basin Property Remediation Program (BPRP)will soon be working. Also, the repository is needed to serve the community’s ICP requirements.Material disposed of at the East Mission Flats Repository will be generated by the BPRP and ICPactivities. After the EMF Repository is full, it will be closed and maintained in perpetuity.

2 PROJECT DESCRIPTION

The East Mission Flats Repository footprint is roughly triangular, covering an area of about 14 acres(TerraGraphics, 2009). The site is approximately 650 feet on the northwest side, 1,600 feet on thesouthwest side, and 1,350 feet on the northeast side. Based on this configuration, the estimatedtotal volume of the repository is approximately 445,000 cubic yards (cy). The cover volume isapproximately 30,000 cy and will be constructed out of clean materials. The side slopes of therepository will be made at a three foot horizontal to one foot vertical (3:1) slope. There will not beany side benches or “steps” except for a temporary road on the northeast side during construction.The top of the repository will be no higher than 2,165 feet, to reduce visual impacts. The top willbe peaked and sloped at 3% from the top to where it meets the 3:1 side slope perimeter to blend inwith the surrounding landscape. The configuration of the repository is depicted in Figure 1.

Idaho Department of Environmental Quality (IDEQ) has acquired a 23-acre parcel of land for thepurpose of providing a repository for wastes generated in the Coeur d’Alene Basin. The property,referred to as the “East Mission Flats Repository” or EMF, is located adjacent to Exit 39 on I-90and Canyon Road, two miles west of Cataldo, Idaho. The site is located approximately one-quartermile northeast of Old Mission State Park, north of the Coeur d’Alene River.

Waste placement will commence at the westernmost corner near Dredge Road and proceed to thenortheast along the northwest boundary. Once an advancing waste pile/slope has been established

295

Figu

re1.

along the northwest boundary of the repository footprint, placement will proceed southeastwarduntil a base pad at elevation 2,140 feet is established. This pad will facilitate a dry work area abovethe average yearly flood elevation. With this pad in place the waste pile will advance from theDredge Road access of the repository back to the Institutional Control Program (ICP) access to anelevation of 2,152 feet. Once the final waste placement has reached elevation 2,152 feet, the topwaste in the repository will be constructed to elevation 2,162.5 feet from the ICP Access end ofthe repository and proceed to the northwest boundary. Construction of the evapotranspiration (ET)cover will proceed from the ICP Access in the direction towards the Dredge Road access and willbe installed after the waste material for the top, as described in Phase 6, is completed.

Waste will be placed in 6- to 12-inch lifts and compacted. The waste placed within the interiorof the repository will be compacted to 90% of optimum while the waste placed at the perimeterwill be compacted to 95% of optimum. Slopes which are complete, that are part of the perimeterprotection embankment will have the protective layer installed prior to seasonal closure for the year.Figure 2 depicts the perimeter protection embankment, the construction pad, and the waste slopes.Exposed slopes that will not have perimeter protection installed will be stabilized for seasonalclosure by a spray-on soil stabilizer prior to winter closure. These design elements will providecontinual protection for the fill during the construction season as well as provide winter protectionfrom severe storms or flooding.

The use of waste material to construct the perimeter of the repository is based upon the following:

• Perimeter slopes will be constructed at a 3H:1V angle to encourage runoff and reduce the amountof time that runoff is in contact with the perimeter slopes.

• The perimeter protection will be installed over the waste and will cover the waste with a minimumof 21 inches of clean material and be installed on the repository prior to seasonal closure.

• The relatively level nature of the existing topography would make it very unlikely that runoffwould leave the site and discharge to surface water.

• Stormwater retention basins will be constructed to collect stormwater within the perimeter ofthe repository.

The perimeter slopes being constructed of waste material, protected from runoff by a minimum21 inch thick protective layer and the other aspects of the repository design provide an acceptablemeans of placing waste at the perimeter of the repository.

3 VISUAL ASSESSMENT AND IMPACTS

The repository will be a sloped mound of soil over most of the 23-acre site. Although the repositorysurface will be revegetated with native grass and shrubs to blend into the surrounding landscape, itwill be visible to users of the surrounding area, including visitors to the Old Mission State Park. Toportray the visual impacts to observers in the local area, visual simulations of the repository fromlocal view points were prepared.

The visual simulations were prepared from six view points. The view points included: two pointson the Old Mission State Park grounds; one point on the Exit 39 overpass; one point on the west-bound I-90 off-ramp at Exit 39; one point along Canyon Road northwest of the site; and one pointin the town of Cataldo, Idaho.

Results of the simulation indicated the site would not be visible from the Canyon Road orCataldo view points. The greatest visual impact would occur to observers on I-90 and on the Exit39 overpass. These observers would have a generally unscreened view of most of the repository.

Smaller impacts were noted from the two view points at the Old Mission State Park. One viewpoint was located near the Visitors Center, and the second view point was near the front steps of theOld Mission. The observation points were first photographed in October 2007, when the deciduoustrees were in full canopy. The October photographs show the repository would be totally screenedby the tree canopy at the Visitors Center, and a small fraction of the top portion of the repositorywould be visible as a minor background element from near the Old Mission front steps.

The two sites on the Old Mission grounds were re-photographed in December 2007, afterthe deciduous trees had dropped their leaves. Short portions of the top of the western end ofthe repository would be visible as a background element through the bare tree branches near the

297

Figu

re2.

Visitors Center. At the observation point near the Old Mission steps, the bare trees of Decemberafforded a slightly expanded view of a small portion of the top of the repository as a backgroundelement between a house and garage.

To address potential concerns regarding visual impact when viewed from I-90, more trees wereplanted between I-90 and the existing power lines. The existing trees in the buffer zone will beleft in place and continue to grow to reduce visual impacts from Canyon Road. Mitigation efforts,including enhanced vegetation screening by planting additional trees between I-90 and the EMFsite and revegetation of the repository slopes will reduce visual contrasts and help the repositoryblend in with the background. Therefore, the overall visual impact is expected to be small.

4 ACCESS ROADS

Once the BPRP is fully operational in the Lower Basin, the repository will be frequented by BPRPcontractor vehicles and other disposal contractor vehicles. As many as 100 trucks per day will haulwaste materials to the repository during a typical Monday through Thursday work week. Due tothis increased use, it is critical to select a safe, efficient, and cost-effective way to circulate thetruck traffic in and out of the repository.

All truck traffic coming to the EMF site will be routed to Exit 39 off of I-90. This is necessarybecause the load limit on the Old Bridge on Canyon Road in Cataldo does not allow for passage ofthe heavy trucks typically used by BPRP contractors. To reduce the expenses related with haulingBPRP waste along Dredge Road and Canyon Road, an access point was selected in coordination withthe Kootenai County East Side Highway District, the Idaho Transportation District, USEPA, andIDEQ. The selected entrance for the BPRP utilizes a small portion of Dredge Road, approximately145 feet after exiting I-90 at Exit 39. The repository will be accessed from Dredge Road by a singlespan precast concrete bridge. Users of the ICP Access Area will enter the repository via the CanyonRoad entrance, which consists of clean gravel and will place their load in the designated dumpingarea. By limiting the access to other portions of the repository and providing a clean gravel accessarea, ICP users will not be required to decontaminate their vehicles upon leaving the repository. Itis anticipated that minimal maintenance will be required to maintain clean access for users of theICP area.

The bridge section will span a total length of approximately 100 feet and be approximately 24feet wide. The width will allow for two 12-foot wide lanes, one for entering and one for exiting therepository. The height of the bridge from ground surface to the bottom of the structural memberswill be approximately 15 feet (bottom elevation to be 2,148.5 ft). This height will allow the watersof a 100-year flood event to flow freely beneath the structure with no interference due to the accessbridge. The elevation at the bottom of the bridge will be above the WSE of a 100-year flood eventas calculated by hydraulic modeling and determined to be 2148.5 ft with the repository in place.As per the Bridge Load and Resistance Factor Design Manual (ITD 2008), all bridges with a clearspan of 20’ or more must allow the flow of a 100-year flood event to pass below the lowest chordof the structure. Though this bridge would not be required to meet ITD standards, these standardswere used as the basis for design with regard to flow clearance.

5 STORMWATER MANAGEMENT SYSTEM

Phase 1 construction will consist of clean fill material and will not require a stormwater collectionsystem though the disturbed areas will have a silt fence installed for erosion and sediment control.During Phase 2, drainage from areas where waste is placed will be collected in temporary bermedareas that will be sized to hold runoff quantities from a 25-year 24-hour rainfall event. DuringPhases 3 thru 7, stormwater runoff will be collected in a retention basin located at the interior ofthe repository. The interior retention basin will be constructed during Phase 3 and will be modifiedas needed. Modifications will be specified in the annual placement and construction plan. Drainagefrom areas outside the waste placement areas (undisturbed areas) will not be collected. Stormwaterrunoff from the fill areas will be managed to reduce sediment and additional contaminant loading tonearby surface water bodies. Residual water will be held and used in the dry periods for use in dustsuppression. Disposal of excessive stormwater will be accomplished by pumping the stormwater

299

into a truck, hauling the stormwater to the CentralTreatment Plant in Kellogg, where the stormwaterwill be treated and disposed.

Waste slopes that will not have additional waste placed against them (final waste slope) willhave the perimeter protection layer installed prior to seasonal closure. Because the construction ofPhase 4 may span several construction seasons, the advancing waste pile will have a temporaryberm constructed at the toe of the exposed slope to collect stormwater runoff. The advancing wasteslope graded with a positive slope to the retention basin, see Figure 1, will also be protected forseasonal closure by means of a spray on soil stabilizer.

6 SLOPE STABILITY

A global stability analysis was performed for the EMF Repository to determine the stability of therepository slopes under various conditions and combinations of conditions. The global stabilityanalysis evaluates multiple possible failure planes through the waste and existing base materialresulting in an analysis that determines the factor of safety against failure for the entire slope.

The global stability analysis was performed using the Slope/W 2007 Version computer modelingsoftware developed by Geo-Slope International Ltd. Slope/W computes the factor of safety of themodeled slope by means of limit equilibrium analysis. Limit equilibrium formulations are basedupon the widely accepted method of slices analysis, where a failure plane is broken into a numberof cross-sectional slices and each slice is analyzed individually for static equilibrium to determinethe tendency for sliding.

The measure of acceptable resistance to slope failure, or acceptable factor of safety, varies withthe mode and conditions related to that potential failure. A factor of safety is calculated by dividingthe forces within a slope which resist failure by the forces within a slope which could cause failure:

This calculated factor of safety is then compared to the acceptable factor of safety for the givenconditions and mode of failure to determine whether the slope is acceptably stable.

7 COVER

The cover design for the EMF Repository uses an ET cover to minimize infiltration into the wastematerials and to ensure reduction or elimination of leaching and contaminant migration to protectgroundwater. The ET cover system consists of two sections: an upper section and a lower section.The upper section contains 6 inches of loam soil to support and maintain the cover vegetation. Thelower section is a 2-foot thick layer of silty clay loam soil that provides water storage during wetperiods for later use by the cover vegetation as well as evaporation into the atmosphere. The ETcover system will be placed over the waste materials in the EMF Repository; therefore, the rootsof the cover vegetation will extend beneath the cover system and remove water from the wastematerials. A schematic of the ET cover for the EMF Repository is illustrated in Figure 1.

The ET cover design for the EMF Repository was evaluated using both one-dimensional andtwo-dimensional computer models. The one-dimensional modeling was performed using the Envi-ronment Policy Integrated Climate (EPIC) model, developed by theTexasA&M Blackland Researchand Extension Center located in Temple, Texas; and the VADOSE/W computer program developedby Geo-Slope International, Inc. of Calgary, Alberta. The VADOSE/W computer program was alsoused to perform the two-dimensional modeling. Based on the one-dimensional modeling results,the efficiency of the ET cover at the base of the root zone using the EPIC model is approximately92.2% and the efficiency of the ET cover at the base of the root zone using the VADOSE/W com-puter program is approximately 91.6%. Based on the two-dimensional modeling results using theVADOSE/W computer program, the efficiency of the ET cover at the base of the root zone is 95.8%.

ET covers are typically constructed in arid or semi-arid climates with average annual precipitationamounts of 15 inches or less. At the EMF Repository site, the average annual precipitation is over30 inches based on daily weather data from the National Weather Service weather station located in

300

Kellogg, Idaho. The EPIC model and the VADOSE/W computer program were used to evaluate theapplicability and appropriateness of an ET cover for use at the EMF Repository site by using regionaland site specific data. Based upon the modeling results, the EMF Repository ET cover significantlyreduces the amount of percolation through the waste materials and reduces potential impacts togroundwater underneath the EMF Repository. These cover modeling results are dependent uponthe soil properties identified for the ET cover. Usually, soil samples are collected from a localborrow source and tested for soil properties important to the proper functioning and performanceof the ET cover. These soil properties are then used as the basis for the ET cover design. A localborrow source with soils appropriate for use in an ET cover does not exist in the vicinity of theEMF Repository site; therefore, soils meeting the specifications will be imported to construct theEMF Repository ET cover.

8 GROUNDWATER QUALITY PROTECTION

Four monitoring wells were drilled at the site during October 2007. One of these four monitoringwells is up gradient of the repository site. The wells reach to about 30 feet below ground surface, inthe upper aquifer. A broader look at the soil samples reveals two general soil types are under the site.The upper 12 to 15 feet below ground surface is made up of light brown silt, clay, and fine sandyclay. The top two to three feet of this material has bright orange-brown streaks in it from miningwaste. Gravel and sand beds are underneath this fine-grained material. These beds are saturated andare the first water bearing zone beneath the site. The monitoring wells were constructed to samplewater from this zone. In addition to the four monitoring wells installed at the site, two monitoringwells were installed to the west and to the south of the repository site during October 2008.

Five groundwater monitoring events were completed between December 2007 and November2008. Results from the five events showed that groundwater meets USEPA drinking water qualitythresholds for the following metals: antimony, cadmium, lead, and zinc. During the five monitoringevents, the only concentration to exceed the applicable USEPA drinking water quality thresholdswas dissolved arsenic in a monitoring well located west of the repository site, which had a con-centration of 0.0148 milligrams per liter (mg/l) compared to a National Primary Drinking WaterRegulation (NPDWR) maximum contaminant level (MCL) of 0.010 mg/l (IDAPA, 2007).

The groundwater monitoring effort will continue through construction and operation and main-tenance, at a minimum, to evaluate whether the repository has an effect on local groundwaterconditions. If future tests indicate the repository is responsible for impacts to groundwater quality,measures will be implemented to address the contamination source.

The groundwater gradient is very low across the EMF site, and studying water level elevationssuggests two possible flow paths to the Coeur d’Alene River. The shortest flow path is south fromthe repository, with groundwater moving east of the Old Mission. The other flow path is westfrom the repository, with discharge to the river west of the Old Mission site (downstream). Tomonitor groundwater movement to the Coeur d’Alene River, one additional monitoring well wasconstructed along each of these possible groundwater flow paths.

Water that seeps through the waste soil is called leachate. The leachate may dissolve metals fromthe waste soil and transport them through the repository material to the native soil underlying therepository. Some of the leachate may mix with groundwater, and could result in elevated metals inthe groundwater. However, the soil at the site appears to remove metals from leaching groundwaterby a process called sorption. Based on the reduction in leachate volume and the demonstrated abilityof the soil at the site to sorb dissolved metals, the potential for significant impacts to groundwaterfrom the presence of the repository is low. Column tests were completed simulating leaching fromthe repository waste material and interactions of leachate with underlying native soil. Results ofthe tests confirm the assessment that impacts to groundwater will be negligible.

Although the top of the repository will be gently sloped to help water flow off and away fromthe soil mound, a very small amount of water will seep into the soil mass. When the waste materialhas reached total design height, an evapotranspiration (ET) cover will be constructed to coverthe waste. Based on results from a cover analysis, the ET cover will reduce the amount of waterpercolating past the base of the root zone from 95.8 percent based upon two-dimensional modelingand 98.3 per cent when using a simplified one-dimensional model based on the wettest annualrecorded and average annual precipitation, respectfully. This amounts to approximately two inches

301

of leachate fully penetrating the ET cover and underlying two-foot root zone in the “wet” year andapproximately one-half inch in the “average” year. Given the height of the waste soil column, upto 32 feet high, the post-placement compaction and top surface grading, it is unlikely a significantquantity of leachate will fully penetrate the waste soil mass and come into contact with underlyingsoil. In addition, as explained above, the underlying clean soil has demonstrated an ability to sorbdissolved metals from leachate.

Building the repository with the ET cover will greatly reduce or eliminate the amount of leachategenerated beneath the repository. Based on Daily weather data from 1970-1973 and 1975-2005from the National Weather Service (NWS, 2006) weather station located in Kellogg, Idaho, thewettest year occurred in 1996 with an annual precipitation amount of approximately 49 inches.Based on the two-dimensional cover modeling results using the wettest year data, the efficiencyof the ET cover at the base of the root zone is approximately 96%. This means the EMF ET coversignificantly reduces the amount of percolation through the waste materials to approximately 2inches, and reduces potential impacts to groundwater underneath the EMF Repository. In addition,the two inches of leachate would then have to pass through the upper existing soil layer at the site,which has shown the ability to effectively remove metals from percolating groundwater.

Leachate generated by precipitation infiltrating through yard waste soil in the repository isnot expected to contain elevated levels of metals. Based on column tests closely approximatingconditions at the proposed repository, arsenic, cadmium, and lead will not be present, and only verylow concentrations of antimony and zinc will be present. The column tests actually indicate thatthe existing native deposits, which include the historic fluvial tailings horizon, have the potentialto generate more metals than yard waste soil, specifically cadmium and zinc. Cadmium, leachedfrom the existing native deposits, may be in the range of the NPDWR MCL, and zinc could exceedthe National Secondary Drinking Water Regulation (NSDWR) MCL.

Additionally, the total volume of leachate generated will be reduced by the construction andestablishment of a vegetative cap, which will reduce infiltration by promoting evapotranspiration.Periodic inundation by flood waters will only saturate a small volume of soil around the perimeterof the repository, and only for a brief period. The extent of infiltration based on the 2008 flood eventindicates that during a standing water event of 75 days, water will penetrate the repository mass 15to 17 feet. The thickness of the saturated zone will range between 0.5 and 0.7 feet. This model wasdeveloped using in-situ natural soil conditions. During repository construction, soil placed on theperimeter will be compacted to ≥ 95% maximum density. This compaction will likely reduce thehydraulic conductivity of the waste soil by an order of magnitude or more. Based on this reducedhydraulic conductivity, it is likely that the flood waters will penetrate two feet or less into the wastesoil mass. Long-term saturation of the base of the repository and the development of reducingconditions are not expected. Because the existing soils generate higher levels of metals than theproposed yard waste, the reduction in infiltration should result in an overall decrease in metalsleached to shallow groundwater, and an improvement in water quality.

9 GEOCHEMICAL EVALUATION OF POTENTIAL IMPACTS TO GROUNDWATER

The potential for impacts to groundwater from placement of contaminated soil at the EMF Reposi-tory has been evaluated by reviewing the available hydrological and geochemical data, includingsite-specific characterization data and relevant information from Coeur d’Alene Basin studies, andconducting column leach tests simulating site conditions.

9.1 Metal sorption

As precipitation infiltrates through contaminated soil placed in the repository, soluble metals willdissolve, resulting in porewater or leachate with elevated metals concentrations. Leachate willcontinue to migrate through the soil column until it reaches and mixes with underlying groundwater.

As leachate migrates through the soil column, dissolved metals will be sorbed to organic matter,clay minerals, and iron/manganese (Fe/Mn) oxides in the native fluvial sediments and in the existingtailings. Each metal will be sorbed to a different degree, depending on site specific characteristicsof the soil/water system, such as soil type, pH, and other factors. The degree or amount of sorptionis described by a term known as the partition coefficient. Partition coefficients have not been

302

determined for the EMF Repository, but literature values for arsenic, cadmium, copper, lead, andzinc, with similar soils, range from about 500 to 15,000. Values of 100 or greater are generallyconsidered to indicate that the metal is essentially completely sorbed by the soil. This is supportedby existing site characterization data. While the Mission Flats area contains dredge materials upto 40 feet thick with extremely high total metals concentrations, the shallow groundwater containsrelatively low levels of dissolved metals.

Additionally, metal concentration profiles in test pits and monitor well boreholes at the reposi-tory site indicate an abrupt change in soil metal concentrations at approximately two to four feetbgs. Concentrations in the upper one to two feet are up to approximately 8,700 mg/kg of lead;3,000 mg/kg zinc; 100 mg/kg arsenic; and 20 mg/kg cadmium. Those concentrations decreaseabruptly to multiples of 10 mg/kg at depths of three to four feet, which is interpreted as the nativesoil horizon. This decrease is almost certainly due to efficient sorption in the upper level of the nativesoils, indicating that metals concentrations in leachate from the EMF Repository will similarly bedecreased as it migrates through underlying soil.

To resolve concerns regarding metal leaching and sorption, a column test was conducted duringSummer 2008 to simulate conditions in the repository following the placement of yard wastesoils. A two-stage column test, in which one column was filled with yard soil and leached withsimulated rainwater, generated leachate representing that of yard waste soil placed at the EMFrepository. The yard soil leachate was subsequently passed through a column of native soil fromthe site, to determine metals attenuation/sorption and the chemical characteristics of leachate thatwill ultimately mix with groundwater.

The test data indicate that leaching of yard waste soil by precipitation will not release any arsenic,cadmium, or lead, and only very low concentrations of antimony and zinc. However, the native soilleached cadmium (at just below the drinking water MCL of 0.005 mg/L) and zinc (exceeding thezinc secondary drinking water standard of 5.0 mg/L).Therefore the yard waste soils pose essentiallyno risk to groundwater, whereas the native soils actually leach more metals to groundwater, althoughnot at levels that pose a human health risk.

Furthermore, it is important to note that after repository construction the volume of leachatepassing through native soil to the first groundwater will be significantly reduced or eliminated.The repository will have an ET cover, and the functioning ET cover will essentially create anumbrella over the 14-acre repository footprint. No water will be percolating through the 14 acresafter construction, where prior to construction an average of 32 inches per year fell to the groundat the repository site. The reduction in leachate volume should benefit the first water-bearing zonebeneath the site.

9.2 Infiltration into the repository

Infiltration of precipitation through the repository will be substantially reduced following comple-tion of ET cap construction. After the repository is full, an engineered ET cap will be constructed,vegetation will become established, and infiltration will be reduced due to plant respiration andevaporation. The average annual post-construction infiltration rate is estimated to be on the orderof from 0 to 7.8% compared to an estimated 75% under current conditions. Metals leaching shouldbe approximately proportional to the volume of infiltrating precipitation, consequently leachingthrough the EMF will be reduced by a factor of about 10 or more, based on cover modeling results.

The EMF site is also subject to periodic flooding during high water of spring runoff events onthe Coeur d’Alene River, resulting in standing water for short periods. Most recently, for example,was June 2008, when the site was flooded to a depth of two to three feet depth for several days.This was considered a 12-year runoff event.

A model to predict the extent of waste soil saturation due to the observed 12-year event wasconstructed using the USGS VS2DT code (USGS, 2009). The VS2DT program is designed tosimulate two dimensional, variably saturated flow. The model was developed to simulate waterflux through the part of the repository perimeter armored with riprap. This scenario was selectedbecause the armored areas will allow the greatest amount of flood water infiltration and modelingthis condition would represent the “worst-case” scenario for flood water intrusion. The armoredareas will be covered with 21 inches of riprap and gravel of very high hydraulic conductivity thatwill essentially place flood water in direct contact with contaminated repository waste material.

303

The armored areas constitute approximately 20% of the running length of the repository perime-ter. The remaining 80% will be constructed with a clean soil storage layer, top soil and a vegetatedsurface. Water transmittal through this material will be significantly less than in the armored areas.By selecting the armored areas for simulation, the model as constructed is conservative in estimationof the degree of flood water intrusion.

The results of the model indicate the 75-day standing water event will result in a maximuminfiltration of 15 to 17 feet into the repository soil. The saturated thickness of the flood waterintrusion ranges from 0.5 to 0.7 feet. This saturated area represents approximately 0.05% of thetotal repository volume.

The model incorporated the hydraulic conductivity value reported from a flexwall permeabilitytest performed on a remolded sample in a laboratory setting. The waste material as-placed will becompacted to ≥95% maximum density, much greater than the test sample. The field compactionto ≥95% maximum density will likely decrease the conductivity at least one order of magnitude.A reduction of this order would reduce the infiltration to less than two feet.

Based on the model assumptions and results, the potential for measureable groundwater impactsfrom lateral migration of flood water into the repository waste soil mass is low for the followingreasons: (1) the volume of saturated soil and metals in the saturated soil is negligible; (2) the yardsoils imported to EMF will have lower metals concentrations than the contaminated soil already atthe site; (3) groundwater monitoring results indicate that water beneath the site currently meets EPAprimary and secondary drinking water standards; and (4) the repository as designed will decreasethe amount of water moving through the native soil to groundwater, thus decreasing the metals loadfrom surface water sources. Furthermore, as column test results indicate, no additional metals areexpected to be leached from the yard soils in any event.

10 CONCLUSIONS

• Visual simulation using photographs was very useful in explaining visual effects to stakeholders.• Installation of an access bridge resolved load limit issues at a lower cost than other approaches.• Stormwater management addressed contaminated stormwater via on-site controls while un-

contaminated stormwater was kept un-contaminated.• Global slope stability was demonstrated using computer modeling so that a stable fill could be

constructed.• An evapotranspiration cover was included in a humid environment because it significantly

reduces that amount of precipitation passing through the waste materials.• Groundwater quality is estimated to be protected because of the reduced amount of precipitation

passing through the waste materials along with the ability of existing subsurface soils to absorbgroundwater contaminants. Perimeters construction reduces floodwater infiltration to a shortinsignificant distance.

REFERENCES

IDAPA, 2007, Chronic Aquatic Life Criteria, IDEQ Water Quality Standards. IDAPA 58.01.02.210. Lastupdate March 2007.

Idaho Transportation Department (ITD), 2008, Bridge Load and Resistance Factor Design Manual.NWS, 2005, Online Precipitation Data, Kellogg and Pinehurst, Idaho.TerraGraphics, 2009. East Mission Flats Repository, 90% Design Report, TerraGraphics Environmental

Engineering, Inc., Moscow, Idaho, June 5, 2009.U.S. Geological Survey (USGS), 2009, VS2DT Code, USGS Survey website, http://wwwbr.cr.usgs.gov/

projects/GW_Unsat/vs2di1.

304


Physical properties of mill tailings as foundation material forwaste repositories, Bunker Hill Superfund Site

Justin S. WoolstonTerraGraphics Environmental Engineering, Inc

ABSTRACT: To support soil cleanup and site remediation activities at the Bunker Hill SuperfundSite (BHSS) in northern Idaho, contaminated soils are removed from residential yards and storedin repositories. With a limited supply of suitable land available for site selection, former tailingsponds and tailings-covered ground are used as repositories. At the Big Creek Repository (BCR)in the BHSS, design methods developed from the results of in-situ and laboratory testing andanalysis of the physical properties of mine and mill tailings have been used to design, construct,and expand this repository. In 2004 the engineering design properties were developed for the BCRas a result of extensive geotechnical testing of the foundation materials. These properties wereretested to demonstrate how the strength properties change as the repository is loaded. The datacollected at BCR are being successfully applied to the expansion of the BCR and the design ofother repositories.

1 INTRODUCTION

This paper presents geotechnical strength data collected over a 6-year span for tailings that havebeen subjected to consolidation due to the placement of a waste repository on top of a tailingsimpoundment, including extensive in-situ and laboratory testing of flotation tailings before load-ing, during loading, and after substantial loading has occurred. The geotechnical strength propertiesof flotation tailings are not typically considered due to their method of deposition and their reten-tion by either a dike or dam. These fine-grained flotation tailings, often referred to as slimes,visually resemble silts and clays when in an unconsolidated, undrained state but exhibit strengthcharacteristics more representative of sands when in a drained state. The results and conclusionsof these tests are being applied to the design of soil waste repositories on top of inactive tailingsimpoundments.

2 SITE HISTORY

In 2004 the U.S. Environmental Protection Agency (USEPA) and the Idaho Department of Envi-ronmental Quality (IDEQ) identified an inactive Sunshine Precious Metals, Inc. (Sunshine) tailingimpoundment as the site for disposal of waste generated by cleanup activities conducted pursuantto the Bunker Hill Mining and Metallurgical Complex Operable Unit 3 Record of Decision (OU3ROD) (USEPA, 2002). The site referred to as the Big Creek Repository (BCR) is designed tosecurely hold soil waste generated from the remediation of the lower Coeur d’Alene River basinarea. As part of the Selected Remedy, repositories are necessary to store the waste generatedthrough the Bunker Hill Superfund Site (BHSS) cleanup process. The site is located adjacent toBig Creek in Shoshone County, Idaho, approximately 0.25 miles south of Interstate-90. The repos-itory is located on the more northerly of two tailings impoundments created by Sunshine, with thesouthern impoundment still in operation. The tailings impoundment was constructed in 1968 witha dam height of 10–40 feet and with slopes of 1.4 horizontal on 1 vertical (1.4H on 1V). The pondinterior measures approximately 14 acres and is filled with flotation tailings. In 1973, the dam wasraised 10 feet.

The tailings impoundment was selected as a waste repository site because: it was available, itwas already contaminated, and it was one of a few parcels of ground flat enough to accommodate a

305

soil waste repository. With the exception of the valley floor, most terrain in the area is mountainous,which is too steep for a stable waste configuration. The original repository design was created in2004 by the U.S. Army Corps of Engineers (USACE), although preliminary waste placement atthe site started as early as 2003. The USACE design placed approximately 250,000 cubic yardsof waste on top of the existing tailings, with the majority of waste placed on the centerline ofthe impoundment berm. This equates to approximately 30 feet of soil waste placed on top of the14-acre footprint of existing tailings material.

In 2005, anticipating the need for additional waste storage capacity, the design was expandedby increasing the height of fill material by an additional 20 feet. At this point the tailings weretested again to determine the new shear strength properties. Testing, both in-situ and laboratory,was again performed in 2009 by Strata Geotechnical Engineering and Material Testing (Strata),in compliance with the conditional approval from USACE, as USEPA’s third-party reviewer ofthe increased height expansion. USACE required the strength of the tailings to be tested prior toexceeding the original USACE design height, in order to verify the design parameters used inanalysis of the increased height expansion.

3 TAILINGS MATERIAL DESCRIPTION

The flotation tailings underlying the repository soil waste are as much as 45 feet thick and varyin grain size from fine-grained sand at one end of the impoundment to very soft, saturated siltreferred to as slimes at the opposite end of the pond, with a transitional zone in the middle. Theinitial repository design performed by USACE described the tailings as “low shear strength tailingsor slimes” (USACE, 2004). The USACE design asserted that as the repository soil waste was added,the pore water pressure within the tailings would increase due to the weight of the soil waste andthen slowly dissipate as the tailings consolidated, resulting in an increase in shear strength. TheUSACE design also asserted that if soil waste was placed too rapidly, the pore water pressure wouldnot have adequate time to dissipate, and the additional weight of the soil waste could cause slopefailure. Due to the perceived potential for slope failure, the repository was loaded slowly usingspecific procedures that allowed for the tailings to consolidate and develop adequate shear-strength.As the repository was filled, pore water pressure and settlement data were monitored to ensure therepository was filled in a safe manner.

When in operation, the tailings impoundment was filled by pumping the tailings slurry into theimpoundment at the southern end and allowing the tailings to settle out from south to north. Decantstand pipes located at the north end of the impoundment allowed excess water to be dischargedinto the adjacent creek. This process deposited a coarser grained tailings beach at the southernend and deposited the finer grained tailings at the northern end of the impoundment (Figure 1).Initial in-situ and laboratory testing of this beach material displayed strength properties typical ofsands; whereas, the fine grained slimes resembled silts and clays. Although USACE addressed theentire site in the initial repository design, they focused on the northern portion containing the finergained slimes material as the critical section for repository stability due to a perceived lack of shearstrength.

USACE extensively tested the tailings and the impoundment dikes with cone penetration tests(CPTs). The northern portion of the tailings impoundment containing the finer grained tailingsserved as the control point for determination of the performance of the tailings material as a foun-dation for the repository as the repository is loaded. In 2005, when the tailings were investigatedagain in preparation for expansion of the repository by means of increasing the final design height,in-situ testing consisted of vane shear tests (VSTs) performed in borings to a depth which corre-sponded to the depths tested by USACE. Golder Associates Inc. (Golder) used these data in thestability analysis of the height expansion assuming that shear strength would gradually increase asthe repository was loaded and as the tailings consolidated (Golder, 2005).

4 IMPOUNDMENT MATERIAL TESTING

USACE tested the impoundment material with a CPT to determine the design parameters for therepository. Golder tested the impoundment material with a VST for the increased height expansion.

306

Figure 1. Sunshine Precious Metals, Inc. (Sunshine) tailing impoundment, 1975.

Strata tested the impoundment material with a CPT to confirm the material strength subsequent tosubstantial loading of the repository.

USACE performed CPTs throughout the impoundment site. Golder focused the VSTs at thenorthern end of the impoundment where the slimes were located, based on the initial data fromUSACE, as did Strata with the CPTs.

In the last round of testing performed by Strata, the waste material stored in the repository poseda problem with regards to accessing the tailings material. The 30-foot thick layer of waste materialincludes debris consisting of chunks of concrete, metal, and other materials which prevented aCPT apparatus from advancing to depth. To address this issue, Strata advanced the borings to adepth that corresponded to the top of the tailings material. Once these depths were achieved, Stratare-tooled the drill rig to advance the CPT apparatus, which provided continuous data from the topof the tailings layer to the bottom of the tailings impoundment.

Strata drilled a total of 11 borings at pre-determined locations that corresponded to the tail-ings material of greatest concern because of strength properties. Approximate boring locationsare illustrated in Figure 2. Borings extended to depths ranging from 20 feet to 53.5 feet belowthe ground surface. Boring B-7a was accomplished to obtain an in-situ sample of mine tailingsat a depth of approximately 53 feet and a CPT sounding was not conducted. Boring B-7b wasaccomplished to conduct a CPT sounding of mine tailings near B-7a to allow for soil parametercomparison.

Strata advanced CPT soundings from just below the overburden soil to depths ranging fromapproximately 55 feet to 73 feet below the existing ground surface. At depths of approximately20–45 feet below the ground surface overburden soil was encountered and below the overburdenmine tailings were encountered, which generally consisted of a clay and silt sized particle mixturewith sand that was tan to gray, soft, and very moist. Mine tailings extended to at least the termi-nation of CPT testing at approximately 55–73 feet below the ground surface. Based on the minetailings’ in-situ moisture and recorded pore water pressure during CPT soundings, TerraGraphicsEnvironmental Engineering, Inc. (TerraGraphics) assumed the tailings were saturated and that

307

Figure 2. Boring locations.

groundwater elevation at the site generally corresponded with the interface between mine tailingsand overburden soil. TerraGraphics recorded penetration resistance on the cone and friction on thesleeve with depth during penetrometer advancement. The near-continuous data profile obtainedfrom CPT soundings was used to help correlate general stratigraphy at depth, soil type, and variousengineering properties of the subsurface materials.

TerraGraphics estimated soil engineering parameters based on published correlations to dataobtained from the CPT soundings. The parameters included soil behavior type, internal frictionangle, and undrained shear strength (cohesion). The following sections describe the methodologyused for estimating each of the requested parameters.

4.1 Soil behavior type

Soil behavior type (SBT) was estimated based on correlations with cone tip resistance and frictionratio (sleeve friction divided by cone tip resistance) as published by Robertson et al. (1986). Basedon these correlations, the mine tailings’ SBT appears to be clay and silt mixture with varyingamounts of sand to the total depths explored. Each sounding was terminated in soil interpreted asclayey to sandy silt or silty sand at depths ranging from approximate 55 feet to 73 feet below theexisting ground surface. Although the slimes exhibit similar properties of clays and are clay-sizedparticles, they do not fully meet the classification requirements for clay.

4.2 Drained friction angle

Drained friction angles of mine tailings in CPT soundings were estimated based on the correlationto cone tip resistance published by Mayne (2001).As expected in a non-homogenous geologicunit, some variability was observed in friction angle correlations both within the CPT soundingsand between different CPT soundings. Table 1 summarizes the minimum, average, and maximumfriction angle, in degrees, estimated for mine tailings in each sounding.

The mine tailings sample collected from boring B-7a and corresponding to the CPT data ofboring B-7b was subjected to a direct shear test to determine the drained soil friction angle of thein-situ mine tailings. The sample yielded a value of 32.5 degrees as shown in Figure 3. These datawere compared to the CPT data of boring B-7b.

The effective friction angle estimated from the CPT results for boring B-7b and correspondingto the depth of sample collection from boring B-7a is 33 degrees compared to 32.5 degrees for thesample tested in direct shear.

308

Table 1. Estimated effective drained friction angle.

Location CPT-# Minimum Ø Average Ø Maximum Ø

B-1 CPT-1 19.5◦ 31.9◦ 38.8◦B-2 CPT-2 24.0◦ 32.1◦ 37.9◦B-3 CPT-3 25.8◦ 31.7◦ 36.0◦B-4 CPT-4 28.4◦ 32.7◦ 38.6◦B-5 CPT-5 28.8◦ 32.9◦ 41.8◦B-6 CPT-6 27.9◦ 32.8◦ 41.3◦B-7b CPT-7 28.2◦ 33.0◦ 40.5◦B-8 CPT-8 30.2◦ 33.3◦ 34.7◦B-9 CPT-9 22.4◦ 33.0◦ 41.9◦B-10 CPT-10 27.9◦ 34.0◦ 38.6◦

Figure 3. Results of direct shear test of tailings material.

4.3 Undrained shear strength

Undrained shear strength (cohesion) of mine tailings in CPT soundings was estimated based on thecorrelation to cone tip resistance and total in-situ vertical stress. This correlation was developedby the Federal Highway Administration (FHWA, 1992). A cone factor (Nk) of 15 was used inthe correlation to undrained shear strength. As encountered with friction angle analyses, some

309

Table 2. Summary of undrained shear strength estimates.

Minimum Su∗ Average Su Maximum SuLocation (tsf) (tsf) (tsf)

CPT-1 0.02 1.59 5.41CPT-2 0.13 1.48 4.44CPT-3 0.25 1.16 2.92CPT-4 0.48 1.47 5.17CPT-5 0.54 1.56 10.21CPT-6 0.41 1.47 9.35CPT-7 0.44 1.51 7.81CPT-8 0.75 1.59 2.19CPT-9 0.01 1.83 10.44CPT-10 0.40 2.01 5.26

∗Su = Shear Strength

Table 3. Comparative material strength properties by test and date.

Waste depth CohesionDate (feet) Test method Ø Average (PSF)

USACE 2003 N/A CPT 0◦ 300Golder 2005 15 VST 32.0◦ 0Strata 2009 30 CPT 33.0◦ 0Strata 2009 30 Direct Shear 32.5◦ 215

variability was observed in undrained shear strength correlations both within the CPT soundingsand between different CPT soundings. Table 2 summarizes the estimated minimum, average, andmaximum undrained shear strength, in tons per square foot (tsf), for fine grained soil in eachsounding.

4.4 Soil strength parameters

The soil strength parameters are based on CPT correlations from published information. Thetwo primary soil strength parameters presented are Effective Friction Angle and Undrained ShearStrength. These parameters are based on drained and undrained conditions, respectively, whichimpact the soil’s strength response to load and stress.

With the purpose of the data collection activities being to confirm the strength characteristicsof the mine tailings as they experience consolidation due to the filling of the repository, the dataobtained in the most recent data collection activities were compared to those activities previouslyperformed. Table 3 presents the data gathered at a common location in the northwest corner of thetailings impoundment by USACE, Golder, and Strata.

The data obtained from the in-situ testing in 2005 and 2009 confirm the assumption of theUSACE design that as the repository is loaded, consolidation would occur in the tailings and resultin greater shear strength in the form of internal friction. It is worth noting that the strength increasewas essentially fully achieved with 15 feet of waste material placed over the tailings; the additional15 feet placed from 2005 to 2009 had little to no effect on the shear strength of the tailings. Asshown in Table 4, settlement essentially had tapered off by 2008. Variations in rates of settlementfrom one location to another are directly related to the fill sequence of the waste that concentrateswaste placement on one end of the repository each year as well as the difference in material typesfrom slimes at one end to beach sands at the other. Settlement monument locations are shown inFigure 4.

310

Table 4. Settlement monument data.

Yearly settlement (feet)2004

# ∗ (Base year) 2005 2006 2007 2008

North End Dike (West) 1 0 −1.09 nd∗∗ −0.45 −0.02(Slimes) Tailings (Slimes) 2 0 −0.20 −0.65 −1.07 −0.2

Dike (East) 3 0 −0.07 −0.16 −0.23 −0.01

Middle Dike (West) 4 0 −0.16 −0.38 −0.85 −0.14(Transitional) Tailings (Transitional) 5 0 nd∗∗ nd∗∗ −0.74 −0.12

Dike (East) 6 0 −0.44 −0.05 −0.65 −0.07

South End Dike (West) 7 0 0.00 −0.35 −0.39 0.00(Beach) Tailings (Beach) 8 0 −0.03 −0.10 −0.06 0.00

Dike (East) 9 0 −0.03 −0.03 −0.03 0.00

∗ See Figure 4 for Settlement Locations∗∗ nd = No Data

Figure 4. Settlement monument locations.

The original design analysis performed by the USACE estimated a total settlement of 3.6 feetwith primary consolidation accounting for 2.1 feet and secondary consolidation accounting for1.5 feet (USACE, 2004). Settlement estimates were based upon the thickest layer of slimes and anoverburden of 35 feet. As stated in the Golder analysis, the less than expected settlement may bedue in part to the parameters used for the settlement analysis were derived from laboratory testson small samples of the slimes, which may have been disturbed during sampling (common withweak materials). Therefore the samples tested and parameters used in the settlement analysis maynot be representative of the tailings within the repository; and the settlement calculation assumes38 feet of uniform tailings slimes, which may not be representative of the thickness of slimes overthe entire repository (Golder, 2005).

311

5 SUMMARY

The geotechnical strength properties of tailings vary significantly from site to site and even withina particular site. Flotation tailings range in size form coarse sands to clay size particles. While thecoarse-grained sands have been found to display consistent geotechnical strength properties similarto those of naturally occurring sands, the fine-grained clay-sized particles often display visualcharacteristics typically associated with clays. Often referred to as slimes as they are commonlyfound under saturated conditions, the clay-sized particles tend to exhibit geotechnical strengthproperties more consistent with sands and silts than clays. This is usually due to the absence ofclay minerals necessary for the material to exhibit clay properties, although this depends on theprocessed material.

The tailings contained within the impoundment located beneath the BCR have proven to be astable foundation for waste placement. This overall stability, however, depends upon the perimeterdikes containing the tailings and the rate and methods used for filling the repository. When con-struction first commenced at BCR the tailings had minimal shear strength. The addition of wasteled to consolidation of the tailings and subsequent strength gain. Although the shear strength cur-rently exhibited by the tailings is more than adequate, the perimeter dike was and still is essentialto the overall stability. The dikes contained the tailings during gradual loading, which allowed porepressure to increase and slowly dissipate in a controlled manner. Even with the current increasein shear strength of the tailings, the perimeter dike serves as a necessary foundation for globalstability.

The data and experience gained in the design, construction, and monitoring of the BCR arebeing applied to other sites where tailings impoundments are serving as the foundation for wasterepositories.

REFERENCES

U.S. Environmental Protection Agency (USEPA), 2002. Record of Decision (ROD) – Bunker Hill Mining andMetallurgical Complex Operable Unit 3 (Coeur d’Alene Basin). September 2002.

U.S. Army Corps of Engineers (USACE), 2004. Big Creek Repository – Design Analysis Report (Final).Golder Associates, Inc., 2005. Big Creek Revised Stability Analyses.Federal Highway Administration, 1992. The cone penetrometer test. Federal Highway Administration

Publication No. FHWA-SA-91-043.Lunne, T., Robertson, P. and Powell, J., 1997. Cone penetration testing in geotechnical practice. E&F Spon,

an imprint of Routledge, New York.Mayne, P.W., 2001. Stress-Strain-Strength-Flow parameters from enhanced in-situ tests. Proceedings of the

International Conference on In-Situ Measurement of Soil Properties and Case Histories, Bali, Indonesia,May 20–24, pp. 27–48.

Roberston, P.K., Campanella, R.G., Gillespie, D. and Greig, J., 1986. Use of piezometer cone data. Proceedingsof theASCE Specialty Conference In Situ ’86: Use of In SituTests in Geotechnical Engineering, Blacksburg,1263–80.

312

Dry Stack/Paste


Dry stack tailings design for the Rosemont Copper project

L. Newman & K. ArnoldRosemont Copper Company, Tucson, Arizona, USA

D. WittwerAMEC Earth and Environmental, Denver, Colorado, USA

ABSTRACT: Dry StackTailings management is a major step toward the future of tailings manage-ment in mining. While the use of dry stack tailings is relatively new to the industry, the technologiesused for the process are well tested in the industry. A unique and significant portion of the RosemontProject is the tailings management system that has been designed to manage 75,000 dry tons perday of material. Tailings slurry will be dewatered and thickened in tailings thickeners and overflowwill be reused while underflow is pumped to the tailings filters. A conveyor system will transporttailings filter cake to a dry stack tailings facility. Using proven technology in a new way gives theRosemont Project the tools necessary for success in Arizona.

1 INTRODUCTION

The Rosemont Property (Property) consists of a group of patented mining claims, unpatentedmining claims, and fee land that covers most of the Rosemont Mining and the adjacent HelvetiaMining Districts. Specifically, the Project is located approximately 30 miles southeast of Tucson,west of State Route 83.

Past and recent exploration activities have confirmed or identified the availability of approxi-mately 600 million tons (MT) of ore, with an estimated project life of approximately 21 years.Copper, silver, and molybdenum will be recovered by grinding and froth flotation, with the principalrecovered minerals being the copper sulfide minerals (bornite, chalcocite, and chalcopyrite) andthe molybdenum sulfide mineral (also referred to as “moly”). The sulfide ores will be processed ata mill throughput of approximately 75,000 dry tons per day or 27 million dry tons annually. Heapleaching, solvent extraction and electrowinning will be used to produce metallic copper at a rateof 10,000 tons annually. These processes are standard to the industry.

The tailings management system makes the Rosemont Project unique. Tailings slurry will bedewatered and thickened in tailings thickeners. Thickener overflow (water) will be pumped to thereclaim water system. Thickener underflow (thickened tailings slurry) will be pumped to tailingsfilters. Conveyor belts will transport tailings filter cake to a dry stack tailings facility. AMEC Earthand Environmental was commissioned by Rosemont to complete a detailed design of the proposedDry Stack Tailings Storage Facility (TSF).

2 BACKGROUND OF FILTERED TAILINGS DISPOSAL

Most of the world’s concentrators or milling operations use conventional tailings processes resultingin tailings impoundments. These impoundments store tailings slurry that is pumped to the impound-ment at solids contents ranging between 25% and 60% depending on the degree of thickening thatis carried out prior to deposition. These engineered impoundments require construction and main-tenance to insure structural integrity for the retention structures. In addition, these facilities mustbe designed and constructed to manage large quantities of water.

Conventional tailings impoundments remain the primary alternative for the majority of oper-ating and proposed mines around the world. These facilities are developed using tailings slurries

315

from the milling process. However, with advances in dewatering technologies over the past fewdecades, tailings slurry is actually only part of a continuum of tailings “states” available to themodern tailings designer. Pressure filters consisting of horizontally or vertically stacked plates,and vacuum filters consisting of drums and horizontal belts are the most common filtration plantconfigurations. Development of large capacity vacuum and pressure filter technology has presentedthe opportunity for storing tailings in an unsaturated state, rather than as conventional slurry orin a paste consistency associated with thickened tailings. The nature of the tailings material, bothgradation and mineralogy, is important when considering filtration. Specifically, high percentagesof <74 µm clay minerals (i.e. not just clay-sized but having clay mineralogy) may hinder effectivefiltration.

Determining the most cost-effective manner to obtain a filtered product consistent with thegeomechanical requirements of the tailings can be a challenge. It is important to anticipate min-eralogical and grind changes that could occur over the life of the project. The candidate filteringsystem(s) must be able to readily expand/contract with future changes at the mine while minimizingeconomical impact. For projects that can use a non-slurried tailings alternative to optimize use ofavailable water and to streamline permitting and/or operating conditions, filtered tailings are oftenan excellent alternative.

Filtered tailings are transported by conveyor or truck and placed, spread and compacted to forman unsaturated, dense and stable tailings stack (often termed a “dry stack”). Dry stack facilitiestypically don’t require a dam for a retention structure and as such, no associated tailings pond.Each project needs to assess the potential applicability for filtered tailings based upon technical,economical, and regulatory constraints. Experience shows the most applicable projects are thosethat have one or more of the following attributes:

– Are located in arid regions, where water conservation is crucial (e.g. Western Australia,Southwest United States, much of Africa, many regions of South America including Chile)

– Have flow sheets where economic recovery (commodity or process agent(s)) is enhanced bytailings filtration

– Are located in areas where very high seismicity precludes some forms of conventional tailingsimpoundments

– Are located in cold regions, where water handling is very difficult in winter– Have topographic considerations that exclude conventional dam construction and/or viable

storage to dam material volume ratios– The operating and/or closure liability of a conventional tailings impoundment are in excess of

the incremental increase to develop a dry stack– Are located in areas where construction materials for conventional dams do not exist or are very

expensive to supply.

Moreover, filtered tailings stacks generally require a smaller footprint for tailings storage (e.g.much lower bulking factor), are easier to reclaim, and can have lower long-term (closure) liability interms of potential environmental impact. Filtered tailings (dry stacks), although new to many miningjurisdictions, are becoming more common both for operating mines and for projects in the evaluationstage (e.g. feasibility). Figure 1 below outlines the continuum of water contents available for tailingsmanagement and includes the standard industry nomenclature. Decreasing water content increasesplacement expense because hauling or conveying is required rather than pumping. However, asthe water content decreases, the tailings are able to be placed in self-supporting structures such asstacks.

Filtered tailings are typically taken to be the dry cake stage. Dry cake material has enoughmoisture to allow the majority of pore spaces to be water filled (70–85% saturation) but not somuch as to preclude optimal compaction of the material.

Often one of the main reasons to select dry stacked filtered tailings as a management option isthe recovery of water for process water supply. This is particularly important in arid environmentswhere water is an extremely valuable resource. Filtering the tailings removes the most water fromthe tailings for recycle when compared with other tailings technologies. This recovery of water hasa cost benefit to the project, which offsets the capital and operating cost of the tailings system. Itshould be noted, water surcharge storage needs to be factored into the design of a filtered tailingssystem. Depending upon the application this can be a small water supply reservoir or tank. Where

316

Figure 1. Tailings Continuum (after Lighthall, Davies, Rice and Martin. 2002).

water is relatively scarce, either year round or seasonally due to extreme cold, sending immensequantities of water to quasi-permanent storage in the voids of a conventional impoundment canseverely hamper project feasibility. By reclaiming the bulk of the process water in or near the mill,far more efficient recycle is achieved.

One of the main advantages of dry stack tailings is the ease of progressive reclamation and closureof the facility. The facility can often be developed to start reclamation very early in the projectlife cycle. This can have many advantages in the control of fugitive dust, in the use of reclamationmaterials as they become available, and in the short and long-term environmental impacts of theproject. Progressive reclamation often includes covers and re-vegetation of the tailings slopes andsurface as part of the annual operating cycle. Dry stack facilities can be developed to closelyapproximate their desired closure configuration with a plan to manage surface runoff. The tailingscan be progressively reclaimed in many instances. In all cases, a closure cover material is requiredto resist runoff erosion, prevent dusting and to create an appropriate growth media for projectreclamation.

The lack of a tailings pond, very low (if any) appreciable seepage from the unsaturated tailingsmass and general high degree of structural integrity allows dry stacks to present the owner/operatorwith a comparably straight-forward and predictable facility closure in comparison with mostconventional impoundments.

3 SITE CHARACTERIZATION

The Rosemont tailings storage facility (TSF) site is located within the southwestern portion ofthe Mexican Highlands basin and range physiographic province within the Santa Rita mountainrange. The Mexican Highlands is a transitional terrain with varied topographic relief separating theSonoran Desert to the southwest and the Colorado Plateau to the north. The Santa Rita mountainrange defines the contact between the Sonoran Desert and the Mexican Highlands sections. Asignificant portion of the site is characterized by rolling grasslands interrupted by densely vegetatedwashes. Relatively steep mountain valleys characterize other portions of the site.

317

3.1 Geology

The geologic units underlying the proposed location for the Dry StackTSF are described in descend-ing order, from the surface downward. The uppermost geologic units include the Gila Conglomerate(Miocene to Pliocene), which locally overlies the Mount Fagan Rhyolite (late Cretaceous). The GilaConglomerate consists of sandstone with variable amounts of pebble-sized clasts ranging fromrounded to subangular. Field investigations intercepted weakly to strongly cemented units from theGila Conglomerate at shallow depths. The Mount Fagan Rhyolite is described as a light coloredash flow tuff with variable amounts of phenocrysts with localized zones of megabreccia with clastsexceeding 3 feet. The Gila Conglomerate and Mount Fagan Rhyolite Formations exceed 200 and1,000 meters, respectively.

The Apache Canyon Formation (early Cretaceous) consists of variable stratigraphy, predom-inately characterized by shales and mudstones, and lesser amounts of micrictic limestone andfeldspathic sandstone. The Willow Creek Formation (early Cretaceous) is characterized as the old-est local formation consisting of sandstone, variably interbedded with medium to large grains, andmudstone. The Willow Creek Formation is further interbedded with several lava flows consisting ofmaficigneous materials from 10 to 100 meters thick. The Apache Canyon Formation greater than400 meters in depth and conformably overlies the Willow Creek Formation which exceeds morethan 2,200 meters in thickness.

3.2 Seismicity

As part of the design criteria, the maximum credible earthquake (MCE) was determined to definethe parameters to be used in the analysis of the Dry Stack TSF, based on a review of the tectonicframework and historical earthquake activity in the region of the facility. The deterministic approachof obtaining seismic parameters involves: (1) identifying the largest potentially active fault closeto the site, (2) estimating the maximum earthquake that the fault is capable of producing, and (3)determining the Peak Ground Acceleration (PGA) that will be produced at the site from this event.

The Santa Rita fault zone is the closest of 27 contributing fault sources within a 200-kilometerradius of the project site with a distance from site of 11.2 kilometers and a length of approximately52 kilometers. The PGA value for a magnitude 7.1 event at the Santa Rita fault zone was determinedto be 0.33 g and was selected as the seismic hazard representing the MCE design event for the projectsite.

3.3 Geotechnical investigations

In support of the design for the Dry Stack TSF, Tetra Tech and Rosemont Copper Company con-ducted several geotechnical field investigations. The field investigations consisted of using severaldifferent methodologies of surface and subsurface exploration across the Rosemont Copper site.The following descriptions of field investigations are constrained to explorations within the DryStack TSF. The objectives of the geotechnical investigations included:

– To define general subsurface conditions for use in evaluation of the Dry Stack TSF stability;– To identify suspect zones that could affect the performance of the Dry Stack TSF; and– To quantify engineering characteristics of the materials incorporated into the Dry Stack TSF.

Historic investigation of the Rosemont Copper property was previously completed by severaldifferent companies and consisted of borings to variable depths. Two phases of field investigationswere completed by Tetra Tech between November 2006 and March 2007 and between May andJuly of 2008. These field investigations included activities such as excavating test pits, drillinggeotechnical borings, performing geophysical surveys, mapping surficial geology, and completinga laboratory testing program.

3.4 Subsurface conditions

There were10 test pits and 38 geotechnical borings completed within or near the vicinity of theDry Stack TSF. The borings were completed with several methodologies, depending on subsurfaceconditions. Either advancing hollow stem augers or an air hammer penetrated the alluvial/colluvial

318

materials overlying bedrock. NQ wireline coring techniques advanced the borings through thebedrock. Samples of the subsurface materials from the geotechnical borings were collected with2-inch and 1 3/8-inch inside diameter spoon samplers in general accordance with the standardpenetration test (SPT) described by ASTM Method D1586. Penetration resistance values (N-values), when properly evaluated, indicate the relative density or consistency of the soils andcan be correlated to internal angles of friction.

The soils encountered at the Dry Stack TSF classify as alluvium/colluvium materials predomi-nately consisting of sands with varying amounts of clay, silt, and gravel. The soils encounteredin the field investigation generally included 1 to 3 feet of topsoil and growth media underlain byalluvial/colluvial materials. The underlying sequence of bedrock consists primarily of the WillowCanyon, Apache Canyon, and Mount Fagan Rhyolite formations. Depth to bedrock varies signifi-cantly across the site as assessed from the test pits, borings, and results from the geophysicalsurveys. Depth to bedrock (as interpolated by field personnel) varies from 0 to over 100 feet withan average depth of approximately 40 feet. The geophysical surveys were subsequently correlatedto subsurface geology as interpreted from historic borings and test pits and place bedrock slightlymore shallow than the borings.

The consistency of the alluvial/colluvial materials encountered in the field investigation variedfrom medium dense to very dense for granular materials and hard for fine grained materials, withthe majority of encountered soils representing very dense or hard conditions as evidenced by theblow count N-values shown on the geotechnical boring logs. The average depth to interception ofvery dense soil is between 10 and 15 feet.

Both Tetra Tech and AMEC obtained select disturbed samples and bench scale tailing samplesfrom the field investigation and pilot plant studies, respectively, for material characterization,hydraulic conductivity properties, and strength parameters. Historic and current laboratory testingcompleted between 2006 and 2009 are listed below:

– Sieve Analysis: ASTM D 422– Hydrometer Analysis: ASTM D 1140– Natural Moisture Content: ASTM D 2216– Natural Density: ASTM D 2937– Atterberg Limits: ASTM D 4318– Specific Gravity: ASTM D 792– Standard Proctor Compaction: ASTM D 698– Modified Proctor Compaction: ASTM D 1557– Direct Shear: ASTM D 3080– One-Dimensional Consolidation: ASTM D 2435– ICU Triaxial Shear: USACE EM 1110-2-1906– Point Load Test: ASTM 5731– Unconfined Compressive Strength: ASTM 7012– Flexible Wall Permeability: ASTM D 5084– Constant Head Permeability: ASTM 2434 (modified apparatus)– Hanging Column and Water Potential: ASTM D 6836– Pressure Plate: ASTM D 6836 & D 2325

3.5 Geochemical tailings characterization

Geochemical testing was completed by Tetra tech on four bench scale samples. The following testswere completed on all or some of the bench scale tailings samples:

– Acid-Base Accounting– Net Acid Generation– pH Testing– Humidity Cell Testing (Kinetic)– Synthetic Precipitation Leaching– Meteoric Water Mobility– Whole Rock Analysis

319

Geochemical testing of the tailings indicates the tailings generally (1) contain less than 0.01percent sulfide-sulfur, (2) can be classified as inert with respect to acid generation, (3) possess highcapacity for acid neutralization, and (4) produce very low metal concentrations in resulting leachate.Furthermore, the acid-base accounting testing indicates the properties of the tailings meet ArizonaDepartment of Environmental Quality (ADEQ) criteria as inert, with total-sulfur concentrationsless than 0.3 percent and a net neutralization potential greater than 0 or a neutralization potentialratio greater than 3 (ADEQ, 1999). Kinetic or humidity cell testing is a laboratory test whichreplicates weathering in an accelerated timeframe. Each week the material subjected to weatheringis rinsed and the resulting solution analyzed for chemical constituents in order to verify possible acidgenerating materials. Test results indicate the tailings are inert and are not anticipated to becomeacid generating.

The synthetic precipitation leaching and meteoric water mobility procedures are primarily con-cerned with the potential for release of chemical constituents, including metals, in both coarseand fine grained materials. The results of each procedure indicate the majority of metal con-centrations were either below detection concentrations or low compared to aquifer water qualitystandards.

Solids liquid separation (SLS) testing was conducted on flotation tailings samples obtained fromthe pilot plant studies. The information collected from the laboratory testing was used to providea general set of data to design and size pressure and vacuum filtration equipment. The followingtests were completed on the pilot plant tailings samples:

– Flocculant Screening and Evaluation– Static Thickening– Dynamic High Rate Thickening– Pulp Rheology– Pressure Filtration Studies– Vacuum Filtration and Washing Studies

4 DRY STACK TAILINGS STORAGE FACILITY DESIGN

The Dry Stack TSF consists of two separate areas referred to as the Phase I and the Phase II DryStack TSF. Phase I will be operated for approximately the first 12 years and can accommodateapproximately 343 million tons of tailings. The Phase II Dry Stack TSF is an extension of thePhase I facility, will operate between Years 12 through 22 and can accommodate up to 253 milliontons of tailings.

4.1 Tailings filtration

For design purposes filtration tests were conducted to evaluate the potential for achieving suitablemoisture contents of the tailings using pressure and vacuum filtering processes and to size filtrationunits for a production rate of 75,000 tons per day. Based on the test results, the use of pressurefiltration produced tailings moisture contents within acceptable limits that were considered to bedischargeable and stackable. Additional filtration tests were conducted for filter sizing and arediscussed in Section 7.

Standard Proctor (compaction) tests were completed by AMEC on the bench scale tailings inaccordance with ASTM Method D 698. The tailings had an optimum moisture content of 15percent (by dry weight) and a maximum Standard Proctor dry density ranging between of 116 and119 pounds per cubic feet. Based on the moisture-density testing completed, a moisture contentnear 15 percent (by dry weight) is desired to optimize tailings placement, compaction efforts andtrafficability of the dry stack tailings surface. However, a moisture content ranging between 12and 18 percent will be acceptable for the placement of the tailings without affecting the overallstability of the Dry Stack TSF, as tailings exceeding 18 percent moisture will experience a rapidloss of strength. Therefore, the recommended moisture range is 15 percent plus or minus 3 percent(by dry weight).

320

4.2 Foundation preparation

Site preparation will include removal of trees, clearing and grubbing, construction of access roadsand salvaging topsoil for future reclamation use. Throughout most of the basin this will typicallyleave a surface of moderate relief composed of slightly weathered bedrock. Throughout the basin, anatural drainage network exists consisting of alluvial channels filled with sands and gravels. Prior toplacement of tailings, the drainages will be cleared and grubbed and the soil or growth media will beremoved and stockpiled in designated areas. Pockets of unsuitable materials within the limits of theDry Stack TSF footprint will be removed at the direction of the engineer. These unsuitable materialsmay include but will not be limited to soft saturated zones, highly organic zones, loose zones, andother potentially deleterious materials. After stripping of the drainages is complete, larger drainageswill be filled with inert rock to facilitate site drainage by providing a network of underflow drains.Remaining drainages will be reworked with suitable fill material prior to tailings placement.

4.3 Rock buttresses

An initial Starter Buttress will be constructed in the lower Barrel Canyon drainage using rockfillto accommodate approximately three months of tailings storage. However, if needed, the dry stacktailings can be placed upstream of the Starter Buttress to allow for additional storage prior to theplacement of the next lift of the Rock Buttress. Concurrent tailings and rockfill placement willoccur throughout the life of the TSF. Rockfill will be advanced ahead of the tailings level in 50-footlifts using upstream construction methods. The Rock Buttresses will have top widths of 150 feetto accommodate two-way haul traffic and outer slopes of 3 horizontal to1 vertical (H:V). The finalouter slopes of the Dry Stack TSF will achieve an overall slope of 3.5H:1V. This configurationwill allow visual screening of the tailings placement activities from Highway 83 and concurrentreclamation of the rock buttress slopes. Slope design mimics existing landforms while providingmaintenance access and drainage over the long term.

4.4 Dry stack tailings placement

As described previously, rockfill will be placed in the lower Barrel Canyon drainage area to providea buttress for the dry stack tailings. Dry tailings will be delivered from the filter plant by conveyorand placed in 25-foot lifts using a stacker against the rock buttress. A dozer will be used to spreadthe dry tailings and the tailings compacted to provide for trafficability of the conveyor/stackersystem and to provide stability in areas near the berm. An abbreviated second conveyor systemprovides redundant stacking capacity to allow temporary disposal of tailings into the upper tailingsarea for placement with dozers if the primary conveyor is inactive for movement or maintenance.

The outer perimeter of each tailings lift beneath the areas of the rock buttress will be placedin 5-foot lifts and compacted with a vibratory smooth drum compactor (or similar equipment) toachieve a density equal to 90 percent of the standard proctor density (ASTM D 698) to provide asuitable foundation for subsequent waste rock buttress construction using upstream constructionmethods. This will also enhance the overall stability of the Dry Stack TSF.

4.5 Flow-through drains

Flow-through drains will be constructed in the significant washes that exist within the Dry StackTSFarea to augment the existing drainage courses and allow them to pass runoff beneath the tailings. Thesize of rock used in the flow-through drains will be approximately 12 inches or less and separatedfrom the tailings using a non-woven geotextile. For construction purposes the flow-through drainswill be constructed to be a minimum of 25 feet thick and have a top width of 150 feet.

5 WATER MANAGEMENT

A two-component system has been designed to separate water that comes into contact with theoperations facilities from stormwater suitable for discharge. For management of run-on water,permanent diversion channels will be constructed at pre-production and approximate productionyear 12 to divert surface runoff around the plant site and eventually around the Phase 2 or North

321

Dry Stack area. This diversion will discharge into the existing drainage northeast of the Dry StackTSF. Runoff that directly comes into contact with the tailings will not be discharged but instead willbe collected using perimeter ditches and by sloping the tailings to drain to low spots or evaporationponds located on the tailings surface. This stormwater will be allowed to evaporate or will bepumped to containment ponds at the plant site.

The dry stack tailings will be placed behind large rock buttresses comprised of pit run rockfill.This will eliminate the potential for erosion of tailings on the outer face. The outer buttresses willalso reduce the visual impact from surrounding areas and allow reclamation of the slopes to beginthe first year of operations. The rock buttress will be raised ahead of the dry stack tailings surface,eliminating the opportunity for discharge from the face of the tailings.

The top surface of the tailings will be shallow, on the order of 0.25 and 0.5 percent. This willminimize the potential for stormwater to transport sediments to the evaporation ponds or intoother drainage features. Erosion of the tailings will be further will be minimized by using specificsedimentation control measures as part of a set of best management practices (BMPs). To controlerosion of the dry stack tailings, the following measures will be utilized:

– Dry Stack TSF compaction – The general placement area may also be compacted to reduce thepotential for dust migration and to enhance the erosion resistance of the tailings from compaction.

– Equipment operations – The Dry Stack TSF shell will be developed in cells versus having place-ment and compaction equipment on the entire tailings surface which will reduce the potentialfor equipment-induced erosion and dust.

– Management of runoff collection/routing areas – For runoff directly from the dry stack tailings,ditches will be constructed along the perimeter of the Dry StackTSF and will have sedimentationtraps and siltation fencing as required for erosion control.

5.1 Perimeter ditches and evaporation ponds

Perimeter ditches will be constructed inside the perimeter of the placed tailings to collect runofffrom the surface of the Dry Stack TSF during operations. The ditches will need to be re-constructedconcurrent with tailings placement and are, therefore, considered a temporary facility duringoperations. As noted, it is expected that tailings in or near the perimeter ditches will be mobilizedduring larger storm events and collected in evaporation ponds. The evaporation ponds will also needto be reconstructed as the tailings surface is raised. Sizing of the evaporation ponds will depend onthe surface area of the tailings for each particular lift as well s any existing, undiverted watershedareas directly upstream of the tailings. Run-on from undiverted watershed areas upstream of thetailings is only an issue on the lower tailings lifts, or until the pit run rock buttresses are constructedon the west side of the dry stack facility.

5.2 Diversion Channels

There are two permanent diversion channels in the tailings design. Diversion Channel No. 1 willbe constructed at Project startup and Diversion Channel No. 2 will be constructed just prior toyear 12 of operation. Diversion Channel No. 1 starts just north of the ultimate pit configuration,intercepting an existing drainage and then continuing around the plant site. The channel eventuallydischarges into a natural drainage approximately 1/2 mile northwest of the plant site, flowing intothe McCleary Canyon, the northern limit of the Phase I Dry Stack TSF. About operations year12, Diversion Channel No. 2 will be constructed to receive up-gradient run-off including overflowfrom Diversion Channel No. 1. These diversion channels, in combination with detention basinsalong the channel alignments, were designed to handle storms of up to the probable maximum peak(PMP) event.

6 DRY STACK TSF ENGINEERING ANALYSES

6.1 Seepage analyses

Analyses were performed to estimate seepage rates, assess distribution of pore water pressure,and evaluate the degree of saturation within the dry stack tailings over the life of the facility. The

322

analyses utilized the finite element method (FEM) based computer program SVFlux Version 2.0.13developed by SoilVision Systems, Ltd. and part of the SVOffice 2009 Geotechnical Modeling suite.

Tailings within the Dry Stack TSF are anticipated to be deposited in an unsaturated condition,at average moisture content of 18 percent (by dry weight) or less. Based on pilot test data, thisas-placed moisture content is readily achievable from the filtration process, which is at or near theoptimum moisture content (as defined from Standard Proctor tests, ASTM D 698).

Meteoric influences were modeled as precipitation and evaporation on the surface of thedeposited dry stack tailings. Meteoric water (delivered as precipitation) will have a small rechargingeffect within the top several feet of the dry stack tailings. However, due to the semi-arid climate atthe Rosemont Project site, it is not likely to have a significant influence on the overall moisture con-tent and seepage through the tailings mass. Overall, evaporation generally produces a cumulativenegative flux across the surface of the tailings.

To assess the behavior of the tailings under unsaturated flow conditions, a series of moistureretention laboratory tests were conducted on representative samples. These tests were used todevelop a soil water characteristic curve (SWCC) for the tailings materials. The SWCC was develo-ped using the knowledge database within the computer program SVFlux. Based on testing data,the saturated moisture content of the tailings is approximately 25 percent (by dry weight) and thefield capacity moisture content is approximately 11 percent (by dry weight). The filtered tailingswill be placed within the Dry Stack TSF at a moisture content of approximately 18 percent orless. Therefore, the tailings will be placed at approximately 7 percent above the residual moisturecontent and approximately 7 percent below the saturation level. Based on these numbers, tailingswill be placed in the Dry Stack TSF as an unsaturated material. A limited amount of seepage willbe generated from the dry stack tailings as the material moisture content drains down from theas-placed value (18 percent) to the field capacity (11 percent). Some of this moisture will also belost to evaporation.

The results of the seepage analyses indicate that as the Dry Stack TSF expands over time, theestimated seepage rate increases to a peak value of approximately 8.4 gallons per minute (gpm) atproduction year 18 and then decreases after this time period. Model results show that the seepagefrom the dry stack tailings is due solely to drainage of pore water as the tailings gravimetric moisturecontent reduces from the as-placed value of 18 percent (by dry weight) to the field capacity valueof 11 percent (by dry weight). Since meteoric water (precipitation) influences are negligible thereis a finite amount of seepage that can occur from the facility.

Although the maximum tailings moisture content at placement is not anticipated to exceed 18percent, higher moisture contents would not significantly affect the peak seepage rates from thefacility since the hydraulic conductivity of the material is the controlling factor in pore waterresponse. Tailings placed in the Dry Stack TSF in excess of 18 percent moisture content will likelybe conveyed towards the center of the facility and spread out to promote evaporation and moisturereduction.

6.2 Stability analyses

Slope stability analyses were conducted in support of the final design of the Dry Stack TSF. Theanalyses required the selection of parameters from previous and current design work and examinedthe stability of the facility under both static and seismic loading conditions for several operationalscenarios.

The stability analyses were conducted at the maximum sections of the facility. Effective and totalstress analyses were conducted for all cross sections analyzed. In order to assess performance of theDry Stack TSF under conservative conditions, the models were developed under the assumptionthat the tailings become saturated in the center of the facility, within 1,100 feet of the upstreamcrest at any elevation. The material in this saturated zone was modeled as having no shear strength,with the material represented by only a unit weight with no internal angle of friction or cohesion.This approach is meant to represent major upset conditions where the tailings become saturatedand unconsolidated due to a temporary loss of filtration performance, reduced drainage paths, orextreme climatic events. However, large saturated zones are not expected within the Dry Stack TSFdue to the tailings being filtered, the arid climate, the subsurface soil conditions, and the proposedstormwater BMPs.

323

Table 1. Slope stability analyses results.

Analysis Static Factor Psuedostatic FactorPhase Modeled of Safety of Safety

Phase I Effective 2.3 1.2Total 1.9 1.0

Phase II Effective 2.3 1.2Total 1.9 1.0

Prescriptive Factor of Safety 1.3 1.0

Table 2. Sample characteristic summary and malvern laser diffraction PSA.

Year 1–3 Colina

Characteristic SummarySolids Specific Gravity 2.95 3.49Liquor Specific Gravity 1.0 1.0Solids Concentration (wt %) 85.0 84.0

Percent PassingP90 (µm) 154 169P80 (µm) 113 127P50 (µm) 47 57P10 (µm) 3.3 3.4Mean Size (µm) 65 73

As summarized in Table 1, the proposed Dry Stack TSF facility is stable under the conditionsnoted above as the computed values meet or exceed the prescriptive factors of safety (as definedin Arizona’s Best Available Demonstrated Control Technology [BADCT] document) for both staticand seismic loading conditions.

7 FILTRATION TESTWORK

The design moisture content requirements necessitated a series of filtration tests to determine theappropriate filtration technology for the Rosemont Project. These tests also were used to determinethe size and the design criteria for the filters. The impact of tailings thickener underflow densityon the filtration rate was also considered.

Bench scale filter tests were conducted on a composite sample that represented expected con-ditions for Years 1–3 as well as for a 4:1 mix of the Years 1–3 sample and a sample of Colina. Thebench-scale tests were used to identify filtration rates and filtration equipment sizing parametersas well as reviewing the effects of feed solids concentrations and flocculant dosages.

Particle size distribution for the Year 1–3 and Colina sample materials were similar. Based onthis particle size similarity and the test results, no additional blend tests were determined to benecessary.

Three types of dewatering equipment were investigated. Based on the results, it was determinedthe FLSmidth Eimco® Automatic Filter Press (AFP) Mark IV™ automatic filter press was theappropriate technology for the Rosemont application. The AFP will result in the fewest operatingfilters with the lowest cake moisture content applicable to the anticipated feed materials.

7.1 Sample characterization

Table 2 shows the result of the characterization work and particle size analysis performed on thetailings samples.

324

Table 3. Pressure filtration results.

Result

Feed Solids Concentration (weight %) 63Filter Media POPR 929MFilter Media Air Flow (standard cubic feet per minute per square foot) 12Cake Moisture (weight %) 16.0Cake Thickness (inches) 2Feed Rate (dry ton per day) 76,000Filtration Pressure (pounds per square inch gauge) 140Feed Time (min) 2Filter Cake Bulk Density (wet pounds /dry pounds per cubic foot) 139.0/118.0Filter Size (square feet) 5,448

7.2 Filtration test results

The filtration tests were performed on the Year 1–3 and Colina feed streams. The feed to the filtersis anticipated to be about 76,000 tons of dry solid per day. The tailings thickeners will be a HighDensity Thickener design. The tailings were fed to the filter at 60–63 weight percent solids, whichis approaching the maximum solid concentration for a high rate thickener in this application.

TheAFP Mark IV™ press filtration was simulated in a small single chamber bench test apparatus.The bench test was conducted by pumping feed slurry into the double-sided chamber. Filtrateproduction was measured with respect to time while the chamber filled. Once the chamber wasfull, the pumping was stopped and a cross-flow air stream was blown from one side of the chamberto the other was used to dewater the cake to the desired moisture. The filtrate production was alsomeasured with respect to time. The filtration tests are summarized in Table 3.

8 FACILITY CLOSURE CONCEPT

The primary goal of the closure/post-closure plan is to eliminate, to the greatest extent practicable,any reasonable probability of further discharge from the Dry Stack TSF and of exceeding aquiferwater quality standards at the applicable point of compliance. Based on the results of the geochemi-cal testing and modeling completed by Tetra Tech, no impact to the regional aquifer water quality isanticipated during the operational, closure, and post-closure periods of the facility. Therefore, thefocus of the closure/post-closure strategy is to minimize erosion and promote landform stability.

Pit run rock buttresses will be placed around the Dry Stack TSF, thus eliminating the potentialfor tailings exposure and subsequent erosion by wind or water on the outer slopes. Capping of thetop surface tailings with pit run rock will also be performed at closure.

Reclamation of the outer slopes of the Dry Stack TSF will begin the first year of operations.As much as practicable the outer slopes will be concurrently reclaimed and will be shaped tocomplement a natural, stable landform terrain. It is anticipated that a system of variable slopepatterns will be constructed on outslopes of the Dry Stack TSF. Stormwater control features willbe incorporated into the landform as the outer buttresses are raised and concurrently reclaimed.

9 CONCLUSION

The plan for the Rosemont Project is to develop an operation that will provide the balance betweenoperational reliability, cost effectiveness, and sustainability, specifically water conservation, whilestill processing copper ore at 75,000 tons per day. Dry Stack Tailings provides that balance andtherefore has become an integral part of the design work currently underway.

Dry Stack Tailings systems are a major step toward the future of tailings management in mining.While these systems are relatively new to the industry, the technologies used for the process are welltested. The use of reliable technology coupled with improvements in filtering equipment make Dry

325

Stack Tailings a more natural fit for new, Greenfield facilities. Sustainable mining techniques inareas around the world – especially in areas where water is scarce – will require that technologicaladvances be made in tailings processing and management.

REFERENCES

AMEC Earth & Environmental, Inc. 2008. Rosemont Copper Company Filtered Tailings Dry Stacks CurrentState of Practice Final Report. November 2008.

Lightall, P., Davies, M., Rice, S., and Martin, T. 2002. Design of Tailings Dams and Impoundments. MineralProcessing Plant Design, Practice and Control Proceedings Volume 2. 1828–1845.

Nowicki, K., FLSmidth Minerals. 2009. Confidential Report on Testing for M3 Engineers/Augusta ResourceCorp. Rosemont Project, Filtration Testing on Flotation Tailings, May 2008.

326


Dry stack tailings – design considerations

J. Lupo & J. HallAMEC Earth and Environmental, Englewood, CO, USA

ABSTRACT: Filtered tailings can be a viable option for managing tailings disposal at mines.Some of the advantages of filtered tailings include: i) reduction in water consumption, as moreprocess water can be recycled; ii) the filtered tailings can often be stacked (often referred as drystack tailings) to reduce the footprint for tailings storage; and iii) the dry stack tailings can oftenbe reclaimed concurrent with placement, thereby reducing reclamation costs. Filtered tailings havethese advantages over slurry, thickened, and paste tailings as the filtration process used (eithervacuum belt or mechanical presses) essentially advances the consolidation process (which can tensof years or more to achieve with traditional slurry disposal) to form an unsaturated cake. The filtercake often has a consistency of moist sand or silty sand, with geotechnical and hydraulic propertiesamenable to stacking and compaction. While thickened and paste tailings may be able to achievebeach slopes in the 3 to 6 percent range, dry stack tailings may be stacked with stable slopes in the20 to plus-30 percent range, with compaction.

This paper presents the primary geotechnical considerations for the design of dry stack facilities.The designs issues discussed in this paper include rate of stacking, stacking height, seep-age/infiltration, and settlement. In addition, this paper addresses some common misconceptionsregarding the geotechnical and hydraulic performance of dry stack tailings based on actual lab andfield data.

1 INTRODUCTION

Filtered tailings can be viable alternative to slurry, thickened, or paste tailings disposal at manymine sites. Some of the key advantages of filtered tailings disposal over the other methods includes:i) minimized water consumption (most of the water is recovered during the filtration process); ii)as discussed later in this paper, filtered tailings often have properties that are amenable to stacking,thereby reducing the land requirements for disposal; and iii) the filtered tailings stack provide astable land form, allowing reclamation to be conducted concurrent with disposal, which can saveboth time and cost for the operation.

Tailings filtration can be accomplished through either vacuum or mechanical press. The filtrationprocess basically accelerates the consolidation process that would naturally occur in other tailingsdeposits (slurry, thickened, or paste). The resulting filtered product is an unsaturated (often between50 to 75 percent saturated) material that is firm with a low compressibility and low hydraulicconductivity. In this state, the filtered tailings are often termed “dry stacked” tailings, since thematerials resemble a moist sand that can be stacked and compacted into a stable landform.

In the design of dry stack tailings facilities, it is important to consider the materials characteristicsthat are unique to dry stack tailings as well as operational and material handling considerations.These issues are discussed in the following sections.

2 DRY STACK TAILINGS CHARACTERISTICS

Before discussing dry stack tailings design issues, it is often useful to review some typicalcharacteristics of filtered tailings.

327

Figure 1. Tailing particle size gradations.

2.1 Particle size

Particle size is not a unique characteristic of filtered tailings. However, it is worth a brief discussionin terms of what particle size gradations have been successfully used at some dry stack operations.Figure 1 presents a plot of particle size gradations from several dry stack tailings projects. Thedata shown on this graph shows that successful filtered tailings projects have been completed onmaterials with very high fines content.

The particle size data presented in Figure 1 illustrates that dry stack tailings projects are notlimited to low fine content tailings. Advances in filtration processes and equipment have it possibleto filter relatively fine materials with good results.

2.2 Shear strength

The shear strength of dry stack tailings will vary, depending on the moisture content and density ofthe tailings, and drainage conditions within the stack. From a design standpoint, it is important torecognize that the density, moisture content, and drainage conditions within the stack are changingas more tailings are placed in the dry stack. In addition, the tailings density and moisture contentmay vary depending on filter efficiency and ore mineralogy. For the most part, dry stack tailingstend to exit the filtration unit near the maximum dry density (based on the Standard Proctordensity) and slightly wet of optimum. At these conditions, the tailings can have a high degree ofshear strength, and are suitable for placement and compaction. Figure 2 presents a plot of shearstrength (plotted in effective stress path space) from five different dry stack tailings subjected toConsolidated-Undrained (CU) triaxial compression tests.

All of the samples tested in Figure 2 were remolded near the density and moisture content ofthe tailings from the filtration units, and then saturated prior to testing. The purpose of the datashown in Figure 2 is to illustrate the relatively high shear strength that can be achieved with fine-grained, filtered tailings (without cement or lime amendment). The ultimate failure surface shown

328

Figure 2. Stress path plot – filtered tailing samples.

in Figure 2 could be described with an effective friction angle of 35 degrees, in terms of a traditionalMohr-Coulomb shear strength criteria.

In addition to the high shear strength, it is evident from the stress path that the filtered tailingsexhibit dilatant behavior under undrained compression (for the five tailings samples shown inFigure 2). This behavior reflects the relative dense condition of the tailings from the filtration units.The observed dilatant behavior in Figure 2 needs to be considered when assessing slope stability,deformations under seismic loading, and liquefaction assessments for dry stack facilities.

For the design of dry stack tailings facilities, it is recommended that the filtered tailings betested over a range of dry densities and moisture contents, to reflect the variability of materials thatcould be delivered to the dry stack. Filtration equipment (presses and belts) can generally providematerials with consistent dry densities and moisture contents, but some variability may occur dueto changes in ore mineralogy, loss of filter efficiency by plugging, or increase in throughput abovethe design level.

Dry stack operations may also experience brief periods when overly wet tailings may need tobe placed onto the dry stack facility (e.g. upset conditions). Depending on how these materialsare delivered (low density, high moisture content, rapid placement rate, etc), filtered tailings withcontractive behavior may be placed within the dry stack. Under shear, contractive tailings couldlead to very low undrained shear strengths within the dry stack. The placement of these materialswithin the dry stack is an important design consideration, which will be discussed in the Section 3.

2.3 Hydraulic conductivity

The hydraulic conductivity (saturated) of filtered tailings tend to be relatively low, typically lessthan 1 × 10−6 centimeters per second (cm/sec). The hydraulic conductivity will be dependent onthe amount of fines, mineralogy of the ore, and density of the material. Figure 3 presents a plot ofsaturated hydraulic conductivity measured from several samples of filtered tailings. The samplesshown in Figure 3, were remolded to the dry density and moisture content of filtered tailings. Thefines content (finer than 0.075 mm) of the materials tested ranged from 60 to 90 percent.

As shown, the measured hydraulic conductivities ranged from 6.7 × 10−6 to 1.7 × 10−8 cm/sec,indicating the tailings placed in the dry stack have low to very hydraulic conductivities. The lowsaturated hydraulic conductivity, combined with being placed as an unsaturated material, minimizesthat quantity of water (e.g. precipitation) that can infiltrate through the dry stack. Section 3 providesadditional discussion on the infiltration potential through dry stack tailings facilities.

2.4 Compressibility

The compressibility of filtered tailings is an important consideration for dry stack design, as itdirectly affects the stacking rate and configuration. As presented earlier, the filtered tailings are

329

Figure 3. Filtered tailings saturated hydraulic conductivity.

Figure 4. Typical tailing porosity with increasing dry stack depth.

typically delivered in an unsaturated state (usually between 50 and 75 percent saturation). However,the degree of saturation of the tailings is a function of porosity, which changes as the depth of thedry stack increases (i.e. lower porosity at the bottom of the stack, and higher porosity at the top ofthe stack). A plot of porosity versus dry stack depth is presented in Figure 4, showing the reductionin porosity due to the compression of the tailings.

330

Figure 5. Saturation versus tailing height.

If the filtered tailings are highly compressible, there is a potential for the porosity to decrease toa point where the dry stack tailings become fully saturated. Once a portion of the dry stack becomesfully saturated, then any additional loading (e.g. adding another lift of tailings), could give rise toexcess pore water pressures. The presence of either saturation and/or excess pore water pressureswithin the dry stack, could give rise to stability issues, therefore it is important to quantify thecompressibility of the tailings.

Figure 5 presents a plot of filtered tailings compressibility from several dry stack projects. Thetailings compressibility has been converted from that shown in Figure 4, to saturation versus depthof tailings. The format shown in Figure 5 can be readily integrated into dry stack facility design.As shown, the compressibility of the filtered tailings can vary from high compressibility (TailingsB and B2) to low compressibility (Tailings A).

The curves presented in the compressibility plot show that for design, saturated conditions withinthe dry stack would be anticipated to develop within 25 m of stack height forTailings B; within 50 mfor Tailings B2 and D; and saturated conditions are unlikely to develop in Tailings A, A2, and C.It must be noted that the development of saturation within a dry stack is not a fatal flaw, rather itneeds to be addressed in the design so that stability of the facility is not negatively impacted.

3 DRY STACK TAILINGS DESIGN CONSIDERATIONS

The discussions presented in Section 2 provided some general observations regarding the charac-teristics of filtered tailings shear strength, permeability, and compressibility. This section discussesintegration of these characteristics into design.

3.1 General design approach

The design of a dry stack tailings facility is a balance between materials handling and constructionwith the geotechnical properties of the tailings. From a materials handling standpoint, the facilityneeds to be designed to support the method of transport and deposition of the tailings to the facility(either conveyor or truck haulage). Therefore, the overall layout needs to include any access roads orconveyor corridors. The site topography and distance from the filter plant may dictate the method ofdelivery, however the tailings properties (moisture content) may also influence this decision. During

331

Figure 6. Generalized dry stack cross-section.

the operation of the dry stack, provisions must be made to handle overly wet tailings, should upsetconditions occur in the filter plant. Therefore, the design of the materials handling systems andequipment should be robust to handle a wide range of materials (unsaturated to saturated tailings).

Placement and compaction of the tailings will also need to consider a wide range of materialsduring operation. In general, it is good practice to place and compact unsaturated tailings along thedownstream end of the facility, forming a structural zone to support the dry stack facility. Behindthe structural zone, the tailings can be placed with no or low compaction. These tailings may alsoinclude overly wet materials that may have low shear strength or high compressibility.

The size and extent of the structural zone can developed based on stability analyses to support andcontain the materials behind the structural zone. The design basis for the structural zone needs toconsider material strengths, static and seismic loading, and potential water management concerns.A generalized cross-section through a dry stack facility is shown in Figure 6.

In some cases, a rockfill buttress or rock mulch may be added along the outside slopes of thedry stack facility to enhance stability, provide benching for slope reclamation, and for erosionprotection of the out slopes.

3.2 Stacking height

The design of the dry stack height is generally based on slope stability analyses conducted onseveral sections through the dry stack facility. The stability analyses need to consider:

• Shear strength of all the materials within the section (tailings, foundation, rockfill buttress, etc).The tailings shear strengths need to be derived based on drained or undrained laboratory testing.The drainage conditions (drained/undrained) assumed in the stability analyses should reflectthose that could develop through operation. Data, such as that presented in Figure 5 can beused to establish if zone(s) of saturation may develop within the dry stack. If saturation may bepresent, then the designer will need to conduct some additional engineering analyses to estimatethe magnitude of excess pore pressures that may be present (if any).

• Presence of a static groundwater surface. As indicated previously, the presence of groundwater isnot fatal flaw (unless the tailings must remain dry for geochemical reasons) in a dry stack design.The design must be developed to provide a stable landform in the presence of the groundwatersurface.

• Seismic loading and residual shear strengths. Seismic loading must be considered for slopestability for operational and post-closure configurations. Depending on the location and materialproperties of the tailings, seismically-induced permanent deformation should also be assessed.These assessments need to consider the dilatant behavior of filter tailings (if it is present) asshown in Figure 2. Material dilatancy may have a significant effect on post-earthquake materialstrengths and development of permanent deformations.

3.3 Stacking rate

Stacking rate needs to be considered in cases where the tailings exhibit a high degree of compress-ibility (see Section 2.4). Highly compressible tailings, if stacked too quickly, may generate excess

332

Figure 7. Example of infiltration through dry stack tailings.

pore pressures, which could compromise stability of the dry stack. If high compressible tailingsare to be included the dry stack design, the following approaches may be considered:

• Include multiple deposition points, so that tailings can be distributed across the facility, effec-tively reducing the stacking rate. This is often the most cost-effective and practical approach tocontrol compressible materials.

• Design a substantial structural fill zone to support low shear strength tailings mass. While thiscan easily be designed, the cost of construction and scheduling needs to be considered as partof the feasibility of this approach.

3.4 Infiltration and surface water management

One of the most common misconceptions with dry stack tailings is that surface water (e.g. precip-itation or run-on) will readily infiltrate through the tailings, and could saturate the stack leadingto failure or sloughing. Experience from actual operating dry stack operations have shown thatinfiltration through the dry stack is not a significant issue, as long as proper surface water man-agement controls are employed. As shown in Figure 3, the saturated hydraulic conductivity offiltered tailings is typically quite low. This, in combination with the fact the tailings are placed inan unsaturated state, results in very low infiltration through the dry stack. Even if water is pondedat the surface, the infiltration rate will be limited by the low saturated hydraulic conductivity.

The rate at which the infiltration front moves through the tailings can be assessed using one ortwo-dimensional seepage models. An example showing the infiltration front moving through a one-dimensional section of dry stack tailings is shown in Figure 7. In this example, water is ponded on

333

the surface of the dry stack. The saturated hydraulic conductivity of the tailings is 5 × 10−7 cm/sec.As shown in the figure, the infiltration front requires a very long time to move through the drystack column, and does not readily infiltrate.

For most dry stack tailings facilities, the majority of the precipitation runs off the surface (due tothe low hydraulic conductivity of the tailings) and can be collected by surface water managementchannels. It is prudent to design the dry stack facility with the upper surface graded to drain wateraway from critical stability areas (e.g. downstream buttresses or structural zones) during the wetseason. Integrating surface water collection channels into the dry stack design can minimize thepotential for infiltration through the stack and provide a way manage surface water from the facility.

4 CONCLUSIONS

Dry stack tailings facilities can be a viable option for tailings management at mining operationsby minimizing water consumption, minimizing the footprint for tailings storage (by stacking), andallowing concurrent reclamation of the dry stack, thereby reducing costs. The design of the dry stackfacilities needs to consider the geotechnical properties of filtered tailings. This paper presents actualtest data from several filtered tailings projects and discusses the characteristics of the materials interms of design. Key characteristics include relatively high shear strength at the dry density andmoisture content of the filtered tailings, low hydraulic conductivity, and compressibility.

Design considerations for dry stack tailings facilities are provided based on performancefrom actual facilities. The design considerations can be used to integrate materials handling andconstruction issues with the geotechnical characteristics of the filtered tailings.

334


Reprocessing of tailings of Chador-Malu iron ore, Iran

H. NematollahiUniversity of Tehran, Iran

ABSTRACT: During the last twelve years of exploitation and production at Chador-Malu ironmine, about 27 Mt of tailing is accumulated containing about 31% Fe. This material could bereground as a feasible reserve. The studies of this tailing showed that by regrinding and reprocessing,using high gradient magnetic separation and flotation, a concentrate of 63.31% Fe and 0.058% P,similar to actual flotation concentrate, could be produced.

1 INTRODUCTION

Chador-Malu iron mine is located at 165 km north east of city of Yazd, in central desert of Iran.The exploration of this reserve carried out during 1966-1970 by NIOC (National Iranian SteelCorporation) under the supervision of ex-Soviet geologists. The reserve was estimated to 400 Mt,containing 55% Fe in average. The recent evaluation of the reserve has approved only 320 Mt of ore.

Exploitation of this mine began on 1997. At the beginning, the ore was processed in three parallellines, each one with 392 t/h capacity. Since beginning of 2008, a fourth line was added. Iron gradeof tailing, especially during the first years of operation has been relatively high and has rangedbetween 42% and 26%, averaging 30.80%.

This research has been carried out on reprocessing of the tailing. The microscopic studies oftailing show the interlock of hematite with phosphate and gangue minerals. Regrinding of the tailingand then reprocessing by using high gradient magnetic separation and finally flotation, 39% of ironcould be recovered with 63.31% Fe and 0.058% P, which is similar to actual flotation concentratein Chador-Malu mine.

2 MINERALIZATION

In this mine, the main mineralized zones occur in igneous and metamorphic rocks, includingultramafic to mafic and intermediate igneous rocks with reported age as Cambrian and retrogressivemetamorphic products of such igneous rocks, in particular greenschist. There are two main ideasabout the genesis of this mineralization: (1) active metsomatic processes, and (2) separation of iron-rich melt from magma and subsequent emplacement of such melt in this geological environment.

3 MINING

The annual mine ore tonnage was initially designed as 7 Mt. This tonnage was obtained on the sixthyear of commissioning when the three parallel processing lines were active. After the installationand commissioning of the 4th line (January 2008), and some modification in the previous lines,the annual ore tonnage mined increased to more than 12 Mt. Mining operation as well as mineralprocessing is carried out by the contractor, Asfalt Tous Co., a Persian contractor. The contract hasbeen renewed every three to five years.

Major equipment currently in use include two 250 mm pull-down rotary blast hole drills, three11 m3 cable shovels, four 13 m3 hydraulic shovels (one diesel and three electric), twelve 120 ttrucks. Recently, thirteen new 120 t trucks were procured for recent and future development ofthis mine. Mining crews work 24 h a day, 355 days a year (10 days of holiday for Persian NewYear “Now-rouz”). The iron ore, after exploitation, is transported by trucks to the gyratory crusher.

335

Figure 1. Simplified flowsheet of Chador-Malu processing plant.

The crushed ore is stocked in stockpile in two alternative blending beds, and then transferred by areclaimer to a conveyor belt, discharged in bins at the beginning of each line.

4 PROCESSING

The simplified flowsheet of Chador-Malu Processing is shown in Fig. 1, for the three initial lines.The major equipments in each line include an autogenous mill (D = 9.7 m, L = 4 m), low intensitymagnetic separator (LIMS) as cobber, ball mill (D = 5.4 m, L = 9.25 m), LIMS as cleaner, highgradient magnetic separator (HGMS), ball mill (D = 3 m, L = 5.9 m), flotation cells for hematitebeneficiation and flotation cells for apatite processing as by-product. The fourth line is a littledifferent with the three previous ones, in the size and power of ball mill, the size and number ofthickeners and the size of filters. There is no section to recover the apatite as by-product.

In each processing line, the ore is ground in the autogenous mill to d80 = 180 microns, then mag-netite content is separated by cobber LIMS, which is reground in the ball mill to d80 = 35 microns,and reprocessed by four stages of cleaner LIMS to produce the magnetite concentrate of 68.75–70% Fe and ≤0.05% P. The rest of this stage is processed by the HGMS to recover the hematitecontent which is dephosphorized, using flotation method to produce the hematite concentrate of60–67% Fe and ≤0.05% P. The final concentrate which is the mixture of magnetite and hematiteconcentrates with ≥67.5% Fe and ≤0.05% P, is used in direct reduction plants, at Isfahan an Ahvaz.The by-product of the processing lines is phosphate concentrate that is produces by flotation fromthe HGMS tailings.

5 TAILING REPROCESSING

During the past 12 years of plant operation, about 73 Mt of iron ore with 54.15% Fe is processed,producing about 46 Mt of concentrate with 67.82% Fe and 27 Mt of tailings with 30.80% Fe. TheX-RD examination of the tailing sample indicates the presence of the following minerals:

Hematite, goethite, ankerite, clinochlore, biotite, dolomite, albite, anorthite, quartz, fluorapatite,kaolinite, amphibole.

A representative portion of the sample is screen analyzed and the different size fractionsare studied, using microscope method. This study showed that in coarse fractions, hematite is

336

Table 1. Flotation results of HGMS concentrate.

GradeWeight Recovery(%) Fe % P % Fe %

Feed (representative sample of tailing) 100 33.85 2.204 100HGMS concentrate 34.20 54.89 0.596 55.46HGMS tailing 65.80 22.91 3.040 44.54Flotation concentrate 24.09 60.0 0.19 42.70Flotation tailing 10.11 42.71 1.563 12.76

Figure 2. Size distribution of feed, overflow and underflow of de-sliming hydrocyclone.

interlocked with gangue minerals, therefore liberation of hematite needs grinding of such materialto d80 = 25µm.

Processing the sample in a HGMS, produce a concentrate with 55.89% Fe and 0.590% P.To increase the iron grade and reduce the phosphorus grade, the HGMS concentrate should beprocessed by flotation method, under the following conditions:

– pH: 10 (adjusted by soda ash)– water glass: 800 g/t– water glass: 800 g/t– fatty acid: 600 g/t– fuel oil: 200 g/t

The result of HGMS and flotation tests are shown in table 1.The above product could be considered as convenient material for blast furnace.Since, numerous direct reduction plants are active in Iran, and the feed to these plants should

contain more than 67.5% Fe and less than 0.05% P, the studies are carried out to obtain a concentrateconvenient for such plants.

One of the obstructions to modify the concentrate quality is the slime content of the ore. So, ade-sliming stage is followed to the grinding stage. The de-sliming is carried out by a hydrocyclonewith a d50 of 4 µm. The size distribution of ground material, slime and de-slimed fractions areshown in Figure 2.

The de-slimed material is passed through HGMS, followed by flotation. The results are shownin Table 2.

337

Table 2. Results of processing of de-slimed material.

GradeWeight Recovery(%) Fe % P % Fe %

Feed (representative sample of tailing) 100 33.54 2.278 100Slime 39.75 34.98 1.576 41.46De-slimed feed 60.25 32.59 2.741 58.54HGMS concentrate 27.81 55.63 0.653 46.12HGMS tailing 32.44 12.84 4.531 12.42Flotation concentrate 20.66 63.31 0.058 39.00Flotation tailing 7.15 33.44 2.372 7.12

Figure 3. Suggested simplified flowsheet for reprocessing of tailings.

Based on the above obtained results, the simplified flowsheet for the tailing reprocessing issuggested as is shown in figure 3.

6 CONCLUSION

Some deficiencies in design of Chador-Malu Iron Ore Concentration Plant has resulted in encoun-tering problems in all four existing production lines. Consequently, part of iron minerals is notliberated, therefore, final iron ore recovery is relatively low and the Fe grade of final tailing is high.To solve the problem, modification of existing produc-tion lines is not recommended, as the majorproblem arises from the size and power of grinding mills and there is no enough space in the Plantto add another mill to the line.

A research program is carried out on the tailing, deposited during last 12 years. Reproc-essingof the tailings is possible and feasible by regrinding it to appropriate size (d80 = 25 µm) andre-processing it by a combination of high gradient magnetic separator followed by flotation process.

By making use of this method, 42.7% of lost iron is recovered as an acceptable iron concentrate.

338

Oil Sands


Suncor Pond 5 coke cap – The story of its conception, testing,and advance to full-scale construction

Patrick Sean WellsSuncor Energy Inc., Fort McMurray, AB, Canada

Jack CaldwellRobertson GeoConsultants, Vancouver, BC, Canada

Joseph FournierSuncor Energy Inc., Fort McMurray, AB, Canada

ABSTRACT: The Pond 5 Tailings Impoundment at the Suncor Energy Oil Sands Mine northof Fort McMurray, Alberta, was turned over to the mine’s reclamation department in 2009 forclosure. At the beginning of 2010, construction of a pond-wide coke cap, part of the proposed finalcover, was begun. In the two years preceding construction of the pond-wide coke cap, the authorsundertook extensive laboratory and field trials of a prototype coke cap. This paper describes thetheoretical and practical work done to formulate a viable and safe coke cap—in effect the cap thatis now under construction across almost the entire pond. We describe field testing to characterizethe tailings; laboratory testing undertaken to characterize the response of the tailings to plannedconstruction procedure; and two prototype covers constructed in the winters of 2008 and early 2009to test and confirm theoretical analyses and designs.

1 INTRODUCTION

Suncor Energy, Inc. plans to close the Pond 5 tailings impoundment at the Suncor oil sands minenorth of Fort McMurray by 2019. This paper describes the design and construction between 2008and 2010 of a coke cap over the soft tailings; the coke cap will serve as the basal layer of the plannedfinal cover that will include a thick sand layer, soil, vegetation, drainage swales, and a lake. Inaddition, the coke cap provides access to the surface of the tailings for equipment and possibleinstallation of wick drains that may be used to promote dewatering and hence consolidation of thetailings.

2 DESCRIPTION OF POND 5

Pond 5 has been in use since the mid 1950s. It is roughly fan-shaped and extends about 3 kmnorth south and about 5 kilometers east west (Figure 1.) The west and north perimeter dikes wereconstructed of tailings sand to a height of upwards of 100 m to enclose a mined-out open pit.The tailings vary considerably. The upper layer is very soft, low strengths (less than 1 kPa) clayeymaterials that are still essentially fluid at void ratios of up to five. The strength and sand contentincrease with depth to the bottom of the pond that is deeper than 50 m. (Figures 2 and 3).

3 THE SAFE COVER

In 2008 and 2009 the authors and many others, recognized at the end of this paper, conceivedof what we called the Semi-Anchored, Floating Experimental (SAFE) Cover. We recognized thatthere is a large volume of coke available at site and the coke is essentially a waste product. Thecoke is light and floats on the fluid tailings.

341

Figure 1. Suncor Tailings Ponds – Fort McMurray, Alberta.

Figure 2. 2009 pond 5 solids section B.

On the basis of successful covering of uranium mill tailings at Wismut, we decided to constructa prototype cover as follows (Figure 4):

Pack or remove winter snow to induce freezing of the upper tailings

• Place a geotextile over the frozen tailings which constitutes a safe access surface. Adjacent rollsof geotextile were overlapped.

• Place a geogrids over the geotextile with the long axis of the geogrids perpendicular to the longaxis of the geotextile roll and parallel to the short access of the section. Adjacent rolls of geogridwere overlapped.

342

Figure 3. Estimated pond-wide void ratio distribution.

Figure 4. 2008 & 2009 access road section.

• Place an initial lift of coke, one-meter thick, using light equipment, primarily SnoCats.• Place subsequent lifts of coke up to three-meters thick using heavier equipment.• Along the Access Road thus constructed continue to transport coke using heavy equipment,

including Moxies that weight up to 60 tons fully-loaded.

4 DESIGN ANALYSES

The concept is simple: the coke floats on the upper tailings (essentially at a factor of safety ofone) and the coke is held “together” by the basal geosynthetics. One almost needs no more thanArchimedes Principle to establish the geometry of the floating mass.

In spite of this simplicity, we undertook extensive FLAC analyses of various sections and differingequipment loads (Figures 5 and 6.) FLAC is a finite element code that enabled us to replicate boththe fluid and soil-like response of the tailings to the loading of the coke cap and imposed equipment.

343

Figure 5. Typical FLAC output showing vertical deformation of the tailings and a result of the presence ofthe access road.

Figure 6. FLAC analysis of heavy load on coke access road

5 2008 & 2009 CONSTRUCTION

In 2008 we constructed a series of coke cap access roads out over beach materials at the south westcorner of the pond (Figure 7). In 2009 we extended a coke cap access road out from the west dike

344

Figure 7. 2010 design overview.

Figure 8. 2010 access road – cross section.

of the Pond and over deep, soft tailings. On the basis of the success of these prototypes, a decisionwas take to proceed to full-scale construction in the winter of 2010.

6 2010 CONSTRUCTION

To provide for pond-wide coke delivery, it was decided to construct a series of access roads acrossthe pond (Figure 8). The design of these access roads is essentially as for the access roads of the2008 and 2009 road, except that a stronger, seamed geotextile was used.

Because the winter was shorter than usual only eleven kilometers of access road was constructed.It is planned to construct the remainder in the winter of 2010 and 2011. Construction proceededwithout any incidents.

ACKNOWLEDGEMENTS

We acknowledge Andy Robertson who knew the Wismut covers and had the insight to apply theideas to Suncor. He supported and encouraged us through all phases.

345

Peter Byrne, now a retired UBC professor, and Ernie Naesgaard, and independent consultantswere there to remind us of the basics of fundamental soil mechanics and to undertake the FLACanalyses.

Gordon McKenna of BGC and Richard Dawson and Erin Olsun of Norwest provided invaluablepeer review and new ideas through all phases of work.

AMEC and their staff compiled the 2010 detailed designs and oversaw construction. We cannotnote them all, so refer to but two: Gordon Pollack and Ed McRoberts who had the courage to takeour small-scale success and turn them into large-scale successes.

None of this would have happened without the Suncor field staff: Neil Jevning and Ivan Deckerwho managed all construction, Joseph Fournier who looked after things from his office overlookingthe pond, and the managers of the Suncor reclamation group, Bill Tully and Mat Le Blanc, whohad to manage us and the large budgets.

346


Treatment of fluid fine tailings with silica

R.H. MoffettE.I. DuPont de Nemours and Company, Wilmington, DE, USA

ABSTRACT: Management of fluid fine tailings remains an environmental challenge to the oilsands industry. Recently enacted ERCB directive 074 requires fine tailings be treated to generatea trafficable surface. The ERCB directive dictates the minimum undrained shear strength of thedeposited fines shall be 5 kPa at 1 year after treatment and 10 kPa after 5 years. Rheologicalmodification of the fluid fine tailings can be easily accomplished by in-situ polymerization ofsilica. Yield stress of the tailings can be increased an order of magnitude within minutes aftertreatment. With minimal dewatering the treated tailings can exceed the one and five year shearstrength requirements. Dewatering of the treated tailings can be accomplished by a number ofdifferent techniques. The in-situ silica polymerization treatment has been demonstrated to workover a wide range of fine tailings compositions. This paper will discuss rheology modification ofoil sands fluid fine tailings using silica and the subsequent dewatering of the tailings.

1 INTRODUCTION

Extraction of bitumen from Canadian oil sands generates vast quantities of fluid fine tailings.Recently enacted ERCB directive 074 specifies performance criteria for the reduction of fluidtailings and the formation of trafficable deposits. The directive requires operators to reduce fluid tail-ings through fines capture in dedicated disposal areas in a manner that creates trafficable deposits.These deposits must have minimum undrained shear strength of 5 kPa within one year of depositionand 10 kPa within 5 years.

Historically oil sands tailings have been placed in ponds or impoundments where the fine tailingsconsolidate with time into what is termed mature fine tailings (MFT). It is estimated it will takecenturies for the tailings in these ponds to reach the consistency of soft clay (Lord 1998). Even ifthe MFT consolidates to 60 wt% solids they yield stress is only about 0.5 kPa which is far short ofthe ERCB 1 year requirement (Wells 1997).

Since the 1990’s oil sands operators have sought to create geotechnically stable fine tailingsthrough the use of NST or CT technology (Non Segregating Tailings and Consolidated Tailingsrespectively). Variations exist for these techniques but basically require a specific blend of finetailings and coarse sand tailings along with a chemical coagulant to prevent separation of the finesfrom the sand after deposition. A number of chemical coagulants have been proposed and testedthrough the years. Gypsum (Omotoso 1999) and carbon dioxide (Mikula 2006) have found themost commercial success for this purpose. Reported drawbacks to the CT and NST process arethe necessity for large volumes of sand which is preferentially used to create dikes, the creation ofponds and lack of robustness (MacKinnon 2007).

Another approach to fine tailings management has been the use of high molecular weight floccu-lants to facilitate solid/liquid separation. These organic polymers can be used in conjunction withmechanical thickeners to generate a more highly concentrated tailings stream to avoid creation oftailings ponds and MFT. Underflows from thickeners are reported to be as high as about 50 wt%solids. However at 50 wt% solids, the thickened tailings have a yield strength of only about 50 Paand will not meet ERCB yield stress requirements without further drying (Lord 1998). The poly-mer thickened tailings must have a solids concentration of nearly 75 wt% to achieve a yield stressapproaching 5 kPa.

347

Flocculation of MFT with high molecular weight polymers with subsequent drainage and dryingin thin lifts has been also trialed. Reportedly this approach will allow MFT to self dewater to about55 wt% solids. However, this treatment technique suffers from requiring very specific mixingconditions to achieve good drainage of the flocculated MFT. Polymer flocculated tailings arereported to be very susceptible to shear thinning (Mikula 2006). Mikula also notes that driedpolymer treated MFT will more easily reabsorb rain water than lime/gypsum treated MFT (Mikula2008a).

Mechanical assisted dewatering of polymer flocculated fine tails has also been evaluated. Cen-trifugation has been reported to require less high molecular weight polymer compared to thin liftdrying, however the chemical cost savings is offset by the large capital expense of the centrifuges.What’s more the solids from the centrifuge need to be conveyed or trucked to the disposal arearather than being simply pumped as the solids cake discharged from the centrifuge is reported tobe quite susceptible to shear thinning. Mikula states the yield stress decreases from approximately6 kPa to about 0.3 kPa after shearing (Mikula 2008b).

2 SILICA FOR FINES TAILINGS TREATMENT

Silica (SiO2) in the form of sand is the major constituent in the mined oil sands ore. Silica mayhowever take numerous other forms, some of which are water soluble while others are waterdispersible. Many of these silica sources can be made to polymerize into three dimensional chainednetworks at relatively low SiO2 concentrations. We have found these silica networks can alsobe formed in-situ within the water present in fluid fine tailings (Moffett 2010). The networksare capable of making major changes in the rheology of oil sands fine tailings at relatively lowsilica concentrations. Upon dewatering and/or drying the yield stress of the silica treated tailingsincreases.

In-situ silica polymerization and network formation can be initiated by the addition of smallamounts of various chemicals along with the silica source. Totally inorganic treatment systems canbe practiced by initiating silica polymerization by the addition of bi or tri-valent metal salts suchas calcium chloride, magnesium sulfate or sodium aluminate. Alternatively, acids or organic esterscan also be used to initiate silica polymerization. Silica network formation can be accomplishedin seconds or over days depending upon the choice of initiator, dose and silica source and optionaluse of accelerants. Perhaps the simplest and safest initiator is carbon dioxide.

The silica network provides strength to the treated MFT while encompassing the water andsolids until dewatering is desired. A small applied pressure will cause the water contained withinthe network to exude. The solids however remain trapped within the silica matrix. Water can alsobe removed from the network through evaporative drying. The ease of water release allows formultiple potential methods to dewater the treated tailings. Some dewatering methods that could bepotentially utilized include; in-pit self-weight consolidation, centrifugation, pipe-line dewatering,belt pressing, and thin lift deposition. Upon loss of water the silica treated tailings consolidate withsignificant reduction in volume

3 EXPERIMENTAL

MFT was obtained from a major oil sands producer. Analysis showed the MFT to contain 27 wt%solids, and 73 wt% water. Particle size distribution was determined by static light scattering usinga Malvern Instruments Mastersizer 2000. Results indicate >90% of the particles are smaller than44 micron (Figure 1). Yield stress of the MFT was determined to be approximately 5 Pa using aBrookfield DV-III+ HB rheometer equipped with a vane spindle rotating at 0.1 rpm. The Brookfieldrheometer can measure yield strength up to 8.8 kPa.

MFT was dosed with various amounts of silica in the form of sodium silicate. Silica polymeriza-tion was initiated within the MFT by addition of gaseous carbon dioxide. The treated MFT sampleswere stored in sealed containers to prevent evaporation. Yield stress was measured approximately24 hours after the treatment (Figure 2). The treated samples were then exposed to the environmentto allow water evaporation. Yield stress and solids concentration of the samples was measured overtime as the water evaporated. The yield stress and solids concentration of an untreated MFT sample

348

Figure 1. MFT particle size analysis by static light scattering.

Figure 2. MFT treated with silica, yield stress versus solids concentration.

was also measured with time. As can be seen in Figure 2 creation of the silica networks significantlyincreases yield stress compared to untreated MFT. Yield stress of the silica treated samples rapidlyincreases with limited drying. Yield stress values meeting the ERCB 1-year requirement of 5 kPaare achievable at MFT solids concentrations of <50 wt.

349

Figure 3. Silica treated MFT compared to gypsum/lime treatment.

Thin lift deposition and evaporative drying of MFT has been discussed by Wells (Wells 2007).To assess the evaporative drying potential of silica treated MFT, 5 and 10 cm thick lifts were pouredonto sand beds in 4 liter plastic beakers and allowed to dry. As a comparison, a 10 cm thick lift ofMFT treated with gypsum and lime following the recipe disclosed by Wells was also prepared. Fulldepth core samples of the MFT for moisture analysis and yield stress measurements were takenover a 9 day period. The drying rate for the silica treated MFT was found to be similar to thatof the gypsum/lime treated material (see Figure 3). However, the yield stress of the silica treatedMFT was approximately an order of magnitude higher than the gypsum/lime treated MFT. Forexample, after 7 days of drying the gypsum/lime treated MFT had a moisture content of 50 wt%and a yield stress of 1414 Pa. Comparatively after the same 7 days of drying the 5 cm silica treatedlift moisture content was 48.2 wt% and the yield stress of >8800 Pa. The 10 cm silica treated liftmoisture content was 43.7 wt% and the yield stress 8520 Pa.

Tailings treated by the CT process dewater through self consolidation. To investigate the potentialof silica treated MFT to dewater via self-weight consolidation a pressure cell was employed. MFTwas treated with sodium silicate at 6.8 kg SiO2/tonne of MFT solids. Silica network formation wasinitiated by CO2 addition and addition of 5.4 kg of gypsum/tonne of MFT solids. The cell wasequipped with a perforated metal plate and coarse filter paper through which water exuded fromthe treated MFT could pass (Figure 4). The cell was filled with 7.4 cm of silica treated MFT ontop of the filter paper. A load of 17.9 kPa was then applied and maintained to the top surface of theMFT while the water passing through the filter paper was collected and weighed. When essentiallyno further water was being exuded, the applied load was increased to 38.6 kPa and then finally to amaximum of 73.1 kPa. MFT solids concentration within the cell at various intervals was estimatedbased upon the mass of water collected (Figure 5). Upon completion of the test the pressure cellwas disassembled and the consolidated solids removed. The solids thickness had been reduced by55% from 7.4 to 3.4 cm. Some free water was found to have drained from the MFT but was trappedwithin the pressure cell. This water was not included in the estimated MFT solids calculations. Asa result, the solids concentration of the MFT left within the pressure cell was found to be somewhathigher than what was estimated based upon the captured water (Figure 4).

In another dewatering experiment 3.7 cm of silica treated MFT was placed between two layersof water saturated tailings sand to simulate in-pit dewatering. The water saturated sand initially

350

Figure 4. Pressure cell.

Figure 5. Dewatering of silica treated MFT under pressure.

contained 83 wt% solids and 17 wt% water. In this test the cell was immediately pressurized to 10 psiand this pressure was maintained for 20 hours. After 20 hours the sand and MFT were removedfrom the cell and the solids content of both the sand and MFT was determined. The solids contentfor both the upper and lower sand layers was found to be 82 wt%. The MFT solids concentration

351

Figure 6. Shear effects on MFT treated with 6.8 kg/tonne SiO2 + CO2 and 5.4 kg/tonne accelerator.

was determined to be 56 wt%. This is essentially the same MFT final solids concentration as shownin Figure 5. Visual observation indicated the MFT did not significantly penetrate into the sand beds.

Polymer treated tailings have been reported to be very susceptible to shear thinning duringtransportation through a pipeline or after deposition. Since the rate of silica network formation canbe controlled over a wide range of time, there is minimal risk of disrupting the silica network in thepipeline carrying the tailings to the dedicated disposal area. There is however a risk that the treatedtailings could somehow be exposed to shear after they have been placed in the disposal area. Togage the impact of shear after silica network formation, MFT was treated with sodium silicate at6.8 or 13.6 kg SiO2/tonne MFT solids. Silica polymerization was initiated in two different manners:

– By saturation with CO2 plus addition of 5.4 kg/tonne of an inorganic accelerator– By addition of 27 kg/tonne of an inorganic accelerator

After treatment the samples were stored in sealed containers. Yield stress of the silica treatedMFT was measured at 24 and 48 hours. After 24 hours a portion of the treated MFT was subjected tointense shear by remixing the solids in a Kitchen Aid planetary mixer for 60 seconds. The remixedMFT was placed in sealed containers and its yield stress was measured at 5 minutes, 24 and 48 hours.The tests indicate the CO2 treated MFT recovered approximately 63% of its unsheared yield stresswithin 24 hours and was essentially unchanged after 48 hours (Figure 6). The magnesium sulfatetreated MFT recovered approximately 51% after 24 hours and 79% after 48 hours (Figure 7). After48 hours storage, both the treated MFT and the remixed MFT were exposed to the environment andallowed to evaporate for 24 hours. The yield stress and solids content were then re-measured. Solidsconcentration increased approximately the same in the remixed and the unsheared samples. Yieldstress was also found to be equivalent between the remixed and unsheared samples. The yield stressof both remixed samples increased by an order of magnitude and was essentially the equal to theunmixed samples. Evaporative water loss was also equivalent between the unmixed and remixedsamples.

352

Figure 7. Shear effects on MFT treated with 13.6 kg/tonne SiO2 + 27 kg/tonne accelerator.

4 CONCLUSIONS

In situ polymerization of silica offers a unique method to treat oil sand fine tailings streams.Formation of silica networks within the water phase of fluid fine tailings can significantly increasethe yield stress. Treated tailings can be easily dewatered by a number of methods including selfweight consolidation and evaporation. Laboratory tests indicate 5 kPa yield strength is achievablewith at less than 50 wt% solids which is significantly wetter than what has been reported for polymertreated tailings in the open literature. Silica treated tailings recover a large portion of their initialyield strength after experiencing high levels of shear, and fully recover with limited drying.

REFERENCES

Lord, E., Liu,Y. 1998. Depositional and Geotechnical Characteristics of Paste Produced from Oil SandTailings,Fifth International Conference on Tailings and Mine Waste ’98, Fort Collins Co, 1998: 147–157.

MacKinnon, M. 2007. Surface Oil Sands Operations Will Affect Water Quality and Impact Options for WaterManagement, Part II: Process Effects, Conrad Water Usage Workshop, November 2007.

Mikula, R. J., Omotoso, O. 2006. Role of Clays in Controlling Tailings Behavior in Oil Sands Processing, ClayScience, 2006, 12 Supplement 2: 177–182.

Mikula, R.J., Munoz, V.A., Omotoso, O. 2008. Water Use in Bitumen Production: Tailings Management inSurface Mined Oil Sands, Canadian International Petroleum Conference, 2008: 1–8.

Mikula, R.J., Munoz, V.A., Omotoso, O. 2008. Centrifuge options for production of “Dry stackable tailings”in surface mined oil sands tailings management, Canadian International Petroleum Conference, 2008: 1–8.

Moffett, Robert H. 2010. U.S. patent application publication 2010/0104744 A1.Omotoso, O.E., Mikula, R.J. 1999. Alternative Consolidated Tailings Chemicals, Canmet division report WRC

99-23, May 1999.Wells, P.S., Riley, D.A. 2007, MFT Drying – Case Study in the Use of Rheological Modification and Dewatering

of Fine Tailings through Thin Lift Deposition in the Oil Sands of Alberta, 10th International Seminar onPaste and Thickened Tailings, March 2007.

353


Filtration tests on PVD filter jackets in fine oil sands tailings

Y. Yao, A.F. van Tol, B. Everts & A. MulderSection Geo-engineering, Faculty of Civil Engineering and Applied Earth Sciences,Delft University of Technology, Delft, The Netherlands

ABSTRACT: The application of Prefabricated Vertical Drains (PVD) is considered to be one ofthe engineering options to accelerate the consolidation of oil sands tailings. However, when installedin a fine tailings pond, the geotextile filter jacket has a risk of clogging or blinding which can lead tothe significant reduction in the discharge capacity of the PVD. Based on the theory, the clogging orblinding of filter jackets can be caused by the fine soil particles and/or residual organic matters. Theobjective of the present research is to evaluate the filtration and clogging behavior of filter jacketswhen applied in oil sands thickened tailings (TT) from Muskeg River Mine, Alberta, Canada.Filtration column tests on three types of geotextiles jackets were conducted in the laboratory tostudy the possible clogging of the filter jackets of PVD. After starting the tests, the discharge flowrate through the jackets decreased gradually over a period of the tests for 2–3 weeks. The filterjackets kept on functioning well during the test. It appeared that the permeability of the consolidatedthickened tailings controlled the thickened tailings/geotextile system.

1 INTRODUCTION

Prefabricated Vertical Drains (PVD), also known as wick drains or band drains, are widely usedto accelerate consolidation of soft clays nowadays. When installed in soft soils, PVD can create ashort path for the trapped pore water to escape and thus speed up the rate of consolidation. PVDare comprised of a drainage core wrapped in a geotextile filter jacket which has two basic filtrationfunctions: first to retain soil particles; and second, to allow water to pass from the soil into thePVD core. As noted by McGown (1976), Bell and Hicks (1980), D.T. Bergado (1996) et al., aneffective geotextile filter jacket in the soil may function as follows, see Figure 1. A small amountof fine particles moves into or through the filter jacket leaving the coarser particles to bridge andarch outside of it. The zone of fine particles immediately behind the soil bridge network is calleda “filter cake”. Once the soil filter is established, no further particle movement will occur and thesoil-geotextile jacket system is in equilibrium; hence, the geotextile filter jacket retains the soilparticles and prevents its migration into the PVD core.

According to literature, the establishment of a stable and effective soil filter by the geotextilefilter jacket depends on the following:

1. The physical and mechanical properties of the geotextile filter jacket e.g. pore size, porosity andcompressibility of geotextiles etc.

2. The characteristics of the soil, e.g. particle size distribution, porosity, permeability, andcohesiveness.

3. External stresses and strains imposed on the soil-PVD system.4. The prevailing hydraulic conditions, e.g. laminar or turbulent flow, unidirectional or reversible

flow, and dynamic or pulsating flow.

For well–graded soil, the selection of geotextile filter jacket which can function properly isessential for the successful application of PVD. For fine tailings, the potential migration of fineparticles, as they are not bonded, during the drainage and consolidation presents a higher risk ofclogging of the filter jackets compared with soils. Thus the application of PVD in fine oil sandtailings needs to be studied.

355

Figure 1. How a filter jacket works (McGown 1976; Bell and Hic.

Oil sands tailings, the waste production of oil sands from the extraction plant, are a mixtureof sand, silt, clay, water and approximately 1% to 2% un-recovered bitumen. Oil sands tailingsare pumped into the tailing ponds where sand is used to build containment structures and beachesand the clay-sized mineral particles flow with water into a settling basin for settling and waterclarification. Thickening of fine tailings in high rate thickeners has been implemented at the ShellAlbian Sands operations and other technologies such as paste thickened tailings, inline flocculationof fluid fine tailings, and centrifugation of flocculated fluid fine tailings are in advanced stages ofdevelopment. There is a strong interest and incentive to accelerate the rate of consolidation of fineoil sands tailings deposits and the application of PVD is one of the options. Liu & McKenna (1998)conducted a pre-feasibility study including laboratory testing and numerical modeling to obtainparameters for the potential application of PVD in Composite Tailings (CT, oil sands tailingswith a relative high sand content and with added inorganic coagulants (e.g. calcium sulphate,aluminum sulphate in order to avoid segregation). Wells & Caldwell (2009) presented field tests ofthe application of PVD in Fort McMurray in a high fines content tailings deposit produced throughat Suncor operations.

At the Shell Albian Sands oil sands operations tailings are classified by hydro-cyclones in whichthe coarse fraction is separated from the tailings. The cyclone overflow, which contains the finesand water, is further processed in high rate tailings thickeners to recover warm water that canbe immediately reused in bitumen processing while producing Thickened Tailings (TT) slurry.Compared with CT, TT has much higher fines content (up to 70% of particles finer than 44 µm)and a high water content which may cause the clogging of the geotextile filter jackets. Bell andHicks (1980) present three different clogging mechanisms for geotextile filter jackets, distinguishedas follows:

1. Clogging of the core, by passage through the jackets of too much fine particles;2. Clogging of the filter jacket by particles trapped within the fabric structure;3. Blinding of the jackets by formation of a filter cake due to filtration of the sludge.

In addition to above mechanisms, the proper functioning of the PVD in fine oil sands tailingscan also be hammered by

356

Table 1. Properties of thickened tailings sample.

Index Value

Water content∗ ASTM D2216 65.0%Bulk density of sludge 1.20 g/ccDensity of solid particles Ultra pycnometer 2.29 g/ccVoid ratio 4.3Liquid limit ASTM D4318 52%Plastic limit ASTM D4318 15%Plasticity index ASTM D4318 37%SFR(44 µm)∗∗ British Standard 0.54Fines content British Standard 65%D30 6.3 µmD50 11 µm

* Water content is defined as the ratio of water mass and total mass, W = w/T.** SFR = Sand Fine Ratio, The ratio of sand content (>44 µm) and fines content (<44 µm).

Figure 2. Comparison of TT used and classified tailings products plotted inTernary Plot (Scott & Cymerman,1984).

4. Decreased permeability of consolidated sludge around the drain.5. Clogging of the filter jacket by the bitumen in the oil sands tailings.

In order to study the performance of geotextile filter jacket of PVD in oil sands tailings, especiallytheTT, laboratory filtration tests were performed to observe the clogging behaviour of filter jackets.Three types of geotextiles were imposed under different pressures.

2 MATERIALS AND METHODS

2.1 Slurry used

The soil used in the test is TT taken from Albian Sands Muskeg River Mine at Fort McMurray,Canada. Table 1 provides a summary of engineering properties of material used. The water contentis defined as the ratio of mass of water and the total mass. Fines are defines as particles smaller than44 µm. Figure 2 presents the comparison between TT used in the test and other oil sands tailings

357

Table 2. Properties of geotextiles used in the experiments.

Items TYPAR D165 TYPAR 5417HS LDFD

Material Polypropylene polypropylene polyesterUnit weight (g/m2 ) 165 140 95Effective area (cm2) 60 60 60Thickness 20 kPa (mm) 0.39 0.40 0.30Elongation at break (%) MD 50 50 39Tensile strength (kN/m) 12.9 11.7 6Opening size O95 (µm) <75 75 75Opening size O95wet (µm) – 75 <65Permeability (mm/s) 11 12 39Permittivity (1/s) 0.3 0.3 –

plotted in the Ternary Plot (Scott & Cymerman, 1984). The TT sample used in the test fits wellwithin the area of classical classification of TT (Shahid Azam & J. Don Scott, 2005).

2.2 Geotextile used

The filtration tests were conducted with three non-woven geotextiles provided by one of the PVDmanufacturers. The filter jacket samples represent the widely used geotextiles that are applied asstandard filter jacket of PVD in the consolidation of soft soils. Table 2 presents the propertiesof geotextile filter jackets used in the tests. These filter jackets are readily distinguished by thedifferences in their unit weight and permeability, i.e. white colored polyester geotextile (LDFD) islight weight and more permeable while the grey colored ones (D165&5417HS) are heavy weightand less permeable. All the three geotextiles were used without any modifications.

2.3 Apparatus used

The apparatus used for testing was indigenously fabricated in the Soil Laboratory of the Departmentof Geotechnology, Delft University of Technology, the Netherlands. See Figure 3, three Perspexglass cylindrical moulds, 100 mm diameter each, were mounted on top of each other and tightlyscrewed, forming the filtration column. The three major parts of the column are pressure chamber,filtration chamber and collection chamber with the height of 300 mm, 200 mm and 100 mm sepa-rately. The system was developed to permit automated measurement of variables monitored duringall phases of a test. The main features of the system consist of the following: (1) Constant pres-sure and pressure monitoring system,(2) Filtration system, and (3) Flow measurement system. Theconstant pressure system using compressed air is maintained at the required value on adjustmentsof the pressure controller located on the air pressure panel. Air at the controlled pressure is thenadmitted to the bladder in the air/water bladder cell thus pressuring the water without direct contactbetween air and water. The pressurized water is then led to the pressure chamber inlets. The filtra-tion chamber is the part where the thickened tailings remains throughout the test. Tailings in thefiltration chamber is dewatered through a geotextile filter supported by a filter screen (wire mesh)placed between the filtration chamber and collection chamber. Pressure transducers are installedin the wall of filtration chamber to monitor the pore pressure in the sludge during the test. Themeasurement of outflow is achieved by weighing the accumulative discharged water, collected ina cup below the outlet. In order to prevent from evaporation, the cup was sealed with plastic paperwith a hole letting the filtrate drop in. The weight of discharged water and the pore pressures in thetailings are monitored and recorded by a Data Acquisition System.

2.4 Test plan

The magnitude of imposed pressure was to simulate the pressure/potential of the depth where PVDwere installed. The fine tailings are essentially a fluid (Wells & Caldwell, 2009). Therefore the

358

Figure 3. Schematic diagram of filtration column.

driving force which promotes water flow from the tailings into the vertical drain is equal to thetotal head (hydraulic potential) within the tailings column, minus the total head within the core ofan individual drain. When the PVDs are installed, for tailings with bulk density (ρt), the expectedfluid pressure hence potential (P) at a depth (ht) in the tailings is Ptailings = γtailinght = ρtght. Insidethe drain the potential Pcore = γwaterhw = ρwghw. In case that the density of water is less than the unitweight of the tailings (in this case 12 kN/m3), there is a net pressure differential between the outsideand the inside of the vertical drain and hence flow of tailings water from the tailings into the wickdrain. Moreover, this process continues until the soft tailings approach their liquid limits and beginto generate effective stresses, acting as a soil as opposed to a slurry. Once the fluid-like tailingsbegin to experience grain-to-grain contact the hydraulic conductivity of the soil-like tailings beginsto exert an influence on the flow of tailings water into the wick drain. This is the phenomenonreferred to as consolidation. The above mentioned pressure conditions are applicable as long asthere is no grain-to-grain contact between the soil particles.

The structure of the apparatus used in this test, is such that once the valve is opened, the water inthe collection chamber is directly connected to open air. Thus the pressure imposed on the bottomside of the geotextile filter jacket is constantly zero. The pressure on the filtration chamber side ofthe filter equals to the pressure applied on top of the sample plus the self-weight of sludge. In the testthe pressures imposed on the sludge samples with geotextile of LDFD and 5417HS were constant20 kPa and 10 kPa respectively. For geotextile of D165, however, the pressure was set initially to10 kPa and doubled when the flow rate became stable. The pressures imposed on the filter jacketswere approximately representing the pressure difference of PVDs at the depth between 5 m and10 m in the pond. Tests were stopped when the flow was negligible or some abnormal phenomenonoccurred. The duration of the tests varied from 20 to 40 days.

2.5 Test procedure

Performing the filtration tests the following sequential procedures are applied:

1. Preparation of the sampleTailings samples were prepared by mixing for half an hour until a homogeneous fluid wasobtained. Geotextile filter jackets were saturated with the help of vacuum pump to expel the airbubbles before the use.

359

2. Installation of apparatusFirst the valve of the collection chamber is closed and filled with de-aired water. The filterscreen, the geotextile filter jacket and O-ring are placed on the top surface of the chamber.Next, the filtration chamber, the middle mould that contains tailing sample is connected to thecollection chamber with rubber O-ring preventing from sidewall seepage. When the mould isfully filled with sludge, place a plastic sheet on the top surface of it and install the next chamber.The plastic sheet is used to temporarily separate the sludge from the water added later. After thetop chamber is full of water, the sheet is cut without disturbing the sample. Last, the filtrationcolumn is sealed and the water inlet is connected to the pressure system.

3. Application of pressureAfter setting up the filtration system, the pressure controller is switched on the air pressure paneland the water pressure is adjusted to the required value. Because of the sensitivity of compressedair supplier, the actual pressure is always gently fluctuating, which needs to be monitored andadjusted if necessary.

4. MeasurementDuring the tests, the height of sludge, the pore pressures at two horizons and the amount ofdischarged water are monitored. After completion of the tests, the water content of the sludgesamples and the increased weight of geotextiles are determined.


3.1 Observations

Figures 4(a)–(c) illustrate the flow behavior of each filter sample and Figures 5 (a)–(c) showthe variation of pore pressures at different depth in the tailings sample with time. In Fig.4c, thetransducer Port 1 was replaced by Port 2.

According to the data, similar trends are observed for all tests. In the first several seconds therewas a continuous flow coming out from the outlet, followed by a sharp decrease of flow finallyresulting in a decreasing. After 10 to 15 days the discharged flow rate became almost constantin each test. The filtrated water contained some fines at the beginning but remained completelyclean after a while, which demonstrated that the jacket functioned correctly, not allowing too muchof fines to get into the core of the drains. The two pore pressure transducers measured the porepressure at the different levels in the sample. The initially recorded small difference is due to thedifference in hydrostatic pressure. In a later stage the effect of bottom drainage and consolida-tion was reflected in the decreasing pressure at the lowest transducer. The average duration ofthe tests is no less than 3 weeks (Test 1 was only 2 weeks). At the end of each test, the flowrate increased as well as the pore pressure of lowest port. For Test 1 the rate of discharged floweven increased by a factor of 10. The explanation of this phenomenon can be found by the obser-vation of the cracks in the sample along the wall of cylinder (Figure 6). It is assumed that thecracks increased the permeability of soil dramatically as a result of the formation of a shortcutfor drainage. On the other hand, at the outer surface some gas bubbles were seen. These gas bub-bles might have disturbed the pressure measuring by the transducer because of poor contact withthe sample.

Comparing (a) and (b) in Fig. 4, we can see that after 10 days both of the flow rate figures wereidentical with a value around 18 ml/day in spite of the significant difference in the permeability ofgeotextiles as well as the different pressure applied. However, for long-term performance of filterjacket illustrated in (a) and (c), the higher imposed pressure (20 kPa, relative to 10 kPa) resultedin a slightly higher flow rate. Moreover, we notice that the flow rate in (c) increased dramaticallywhen the pressure was doubled. Finally, according to Fig. 5, the pore pressures are still decreasingat the end of the tests.

3.2 Analysis of soil sample

Figure 7 provides the water content distribution of the soil samples with different geotextile filterjackets at the end of the tests. Due to sedimentation and consolidation the maximum solid content of

360

Figure 4. The flow behavior of each filter sample: (a) LDFD (b) D165 (c) 5417HS.

soils was obtained at the bottom part near the filter jacket and the height of soil samples decreasedfrom 20 cm to 9.5 cm. The solid content on the surface (top of sample) was the lowest, almost equalto the initial value.

Figure 8, taken from Sobkowicz & Morgenstern, (2009) presents the void ratio and permeabil-ity of fine tailings plotted in the Ternary Plot of Oil Sand Tailings (Scott & Cymerman, 1984)

361

Figure 5. The pore pressures of the sludge with each filter sample: (a) LDFD (b) D165 (c) 5417HS.

respectively. In this test, at the initial water content of 65% the void ratio of soil is about 5 and thepermeability is 10−4 cm/s which equals to 10−6 m/s. However, at the final water content of 30%(void ratio about 1) the permeability significantly decreases to 10−7 cm/s = 10−9 m/s which differsa factor of 1000 (Fig. 8b). It can be summarized that due to the consolidation a layer of consolidatedmaterial with extremely low permeability is formed and controls the flow rate.

362

Figure 6. The shrinkage and cracks occurred on the surface of soil sample.

Figure 7. Distributions of water content along the height of soil sample after the test (distance measured fromoriginal top of the sample; 20 cm is bottom of sample).

Figure 8. Void ratio (a) and hydraulic conductivity (b) of fine tailings plotted in the Ternary Plot of Oil SandTailings (Scott & Cymerman, 1984).

363

Figure 9. Geotextile filter jacket before (left) and after the test (right): (a) LDFD (b) 5417HS.

Table 3. Weight comparisons of filters before and after the test.

Geotextile filters LDFD 5417HS D165

Dry weight before test (g) 1.49 1.92 2.33Dry weight after the test (g) 1.77 2.28 2.76Increased by (%) 18.8 18.8 18.5

3.3 Analysis of geotextile filters

Figure 9 compares the clogging phenomenon of filters before and after the use. It is visible thatsome bitumen was accumulated at the edge of the filter named LDFD, which is due to the rubberO-ring used in the first test. This ring was not thick enough to prevent seepage between filter jacketand sidewall. For the following tests a double O-ring was used and the seepage was successfullyavoided (Fig. 9b). It seems that the filters were not seriously clogged; only the color in the effectivearea was darkened by soil particles.

Table 3 compares the dry weight of all the filter jackets before and after the test. The increasingweight of filters during the test demonstrated a certain amount of soil particles were trappedamong the geotextile fibers. Additionally, it is interesting to find that for all the filters, the relativeincrements are almost identical, which may prove that in this test the difference of the threegeotextiles does not affect the clogging behavior.

The authors planned to use the same test apparatus to measure the hydraulic conductivity of thefilter after the completion of the test using constant head method, however, it was found impossibleto determine the flow rate with sufficient accuracy because the filter was still too permeable.Therefore the permeability of the jacket does not play a role in the consolidation process.

The performances of filter jackets with respect to the possible clogging mechanisms in the testas mentioned before are evaluated as follows:

1. The geotextile filter jackets were able to retain the tailings particles. Only a few fine particlesmigrated through the filter jackets during the beginning of the test. Clogging of the cores ofPVDs is very unlikely to happen in reality.

2. The increased weight of filter jacket after the test demonstrates that some particles were trappedamong the fibers of the geotextile jackets, however, this phenomenon was not significant toseriously reduce the permeability.

3. Blinding of the filter jackets by formation of a clear filter cake that stuck to the surface of thefilter jackets was not found when the samples were dismantled from test apparatus.

4. The solid content around the filter jackets increased during the drainage and consolidation and thepermeability of the soil decreased along with the void ratio. After the completion of the teststhe permeability was reduced by a factor of about 1000. Whether the compressibility of thisconsolidated soil decreased with a similar or a minor order has not yet been determined. Thisphenomenon will determine the consolidation coefficient during the test and finally determinethe duration of the consolidation process and the performance of the PVD system. To concludeon this important point a proper analysis of the consolidation process has to be made.

5. The residual bitumen did not accumulate on the filter jackets after the test. This was concludedafter visual inspection of jackets.

364

Finally, the fact that the used filter jackets were still very permeable after the tests confirms thatthey are not clogged by oil sands TT in the filtration tests.

4 CONCLUSIONS

The behavior of oil sands thickened tailings with different geotextiles have been studied. Thefollowing major conclusions may be derived.

1. Permeability of the sludge/geotextile system decreases with increasing solid content of soilsample irrespective of the type of geotextile used and pressure applied on the system.

2. The bitumen as well as the fines in the sludge do not cause serious blinding or clogging of filterjackets. During the test period as long as 40 days of application, the filter jackets functionedproperly.

3. The layer of consolidated fine tailings adjacent to the jackets decreases the permeability of thesystem dramatically, but does not effect functioning of the filter jacket. In other words, it is theconsolidation behavior controlling the permeability characteristics of TT; the filter jackets donot control the long-term filtration performance of the system.

4. The soil sample under higher pressure gains higher solid content compared with others with thesame filter jacket. There is no direct relation found in the test between the pressure differenceand clogging behavior of filters.

5. In this study it is found that gas production may occur, but the produced gas will escape soonand not hamper the consolidation process in a considerable way.

This study which evaluates the performance of geotextiles with oil sands fine tailings has practicalsignificance in the design of PVD for tailing pond applications.

REFERENCES

ASTM (2000)annual book of standards. Philadelphia: American Society for Testing and Materials.Azam, S. & Scott, J. D. 2005. Revisiting the ternary diagram for tailings characterization and management.

Waste Geotechnics, December 2005:43–46.Aziz, A.A., Mohammed, T.A. Omar, H. 2008. Filtration performance of a silt/geotextile system within a triaxial

permeameter. Proceedings of the 4th Asian Regional Conference on Geosynthetics, 2008:415–419.Liu, B.Y. & McKenna, G. 1998. Application of Wickdrains in consolidated tailings (personal communication).McGown, A. 1976. The properties and uses of permeable fabric membranes. Proceedings of the Workshop on

Materials and Methods for Low Cost Road, Rail and Reclamation Works, Lee, Ingles and Yeaman, Eds.published by the University of New South Wales, Leura, Australia, September, 1976: 663–710.

Pollock, G.W., Mc Roberts, E.C., Livingstone, G., Mc Kenna, G.T. Matthews, J.G. 2000. Consolidationbehaviour and modelling of oil sands composite tailings in the Syncrude CT prototype, Tailing and MineWaste ’00. Rotterdam: Balkema.

Rao, G.V., Gupta K.K. Pradhan, M.P.S. 1991. Long-term filtration behavior of soil-geotextile system.Geotechnical Testing Journal, Vol.15, No.3: 238–247.

Sobkowicz, J.C. & Morgenstern, N.R. 2009. A Geotechnical perspective on oil sands tailings. Tailings & MineWaste ’09. Edmonton: University of Alberta.

Scott, J.D. & Cymerman, G.J. 1984. Prediction of viable tailings disposal methods. ASCE Proceedings of asymposiuim on sedimentation and consolidation models: 522–544. San Francisco, California, USA.

Wells, P. S. & Robertson, J. C. Vertical “wick” drains and accelerated dewatering of fine tailings in oil sands.Tailings & Mine Waste ’09. Edmonton: University of Alberta.

365


Suncor oil sands tailings pond capping project

Gord Pollock & Xiteng LiuAMEC Earth & Environmental, Edmonton, Alberta, Canada

Ed McRobertsAMEC Earth & Environmental, Vancouver, British Columbia, Canada

Keith WilliamsAMEC Earth & Environmental, Denver, Colorado, USA

Patrick Sean Wells & Joseph FournierSuncor Energy Inc., Fort McMurray, AB, Canada

ABSTRACT: After successful field trials, Suncor began construction in January 2010 of a floatingcover on Pond 5. The surface area of the soft tailings in Pond 5 is approximately 200 Ha. The coveris intended to provide a trafficable surface which will facilitate further reclamation activities,ultimately leading to the closure of Pond 5. The cover is being constructed with geosynthetics andpetroleum coke. Coke is a byproduct of the oil sands process, and has the benefit of being lighterthan typical earthen materials thereby allowing a cover to be placed over low strength tailings. Thispaper will provide an overview of the design, initial construction progress, and lesson learned todate.

1 INTRODUCTION

As part of the extraction process at the Suncor Oil Sands Mine, various tailings managementfacilities have developed over time. Pond 5 is being reclaimed and the placement of a coke capis the first step towards obtaining a trafficable surface to allow further reclamation activity to beconducted. This paper discusses the design of the coke cap, the construction during the 2010 winter,and assessment of the performance of the construction to date.

2 POND 5 OVERVIEW

Figure 1 shows an aerial photo taken in the summer before construction of the cap commenced.Pond 5 is contained by dykes and the Waste Dumps overlying perimeter pit walls to the southwest.Pond 5 contents historically consisted of a water cap overlying a layer of gypsum enriched maturefine tails (MFT), underlain by CT. In preparation for pond reclamation by the engineered coke cover,some of the MFT in Pond 5 was replaced with tailings sand. The removed MFT was transferred toPond 6. Prior to starting the construction of the coke cap, the water cap in Pond 5 had largely beeneliminated. The key parameters for Pond 5 tailings coke capping project are solids content andstrength. AMEC reviewed information provided by Suncor that were collected in 2008 and 2009 tomake the following conclusions. Pond 5 as exposed has a surface layer of MFT at solids contents oftypically 40% to 60% grading to a Soft-CT of 70% solids content or more at depths of 20 to 40 feet.Typical sections through the contents of Pond 5 developed from the 2008 data collected by Suncorare shown on Figure 2. A summary of all direct measurements of MFT solids contents expressedas bulk density for field sample sites from the 2008 field campaign is provided in Figure 3. It canbe seen that at any given location the densities with depth are variable, and relative to the coke capdensities (∼1.3 t/m3) the variations can be significant. The figure also indicates three trend lines

367

Figure 1. Pond 5 aerial photo (June 1, 2009).

Figure 2. Pond contents from Suncor block model.

368

Figure 3. Tailings density with depth.

(lower, median and upper bound) of density with depth which have been used for several designpurposes.

The CPTu and vane shear data from 2008 and 2009 indicate that the MFT and soft CT mobilizevery little to no effective stress. Summary plots of the CPTu and vane data with depth are shown onFigure 4 for 2008 and 2009 data. (Also included on Figure 4 is the relationship used in the FLACanalysis discussed below).

3 FIELD TRIALS

Suncor conducted two field trials prior to the design and construction of the full scale coke cap. ThePhase 1 trial was conducted on a beach at the South-East corner of Pond 5 as shown on Figure 1.The Phase 2 trial at the west end of Pond 5 was conducted on MFT which was more representativeof the foundations conditions which would be encountered for the full scale operation. A detaileddiscussion of these trials is presented in a companion paper at this conference (Caldwell, Wells &Fournier, 2010).

The cover for the trial areas consisted of a light weight geosynthetic to act as a barrier betweenthe MFT and overlying road fill. A layer of geogrid was placed on the geosynthetic. The road fillon top of the geosynthetics consisted of 3.5 m to 4 m of coke. Coke is a byproduct of the oil sandsprocess, and has the benefit of being lighter than typical earthen materials with high strength (inthe order 40◦ to 45◦) making it an ideal material for a tailings cover.

Primary findings from the trials influencing the full scale design included:

• It is possible to construct up to 3.5 m to 4 m of coke on geosynthetic over MFT and back a cokeloaded 40 ton articulated haul truck on to it.

• Coke sinks to a level not much higher than the surrounding MFT for both 3.5 m and 1 m lifts.• Field bulk densities of coke measured varied from 0.99 t/m3 to 1.40 t/m3 with an average of

1.23 t/m3.• Woven geotextile seems to act as a semi-impermeable barrier, retaining MFT fines, which then

develop a consolidated skin impeding water flow.• Where coke was not compacted by heavy truck traffic, it became untrafficable under light vehicle

traffic when saturated.

369

Figure 4. CPT and vane data with depth for 2008 and 2009.

The differences between the field trials and the full scale operation which need to be accountedfor in the design include:

• Scale. The Phase 2 field trial (which occurred on MFT) represents an area of approximately17,000 m2, compared to the more than 2,000,000 m2 which will eventually need to be covered.

• Up to 4 m of coke was placed and the commercial design was to be based on 3 m.• During the trial, a single haul truck was used, was operated at a very low speed, and was backed

out on to the trial area to eliminate the need for turning. During the construction of the full scalecap, it was necessary to design for multiple trucks which would travel at a reasonable speed(15 km/hr to 25 km/hr) and would be turning around for each dump load.

• Rate of construction for the full scale cap would be significantly faster than for the field trial.• Given the large area to be covered, there will be varying foundation conditions potentially giving

rise to differential settlements and induced stresses. It was unclear whether the trial foundationtailings represented a lower bound of MFT densities that will be encountered over the pond.

4 DESIGN BASIS AND APPROACH

The design of the full scale cover also utilized coke because of its beneficial properties for cappingsuch as its light weight with a high friction angle. The specific gravity of coke particles variesaccording to the amount of vugs or entrained air does vary according to coke particles size. Testing

370

reports specific gravities of 1.41 for raw coke and 1.46 and 1.47 for particles crushed 1 to 4 times.As coke particles typically have a specific gravity greater than that of water they will therefore sinkin water. The dry unit weight of coke at Optimum SPD is in the region of 1.0 t/m3 or a unit weightof 9.8 kN/m3, and for a specific gravity of 1.45 the void ratio (excluding vugs) would be 0.45. Atthis void ratio water saturated coke has a density of 1.310 t/m3 or a unit weight of 12.9 kN/m3. Ifcoke becomes saturated with MFT which has a density greater than water, then the density of thecoke/MFT system can be greater than MFT. Therefore it can be understood that on a self weightbasis water saturated coke will sink to some level in MFT. For example, coke voids saturated withMFT at a solids content of 40% and density of ρ = 1.34 t/m3, the coke/MFT mixture would have atotal bulk density of ρ = 1.42 t/m3 or 13.93 kN/m3. Thus the coke/MFT mixture would sink into theMFT if fully MFT saturated, but would float on MFT at 40% solids content if saturated just withwater. Based on available testing coke has been assigned strength parameters of c′ = 0, �′ = 40degrees.

The design approach was predicated on several simple premises. Firstly that coke has a lowerbulk density than most of the soft tailings in Pond 5 and if “encapsulated in a bag” will sink to a levelconsistent with the coke bulk density and superimposed vehicle weight. This invokes Archimedes’principle. The local stability of the coke at the edge of the road, under vehicle loadings, and mudwave drag is obtained by containing the coke in a continuous geosynthetic layer. Constructionis facilitated by placing the geosynthetic on temporarily frozen tailings, consistent with practiceelsewhere. The identified instability mechanisms that the design must consider are lack of buoyancy,particularly under the types of dynamic loading imparted by construction equipment, which cancause the road to sink into the underlying MFT, MFT intrusion into the coke, and hydraulic fracturingof the coke causing MFT upflow into and over the coke. In order to maintain buoyancy requiredfor cap integrity, MFT intrusion or upflow into the coke must be prevented by the inclusion ofa suitable geotextile to separate the coke from the underlying MFT. MFT can also be deformedlaterally from a loaded area creating mud wave drag on the base of the geosynthetic layer. Thiscould cause separation of overlapped seams or tearing of poorly stitched seams.

Instability of the coke cap could be induced by some combination of excessive settlement ofcap into tailings, mudwave drag, differential movement of the soil and geotextile, local punchingshear from haul equipment, impacts from frozen tailings that subsequently thaw, failure of thesewn geotextile seams, pond surface slope causing long shallow skin flow type movements, anddegradation of the geotextile over time by the underlying tailings (likely not an issue, given thecomposition of the tailings and the type of geotextile specified in the design). Preventing thegeotextile from tearing generating a long rip is of particular importance as a tear could result in avery brittle failure, with little time to react with obvious safety concerns.

The design criteria for the Pond 5 coke cap roads include both objective and subjective criteria,as follows:

Overall Performance Objective: Maintain structural integrity of the road, and keep it above thesurrounding MFT, while continuing to provide a trafficable surface and safe access across Pond 5for the specified construction equipment (current design basis being a coke loaded 40 ton truck).

Factor of Safety (FoS): Although only applicable to certain loading conditions and not the overallsystem, a factor of safety of 1.5 was applied to the loading that the geotextile must resist, includingat seams. The loading by haul trucks must consider dynamic effects, and an empirical load factorof 2.0 was used in the design to account for this and all other physical interactions between thetruck tires and the coke road that cannot be modeled directly in the 2-D FLAC model.

Drainage of Coke: While a drainpipe system was planned for the coke roads, the design assumedthat all coke except for the top 0.5 m was saturated.

Settlement of Coke: Settlement of coke along the road should not be more than the elevation ofthe mudwave outside of the coke road edge. This is to prevent back-flow of MFT over the road.

Construction Direction: In order to guard against overall instability of the road system, con-struction should proceed “uphill” against the regional dip of the pond bottom/exposed MFTsurface.

Geotextile: In addition to the tensile capacity, the geotextile and seams must prevent intrusionof MFT into the coke. While some coke gradations might be expected to prevent intrusion, speci-fication of a suitable particle size gradation was not practical. Visual inspection of the coke is to beundertaken on a continual basis to ensure that the coke is well graded to provide a backup againstthe MFT penetrating into the coke in the event of a geosynthetic flaw.

371

Observational Method: The fluid nature of the fine tailings creates a design section of coke,geosynthetic and MFT which all have significantly different strength and stiffness characteristicsfrom one another. The very nature of this cover, combined with the variability of the MFT deposit’sstrength and density across Pond 5, necessitates a design approach that includes analyses, obser-vation and periodic adjustment. Therefore, the Observational Method was considered to be a keycomponent of the design which, at minimum, requires performance monitoring during constructionas part of the construction QA/QC process, and the ability to invoke changes to the design section,if required, on a contingency basis.

Key variables affecting the coke thickness required for the road design section include:

• MFT profile (density, strength, frozen/thawed)• Final loading scenario, including:

o Size, operating speed and braking of trucks and other required equipmento Requirements for trucks to end-dump, turn and/or park temporarily

• Geosynthetics at the base of the road only or the ability to include additional geogrid higherup within the coke section (which may allow the use of larger trucks and/or higher operatingspeeds)

• Geosynthetic tensile strength and stiffness• Interface strengths between the various components of the design section (i.e. coke/geogrid,

geogrid/geotextile, geotextile/MFT)• Coke width and saturation levels

Depending on observed field performance in relation to predicted behaviour through modeling,the primary adjustment that could be made in the field is the coke thickness. The geosynthetic atthe base of the road will already be in place, and must be selected in advance to be sufficientlystrong (including across seams) for the expected loads in order that it will not tear under loading.However, additional geosythetics could be considered higher up in the coke section, if required tomeet the performance objectives.

The basic method of analysis used in design was stress analysis using FLAC. It was recognizedfrom the outset that the MFT (or sludge) would not behave as a typical soft soil as MFT at the lowdensities encountered mobilizes very low strength and essentially zero effective stress, and largedeformations were expected. FLAC could also combine various instability mechanisms such asmud wave drag, punching shear failure from haul equipment, and tensions induced by differentialmovement, as these effects could be additive. The FLAC model allows one to evaluate buoyancy,estimate the maximum coke settlement and predict the tension induced in the geotextile. TheFLAC software was selected because it has the capability of handling both large strain and materialinterfaces. Several of the mechanisms listed earlier that could cause tension to be induced in thegeotextile are embraced in the FLAC procedures, but a good understanding of the mechanismsis required in order to interpret FLAC results. Buoyancy of the coke road and an estimate ofthe maximum coke settlement that will occur can also be evaluated through the 1-D applicationof Archimedes’ principle. This allows one to evaluate the amount of road that will be above thetailings (or conversely how much will be submerged) as a function of tailings density and assumedcoke density. The amount of coke above the surrounding tailings can vary significantly dependingon the density of the tailings. The Archimedes approach does not consider strength of the tailings,but is considered to work well for coke roads over MFT because the truck loads will overcome theshear strength of the MFT, given the thixotropic nature of MFT.

Significant simplifications that were made in analysis included using theoretical elastic solutionsto convert 3-D truck loads into 2-D line loads, modeling geofabrics as beam elements with zeroflexural stiffness and using a Mohr-Coulomb constitutive model with residual strength to modelMFT. Effort was taken in allowing for slippage (relative movement) at the coke/geotextile andgeotextile/MFT interfaces to occur. Theoretically, all coke could be placed on unfrozen MFT withthe use of geotextile as a separator. The construction of the coke cap was modeled in FLAC inaccordance with the expected construction sequence. However, to improve the numerical stabilityof the FLAC model, the model was constructed by placing the initial 1 m thick lift of coke on frozenMFT and then placing the consecutive coke lifts with all MFT assumed to be thawed.

The selection and placement of the geosynthetic materials was based on FLAC analysis butconstrained by overall schedule, operational requirements and judgement The products selected

372

Figure 5. Plan view of road and cell layout.

were a High Strength Woven Geotextile [HSWG] overlain by a geogrid. The ultimate tensilestrength of the HSWG and geogrid were 82.5 and 27.5 kN/m respectively. The HSWG providedboth a strength function, the ability to be stitch-seamed, and a reliable barrier against MFT intrusioninto the coke. By itself the HSWG with perfect seams was modeled as being acceptable for the 40tonne truck design load. However as the work was to be executed in the coldest winter months overice a geogrid was also placed over the HSWG as an additional line of defense.

Once a road is built, a primary direction of instability is across the road from vehicle wheelloading. However the geosynthetic must “float” the road and traffic in both directions spreadingout truck weight in a snow-shoe effect. In addition underlying mud wave drag along the roadtowards the working face can also put considerable tension into the geosynthetic in that direction.A clear advantage of a HSWG is the ability to stitch-seam the rolls. “Seaming” geogrid by insertionof a bar was not considered practical. Governance literature clearly establishes the necessity forstitched seams for covering low density sludges. It was considered that an unseamed geotextile willdevelop little friction on an overlap joint. If pressurized MFT intrudes into the overlap there willlikely be little to no strength mobilized. Therefore part of the design required that the HSWG bestitched.

5 ROAD AND CELL APPROACH

A road and cell approach has been adopted for the cover as opposed to a full frontal advancemethod. The intent behind the design is to construct the roads initially, preferably in the wintertime, followed up by construction of the cap in the cell areas. The roads still need to be able tosupport loaded truck traffic, however the cells need only to support the wick drain equipment. Theroads would be placed in a rectangular configuration as shown on Figure 5. The spacing betweenroads is governed by maximum push distance for snow cats and is 700 ft centerline to centerline.Cross roads will be placed at about 1350 ft as currently indicated, but could be revised once cellcover issues are resolved. The total rib width is 100 m (330 ft) with a central 30 m (100 ft) mainroad and two – 30 m (100 ft) wide support roads.

The road and cell approach has been adopted for the following reasons:

1. Geotextile Stitching In order to make the maximum placement progress, winter field geotextilestitching can be significantly reduced by concentrating on road development.

373

Figure 6. Road design cross section (vertically exaggerated).

2. Squeezing to Cells By leaving cells to be infilled at a later date, there is improved potential oflateral “squeezing” of mobilized material into a cell, rather than displacing MFT towards a dykeor beach as with a full frontal advance.

3. Removing MFT By leaving uncovered MFT in cells, if locally very low solids MFT isencountered then it may be possible to access cells to pump out this material and move itelsewhere.

4. Geotextile placement over cells Geotextiles could be pulled across the cells from one road to theadjacent one. This may further reduce the amount of field stitching required. The best methodto pull geotextile across MFT has not yet been determined and may require some trial and error.

5. Reduced Coke Volumes The amount of coke needed in a cell is only the minimum required tosupport the equivalence of the wick drain installer. The amount of coke needed has not beenfinalized pended obtaining all details of the potential equipment to be used in the wick draininstallation.

The best method to cover the cells and connect the cell geotextile to the road geotextile has yetto be determined. The construction for the first season focused on the roads. The connection maybe stitched or over-lapped with a geotextile/coke/geotextile overlap.

6 DESIGN CROSS SECTION AND OTHER DETAILS

The cross section for the roads is shown on Figure 6.The main elements of the design are:

• High Strength Woven Geotextile directly on the tailingso 100 m wide. 5 m extra on each side for mud wave development.o 4.5 m × 100 m long panels seamed perpendicular to road alignmento 82.5 kN/m strength including seamso Seams are to be stitched

• Biaxial Geogrid on the geotextileo 75 m wide (single roll)o Overlapped 1 m with ties only

• 3 m of Coke on Geosyntheticso Constructed in 1 m liftso Bottom lift is 90 m wide

• Corrugated HDPE drain pipe along centerline of roadso 36 inch diameter sumps installed at each intersectiono Water removed with portable pumps as required

7 CONSTRUCTION

As discussed above, with the road and cell approach, the initial construction focused solely onthe roads. To provide a construction platform for geosynthetic installation, the construction wascommenced during freezing conditions.

374

Figure 7. Photo showing coke placement with 40 ton truck and D6 dozer.

Winter construction activities began by packing snow on the pond surface in December 2009.HSWG anchor construction began on January 16, 2010 and continued to March 11, 2010, whenthe road alignment NS-2 was completed. All construction activities were suspended in late March2010 due to the early warm spring weather. Visual inspections, installation and monitoring of theinstrumentation, dewatering and monitoring of the water levels, and surveying continued. In earlyMarch, dewatering of the pond surface water was initiated to coincide with the onset of the earlyfreshet to maintain a trafficable and safe working surface so that the construction activities couldcontinue.

There were several phases of construction that occurred simultaneously and in some cases sequen-tially where only one or two tasks could be performed at a single time. Items of work included thefollowing:

• Ice thickening and snow compaction• HSWG anchors,• Installation of HSWG,• Installation of geogrid,• Installation of drainage pipework and sumps,• Installation and monitoring of instrumentation,• Placement of coke material, and;

Figure 5 provides an aerial view of the Pond 5 coke roads that were completed during the 2010winter construction season. Figure 7 shows the dumping of the coke with a 40 ton articulated truckand the spreading of the coke on the geosynthetics with a D6 dozer. Good performance of the roads,likely due to ice, led to the haul trucks being permitted to travel on 2 m of coke.

The following quantities were placed during the construction season:

• Roads: 4,500 m (area of 450,000 m2)• Coke: 844, 077 m3

• Geotextile: 550,000 m2

• Geogrid: 510,000 m2

• Pipe: 4500 m

The following is summary of the observations and lessons learned during the 2010 winterconstruction season:

1. The presence of ice allowed for a good working surface to install geosynthetics (HSWG andgeogrid) and resulted in a better than anticipated support for placement of coke material on theroads. Based upon empirical evidence from the construction works, if the thickness of frozenlayer was over 15-inches (0.38 meters) thick, it provided an adequate working surface and if thelayer was less than 10-inches (0.25 meters) thick, cracking and breaking occurred.

375

2. In general, installation of HSWG went well (high quality materials, seam strengths exceededstrength requirements) and was not a limiting factor during construction of the roads. Hand heldsewing machines proved to be more versatile for field seaming of the HSWG.

3. The installation time for geogrid was quite small in comparison to the HSWG installation andcoke material placement. Based on field observations, it was important to make sure that thegeogrid was shingled correctly and that the advancing coke placement was placed in the shapeof a crescent moon to avoid wrinkling. By pushing out coke material out on the edges first, theweight of the coke material would minimize wrinkling in the middle of the geogrid. This alsohelps trap MFT under the coke.

4. The effort and time required to install the drainage collection system (collection pipe and sumps)was minimal compared to other activities and was not a limiting factor on the overall progressof the road advancement. Sump construction was delayed until after the placement of coke toavoid slowing the production of coke placement along the roadway.

5. Due the large volume of coke material required to cover the roads and the necessity to userelatively small haulage trucks, placement of coke material became the limiting factor for theconstruction schedule. To increase the advancement of coke, trial tests were performed thatallowed for increased haul speeds and decreased distance between the lifts as well as the use oflarger dozing equipment to place the coke material. It should be noted that these approvals werebased on having an adequate frozen layer below the access roads and were subject to change ifthe frozen tailings layer became thinner or if it was observed that the frozen tailings capable ofsafely bearing the loads. These changes significantly increased the rate in which the coke couldbe hauled and placed along the road alignment, thereby increasing the rate of road advancementas well as the volume of material that could be hauled.

8 MONITORING

Instrumentation was installed and read by MDH Engineered Solutions. AMEC personnel conductedregular visual monitoring on a continual basis. The following instrumentation and monitoring wasundertaken both during the winter construction season and continuing into the post-constructionsummer season.

• Visual observations (including ruts, excessive deformation cracks, potholes, MFT intrusion tothe surface, mud wave development)

• Surveying the surface of the coke• Surveying Settlement Plate Markers to measure the elevation of the base of the coke• Remote Cell Markers to measure the heave in the cell areas• Measuring water levels within standpipes and sumps• Measuring deflections in Shape Accelerometer Arrays• Measuring stresses in pressure plates• Measuring strains in optical strain gauges

A research station was installed in one of the roads which included pressure plates, optical straingauges, piezometers, and SAA’s were installed. Loading trials were conducted at the researchstation. A discussion of the results are beyond the scope of this paper and will be the subject of afuture paper.

9 PERFORMANCE

Given the purpose and function of the coke cap described above, the performance of the roadwas assessed with respect to two main categories: settlement and trafficability. Other categoriesin which the performance was also evaluated, which in certain ways are subsets of the above twocategories, include: integrity of the geosynthetics, MFT intrusion, mud wave activity, and waterlevels.

376

Figure 8. Profile along NS2.

Figure 9. Cross section at south end of NS2.

9.1 Settlement

One of the main performance criteria for the successful construction of the coke roads was whetherthe roads established during frozen pond conditions remained floating while subjected to loadingby coke and the haul trucks. The road was designed to maintain the surface of the coke abovethe surrounding MFT levels at all times; that is, submergence of any section of road would notmeet design intent. Surveying was done during and after construction for comparison with designestimates of deformation behavior.

Figure 8 shows the top of the coke profile along NS1 with the top of the tailings at the side of theroad. Figure 9 is a vertically exaggerated cross section at the south end of NS2 showing the surfaceof the coke plotted with the measurements from the settlement plates. Both of these figures aretypical of the plots generated from the surveying data and indicate that the surface of coke roadsremained above the tailings at all times.

Figure 10 shows a typical settlement plot with time at one section on NS2. Also shown on theplot are different events with time including coke placement and thaw.

Figure 11 shows a comparison of the range of measured settlements measured at each stationrelative to the FLAC and 1D Archimedes predictions. Two things are evident from the figure:

• There is a wide range of measured settlement over the project. It is possible that this is due tothe variability of the tailings and the variability in truck traffic.

• While the models in the range of the measurements, they slightly under predict the settlement.This is possibly due to deformations caused by repetitive loading which is not modeled byArchimedes and only partially modeled by FLAC.

377

Figure 10. Settlement with time on at station 09 on NS2.

Figure 11. Comparison of measured to predicted settlements.

Further work is ongoing to determine the influence of the variability of the tailings propertiesas well as the influence to repetitive loading.

9.2 Trafficability

The other main performance criterion was whether the road remained trafficable for the nature ofequipment being used both after construction and after thaw. That is, were there significant cracksand/or soft spots which caused travel along the road to be unsafe or impede traffic in any fashionduring non-winter (frozen subgrade) conditions? Visual monitoring was done on a continuous basisto evaluate compliance with this criterion.

378

During the winter construction season, the roads performed very well, likely aided by the iceand frozen tailings. Due to the good performance, construction was allowed to proceed with the40 ton haul trucks driving on the 2 m lift as opposed to the 3 m lift as per the design. Only one softarea was encountered, which was later on correlated to thin tailings underneath.

After the frost had left the tailings, a coke loaded 40 ton truck was used to proof roll the 2 mthick road. The vast majority of the road exhibited excellent behavior with two exceptions, onebeing the area just mentioned.

9.3 Geosynthetic integrity

Although direct measurement of the stress level in the geosynthetics is still a work in progress,general observations indicate that the geosynthetics performed their intended function. That is,strength to support coke and haul trucks, as well as, in the case of the geosynthetics, a barrier toprevent MFT from intruding into the coke.

10 CONCLUSIONS

Coke roads were successfully constructed across Pond 5 at the Suncor oil sands site as part of thefirst stages of establishing a cap for reclamation purposes. The roads performed as intended withrespect to settlement and trafficability under heavy truck loads.

The light weight, high strength combination of the coke makes it a very useful material forcapping the low density, weak tailings encountered at the Suncor site.

The geosynthetics performed their function with respect to tailings separation and road support.At this time it is not clear what the exact level of stresses were that were mobilized in the geosyn-thetics. However, work is ongoing (including load trials and further instrumentation) to obtainadditional data.

ACKNOWLEDGEMENTS

The authors would like to recognize the valuable contributions of the following individuals on thisproject:

Mr. Matt LeBlanc (Suncor Energy Inc., Fort McMurray, AB)Mr. Bill Tully (Suncor Energy Inc., Fort McMurray, AB)Dr. Michael Davies (AMEC, Vancouver, BC)Mr. Dave MacDougall (AMEC, Fort Mcmurray, AB)Mr. Ayman Abusaid (AMEC, Edmonton, AB)Mr. Matt Haley (AMEC, Denver, CO)

REFERENCE

J. Caldwell, P.S. Wells & J. Fournier, 2010. Suncor Pond 5 Coke Cap – The story of its conception, testing,and advance to full-scale construction, Proceedings from Tailings & Mine Waste ’10 Conference, CRCPress/Balkema, Vail, Colorado, Oct 2010.

379


Review of oil sands tailings technology options

C.B. PowterOil Sands Research and Information Network, School of Energy and the Environment,University of Alberta, Edmonton, Alberta, Canada

K.W. Biggar & M.J. SilvaBGC Engineering Inc., Edmonton, Alberta, Canada

G.T. McKenna & E.B. ScordoBGC Engineering Inc., Vancouver, British Columbia, Canada

ABSTRACT: The Oil Sands Research and Information Network commissioned BGC Engineer-ing Inc. to conduct a review of 34 oil sands tailings technology options. BGC’s report provided anoverview of oil sands processing and tailings characteristics and: summarized what is known and notknown about each option; described the state of commercial readiness of the technology; and identi-fied gaps in the knowledge. The 34 technologies fall into five main categories: Physical/MechanicalProcesses, Natural Processes, Chemical/Biological Amendments, Mixtures/Co-disposal, and Per-manent Storage. The review concluded that there is no single method of dewatering which workswell for all tailings. Similarly, experience has shown that there is unlikely to be one unique solutionto the problem of tailings disposal.

1 STUDY BACKGROUND

Environmental management of oil sands tailings is a key challenge for industry and provincialregulators. Although industry has made progress, through significant investments in research anddemonstration, regulators took a major step in 2009 to drive more rapid technology adoption. TheEnergy Resources Conservation Board’s Directive 074 (ERCB 2009) requires that oil sands opera-tors deposit a significant portion of their annual production of fine tailings in Designated DisposalAreas, which must be formed in a manner that ensures trafficable deposits. The performance criteriaare based on the strength of the deposit. The minimum undrained shear strength of 5 kilopascals(kPa) must be attained for the material deposited in the previous year, and the deposit must beready for reclamation within five years after active deposition has ceased. The deposit will have thestrength, stability, and structure necessary to establish a trafficable surface. The trafficable surfacelayer must have a minimum undrained shear strength of 10 kPa.

1.1 The issue

Four active oil sands processing facilities (Suncor Energy Inc., Syncrude Canada Ltd., Shell CanadaLimited, and Canadian Natural Resources Limited) are currently producing tailings at mines northof Fort McMurray, Alberta, Canada. Additional mines are in various stages of the regulatoryor planning process. As of 2008, about 750 million cubic metres of Mature Fine Tailings existwithin the tailings ponds. If there is no change in tailings management, the inventory of fluidtailings is forecast to reach one billion cubic metres in 2014 and two billion in 2034 (Houlihan andHaneef 2008). As of the end of 2009, there are more than 130 square kilometres of tailings ponds(Government of Alberta n.d.). The media and environmental organizations frequently make noteof the fact that these tailings ponds are so large they can be seen on satellite photos and can easilybe found on Google Earth.

The main objective in treating the oil sands tailings is to remove water so that a trafficable load-bearing surface can be produced within a reasonable time-frame to allow subsequent reclamation.

381

Many technologies have been suggested and tried, but they have been rejected for lack of technicalor economical feasibility. However, technologies that were previously thought to be uneconomicmay now be seen as a viable option given the requirements of Directive 074 and changing economicconditions.

The search for a viable tailings dewatering technology will intensify as the industry is required tomeet the objectives of Directive 074, the already large quantities of liquid waste products generatedby the oil sands industry grows and tailings storage facilities fill nearer to capacity. With no uniqueand acceptable solution yet in sight, research is now focusing on schemes which utilize more thanone technology and combining them into a disposal package.

1.2 Previous reviews

Industry reviews tailings management options as part of environmental impact assessment work andregulatory applications. Recent examples include Syncrude Canada Ltd. (2008), Suncor EnergyInc. (2009) and Total E&P Joslyn Ltd. (2010).

In addition to these company-specific assessments there have been a few comprehensive assess-ments of technology options (Devenny 2009, Fine Tailings Fundamentals Consortium 1995, Flint2005, Fuhr et al. 1993).

1.3 OSRIN and BGC

The Oil Sands Research and Information Network (OSRIN) is a university-based, independentorganization that develops the best available knowledge about returning landscapes and waterimpacted by oil sands mining to a natural state and gets that knowledge into the hands of those whocan use it to drive breakthrough improvements in reclamation regulations and practices. OSRIN isa project of the University of Alberta’s School of Energy and the Environment and was funded bytwo grants from the provincial Department of the Environment and one from the Canada School ofEnergy and the Environment Ltd. One of the core areas that OSRIN is focused on is reclamation ofoil sands tailings, with particular emphasis on reducing the footprint and impact of tailings pondsand disposal areas.

Part of OSRIN’s function is to identify gaps in knowledge or work at undertaking a comprehensivereview of the available knowledge. To do this, OSRIN will commission scoping studies – studieswhich look at the potential technologies and approaches that offer solutions to challenges in oilsands reclamation. A scoping study provides an overview of the key information relating to anissue or research need.

BGC Engineering Inc. (BGC) is an international consulting firm specializing in geotechnicaland water resources engineering and applied earth sciences. Their core services are mining, energyand transportation. BGC’s mining team offers a comprehensive range of services for the industrycovering every phase of mining and has worked on a wide variety of mineral, metal and coal mineslocated in diverse project settings and environments.

1.4 Terminology

The following terms are used in this paper.ConsolidatedTailings:A non-segregating mixture of chemically amended fine and coarse tailings

that consolidates relatively quickly into solid landforms. The purpose of producing ConsolidatedTailings is to consume both legacy fines (Mature Fine Tailings) and new fines (Thin Fine Tailings)to create a land surface reclaimable to predominantly upland vegetation. To this end, ConsolidatedTailings has a sand:fines ratio that is less than about 5.5:1 (to permit useful levels of fines capture)to greater than about 3:1 (to allow rapid consolidation). Consolidated Tailings starts as a slurry andends as a semi-solid, loose, silty sand deposit that is dense enough and strong enough to supporthydraulic sand capping.

Disposal Area: An area dedicated solely to the deposition of captured fines using a technologyor a suite of technologies.

Fines: Solid grains with diameters less than 44 µm grain size (as measured with a wet sieve).Mature Fine Tailings: Fine tailings that have settled and dewatered to approximately 30 wt%

solids over a period of about three years after deposition. The rate of consolidation beyond thispoint is substantially reduced. Mature Fine Tailings behave like a viscous fluid.

382

Oil sands: A sand deposit containing a heavy hydrocarbon (bitumen) in the intergranular porespace of sands and fine-grained particles. Typical oil sands are composed of approximately 10 wt%bitumen, 5 wt% water, 85 wt% coarse sand (>44 µm) and a fines (<44 µm) fraction, consisting ofsilts and clays.

Sand:Fines Ratio: The weight of dry sand (>44 µm) to the weight of dry fines (<44 µm).Shell: Shell Canada LimitedSuncor: Suncor Energy Inc.Syncrude: Syncrude Canada Ltd.Tailings: A by-product of oil sands extraction typically comprised of process water, sands and

clays, with minor amounts of residual bitumen.Thickened Tailings: Tailings with added flocculant to cause the active minerals to bind together

and settle rapidly.Thin Fine Tailings: The segregating fines portion of the conventional tailings that are carried

into the tailings pond from the tailings slurry deposition point with the process water, which arepredominantly clay or clay sized (<2 µm) particles.

Whole Tailings: The complete tailings stream that exits the plant, comprised of water, sands,fines and residual bitumen.

2 OIL SANDS PROCESS AND TAILINGS

Athabasca oil sand (McMurray Formation) is a mixture of bitumen, mineral matter and water invarying proportions. The bitumen content ranges from 0 wt% to 19 wt%, averaging 12 wt%; watervaries between approximately 3 wt% to 6 wt%, increasing as bitumen content decreases; mineralcontent, predominantly quartz sands and silts, and clay, varies between approximately 84 wt% to86 wt%. Clays are present in the McMurray bitumen-containing deposits in discontinuous beds orbands varying from 1 cm to 15 cm in thickness (Chalaturnyk et al. 2002).

Oil sands processing begins with crushing of the excavated ore. The crushed ore is then condi-tioned with warm to hot water, steam, and process aides such as caustic (NaOH) or sodium citrate(Shell only) and hydrotransported via pipeline to the extraction plant. Bitumen is separated from thecoarse fraction as a floating froth in large gravity-separation vessels. The bitumen froth is furtherprocessed to remove fine solids. Typical bitumen recoveries range from 88% to 95% depending onoil sands grade and origin (Beier et al. 2009). Tailings include a mixture of solids (sand, silt andclay), water and residual bitumen. It is a common practice for the surface mined oil sands industryto define fines as mineral particles smaller than 44 µm, whereas geotechnical fines are defined assmaller than 75 µm. The tailings slurry is approximately 55 wt% solids (with the solids comprising82 wt% sand [Suncor and Syncrude, Shell’s tailings contain more sand] and 17 wt% fines). Upondeposition, the tailings stream segregates with the coarse fraction forming beaches and the finesflowing with the water stream (8 wt%) into the settling basin where the solids settle gradually toform a densified zone of fine tailings at depth (known as Thin Fine Tailings). After a few years thefines settle to 30 wt% to 35 wt% and are referred to as Mature Fine Tailings. Mature Fine Tailingswill remain in a fluid-like state for decades because of their very slow consolidation rate (Kasperski2001, MacKinnon 1989). Significant portions of the fines remain in suspension after depositionresulting in a tailings management challenge for the industry.

Oil sands tailings are not a consistent product. The volume of solids, fines and bitumen can varyover a wide range, depending on variations in the ore from the mine, and on various operatingand upset conditions within the extraction plant. In summary, for every unit volume of bitumenrecovered, there are 7 to 8 volume units of wet sand and Mature Fine Tailings that need to behandled, and 10 volume units of water (recycle and make up) that are pumped around the system(Flint 2005). About 65% of the water used in the extraction process is recycled. The balance –about three cubic metres of water per cubic metre of bitumen – is trapped in the tailings pond (Flint2005). This water is responsible for continually rising pond volumes.

The major clay components of the McMurray Formation are 40 wt% to 70 wt% kaolinite; 28 wt%to 45 wt% illite and 1 wt% to 15 wt% montmorillonite (Chalaturnyk et al. 2002). It is believed thatmontmorillonite and illite clays are responsible for the processing and compaction problems inoil sand extraction and tailings disposal. These two clays, degraded by weathering or the actionof caustic soda additives, and coated with bituminous residues appear to be the main cause of the

383

gel-like structure formation in the hot water extraction tailings and for the ion exchange mechanismin the tailings ponds (Chalaturnyk et al. 2002).

Clay minerals, in the presence of caustic soda, possess an enhanced negative surface charge whichpromotes dispersion of the particles, inhibiting their settling and consolidation. Addition of sodiumhydroxide in the extraction process, therefore, results in a paradox. Dispersion of the clays, whichis necessary for efficient bitumen extraction by flotation, prevents rapid dewatering (sedimentationand consolidation) of the tailing’s clays. Adding sodium ions to the oil sand extraction processessentially creates a very undesirable condition as far as tailings disposal is concerned.

The dispersant effect of sodium ions (monovalent) can be counteracted and controlled to someextent by the addition of calcium ions (divalent). This cation exchange process, and the affinity ofcalcium ions for the clay surface, plays an important role in almost all tailings treatment strategies(Mikula et al. 2008).

The water holding capacity of Mature Fine Tailings, or the slow consolidation rate, is governedby the surface properties of the minerals. The forces that affect colloidal particles in suspension anddetermine the final settled volume, permeability, and strength of the material have four essentialcomponents (Fine Tailings Fundamentals Consortium 1995): electrostatic, steric, Van der Waals,and hydration. A knowledge and understanding of these components will help explain why so manyconventional solutions to the clay tailings disposal problem have been unsuccessful in the oil sandsindustry.

Some basic geotechnical properties of Mature Fine Tailings are summarized as follows (FineTailings Fundamentals Consortium 1995):

– The mean particle size of the fine tailings is between 5 and 10 µm;– The average solids content of Mature Fine Tailings is about 33 wt% which is an average void

ratio of 5;– The permeability of the Mature Fine Tailings is in the range of 1 × 10−6 to 1 × 10−9 m/s which

accounts for its slow rate of consolidation;– The liquid limit ranges from 40% to 75%;– The plastic limit ranges from 10% to 20%;– The viscosity varies from 0 to 5000 cP and it increases as time passes (after it is disturbed or

deposited); and– Typical values of shear strength (primarily from vane shear tests) vary from 0.2 to 2.5 kPa.

3 STUDY METHODS

OSRIN commissioned BGC to conduct an in depth scoping study of the state of knowledge relatedto oil sands tailings treatment technologies. The objective of the scoping study was to help facilitateresearch and information sharing among a variety of stakeholders and establish an understandingof the status of tailings treatment technology in the Athabasca Oil Sands Region.

BGC reviewed peer-reviewed literature and grey literature to gather information on the tech-nologies. A key observation from the initial information gathering phase of the project was thatwhile a large amount of work has been undertaken by various research organizations and oil sandsmine operators to characterize oil sands tailings materials and techniques for efficiently removingwaters, much of this work is not easily accessible to the public so these extensive efforts are notproperly acknowledged. A number of these scientific and engineering advances are published indiverse journals and specialized conferences making it a challenge to compile the important litera-ture. Other work has not been published in the open literature, but rather is contained in reports thatmay be more difficult to locate and obtain. Other components of the research information gatheredby the oil sands operators from field demonstration tests are proprietary and not accessible to thepublic.

Thirty four technologies were identified for analysis. They fall into five main categories, withmany of the technologies representing variations on a theme:

– Physical/Mechanical Processes,– Natural Processes,– Chemical/Biological Amendments,

384

– Mixtures/Co-disposal, and– Permanent Storage.

BGC produced a report (BGC in prep.) that includes an overview of the oil sands extractionprocess that produces tailings, an overview of tailings properties, and for each of the 34 technologies:

– A brief description of the technology;– Pros, including benefits and advantages;– Cons, including challenges and disadvantages;– Knowledge gaps, including an assessment of what is missing and what research needs to be

conducted to take the technology to the next level or stage. The information provided is basedon available published literature and the authors’ opinions; and

– Stage of Technology: providing an assessment of the maturity of the technology. Four stages areconsidered: Basic research, Applied research and demonstration, Commercial demonstration,and Mature (operates commercially).

The report also contains a bibliography with over 175 references.

4 THE TAILINGS TECHNOLOGIES

A brief overview of the category and the key technologies in each category is provided below;details are presented in BGC (in prep.).

4.1 Physical/mechanical processes

Physical/mechanical processes involve the use of a variety of technologies to separate the waterfrom the solids. The 12 processes reviewed include:

– Filtered whole tailings: Filtering can take place using pressure or vacuum force. Drums, hori-zontally or vertically stacked plates and horizontal belts are the most common filtration plantconfigurations.

– Cross flow filtration of whole tailings: In cross flow filtration, the feed is passed across thefilter membrane (tangentially) at positive pressure relative to the permeate side. A proportionof the material which is smaller than the membrane pore size passes through the membrane aspermeate or filtrate; everything else is retained on the feed side of the membrane as retentate.With cross flow filtration the tangential motion of the bulk of the fluid across the membranecauses trapped particles on the filter surface to be rubbed off. This means that a cross flow filtercan operate continuously at relatively high solids loads without blinding.

– Filtered coarse-fraction tailings: This technology consists of filtering and dry stacking of thecoarse fraction (cyclone underflow tailings) of the tailings slurry. Variants include adding somefines to the mix prior to filtration. Cyclone underflow tailings are usually stripped of some finesand water, and are not too dissimilar, although somewhat more variable, in composition thancomposite/consolidated tailings (Sobkowickz and Morgenstern 2009).

– Filtered thickened tailings: This technology includes filtration and dry stacking of thickenerunderflow (predominantly fines). Other fine tailings streams may include centrifuge fine tailingsand Mature Fine Tailings. This technology has been proposed, but seems impractical due to thehigh fines content.

– Centrifuge: Centrifuges are used extensively in oil sands froth treatment, but they have notbeen used to process Mature Fine Tailings. Centrifuge technology to produce dry tailings wasevaluated in the past with some success, but the cost was unacceptable at that time. A betterappreciation of the long term costs of Mature Fine Tailings storage has prompted a re-evaluationof this technology. The use of additives to improve centrifuge performance has significantlyimproved the results which can be achieved. Centrifuge technology has been developed at benchscale at CANMET (Mikula et al. 2008, 2009) on Athabasca oil sands fluid fine tailings and hasbeen successfully piloted in demonstration plants on Syncrude’s Mature FineTailings (Fair 2008).

– Thermal drying of Mature Fine Tailings: This technology consists of heating Mature Fine Tail-ings in an oven/kiln to reduce the moisture content of Mature Fine Tailings. Thermal drying can

385

remove water from tailings to a significantly higher degree than all other dewatering processes.Solids reaching 90 wt% are attainable, but the tailings are typically dewatered to a minimum of18 wt% to 20 wt% solids before being directed to the drying facility (BCI Engineers & Scientists2007). This technology has been suggested, but not seriously researched because it is consideredimpractical in cold regions.

– Electrical treatment: This technology is the application of a direct current electric field to a clayslurry. The electrical field causes negatively charged clay particles to migrate to the positive(anode) electrode, resulting in accelerated sedimentation.

– Blast densification: Blast densification is predominantly used by the geotechnical communityto pack loose, saturated, medium to coarse materials. Explosive compaction is carried out bysetting off explosive charges in the ground. The energy released causes liquefaction of the soilclose to the blast point and causes cyclic straining of the soil. This cyclic strain process increasespore water pressures and, provided that strain amplitudes and numbers of cycles of strainingare sufficient, the soil mass liquefies (i.e. pore water pressures are temporarily elevated to theeffective vertical overburden stress in the soil mass so that a heavy fluid is created) (Gohl et al.2000). Liquefaction of the soil, followed by time-dependent dissipation of the excess waterpressures, causes re-consolidation within the soil mass.

– Wick drains: Prefabricated vertical drains (also called wick drains or band drains) greatly facili-tate the dewatering process, by providing a suitable conduit to allow the pore water to escape veryquickly. Vertical wicks can be economically installed in tailings deposits at close spacing, short-ening the flow path of the water, and thereby expediting the consolidation process. Consolidationof soft cohesive soils using wick drains can reduce settlement times from years to months.

– Surcharge loading: Surcharge loading offers a time-tested procedure for accelerating the consol-idation and dewatering process and increasing the rate of strength gain of poorly-consolidatedclays, but the low shear strength of high water content tailings usually makes it difficult to applythe surcharge without causing a stability failure and consequent mud wave. Surcharging hasbeen theoretically interesting but totally impractical as a process for dewatering and compactionof tailings.

– Placement of Consolidated Tailings under Mature Fine Tailings: Placing Consolidated Tailingsunder Mature Fine Tailings changes the release water composition by reducing the concentra-tions of calcium and sulphate ions, and the electrical conductivity; the Mature Fine Tailings hasa steady densification rate; and the Consolidated Tailings densification rate remains the same.

– Increasing tailings density: This technique aims to reduce segregation by increasing solids con-tent of tailings sand slurries in the pipeline prior to beaching. This technology has been suggested,but not seriously researched.

Dewatering using physical/mechanical processes (centrifuges, filter presses, vacuum filters,thermal drying, electrical, etc.) involves costly machinery/equipment and the results are oftenpoor. In particular, oil sands tailings have both high fines and residual bitumen so that it is difficultwithout chemical reagents to effectively dewater by pressure filtration. High investment costs arenecessary to obtain satisfactory results. There are also two little-considered drawbacks which arecommon to all mechanical dewatering processes; they produce a cake which must be transportedto the disposal site and the cake must be stacked on arrival at the disposal site. The long termobjective of this class of technologies is to develop stackable tailings, with significant benefitsfor land reclamation. From this group, centrifugation, and wick drains combined with surchargeloading seem to have good potential for dewatering tailings at a large commercial scale. Syncrudeis currently conducting a field research demonstration using centrifuge dewatering technology.Suncor is testing the use of wick drains on one of their Consolidated Tailings ponds.

4.2 Natural processes

Natural processes involve using environmental or geophysical processes to remove water fromsolids. The five processes reviewed include:

– Sedimentation/self-weight consolidation: Sedimentation and consolidation are natural dewater-ing processes that use the force of gravity to separate the suspended solids from the tailingsstream.

386

– Evaporation/drying: This technology consists of depositing Consolidated Tailings or MatureFine tailings in thin lifts and allowing the lifts to desiccate (remove moisture) by evaporativedrying. Promotion of natural drying is generally considered the most cost-effective means ofdewatering fine-grained material.

– Accelerated dewatering: This technology involves the deposition of a large volume of Mature FineTailings in a dedicated cell and the subsequent use of evaporation and rim ditching to acceleratedewatering to create a final deposit of suitable density to support dry landscape reclamation.The process relies on the evaporation rate from the Mature Fine Tailings deposit exceeding therate of water release, thereby allowing a crust to form on top of the deposit (Lahaie 2008). Thiscrust in turn allows a ditching network to be established that promotes further dewatering andcrust thickening as the ditch is worked toward the bottom of the deposit.

– Freeze/thaw: This technology consists of depositing Consolidated Tailings or Mature Fine Tail-ings in multiple thin layers which are allowed to freeze and then the frozen mass is allowed tothaw the following summer. The freezing cycles causes consolidated soil-like “peds” to form,developing a fissured structure throughout the deposit which quickly drains when thawed. Aconsiderable amount of water is released when thin layers (5 to 15 cm) of Mature Fine Tailingsare subjected to freeze-thaw cycles (Dawson & Sego 1993, Proskin 1998).

– Plant (evapotranspiration) drying: Suitable plant species can dewater tailings by transpirationthrough the leaves and associated root systems (Silva 1999). Plants can transpire large quan-tities of water during the growing season; the rate of water loss may exceed that of free waterevaporation and continue long after the surface has become dry.

Natural dewatering technologies are looked upon more favorably than most since they are depen-dent on the use of natural processes for their effectiveness, which essentially provides free energy.However, Fort McMurray’s northern location with its associated cold weather and moisture deficitmake success of these technologies very unpredictable.

Sedimentation and self-weight consolidation give only limited densification because the lowstress imposed by the low buoyant weight of the fined-grained particles is insufficient to overcomethe gel strength of the thixotropic tailings structure. The rate of dewatering is therefore slow and theextent of dewatering may be poor within a reasonable time-frame. Dewatering of fines by gravityprocesses actually appears to stop completely at approximately 30 wt% solids content. Developmentof a high solids content material which could be strong enough for use as a self-supporting mineback-fill is not viable solely by self-weight consolidation of current tailings streams.

Evaporation/Drying and Accelerated Dewatering offer considerable promise especially whenthey are considered as part of a treatment package. These technologies rely on solar radiation andwind action to accelerate the rate of water evaporation from soils. Accelerated dewatering uses therim-ditching technique to enhance the evaporation of tailings layers. Low pressure vehicles pullingploughs can be used to increase the exposed surface area and promote surface drainage by creatingshallow ditches and encourage desiccation by evaporation. Syncrude is currently conducting fieldtests to evaluate the suitability of the Accelerated Dewatering technology. Suncor is conductingfield tests of the evaporation/drying dewatering technology.

Freeze-thaw seems ideally suited to the geographical location of Fort McMurray. This techniquemay be useful as an adjunct to other processes, such as evaporation and accelerated dewater-ing. However fluid management at very cold temperatures and vagaries of the weather remain achallenge.

4.3 Chemical/biological amendments

Chemical/biological amendments involve changing the properties of the tailings to remove water.The seven technologies reviewed include:

– Thickening process: Thickened tailings technology involves rapid settling and sedimentation ofsuspended fines within a process vessel through the addition of chemicals that aid in flocculatingthe fines solids. Thickeners incorporate moving components, such as rakes, which shear flocsand promote removal of entrapped water. Normal thickener processing yields a density of about30 wt% solids. Higher densities are reported but may be due to the addition of sand.

– In-line thickening of tailings: This technology consists of injecting and mixing coagulants intothe high fines content cyclone overflow tailings in an in-line multi stage fashion. Conceptually

387

by binding fine particles at low solids content into flocs, the hydraulic conductivity is increased,tortuosity is decreased and the mass of the failing flocs is increased.

– Whole tailings coagulation: In this technology, a coagulant is added into the tailings pipeline togenerate whole-tailings Consolidated Tailings or partially segregating Consolidated Tailings.

– Whole tailings flocculation: In this technology, a flocculant is added into the tailings pipeline togenerate whole-tailings Consolidated Tailings or partially segregating Consolidated Tailings.

– In-situ biological treatment: In this technology, inoculation or enhancement of bacterial actionare used to densify Mature Fine Tailings or fine tailings. Biological methane production canaccelerate densification of Mature Fine Tailings by generating channels in the tailings wheregas bubbles rise. These channels can then allow drainage of water due to excess pore pressureswithin the tailings mass.

– In-situ chemical treatment: This technology consists of injecting and mixing chemical reagentsinto Mature Fine Tailings in situ. Chemical additives injected in tailings ponds can increase theefficiency of the consolidation process by changing the pH or by promoting coagulation and/orflocculation.

– Reduce dispersion of fines in process: This technology consists of changing the tailings waterchemistry to reduce the amount of fines dispersion and trap more fines within the tailings sand.

Chemical amendment technologies appear to offer a unique benefit because they are capableof altering the properties of the clays responsible for tailings formation and thus they offer anew dimension in dewatering technology. Chemical amendment may also be used to assist otherdewatering methods. For example, the rate of sedimentation during mechanical thickening andcentrifugation may be accelerated significantly by the use of coagulants and flocculants. The use ofchemical additives has been studied for many years, but this technology has not been accepted forreasons of operational practicability and cost-effectiveness. Recent research conducted on polymershas show promise that this technology can be effective to dewater tailings. Suncor and Syncrudeare using polymers in their centrifuge, thickener and evaporation/drying field tests to improve theperformance of the technologies. There remain uncertainties regarding the potential environmentalimpacts of chemical additives, especially polymers.

4.4 Mixtures/co-disposal

Mixtures/co-disposal technologies involve mixing tailings streams with a variety of available soilmaterials and waste products to increase tailings density. The seven technologies reviewed include:

– Consolidated Tailings: technology involves mixing densified extraction tailings (coarse sandfrom cyclone underflow tailings) and Mature Fine Tailings with an amendment (typically gyp-sum) to create a non-segregating slurry, with subsequent discharge into a tailings pond to forma rapidly consolidating, soft, cappable deposit.

– Mature Fine Tailings spiked into whole tailings: This technique consists of injecting MatureFine Tailings into a fresh tailings stream to form a segregating slurry with a high fines content.The concept is that a high proportion of fines can be captured in the beach formation followinghydraulic discharge.

– Mixing Mature Fine Tailings with Clearwater overburden: This technology consists of mixingMature Fine Tailings with Clearwater clay to form a semi-solid mixture suitable for storage inpolders. The Clearwater formation clay is very dry and has a considerable affinity for water.

– Mixing Mature Fine Tailings with other overburden: This technology consists of mixing MatureFine Tailings with glacial materials (tills, clays, sands) to form a semi-solid mixture suitable forstorage in polders. This technique is similar to the previous one, but the glacial materials havea lower water absorption capacity.

– Mixing Mature Fine Tailings with reclamation materials: This technology consists of mixingMature Fine Tailings with salvaged peat soils to form a semi-solid mixture suitable for earlyreclamation.

– Mixing Mature Fine Tailings/Consolidated Tailings with coke: This technology consists ofmixing Mature FineTailings/ ConsolidatedTailings with coke from the bitumen refining process.

– Mixing thickened tailings with sand: This process involves mixing tailings sand with thickenedtailings to form a non-segregating mix suitable for poldering. The deposit would be furtherdewatered via evaporation and then capped with a layer of overburden and seeded.

388

From the mixture/co-disposal group, only Consolidated Tailings is operating commercially(Suncor in 1995 and Syncrude in 2000). However, operators have experienced difficulty mak-ing, transporting and placing on-spec Consolidated Tailings; there is a tendency for the fines tosegregate leaving a weak, low sand:fines ratio material that is difficult to reclaim. The other issueis that supplemental sources of sand will likely be needed to treat the legacy Mature Fine Tailings.Consolidated Tailings is the most mature technology currently in practice. Suncor is planning toreplace this technology with their Tailings Reduction Operation and Mature Fine Tailings dryingtechnology in the near future (Suncor 2009).

4.5 Permanent storage

Permanent storage technologies acknowledge the complexity and cost associated with tailingstreatment and instead opt to store tailings above or belowground in their original form. The threetechnologies reviewed include:

– Mature Fine Tailings water-capped lake: This involves placement of Mature Fine Tailings ina mined out pit, followed by introduction of a water cap over this deposit. The water used inthe cap may range from natural surface waters to various blends of fresh and process-affectedwaters. The water is initially at least five metres in depth, but increases over time as Mature FineTailings consolidate.

– Pit lake storage: The Pit Lake technology is similar to the Mature Fine Tailings water cappingconcept. The difference is that the Pit Lake will have three distinct layers: a bottom substrate,soft tailings (on top of the bottom substrate, it can be Consolidated Tailings, Mature Fine Tailingsor Thickened Tailings), and the water cap.

– Storage of Mature Fine Tailings in underground caverns: This technology consists of injectingMature Fine Tailings in underground caverns or deep wells where future contact with humansis unlikely. It is doubtful whether government regulatory bodies would approve this technology.This technology has been suggested, but not seriously researched.

Syncrude will be the first operator to demonstrate the Mature Fine Tailings water-capped laketechnology with their Base Mine Lake site. The water cap will be placed in 2012 and detailedmonitoring will begin aiming to answer many uncertainties regarding function and success includ-ing water quality and toxicity, sustainability and liability. Other pit lakes have been proposedto permanently store Mature Fine Tailings, Consolidated Tailings or Thickened Tailings creatingself-sustaining aquatic ecosystems (Clearwater Consultants 2007).

5 CONCLUSIONS

A number of technologies have been considered for treatment and/or storage of oil sands tailings(whole tailings or the individual tailings components). Consolidated Tailings is the primary treat-ment method for the majority of the tailings currently produced although other technologies suchas use of flocculants are also being used. A demonstration of the Mature Fine Tailings water-cappedlake will be started in 2012.

Selection of viable technologies must consider the trade-offs between of a number of factors,including:

– Technical feasibility (does it work?);– Dewatering efficiency;– Optimizing moisture content for pumping or other types of transport;– Winter operation (Fort McMurray’s long term mean annual temperature is 0.7◦C, with January

being the coldest month (−18.8◦C) and July the warmest (16.8◦C);– Provision of an appropriate substrate for subsequent, quick reclamation;– Operational practicability (is it practicable on a large scale?);– Does it address legacy tailings or only those currently produced;– What is the quality of the water expressed from the tailings (is it suitable for recycle and use in

the production plant or for discharge to the environment);

389

– Cost-effectiveness (is it affordable?) – costs are often expressed in $/bbl to allow comparisonsbetween technologies and an assessment of the incremental impact on profit margins;

– Robustness (can it deal with the variability in the tailings consistency); and– Does it produce impacts that are worse that the original problem (i.e. are there unintended

consequences such as the production of more saline water or the potential impacts of chemicaladditives).

BGC provided an assessment of the knowledge gaps for each technology (BGC in prep.). Beloware listed general technology gaps identified in the overall tailings treatment processes (some ofthem proposed by Flint 2005):

– Quantification and modeling of fine tailings dispersion caused by the addition of sodiumhydroxide in the extraction process;

– A unified sedimentation-consolidation model to predict settlement of material deposited intailings ponds;

– Mature Fine Tailings morphology and characteristics;– Sand, clay, organics and water interactions in tailings;– Role of chemical additives (polymers) in modifying tailings properties;– Pumping of high solids content materials;– Better mechanical dewatering means;– Pond emission quantification, characterization and reduction; and– Quality of the water expressed from tailings as it impacts use in the extraction process or discharge

to the environment.

REFERENCES

BCI Engineers & Scientists Inc. 2007. Rapid dewatering techniques for dredged lake sediments. Literaturereview and summary report. Prepared for: St. Johns River Water Management District, Palatka, Florida,USA.

Beier, N., Alostaz, M. & Sego, D. 2009. Natural dewatering strategies for oil sands fine tailings. In Tailingsand Mine Waste ’09. Banff, Alberta, Canada. University of Alberta, Department of Civil & EnvironmentalEngineering, Edmonton, Alberta, Canada.

BGC Engineering Inc. in prep. State of knowledge: Oil sands tailings technology. Oil Sands Research andInformation Network, School of Energy and the Environment, University of Alberta, Edmonton, Alberta,Canada.

Chalaturnyk, R.J., Scott, J.D. & Ozum, B. 2002. Management of oil sands tailings. Petroleum Science andTechnology 20(9): 1025–1046.

Clearwater Consultants. 2007. End Pit Lakes – Technical Guidance Document. Cumulative EnvironmentalManagement Association. End Pit Lakes Subgroup, Fort McMurray, Alberta, Canada.

Dawson, R.F. & Sego, D.C. 1993. Design concepts for thin layered freeze-thaw dewatering systems. CanadianGeotechnical Conference. pp. 283–288.

Devenny, D.W. 2009. A screening study of oil sand tailings technologies and practices. Prepared for AlbertaEnergy Research Institute, Edmonton, Alberta, Canada.

Energy Resources Conservation Board (ERCB). 2009. Directive 074: Tailings Performance Criteria andRequirements for Oil Sands Mining Schemes. Alberta Energy Resources Conservation Board, Calgary,Alberta, Canada. 14 pp.

Fair, A. 2008. The Past, Present and Future of Tailings at Syncrude. In First International Oil Sands TailingConference. Oil Sands Tailings Research Facility, University of Alberta, Edmonton, Alberta, Canada.

Fine Tailings Fundamentals Consortium. 1995. Advances in oil sands tailings research. Alberta Departmentof Energy, Oil Sands and Research Division, Edmonton, Alberta, Canada.

Flint, L. 2005. Bitumen recovery technology, a review of long term R & D opportunities. LENEF Consulting,Calgary, Alberta, Canada. 210 pp.

Fuhr, B.J., Rose, D.L. & Taplin, D. 1993. Catalogue of technologies for reducing the environmental impact offine tailings from oil sand processing. Alberta Conservation and Reclamation Council Report No. RRTAC93-3. 63 pp.

Gohl, W.B., Jefferies, M.G., Howie, J.A. & Diggle, D. 2000. Explosive compaction: design, implementationand effectiveness. Geotechnique 50(6): 657–665.

390

Government of Alberta. Alberta’s Oil Sands. Fact Sheets – tailings management. (website last viewed April2010).

Houlihan, R. & Haneef, M. 2008. Oil sands tailings: Regulatory perspective. In First International Oil SandsTailing Conference. Oil SandsTailings Research Facility, University ofAlberta, Edmonton, Alberta, Canada.

Kasperski, K.L. 2001. Review of research on aqueous extraction of bitumen from mined oil sands. CANMETWestern Research Centre, Natural Resources Canada, Devon, Alberta, Canada.

Lahaie, R. 2008. Syncrude Canada Ltd. – New Tailings Concepts. In First International Oil Sands TailingsConference. Oil Sands Tailings Research Facility, University of Alberta, Edmonton, Alberta, Canada.

MacKinnon, M.D. 1989. Development of the tailings pond at Syncrude’s oil sands plant: 1978–1987. AOSTRAJournal of Research 5: 109–133.

Mikula, R.J., Munoz, V.A. & Omotoso, O. 2008. Centrifuge options for production of “dry stackable tailings”in surface mined oil sands tailing management. In Canadian International Petroleum Conference. Calgary,Alberta, Canada.

Mikula, R.J., Munoz, V.A. & Omotoso, O. 2009. Centrifugation options for production of dry stackable tailingsin surface mined oil sands tailings management. Journal of Canadian Petroleum Technology. 48: 19–23.

Proskin, S.A. 1998. A geotechnical investigation of freeze-thaw dewatering of oil sands fine tailings. In Civiland Environmental Engineering. University of Alberta, Edmonton, Alberta, Canada.

Silva, M.J. 1999. Plant dewatering and strengthening of mine waste tailings. In Civil and EnvironmentalEngineering. University of Alberta, Edmonton, Alberta, Canada.

Sobkowicz, J. & Morgenstern, D.N.R. 2009. A geotechnical perspective on oil sands tailings. In Tailings andMine Waste ’09. Banff, Alberta, Canada.

Suncor Energy Inc. 2009. Application for tailings reduction operations. Suncor Energy Inc., Fort McMurray,Alberta, Canada.

Syncrude Canada Ltd. 2008. Application for approval of the Southwest Sand Conversion Project. SyncrudeCanada Ltd., Fort McMurray, Alberta, Canada.

Total E&P Joslyn Ltd. 2010. Tailings management. In Joslyn North Mine Project: Project Update. Total E&PJoslyn Ltd., Calgary, Alberta, Canada. pp. 6-1 to 6-14.

391


Case study: Sand capping of weak tailings at Suncor’s Pond 1

E. Olauson & R. DawsonNorwest Corporation, Vancouver, BC, Canada

P.S. WellsSuncor Energy, Inc., Fort McMurray, AB, Canada

ABSTRACT: As one of the final phases of surface closure operations on Pond 1, due to becompleted in 2010, Suncor has been undertaking the capping of an 11 ha area of soft tailings. Afterseveral years of investigation using cone penetration test methods and sampling to determine thegeneral geotechnical properties of the tailings (strengths in the range of 1 to 8 kPa were measured),a mechanically placed, geogrid-reinforced sand cap was designed for the area. The objective of thecap was to create a trafficable surface on the pond and to apply a surcharge to promote consolidationof the underlying material. The design basis included carrying out the construction work in thewinter months, when the benefits of frozen tailings could be realized. Construction progressedfrom January to April 2010 and involved using small equipment (D3 and D6 dozers) to place thetwo 1 m sand lifts. The cap provides a trafficable surface for carrying out reclamation activitiesand monitoring tailings consolidation and strengthening.

1 INTRODUCTION

The Suncor Energy, Inc. (Suncor) oil sands operation is located approximately 25 km north of FortMcMurray, Alberta. As part of the initial start-up operation in 1967, Pond 1 was constructed on theshores of the Athabasca River with the first containment structure extending from the shoreline andacross an island known as Tar Island. Now known as Tar Island Dyke, this forms the east and southcontainment structures of Pond 1 as shown in Figure 1. Pond 1 is the first tailings pond in the oilsands industry and was in active use between 1967 and 1999. Seeding and planting of the outer dykeslopes was conducted in the 1970’s and 80’s as part of early reclamation trials. Following severalyears of closure planning and infilling trials, full scale closure operations began in 2007. Dredgingof Mature Fine Tailings (MFT) and replacement with hydraulically placed sand was completed inlate 2009. These efforts were part of Suncor’s commitment to achieve the closure of Pond 1 to atrafficable surface by the end of 2010.

One area of the pond that was not suitable for dredging and replacement operations is knownas the Plant 4 beach, as shown in Figure 1. These tailings are the result of froth treatment andare characterized by predominantly silt-sized particles, very low consolidation rates, and elevatedhydrocarbon contents. It was decided to keep the deposit in place rather than dredge it to anotherlocation, and this required improving the trafficability of the materials in order to provide a longterm, stable surface for final reclamation. Several methods of improving the trafficability wereexamined. The focus of this paper is on the characterization of the weak froth tailings, the trials toimprove the strength of the tailings, and the final cap design and construction activities carried outas part of the capping work.

2 BACKGROUND CAPPING STUDIES

Suncor has been carrying out research and development activities for several years to evaluatepractical tailings capping alternatives for their tailings ponds. Suncor conducts annual investigations

393

Figure 1. Suncor pond 1.

of its tailings ponds, which include surveys of the pond bottom using spiked weights, as well asdetailed sampling and geotechnical vane strength measurements at several predetermined locations.Data from these investigations has been used to develop several tailings capping research projectsand has helped to guide reclamation planning.

In 2007, Norwest Corporation carried out a series of experiments related to weak tailings capping,tailings stabilization, and tailings dewatering. These experiments focused on methods of creatinga trafficable surface on the weak tailings and techniques for decreasing the consolidation time ofthe tailings by enhancing drainage and applying surcharge loads. The experiments and results aresummarized below.

1. Coke capping: Petroleum coke was rained-in through a water cover over weak tailings (44%solids with no measurable vane strength) in a small test cell. The density of the coke cap wasless than the underlying tailings, thus the cap was able to float on the tailings, which essentiallybehaved as a heavy fluid.a. Plate bearing tests: These tests were carried out on the coke cap to determine the strength

of the floating cap system. The results showed that for a given cap thickness, the ultimatebearing pressure decreased rapidly as the size of the bearing plate increased. When these tests

394

were conducted on stronger tailings (70% solids and vane strengths ranging from 2 to 8 kPa),the ultimate bearing capacity was about two times greater than for the weaker tailings.

b. Dewatering: Prefabricated vertical drains (or wick drains) were investigated to determinetheir ability to dewater the tailings beneath the coke cap. An increase in the solids contentof the tailings directly adjacent to the wick drain (within 2.5 cm) was observed, however nofurther dewatering of the tailings was noted.

2. Sand capping: A sand cap was hand shoveled over tailings at 39% solids with no measurablevane strength. Geotextile and geogrid were placed at the tailings surface as reinforcement andwick drains were installed through the geosynthetic reinforced sand cap. During installation,MFT began squeezing up through the holes created by the mandrel until eventually the entiresand cap was overlain by a shallow pool of MFT.

3. In-situ stabilization: Shallow soil mixing is a proven technique for remediating industrial slurries.A series of field trials was carried out using imported fly ash and in-situ mixing with an excavatorin small 5 m by 5 m test cells. The work was carried out under the auspices of a commercial soilmixing contractor, using experienced operators and proven methods. It was demonstrated thatthese techniques could produce caps of varying strength, depending on the mix design (uniaxialcompressive strength values in the range of 170 to 650 kPa were measured). The high cost ofimporting the ash along with transportation logistics were considered to be impediments to largescale applications.

During 2008 and 2009, Robertson GeoConsultants Inc. and Suncor completed further researchand trials into floating coke caps and tailings dewatering (Wells & Caldwell, 2009). Small-scalecoke caps were placed over approximately 2 ha of weak tailings contained in Suncor’s TailingsPond 5 during the winter months, using geogrid and geotextile as reinforcement. Wick drain fieldswere installed and target depths were varied to determine how the flow changed with the variousgradients. This work helped to develop the design basis for Suncor’s mechanically placed Pond 5coke cap, constructed over approximately 100 ha of weak tailings during the winter of 2010. Thiswork is not discussed in this paper.

3 TAILINGS CHARACTERIZATION

Figure 2 shows the sand capped area with the zones of weakest tailings identified, along with ageneralized strength profile from a hole near the centre of the pond, and Figure 3 shows a generalizedcross-section. The area was characterized through several investigations using cone penetration testmethods, vane strength measurements, and sampling. Table 1 summarizes the general geotechnicalproperties of the identified tailings.

The Plant 4 tailings are a froth treatment tailings deposit and are generally composed of avery narrow range of silt sized particles (between approximately 0.08 and 0.015 mm), with minoramounts of clay and sand. Mineral contents in the hydraulically placed deposit are typically around50% near surface and gradually increase with depth. The Plant 4 tailings contain significant amountsof bitumen, ranging from 7 to 20%. Consolidation behaviour of this material may be difficult topredict given the multi-phase pore fluids. Undrained strengths range between about 1 and 8 kPa.The Plant 4 tailings at the surface formed a drying crust that could produce mounds up to 1 m inheight. Figure 3 shows that in the centre of the pond, the Plant 4 tailings extend from surface toapproximately 10 m below the tailings surface.

MFT was identified below the Plant 4 tailings in most areas, and along some of the edges ofthe pond. The MFT is more fluid-like and contains more clay than the Plant 4 tailings, and ballpenetrometer and vane strengths measured were around 0 to 2 kPa. The MFT is up to 14 m thick,with typical mineral contents of 35–50% near the surface.

4 CAP DESIGN BASIS

There were two primary goals for the Plant 4 tailings capping design. The first was to create atrafficable surface by the end of 2010 upon which reclamation cover materials could be placed.The second was to apply a surcharge load to promote consolidation of the underlying tailings.

395

Figure 2. Sand capped area and typical strength profile through the centre of the pond.

Figure 3. Generalized cross-section A-A’.

While petroleum coke best meets the trafficable surface criteria, the low bulk density coupled withthe potential for chemical interactions between the froth tailings pore fluids and the coke madeit unsuitable for use. As such, the cap was designed with sand as the capping material, placedmechanically over frozen tailings using small equipment (D3 and D6 size dozers).

There is significant experience with the use of ice covers on lakes and rivers for transportationpurposes (reviews of this experience provided by Gold, 1971 and US Army Corps of Engineers,1996). Failures in ice covers largely occur due to the stresses associated with the maximum bendingmoment, where the ice cover floating on water can be considered as a plate on an elastic foundation.The ice thickness required for a given load is related to the square root of the load by a constant,A, as shown in Equation 1.

where h = ice thickness; A = empirical constant; P = load.

396

Table 1. Summary of typical Plant 4 tailings and MFT properties.

Property Plant 4 MFT

Bitumen content, % 15 5Mineral content, % 50 35–40Silt content, % 85 50Clay content, % 15 50Undrained shear strength, kPa 2–5 0–1

Figure 4. Access guidelines for frozen materials.

Figure 4 shows the relationship between ice thickness and applied load derived from guidelinesproduced by Work Safe Alberta (2009), for A values of 3.5 and 6. The A value of 3.5 representsa “low risk” according to Work Safe Alberta and requires minimal hazard controls (manual icemeasurements and quality assessments, routine worksite observations, and repairs and maintenanceas needed). Gold compiled data on observed failures of ice covers and found that an A value of 3.5envelopes more than 90% of the case histories. The design line used for the construction of the sandcap on Pond 1 follows the line calculated with an A value of 3.5, which shows that for a D3 dozer,0.5 m of ice is required. Ice thicknesses were frequently measured as the cap placement progressedacross the entire pond. The authors have experience at other oil sands operations showing that theice guidelines are suitable for some frozen tailings materials.

Geogrid reinforcement was rolled out over the entire frozen tailings surface prior to the place-ment of the first sand lift, and a second layer was placed between the first and second lifts. Thespecified grid was Tensar BX1200, with manufacturer specified ultimate tensile strengths of 19.2and 28.8 kN/m in the machine and cross-machine directions, respectively. Design calculations fora reinforced cap on thawed ground showed small deformations for light-weight equipment loadingconditions; thus the geogrid provides additional support in case of weak frost and/or thin frostground conditions.

397

Figure 5. Ice thickness measurement locations.

5 CONSTRUCTION ACTIVITIES

Construction of the sand cap began in mid-January 2010, when daily high temperatures in FortMcMurray had averaged around −15◦C for over a month. Ice thicknesses were measured prior tostarting work and throughout the construction period. The results are shown on Figure 5 and showthat there were some areas where frost thicknesses were less than the specified 0.5 m.

Geogrid was deployed using argos, from the east and west sides of the pond, and progressednorth. Overlaps were specified as 1 m and 6 m, for the sides and ends of the rolls, respectively, andzip-ties were used to hold the overlaps in place. Six metres of geogrid was also deployed on thebeach and covered with piles of sand, to act as anchors. Sand placement using D3 dozers followedshortly after the geogrid was rolled out. Sand stockpiles were placed around the perimeter of thepond and the dozers pushed the sand from both the east and west sides (up to 150 m), progressingtoward the north.

During sand placement, a small heave in the ice was consistently noted in front of the advancingsand push. This heave would crack at some locations, as shown in Figure 6, when an area with thinfrost was confined by the sand pushes, but was capped with minimal change to the constructionmethod. It was determined that pushing an initial lift of approximately 0.6 m was the preferredmethod by the dozer operators, as it would generally not crack the ice and was also a manageablequantity of sand to push out the significant distances required. An additional technique learnedduring the construction phase was to continuously advance the sand lift as one front, avoidingcreating long roads or peninsulas out into the pond. These peninsula roads were less stable andsignificant cracking was observed around them.

The 1 m sand cap was completed at the beginning of March 2010. A second layer of geogrid wasrolled out and then covered with a second metre of sand. A slotted pipe was placed down the centreof the capped area to act as a drainage conduit and three vertical riser pipes were placed along thispipe to act as collection points, where water could be pumped out.

The second sand layer was placed mainly with D6 dozers, as it was determined with some initialtrials that there was minimal deflection (less than 5 cm) in the first lift with this heavier equipment.

398

Figure 6. Example of typical deformation and cracking in front of sand lift that was successfully covered.

Deflection testing also showed minimal deflection with loaded 40 ton dump trucks near the edges ofthe capped area. This allowed for dumping at the crest of the cap instead of at the base, significantlyincreasing productivity.

The final approximately 1 ha of the second metre sand cap was completed using D3 dozers, asincreased movement and cracking was observed in the first metre lift at the north end of the pond,where thinner frost zones had been noted. This cracking and deformation can likely be attributedto some melting of the frozen tailings. During the construction of the first lift, temperatures roseabove 0◦C for a period of approximately two weeks. These temperatures combined with sunnyconditions and the exposed black geogrid likely contributed to the melting. The last area to becovered with the second sand lift was also likely composed of the weakest tailings, pushed to thisarea by the surrounding sand cap. These weak tailings were not able to support D6 traffic and assuch, deformed and cracked.

6 CURRENT POND STATUS

Since completion of the sand cap in early April 2010, Suncor has proceeded with placement ofapproximately 0.5 m of reclamation cover material over the capped area, using D3 and D6 dozersto spread the material. The reclamation strategy is to seed this material with a selection of nativegrasses.

Piezometers have been installed in the upper portions of the sand capped tailings, to monitorthe temperature and pore pressures beneath the cap and a settlement monitoring program is alsobeing developed. This program will help to predict the consolidation times and total settlementof the deposit, as well as identify the multi-phase pore fluids that may be expelled during theconsolidation process. It is expected that the capped area will settle significantly and will requireadditional material placement, to avoid creating a topographic low that will collect water.

7 SUMMARY

A trafficable surface was successfully constructed over 11 ha of weak tailings in Pond 1. The weaktailings deposit was up to 24 m thick and in-situ testing and sampling showed near surface solidscontents of approximately 50% and strengths in the range of 1 to 8 kPa. The cap was designed tobe placed with light equipment over a 0.5 m frost cap. It consists of two 1 m thick layers of sandseparated by biaxial geogrid reinforcement. The cap provides a trafficable surface for carrying outreclamation activities and monitoring tailings consolidation and strengthening.

399

REFERENCES

Gold, L.W., 1971. Use of ice covers for transportation. Can. Geotech. J. 8: 170–181.Government of Alberta, Work Safe Alberta, 2009. Best Practice for Building and Working Safely on Ice Covers

in Alberta. Publication No. SH010, October 2009.Wells, P.S., Caldwell, J. 2009. Vertical “wick” drains and accelerated dewatering of fine tailings in oil sands.

Proceeding of the 13th International Conference on Tailings and Mine Waste, 1–4 November 2009, Banff,Alberta, Canada.

US Army Corps of Engineers, 1996. Engineering and Design, ICE ENGINEERING. Department of the Army,Manual No. 1110-2-1612, December 31, 1996. Washington, DC, United States.

400


The use of geosynthetics in the reclamation of an oil sandstailings pond

Chris AthanassopoulosCETCO, Hoffman Estates, Illinois, USA

Patrick Sean WellsSuncor Energy Inc., Fort McMurray, Alberta, Canada

Serg TrincaNilex, Edmonton, Alberta, Canada

William UrchikCETCO, Pendleton, New York, USA

ABSTRACT: Suncor Pond 1, near Fort McMurray, Alberta, serves as an industry milestone — thefirst surface reclamation of an oil sands tailings pond. During the design and infill operations, a low-permeability hydraulic barrier was identified as an element of the reclamation cap needed to limitpercolation. A plastic-laminated Geosynthetic Clay Liner (GCL) was selected as the barrier layerin lieu of a traditional compacted soil liner. The GCL was selected for several reasons, including:low hydraulic conductivity (5 × 10−10 cm/sec), material consistency and quality control, simplicityand speed of installation, and ability to install in extreme cold temperatures. The paper will discusseach of these design and installation considerations, as well as providing performance data fromother similar applications. Construction of the geosynthetic capping system began in late 2009,with capping operations to be completed in 2010.

1 BACKGROUND

Full scale closure and reclamation activities have been underway at Suncor’s Pond 1 since 2007.Densified tailings (DT) sand has been beached from north to south, progressively displacing maturefine tailings (MFT) to the south end of the deposit for pumping to other ponds. Infilling wascompleted in late 2009, with landform construction and capping progressing into 2010.

The Pond 1 reclamation landscape will feature a small shallow wetland in the southwesterncorner of the site connected to a series of swales intended to collect surface water from throughoutthe pond and transport the flows to the wetland and eventually away from the area. Hummocks andmounds at various scales will promote surface water drainage between the swales as well as addlandscape diversity to the beach area. These features are shown in Figure 1, with a reclamation caparea of approximately 244 hectares. Reclamation material will be placed over the site to promoteplant growth and establish a boreal forest environment and also provides the primary barrier againstinfiltration of water from the surface. The cap cross-section consists of (from top to bottom), a500-mm thick vegetated cover soil layer, a 610-mm coarse sand layer, a low-permeability layer,and a prepared sand subgrade (Figure 2). The intent of the low-permeability layer was to limitsurface water percolation into the capped sands until the vegetation and reclamation soil layersare sufficiently established. Initially, the design called for the low-permeability layer to be a thickcompacted soil liner, with a hydraulic conductivity of 1 × 10−7 cm/sec. However, a geosyntheticclay liner (GCL) was selected as the hydraulic barrier layer in the Pond 1 reclamation cover in lieuof the compacted clay liner. A technical comparison of two types of hydraulic barrier layers, andthe rationale for selecting a GCL, is presented in the following sections.

401

Figure 1. Site map.

Figure 2. Typical cap cross-section.

2 COMPACTED CLAY LINERS

Compacted soil liners have been traditionally used as barrier layers in landfill and mine closures tolimit the infiltration of surface water into the buried waste. Soil liners are often selected because anadequate borrow source is located nearby. Assuming a clay-rich soil, a low hydraulic conductivityliner can be achieved if the soil is compacted within a specific range of water contents and dryunit weights. Compacted clay liners are generally thick, usually between 0.6 to 0.9-meters thick,and cannot be accidentally punctured, like thinner geosynthetics can. However, compacted clay

402

liners can be difficult to construct and are subject to deterioration from various factors, includingdifferential settlement, desiccation, and freeze-thaw action (Koerner and Daniel, 1993).

A factor of particular importance in cold weather regions like Northern Alberta is resistanceto freeze-thaw cycles. Numerous studies have shown that freezing of compacted clay liners canproduce significant increases in hydraulic conductivity [including, Erickson et al. (1994), Bensonand Othman (1993), and Kraus and Benson (1994)]. Kraus and Benson performed hydraulicconductivity tests on specimens obtained from compacted clay liner test plots before and aftera winter season. Results indicated that the compacted clay liners had an increase in hydraulicconductivity of several orders of magnitude, from approximately 1 × 10−8 cm/sec to greater than1 × 10−5 cm/sec. Extensive crack networks were present in the after-winter specimens. Thesecracks serve as preferential flow paths and are the primary cause of the high measured hydraulicconductivities. During freezing, cracks form in the clay due to the formation of ice lenses. As thetemperature increases, the ice thaws, and voids in the clay are left behind, allowing preferentialpathways for flow. Similarly, Benson and Othman (1993) found that clay hydraulic conductivityincreased by 2 orders of magnitude after three to five freeze-thaw cycles. Erickson et al. (1994)saw increases in CCL field test pads of up to 4 orders of magnitude.

Even without such long-term environmental factors, soil liners are susceptible to erratic fieldperformance due to the many hard-to-control variables involved in their construction. Factorssuch as borrow source characteristics (e.g., clay variability, clay clods, or excessive amounts ofgravel), moisture content, compaction equipment/procedures, inter-lift bonding, and slopes, causepractical difficulties which result in fluctuating permeability values in the field. For example,Rogowski (1990) constructed a homogeneous compacted clay liner over a small, 0.05-acre area,using specifications commonly used in constructing liners. Based on leakage rate measurementsthrough the clay, the actual hydraulic conductivity of the liner varied by 4 orders of magnitudethroughout the test area.

3 GEOSYNTHETIC CLAY LINERS

Geosynthetic clay liners (GCLs) are bentonite clay-based liners that often used as a substitutefor compacted clay in solid waste and mining applications. Koerner and Daniel (1993) evaluatedthe differences between GCLs and compacted clay in terms of three technical issues: Hydraulic;Physical/Mechanical; and Construction. They determined that GCLs are equivalent, or superiorto, compacted clays in regards to hydraulic issues and physical/mechanical issues. In regardsto construction issues, the authors determined that, because they are thinner, GCLs are moresusceptible to puncture damage and lateral squeezing/thinning during construction than compactedclay. However, they also noted that this type of installation damage can be limited with soundconstruction practices.

A plastic-laminated GCL was selected as the hydraulic barrier layer in the Pond 1 reclama-tion cover. The GCL consists of 3.6 kg/m2 (0.75 lbs/ft2) of sodium bentonite clay, needlepunchedbetween woven and nonwoven geotextiles. A 0.1-mm (4-mil) HDPE plastic geofilm was laminatedto the nonwoven geotextile for improved hydraulic performance. Overall, the GCL was selectedfor several reasons, including: low hydraulic conductivity (5 × 10−10 cm/sec), lack of nearby clayborrow sources, material consistency and quality control, simplicity and speed of installation, andability to install in extreme cold temperatures. These considerations are discussed in more detailbelow.

3.1 Hydraulic performance

The theoretical hydraulic performance (i.e., leakage) of either a compacted clay liner or a GCL canbe estimated using Darcy’s Law, which states that the flow through a porous medium is proportionalto the hydraulic head and the hydraulic conductivity. Figure 3 presents the results of these theoreticalcalculations. The compacted clay liner was assumed to have a thickness of 0.6-m and a hydraulicconductivity of 1 × 10−7 cm/sec. The plastic-laminated GCL was assumed to have a thickness of1 cm and a hydraulic conductivity of 5 × 10−10 cm/sec. A comparison of the graphs in Figure 3shows that a GCL is the superior hydraulic barrier, expected to only allow a small fraction ofthe percolation expected through a conventional compacted soil cover. Although these are only

403

Figure 3. Theoretical percolation estimates through selected cap alternatives.

theoretical calculations, based on several assumptions, the calculated values do seem consistentwith measured field data, as discussed below.

3.2 GCL field performance at past similar sites

Benson et al. (2007) present a case history describing the hydraulic performance of a final coverfor a coal ash landfill, where the barrier layer consisted of a composite GCL in lieu of a compactedclay layer. The site, which is located in southwestern Wisconsin, receives 892 mm of precipitationper year. The composite GCL installed at the Wisconsin site is very similar to the material usedto cap Suncor Pond 1. The GCL contained 3.6 kg/m2 of granular sodium bentonite, was encasedbetween nonwoven and woven geotextiles, and was laminated with a 0.1-mm thick polyethylenegeofilm. The cover profile consists of a 760-mm-thick vegetated surface layer (silty sand), theGCL, and a 150-mm-thick layer of interim cover soil (silty sand) placed over the ash.

Two 4.3 × 4.9 m pan lysimeters were installed beneath the cover to monitor the percolation rate(discharge from the base of the cover). The lysimeters were filled with pea gravel and drainedto a still well, which was periodically pumped to determine the volume of water collected by thelysimeter. Two separate plots were constructed: the first had the laminated GCL installed withthe geofilm downward; the second had the geofilm oriented upward. Percolation rates remainedlow in both lysimeters. Over a five-year period, the average measured percolation rates for thetwo lysimeters were 2.6 mm/yr and 4.1 mm/yr. These percolation rates represent less than 0.5%percent of precipitation, indicating that the GCL is serving as a very effective hydraulic barrier atthe Wisconsin landfill site.

3.3 GCL cold weather performance

Northern Alberta has long, very cold winters, with an average winter temperature of −19◦C, and arecord low temperature of −50◦C. Clearly, the ability to withstand and perform in cold weather wereimportant considerations when selecting the low-permeability liner for Pond 1. Kraus et al. (1997)found that after being frozen and thawed 20 times, GCLs maintained a low hydraulic conductivity.A similar evaluation of freeze-thaw resistance of GCLs was performed in 2006 by the Idaho National

404

Table 1. Cold-weather testing results.

Room temperature Cold weatherProperty (22◦C) (−40◦C)

Puncture resistance 552 N 640 N(ASTM D4833) (σ = 35 N) (σ = 18 N)Tensile strength 480 N 561 N(ASTM D6768) (σ = 45 N) (σ = 9 N)Brittleness temperature by Impact – Passed(ASTM D1790) (no signs of cracking)Low temperature flexibility – Passed(ASTM D1970) (no signs of cracking)

Engineering Environmental Laboratory (INEEL). GCL samples were exposed to repeated freeze-thaw cycles in the laboratory at pressures encompassing final cover (20 kPa) and bottom liner(60 kPa) applications. Samples were tested in the laboratory after 3, 9, 15, 21, 30, 45, 75, 100, 125,and 150 freeze-thaw cycles. Hydraulic conductivity testing found no appreciable changes, evenafter 150 freeze-thaw cycles. Examination of the GCLs while frozen and after thawing reveal thatice segregation does occur in GCLs, but the cracks formed during ice segregation close when thebentonite thaws because the thawed bentonite is very soft and compressible.

Additionally, to determine whether the plastic-laminated GCL could be handled and installedin extreme cold temperatures without undue risk of damage, Precision Geosynthetic Laboratorieswas contracted to perform a low-temperature testing program on the material in their Anaheim,California laboratory. Specifically, the following tests were performed:

• ASTM D6768, Tensile strength. Tests performed at both room temperature and after coolingto −40◦C.

• ASTM D4833, Puncture strength. Tests performed at both room temperature and after coolingto −40◦C.

• ASTM D1790, Brittleness temperature of plastic sheeting by impact. Samples were first cooledto −40◦C, wrapped in a closed loop on an anvil, struck with a 6-lb swinging arm, and thenchecked for signs of cracking.

• ASTM D1970, Low-temperature flexibility. Samples were cooled to −40◦C, bent repeatedlyaround a 1-inch mandrel, and then check for signs of cracking.

The laboratory test results in Table 1 show that samples of the plastic-laminated GCL cooledto subzero temperatures (40◦C) did not experience any reduction in tensile or puncture strength,and passed both the ASTM low-temperature brittleness and flexibility tests. These data, togetherwith the freeze/thaw data cited in the literature, indicate that GCLs can withstand the rigors ofcold weather installation. Installation of a compacted clay liner or a polyethylene geomembranewould not likely have been possible for temperatures less than 0◦C. The use of a GCL thereforeallowed construction to proceed through the cold winter months, reducing the overall constructionschedule, as discussed further below.

4 OTHER PRACTICAL CONSIDERATIONS

In addition to the technical considerations outlined above, the GCL also offered several practicalbenefits:

• Freight. GCLs are packaged and delivered in rolls. One truckload of GCL rolls can coverapproximately 3,350 square meters (36,000 square feet). By comparison, over 200 truckloads ofclay, hauled from a borrow source 10 km away, would have been required to cover the same areawith a 0.6-meter thick compacted clay liner. Over the entire Pond 1 cap area, this represents adramatic difference in the number of trucks; tens of thousands of truckloads of clay compared

405

Photograph 1. GCL installation.

to only a few hundred truckloads of GCL. In addition to the obvious safety and logistical issuesinvolved with routing such a large number of trucks to the job site, time was also a factor: Themajority of the GCL was delivered within a 1-month timeframe, which would not have beenpossible with clay.

• Speed/Ease of Installation. Since GCLs are supplied in large rolls that are simply unrolled into4.57-m × 45.7-m rectangular panels in the field, they are more straightforward to install com-pared to compacted clay liners, which require careful moisture conditioning and compaction.GCLs are seamed by overlapping adjacent panels 0.15- to 0.3-m and applying supplementalgranular bentonite (0.4 kg/m) to the overlap area. In addition, a pneumatically-powered geosyn-thetic installation device was used to deploy the GCL at Suncor Pond 1. The installation devicewas mounted on a large-capacity tractor, as shown in Photograph 1. As the tractor operator droveforward, a ground operator used a control cable to unroll the GCL onto flat panels. Using thisequipment, more than a hectare per day of GCL was safely deployed.

• Quality Control. Since GCLs are manufactured under controlled factory settings, they are muchmore consistent materials requiring less on-site quality control testing compared to compactedclay liners. Manufacturing quality control testing is performed at the plant in accordance withASTM D5889, Quality Control of GCLs. In contrast, compacted clay liners have high inherentvariability, requiring numerous and frequent field tests, including Atterberg limits and soilparticle size (1 every 800 m3), water content and density (13 tests per hectare per lift), andhydraulic conductivity (3 tests per hectare per lift).

• Schedule. A major project driver was schedule. As indicated previously, at low ambient tem-peratures, it would have been possible to construct a compacted clay liner. Cold weather delaysrelated to compacted clay would have pushed the project completion date well past 2010. Theuse of a GCL allowed construction to proceed through the cold winter months, reducing theoverall construction schedule.

5 SUMMARY

Geosynthetics were critical in successful design and completion of the Suncor Pond 1 reclamationproject. A GCL was selected as the low-permeability hydraulic barrier in the cover system forseveral reasons, including: low hydraulic conductivity (5 × 10−10 cm/sec), lack of nearby clay

406

borrow sources, material consistency and quality control, simplicity and speed of installation, andability to install in extreme cold temperatures. Construction of the geosynthetic capping systembegan in late 2009, with capping operations to be completed by September 2010. Because of thetight deadline, much of the construction activities took place during the cold winter months inlate-2009 and early-2010. This would not have been possible if a traditional compacted clay linerwas used.

REFERENCES

Benson, C.H., and Othman, M.A. (1993). “Hydraulic Conductivity of Compacted Clay Frozen and ThawedIn-situ,” Journal of Geotechnical Engineering, ASCE, Vol. 119, No. 2, pp. 276–294.

Benson, C., Thorstad, P., Jo, H., and Rock, S. (2007), Hydraulic performance of geosynthetic clay liners in alandfill final cover. J. Geotech. Geoenviron. Eng., 133(7), 814–827.

Erickson, A.E., Chamberlain, E.J., and Benson, C.H. (1994). “Effects of Frost Action on Covers and LinersConstructed in Cold Environments,” Proceedings of the 17th International Madison Conference, Madison,WI, pp. 198–220.

Koerner, R.M., and Daniel, D.E. (1993). “Technical Equivalency Assessment of GCLs to CCLs,” Proceedingsof the 7th GRI Seminar, Philadelphia, PA, pp. 255–275.

Kraus, J.F., and Benson, C.H. (1994). “Effect of Freeze-Thaw on the Hydraulic Conductivity of BarrierMaterials: Laboratory and Field Evaluation,” Environmental Geotechnics Report 94-5, University ofWisconsin-Madison.

Kraus, J.F., Benson, C.H., Erickson, A.E., and Chamberlain, E.J. (1997). “Freeze-Thaw Cycling and HydraulicConductivity of Bentonitic Barriers,” ASCE Journal of Geotechnical and Geoenvironmental Engineering,Vol. 123, No. 3, pp. 229–238.

Podgorney, R. K., and Bennett, J. E. (2006) “Evaluating the Long-Term Performance of Geosynthetic ClayLiners Exposed to Freeze-Thaw,” Journal of Geotechnical and Geoenvironmental Engineering, Vol. 132,No. 2.

Rogowski, A.S. (1990). “Relationship of Laboratory- and Field-Determined Hydraulic Conductivity in Com-pacted Clay Layers,” EPA/600/S2-90/025, USEPA Risk Reduction Engineering Laboratory, Cincinnati,OH.

407


A new approach to oil sand tailings management

L. Lawrence & Z. AliSuncor Energy Inc. Fort McMurray, AB, Canada

ABSTRACT: This paper describes Suncor’s operational scheme change from a ConsolidatedTailings (CT) based operation to a Tailings Reduction Operation (TRO) which incorporates MFTdrying as a fines capture technology. The risks associated with CT operations and the benefits ofthis new approach at Suncor will also be discussed. By adopting this new tailings managementstrategy Suncor believes it will not only address the challenges of a CT operation but will alsosatisfy the objectives of Directive 074 (D074) which was approved by the Energy Resource andConservation Board (ERCB) of Alberta on February 3rd, 2009. The success of the TRO conceptwill prove beneficial for large scale, long term economic investment in the oil sands industry byreducing the impact on the environment associated with oil sand mining operations.

1 INTRODUCTION

One of the primary challenges facing the oil sands industry is the production of Mature FineTailings, commonly known as MFT. This is one of the waste products resulting from the bitumenextraction process. MFT is a fluid tailings which forms when tailings from the primary extractionprocess is discharged into tailings ponds. The coarse sand settles out while the clay particles orfines remain suspended in water to form MFT comprising 30% solids and 70% water by weight.If left untreated, the continuous production of oil sand will result in an increasing MFT inventoryand therefore an increasing need for additional tailings containment (i.e. additional tailings ponds)through the life cycle of an oil sand operation. This in turn results in an expanding environmentalfootprint through increasing land disturbance and numerous tailings ponds which are costly toreclaim within the timeframe of the operation’s life cycle.

Recognizing this risk, the Energy Resource Conservation Board (ERCB) of Alberta issuedDirective 074 which set out the new requirements for the regulation of tailings operations with thefollowing objectives:

• To minimize and eventually eliminate long-term storage of fluid tailings in the reclamationlandscape;

• To create a trafficable landscape at the earliest opportunity to facilitate progressive reclamation;• To eliminate or reduce containment of fluid tailings in an external tailings disposal area during

operations;• To reduce stored process-affected waste water volumes on site;• To maximize intermediate process water recycling to increase energy efficiency and reduce fresh

water import;• To minimize resource sterilization associated with tailings ponds; and• To ensure that the liability for tailings is managed through reclamation of tailings ponds.

In 1995 Suncor identified MFT management as a strategic priority and addressed the issue ofan increasing MFT inventory by implementing the production of Consolidated Tailings (CT) asa fines capture technology. CT is formed by combining coarse sand with MFT recovered froma settling pond at a defined sand to fines ratio (SFR) together with a small amount of gypsum.The resulting mixture, which is hydro-transported to a CT disposal pond, forms a non-segregatingdeposit. Although CT has been relatively successful in capturing fines to date, Suncor has decidedto convert from a CT based operation to a Tailings Reduction Operation.

409

This paper discusses the long term tailings management risk associated with CT and the benefitsof TRO at Suncor’s oil sand operation. It is believed the concept of TRO will increase the environ-mental and economic sustainability of Suncor and the oil sands industry by minimizing the impacton the environment.

2 THE IMPACT OF CT ON SUNCOR’S OIL SAND OPERATION

The application of CT to manage the MFT inventory at Suncor offered the following benefits:

• The CT process liberates water bound in the MFT as CT release water. The return of this water tofree water inventory results in less water demand (i.e. less water withdrawal from the AthabascaRiver, per unit of ore mined).

• MFT incorporated into CT occupies less volume than in its previous state (because of theliberation of bound water) and thereby adds to the tailings storage capacity.

However, CT also presented significant risk to Suncor’s long term tailings management plan interms of cost and the availability of overburden dyke construction material. Since CT relies ontailings sand for fines capture, the construction of overburden dyke structures was a necessityto meet the long term demand for fluid tailings containment. Not all overburden material canbe utilized as dyke construction material and therefore Suncor developed an overburden materialclassification system which categorizes the various types of overburden or waste material accordingto geological horizon and compaction specification. The overburden material types are defined inthe table below:

OverburdenMaterial Dyke and StructureClassification Material Characteristics Material Location Construction Suitability

G-spec Low shear strength, high moisture Muskeg and upland Not suitable: Requirescontent Pleistocene silts, sands soils between 15% and 20%and clayey tills K or B spec padding for

equipment accessibility

K1-spec Up to 50% high plastic Clearwater Top of the Clearwater Suitable except in cases ofor Clearwater derived till material, formation down to high moisture contentblended with low plasticity the top of the Uppermaterials (tills and lean oil sands) McMurray

K2-spec Up to 100% plastic Clearwaterderived till material

B1-spec Blend of low to medium plastic clay Top of the Upper Suitabletills, lean oil sands and gravels, McMurray to the topless than 30% Kcw and less than of mineable ore10% high plastic Clearwaterof Clearwater derived till material

B2-spec Blend of low to medium plasticclay tills, lean oil sands andhigh fines sandy tills, less than30% Kcw and less than 10%high plastic Clearwater ofClearwater derived till material

Figure 1 below shows the percentage of overburden material type available on average withinthe Millennium pit footprint at Suncor’s current oil sand operation.

As can be seen, a large portion of the overburden is classified as G-spec material which isunsuitable for dyke construction. Therefore in order to expose sufficient B and K spec material

410

Figure 1. Average overburden waste percentage.

required for dyke construction to form an in-pit CT pond that satisfied Suncor’s tailings containmentrequirements, significant volumes of G-spec material had to be mined in advance. This resulted inlong hauls of the G-spec material to large external waste dumps to maximize the tailings containmentof the in pit pond thereby significantly increasing operating costs.

In addition to the high cost of external waste dumping Suncor had to contend with the issue ofmaterial utilization which is defined as follows:

MATERIAL UTILIZATION = B + K PLACED

B + K + G MINED

Material utilization is the percentage of overburden that is placed in B-spec or K-spec zonesof tailings dyke structures and in order for Suncor to meet its tailings containment requirements,a minimum material utilization target of 35% was required to be achieved. However as the minedeveloped and more overburden information became available through operational experience andgeological modeling, a shortfall of B and K spec material available for dyke construction wasidentified. This was due to:

• Lower than expected K-spec capture resulting from poor shovel footing;• High inherent moisture content resulting in non-compliant dyke-spec material,• Higher than planned weather events resulting in placement delays,• Higher than planned volumes required for other requirements (i.e. dump padding, in-pit roads

and ramps)

In light of this shortfall, material utilization had to be reduced to 25% by reducing the height ofone of its in-pit CT tailings pond from 370 to 350 masl. The 20 m reduction in dyke elevationsignificantly reduced the volume of dyke-spec overburden required to construct the required dykesfor containment which brought the dyke-spec demand in line with availability. This however,resulted in a smaller tailings pond and insufficient tailings containment to meet Suncor tailingscontainment requirements in the long term.

Suncor therefore decided to adopt a new tailings management strategy to address the risk andincreased cost associated with its CT operation by converting to TRO which incorporates MFTDrying (MFTD) as a fines capture technology. This conversion also afforded Suncor the opportunityto adopt a tailings management strategy in line with the objectives as set out in D074.

3 TRO AT SUNCOR’S OIL SAND OPERATION

UnderTRO, extraction tailings will be deposited as segregating tailings (also called RegularTailingsor Conventional Tailings) into in-pit beaching areas initially enclosed by overburden starter berms.As the regular tailings is continuously discharged the coarse sand settles out to form beaches whichis sequentially raised over time through the step-over cell construction technique to form a solidsand structure or sand dump. The remainder of the fines will segregate to form Thin Fine Tailings orTFT (5% solids by weight) which will be transferred to a settling pond for maturation to form MFT

411

Figure 2. TRO material flow diagram.

(30% solids by weight). After maturation, MFT will be subjected to the MFTD process which isoperated independently of ongoing beaching operations and is sized to consume all of the ongoingMFT generation as well as some legacy MFT. In the MFTD process, MFT is mixed with a polymerflocculent and then deposited in thin layers over beaches or other suitable areas. These designatedareas are called Dedicated DisposalAreas or DDA’s. The deposit initially dewaters to approximately60% solids by weight through flocculation which then further dries through evaporation to form adeposit which is approximately 80% solids by weight. The dried MFT can either be reclaimed inplace or rehandled to waste dumps as defined by the Mine and Tailings Plan. Figure 2 below showsa basic illustration of the TRO process.

The MFTD process is considered to be a technological breakthrough that will enable Suncor todirectly and proactively address one of the biggest challenges of the oil sands industry which is theliability of increasing MFT inventories. MFTD results in the treatment of fines in DDA’s eliminatingthe risk of fluid tailings in the long term associated with oil sands operations. The benefit of DDA’sare that they can be located in areas already disturbed by mining and tailings operations whichminimizes the impact on the environment. Figure 3 illustrates current and planned DDA’s (yellowareas) part of Suncor’s current TRO Mine and Tailings plan. Areas considered suitable for MFTDoperations include:

• Exposed beaches on existing tailings ponds (Ponds 5, 6, 8a and STP)• Cleared areas within the mining footprint ahead of mining (DDA1), and

• Future areas on top of Sand dumps constructed to full height (Sand Dump 8 and Horseshoe)

In order for Suncor to adopt this new tailings management approach, the tailings infrastructurewill be modified from the current CT operation to a Tailings Reduction Operation as shown inFigure 4. Regular Tailings will be deposited in pit behind the mining operation from two Extrac-tion plants to construct the first sand dump (Sand Dump 8). The TFT will be transferred to STPwhich will serve as a settling pond where the TFT will mature to form MFT. Water will then betransferred to Pond 7 where it will be distributed to Upgrading for cooling and then distributedto Extraction as part of the Hot Water Process. Once Sand Dump 8 (SD8) is complete new sand

412

Figure 3. Planned and established DDA’s at Suncor’ oil sand operation.

dumps (SD9 and SD10) will continuously be constructed behind the advancing mining operationinfilling the mined out void. Although the cost of implementing this change is significant, Suncorbelieves the benefits of TRO far outweigh the cost. The adoption of this approach will be extremelybeneficial to Suncor in terms of its commitment to sustainable development which is one ofits core values.

4 THE BENEFITS OF TRO

The fundamental difference between TRO and CT is that TRO largely eliminates the need for largeoverburden dyke structures for containment and enables the independent management of sand andfines through:

1. The construction establishment of DDA’s which allows for the direct treatment of MFT fluidtailings into reclaimable deposits

2. The construction of Sand Dump structures which eliminates the need for large overburden dykestructures

413

Figure 4. TRO infrastructure layout.

This in turn decouples the mining operation from the tailings operation resulting in tailingscontainment no longer being dependent on the availability of overburden which eliminates the riskof material utilization and expensive pre-stripping associated with CT operations at Suncor.

A further benefit of TRO is that Sand Dumps and DDA’s do not require consolidation time asis the case with CT ponds and can easily be incorporated into the reclamation landscape therebyfacilitating progressive reclamation. Figure 5 clearly shows the relative difference in the timerequired to reclaim the landscape affected by surface mining between a TRO and CT operation. Inthe case of TRO, the operation’s life cycle can be reduced from approximately 30 years to 10 years,from the time mining commences to the time that a landscape is available for replanting and fullreclamation, by constructing sand dumps and DDA’s as opposed to tailings ponds.

414

Figure 5. CT vs. TRO life cycle.

5 CONCLUSION

The implementation of TRO will reduce Suncor’s tailings management risk profile while affordingthe opportunity to adopt a tailings management strategy in line with the objectives of D074 withthe following benefits:

• It reduces the need for long term storage of fluid tailings• It reduces the need for external fluid containment

• It enables the progressive reclamation of the disturbed landscape rapidly increasing the paceof reclamation

It will ultimately lead to the long term economic and environmental sustainability of Suncor andthe oil sands industry making the previously perceived “dirty” oil a lot more attractive to the public,environmental groups, potential investors and international stakeholders.

REFERENCE

ERCB, (2009). Directive 074 –Tailings Performance Criteria and Requirements for Oil Sands Mining Schemes.

415

Environmental issues


A landscape design approach for the sustainable reclamation activitiesof a post-mining area in Cartagena, SE Spain

Sebla Kabas, Ángel Faz, Dora M. Carmona, Silvia Martínez-Martínez & Raúl ZornozaTechnical University of Cartagena, Murcia, Spain

Jose A. AcostaTechnical University of Cartagena, Murcia, SpainUniversity of Amsterdam, The Netherlands

ABSTRACT: Utilization of Cartagena-La Unión Mining District (SE Spain) for more than 2500years created deleterious effects on the environment and human health because of its chemical andphysical characteristics. These effects can be ameliorated by the help of sustainable reclamationactivities which can be seen as a required solution in order to return the mining area to an acceptableenvironmental condition. Sustainable reclamation activities, which require an extensive ecologicalknowledge, have to consider landscape values and functionality of the area, with the attention tothe historical, socioeconomic and cultural aspects.

1 INTRODUCTION

Throughout the world, mining has created landscapes of both wasteland and wonder, profoundlyaltering the earth’s surface. The necessity to reclaim these post-mined landscapes raises importantissues of ecology, society, and aesthetics. Planning for ecological processes or new land uses wereabsent in the process of mining reclamation, but today, however, with new regulations and height-ened environmental concern for post-industrial and mined lands, designers are beginning to playa vital role in the design of reclaimed mines. Consequently, new challenges for both scientists anddesigners include rethinking reclamation strategies across disciplines, and designing for evolvingsocial and ecological landscapes that confront our perception of mining and its reclamation (Fisher,2006).

Cartagena-La Unión Mining District is one of the post-mining areas in Iberian Peninsula in whichintense metal mining activity had been under operation for more than 2500 years. The activitieswere ended in 1991. The area is located in Murcia Province, SE Spain (Fig. 1) and covers 50 km2,including five population nuclei with around 20,000 total populations. The surroundings of themining area have a population of more than 200,000 inhabitants, including the city of Cartagenaand the rest of its Municipality. During the summer, the population dramatically increases becauseof tourists who come to the Mar Menor Lagoon and Mediterranean Sea beaches. In the north, agri-cultural areas; in the west, petrol refinery; in the south, Mediterranean Sea; and in the north-east,one of the most important lagoon of Mediterranean Area, Mar Menor, is located. The altitudesof the area occur between 0 and 500 m above sea level. The climate is semiarid Mediterranean,long warm summers and short moderate winters with mean annual temperature of 18◦C and meanannual rainfall of 275 mm.

Formation of the area and regional climate conditions determine a natural flora adapted todrought and high temperatures which are only found only in hills and small sites that contain smallformations of pine trees (Pinus halepensis) and groups of typically Mediterranean brushes withxerophytic characteristics. Some of the thicket plant species are endemic in this zone and thereforehave a high botanic interest (Conesa et al. 2008).

The mines constituted the only economic activity during hundred of years for local people.Mining-dependence caused a lot of population fluctuations with a decline from 30,000 citizens in1900 to 13.900 in 1991 when the mining activity ceased.

419

Figure 1. Cartagena-La Unión Mining District (Conesa, 2008).

But the environmental impacts of the long-history of mining activities in southeast Spain includelarge areas of soils characterized by strong acidification processes, high salinity and accumulationof metals. These mining activities have generated high amounts of sterile materials for many years;the wastes are accumulated in pyramidal structures called tailing ponds. Throughout the area thereare 85 tailing ponds which contain materials of high Fe-oxyhydroxides, sulphates, and potentiallyleachable elevated contents of heavy metals (mainly Cd, Pb, Cu and Zn) due to extreme acidicconditions.

As a consequence, these mine soils have scarce or null vegetation due to very poor propertiesof soil such as extremely low organic matter, and most of the natural vegetation formations aredegraded due to the atrophic use of the soil. While unvegetated tailings have been exposed to eoliandispersion in the semiarid climate conditions of the region, on the other hand for a long time, thesemine residues have been transported downstream during periods of high rainfall and runoff; as aconsequence the pollutants have migrated long distances.

It is also possible to see some deleterious examples related to the public safety. For example inthe year 1972, the mine tailing Brunita collapsed due to the strong rainfalls, spreading the materialsin nearby areas and killing one person (Orozco et al. 1993). According to Ortega (1993) some ofthe mine tailings are more than 20 m in height and cover an average area of 40,000 to 80,000 m2,are unstable and difficult to eliminate. Close to two children school, one acid mine tailing wasreported with high concentrations of zinc, lead and arsenic (Conesa et al. 2008). Because of thewaste from mining operations was discharged directly into the inner part of the bay for more than30 years, polluting the Mediterranean Sea for a radius of several kilometers, Portman Bay wasreported the most contaminated bay in the entire Mediterranean by Martinez-Frias (1997).

Remediation of these areas is required as these sites present an environmental hazard, especiallyfor agricultural and fisheries in the surrounding area. Recently the use of plants for removing heavymetals from soil has received increased attention which is generally called phytoremediation (Garcíaet al. 2003). But some soils are so heavily contaminated, like in Cartagena-La Unión, that removal ofmetals using plants would take an unrealistic amount of time, so that an alternative phytoremediationtechnique phytostabilization has been emerged. Phytostabilization is the use of metal tolerant plantspecies to immobilize heavy metals through absorption and accumulation by roots, adsorptiononto roots, or precipitation within the rhizosphere, creates a new ecological condition in the area,

420

prevents erosion and spreading of pollution, reduces metal mobility and bioavailability for entryinto the food chain. Using metal-tolerant species for stabilizing mine spoils also could provideimproved conditions for natural attenuation (Wong, 2003). Phytostabilization needs to use the nativeplant species in which the mine tailings are found, appropriate to the harsh climate of semiaridenvironments and native plants also avoid introduction of nonnative and potentially invasive speciesthat may result in decresing regional plant diversity (Mendez & Maier, 2008). Utilization of naturalplant species, especially in post-mining areas, is also necessary for landscape design for creatinga self-sustainable system. There is a growing evidence that phytostabilization can facilitate therestoration of mining degraded land (Li, 2006). Several phytostabilization experiments held in thearea showed successful results such as recuperation of some soil properties. In Zanuzzi et al. (2009),Carmona et al. (2010), it is possible to see the information related to phytostabilization trials inthe area and their results. Though the objective of the phytostabilization studies on the area mostlyrelies on the achievement of recuperation of soil properties, the final object is to create a newland use. By considering this in our aspect we tried to combine landscape design with solutions orsolution processes of environmental risks. Based on this knowledge; our purpose is to try to providea general approach related to the integration of landscape design and reclamation techniques.

2 MINE RECLAMATION & SUSTAINABILITY

Within the mining concept, reclamation refers to the general process whereby the mined wastelandis returned to some forms of beneficial use. Other similar concepts, which are used as synonymsof reclamation, are restoration and rehabilitation; in the abstract have different meanings.

While restoration refers to reinstatement of the pre-mining ecosystem in all its structural andfunctional aspects; rehabilitation means the progression towards the reinstatement of the originalecosystem. In fact; restoration and rehabilitation are impossible to realize in abandoned miningareas, because turning back to their historical trajectory, especially in an area such as Cartagena-LaUnión Mining District, very old and full of human prints, can only be an imaginary way of thinking(Mchaina, 2001; Fisher, 2006). According to Schulz & Wiegleb (2000), a pre-mining situationdoes not exist and even if it did, it could never be restored. However reclamation does not attemptto restore the landscape to its historical trajectory; rather it converts the land to a new land-useand creates new ecological conditions. So that the concept of “reclamation” combines all measuresneeded to make surface-mined landscapes productive and visually attractive again (Mchaina, 2001;Fisher, 2006).

Mine reclamation has to be formed an integral part of the planning process which includesfeasibility studies and environmental impact assessment for new mines. Following objectives haveto be provided by the help of reclamation (Mchaina, 2001):

1. To allow a productive and sustainable post-mining use of the site which is acceptable to allstakeholders

2. To protect public health and safety3. To alleviate or eliminate environmental damage and as a result encourage environmental

sustainability4. To conserve valuable attributes, and5. To minimize adverse socio-economic impacts.

Accomplishment of these objectives by mine reclamation is directly related to the sustainablecharacter of mining reclamation activities.

The definition of sustainability, approved by FAO in 1998, is the handling and conservation ofnatural resources and the orientation of technological and institutional change so as to ensure thecontinuous satisfaction of human needs for present and future generations (Leitão & Ahern, 2002).

IUCN (1992) states; sustainable development is widely accepted as a strategic framework fordecisions on the future use of land. The principles of sustainable development imply that in develo-ping land, ecological, social and economical functions are balanced in space and time to maintaintheir potential to deliver goods and services to future generations (Termorshuizen et al. 2007).

In this context, decision making is, in the first place, on attributing targets for nature conservation,quality of life or economic welfare to the landscape region. Secondly, it includes the assessment of

421

ecological, social and economic values and their interactions. Thirdly, decisions are made on theallocation of land use functions (Termorshuizen et al. 2007).

3 LANDSCAPE VALUE & FUNCTION

To evaluate a specific form of resource management, for example, the maintenance of valuablecultural landscapes, the contributions of many different viewpoints related to history, inhabitants,legislations and etc. must be considered (Gómez-Sal et al. 2003). Landscape reclamation and designis also the subject of resource management and requires the consideration of landscape values suchas natural value, historical value, productive value, socio-economic value.

The natural components of the landscape form the basis of all resources and ecological function-ing of the landscape. Increasing fragmentation and losing their connectivity causes malfunctioningand, consequently, restructuring of the environment. Many natural values remain as isolated relictslost in the superimposed landscape structured by man in a different way. For example, industrialchanges deform the structures, and thus their functioning, of the existing landscapes. In someplaces, the landscape was even wiped away entirely to create a completely new landscape. Eco-nomical rationalization controls it and all regional diversity and the identity of landscapes becomeunrecognizable. The spirit or the character of the place is lost. Conservation of biological andgeological/geomorphologic remains is a first value to protect. This can be achieved by creatingbuffer zones and connected isolated units by corridors in order to keep their functioning going. Land-scape reclamation and creation are, therefore, additional instruments (Antrop, 2000). Sklenickaet al. (2004) also supports the value of natural components by asserting biological diversity (atall levels) and speed (success) of revitalization are the two key criteria for post-mining landscapereclamation.

Many abandoned mines, despite their altered natural characteristics, have significant historicalvalue and should be protected from destruction, vandalism, and theft. The U.S. National HistoricPreservation Act established the National Register of Historic Places (NRHP) as a federal listing ofcultural resources worthy of preservation. The NHRP is maintained by the National Park Service,and to be eligible for listing, abandoned mine lands must be demonstrated to have significanceto history, architecture, engineering, or culture. The NHRP nomination process uses additionalcriteria in order to determine the historic significance of sites, buildings, structures, and objects.Besides meeting one or more of the NHRP criteria, a mine site generally must also be at least50 years old, and have integrity of location, design, setting, materials, workmanship, feeling, andassociation in order to be eligible for inclusion. If a site has been compromised by significantalterations, it may not be eligible (CAM, 2000). According to Antrop (2000) landscapes that aremade by society reflect the changing society and attitude towards the environment. Landscapesreflect the superposition of all attempts man makes to adapt the environment to improve livingconditions. The landscape is full of past memories, which still have a strong symbolic value. Thiscan be seen clearly when they are exploited as tourist attractions.

Related to the historical-cultural value of Cartagena-La Unión Mining District several authorsasserted different ideas. Manteca & Berrocal (1997) related the mining elements with archaeo-logical and geological value in order to justify the creation of a geo-mining and archaeologicalpark in the region. According to those authors the mining heritage elements in the Mining Districthave two components regarding to their value: 1) an educational value to raise public awarenessabout cultural, natural and historic resources and to a didactic tool for researchers and 2) theirsocio-economic value, since they can constitute an important tourist attraction because of theirgeographic situation near an important tourist focus. Rodríguez-Estrella et al. (2003) proposed theuse of mine tailings in didactic works because in these structures it is possible to identify sedimen-tary structures very similar to old rocks although the age of the tailing is less than hundred years.As a consequence, these authors consider these tailings like “natural laboratories with high didacticand scientific interest”. The most important proposal, mentioned in La Verdad de Cartagena (2005)and has already been made real, is based on the creation of a thematic park about mining, that isconsidered the local mining heritage site with highest historic interest (Conesa et al. 2008).

Aesthetics and ecological sustainability are also two highly regarded landscape values (Gobster,1999). In developed societies aesthetics plays a relevant role in public acceptance of landscapeinterventions. The understanding of aesthetic aspects of landscape and clarification of its relation

422

to sustainable principles in a reclamation project presents itself as a useful tool, resolving potentialconflicts between ecology and sustainability principles and public perceptions and expectations(Rodrigues-Ramos & Panagopoulos, 2007).While the sense for beauty is universal, the expressionof beauty may differ between regions, cultures and periods. Aesthetics are also found in the waysociety organized the landscape during history. The most striking examples are found in gardening;landscapes that are considered having ‘outstanding beauty’ are appreciated, receive a special legalstatus and are sometimes protected. Cleanness and a well-maintained appearance of the landscape,the durability which is expressed in its old age (represented by monuments) and its naturalness (asa symbol for the slow evolution and growth) are positive landscape characteristics and generallyappreciated by people. For example, degraded and derelict land is without order, is not clean and notwell maintained, so it receives only a poor value and ‘attracts’ spontaneous waste dumping of anykind and, thus, reinforces the degraded character (Antrop, 2000). Positive landscape characteristicscan provide an easier adoption of the local people to the new land use. Once adoption is provided,protection and maintenance can be realized by the pleasure of local people.

High level ecosystems have a multifunctional feature which often serves more than one function.But generally, there are one or two functions outstanding. Biological production function includescultivation, forestry, stock raising, fisheries and other agricultural production-oriented uses of land.Cultural support function includes urban constructional purposes, such housing and infrastructureuses. Environmental service function includes greening and tourist areas. Though such activitiesare generally not part of pre-mining land uses, they are an important part of holistic land-use plansand serve to implement the environmental service function of the landscape system. This functionis applied to degraded areas where agriculture cannot be developed, like in our case. Green areadevelopment and rehabilitation supports the implementation of the environmental service function,whose fundamental aim is to eliminate pollution in the mining area. Owing to the special statusgiven to green areas, special technology guarantees and meticulous care are needed at each stage ofdevelopment from feasibility analysis to land levelling, species selection, planting and managing. Iftourism function can be concentrated in those areas, it can also be possible to supply local industry(Wang et al. 2001).

In order to create a successful and sustainable reclamation design it is important to recognize andinterpret the historic and cultural significance of the landscape, collaboration of different ecologicaldisciplines and to understand how “landscape ecology and design can invent alternative forms ofrelationships between people, place, and cosmos, so that landscape design projects become moreabout invention and programs rather merely corrective measures of restoration”. Without ecologicalknowledge, new sustainable landscapes cannot be achieved (Hüttl & Gerwin, 2005; Loures &Panagopoulos, 2007).

4 DESIGN STRATEGY

Derelict and degraded industrial areas can be filled with a new spirit and can be made worth living bykeeping visible the spirit of existing site, by applying design strategies that contribute to economicprosperity, social cohesion and environmental quality (Loures & Panagopoulos, 2007). Punter(2002) draw the limits with five fundamental principles that have to be integrated by landscapereclamation design: protect and conserve the quality of landscapes; develop a clear vision andstrategy for an area; apply collaborative design principles; allow resources for long-term aftercareof new landscapes; enhance biodiversity, social stability and economic development (Loures et al.2006). While these principles aim to preserve the innate medium by giving a new vision to the area,also they present the necessity of sustainability.

At this point also it is important to indicate that planning of land use cannot be restricted to thedetermination of the uses of each field or land parcel. Landscape is not something to be used onlyby the landowners, but also by temporary visitors: recreants, tourists, and neighbors. Landscapeis multifunctional and the design must be taken into account as a whole as well. Because eachelement only gets its meaning, significance or value according to its position and relationshipwith the surrounding elements. It should not be forgotten that changing one element always meanschanging the whole in some way (Antrop, 2000).

Leitão & Ahern (2002) after their comparison related to several ecology based plannings, col-lapsed the stages included by each method into five basic phases: (1) setting goals and objectives,

423

Figure 2. Scheme of the landscape design approach in problematic areas.

(2) analysis, (3) diagnosis, (4) prognosis, and (5) syntheresis, or implementation. Some includedmore recent planning methods and tools such as spatial modeling and GIS, alternative scenariosand simulation techniques, monitoring, and the participation of stakeholders and public in generalinto the planning process were found fundamental components of sustainable landscape planning.

Design strategy in our case (Fig. 2) has a similar mechanism with the ecology based planningsthat are mentioned above. But in our approach determination of objectives takes place after thestep of diagnosis instead of being in the beginning, because objectives are wanted to be establishedrelated to the characterization of the area. This consideration can be explained with an approachwhich can be applied in a situation of illness. Priorities should be directed to the recovery of theperson or living organism. On the other hand determination of the objectives without knowing thecharacterization of the area can hinder the emerging of optimal potential land use. Giving a functionto the area according to its present situation is the base idea. Determination of landscape valuesby examining landscape components is one of the step leads us to the function, but also solutionways for environmental risks can canalize us to the same point. This feature can come forward inproblematic areas, like in Cartagena-La Unión, and can be formed according to the feature of theproblem.

The problematic features of the area can be transformed to a functional stage by providing theirassociation with landscape components. With a controlled usage such as tampon zones, fencesetc., physical character of tailing ponds or color of acid mine drainage pools can be seen aestheticand can provide a touristic attraction centre in the region. Solutions for environmental risks can becombined with the main landscape value or specifically with one of the landscape components. Forexample in our case phytostabilization can be seen the way of solution for various environmentalproblems of the area, which can immobilize the heavy metals, thus prevent spreading of pollutantsto the surroundings and while the new vegetation cover creates a new ecological condition also itcan get erosion and participation of heavy metals to the food chain under control. Phytostabilizationprocess itself has a didactic value which can be utilized in a thematic park in order to improve theknowledge of visitors. Also mining elements (tools, shafts, etc.) can be represented by providingtheir safety; their educative and museum effect can be helpful in the reforming of the area. In this

424

situation didactic value can be the main value of the area which is prominent in all, or a sub-valuewhich is auxiliary for the main value.

Transformation of the ultimate decision of function to the landscape design has to be realizeddelicately by professions from different disciplines.

5 CONCLUSIONS

Cartagena-La Union Mining District because of its fragmented land form has a potential for differentkinds of land use. But related to the physical and chemical hazards in the tailing ponds whileagriculture cannot be developed, environmental service function which includes recreational andtourist areas emerges as an alternative land use for these ponds. With these alternative land usesit is possible to carry out the elimination of pollution in the area simultaneously. Regarding tophytostabilization which is planned to use in our case in order to eliminate pollution and achieveself sustainable revegetation, it is thought to make a combination between potential land values andresolution advisories for environmental risks. According to the values and solutions, recreationalarea can serve for different aims such as thematic park or geo-mining and archaeological park etc.Each tailing pond can have its own characteristics and has to be investigated in order to determinethe potential abilities.

Beside the considerations such as recovery of environmental conditions, generation economicopportunities, maintenance of historic and cultural values, development of aesthetic values; land-scape design and planning has to establish relationships between people, place, and cosmos. Studyareas have to be taken into account with their surroundings regarding to the holism principle.

Landscape design approach, which is emerged related to phytostabilization studies in Cartagena-La Union Mining District, required associations between landscape values and solutions. Byconsidering this association a general approach was tried to explain in order to light the wayfor similar cases.

ACKNOWLEDGEMENTS

This work has been funded by the Project No: CP-IP 213968-2 IRIS, funded by the European UnionFP7. R. Zornoza acknowledges a “Juan de la Cierva” contract from the Ministry of Science andInnovation of the Government of Spain. J.A. Acosta acknowledges a grant from Fundacion Senecaof Comunidad Autónoma de Murcia (Spain).

REFERENCES

Antrop, M., 2000. Background concepts for integrated landscape analysis. Agriculture, Ecosystems andEnvironment 77 (2000) 17–28.

CAM (California’s Abandoned Mines), 2000. A Report on the Magnitude and Scope of the Issue in the State,Volume I, US Department of Conservation Office of Mine Reclamation Abandoned Mine Lands Unit.

Carmona D.M., Faz, Á., Zornoza, R., Büyükkiliç, A., Kabas, S., Acosta J.A., Martínez-Martínez, S., 2010Influence of inorganic and organic amendments for mine soils reclamation on spontaneous vegetationcolonization and metal plant bioaccumulation. 19th World Congress of Soil Science, Soil Solutions for aChanging World. 1–6 August 2010, Brisbane, Australia.

Conesa, H. M., Schulin, R., Nowack, B., 2008. Mining Landscape: A Cultural Tourist Opportunity or AnEnvironmental Problem? The Study Case of the Cartagena-La Union Mining District (SE Spain). EcologicalEconomics 64 (2008) 690–700.

Fisher, T., 2006. TheArt and Science of Mining Reclamation: An IntegratedApproach to the Design of the Post-mined Landscape. Master of Landscape Architecture Department of Landscape Architecture University ofOregon.

García, G., Faz, Á., Conesa, H.M., 2003. Selection of autochthonous plant species from SE Spain for soil leadphytoremediation purposes. Water, Air and Soil Pollution: Focus 3: 243–250, Netherlands.

Gobster, P.H., 1999. An ecological aesthetic for forest landscape management. Landscape Journal, 18, 1999,pp. 54–64.

425

Gómez-Sal, A., Belmontes, J. A., Nicolau, J. M., 2003. Assessing landscape values: a proposal for amultidimensional conceptual model. Ecological Modelling 168 (2003) 319–341.

Hüttl, R. F., Gerwin, W., 2005. Landscape and ecosystem development after disturbance by mining. EcologicalEngineering 24 (2005) 1–3.

Leitão, A.B., Ahern, J., 2002. Applying landscape ecological concept and metrics in sustainable landscapeplanning. Landscape and Urban Planning 59 (2002) 65–93.

Li, M. S., 2006. Ecological restoration of mineland with particular reference to the metalliferous mine wastelandin China: A review of research and practice. Science of the Total Environment 357, 38–53.

Loures, L., Horta, D., Santos, A., Panagopoulos, T., 2006. Strategies to Reclaim Derelict Industrial Areas.WSEAS Transactions on Environment and Development. Issue 5, Vol. 2, ISSN: 1790-5079.

Loures, L., Panagopoulos, T., 2007. Sustainable reclamation of industrial areas in urban landscapes. SustainableDevelopment and Planning III. WIT Transactions on Ecology and the Environment, Vol 102, ISSN 1743-3541.

Martinez-Frias, J., 1997. Mine waste polluted Mediterranean. Nature, Vol. 388, p. 120.Mchaina, D. M., 2001. Environmental Planning Considerations for the Decommissioning, Closure and Recla-

mation of a Mine Site. International Journal of Surface Mining, Reclamation and Environment 2001, Vol.15, No. 3, pp. 163–176.

Mendez, M.O., Maier R.M., 2008. Phytostabilization of mine tailings in arid and semiarid environments –An emerging remediation technology. Environmental Health Perspectives. Volume 116, Number 3, March2008, 278–283.

Orozco, J.M.M., Huete F.V., Alonzo S.G., 1993. Environmental problems and proposals to reclaim the areasaffected by mining exploitations in the Cartagena mountains (southeast Spain) (1993) Landscape and UrbanPlanning, 23, pp. 195–207.

Rodrigues-Ramos, B., Panagopoulos, T., 2007. Integrating Aesthetic and Sustainable Principles in StreamReclamation Projects. WSEAS Transactions on Environment and Development. ISSN: 1790-5079, Issue11, Volume 3, November 2007.

Schulz, F., Wiegleb, G., 2000. Development options of natural habitats in a post-mining landscape. LandDegradation and Development, 11: 99–110 (2000).

Sklenicka, P., Prikryl, I., Svoboda, I., Lhota, T., 2004. Non-productive principles of landscape rehabilitationafter long-term opencast mining in north-west Bohemia. March 2004, pp. 83–88.

Termorshuizen, J. W., Opdam, P., Brink, A., 2007. Incorporating ecological sustainability into landscapeplanning. Landscape and Urban Planning 79, 374–384.

Wang, Y., Dawson, R., Han, D., Peng, J., Liu, Z., Ding, Y., 2001. Landscape Ecological Planning and Designof Degraded Mining Land. Land Degradation & Development. 12: 449–459.

Wong, M.H., 2003. Ecological restoration of mine degraded soils, with emphasis on metal contaminated soils.Chemosphere 50 (2003) 775–780.

Zanuzzi, A., Faz, A., Loring, T. 2009. Recommendations for the phytostabilization of acidic mine tailingsfrom SE Spain. Land Degradation and Rehabilitation–Drylands Ecosystems. Advances and GeoEcology40. Catena Verlag publishers GMBH, pp. 377–389.

426


Priority setting in Idaho’s Coeur d’Alene Basin

Bill AdamsU.S. Environmental Protection Agency

Daniel R. PitzlerCH2M HILL

ABSTRACT: The U.S. Environmental Protection Agency (EPA) Region 10 is in the process ofdeveloping a comprehensive cleanup plan in the Upper Basin of the South Fork Coeur d’AleneRiver. The cleanup is a complex project addressing widespread contamination from over a hundredyears of mining activity of multiple constituents over a broad geographical area in hundreds oflocations affecting land and water, with significant uncertainty and many stakeholders. This paperreports on the approach EPA is using to make decisions and prioritize actions for this project, whichincludes three main elements: 1). the adaptive management paradigm, in which future actions arerefined based on what is learned from initial actions; 2). a Multi-attribute Utility Model (MAU)which prioritizes actions in accordance with multiple objectives; and 3). a predictive analysis toestimate water quality goals at a distant point after remedial actions are taken.

1 INTRODUCTION

The Bunker Hill Mining and Metallurgical Complex Superfund site (“Bunker Hill Superfund site”)is located in Coeur d’Alene Basin of northern Idaho and eastern Washington (Figure 1). Miningand smelting in the Coeur d’ Alene Basin began in the 1880s, and the area became one of theleading silver-, lead- and zinc-producing areas in the world. After over 100 years of mining,milling, smelting and related activities, the Coeur d’Alene Basin now contains high levels ofmining contamination in the soil, sediment, surface water, and groundwater. The site was listedon the National Priorities List (NPL) in 1983. It includes mining-contaminated areas in the Coeurd’Alene River corridor, adjacent floodplains, downstream water bodies, tributaries and fill areas.The metals contamination poses both human health and environmental risks.

As a result of past mining, milling, and smelting practices, substantial portions of the Basincontain elevated concentrations of lead, zinc, cadmium, and other metals. Elevated concentrationsof metals resulted primarily from the discharge or erosion of over 62 million tons of mill tailingsand other mine-generated wastes into rivers and streams which, in turn, carried these wastes intodownstream streambeds, floodplains, and shorelines throughout the Upper and Lower Basins.Contaminated media include surface water, groundwater, soil, and sediments. Contaminants ofconcern are metals, particularly lead, arsenic, cadmium, and zinc. Some significant details on thehistory of milling and tailings disposal practices in the Basin are summarized below.

• Approximately 62 million tons of tailings were discharged to the Coeur d’Alene Basin aftermining began.

• Tailings were frequently used as fill material for residential and commercial constructionprojects.

• Until 1968, tailings tended to be discharged directly into the South Fork of the Coeur d’AleneRiver (SFCDR) or its tributaries. Most of the tailings were transported downstream, particularlyduring high-flow events, and deposited as solid tailings or as tailings/sediment mixtures in beds,banks, and floodplain areas. Since 1968, all mill tailings have been placed in impoundments orreturned as fill to provide structural support to active mines.

• Mining activities generated large volumes of waste rock and discharged water from mineopenings (adits) that contained high concentrations of metals.

427

Figure 1. Bunker Hill Superfund Site Location.

• Particulates released to the air from smelting operations contained high concentrations of metals.The particulates were transported by the wind and were deposited throughout the Bunker HillBox area (Box). (The Box is a 21-square-mile area surrounding the former smelter complex.)

• High concentrations of metals are pervasive in the soil, sediments, surface water and groundwaterin the Basin; these metals pose substantial risks to the people, plants, and animals that inhabitthe Basin.

• Elevated blood levels in children living in the Basin have been documented for more than 15years.

• Migratory birds and mammals have died due to ingestion of lead-contaminated soil and sedimentsin the Basin.

• Contamination from mining activities to surface water, soil, sediments, and biotic tissues havecaused increased mortality and decreased survival, growth, and reproduction to various plantand animals, particularly fish and waterfowl (Stratus 2000; EPA 2001b, c).

EPA has been conducting cleanup actions at the site since the 1980s. To date, remedial activitiesat the Bunker Hill Superfund Site have focused primarily on human exposure. Remedial actionshave included, but are not limited to:

• Hillside erosion control work, hydroseeding, terracing, and revegetation (over 1,000,000 treesplanted since 1992).

• Source removal and creek restoration and reconstruction.• Surface drainage improvements.• Upgrades to the Central Water Treatment Facility (CTP).• Removal and consolidation of approximately 4 million cubic yards of contaminated materials.• Demolition of industrial complex structures.• Capping of more than 800 acres to eliminate direct exposure to contaminants.• Removal of approximately 1.4 million cubic yards of contaminated materials from stream banks,

mine sites, and floodplains, and placement in repositories.

428

• Institutional controls have been put into place in the Upper Basin and other areas to ensure theprotection of the human health remedies.

• Cleanup of over 5,000 residential properties and rights-of-way.

2 PROPOSED CLEANUP PLAN

Over the past two years, EPA and its partners have been evaluating Bunker Hill Superfund sitecleanup activities in operable units (OUs) 1, 2 and 3 to develop a comprehensive, prioritizedcleanup approach for the Upper Coeur d’Alene Basin. The Upper Basin includes the South Fork ofthe Coeur d’Alene River and its tributaries downstream to where they combine with the North Fork,and the Box. The new cleanup approach is being undertaken in part to address National Academy ofSciences recommendations, to incorporate improved knowledge of the Upper Basin and Box and tomove forward on OU 2 Phase II cleanup activities. For example, additional data have been collectedduring the ongoing monitoring program, actions completed as selected mine and mill sites, andsite-specific studies (EPA 2004). This effort led to development of a Focused Feasibility Study(EPA 2010), and will identify and select additional remedial actions for the Upper Basin and Boxin an amendment to the record of decision (ROD) for the site (EPA 2002). The ROD amendmentis also planned to include actions to protect the human health remedy by addressing local drainageissues. Implementation of the new cleanup plan will provide a framework for reducing the amountof metals getting into streams and meeting environmental benchmarks that are necessary to protectfish, wildlife and provide a better quality of life for residents.

This plan is intended to be more comprehensive in nature in order to provide for human andecological protection for surface water. This plan will include actions at a larger number of Mineand Mill sites and sources areas (over 300 sites), removals of tailings in floodplains, protectionof existing remedies, treatment of groundwater from selected areas in order to reduce metal load-ing to surface water, and hydraulic controls to isolate contaminated groundwater for extraction.Simultaneously with this work, EPA is collecting additional information for the Lower Basin. Thiswork begins to address the complex interactions of sediment transport and deposition; sourcesand mobility of lead; river, lake, and floodplain interactions; and their relevant to environmen-tal risk and remediation. This work could lead to a future ROD Amendment for this portion ofthe site.

The outcome of this effort will be an Implementation Plan that will prioritize and guide theactions selected in the Upper Basin ROD Amendment. This will be a separate “living document”that will lay out a strategy for identifying priority projects that would be implemented in the nearterm. The Implementation Plan will be a component of an overall Adaptive Management Planfor the Upper Basin. Adaptive management is a process in which decisions are made as part ofan ongoing science-based process. It involves testing, monitoring, and evaluating applied strate-gies, and incorporating new knowledge into management approaches that are based on scientificfindings. The Implementation Plan and the Adaptive Management Plan will be tools to help EPAand others make better decisions as more information becomes available on the effectiveness ofinitial cleanup actions, and will provide the framework for the implementation of future actions.The Implementation Plan will be updated and modified on a regular basis to guide future decisionmaking and to determine when sufficient actions have been taken to meet the objectives of theUpper Basin ROD Amendment.

The Implementation Plan will consider a number of key factors such as metals loading to surfacewater, the potential for recontamination of clean areas, and the degree to which each remedial actionis expected to reduce risks to human health and the environment. Other factors to be consideredinclude whether water treatment would be required, whether repository space is needed, whetherrestoration work is planned, construction staging and design needs, coordination with local infras-tructure or public works projects, potential environmental issues associated with the actions (e.g.,the impacts of access roads), erosion potential, accessibility to children, and local community andstakeholder input. Another important consideration will be the amount of funding available onan annual basis. Consideration of these factors will help guide this important cleanup work andprovide transparency on how cleanup decisions will be made, the expected outcomes, and progresstowards meeting the objectives of the Upper Basin ROD Amendment.

429

3 PROBLEM STATEMENT

EPA faces a number of challenges as it seeks to prioritize a large list of sites that represent primarilya risk to the environment, including:

• Where to start• How to take action when significant uncertainty is present

– Contaminant pathways and recontamination potential– The extent of contamination at individual sites– The effectiveness and cost of various treatment options– The amount of money that will be available for cleanup

• How to ensure that it gets the “most bang for the buck”• How to make use of lessons learned from cleanup actions and incorporate most recent data

For the ecological remedy, EPA identified three priority issues for environmental protection thatfocus cleanup activities on reducing exposures to dissolved metals (particularly zinc and cadmium)in rivers and streams, lead in floodplain soil and sediment, and particular lead in surface water.

3.1 Sites

The ecological cleanup consists of more than 300 sites. Over the course of many years, EPAhas developed a database with information about each site including its location, access issues,ownership, distance to water bodies, habitat, contaminant levels, proposed remedy, projectedremedy effectiveness, and estimated costs. One key issue is the enormity and cost of characterizingthat many sites to a level sufficient for decision making. There is considerable uncertainty aboutmuch of the available information about the sites.

The remedies proposed range from passive and active water treatment, excavation of wasterock, tailings, and floodplain sediments, hydraulic isolation of contaminated materials, capping,regarding, and revegetation of consolidated wastes, and stream and riparian improvements.

3.2 Stakeholder objectives

Additional complexity was added to the ecological cleanup by the many different organizations andindividuals with stakes in the outcome. These stakeholders represent a wide range of views on whatfactors are the most important for prioritization of the cleanup work in the basin. EPA works closelywith a wide range of stakeholders, including state and federal, tribal, and local governments andcommunities, to identify and implement cleanup actions in the Coeur d’Alene Basin. This includescoordination with the Coeur d’Alene Basin Environmental Improvement Project Commission (theBasin Commission), which was established by the Idaho State Legislature. The Basin Commissionis composed of federal, state, tribal, and local governmental stakeholders, and its purpose iscoordination of cleanup activities, environmental restoration, and related measures in the Basin.Stakeholder Committees include a Citizens Coordinating Council (CCC), Technical LeadershipGroup, Project Focus Team (PFT) that provided technical and citizen perspectives into the decisionmaking process. EPA serves as the federal government’s representative to the Basin Commission.EPA will continue to be responsible for seeing that cleanup actions in the Coeur d’Alene Basinmeet the goals and requirements of decision documents and the Comprehensive EnvironmentalResponse, Compensation, and Liability Act of 1980 (CERCLA).

3.3 Development of the adaptive management and structured decision making framework

The uncertainties and complexities of this project noted above resulted in many challenges at theonset of the project to develop a unified understanding of key project issues and an effective frame-work for evaluation. In response, EPA used structured decision techniques from the field of decisionanalysis. A series of facilitated workshops were held with senior staff to explore objectives, keydecisions, data availability and quality, boundary conditions, stakeholder roles, and uncertainties.The results of these workshops was a significant improvement in the understanding of the problem

430

at hand, and a roadmap and framework for subsequent action, including the use of adaptive man-agement and a structured approach to decision making that includes multi-attribute utility (MAU)for prioritization.

4 ADAPTIVE MANAGEMENT AND DECISION MAKING APPROACH

4.1 Background and the case for adaptive management

The Comprehensive Remedy for the Coeur d’Alene Basin identifies EPA’s best estimates regardingimplementation of and expenditures for final human health and environmental cleanup actions forOU 3 (Grandinetti 2007). It recognizes that additional actions beyond the human health remedywill be required for protection of the environment. Thus, EPA developed a comprehensive remedyapproach to ecological protection in the Basin. EPA faces a number of challenges in managingremedy implementation and prioritization of actions, including the following:

• Budgets are relatively well known for any current budget cycle but are highly uncertain in futureyears. Increased funding levels are likely, but the magnitude and duration of that increase isunknown.

• The final, available budget and the number of years required for implementation is uncertain.• While much study has been done, there is still much to learn about contaminant sources, methods

of contaminant transport, and the cost and effectiveness of remedies.

In light of these uncertainties, EPA evaluated various evaluation frameworks (CH2M HILL 2008),and embraced MAU and the adaptive management paradigm under which decision making will beiterative (i.e., results will be evaluated and actions adjusted on the basis of what is learned). EPA’secological cleanup priorities will be addressed as follows:

1. Develop MAU model to prioritize the Upper Basin ecological cleanup actions outlined in theComprehensive Remedy and those identified for the Box.

2. Adjust the results of the MAU model to account for other factors such as implementation effi-ciency, recontamination potential, developing fish corridors vs. pristine habitat, opportunitysites with pending development proposals, potential for revised actions to remove additionalmetals, stakeholder impacts, and human health benefits.

3. Adjust activities and funding levels in a manner that reflects the MAU model results and bestmeets long-term objectives based on the judgment of EPA staff.

4. Modify and adjust the approach and workplan after consulting with EPA management and keystakeholders, such as the State of Idaho.

5. Modify the approach at appropriate times as EPA learns more about the uncertain aspects of thecleanup.

EPA’s implementation plan will address near-term resource needs for three main categories ofactivities:

– Monitoring (including status/trend and project effectiveness)– Predesign (including studies and developing the Lower Basin CSM)– Remedial design

4.2 Multi-attribute utility model

MAU is a quantitative technique for making decisions that involve multiple financial, environmen-tal, and social objectives (Keeney & Raiffa 1976). MAU is a useful framework for prioritizationbecause it is theoretically sound, relatively simple in concept, and it provides insight to decisionmakers and stakeholders. It provides an explicit understanding of why one action is preferred overanother—that is, the extent to which an action is consistent with the preferences of the decisionmaker. It is also an excellent approach for communicating with external stakeholders. EPA involvedstakeholders in the MAU approach, including discussions about the objectives hierarchy, scoringof actions against objectives, and the relative importance of objectives. The output from the processis visual and communicates results clearly.

431

Table 1. Objectives hierarchy and performance measures.

Objectives Hierarchy Performance Measure

1. Reduce dissolved metals (zinc) in surface waters to (Reduction in metal loading by remediating theprotect ecological receptors (fish) source) × (river miles from source to Harrison,

Idaho)2. Reduce particulate metal (lead) in soils/sediments inlakes, wetlands, riparian, and riverine areas to protectecological receptors (waterfowl, other wildlife)

2a. Protect receptors in vicinity of source area Acres of lake, wetland, riverine, or riparian habitataddressed by the action

2b. Protect receptors downstream (Pounds per day of particulate lead removed byremediating the source) × (river miles fromsource to Harrison, Idaho)

The MAU prioritization approach consists of the following steps:

1. Define objectives for prioritization.2. Determine the list of activities to be prioritized.3. Develop a method of measuring how well each activity meets the objectives, then score each

activity accordingly and normalize each score on a 0–1 basis.4. Develop weights that represent the relative importance of each objective.5. Develop a total score for each activity, representing the total “value” of that activity toward

cleaning up the Basin, calculated as a weighted average of scores and weights (assuming linearvalue measures).

6. Prepare a value-cost ratio for each activity.7. Rank each activity on the basis of value and value-cost ratio.8. Test the sensitivity of the rankings to differences in relative weights.

The model is a spreadsheet-based tool that can be updated easily as new data are developed,adjusted to reflect different opinions about the importance of various factors, and can be used toshow benefit to dollar spent. The objective of the MAU model is stated as follows:

Prioritize remedial actions in the Upper Basin and Box to identify a smaller number of sites fornear-term implementation and provide insight into managing this project.

The objectives hierarchy and performance measures for the model are shown in Table 1. Asindicated, the model will identify those sites that have the highest potential for improving surfacewater quality for ecological receptors by reducing dissolved metals and improving soil sedimentquality for ecological receptors by reducing particulate metal. The assessment of how well eachaction performs on these criteria was developed using the projected results of proposed remediesat each site based on existing data.

The initial weights assessed by the project team, and validated with the PFT, are approximatelya 60–40 split of relative importance for the two main objectives. Sub-objective 2b was viewedas being slightly more important than Objective 2a. The sensitivity of results to changes in theseweights is an important part of the initial MAU analysis.

The resulting value score for each action is divided by the current estimate of the long-term lifecycle cost of each action (in present value terms) resulting in a value-cost ratio. Actions are thenranked in descending order of value-cost ratio.

As indicated above, there are a number of other factors that are important to consider whendeciding which actions should be undertaken first, and how much money should be spent onremedial actions versus monitoring, predesign, and design. These factors are considered whendeveloping specific implementation plans. As time goes on, better information will be availableto refine value scores and costs. This framework is flexible and can be updated and modifiedconsistent with the principles of adaptive management.

432

4.3 Predictive tool

A predictive tool was developed during the Remedial Investigation/Feasibility Study (RI/FS) forthe Coeur d’Alene Basin (EPA 2001a), to provide point estimates of post-remediation surface waterquality. Specifically, the tool predicted average dissolved zinc loading in the SFCDR at Pinehurstand in the Coeur d’Alene River at Harrison, Idaho following the implementation of differentremedial alternatives identified in the RI/FS. However, it would be time consuming and extremelycomplex to use the original predictive tool to estimate the combined contribution of hundreds ofcontaminated sites on water quality. The process was further complicated by the need to evaluatethe time-frame for achieving the ambient water quality criteria (AWQC) at Pinehurst and Harrisonfollowing remediation. This evaluation also required estimates of source depletion or decay overtime.

As part of remedy planning and implementation, EPA evaluated different uses of the predictiveanalysis to help prioritize cleanup actions. EPA recognized the importance of being able to presentthe technical basis for prioritizing cleanup actions in a manner that is understood by the publicand other key stakeholders. EPA has assessed additional information, including new data collectedsince the RI/FS that could be used to predict cleanup effectiveness and prioritize cleanup actions.As a result, EPA (2006) developed a simplified approach to the predictive analysis. The simplifiedtool allows for the evaluation of source areas and the potential benefits of specific remedial actionsfor smaller segments of a stream versus the aggregated source areas and remedial actions evaluatedusing the predictive analysis. The simplified version also includes the evaluation of particulatelead loading. This is of particular importance because a significant amount of particulate leadtransported downstream to the Lower Basin is derived from sources within the Upper Basin. Thegoal is for this simplified approach to be more readily understood and reproducible by a variety oftechnical staff and interested parties within the Basin. This will be helpful as additional data arecollected and information regarding the impact of sources and remedial actions on water qualityare further refined.

The simplified version of the predictive analysis uses data collected synoptically (during thesame time period) at two monitoring locations to determine the gain in metal loading attributable tosources located between the upstream and downstream monitoring locations. Similar to the originalpredictive tool, sources of dissolved zinc loading associated with adits, seeps, groundwater, andother discrete discharges are separated out because they represent a direct input of loading to thestream in the reach. For particulate lead, loading from groundwater is considered to be minimalbased on groundwater quality data collected within the Basin and therefore is not assigned acontribution to load. In the analysis, if data is not available for a given discrete discharge source,the contribution of that source is assumed to be zero for both dissolved zinc and particulate lead.

After the removal of the contribution from direct inputs of loading to the stream reach, allremaining sources between the two monitoring locations are then segregated and identified basedon their source material type. The source material types were identified originally in the predictiveanalysis (URS 2001). For each source material type, a relative loading potential (RLP) for dissolvedzinc was developed in the predictive analysis. The RLP of a given source type is an index of theaverage contribution of metal load from that source type to the river per cubic yard of sourcematerial. The RLP therefore provides an estimate of the relative propensity of a source type tocontribute metal load to the river. Source types with the highest propensity to contribute metalload were assigned an RLP of 1.0 and other source types were scaled proportionally. For example,adits were assigned an RLP of 1.0 because their loading is contributed directly to the river. Table 2presents the RLPs for source types within the Upper Basin from the predictive analysis. For thesimplified version of the predictive analysis, the RLPs identified in the predictive analysis fordissolved zinc sources were retained as the best available estimates. The RLPs for dissolved zincloading were retained for particulate lead with the exception of deeper floodplain sediments. Thiswas done because to a degree, the RLPs for dissolved zinc were based on the propensity for sourcematerials to come into direct contact with surface water and storm water. It is recognized that theRLPs for particulate lead will need to be revised as data are evaluated, but at this time, it is believedthat they represent the best available RLP estimates for particulate lead.

For each source type, the RLP and the volume of the source type were multiplied to obtain arelative total volume. From the relative total volume, the percentage contribution and estimatedmetal load contribution for each source type were calculated.

433

Table 2. Relative loading potentials.

Dissolved Zinc Relative Total Lead RelativeSource Material Type Loading Potential* Loading Potential

Adit Drainage** 1.0 1.0Floodplain Sediments 0.795 0.795Deeper Floodplain Sediments 0.166 0Impounded Tailings, inactive facilities 0.143 0.143Impounded Tailings, active facilities 0.143 0.143Unimpounded Tailings 0.404 0.404Waste Rock with Loading Potential*** 0.059 0.059Waste Rock without Loading Potential*** 0.003 0.003

*All Relative Loading Potentials from Predictive Analysis (URS 2007), except as noted.** Includes seeps.*** Loading potential of waste rock as reported in the RI/FS (URS 2001).

Table 3. Remedial effectiveness factors.

Source Type and Remedial Action Dissolved Zinc Remedial Effectiveness Factor*

Adit DrainagePassive Treatment 0.11Active Treatment 0.01

Floodplain SedimentsExcavation/disposal in repository 0.01Hydraulic isolation at discrete facilities 0.18Hydraulic isolation of stream reaches 0.25

Tailings: Impounded, Inactive FacilitiesHydraulic Isolation 0.22

Tailings: UnimpoundedCap 0.05Excavation/disposal in repository 0.01

Waste Rock: With Loading PotentialRegrade/cover 0.46Cap 0.05Excavation/disposal in repository 0.01

Waste Rock: upland, little loading potentialRegrade/cover 0.46

*Dissolved Zinc Remedial Effectiveness Factors from Predictive Analysis (URS 2007).

For each source, the remedial actions and their associated remedial effectiveness factors wereidentified. Dissolved zinc remedial effectiveness factors were developed as part of the predictivetool and are summarized by source type and remedial action in Table 3. The remedial effectivenessfactors indicate the degree to which loading from the contaminant source will be present followingremediation on a scale from zero to 1. Therefore, remedial actions anticipated to be the mosteffective are assigned remedial effectiveness factors closer to zero and for sources where no actionis selected a remedial effectiveness factor of 1.0 is assigned.

Remedial effectiveness factors for particulate lead are also presented in Table 3. The remedialeffectiveness factors for particulate lead were revised from the dissolved zinc remedial effectivenessfactors based on professional judgment to reflect the estimated impact of remedial actions on therelease of particulate lead from source materials. Remedial actions involving the removal andcapping of source materials would be expected to have the greatest impact on reducing particulatelead releases from source materials. Actions that focus on limiting the transport of metals throughgroundwater/surface water interaction or groundwater movement (such as, hydraulic isolation and

434

groundwater cutoff walls) would not be anticipated to have an appreciable impact on particulatelead release. Lead is typically either not present or present at low concentrations in groundwaterand then only in areas with relatively acidic groundwater conditions.

The estimated dissolved zinc load contribution for each source is then multiplied by the remedialeffectiveness factor for the selected remedial action to estimate the dissolved zinc loading associatedwith each source that would be expected to remain following remediation. The resulting zinc andlead load potential for each remedial action were used as inputs to the MAU model.

4.4 Stakeholder engagement

The EPA and State used a variety of different forums to communicate with stakeholders includingregular meetings with the CCC and PFT, newsletters, public open houses, and public serviceannouncements in local newspapers and radio stations. The PFT and Natural Resource Trusteeswere engaged on a regular basis during development of the predictive tool and MAU model toprovide guidance and feedback on objectives and how the models are used. Many PFT and Trusteemembers had intimate knowledge of sites, the likely effectiveness of remedial action, and thepotential for win-win solutions that could also result in habitat restoration. Their input led to manymodifications and adjustments to help ensure that actions taken will result in as much benefit aspossible.

In addition to stakeholder input, EPA’s National Remedy Review Board has reviewed the BasinFFS and recognized the importance of an Implementation Plan for such a large project. The Boardalso commented that final decision documents should include a description of the procedures fordeveloping and updating an implementation plan.

4.5 Implementation approach

Implementation plans were developed for two assumed funding streams: $15 m/year; $25 m/year.The plans were established by year for the first five years, under the assumption that future planswould be developed based on lessons learned during the initial five years. The plans were developedusing MAU principles accounting for other important factors such as: implementation efficiency,Recontamination potential, developing fish corridors vs. pristine habitat, opportunity sites withpending development proposals, potential for revised actions to remove additional metals, stake-holder impacts, and human health benefits. A diagram that shows the process that will be used tomanage implementation is shown in Figure 2.

The initial plans were modified and adjusted after consulting key stakeholders, TLG, CCC,Community etc., then adaptive management principles are used to monitor and modify as more islearned.

The plans assume that BEMP for monitoring would continue each year. Development of aregional repository was placed on the critical path because it is needed for cleanup of many mine-millsites: a new repository was assumed to be online in 3–4 years.

Remedy protection was also a high priority during implementation. There are some remediesthat are at risk of failure during flood events. Priority areas for protection were developed usingcriteria developed by EPA, DEQ, and local officials.

Water treatment is required to manage the largest zinc loading sites. This will require an expansionof the Central Treatment Plant (CTP) and developing an Upper Basin Influent Pipeline. The CTPexpansion is on the critical path so that two key actions (the Box drains and Woodland Park drains)can proceed. The influent pipeline is somewhat less critical because it is needed for somewhat moreremote sites requiring water treatment.

Mine and mill sites are addressed as funds are available. However, some sites are included inthe initial five-year plan to “show progress” and obtain data about cleanup effectiveness that caninform future actions. The sites included show high value and value to cost ratio in the MAU model.

5 CONCLUSIONS

The ecological cleanup of the Coeur d’Alene Basin is complex with significant uncertainty aboutindividual sites and remedy effectiveness, substantial costs that are also uncertain, engaged

435

Figure 2. Implementation plan approach.

stakeholders, and multiple, conflicting objectives. EPA used a predictive tool, a MAU model,and structured venues for receiving stakeholder feedback to develop an approach for initial imple-mentation. These processes and tools have proven to be effective methods of prioritizing actions inthe face of daunting complexity and uncertainty. As actions are completed, EPA will use adaptivemanagement to monitor the effectiveness of those actions and refine the predictive tool and MAUmodel for use in subsequent implementation planning.

REFERENCES

CH2M HILL. 2008. Coeur d’Alene Basin Comprehensive Remedy: Overview of Planning and Prioritiza tionTools, Technical Memorandum. Prepared for U.S. EPA. September 12, 2008.

EPA. 2001a. Coeur d’Alene Basin Remedial Investigation/Feasibility Study. Prepared by URS-Greiner Inc.and CH2M HILL. October 2001.

EPA. 2001b. Human Health Risk Assessment for the Coeur d’Alene Basin Extending from Harrison to Mullanon the Coeur d’Alene River and Tributaries. Prepared by Terragraphics Environmental Engi neering, Inc.June 2001.

EPA. 2001c. Technical Memorandum, Interim Fisheries Benchmarks for the Initial Increment of Remed iationin the Coeur D’Alene Basin, Final. September 2001.

EPA. 2002. Record of Decision, Bunker Hill Mining and Metallurgical Complex Operable Unit 3 (Coeurd’Alene Basin), Shoshone County, Idaho. EPA DCN:2.9. September 2002.

EPA. 2004. Basin Environmental Monitoring Plan – Bunker Hill Mining and Metallurgical Complex, OU3.March 2004.

EPA. 2006. AnalyticalTools toAssess Benefits of CleanupAlternatives on SurfaceWater Quality in the Future –Strawman Concept and Proposal. November 27, 2006.

EPA, 2010. Draft Focused Feasibility Study Report for the Coeur d’Alene River, Bunker Hill Mining andMetallurgical Complex Superfund Site.

436

Grandinetti, C. 2007. Report of the United States Environmental Protection Agency’s Comprehensive CleanupApproach for the Coeur d’Alene Basin. June 14, 2007.

Keeney, R. & Raiffa, H. 1976. Decisions with Multiple Objectives. Cambridge University Press.URS. 2001. Probabilistic Analysis of Post-Remediation Metal Loading, Technical Memorandum (Revi sion 1).

Prepared for U.S. EPA. September 2001.Stratus Consulting. 2000. Report of InjuryAssessment and Injury Determination: Coeur d’Alene Ba sin Natural

Resource Damage Assessment. Prepared for the U.S. Department of the Interior, the U.S. Fish and WildlifeService, the U.S. Department of Agriculture, the U.S. Forest Service, and the Coeur d’Alene Tribe.

URS. 2007. A Predictive Analysis for Post-Remediation Metal Loading, Coeur d’Alene Basin RI/FS TechnicalMemorandum. Prepared for U.S. EPA. October 2007.

437


Acid mine drainage as a sustainable solution to eliminate riskand reduce costs

James CormierXstrata Zinc, Brunswick Mine, New Brunswick, Canada

ABSTRACT: Acid Mine Drainage (AMD) is used to geochemically sequester dissolved sulfidesin process water. The sulfides are a product of sulfide reduction in a reclaim process water pond.They create sustainable development risks as they produce hydrogen sulfide, other Total ReducedSulfur compounds (TRS), and increase the alkalinity of the water.

1 DESCRIPTION AND HISTORY

Brunswick Mine commenced operations in the early 1960’s and to date has extracted more than135 million tonnes of ore from the deposit. Located 20 kilometers southwest of Bathurst, NewBrunswick, Canada, the site continues to be one of the largest lead/zinc mines in the world,producing zinc, lead, copper and bulk concentrates. Brunswick Mine employs 950 people, and isexpected to cease operations within 2 years, when the orebody is depleted.

Currently, Brunswick operates at a planned mining and milling rate of 9,200 tonnes per day ofcomplex lead, zinc, copper and silver sulphide ore. Four concentrates are produced by selectiveflotation.

The major mining methods at Brunswick were summarized with the assistance of Barb Rose,P.Eng, Senior Design Engineer; Brunswick’s mining methods have progressed from blast hole openstoping during the 1960’s, to cut-and-fill mining in 1970, and open stoping method with primarystopes and secondary pillars in 1985. A modified Avoca method has also been employed in areaswith weak wall rock. Since 1996 the operation has been in transition to the use of smaller stopesand pillarless mining. This enables better control of the ground around the open stopes, and limitsstress concentrations that in the past have led to instability of the remaining pillars. The use of pastebackfill has allowed full pillarless mining as well as destress blasting of highly stressed groundwith access via tunnels in paste to recover the remaining ore (Rose, personal communication, June2010).

In 1998, a Semi-Autogenous Grinding (SAG) Mill replaced the crushing, rod mill and primaryball mill circuits in the concentrator. In 2000 implementation of a paste backfill process enabledthe conversion of the backfill quarry to a retention pond to increase the amount of reclaim waterused for the milling process. In 2005, another project was executed to expand the storage volumeof the “reclaim quarry” to allow for a sustainable treatment of thiosalts.

Figure 1 provides a simplified flow sheet for the concentrator. The overall flow sheet involvesSAG milling, secondary ball milling, selective flotation, and dewatering by thickening, filteringand drying. The complexity of the flow sheet evolves from the requirements for extremely finegrinding and the production of the four concentrates: Zinc (Zn), Lead (Pb), Copper (Cu) and Bulk(Pb + Zn) through metallurgical beneficiation (Deredin, 2008).

Since 2005, consumption of pH modifying reagents used in the ore milling process has con-sistently increased. During 2007 and 2008, odors of total reduced sulfur compounds from thereclaim quarry became an issue as these odors impacted the air quality in the underground work-ings, through the ventilation system. During 2009, the concentrations of these odors increased toa level where the entire mine site, as well as the surrounding community (distance of 13 km) wereimpacted for several weeks, until a solution was developed.

439

Figure 1. Concentrator flow sheet (Roberts, 2008).

2 WATER QUALITY

2.1 Recycling rate

Over the years, Brunswick Mine has evolved their processes to reduce the fresh water consumptionof the milling processes. Jean-Guy Paulin, Six Sigma Black Belt and Energy Management hasbeen appointed to champion the processes related to water management. The evolution of wateruse at Brunswick has included the implementation of recycling loops, directing the tailings to thetailings basin as a 30% solids solution, and the water is reclaimed back to the mill, for re-use.Prior to 2001, the reclaim system had very limited storage capacity and the mill’s ability to recyclewater was constrained by the flow from the tailings decant tower. To remove this constraint and tomaximize the recycling rate, a reservoir was required to balance the peak flow requirements, andensure the milling requirements would not be compromised. In 2001, a paste backfill system forfilling underground voids was constructed. This made the rock quarry redundant, and enabled it tobe converted to a reservoir for the storage of process water. This quarry measures approximately500 meters by 500 meters, to depth of 20 meters. Storage of 1.25 million Cubic meters (m3) wasachieved using a depth 5 meters of process water. This enabled the mill to increase the recyclingrate from 48% to 64%, while reducing the fresh water withdrawal rate from the Nepisiguit River by20%. This recycle rate increase also contributed to a step change in concentrations of parametersin the mill process water, since the reduction in fresh water used reduced the dilution of the processwater (Roberts 2008).

2.2 Thiosalts

The milling process creates partially oxidized intermediate sulfur compounds, these compounds,(thiosalts) are in the form SnO−2

m . The majority of thiosalts are Thiosulphate (S2O−23 ), Trithionate

(S3O−26 ), and Tetrathionate (S4O−2

6 ). Thiosalt neutralization is not achieved in the conventional limeneutralization process used in the Brunswick Mine Effluent Treatment Plant (ETP). Technologyto treat thiosalts in water involves completion of the reaction by oxidation to stable sulfates, andneutralizing the resulting acidic solution. During the spring and summer months, the thiosalts areoxidized in the receiving water of the Brunswick Mine effluent. This oxidation resulted in pHdepression in the natural environment, and consequently impacted Little River.

440

Figure 2. Average volumetric flow rate (l pm) and recycling rate (Paulin, personal communication, June,2010).

Figure 3. Site water balance (Roberts, 2008).

441

Figure 4. Alkalinity in the reclaim water (Cormier, 2010).

In 2005, a thiosalt concentrations management project commenced to reduce the risk of pHdepressions downstream of the effluent discharge. This project consisted of increasing the level inthe reclaim quarry to 10 m, and storing up to 2.5 million m3 of process water until the concen-trations of thiosalts could be effectively treated with hydrogen peroxide (H2O2) which completesthe oxidation of the thiosalts prior to treatment at the ETP. A proportional trend showed Thiosaltconcentrations have increased as the recycling rate has increased over the years.

2.3 Sulfates

The process water used in the milling process has a high concentration of sulfates. In 2001,sulfate concentrations in process water were measured at 2 g/L (2,000 mg/L). Over time, theseconcentrations gradually increased to the 2010 levels measured at 4 g/L (4,000 mg/L).

2.4 Alkalinity

Measurements of alkalinity in the reclaim water began in the fall of 2001. The reduction in freshwater, and increase in recycling rate had a slight effect on the alkalinity of the reclaim water, butconcentrations were relatively stable at 200 mg/L. In 2004, the alkalinity decreased as the quarrylevel was raised and more water was put into storage. The alkalinity remained at a low level untilearly in 2007, when the alkalinity increased steadily from less than 50 mg/L to the peak valuesmeasured in the fall of 2009 at over 1000 mg/L. Figure 4 portrays the historical alkalinity values.

The peak alkalinity values measured in the fall of 2009 resulted in risks to production. Thequantity of soda ash required for the milling process was threatening to exceed the rate at whichsoda ash could be delivered. Although production shutdowns due to lack of reagent availabilitydid not occur, a difficult situation was created for procurement to manage. It is possible that if therate of increasing consumption had continued the delivery and availability of pH modifiers wouldcreate a constraint to production rates.

2.5 Comparison of plant tailings and quarry reclaim concentrations

Historically, the quarry reclaim water is lower in thiosulfate concentrations than the concentrationsof thiosulfate in the plant tailings. The relative difference remains constant over time, sulfate

442

Table 1. Concentrations of thiosulfate and sulfate in the reclaim quarry and tailingsstreams (Cormier, 2009).

Thiosulfate Sulfate in Thiosulfate Sulfatein Quarry Quarry in Tailings in Tailings(mg/L) (mg/L) (mg/L) (mg/L)

April 2002 500 2263 983 2212Sept 2005 1314 3470 1501 3500Sept 2009 1870 4722 2233 5121

levels are relatively consistent. The concentration has increased consistently in both thiosulfate andsulfate, resulting in double the concentrations in 2009 compared to the concentrations in 2001.

2.6 Other parameters

Research and data compilation indicates that the pH of the reclaim system does not show anysignificant trends. Occasional measurements made for dissolved oxygen have indicated anoxicconditions in the lower layer of the reclaim quarry.

3 REDUCED SULFUR

For many years, odors of total reduced sulfur (TRS) compounds were detected in the immediatevicinity of the reclaim quarry. In the fall of 2007, the concentrations of these odors increased whichresulted in TRS odors, including H2S which were detected throughout the mine site. These odorsresulted in production impacts, since the ventilation system for the underground operations waswithdrawing air from the surface that contained concentrations of TRS compounds that were abovehuman threshold detection levels of 20 parts-per-billion (ppb) (Gerits 2009). Within a few days,the concentration of TRS odors returned to lower levels, before the determination of cause wascompleted. A similar episode occurred in the fall of 2008.

The growth cycle of sulfate reducing bacteria (SRB) begins in the lag phase (a period of slowgrowth when the cells are adapting to the high-nutrient environment and preparing for fast growth).The lag phase has high biosynthesis rates, as proteins necessary for rapid growth are produced.The second phase of growth is the exponential phase. The exponential (log) phase is marked byrapid exponential growth. During log phase, nutrients are metabolized at maximum speed until oneof the nutrients is depleted and starts limiting growth. The final phase of growth is the stationaryphase and is caused by depleted nutrients.

The concentrations of nutrients Sulfates and Thiosalts in this biosystem (reclaim quarry) wereincreasing over time. This was due to the fact that the minimum operating low level in the quarrycontained 40% of the volume. Another factor was the lack of dilution of fresh water addition.The effect of increased recycling rates has reduced the fresh water consumption. As a result, theconcentrations of sulfates and thiosalts increased more than double their concentrations than in2001.

The conditions in the quarry (concentrations of sulfates and thiosulfates, lack of dissolvedoxygen, pH, and temperature) enabled the Sulfate Reducing Bacteria (SRB) populations to reducesulfate in large amounts to obtain energy and expel the resulting sulfides as waste; this is knownas dissimilatory sulfate reduction.

Sulfides are formed by disproportionation of Thiosulfates (Gerits 2009):

Or sulfate reduction:

443

Figure 5. Concentrations over time of thiosulfate (left) and sulfate (right) in plant tails and reclaim quarry(Cormier, 2010).

Figure 6. H2S(g) in quarry water and air at (H2S) of 0–300 mg/L (Gerits, 2009).

In the fall of 2009, following an unseasonably dry August, the concentration of TRS gases fromthe reclaim quarry were much higher concentrations that had been observed in previous autumns.The concentrations were at such levels that inquiries were received at the mine site for the localcommunity (>10 km South of the mine site), depending on wind direction. All employees on themine site were easily able to detect these TRS odors, as the odor threshold of 20 ppb was exceededon a regular basis.

The decrease in thiosalt concentrations between the plant tailings and reclaim quarry indicatea disproportionation of thiosulfate (S2O−2

3 ) is occurring. The rate of reduction of concentrationof thiosulfate is relatively constant over the years, but the concentration has doubled since 2001,resulting the twice the quantity of thiosulfates available for reduction by SRB to dissolved sulfides.The average monthly reduction in thiosalts of 500 mg/L (95% CI 461, 561) represents a tremendousquantity of thiosulfates available for reduction by SRB to dissolved sulfides.

This resulted from concentrations of dissolved total sulfides (passing 0.45 micron) in the reclaimwater at 280 mg/L. Thiosulfate and sulfate concentrations also reached the peak measured values.The sulfide concentrations grossly exceeded the solubility in water, and TRS odors were detecteddownwind of the Mine for several kilometers.

The reported partial pressure (atm) and volumetric concentration (ppm) of H2S(g) in the airrefer to a closed system. At the time TRS odors from the quarry was detected in the surrounding

444

community, the dissolved sulphide concentration in the quarry water had reached 280 mg/L cor-responding to a concentration of 2440 ppm in the air (closed system). This implies that even afterapproximately 1:2 million times (i.e. 1,926,230) dilution in the air, the H2S(g) from the quarrycould still be detected in the surrounding community (Gerits 2009).

4 INCIDENT MANAGEMENT

4.1 Health and safety

Completion of a risk assessment determined that the risk of these TRS odors entering the under-ground work environment was high. Employees in the underground environment rely on theirolfactory system to detect the stench gas (ethyl mercaptan) released to alert them of an emergency.The TRS odors also could be confounded to be odors from a fire in the underground environment.The concentrations measured on site were as high as 200 ppb instantaneous, with an hourly averagepeaking at 20 ppb. These concentrations are below published exposure limits, but short term expo-sures at these concentrations have been reported to cause headaches and nausea. The Joint Healthand Safety implemented a code of practice to reduce these risks of confusion over the odors, andactions to take when odors were detected. An employee communication program was developed.All employees were informed of the situation, and the code of practice. During September andOctober there were more than 30 hours of lost production impacted by the odors, which resultedin a calculated loss of $300,000.

4.2 Environment & Community

A full disclosure presentation was made to the New Brunswick Department of the Environment.Several inquiries were received from residents at Nepisiguit Falls, 8 km south of mine site andRose Hill 10 km. On September 17th 2009, reports of TRS odors in Allardville (25 km East) werereceived. Prevailing winds from the mine site are from the West and Northwest, causing TRS odorsto be carried downwind to the communities East and Southeast of the Mine.

Sulfide concentrations in water can be acutely lethal to fish, (Smith, 1974), and are recognizedto be contributors to toxicity observed in effluent failing to meet the 96-Hour acute lethality test,required by the Metal Mining Effluent Regulations (MMER). The concentrations of sulfides werealso a risk to compliance with the MMER. The reclaim quarry level is lowered in the fall andwinter seasons, which enables adequate storage water in the reclaim quarry for the spring freshet.The water is typically pumped and/or siphoned into the water management system. The watermanagement system consists of a series of collection ponds, which discharge into an 820,000 m3

buffer pond. The water is pumped from the buffer pond to the high density sludge (HDS) effluenttreatment plant. The ETP is capable of treatment rates as high as 60,000 lpm. Treatment consistsof lime neutralization and clarification prior to discharge to the Little River.

5 DEVELOPMENT OF A SOLUTION

The response from the employee communication sessions was overwhelming, many employ-ees offered suggestions for potential solutions. A list of the opportunities was developed whichincluded aeration, oxidation, chemical treatment, sulfide specific chemicals, odor masking agents,de-odorizers. All solutions presented risks, including cost and effectiveness uncertainties. Per-sonnel from Brunswick, Xstrata Zinc, and a 3rd Party environmental consulting firm CH2MHillcollaborated to analyze the options. During the analysis of these options it was it was discoveredthat acid mine drainage (AMD) could treat the aqueous sulfide which was produced by the sulfatereducing bacteria. It has been recognized for many years that sulfide bioreactors could successfullybe used to treat AMD. At Brunswick Mine, AMD has been treated by the HDS process. A fullscale SRB bioreactor has never been developed. The natural sulphide production by SRB in thequarry created an opportunity to reduce sulfide concentrations with AMD. This was confirmed inconsultation with CH2MHill’s Associate Consultant Robert (B.T.) Thomas, Ph.D.

445

Brunswick Mine has two sources of AMD, the underground operations at the Brunswick #12 mine, and the Brunswick # 6 open pit, which has been closed since 1984. Bench tests wereconducted with reclaim water and AMD, and odors were eliminated at concentrations as low as0.1% (v/v) AMD. AMD from # 6 open pit typically contains 1,100 mg/L dissolved iron (400 mg/LFerrous), and 200 mg/L dissolved zinc with a pH level of 3.0. AMD can be effective in removalof sulfides in water, due to the fact that sulfide has a naturally high affinity for complexing withcertain metals, thereby causing an immediate reaction producing stable metal sulfides. The reactionof dissolved metals (Ferrous Iron and Zinc) and aqueous sulfides produces insoluble metal sulfidesZnS and FeS. Removal of dissolved sulfides in water would eliminate the odors, by lowering thesulfide concentration below the solubility levels in water. Bench testing showed that concentrationsat 5% (v/v) AMD lowered sulfide concentrations to 100 mg/L, and maintained the pH near neutral(pKa of H2S is 7).

The concept of adding AMD from the # 6 open pit to the quarry was evaluated through a riskassessment approach, and it was determined that the overall risk ranking was very low. An interimsystem was quickly implemented to re-direct theAMD from the pipe discharge to the reclaim quarry,a distance of 5 km. During the installation of this system, trucks were contracted to transport theAMD. Jan Gerits, Ph.D, Senior Environmental Geochemist with LORAX Environmental visitedthe site on September 24th, shortly after AMD additions had started, and developed a technicalmemorandum on the remediation of the reclaim water. In November, the pumping system at the# 6 pit was de-commissioned for the winter. A project was developed to permanently pump theAMD from # 12 underground mine. This system was commissioned on February 25th, 2010. It wascapable of delivering 1000 l pm of AMD, containing typical concentrations of Iron at 1,000 mg/L(600 mg/L Ferrous) and 1,500 mg/L Zinc.

6 RESULTS

6.1 Chemistry

The addition of AMD to the quarry had an immediate positive effect. The TRS odors were imme-diately reduced and eventually eliminated. The only constraints were mixing and volumetric flowrate. The color of the water changed from an opaque black to a light orange. Concentrations of sul-fides were reduced by 40% as of June 30th, 2010, reaching concentrations of 150 mg/L. Alkalinitywas also reduced (−66%) with levels reaching 350 mg/L while suspended solids levels remainedconstant.

Dissolved concentrations of Fe and Zn were reduced to less than 1 mg/L. pH levels have remainednear 7 despite the addition of 1,000 l pm ofAMD at pH 3. Figure 7 shows the pH fluctuating between6.8 and 9 the alkalinity and sulfides have been reduced significantly.

Concentrations of both thiosulfates and sulfates have also reversed their increasing trends, andhave been declining steadily since the AMD additions were implemented. This tread is illustratedin Figure 5, where the concentrations of both thiosulfates and sulfates had been increasing, priorto October 2009, and decreasing since.

Biweekly samples at 4 locations, and at 3 meter depth intervals revealed stratification below 6meters from the surface. At these depths anoxic conditions exist and Sulfide levels are 200 mg/L.This remains a risk for a short-term episode if an annual mixing event were to occur as the reclaimquarry behaves like a meromitic lake. Water quality parameter levels in the water column at 6 metersdepth are consistent throughout the reclaim quarry. Alkalinity and sulfide concentrations showa strong linear correlation and the reduction in sulfides results in alkalinity levels returning toconcentrations similar to 2007 levels. The addition of 1000 lpm of AMD has not significantlyimpacted the pH of the reclaim quarry water, as shown in Figure 7, but the AMD additions havereduced the sulfide concentrations, which has resulted in decreased alkalinity.

7 EFFLUENT QUALITY

The sulfide concentration in the reclaim quarry will have an effect on the sulfide concentrationsin the ETP effluent, when the level in the reclaim quarry is lowered. In November, when the # 6

446

Figure 7. Alkalinity and sulphide concentrations (Cormier, 2010).

Figure 8. Annual consumption of sulfur dioxide (Cormier, 2010).

447

Figure 9. Annual consumption of soda ash (Cormier, 2010).

Open Pit pumping system was de-commissioned for the winter, sulfide levels had reached a low of211 mg/L. In the elapsed time between the de-commissioning of the # 6 Open Pit pumping system,and the commissioning of the # 12 Mine pumping system on February 25th, the dissolved sulfideconcentrations were measured at 250 mg/L. The variation of sulfide concentrations in the reclaimquarry is shown in Figure 7.

In March 2010, an effluent sample collected from the discharge of the ETP did not meet therequirements for the 96-Hour acute lethality test. An Acute Lethality Response Protocol (ALRP)investigation conducted with Lesley Novak, M. Sc., Vice President, Senior Aquatic Toxicologist,Aquatox Testing and Consulting of Guelph, Ontario determined that dissolved sulfide concentra-tions in the effluent were the root cause of the toxicity (Novak, Personal Communication, March2010).

Shortly after the # 12 Mine pumping system was fully operational, the concentrations of sulfidesreturned to the previous minimum concentrations of 210 mg/L and by the end of March concen-trations were 200 mg/L. Also at the end of March concentrations of sulfides in the ETP Effluentwere less than the detection limit of 0.02 mg/L. Samples collected from the effluent showed mor-tality in the 96 Hour test to have dissolved H2S concentrations of 3–5 mg/L (Novak, PersonalCommunication, March 2010).

7.1 Reagent consumption

The reduction in alkalinity (60% as of June 30th) has resulted in a huge impact on the consumptionof pH modifying reagents used in the milling process. Sodium Carbonate (Na2CO3) and SulfurDioxide (SO2) were reduced by 20 and 25% respectively. This reduction has resulted in a costsavings of $ 970,000 at the end of the 2010 second quarter, and is projected to save approximately$ 1 million in the 3rd and 4th Quarters, for an annual savings of $ 2.0 million. Further reductionsin alkalinity should continue to positively impact savings in the future.

Addition reagent consumption reductions were observed at the effluent treatment plant (ETP).The removal of dissolved Fe and Zn, as well as the reduction in alkalinity of the reclaim quarrywater has reduced the lime demand at the ETP. This has resulted in an additional savings of $ 30,000through June 2010 from reduction of lime consumption.

448

8 SUSTAINABLE DEVELOPMENT SOLUTION

A health, environmental, community and production risk was not treated effectively and an extendedincident occurred. The sustainable solution to eliminate this risk was treatment with a waste. Thetreatment with a waste was very effective in eliminating the risk and also resulted in an improvementin water quality for water recycled in the milling process. The improved water quality significantlyreduced consumption of chemicals. The reduction in consumption of these chemicals has reducedoperating costs by $ 970,000 in the first 6 months of 2010. This is a prime example of what WilliamMcDonough principles of design for sustainability, with one of his foundations of a cradle-to-cradledesign, waste equals food (McDonough 2002).

9 CONCLUSION

This case study exhibits a successful application of the principles of sustainable development inmining. Although the details of the application are specific to Brunswick Mine, the strategy usedto develop a solution is applicable to any sustainable development risk.

Many components of Xstrata’s sustainable development management system were applied.Effective communication and engagement throughout the incident management process led tothe development of a list of potential solutions. Evaluation of the potential solutions using a rig-orous risk and change management system led to the selection of a sustainable solution. Effectiveassessment of the solutions resulted in the determination of the best option of using a technol-ogy that was well established in mining, (sulfide bioreactors are used to treat AMD). In this casesome counter-intuitive thinking was required to apply the process in reverse and use AMD as thetreatment for the un-intended product of the reclaim quarry sulfide bioreactor.

As a matter of sustainable development, the use of a waste to treat the waste from another processis classic use of the “waste equals food” concept. As a matter of sustainable development, the useof a waste to treat the waste from another process is classic use of the “waste equals food” concept.

ACKNOWLEDGEMENTS

Many people contributed to the success of this project, including but not limited to the following:employees at Brunswick Mine, Robert (BT) Thomas (CH2MHill), Florian Reher (Venture Engi-neering and Construction), Lesley Novak (Aquatox Testing and Consulting), Paul Deveau (XstrataZinc Canada), Rick Schwenger (Xstrata Zinc Canada) and Christina Vink (Memorial University).

449


Chemical compound forms of cadmium in uranium tailingsof Schneckenstein

T. NaamounFaculty of Sciences, Sfax, Tunisia

B. MerkelTU Bergakademie Freiberg, Germany

ABSTRACT: To outline the behaviour of any hazardous heavy metal in soils and hence its toxicitytoward terrestrial biota, an understanding of the ways of its binding or of its specific chemical formsin the material of investigation are required. The present study sought to evaluate the release andmobility of cadmium from abandoned uranium residues. For that purpose, a seven steps sequentialchemical extraction was the tool to define its different compound forms in the processed sediments.The results show that most of cadmium seems to be bound to the host minerals such as hornblendeand chlorite. Also, from 16 to 69% the non residual cadmium is in association with the biogenicphase. Eleven to 35% is linked to carbonates. Nineteen to 100% of cadmium is exchangeable.

1 INTRODUCTION

Between 1947 and 1957, the German Democratic Republic (DDR) developed the third largesturanium mining province of the world after the US and Canada. During peak times, the GermanCompany “Wismut produced 7000 tonnes of uranium per year. And, its total production was in theorder of 220,000 tonnes. After the German unification at the end of 1990, uranium mining wasshut down. Hundreds of millions of tonnes of radiating waste rock and uranium mill tailings areleft over, presenting health risks through release of radon gas and contaminated seepage.

According to Salvarredy-Aranguren, et al., (2008) and Nriagu, (1996), mining activities areknown to be a major source of environmental heavy metal contamination. As a result water course;stream sediments and alluvial plains have been and continue to be contaminated with metals.Thus, the German Federal government funded a remediation program with a budget of about9.3 billion US$.

The evaluation tailing area at the Schneckenstein site was carried out between 1995 and 1997within the frame of a technical cooperation between the Chair of Hydrogeology of the TechnicalUniversity of Mining and Technology Freiberg and the department of Ecology and EnvironmentalProtection of the Technical University Dresden. The investigation includes measurements of pre-cipitation and evaporation as well as surface runoff and the chemical analyses of the surface andground water. Also the project was extended and co-financed by the German research community.Different measurements including physical properties of soils; radioactive elements determinationas well as geochemical exploration are carried out.

The behavior of metals in soils (e.g., mobility, bioavailability) cannot be reliably predictedon the basis of their total concentrations (Tack & Verloo, 1995; Van Peijnenburg et al., 1997;Krishnamurti & Naidu, 2003; Meers, et al., 2007). Also, the uptake and toxicity of many metalsshow marked dependence on speciation of the metals and these responses often correlate best withthe activity of free metal ions (Laxen & Harrison, 1981; Knight & McGrath, 1995; Parker & Pedler,1997; Prokop et al., 2001). The process of identifying and quantifying these different species ofmetals in a sample is referred to as speciation. According to the findings of Janssen et al. (1997)and Janssen et al. (1997), the bioavailability of metals in soils or sediments is often expressed interms of concentration in a water phase.

451

Figure 1. The area of investigation, uranium tailings, Schneckenstein.

The present study sought to outline the chemical behaviour of contaminant cadmium in tailingsediments. This paper focuses on the chemical fractionation of cadmium in uranium residue byusing extraction chemical procedure.

2 THE STUDY AREA

The area of investigation, the uranium tailing Schneckenstein, is located in the Boda valley northof the village of Tannenbergsthal/county of Vogtland, southwest of Saxony (Figure 1).

The Boda valley is boardered by the Runder Hübel (837 m above sea level) and the Kiel (943 ma.s.l) to the Northwest and Southeast respectively. The area is situated within the southern branchof the Boda valley at an altitude of 740 to 815 m above sea level.

Geologically, the area belongs to the transverse zone of the southern Vogtland and the westernOre Mountains; a subunit of the anticlimal zone of the Fichtelgebirge – Ore Mountains. It is locatedon the Southwest border of the Eibenstock granite, a biotite -syenogranite which belongs to theyounger granites (J G 1). It is medium to coarse-grained, serial-porphyric and tourmaline bearing.Its intrusion is related to the Upper Carboniferous series.

Also, the investigated sites lie in the area of the watershed between the Zwickauer basin and theEger. It belongs to the highest precipitation area of the entire Erzgebirge region. Depending on thealtitude, 960 to 1160 mm precipitation fall annually and the study area receives an average annualprecipitation of 1053 mm.

3 METHODOLOGY

Based on the contrast of their mineral contents and metal concentrations, nine samples were selectedfor the procedure. They were freeze dried. To avoid any contamination, a polyethylene spoon wasused. They were ground in an agate mortar till to a size of ≤ 63 µm and homogenised. A seven stepsextraction procedure was used following the Salomon & Forstner (1984) findings. The leachingscheme and reagents are illustrated in Figure 2.

452

Figure 2. Sequential extraction – scheme.


The total cadmium content ranges from 0.23 and 6.8 ppm in the sediments. Between 5 and 100%of this concentration is of lithogenous origin and most of the cadmium seems to be bound to thehost minerals such as hornblende and chlorite (Figure 3).

Further, although there is an appreciable affinity of Fe oxides for cadmium (Kinniburgh, 1976;Pickering, 1979; Cronan, 1976; Li, 1982; Tiller et al., 1984) at high pH values, cadmium wasnot found in association with the nodular hydrogenous fraction and this may be due to its highertendency for complexation with other species other than Fe and Mn oxides and hydroxides.

Despite the absence of cadmium in the carbonatic phase in two samples, a high amount of its nonresidual fraction is found in association with the phase in the range ∼11 to 35%. This is due mostlyto the tendency of cadmium in its cadmium2+ ionic form to co-precipitate with CaCO3 (McBride,

453

Figure 3. Uranium tailings schneckenstein selective extraction procedures (cadmium Element).

Figure 4. The relationship between Sulphur and the cadmium amount bounds to the Sul./Org. phase(Cr = 0.86).

1994) and to precipitate in form of cadmiumCO3 as well (Brookins, 1988; McBride, 1994) at pHabove 7. These results show the high affinity of Cadmium to carbonates.

A considerable amount of the non residual fraction of cadmium is associated with the exchange-able phase varying from ∼19 to 100%. These findings are in agreement with that of Andersson,(1975); Stover et al., (1976); Salomons & Forstner, (1980) & Forstner, (1986) affirming the sig-nificance of cadmium in the mobile fraction which is due to the high selectivity of clays forcadmium2+.

In addition, at neutral to alkaline medium cadmium tends to precipitate into sulfide minerals oras CadmiumCO3 and to co-precipitate with CaCO3 as well. Its solubility under the above conditionsis low. Therefore not any cadmium is in connection with the pore water phase.

454

5 CONCLUSION

It may be concluded that most amount of the non residual Cadmium is associated to the high solublephases; carbonatic and exchangeable. Also, an alkaline chemical medium reduces the concentrationof Cadmium in pore waters. This is because at high pH values, cadmium tends to form carbonatecomplexes. As well, anoxic conditions reduce the solubility of cadmium. Under these conditions,it tends to form cadmiumS complex independent of the pH values.

REFERENCES

Andersson, A., (1975): Relative efficiency of nine different soil extractants. Swed. J. Agric. Res., 5, 125–135.Chester, R. (eds.). Chemical Oceanography. Academic, London. New York, 5, 217–263.Cronan, D. S., (1976): Manganese nodules and other ferromanganese oxide deposits. In : Riley, J. P. &Fernex, F. E. et al., (1986): Behaviour of some metals in surficial sediments of the northwest mediteranean

continental shelf. In: SLY, P. G. (ed.): Sediment and water interactions. Springer- Verlag.Forstner, U., (1986): Cadmium in sediments. In: Mislin, H. & Ravera, O. (ed.). Cadmium in the environment.

Birkhäuser- Verlag, Germany.Janssen, R.P.T., et al., (1997a). Equilibrium partitioning of heavy metals in duch field soils, II: prediction of

metal accumulation in earthworms. Environ. Toxicol. Chem. 16, 2479–2488.Janssen, R.P.T., et al., (1997b). Equilibrium partitioning of heavy metals in duch field soils, I: relationship

between heavy metal partition coefficients and soil characteristics. Environ. Toxicol. Chem. 16, 2479–2488.Kinniburgh, D. G. et al., (1976): Adsorption of alkaline earth, transition, and heavy metal cations by hydrous

oxide gels of iron and aluminium. Soil Sci. Soc. Am. J., 40, 796–799.Knight, B., McGrath, S.P., 1995. A method to buffer the concentrations of free Zn and Cd ions using a cation

exchange resin in bacterial toxicity studies. Environ. Toxicol. 14, 2033–2039.Krishnamurti, G.S.R. & Naidu, R., (2003). Solid-solution equilibria of cadmium in soils. Geoderma 113,

17–30.Laxen, D.P.H. & Harrison, R.M., (1981). The physicochemical speciation of Cd, Pb, Cu, Fe and Mn in the

final effluent of a sewage treatment works and its impact on speciation in receiving river. Water Res. 15,1053–1065.

Lee, F. Y. & Kittrick, J. A., (1984a): Elements associated with the cadmium phase in a harbor sediment asdetermined with the electron beam microprobe. J. Environ. Qual., 13 (3), 337–340.

Li, Y. H., (1982): Interelement relationship in abyssal pacific ferromanganese nodules and associated pelagicsediments. Geochim. Cosmochim. Acta., 46, 1053–1060.

Lu, C. S. J. & Chen, K. Y., (1977): Migration of trace metals in interfaces of seawater and polluted surficialsediments. Environ. Sci. Technol., 11, 174–182.

McBride, M. B., (1994): Environmental chemistry of soils. Oxford University Press, New York, USA, 406 pp.Meers, E. et al. (2007). Comparison of cadmium extractability from soils by commonly used single extraction

protocols. Geoderma 141 (2007) 247–259.Nriagu, J. O. & COKER, R. D., (1980): Trace metals in humic and fulvic acids from Lake Ontario sediments.

Enviro. Sci. Technol., 14, 443–446.Nriagu, J.O., (1996). A history of global metal pollution. Science 272, 223.Parker, D.R. & Pedler, J.F., (1997). Reevaluating the free-ion activity model of trace metal availability to higher

plants. Plant Soil 196, 223–228.Pickering, W. F., (1979): Copper retention by soil/sediment components. In.: Nriagu JO (ed) Copper in the

environment. Part I. Ecological cycling. John Wiley, New York, 217–253.Prokop, Z. et al., (2001). Mobility, bioavailability, and toxic effects of cadmium in soil samples. Environmental

Research 91 (2003) 119–126.Salomons, W. & Forstner, U., (1984): Metals in the hydrocycle, Springer Verlag, Berlin-Heidelberg, 349 pp.Salvarredyal-Aranguren, M. A. et al., (2008). Contamination of surface waters by mining wastes in the Milluni

Valley (Cordillera Real, Bolivia): Mineralogical and hydrological influences. Applied Geochemistry 23(2008) 1299–1324.

Tiller, K. G. et al., (1984): The relative affinities of Cd, Ni and Zn for different soil clay fractions and goethite.Geoderma, 34, 17–35.

Van Pejinenbur, W.J.G.M., et al., (1997). Conceptual framework for implementation of bioavailability of metalsfor environmental management purposes. Ecotoxicology and Environmental Safety 37, 163–172.

Wallmann, K., (1992). Solubility of cadmium and cobalt in a post-oxic or sub-oxic sediment suspension.Hydrobiologia. 235/236, 611–622.

455


Uranium residue impacts on ground and surface water resources at theSchneckenstein site in East Germany

T. NaamounFaculty of Sciences, Sfax, Tunisia

B. MerkelTU Bergakademie Freiberg, Germany

ABSTRACT: In the former uranium mining areas of Saxony and Thuringia there are still manysources of possible contamination due to uranium, thorium, radium and other heavy metals andmetalloids. Due to the increase concerns about the environmental pollution problems, it is soimportant in waste disposal management to perform an accurate chemical monitoring. This papersummarizes groundwater and superficial water analyses as well as geochemical modelling ofuranium and other pollutants which were undertaken in order to evaluate the contamination risk ofgroundwater in the vicinity of the Schneckentein site.

1 INTRODUCTION

Uranium mines discharge voluminous residues. Generally, they are impounded behind large dams,and when mining ceases, these impoundments are capped, revegetated, and left as permanent fea-tures on the landscape (Ritcey, 1989). Because the tailing impoundments are long-term repositories,the chemical processes occurring within them are of environmental significance because they gov-ern the chemistry of the discharges from the impoundments. Once the tailings have been capped,it becomes difficult to infer the nature of these internal chemical processes (Craw, 2003). Also,mining waste can continue to adversely affect the environment for centuries after the abandonmentof the mine itself, contaminating the soil in the immediate vicinity and also being dispersed byboth surface and ground waters (Sarmiento et al., 2009). Sharma, (1994) stated that mining ofmineral ores and disposal of resulting waste pose a significant risk to the groundwater. Accordingto Johnson and Hallberg, (2005) groundwater pollution from sources such as waste disposal sites,is a worldwide problem. In 1989, approximately 19,300 km of rivers and 72,000 ha of lakes andreservoirs throughout the world were severely affected by mining effluents. The long-term envi-ronmental impacts of mine tailing leachates have increasingly raised public concern. It is of primeimportance to assess their potential environmental risks and toxicity as well to establish a properpollution management plans (Lee, 2006).

In the southwest of Saxony in eastern Germany, the Mining Company WISMUT processedapproximately 1.2 Mt of uranium ore between opening in 1947 and final closure in 1957. TheSchneckenstein mill produced 660 tons of uranium at a capacity of 150.000 tons of ore per year.The resultant 1.96 Mt of tailings had been deposited in an elevated 108-ha two decantation basinsadjacent to the site. The total volume of the residue is about 700,000 m3 (Merkel, et al (1998).These tailings may act as a storage pool for toxic elements, which get into the surrounding riversby drainage and to the groundwater by infiltration.

In the nineties, after the German reunification, an extensive research program was undertakento improve the understanding of the geochemical evolution of uranium tailings and to developmore effective predictive and remedial techniques for the environmental problems associated withuranium tailings (Hurst & Glaser, 1998).

This paper presents results from investigation of groundwater and superficial water pollution atthe Schneckenstein site.

457

Figure 1. The area of investigation, uranium tailings, Schneckenstein.

2 THE SCHNECKENSTEIN AREA

2.1 Location

The Schneckentein tailings are located in the southwest of Saxony, in the Boda valley approximately3 km from the village of Tannenbergsthal county of Vogtland, Germany (Figure 1). The Boda valleyis bordered by the Runder Hubel (837 m a.s.l) and the Kiel (943 m a.s.l) to the Northwest andSoutheast respectively.

2.2 Geology

The area belongs to the transverse zone of the southern Vogtland and the western Ore Mountains;a subunit of the anticlinal zone of the Fichtelgebirge – Ore Mountains. The area of investigationis located on the southwest border of the Eibenstock granite. The Eibenstock granite could beconsidered a biotite-syenogranite and belongs to the group of the younger granites (JG1). It ismedium to coarse-grained, serial-porphyric and tourmaline bearing. Its intrusion is related to theUpper Carboniferous series. The Eibenstock granite is covered with a weathered surface layer.Southwest of the tailings follows the contact zone with quartz-schist.

2.3 Hydrogeology

The site of investigation lies in the area of the watershed between the Zwickauer basin and theEger basin. It belongs to the highest precipitation area of the entire Erzgebirge region. It receivesan average annual precipitation of 1050 mm. The Boda Bach is considered as the most importantReceiving River and all other streams flow into it and subsequently into the ‘Pyra’ river in theTannenbergsthal area (Figure 2).

3 EXPERIMENTAL

3.1 Field activity and sampling

Water samples were taken at several sampling points as indicated in the Figure 3; Table n◦1. Ehand pH were measured in the field with a pH/Eh-meter (WTW® pH 323) in a closed flow-cell.The pH-meter was calibrated with Merck buffer solutions (CertiPUR, pH 7.00 and 10.00). TheEh-meter controlled against a redox buffer solution (Mettler Toledo, Eh 439, accepted tolerance5%) and corrected for temperature and to the standard hydrogen electrode. Total alkalinity wasdetermined by titration (Hach® titrator). All samples were filtered (0.2 µm; regenerated cellulose),anion sample aliquots were stored at +4◦C in darkness prior to analysis. For cation analyses, samplealiquots were acidified to pH < 2 with HNO3 (Suprapure®) and stored frozen.

458

Figure 2. The water network, above and subterranean hinterland of the tailings sites (scale: 1:50.000).

3.2 Activities in laboratories and analytical methods

Due to the low permeability of most of the processed material, only the tailings cover was analysed.Thus, no more than three water samples were extracted by means of a high pressure device. Theelectrical conductivity, the redox potential and the pH values were immediately measured. Alsothe samples were filtered by a 0.2 µm membrane filter as well stabilised by diluted nitric acid(1 ml acid/100 ml water) until a pH ∼ 2 and finally filled in polyethylene bottles then stored in arefrigerator. The metal contents were determined by means of ICP-MS equipment. Their detectionlimits are given in the table n◦2.

3.3 Hydrochemical model

The computer program PHREEQC (Parkhurst, 1995) windows version was used. The software isdesignated for the simulation of a wide range of geochemical reactions including mixing of water,dissolving and precipitating phases to achieve equilibrium with the aqueous phase and effectsof changing temperature. Also, it indicates mineral species and provides estimates of elementconcentrations which had not been determined analytically as well as of molalities and activitiesof aqueous species, pH, pe and saturation indices.

459

Figure 3. Geographical water samples positions.


4.1 Surface water and groundwater

4.1.1 pH-valueFor most surface waters, the lowest pH values were recorded during the winter season. This ismay be due to further acidification of rain waters caused by air pollution attributable to sulphurdioxide and nitrogen oxide. A remarkable increase of pH is recorded after the winter. Independent

460

Table 1. The geographical position of the sampling points.

Sampling point Coordinate (Bessel) Altitude (m; a.s.l) Description

PNP1 4532825 815 Hanggraben5587620

PNP2 4532825 825 GWM15587590

PNP3 4532440 765 GWM25588360

PNP4 4532530 741 GWM35588510

PNP5 4532490 752 Spring5588450

PNP6 4532680 772 Drainage IAAI5588430


PNP8 4532590 748 Drainage IAAI*5588480

PNP9 4532560 745 Drainage IAAI*5588485

PNP10 4532540 745 Drainage IAAI/Spring*5588480

PNP11 4532510 747 Drainage IAAI/Spring*5588470


PNP13 4532540 740 Bodabach “middle-sprint”5588540

PNP14 4532580 732 Bodabach “lower reaches”5588580

PNP15 4532660 780 Drainage IAAII5588180

PNP16 4532680 805 Influx Hanggraben5587940

PNP17 4532920 774 Bach Vergleichstal5588660

PNP18 4532670 722 Spring5588670

PNP19 4532630 810 Spring5586890

PNP20 4532555 735 Spring5587535

PNP21 4532845 693 Spring5588240

PNP23 4532725 753 Hanggraben5587375

of the season, drainpipes supply near neutral waters. In addition, the localisation of the dischargeof the drainage water is well demonstrated by near neutral pH throughout the year at the samplingpoint PNP14 (Bodabach “lower reaches”), where the water-mixing between rain and drainage wateroccurs. The measured pHs (pH < 5) at the two groundwater sampling points (PNP2, PNP3) indicatethe mobilisation of heavy metals such as Mn, Ni, Zn, and Al.

4.1.2 Electrical conductivityAt the sampling point PNP1 (Hanggraben), the average EC is around 80 µS/cm. Thus, the mentionedwater is classified after Hutter (1990) as rain water; snowmelt. At the other surface water sampling

461

Table 2. Detection limits of each element.

Element Limit of Detection(µg/l)

AlAs 10Ba 1Cd 1Co 1Cr 1Cu 1Fe 100Mg –Mn 1Ni 1Pb 1Th 1U 1V 1Zn 10

points (PNP5, PNP16, PNP17 and PNP18) the recorded EC is slightly more and the average ECvaries between 120 µS/cm at PNP17 to 160 µS/cm at PNP5. These waters are classified as poorlymineralised superficial waters (Hutter, 1990). The drainage waters show a remarkable increase inthe EC values. The recorded average EC vary from ∼215 µS/cm at the sampling points PNP10(pipe 7) and PNP15 (drainage IAAII) to ∼690 µS/cm at the sampling point PNP8 (pipe 4). Thesewaters are slightly mineralised (Hutter, 1990). At the surface water sampling point PNP14, theaverage EC is ∼250 µS/cm and the analysed water is also classified as slightly mineralised. Thisis probably due to the water mixing between the surface water and drainage water at the samplingpoint. The groundwater at the GWM1 (PNP2) indicate a low EC with an average of 45 µS/cm andis classified as rain water; snowmelt (Hutter, 1990). The GWM2 and GWM3 show an increase ofEC. It reached an average of 205 µS/cm and 155 µS/cm at GWM2 and GWM3 respectively. Theseground waters are slightly mineralised (Hutter, 1990).

4.1.3 Redox potentialAccording to Gotschalk, (1997), at all surface waters, the measured Eh values are below 400 mV.These conditions reduce iron, nitrate and manganese. However, high Eh values were recorded forsome of the sample points (PNP13, PNP14, PNP18, PNP20) all the year round (Kutschke, 1998).The values range from 422 mV at PNP 14 to 590 mV at PNP18. Therefore the reduction of nitrateand manganese are expected. At other sampled points (PNP1, PNP21, PNP23), the reduction ofiron is probable since Eh values are below 400 mV.

Using sample point PNP7, it was possible to monitor the level of iron reduction throughout thesampling period. The Eh value is below 400 mV the year round and iron is highly mobile (Gotschalk,1997). The groundwater samples show a variable redox potential: The Eh ranges from 277 mV atGWM (2; 3) to 547 mV at GWM 2 and from 195 mV at GWMIV to 590 mV at GWM 2 (Gotschalk,1997; Kutschke, 1998). Therefore depending on the climatic factors such as temperature as wellas on the season, the reduction of iron together with manganese and nitrate can occur for mostinvestigated waters.

4.1.4 Major components4.1.4.1 SodiumFor all sampling points, the recorded sodium concentrations by Gotschalk, (1997) and Kutschke,(1998) were nearly the same. At the Hanggraben (PNP1) as well at PNP19, the Na+ contentis comparable to that of the rain water of the Freiberg region. The average concentrations areequal to 2.4 and 2.9 mg/l at PNP1 (Gotschalk, 1997; Kutschke, 1998). At PNP19 the average

462

Na+ concentration is 3.7 mg/l (Kutschke, 1998). At the other superficial waters (PNP13, PNP15,PNP16, PNP17, PNP18) the Na+ content is slightly higher and its concentration lies between3.2 mg/l at PNP17 and 5.0 mg/l at PNP18 (Gotschalk, 1997). At PNP20 and PNP21, the averagesare 10.2 and 13 mg/l respectively. The variation in content is probably due to the slight weatheringof the silicates. The mixing water at PNP14, shows an increase in the Na+ concentration attainingan average of 19.6 mg/l.

In the drainage waters, the Na+ content increases remarkably. For instance at PNP9; 10; 11;12, its average concentrations ranges from 17.5 mg/l at PNP10 to 50.5 mg/l at PNP9, whereas atPNP6; 7; 8, it is considerably higher varying between 89.9 mg/l at PNP8 to 100 mg/l at PNP6. Thisis may be due to the drainage water contact with high soluble Na+ containing minerals like halite(NaCl). The relative high Na+ concentration is one of the factors that cause the increase of EC atthese points as mentioned previously.

At GWM1, the Na+ concentrations are very low with an average value roughly 1.7 mg/l. This isclose to that of rain water, whereas at GWM2 and GWM3, the average contents of Na+ are nearlythe same reaching 8.8 mg/l at the first mentioned point (GWM2) and 11.3 mg/l at the second one(GWM3). These values lie in the range of the Na concentration of the ground waters of the Freibergregion (Kolitsch, 1996).

4.1.4.2 PotassiumAlthough the potassium content is remarkably higher in the drainage waters, its measured values arevery close to each other and very low in most sampling points. The maximum average concentrationof 6.5 mg/l was recorded at PNP6. The average concentration of 0.4 mg/l recorded at GWM1 doesnot exceed the potassium content in rain waters and it is lower than the concentration in groundwaters estimated by Hutter, (1990). This is due to the low solubility of potassium containingminerals such as potash feldspar as well as due to the high adsorption capacity of the grounds forpotassium by cation exchange.

4.1.4.3 CalciumThe average Ca concentration is almost the same in most superficial waters varying between7.2 mg/l at PNP1 and 12 mg/l at PNP18. It is slightly more at the GWM2; 3 reaching 14.5 and12.4 mg/l respectively. However, at GWM1 it is close to 3 mg/l which coincide with the Ca levelin the rain waters. In drainage waters it is remarkably higher, ranging between 29 mg/l at PNP7and 71.3 mg/l at PNP9. This is probably due to the weathering of plagioclase. At PNP14; 16, theaverage Ca contents are 26 and 18 mg/l respectively and this is may be due to the mixing of waterwith different Ca contents. And Ca is the main cation in all waters.

4.1.4.4 MagnesiumIn most surface water and groundwater samples, its average content lies under or slightly exceedsits maximum concentration in the rain water. It varies between 1.5 mg/l at the GWM1 and 6.1 mg/lat GWM2. In drainage waters, it is relatively low, ranging between 4.1 mg/l at PNP10 and 20.8 mg/lat PNP9.

4.1.4.5 SulphateIn the surface waters, the concentration varies from 17 mg/l at Hanggraben (PNP1) to 80 mg/l atPNP14. At GWM1, these values lie in the range of the sulphate content in the rain water in theFreiberg region. It doesn’t exceed 4.3 mg/l (Winkler, 1998). At the other two points; GWM2 andGWM3, the recorded average concentrations are 70.2 and 45.8 mg/l respectively. These two valuesare in the range of its content in the groundwaters of Freiberg and its environs (Baacke, 1995). Indrainage waters, the concentrations are remarkably higher and this is probably due to the solubilityof some sulphide containing minerals such as iron sulphides. The recorded mean values in thesepoints vary between 55.7 (PNP10) and 330 mg/l (PNP6).

4.1.4.6 NitrateThe measured values vary between 2.2 mg/l (PNP8) and 11.5 mg/l (GWM1), whereas at PNP7, theconcentration doesn’t reach its limit of detection (0.5 mg/l). All these values lie in the range of itscontent in the rain waters of Freiberg and its surroundings (Winkler, 1998).

463

Figure 4. Classification of water quality using piper diagram according to the data of Gotschalk (1997).

4.1.4.7 ChlorideAt all sampling points, the chloride content is very low and lies in the range of rain waters (Scheffer& schachschabel, 1992). In superficial waters, the average concentration ranges from 2.6 mg/lat PNP1 to 6.0 mg/l at PNP14, whereas in drainage waters, it is slightly higher varying between5.2 mg/l (PNP10) and 14.5 mg/l (PNP8). In groundwaters, the concentration lies between 2.1 mg/lat GWM1 and 6.9 mg/l at GWM2.

4.1.4.8 FluorideThe average content varies between 0.1 mg/l (PNP1; 13) and 1.3 mg/l (PNP7). At PNP2, 6; 15; 16;17; and 18 the concentration was below detection limit. However, at PNP9 the concentration is2.5 mg/l.

4.1.4.9 Hydrogen carbonateThe average content of hydrogen carbonate in the area is low in most analyzed samples. In thesuperficial waters, it varies between 18.3 mg/l (PNP5; 18) and 53.7 mg/l (PNP14), whereas inground waters it ranges from 21.4 mg/l (GWM2) to 24.4 mg/l (GWM1). In the drainage waters itis clearly higher, varying from 21.4 mg/l at PNP15 to 139.3 mg/l at PNP8.

4.1.4.10 Classification of water quality using piper diagramBased on the trilinear diagram and according to the data recorded by Gotschalk, (1997), watersamples from locations PNP6, PNP7 and PNP8 belong to the same water type “alkaline water”with prevailing sulphate-chloride (Figure 4). In addition, according to Morgan & Winner, (1962)and Back, (1966), the above waters are of sodium or potassium type. On the other hand, so faras the potassium and chloride contents are negligible with regard to that of sodium and sulphaterespectively, these samples may be classified as alkaline waters with prevailing sulphate-sodium.

On the basis of the two classifications, GWM1 belongs to the water type with prevailingbicarbonate. This is due mostly to the reaction of carbon dioxide with carbonates.

464

According to Morgan &Winner (1962), the PNP15 is classified as normal earth alkaline waterwith prevailing sulphate.

Also, almost all other samples belong to the earth alkaline water with prevailing sulphate andchloride. As well, according to Morgan & Winner (1962) and Back (1966), the samples fromlocations PNP15, PNP16, PNP17 and the PNP18 may be classified as calcium sulphate type.

It is noteworthy to point out that almost all samples show low salinity as demonstrated by thelow chloride content in these samples.

Generally, the data recorded by Kutschke, (1998) shows the same trend for most analyzedsamples.

4.1.5 Trace Metals4.1.5.1 AluminumAt GWM1: 2 as well as at superficial waters, under the recorded pH and Eh values, aluminium ishighly mobile. At drainage waters, most of the aluminium is in the form of the Al(OH)3 compound.

The recorded Al average concentration varies between 0.2 mg/l at PNP10 and 1.50 mg/l at PNP8.However, at PNP21 and PNP23 the Al concentration lies in the range of Al uncontaminated waterwith values of 0.11 and 0.05 mg/l respectively.

4.1.5.2 ArsenicIn surface waters, the average arsenic concentration varies between 2.9 µg/l (PNP13) and 12.3 µg/l(PNP1). In groundwater, it ranges from 0.82 µg/l (GWM1) to 8.0 µg/l (GWM3). In the drainagewaters, arsenic is considerably higher at PNP7 and PNP8 varying between 4.4 µg/l at PNP12 and126 µg/l. At the sampling point PNP14, however, about 21 µg/l of arsenic was recorded and this isdue to the water mixing as mentioned above.

Based on Brookins, (1988), under the recorded pHs-Eh conditions, most of the arsenic ingroundwater is expected to be H2AsO4-.

4.1.5.3 BariumIn all analysed samples, the Ba content is very low. It varies between ∼16 µg/l (PNP1) and 29 µg/l(PNP4). According to Brookins, (1988), the recorded pH and Eh values favour the formation ofthe immobile BaSO4 compound.

4.1.5.4 CadmiumThe pH and Eh conditions of the study area favour the presence of cadmium on its mobile formaround the year. In groundwater samples, its concentration varies between 0.6 µg/l (GWM1) and2.3 µg/l (GWM2). At PNP17 and PNP18, it attains 7.7 µg/l and 4.4 µg/l respectively. At PNP6; 7;and 8 the recorded values range from 3.3 µg/l (PNP8) to 4.6 µg/l (PNP7).

4.1.5.5 ChromiumThe maximum recorded Cr concentration doesn’t exceed 5 µg/l. This is due to the pH and Ehconditions of the area that favours its presence in very insoluble form Cr2O3. Moreover, at the pHrange ∼5–13.5, most of chromium tends to crystallise to the mentioned compound.

4.1.5.6 CopperIn superficial water, the average copper content varies between 10 µg/l (PNP13) and 139.4 µg/l(PNP1), whereas in groundwater, it ranges from ∼8 µg/l (GWM1) to 93 µg/l (GWM2). In drainagewater, it has remarkable great variation between 23 µg/l at PNP10 and 345 µg/l at PNP6.

4.1.5.7 ManganeseAlthough, the Eh-pH conditions of the area favour the high mobility of Mn i.e. its presence in theMn2+ ionic form, the average manganese concentration is not important in all superficial watersamples. It varies between 26 µg/l (PNP18) and 156 µg/l (PNP15). In groundwater, it is alsorelatively low ranging from 88 µg/ l at GWM1 to 559 µg/l at GWM2 and the last value slightlyexceed the maximal Mn concentration recommended by the WHO, (1996). In drainage water, itincreases remarkably, attaining 732 and 1220 µg/l at PNP8 and PNP7 respectively. This is due

465

probably due to the solubility of Mn bearing minerals, whereas at other points it varies between44 µg/l at PNP10 and 495 µg/l at PNP9.

4.1.5.8 NickelThe pH-Eh conditions of the study area favour the presence of nickel in its mobile Ni2+ ionic form.In superficial water, Ni varies between 3 µg/l at PNP17 and 34 µg/l at PNP5; 14. In groundwatersamples, the values range from 3 µg/l (GWM1) to 51 µg/l (GWM2), whereas in drainage waters,it varies between 7 µg/l at PNP10 and 251 µg/l at PNP9.

4.1.5.9 LeadUnder the pH-Eh conditions of the area Pb tends to combine with sulphate or carbonate ions toform PbSO4 (Anglesite) or PbCO3 (Cerussite). In superficial water, it varies between 0.01 µg/lat PNP13 and 2.8 µg/l at PNP17. In groundwater, it ranges from 0.6 µg/l (GWM3) to 1.5 µg/l(GWM1), whereas at drainage water, it varies from 0.3 µg/l at PNP11 to 12 µg/l at PNP7.

4.1.5.10 ZincAt superficial waters, its average content lies between 77 µg/l (PNP1) and 664 µg/l (PNP17). Ingroundwater samples, it varies between 34 (GWM1) and 246 µg/l (GWM2), whereas, in drainagewater it ranges from 97 µg/l at PNP7 to 600 µg/l at PNP9.

However, according to Brookins, (1988), the pH-Eh conditions of the prospected area, most ofZn is expected to be in the ZnS (Sphalerite) immobile form.

4.1.5.11 ThoriumDue to the low solubility of thorianite (ThO2) and ability of thorium ion to be absorbed by particulatesurfaces, the solubility of thorium in natural waters is extremely low (Ivanovitch & Harmon, 1992).Therefore, the average thorium content is very low in all analysed samples varying between 0.02 µg/lat PNP11; 16 and 0.21 µg/l at PNP8.

4.1.5.12 UraniumIn superficial water, the average uranium content varies between 0.4 µg/l at PNP2 and 193 µg/l atPNP14. In groundwater, it ranges from 0.4 µg/l (GWM1) to 8.3 µg/l (GWM2). In drainage watersamples, it lies between 47 µg/l (PNP15) and 1254 µg/l (PNP8). These values emphasise the highmobility of uranium in the area and therefore a potential source of contamination.

4.2 Extracted pore water

4.2.1 pH ValuesThe three analyzed samples show a near neutral pH (Figure 5). It increases slightly with depth. Itvaries between 7.4 and 8.4 with a mean value of 7.9. These recorded values are close to that of thedrainage water. Also, the increase of pH is probably due to the increase of the carbonate contentwith depth. Thus, these values state that for the first five meters of depth, the analyzed water isin contact with only untreated heap material. Moreover, under these chemical conditions, most ofmajor and trace metals have a low solubility.

4.2.2 Redox potentialThe Eh value ranges from 406 to 430 mV with a mean value around 420 mV and a standard deviationof 12 mV. According to Wagman et al., (1982), the measured Eh–pH values favour the mobilityof arsenic in its HAsO2−

4 ionic form. Under the same conditions, cadmium and nickel are highlymobile onto their Cd2+ and Ni2+ ionic forms respectively (Wagman et al., 1982), whereas cobaltand chromium are immobile. Cobalt tends to form CoCO3 and Co3O4 complexes (Garrels & Christ,1965). Chromium tends to be in its very insoluble Cr2O3 form (Wagman et al., 1982). Also, underthese conditions, copper and zinc are high mobile. As said by Garrels & Christ, (1965), coppertends to precipitate onto its CuO immobile form. According to Wagman et al., (1982), most of theZn content should be in its immobile ZnCO3 and ZnO complexes but some zinc may be in its Zn2+ionic form.

466

Figure 5. The change with depth of pH values.

Figure 6. The change with depth of EC.

4.2.3 The electrical conductivityThe electrical conductivity values lie between 530 and 1014 µs/cm with a mean value closeto 717 µs/cm (Figure 6). Based on Hutter (1990), the analysed water is a good mineralisedgroundwater.

4.2.4 Major componentsThe hydrogen carbonate content lies between 24.4 and 42.7 mg/l with a mean value of 32.5 mg/land a standard deviation of 9.3 mg/l.

The fluoride concentration varies between 1.4 and 1.8 mg/l with a mean value of 1.6 mg/l and astandard deviation of 0.2 mg/l.

The chloride ranges from 3.2 mg/l to 7.7 mg/l with a standard deviation of 2.2 mg/l. The meanvalue of the nitrate content is 0.9 mg/l with a standard deviation of 0.4 mg/l.

The sulphate content ranges from 150 to 250 mg/l with a mean of 198 mg/l and a standarddeviation of 50 mg/l.

The sodium and potassium contents vary from 8.9 to 19 mg/l and from 2.7 to 13 mg/l with meanand standard deviation values of 14 and 7 mg/l and 5 mg/l for both elements.

The calcium and magnesium concentrations range 40 to 69 mg/l and 13 to 30 mg/l with meanvalues of 57 and 22 mg/l and standard deviation of 15 and 8 mg/l respectively.

467

4.2.5 Heavy and trace metalsThe aluminum concentration varies between 275 and 1030 µg/l with a mean value of 571 µg/l and astandard deviation of 403 µg/l. Some of its content seems to be in the form of aluminium-hydroxideparticles in the water solution despite the process of filtration. Thus, in spite of the low mobility ofaluminium under the recorded pH, its content is found to be high in all samples.

The arsenic concentration lies between 49 and 105 µg/l with a mean value close to 82 µg/l anda standard deviation of 29 µg/l.

Cadmium content decreases with depth varying between 6 and 19 µg/l with a mean value of11 µg/l and a standard deviation of 7 µg/l.

The manganese concentration is found to be high. It lies between 1319 and 4000 µg/l with amean value roughly 2806 µg/l and a standard deviation of 1364 µg/l. This may be a sign of chemicalreducing conditions. Also, it may reflect the presence of a high bacterial activity.

As expected, the nickel content is relatively high. It decreases with depth. It varies between 83and 149 µg/l with a mean value of 107 µg/l and a standard deviation of 36 µg/l.

The Lead content is slightly high. It ranges from 6 and 21 µg/l with a mean value of 11 µg/l anda standard deviation of 8 µg/l.

Uranium is under oxidizing conditions. Thus, its content is found to be tremendously high. Itincreases with depth. It varies between 55 and 949 µg/l with a mean value of 423 µg/l and a standarddeviation of 467 µg/l.

Under the recorded chemical conditions, barium tends to form BaSO4(S) (Brookins, 1988). Itsconcentration ranges from 153 µg/l and 186 µg/l with a mean value of 173 µg/l and a standarddeviation of 18 µg/l.

The cobalt content is low and nearly the same in the different samples and decreases with depth.It ranges from 37 to 47 µg/l with a mean value of 42 µg/l and a standard deviation of 5 µg/l.

The Chromium content varies between 5 and 17 µg/l with a mean value of 9 µg/l and a standarddeviation of 7 µg/l.

The copper concentration ranges from 108 to 186 µg/l with a mean value of 144 µg/l and astandard deviation of 39 µg/l. These recorded values assert the expected behaviour of Cu under theabove mentioned conditions.

The iron concentration decreases with depth. Its total content varies between 60 and 150 µg/lwith a mean value of 90 µg/l and a standard deviation of 52 µg/l. Also, the measured Eh values arein agreement with the iron species determination. And iron is under oxidising conditions.

As expected, the water solutions do not show any Zn enrichment. Its content varies between199 µg/l and 1370 µg/l with a mean value of 930 µg/l and a standard deviation of 637 µg/l.

4.3 Hydrogeochemical model

As illustrated in (Figure 7), for the three extracted pore water samples, the calculated Eh values areconsiderably lower than the measured ones. In addition, the tailing environment becomes underpost aerobic or reducing conditions with depth mainly at the interval of depth that coincide with thebeginning of the processed material. This is due probably to the presence of an appreciable amountof humic substances in the processed material on one hand, and to the absence of any supplying ofoxygen on the other. The considerable difference between the recorded and the calculated valuesis probably due to the influence of the climatic factors onto such measurements.

5 CONCLUSIONS

It may be concluded that for the first five meters of depth, the water is in contact with only untreatedheap material with a near neutral pH. Also the Eh-pH conditions favour the solubility of As, Cd andNi. Al, Mn, Ni, Pb and U contents were obviously high. However, the Eh value is the main factorcontrolling the solubility of Mn, Fe and U. The EC values are remarkably higher in the processedmaterial than in the heap material. The hydrochemical model PHREEQC shows high solubility ofmost elements. It also shows the change in chemical conditions between the heap materials andthe tailing sediments characterised by the considerable decrease of the Eh values with depth, whichindicates the change of the medium to post aerobic or anaerobic conditions.

468

Figure 7. The change with depth of Eh; Calculated and measured values.

REFERENCES

Baacke, D., (1995), Geochemie untertägiger Stofflüsse in Stollnwässern der Grube Freiberg. Unveröfftl.Diplomarbeit, Institut für Mineralogie der TU Bergakademie Freiberg.

Brookins, D.G., (1988). Eh-pH diagrams for geochemistry. Springer-Verlag, 175 pp.Craw, D., (2003). Geochemical changes in mine tailings during a transition to pressure–oxidation process

discharge, Macraes mine, New Zealand. Journal of Geochemical Exploration 80 (2003) 81–94.Garrels, R. M. & Christ, C. L., (1965). Minerals, solutions and equilibria. Harper and Rowley, New York,

453 pp.Gotschalk, S., (1997).Hydrogeologische Untersuchungen am Uran-Tailing Schneckenstein. Unveröfftl.

Diplomarbeit, Institut für Geologie der TU Bergakademie Freiberg.Hurst, S. & Glaser, K., (1998). Uranbergbausanierung-eine Herausforderung an die Fachgremien, Uranium

Mining and Hydrogeology II, Proc. Of the Intern. Conference and Workshop, Freiberg, Germany, VerlagSven von Loga, Köln.

Hutter, L., (1990). Wasser und Abwasseruntersuchung. Laborbücher Chemie, Diesterweg/Salle, Frankfurt amMain, Germany, 511pp.

Ivanovitch & Harmon, R. S., (1992). Uranium series disequilibrium. Applications to Earth, Marine andEnvironmental sciences. Second edition, Clarendon Press, Oxford, 910 pp.

Johnson, D.B., Hallberg, K.B., (2005). Acid mine drainage remediation options: a review. Sci. Total Environ.338, 3–14.

Kolitsch, S., (1996). Hydrogeologische Untersuchungen in der Himmelfahrt Fundgrube. Unveröfftl.Diplomarbeit, Institut für Geologie der TU Bergakademie Freiberg.

Kutschke, S., (1998). Weiterführende hydrogeologische Untersuchungen an der Industriellen AbsetzanlageSchneckenstein. Unveröfftl. Diplomarbeit, Institut für Geologie der TU Bergakademie Freiberg.

Lee, S., (2006). Geochemistry and partitioning of trace metals in paddy soils affected by metal mine tailingsin Korea. Geoderma 135 (2006) 26–37.

Merkel et al., (1998). Natural leaching of uranium from the Schneckenstein Uranium mine Tailing, UraniumMining and Hydrogeology II, Proc. Of the Intern. Conference and Workshop, Freiberg, Germany, VerlagSven von Loga, Köln.

Parkhurst, D. L., (1995):. PHREEQC, a computer program for speciation, reaction-path, advective-transport,and inverse geochemical calculations. Water-resources Investigations report 95-4227. Lakewood, Colorado.

Ritcey, G.M., 1989. Effluent treatment for environmental control. Tailings management: problems andsolutions in the mining industry. Process Metallurgy Report, vol. 6. Elsevier, Amsterdam, pp. 411–574.

Sarmiento A. M., et al., (2009). Hydrochemical characteristics and seasonal influence on the pollutionby acidmine drainage in the Odiel river Basin (SW Spain). Applied Geochemistry 24 (2009) 697–714.

Scheffer & schachschabel, P., (1992). Lehrbuch der Bodenkunde. 13. Auflage, Ferdinand EnkeVerlag Stuttgart.

469

Sharma, R.S., (1994). Some aspects of liquefaction of tailings dams. MSc thesis, Imperial College of Science,Technology & Medicine, London, University of London.

Wagman, D. D. et al., (1982). The NBS tables of chemical thermodynamic properties. Selected values forinorganic and C1 and C2 organic substances in SI units. J. Phys. Chem., 11 (2), 392.

Winkler, C., (1998). Verfolgung des vertikalen Migrationsweges und Bilanzierung ausgewählter Elementeentlang der mineralisierten Gangzone des “Schwarzen Hirsch Stehenden Nord”. Unveröfftl. Diplomarbeit,Institut für Mineralogie der TU Bergakademie Freiberg.

470

ANDMINE

WASTE‘10

Tailings and Mine Waste ’10 contains the contributions from the thefourteenth annual Tailings and Mine Waste Conference held byColorado State University of Fort Collins, Colorado in conjunctionwith the University of Alberta and the University of British Columbia.The purpose of these conferences is to provide a forum for discussionand establishment of dialogue among all people in the mining industryand environmental community regarding tailings and mine waste.

Tailings and Mine Waste ’10 includes over 40 papers which presentstate-of-the-art papers on mine and mill tailings and mine waste, aswell as current and future issues facing the mining and environmentalcommunities. This includes matters dealing with technical capabilitiesand developments, regulations, and environmental concerns.

TAILINGS

TA

ILIN

GS A

ND

MIN

E W

AST

E ‘10

an informa business

cover_hinshaw10_LG.qxd 13-09-2010 15:33 Pagina 1