kegs pdac symposium 2014 booklet preliminary · geophysical results and the geological knowledge...

KEGS Geophysical Symposium 2014

Integrating Geophysics and Geology

Saturday, March 1st, 2014 Intercontinental Hotel, Toronto

Delegate program generously sponsored by:

We thank the symposium sponsors

Continental Breakfast

Reception

Morning Coffee Break

KEGS PDAC Symposium 2014 - Program Integrating Geophysics and Geology

Saturday, March 1st, 2014, Intercontinental Hotel, Ballroom A,

225 Front Street West, Toronto

Schedule

Oral Program

Time Presenters Title 8:00 Continental Breakfast and Poster setup 8:40 Edna Mueller-Markham,

President of KEGS Opening remarks

Physical Properties Chairs: David Hatch, Gedex, and Edna Mueller-Markham, PGW

8:45 Greg Hodges, CGG, and Laurie Reed, Reed Geophysical Consultant

A user’s guide to integrating geophysics and geology

9:10 Bernd Milkereit, University of Toronto Petrophysical and seismic Data - the perfect combination for the integration of geology and geophysics

9:35 Mark Shore, Magma Geosciences

The unifying links between geophysics and other earth sciences

10:00 Vince Gerrie, DGI Geoscience Physical properties - The quantitative link between geophysics and geology

10:25 Coffee Break and Poster Viewing Business Perspective

Chairs: Alan King, Geoscience North, and Rob Harris, Geonics 10:45 Ken Witherly, Condor Consulting The importance of bringing geology and

geophysics together to build mineral system models

11:10 Hernan Ugalde, Paterson, Grant & Watson

Case studies in integrated geological and geophysical 3D modelling: Value added to exploration and mining projects

11:35 Richard Osmond, Globetrotters Resources

Exploration success – A business model and case study

12:00 Nasreddine Bournas, Geotech An integrated approach for assessing the mineral potential of the Liptako Metallogenic Province (East side of the Fleuve Niger)

12:25 Katherine McKenna, GPX Surveys Introduction to the Ground Geophysical Survey Safety Association (GGSSA)

12:30 Lunch and Poster Viewing Mining Case Histories

Chairs: Luise Sander, SGL, and Claire Samson, Carleton University 1:15 Taronish Pithawala, Geosoft Geologically-constrained magnetic inversion at

the Great Whale Iron Property 1:40 Mathieu Landry, Glencore Geophysics & geology; an inevitable match for

success – Examples from Raglan Mine

Posters

• Discussion of the Terraquest fixed-wing XDS VLF system – Implications for uranium exploration in sandstone basins Jeremy Brett, MPH Consulting, and Peter Walker, GeoAlgorithms

• Airborne multi-sensor geophysical mapping of 3D geology

Brenda Sharp, Jurriann Feijth, Jean Lemieux, and Asbjorn Christensen, CGG

• Helicopter AFMAG (ZTEM) survey results over epithermal gold and gold-skarn deposits in the Guerrero Gold Belt, Mexico Jean Legault, G. Plastow, C. Izarra, S. Zhao, Geotech, and G. Kearvell, Newstrike Capital

• Integration of geology and deep DCIP/MT technology on the Junior Lake property, Northwestern

Ontario Darcy McGill, Michele Tuomi, and Mehran Gharibi, Quantec

• Optimization of ground electroprospecting survey techniques for mining exploration Igor Ingerov, AGCOS, and E. Ermolin, National Mineral Resource University

• The use of reduction to the pole versus total gradient for the interpretation of aeromagnetic data:

pros and cons Martin Bates and Marianne McLeish, Sander Geophysics

• VPMA methodology: dating a rock from its magnetic anomaly only

Renato Cordani, Reconsult

2:05 Julie Palich, Caracle Creek International Consulting

Revitalizing near mine and regional exploration, Granduc Property

2:30 Bill Morris, McMaster University Geophysical Processing and Interpretation with geological controls: Examples from Bathurst Mining Camp

2:55 Coffee Break and Poster Viewing Integration Tools

Chairs: Michel Chouteau, École Polytechnique de Montréal, Elizabeth Baranyi, Geosoft 3:15 Desmond Fitzgerald, Intrepid

Geophysics Reducing ambiguity in 2 & 3 D geology objects derived from potential field observations

3:40 James McNae, RMIT University Using implicit geological constrains in 3D EM modelling

4:05 Peter Lelievre, Memorial University Newfounland

Choices for effectively incorporating geological constraints into geophysical inversion

4:30 John McGaughey, Mira Geoscience A framework for the quantitative integration of geological and geophysical data

ORAL PROGRAM

A User’s Guide to Integrating Geophysics and Geology Abstract for KEGS 2014 By Greg Hodges, Laurie Reed There always has been, and still is, a problem integrating geophysics into geology: getting geologists to use geophysics and getting geophysicists to use geology in their interpretation, and express their results in terms that geologists can use. Not enough geologists have a good understanding of what geophysics measures, how to understand the geological meaning of the measurements, and how to use them in an effective exploration project. But it’s not enough to say that geologists need to learn geophysics: many geophysicists often do not think (or communicate) in geological terms. The two groups must learn enough of each other’s fields to ensure overlapping expertise, exchange of information, and full integration. But as the minority, geophysicists need to facilitate the communication by understanding geology and the needs of the geologists.

Rock properties are the indispensable link between geophysics and geology. To use geophysics explorers must be able to make the connection between geology and the changes in rock properties. Geophysicists must know the rock property measurements inherent in their data, and all explorers should know the rock types and changes that affect the properties, so they can see the geology that is apparent in the geophysical data. Every explorer should know what can (and cannot) be distinguished by magnetic susceptibility, electrical resistivity, density, velocity, and radioelement concentration. It’s about the rock property measured, much more than the method of measurement.

Going into the exploration project, explorers need to predict how the targeted changes to lithology and structure will change the rock properties in a way that might be measurable. The “Four Cs” of geophysical detection need to be considered: is there enough Concentration and suitable rock property Contrast; what is the effect of Cover; and what else might Confuse the interpretation? Geophysicists need to understand the target and host geology, and all the factors that will affect these detection factors: property contrast, concentration relative to the method footprint, the amount and effect of cover, and both geological and ambient noise. Interpretation of the geophysical results should be a process of interpreting the data to determine the rock property changes that exist, and from that predicting the geological changes that caused them. The best interpretations are focused on geology and include the known geology: “What geology caused this change to the geophysical data?” The results should be expressed - and presented - in a style as close to geological as possible. The geophysicist has to ask whether the results will be comprehensible to the geologist, and, if not, provide the proper interpretation. Most of us know of many times when IP pseudo-section “pant-leg” anomalies were targeted, wasting money and credibility.

Geophysicists also have to consider the planned use of the data – will the model provide the information that is needed to follow up? The most useful model is not the fanciest, but the one that best fits the geology and provides the required geological information most cost-effectively. Often, the model still needs interpretation to be understood in geological perspective. The geophysical work is not done when rock property model is developed, even if the ambiguous model-to-data correlations are resolved. The real integration of geology comes when converting the rock property model to geology – lithology and location and alteration state. This stage of interpretation requires a solid understanding of influences on rock properties, the relevance of structure to the target, and economic geology; and most of all, knowledge of the local rocks. Multiple rock property measurements from multiple geophysical surveys provide more than one look at the geology. However, these measurements need to be integrated into a single interpretation. There is only one “geology” at any location, so differing or conflicting geophysical interpretations from different data sets only show that the process is incomplete. A solid understanding of the interaction between the geophysical measurement and geology is needed to understand the different pictures each method map paint of the geology, and how to rationalize or overlap these. Where different methods present different pictures of the geology, it suggests that they are sensitive to different parts of the geology – obviously different rock property changes, but also different depths or layers in the geological column. A rock layer may be transparent to one method, and control the signal on another. Different rock properties may change at different places across a transitional alteration zone. The final Interpretation should integrate all the geophysical results and the geological knowledge into a single, cohesive, interpretation of the geology. It is critical to understand the limitations of each geophysical method and the geological data. Geophysics maps rock properties well over an area, but doesn’t identify the rock. Geology identifies the rock at the sample point (hopefully!), but not between the outcrops. The two characteristics are complementary, if used well. Rock properties are the link between them. Not until both stages are complete - geophysics to rock property, rock property to geology - is the interpretation complete. Effective use of geophysics is a matter of understanding the effect of geological processes on rock properties, and hence the effect on the geophysical data. It requires combined interpretation of multiple geophysical data sets into multiple rock property models, into a single geological description, integrating the available geological information. Geologists need to understand the effect of geological changes on the rock properties, and geophysicists need to think in terms of rock when interpreting their data. No geophysicist should graduate into exploration without a good education in geology, and no geologist should graduate without exposure to applied geophysics.

Petrophysical and Seismic Data - the Perfect Combination for the Integration of Geology and Geophysics Bernd Milkereit*, Ramin Saleh University of Toronto, Toronto, ON, CANADA Summary Densities and compressional (Vp) and shear (Vs) wave velocities provide a robust framework for the lithological identification and interpretation of most common sedimentary, igneous and metamorphic rocks. The well-known Nafe-Drake curves for silicate rocks show many ores have adequate impedances contrast compare to their common hosts, suggesting the feasibility of detecting ores using high-resolution reflection techniques. Generally to solve a scattering problem, modeling approach is the key and this is only possible if a geological model can be translated into a petrophysical model. Examples of comprehensive petrophysical databases for seismic imaging will include permafrost settings, gas hydrate reservoirs and massive sulfide deposits. More recently a 3D petrophysical model of deep mines become available, that can be used for geotechnical and monitoring purposes. Introduction Although the seismic reflection method showed a great success in oil and gas industry, it was not conventionally in use for mining purposes. The main reason was the success of electromagnetic and potential field methods as a cheap and an easy method for exploration purposes. In the past decades the demands for deep resources increased and that required a use of seismic methods for deep mine exploration (Debicki, 1996). 2-D and 3-D surveys have successfully detected and imaged large massive sulphide deposits such as the magmatic and volcanic massive sulphide (VMS) deposits. In addition, other types of deposits have a better chance to be imaged indirectly. Examples are exploration of lode gold and porphyry deposits by reflections from alteration haloes, unconformity Uranium deposits by haloes and basement offsets, and Mississippi Valley-type (MVT) deposits by white spots in otherwise reflective carbonates (Salisbury&Snyder, 2007). Most ore deposits are often fairly small (<1 km across), they commonly appear as diffractions rather than reflections on seismic reflection profiles. In addition, structures in hardrock environments are often complex and steeply dipping, that needs use of sources with frequency of 100 Hz and above and high fold surveys to be resolved. As a result modeling studies have a significant role in optimizing the design of survey and also in interpretation. The main reason is the complex patterns of diffraction from dipping reflectors, which can arrive at unusual locations at

the surface. The modeling of elastic seismic waves requires detailed 3D petrophysical data. Petrophysical information is obtained through lab core samples and in situ borehole logging measurements. Note that, compiling borehole-logging data into a 2D/3D petrophysical model mostly involves intense statistical analysis. Upon the availability of such petrophysical models there are many possible applications. For example, forward seismic modeling can be used to study resolvability of different geology settings, or seismic imaging techniques can help to design more efficient seismic surveys. Also inverse methods can be implemented to reverse time migration problems or AVO studies can be used for reservoirs characterization.

T=0.070 s S−waves

Distance [m]

Dept

h [m

]

200 400 600

100

200

300

400

P−waves

Distance [m]

200 400 600

100

200

300

400

−4000 −2000 0 2000 4000 −4000 −2000 0 2000 4000

Seismic Geophysics

P−wave model

Distance [m]

200 400 600

100

200

300

400

5000 5500 6000

S−wave model

Distance [m]

200 400 600

100

200

300

400

2500 3000 3500

Density model

Distance [m]

200 400 600

100

200

300

400

3000 3500 4000 4500

Petrophysics

Geology

Figure 1: Geophysics, and geology can be linked to each other trough petrophysics information. Contrasts in the elastic-wave velocities and densities of rocks are the most important physical properties influencing the overall seismic response of an orebody. The wave velocities depend on the in-situ elastic properties, which in crystalline metamorphic and igneous rocks are controlled by mineralogical content, damage, stress, in-situ fractures, pressure, and saturation. The elastic properties of core samples of rocks at standard conditions usually will not be representative of those for the same material at depth within the earth. Elastic-wave and density geophysical wireline logs provide important information to constrain the in-situ elastic properties, but a practitioner must be aware of what effects can influence the final interpretation of the logs (Salisbury et. al., 2003). Physical properties of rocks may vary in depth and can show some complexities, which should be taken into account when interpreting seismic reflection data. Stress changes in a formation affect its porosity. Stress-induced fractures will affect seismic wave propagation. The fractures, in general, reduce the stiffness of the rock matrix, thus a decrease in seismic wave velocities is expected. Seismic data is also sensitive to anisotropy effects due to

different orientations of cracks in the rock. It is expected that an increase in stress will lead to changes in P-wave velocity by 10% and that can be observed in travel time measurements. The problem is more complicated if the attenuation is present in the medium, for such a case the porosity and compressibility petrophysical information are also needed (Table 1).

Compressional & Shear wave velocity and Density

Shear & Bulk Modules Porosity and Permeability

Attenuation Stress

Table 1: Important physical properties of rocks and other effects influencing the overall seismic response of an ore-body. It is also important to evaluate the limits of seismic method to image the target in both horizontal and vertical direction. The main factors that control the resolution of seismic reflection method are the diameter (d), thickness (t) and depth (z) of target and also the average velocity of the formation and dominant frequency of the source, which is used in the survey (Salisbury et. al., 1996). The resolution is decreased with depth of investigation and increased as dominant frequency of the survey increased. Thus based on application and scale of targets deal with in seismic method, it is needed to use different tools working in different range of frequencies (Figure 2).

Figure 2: Resolution of different geophysical methods Method With the recent availability of detailed 3D petrophysics models of complex settings, in addition to the development of efficient numerical techniques and parallel computation facilities, a solution for the propagation of seismic waves for very strong elastic contrasts and high frequency sources is achievable. In order to model the effects of high elastic contrasts on the propagation of seismic waves within a 3D model, a full elastic/viscoelastic finite difference time-domain modeling package is used [Bohlen, 2002]. Also some examples illustrated using spectral element method (Komatitsch, D., and J. Tromp, 1999).

Underground mine model In a mine, the presence of very strong elastic contrasts, such as massive orebodies, tunnels, stopes and infrastructure have a significant impact on the propagation of seismic waves. For such heterogeneous medium travel time and amplitude of seismic waves derived from the conventional constant velocity models are inadequate. Since the complexity of seismic wave propagation can affect the distribution of energy significantly, the use of a more accurate model is required to predict the ground motion. For example, the conventional empirical method used to calculate peak particle velocities and accelerations (PPVs/PPAs), tends to underestimate the intensity of seismic waves in stopes or areas close to blast sites, which could be corrected if a more realistic model was implemented. The results show the complexity of scattered seismic waves due to effects of strong elastic contrasts will be illustrated using a 2D/3D finite difference modelling method (figure 3).

Figure 3: The complex scattered waves from a concret lens diffractor. Both amplitude and travel time of the wave are impacted. This very heterogeneous medium results in a complexity of wave propagation with variations in amplitude, travel time, and phase. The significance of these effects strongly depends on the size, shape, petrophysical properties and the frequency content of seismic source. The modelling results verify the amplification effects at regions with high Vp/Vs ratio (such as voids or cemented backfill) for both P and S wave amplitudes. Sources of higher frequency content displayed changes in polarity and/or creation of shadow

zones. The observed complexity of the wave propagation suggests the necessity for further consideration of such effects in the determination of focal mechanisms and full moment tensor inversions. Conclusions It should be possible to detect and prospect for ores using high-resolution reflection techniques if the deposits meet the size, thickness, and presentation constraints required for reflection or diffraction. Due to the small size of most deposits, the structural complexity of hard rock environments, it is needed to carefully design surveys using high frequency sources to identify both reflections and diffractors. The modeling study plays an important role here, and to do such a study the key is the availability of petrophysical data to bridge between geology and geophysics. References Bohlen, T., 2002, Parallel 3-D viscoelastic finite difference seismic modelling. Computers and Geosciences, 28 (8), 887-889. Huang, J. W., J. M. Reyes-Montes, and R. P. Young. "Passive three-dimensional microseismic imaging for mining-induced rock-mass degradation."Rock Mechanics for Resources, Energy and Environment (2013): 135. Komatitsch, D., and J. Tromp, Introduction to the spectral-element method for 3-D seismic wave propagation, Geophys. J. Int., 139, 806–822, 1999. Salisbury, M., and D. Snyder. Application of seismic methods to mineral exploration. Mineral deposits of Canada: A synthesis of major deposit types, district metallogeny, the evolution of geological provinces, and exploration methods: Geological Association of Canada, Mineral Deposits Division, Special Publication 5 (2007): 971-982. Salisbury, M., Harvey, C., and Matthews, L. (2003) 1. The Acoustic Properties of Ores and Host Rocks in Hardrock Terranes. Hardrock Seismic Exploration: pp. 9-19. Salisbury, Matthew H., et al. "Physical properties and seismic imaging of massive sulfides." Geophysics 65.6 (2000): 1882-1889. Schmitt, D., Mwenifumbo, C., Pflug, K., and Meglis, I. (2003) 2. Geophysical Logging for Elastic Properties in Hard Rock: A Tutorial. Hardrock Seismic Exploration: pp. 20-41. Valley B., Milkereit, B, Pun, W., Pilz, M., Dusseault, M., Thibodeau, D.Forsythe, A. Rock mass change monitoring in a sill pillars at Vale’s Coleman mine (Sudbury, Canada). Proceedings of CIM, 8p., 2012 .

��

�� -�� H� !�

"��#��#�$�� #�%��&��#�� &��'�-��%�� &��#��#��#��&��&��#�� (��&��&��#��%��#�� %�%&��#��#��)��&�� &��&��&��#��&��#�� &�%&�� #��&��)��#�� &��%��#�#��& �� &��&�#��&��#��#��!��&��#��&��#��#��'�

-�� &&��&�� &&� �� &��#��#��&��&��#��%��#� ��&�� #�� #��& �� '�*��$� �&��#�#��#�� &�� &��&��&��#�� &��&&��&&��&��#��&��%�� &�&�+�&��#� ��&�!��#�� &&#�� &�� &�� '�"��&��&��&&��#��#��#��&��#��&��#��#�� &��#�� &&��&��&&��#�&��%&%#'�

-�� &�� #��#��#��%� ��&��&&�� &��&&�+��#��%��&��#��#��%��%� ��$'�� &�� &&�� &��&&��$��%��#��%��%�%��&��#��%��%�� #�� #�&�+��#��#��%� ��&�(�� &&��&��%�&��&��'�"��&��&�%��&�#��$�� &�(��&&��'�

-��&��&��#�#�� #�#��& ��&��&��&��%&� ��#�� #��%� ��&��&&��#��#��,��&��#��%�%�&��&��&&��#��#�$�&�� &&�+� �� &�'�� &��#��%��#�� %��+-��&��#�#��&��#��&��#�� #��#��)��#��.��#�� %�&��&��%��&��&��,��&��&��&�� &��#��'�

��%�� &#�� $��#��'�/�%�� #��&��%&��#%&�#��&��&��,��&��%��#��#��'�0��#��&��&�� % ��1+�� ,��%�� 23� #&��&��#�23��&� ��&��&��#��&��&&�� &��#��&�#��&�� #��%��+#��&��$�&�� #�'�

Physical Rock Properties – The Quantitative Link between Geophysics and Geology Vince Gerrie*, DGI Geoscience Inc. Chris Drielsma, DGI Geoscience Inc. Patrick Hooker, DGI Geoscience Inc. Roxanne Leblanc, DGI Geoscience Inc. Pamela Patraskovic, DGI Geoscience Inc. Summary Comprehensive physical rock property measurements provide the critical quantitative link between geology and geophysics, thereby improving 3-D geologic modelling. Machine learning, cluster analysis, and classical statistics are used to classify the rock into petrophysical (physical rock properties) domains and benefits of this classification scheme are discussed. Traditional lithological classification schemes, i.e. visual core logging, are compared to the rock property domains and methods to link geology to geophysics are explored. In addition, establishing relationships between petrophysical data and geometallurgical, geochemical, and geotechnical parameters to create proxy relationships are investigated and select case studies presented. Additionally, in-situ structural measurements combined with structural analysis can aid in the development of structural models. Understanding the distribution of fault zones and their relationship with mineralization, can lead to better exploration targeting, and decreased operational risk. Introduction Using in-situ physical rock property measurements to quantitatively link geology to geophysics is a standard practice in the oil and gas industry, but to date this approach has had limited adoption in the minerals industry. Although the technology exists to acquire high quality petrophysical data, it is not currently a routine part of the traditional mineral exploration and mining workflow. Innovative approaches aimed at maximizing value from petrophysical data are explored through select examples and case studies across a variety of geoscience disciplines such as geology, geophysics, ore delineation, geotechnical engineering, rock mechanics, and geometallurgy. In-situ physical rock property measurements have attributes (multi-parameter, high data density, quantitative) that are ideally suited to data driven approaches, however there are several challenges that arise when using petrophysical domains to link geology to geophysics. Four key issues are discussed with respect to petrophysical domains, linking

geology to geophysics, establishing proxy relationships, and in-situ structural measurements:

1. Data Quality and Consistency: Data driven approaches can only be effective if the input data is consistent and variation in the data set is not artificially induced via instrument drift, calibration errors, and changes in instrumentation over the life of a project etc.

2. Traditional lithological classification schemes rely on a potentially subjective visual inspection of drill core leading to inconsistent results over the lifetime of a project. There is often a lack of correlation between traditional geologic classification and the petrophysical data. A high degree of variance within each lithological unit can often be found across the physical rock properties; such as magnetic susceptibility, density, induced polarization, resistivity, and so forth.

3. Many ex-situ measurements such as geotechnical, geochemical and geometallurgical parameters can be time and cost intensive over very small sample sizes. This often results in sparse and non-representative sampling, and analysis that lacks statistical relevance and accuracy into the geologic model. It can be challenging to relate continuous (densely sampled) petrophysical data to sparse ex-situ measurements.

4. Insufficient measurement and/or collection of structural orientation data. Although there has been an increase of using various core orientation techniques, they are limited by the application and core recovery.

Methodology Instrument Calibration and Data Validation In order to collect valid, consistent data ideally suited to quantitative analytical techniques, proper calibration procedures must be followed. These procedures include establishing a calibration borehole with known well-studied geology, lab measured physical properties and validated geophysical reference logs. Probes must be calibrated periodically at such a site.

Physical Rock Properties – The Quantitative Link between Geophysics and Geology

During collection, the probes must be checked daily for any drift or error that may occur, using either a locally established reference borehole or calibration / test jigs. A thorough process in regards to calibration and quality assurance is key in acquiring quantitative petrophysical data. Petrophysical Domains Since in-situ acquired physical rock properties are densely sampled multi-parameter datasets, they are suitable for analysis via machine learning, cluster analysis and classical statistical techniques to create quantitative petrophysical domains. These domains can then be analyzed and compared with the logged geology by using frequency analysis, principle component analysis (PCA), cross-correlation, and other statistical methods. Viewing and interpreting the domains in 3-D space allows improvements to be made to the existing geologic model and a greater project level understanding. Dataset Integration – Proxy Relationships Using machine learning techniques, geometallurgical, geochemical and geotechnical parameters can be correlated to in-situ physical rock properties to establish proxy relationships. These relationships allow proxy predictions to be made from in-situ rock property data to estimate values where no ex-situ measurements were conducted, or to get quick estimates for results that are traditionally slow to obtain in a lab. For example, geometallurgical milling parameters that are time and cost intensive to collect can be estimated in advance using downhole physical rock properties once a proxy relationship has been established. Structural Measurements In-situ structural measurements can be acquired using acoustic (ATV) and optical (OTV) televiewers. Each probe will generate a 360° image of the borehole wall. The ATV generates an acoustic pulse that reflects from the borehole wall back to a receiver; generating two images: the amplitude of the signal and the travel time. On the other hand, OTV uses a high-resolution digital camera to create a RGB image of the borehole wall. ATV is superior for identifying and classifying fractures, stress induced breakouts and fault zones, whereas OTV excels at identification of bedding/banding/foliation, veins and lithological contrasts. These images are processed, features are identified and classified using a standard scheme or one defined by project objectives. These features are oriented using tilt and azimuth measurements acquired by on-board motion sensors. Oriented features can be plotted on

stereonets or polar frequency graphs (rose diagrams) to analyze the trends in the data. Ultimately, the data can be interpreted and analyzed to create accurate 3-D structural models that define the structural controls on mineralization, and in conjunction with geotechnical logging, define geotechnical and rock mass domains. Conclusions Using physical rock properties to create petrophysical domains and proxy relationships allows for a greater understanding of the geologic setting. The petrophysical domains are an objective classification scheme that can help refine a geologic model. The proxy relationships increase the understanding in areas where there may not have been sufficient data collected in the past or assist in finding the most effective and efficient approach to getting answers project-wide, whether the questions are geophysical, geotechnical, geochemical or geometallurgical in nature. Oriented structural measurements acquired in-situ from the televiewers can lead to an increased comprehension of the structural trends of the environment and mineralization. The increased geotechnical knowledge provides better risk assessment; the structural knowledge better targeting; in combination a better geologic model.

Integrating geology and geophysics- The importance of bringing geology and geophysics together to build mineral system models Ken Witherly, Condor Consulting, Inc. Summary As the discovery of shallow, high grade deposits for essentially all commodities continues to decline (Schodde 2013, Doggett 2013), explorers are having to increasingly search for deposits at greater depth and with increased amounts of cover material. As a result, the direct detection of deposits becomes problematic and there needs to be a much greater reliance on secondary or tertiary signatures or halos of deposits to define the likely presence of the target. The mineral system approach is a means by which the overall environment in the earth which has been changed due to emplacement of a deposit in both space and time is characterized at a variety of scales. If we are able to understand the reasons for these changes and then systematically track them in the earth, explorers have a powerful new tool to identify deposits at depth. For this approach to be successful it will require a degree of integration between geology, geophysics and geochemistry not previously undertaken. Introduction The concept of a mineral system (Wyborn et al. 1994) was the outgrowth of similar work in the hydrocarbon industry dating from the earlier 1990s. In summary form, the key points were defined as:

1. Sources of the mineralizing fluids and transporting ligands

2. Sources of metals and other ore components 3. Migration pathway 4. Thermal gradient 5. Energy source 6. A mechanical and structural focusing mechanism at

the trap site 7. A chemical and or physical traps for ore

precipitation 8. Preservation (added later)

In graphic form, the main concepts are captured in Figure 1. Figure 2 shows a more ‘real world’ concept of the how the mineral system might actual work. Note, that in much of the discussion, the exact nature of the deep heat source is not explained in detail other than to invoke that such anomalies in the middle crust are required to drive the mineralization process in the shallow crust. From an exploration perspective, Figure 3 tries to capture and define some of the facets in practical scales. To help the various scales to be more readily conceived, the terms

Figure 1: Outline of mineral system.

Figure 2: Mineral system in ‘semi-real life’. target (or deposit), footprint and footpath are used to convey the area of actual deposit itself, the near-deposit environment and the extended halo that represents the pathway along which the mineralizing fluids passed during the formation of the deposit. Initial efforts by researchers (Wyborn et al. 1994) to define mineral systems involved

Building Mineral System Models

2

Figure 3: Mineral system with suggested scales of various facets of the system. the search for geological evidence but current efforts draw in geochemical and geophysical surveys as well. Regional surveys are seen as particularly important as they offer the best chance to see large-scale but possibly subtle responses that could be evidence of large changes in rock properties associated with paleo-thermal events. Examples Examples of the expressions are mineral systems is very much a work in progress, especially at the footprint and footpath scales. Much of the current work is empirical in nature as in most instances; direct observation of suspected signatures is impossible and understanding has to be developed through inference and laboratory studies. Unfortunately, even at the footprint scale, not a great deal of attention has been paid to rocks deemed outside the immediate ore zone, often due the cost of sampling. This means that geochemistry and geophysics are often the proxy’s which are available to try and understand the processes which have taken place. Figure 4 shows aeromagnetic coverage over what is interpreted to be a paleo-subduction zone in southern Alaska. Adjacent to this trend are a series of magnetic highs, one of which hosts the world class Pebble porphyry copper-gold deposit (Anderson et al. 2013). This is the suspected indication of what could be termed a mineral system footpath. Modeling of magnetic data over the Bingham deposit in Utah is shown in Figure 5 (from Steinberger et al. 2013). This work suggests that a mafic body several kms below the surface was the original chamber from which the mineralized stocks were derived. MT and seismic results over the Olympic Dam IOCG deposit in South Australia are shown in Figure 6 (after Hronsky 2011). A deep conductive plume extends below

Figure 4: Aeromagnetic coverage over southern Alaska.

Figure 5: Aeromagnetic image and 2.5D modeling of Bingham deposit, Utah. the deposit (warm colors) which as well, is seismically featureless. While appearing to be a ‘compelling’ indication of a deep-seated root system (aka footpath) for the Olympic Dam deposit, an assessment of similar surveys in the Gawler Craton over other similar deposits show what are deemed ‘mixed results’ (Hayward per. comm. 2014). This shows that while empirical results are vital, considerable more work is required to integrate field observations into a solid framework of predictive theory.

Figure 6: MT and seismic across Olympic Dam deposit, South Australia.


3

Evolution of Porphyry Copper Exploration; Linking Geology and Geophysics...Slowly To succeed, the minerals system approach will require the integration of geophysical observation with geological thinking at scales not previously attempted. As an indication of the challenges involved a short review of the development of how geophysics as applied to porphyry copper exploration is provided. The first commercial used of induced polarization as a technique to map porphyry copper deposits was detailed in Baldwin (1959) and shows survey results obtained in 1952 over the Caujone deposit in Peru (Figure 7). There is a clear mapping of the pyrite halo surrounding the central copper rich core; a style of alteration which would be later described almost 20 years later in Lowell and Gilbert (1970).

Figure 7: Chargeability results from Caujone deposit, Peru; 1952. In 1966, Rogers et al. (1966) a summary was published which showed a number of case studies using primarily the IP technique to explore for porphyry coppers deposits, largely drawn from the US south west. Following this paper however, there were basically no significant case studies presented by representatives of American companies or about US deposits, even though much of the exploration focus for porphyry deposits was in that region. During this time however (1960s-1970s), a number of US companies invested considerable time and effort at building specialized IP requirement and conducting proprietary studies on porphyry deposits and IP technology (i.e. Ware 1979, Halverson, et al. 1989 and Nelson 1997). The major case studies on porphyry coppers during that time were Hallof and Winniski (1971) on work in 1966-68 at the Lakeshore, AZ deposit, Fountain (1972) on the results of a number of IP surveys in BC in the 1960s, Pelton and Smith (1977) on IP surveys in the Philippines, Witherly (1979) on the Island Copper, BC deposit and Coolbaugh (1979) on the La Caridad, Mexico deposit.

The Pelton and Smith (1977) and Coolbaugh (1979) papers were especially important. In Pelton and Smith, the authors linked for the IP responses to Lowell and Gilbert (1970) and Gilbert and Lowell (1974). This was the first published reference to the alteration pattern diagnostic to many porphyry deposits and first observed in an IP surveys 1952, 25 years earlier (Baldwin 1959). The Coolbaugh (1979) paper is considered significant in that it is links the first reference found to using the Lowell and Gilbert style model to the discovery of a porphyry copper deposit. This was not actually defined in the cited paper but was noted in Gilmour et al. 1995 that the Asarco’s Mexican exploration group made effective use of the IP alteration model to help define the economic ore deposit at La Caridad. In 1982 (Witherly 2013), an IP-resistivity survey over the then undeveloped Escondida (Chile) porphyry copper deposit showed a strong resistivity low over the deposit area. Then in 1993 (Nickson 1993), during the course of conducting geophysical surveys over the Collahuasi area, northern Chile, it was observed that the chalcocite enrichment blanket was conductive enough to be mapped with a TEM survey. These results are shown in Figure 8. As a diagnostic response in porphyry copper exploration however, this attribute has not been formally defined in published literature, even though the initial field observations were made over 30 years ago.

Figure 8: TEM and DC-resistivity over Ujina deposit, Chile. Conclusions The mineral systems approach is intended to improve the chances to locate significant mineral deposits under greater depth of burial than has been required in the past. To succeed, there will be required a very close coupling between geophysical observations and geological modeling in order to define new approaches to exploration practice.


4

Based on the example presented for IP and porphyry copper exploration, the linking of geophysical observations to geological modeling and exploration practice can take a number of decades and many of the important developments even after >65 years of practice remain poorly documented. This appears to be as a result of both company secrecy and major changes in company structures over time which resulted in a loss of hard-won knowledge. References Anderson, E.D., Hitzman, M.W., Monecke, T., Bedrosian, P.A., Shah, A.K. and Kelley, K.D. (2013), Geological Analysis of Aeromagnetic Data from Southwestern Alaska: Implications for Exploration in the Area of the Pebble Porphyry Cu-Au-Mo Deposit; Economic Geology v. 108, pp. 421-436. Baldwin, R. (1959), Overvoltage Field Results; in Overvoltage Research and Geophysical Application; J Wait, ed. Pergamon Press 1959. Coolbaugh, D.F., (1979), Geophysics and Geochemistry in the discovery and development of the La Caridad porphyry copper deposit, Sonora, Mexico; in Geophysics and Geochemistry in the Search for Metallic Ores; Peter J. Hood, editor; Geological Survey of Canada, Economic Geology Report 31, p. 721-725, 1979. Doggett, M. (2013), The Challenge of Creating Value through Exploration; presented at ProExplo 2013, 19th-21st May, Lima, Peru David K. Fountain (1972) Geophysical Case Histories of Disseminated Sulfide Deposits in British Columbia Geophysics Feb 1972, Vol. 37, No. 1, pp. 142-159 Gilbert, J.M. and J.D. Lowell (1974), Variations in zoning patterns in porphyry ore deposits, CIM Bull. V.67, pp.99-109. Gilmour, P., Andrew, R.L., Bernstein, M., Maxwell L. and Morrissey, C.J., ( 1995), Porphyry Copper Deposits: History, Recent Developments, Exploration, Economics; in Pierce, F.W., and Bolm, J.G., eds., Porphyry copper deposits of the American Cordillera: Arizona Geological Society Digest 20, p. 128-155. Hallof, P.G. and Winniski, E., (1971), A Geophysical Case History of the Lakeshore Ore Body Geophysics Dec 1971, Vol. 36, No. 6, pp. 1232-1249. Halverson, M. O., Kingman, J. E.E, and Corbett, J.D., (1989) Advances in IP Technology: Telluric Cancellation and High Spatial Resolution Arrays p.183-191 in

Proceedings of Exploration '87: Third Decennial International Conference on Geophysical and Geochemical Exploration for Minerals and Groundwater, edited by G.D. Garland, Ontario Geological Survey, Special Volume 3, 960p. Hayward, N. (2014), Personal communication; February 5, 2014. Hronsky, J. (2011) Minerals System Thinking: An Emerging New Paradigm for Exploration Targeting; presented at AMEC Convention, 30 June 2011, Perth Australia. Lowell, J. D. and Gilbert, J.M. (1970), Lateral and Vertical Alteration-Mineralization Zoning in Porphyry Ore Deposits, Economic Geology v. 65, pp. 373-408 Nickson, R. (1993); Chile Porphyry Copper Geophysical Test Program; Quantec Ltd. internal report. Pelton, W.H. and Smith, P., (1976), Mapping Porphyry Copper Deposits in the Philippines With IP, Geophysics Feb 1976, Vol. 41, No. 1, pp. 106-122. Rogers, G. R. , Maillot, E. E., Sumner, J. S., Brant, A. A., Hansen, D.A., Barr, D. A., Hallof, P.G., Heinrichs, W. E., Ludwig, C. S. Lacy, R. J., Morrison, B. C., Forwood, P. S., Roberts, J. B., Seigel, H. O., Rogers, G. R., (1966) The Search for Disseminated Sulfides; in Mining Geophysics Volume 1, Case Histories General Series Editors: Don A. Hansen, Walter E. Heinrichs, Ralph C. Holmer, Robert E. MacDougall, George R. Rogers, John S. Sumner, Stanley H. Ward; Society of Exploration Geophysicists 1966. Schodde, R. (2013), Long Term Outlook for the Global Exploration Industry-Gloom or Boom?; presentation at the Geological Society of South Africa GeoForum 2013 Conference; 2nd-5th July 2013, Johannesburg. Steinberger, I., Hinks, D., Driesner, T., and Heinrich, C.A. (2013). Source Plutons Driving Porphyry Copper Ore Formation: Combining Geomagnetic Data, Thermal Constraints, and Chemical Mass Balance to Quantify the Magma Chamber Beneath the Bingham Canyon Deposit; Economic Geology v. 108, pp. 605-624. Ware, G.H. (1979) In�situ induced�polarization and magnetic susceptibility measurements—Yerington mine; Geophysics Aug 1979, Vol. 44, No. 8, pp. 1417-1428 Witherly, K.E., (1979), Geophysical and geochemical methods used in the discovery of the Island Copper Deposit, Vancouver Island, British Columbia; in Geophysics and Geochemistry in the Search for Metallic


5

Ores; Peter J. Hood, editor; Geological Survey of Canada, Economic Geology Report 31, p. 685-696, 1979. Witherly, K. (2013), Seeing Deep and Staying Focused Challenges for Exploration in Northern Chile; presented at PDAC DMEC workshop March 6, 2013, Toronto Canada. Wyborn, L.A.I., Heinrich, C.A., and Jaques, A.L., (1994), Australian Proterozoic mineral systems: essential ingredients and mappable criteria. Australian Institute of Mining and Metallurgy Annual Conference, Melbourne, Proceedings, pp. 109–115. Acknowledgments On the topic of mineral systems, Cam McCuaig and Jon Hronsky are thanked for providing much of the initial insights into the subject. On the history of IP to porphyry copper exploration, Chris Ludwig, Randall Nickson, Hunter Ware, Dave Fountain and Phil Nelson are thanked for pointing out some of the important road signs along the journey. Finally, Graham Closs is thanked for providing his invaluable assistance as author’s frequent ‘Dr Watson’.

Case studies in integrated geological and geophysical 3D modelling: Value added to exploration and mining projects Hernan Ugalde1, Iris Lenauer2, Jean-Francois Ravenelle2, Anna Fonseca2 and Ivo Vos2 1Paterson, Grant & Watson Limited, Toronto, ON 2 SRK Consulting (Canada) Inc., Toronto, ON Summary The combination of geophysical models with litho-structural models represents a valuable tool for increasing the understanding of subsurface geometries. Better subsurface models add value to mineral exploration projects. Geophysical data is used to enhance and validate litho-structural models. The regular distribution of geophysical data allows lithologies and faults to be continued into areas not exposed. The geophysical signal calculated from the model geology is compared with the observed signal. Discrepancies between modeled and observed signals highlight areas where refinements of the geological model are required. In the presented cases, limitations imposed on litho-structural model by irregular data distribution are counteracted by the regularly spaced data from geophysical surveys. The case studies present examples of how iterative modeling from geological and geophysical data will result in an improved final product. Applications are in determine the position of rocks of distinct physical properties, checking the geometry of faults and extending mapped structures into inaccessible/covered areas. Introduction Geophysical and geological subsurface models are distinct in terms of their data distribution and resolution, modelling workflows, and the time of their creation in the mine life cycle. As a result of this, the district- to deposit-scale 2D-4D models developed in support of mineral exploration and mining projects suffer from a poor integration between geology and geophysics. This contribution aims to demonstrate the advantages of increased cross-disciplinary collaboration in 2D-4D subsurface modelling. A distinct advantage of using data acquired through geophysical surveys for modelling is the even distribution of data over an area of interest. For this reason, geophysical models may be constructed for early-stage exploration programs where little information on the subsurface geology is available. Thus, 3D inversion of geophysical data and semi-automatic interpretation routines are becoming increasingly popular for building geophysical models at this stage of a project lifecycle; though these do not always result in geologically accurate models.

From early-stage exploration programs onward, increasingly more geological data is collected for a given project. In contrast to geophysical data, geological data are commonly irregularly distributed as a result of the distribution of boreholes, or the availability of underground or surface exposures. The use of geological data for the construction of a 2D-4D geological model therefore constrains the model extents both laterally and at depth. Despite the complementary nature of geophysical and geological data, available geophysical data are often not integrated in geological models (and vice versa). Without their proper integration, reconciliation of geological data (faults, contacts, stratigraphy) with geophysical data (maps, sections and block models) will remain a challenge. Method In this contribution we present case studies where we integrated existing geological and geophysical data to develop 4D district- to deposit-scale litho-structural models. In the applied workflow, structural geology data (whether collected from oriented core or exposures, or interpreted from geophysical data) form the backbone of each of the litho-structural models. Subsequently, geophysical section models are developed to validate and enhance the litho-structural model (or vice versa, where only geophysical data are available). Our results demonstrate the added value of a streamlined workflow that integrates geological and geophysical data at any exploration or mining project. Case Studies (1) In area of a developed mine, vast amounts of geological data are available in the form of boreholes, surface and subsurface maps. From these data sets a simplified regional model was created to better understand the orientation of and the displacement on normal graben faults as well as the distribution and geometry of igneous rocks (Figure 1). The geological model is well constrained in data dense areas of drilling and underground working and is the result of interpretation and interpolation in areas where only surface observations exist. Aeromagnetic data

Litho-structural geophysical modelling

was used to test the validity of the current geological model, especially the geometry of faults at depth. Cross-sections extracted from the geological model served as a basis for calculating the magnetic signature of modeled geology. Magnetic susceptibilities for each of the outcropping lithologies were varied to best match the amplitude of the observed signal. In one section, the calculated signal closely resembles the observed signal in areas where igneous rocks are exposed at surface (Figure 2B). A second section was extracted from the geological model and its modeled magnetic signal was calculated using the same rock properties as in the first section. In areas of surface exposure of the igneous rock, the calculated signal deviates significantly from the observed signal (Figure 2C). To adequately replicate the magnetic signal of the igneous rock, different susceptibility values are required for each section. This suggests variations in composition of the lithology marked as a single unit in the geological model. Variations in peak/low width and position between the calculated and the observed signal indicate a mismatch between the modeled and the “observed” geometry of lithologies. The thickness, dip and extent of rock units, as well as the dip of faults are modified to better replicate the observed magnetic signature. Using expected values for various compositions of volcanic and igneous rocks, we were able to determine to position of rock units at surface and depth. Furthermore the forward modelling of magnetic data highlighted internal variations in rock units, provoking continuing studies into differentiation of previously lumped units. Correlation of litho-structural models with geophysical models can be enhanced by acquisition of physical rock properties of each of the modeled lithologies. (2) In a greenfields exploration example, knowledge of local geology is mainly sourced from surface mapping. Aeromagnetic and EM data acquired in the area provide insight into the subsurface geometry of a structurally complex fold and thrust belt (Figure 3). This fold and thrust belt has been interpreted as a series of parallel thrust faults as well as anti- and syncline trains (Figure 4). In this model we test for the possible locations of thrust faults dipping in the opposite direction. A series of sections are created from the surface geology maps. By simplifying the lithologies into physically distinct units, the area is subdivided into metasandstones, metagabbros and –iron formations. Magnetic and EM data are modeled on these sections to determine the dip and position of conductive units and units with high magnetic susceptibility (i.e., iron formations). In the CDI EM section (Figure 4A) graphitic metasediments are modeled as inclined slabs. The distribution and geometry of the gabbroic units and iron formations are modeled from the magnetic data. The dip and extent of these physically

distinct units provide constraints on the possible fault geometries.

Figure 1: Surface geology (A) and aeromagnetic data (B) draped over topography. (A) Igneous rocks (pink) are exposed at surface, though their geometry at depth is not well understood. (B) Igneous rocks are marked by a strong magnetic signal. (3) A case study investigating the regional extents of the structures around a developed mine site was based on a detailed fault model constructed from borehole data and a detailed magnetic survey. The structural interpretation is limited to areas of good data coverage. Extrapolation of the model away from areas of dense data coverage was hampered by edges of magnetic survey grids. In places where resolving the question of where structures are continuing is critical, geophysical data can provide important constraints. However, just “regular” TMI and 1VD grids proved to be insufficient for allowing a clear structural visualization over the covered area (Figure 5A). The original 200 m loose drape survey data was downward continued to 50 m elevation above the terrain, then it was microlevelled and improved grids were generated (Figure 5B). This provided with additional contrast and data texture that allowed for possible location and orientations of faults to be determined (Figure 5C).


Figure 2: Forward modelling of sections. (A) Location of modeled sections over the Amplitude of the Analytic Signal of the TMI data. (B) Signal calculated from the modeled geology (red line) matches the observed signal (black line) in areas where igneous rock (pink) is exposed at surface. (C) Signal calculated from the modeled geology (red line) matches the observed signal (black line) in areas where volcanic rocks (yellow, green and orange) are exposed at surface. Identical physical rock properties were used for both sections.

Figure 3: (A) Geological map showing distributions of gabbro (light brown), metasediments (dark brown and yellow) and iron formation (purple). (B) Second vertical derivative of reduced-to-pole magnetic data shows distribution of iron formation and main structures.


Figure 4: (A) Section showing surface geology and CDI EM section. (B) First pass interpretation of geology. (C) Simplification of lithologies into units with similar rock properties. (D) Modeled section showing thrust and backthrust faults. Conclusions This work reinforces the idea that physical rock properties are critical for the proper geophysical control of any geological modelling program. Although one can associate a small range of physical properties to the observed lithologies on surface and then iterate these properties within that range until the amplitude of computed and observed geophysical anomalies match, this would only work on a situation where the geological model is well known and constrained by boreholes. In all the more common situations where the dip of the geological units and their distribution at depth is unknown, the above methodology is insufficient. It us often thought that the input from geophysics ends as soon as detailed 3D geological models are available from boreholes and surface mapping. However, geophysical models can successfully add value to these existing litho-structural models by:

1) Refining the distribution and geometry of rock units;

2) Allowing structural interpretations (fault orientations) to be tested;

3) Subdividing units based on their magnetic signature;

Finally, advanced geophysical processing beyond what is commonly distributed with the survey package can aid on the visualization of structures and contacts on areas of weak signal to noise ratio and little geological information on surface.

Figure 5: Structural interpretation limited by the extent of high resolution data. (A) Aeromagnetic data prior to reprocessing. (B) Downward continued data. (C) Validation of E-striking shear zone and continuation beyond the observed extent.

Exploration Success – A Business Model and Case Study R. T. Osmond1*, M. Montoya R2. and P. Pare3 1. GeoVision Geosciences Inc., Maple Ridge, BC, Canada. 2. GlobeTrotters Resource Group Inc., Lima, Peru. 3. Anglo American Exploration Canada, Vancouver, BC, Canada Summary

Historically making exploration discoveries in the mining industry has been difficult and very high risk. In the past decade alone the mining industry spent over $100 billion in exploration worldwide with just 524 new discoveries made. Given the extremely high cost of discovery and the dismal track record of exploration success, it is worth analysing the key risk factors involved in making an economic discovery to build a better risk model for exploration success. Outside of the brownfields environments, the key risk factor impacting exploration success is geological risk where all other risks are considered secondary to exploring in the right geological environments. Using this risk model it is possible to build an exploration strategy around improving the odds of exploration success by minimizing geological risk in higher geological risk environments. In these environments it is critical to recognize at an early stage important regional geologic features with similarities to those in known mining districts in other parts of the world and to build the necessary regional geological, geochemical and geophysical datasets to effectively target down to the property scale. This risk model also has implications for RnD where the most significant advances in regional scale geological mapping and property scale targeting have come through the development of satellite based multispectral scanners such as ASTER which is capable of remotely mapping hydroxyl minerals associated with hydrothermal alteration. This technology was used by GlobeTrotters Resource Group as part of their exploration strategy to reduce overall exploration risk which has led to the discovery of two new early-stage Cu-Mo porphyry projects in Central Peru. Introduction

It has always been said that that the chances of making an economic discovery in the mining industry is likely 1 economic discovery for every 1000 prospects tested. However, research by Kreuzer (2007) suggests that there is no real consensus on the exact number of prospects that have to be tested before an economic discovery is made but on average a least 20 to 100 skillfully selected prospects are needed in brownfields exploration (1% to 5%) and between 200 to 3,333 prospects in greenfields exploration (0.5% to 0.03%). Regardless of the case, these are very dismal odds particularly for junior mining

companies who tend to focus primarily on greenfields exploration and rely heavily on capital financings to cover their exploration expenditures. Based on research by MinEx Consulting (Schodde, 2013), a total of $116 billion was spent globally to discovery 524 new economic non-ferrous mineral deposits over the past 10 years (2003-2012) placing the industries all-in exploration costs per unit economic discovery at $221.4 million. Given the extremely high cost of discovery and the poor track record of exploration success, it is worth analysing the key risk factors involved in making an economic discovery to build a better risk model for exploration success particularly for outside of brownfields environments. Exploration Success – Risk Assessment

After 20 years of working in the mineral exploration industry on early stage to advanced stage projects and also participating in several exploration risk assessments on global, country and regional scales, it became increasingly clear that there are many risks associated with making an economic discovery. At the same time, it also has become clear that several of these risks stood out as being significant in the exploration process. The following table summarizes a personal top-ten list of key risk factors affecting exploration success: Risk Description Impacts Economic (Ec) Commodity price F+Ge+Ex Time (T) Time to discovery F Financial (F) Cost of discovery Ge+Ex Political (P) Safe-security-permits Ge+Ex Social (S) Social license Ge+Ex Human (H) Expertize/management Ge+Ex Geographic (Gr) Location/access Ge+Ex Geological (Ge) Right environment Discovery Exploration (Ex) Effective targeting Discovery Environmental (En) Access and work Ge+Ex Table 1: Key risk factors associated with exploration success. Upon reviewing these key risk factors, it was clear that most are directly impacting the ability to acquire properties in favourable geology (geological risk) and to carry out exploration work to effectively drill target on these properties (exploration risk). Both of these risks have a direct impact on discovery. Other factors such as financial which can be the single largest risk to exploration success

Exploration Success – Business Model and Case Study

have the largest impact on the ability to do work and acquire properties in favourable geology. As such, even though they can individually have significant impacts, they are considered secondary to managing geological and targeting risks. Based on this assessment, it is personally felt that it is possible to formulate the probability of exploration success as follows: Exploration Success = Geological x Exploration x Secondary Intuitively, this also makes a lot of sense as exploration risk and all secondary risk factors outlined in Table 1 are only relevant if the geological environment is favourable for discovery. As such, mining companies tend to look for ways to reduce their overall risk and improve their chances of exploration success by first looking to acquire properties in low geological risk environments. This is supported by the fact that a significant portion of global exploration expenditures by major mining companies is spent drilling near-mine and brownfields projects in well-endowed mining districts.

Most junior mining companies conduct exploration on substantially lower annual budgets than the majors and generally cannot compete for properties in brownfields environments. A large portion of junior companies explore just outboard of known brownfields environments in moderate geological risk environments. Many junior miners acquire properties with old prospects or historic small scale operations that were explored or mined in the past when metal prices were significantly lower than what they are today. These properties generally have already been evaluated to some degree by the majors for potential tier 1 or 2 discoveries but in some cases, such as the discovery of the Pebble (Northern Dynasty Minerals, 2013) and Oyu Tolgoi (Rio Tinto Mongolia, 2013) Cu-Mo porphyry deposits, can still yield significant exploration discoveries.

True greenfields exploration is considered to have the lowest probability of exploration success and the highest geological risk where the right geological setting for the development of an economic deposit has not yet been proven. In this environment junior mining companies tend to have a very difficult time finding source funding to conduct exploration and as a consequence most greenfields early stage discoveries are made by local prospectors following surface indications of alteration or mineralization. At this stage, if the new prospect shows exploration potential with similar geology, alteration and mineralization to other known economic deposits, junior and sometimes major mining companies are willing to take the risk to do the early stage exploration. Implications – Exploration Strategy

The key to increasing overall exploration success is through the understanding of geological risk in selecting where to explore and how to effectively drill target once in that right geological environment. As previously mentioned, most major mining companies do this by operating low risk geological environments while paying hundreds of millions to billions of dollars to acquire new economic discoveries outside of brownfields geological environments. This overall strategy is not considered a cost effective approach to mitigating the financial risks related to exploring in high geological risk environments. Most countries have been mapped at a sufficient scale to understand their regional geologic and tectonic settings. Additionally with today’s understanding of the geological and metallogenic processes associated with the formation of economic deposits, experienced geologist are able to recognize greenfields geological environments with similar geology and tectonic settings to known mining districts in other parts of the world. A good example of this was Anglo American’s global Ni targeting initiative carried out in 2003 that recognized Northern Finland as an unexplored portion of the Baltic Shield with similar geology and tectonic setting to the world class Pechenga Ni-Cu deposits found on the Kola Peninsula of Russia. This greenfields targeting initiative provided Anlgo American with an early entry into Northern Finland which eventually led to the discovery of their new Sakatti Ni-Cu deposit in 2009.

Even though similar geological and tectonic setting to other known mining districts have been recognized in unexplored regions of the world, many exploration companies are still reluctant to take the financial risk to conduct exploration to target down to a property scale in greenfields environments. Companies still require some prior indications of alteration or mineralization as an initial focus to justify the early stage exploration spending required to generate property scale targets in these environments. At the same time after an initial discovery is made, the perceived geological risk is substantially reduced which often leads to a major staking rush and hundreds of millions of dollars being spent acquiring surrounding properties and conducting exploration over a 3-5 year period following the initial discovery. Kerr (2003) reports that in the 3 years following (1995-1997) the initial Voisey’s Bay Ni-Cu-Co discovery in Northern Labrador, a total of $220 million was spent on exploration in Labrador.

In the years leading up to the Voisey Bay discovery, total exploration expenditures in Labrador were less than $1 million per year. A large portion of those expenditures were spent by Falconbridge who had recognized Labrador as an unexplored region of the rifted margin of the Superior Craton of North America a major tectonic setting known to host the Raglan Ni-Cu-PGE deposit in Northern Quebec and the Thompson Ni-Cu deposit in Northern Manitoba (Eckstrand, 2007). Falconbridge had been conducting small reconnaissance scale greenfields exploration programs


throughout Labrador for 3 years prior to the Voisey’s Bay discovery. They had outlined the Voisey’s Bay area as exploration target 22 out of 55 on their list of priority targets to evaluate for all of Labrador based primarily on an anomalous 0.2% Ni lithogeochemical sample in a troctolite (personal communication). Unfortunately, Falconbridge never got the opportunity to evaluate target 22 before the discovery of Voisey’s Bay by two local prospectors, Al Chislett and Chris Verbiski, in 1993 while conducting diamond exploration for Diamond Fields Resources throughout Northern Labrador (personal communication).

The DIGHEM airborne geophysical survey completed in 1995 over all of Diamond Fields Resources properties outlined an obvious strong more than 3 km long E-W trending airborne EM anomaly related to the magmatic Ni-Cu sulphides within the troctolite dyke hosting the main Voisey’s Bay Ovoid (31.7 MMT), Discovery Hill (7.3 MMT) and Ried Brook (19.0 MMT) deposits (Balch, 1998). The DIGHEM survey had effectively outlined more than 50 MMT of Ni-Cu-Co sulphide mineralization in the Voisey’s Bay area. Later a detailed ground gravity survey was carried out over the Main Block which outlined the Ovoid deposit as a strong 4 mGal Bouguer gravity anomaly and also mapped out the property scale distribution of dense olivine bearing troctolite intrusions known to host the Voisey’s Bay deposits (King, 2007).

Given the obvious geophysical responses associated with the Voisey’s Bay deposits and related troctolite intrusions hosting these deposits, early stage property scale targeting using commercially available airborne EM geophysical methods would likely have led to the Voisey’s Bay discovery by Falconbridge. However, this would only have been the case if the troctolite intrusions of the Nain Plutonic Suite were recognized as a favourable geological environment for the formation of economic magmatic Ni-Cu-Co sulphide deposits.

Based on the significant value generated by the discovery of a Tier 1 economic deposit like the Voisey’s Bay deposits, it seems very surprising that most mining companies invest very little to generate the necessary regional scale geological, geophysical and geochemical datasets at an early stage to more effectively target in greenfields geological environments. Implications – Research and Development (RnD)

The key to reducing geological risk in most greenfields environments is obviously to develop a better understanding of the geology and metallogeny of a region to help target down to the property scale. However, geological field mapping is considered the one of the most challenging aspects of greenfields exploration in many parts of the world. Field mapping often requires a long term commitment from expert geologists and in many cases is

very difficult to complete due to challenging geographic conditions such poor climate, high altitude or thick surficial cover. In these environments large scale regional geochemical surveys in conjunction with regional scale geophysical surveys provide an excellent more cost effective alternative to geological mapping.

This approach was taken by the Finland government with the in-house development of their GTK airborne geophysical system back in 1972 (Hautaniemi, 2005). The GTK system was mounted on a fixed wing aircraft and capable of simultaneously collects frequency domain EM, total magnetic intensity and radiometric geophysical data. During the period between 1972 and 2004 the Finland government surveyed the entire country at 200 m line spacing. Anglo American acquired the GTK geophysical datasets from the Finland government during the early stages of their generative targeting initiative. As it turned out, the initial selection of the MOS-8 GTK target eventually led to the Sakatti Ni-Cu-PGE discovery by Anglo in 2009. After the completion of the Finland surveys in 2004, the government made the system commercially available at a survey cost of roughly $30 per line km and was eventually sold to SGL in 2008.

Some of the most significant developments in low-cost regional geological mapping and direct targeting have come from remote sensing related to the development of Landsat and ASTER satellite technologies as well as the development of high resolution airborne hyperspectral mapping systems. These sensors operate as surface scanners designed to remotely measure discrete bands of infrared spectral energy reflected and emitted from the ground (Reference).

The ASTER (Advanced Spaceborne Thermal Emission and Reflection Radiometer) multispectral scanner was launched onboard NASA’s Terra spacecraft in December of 1999 as a collaborative effort between Japan’s Ministry of the of International Trade and Industry (MIDI) and the USGS (ERSDAC, 2005). Each ASTER scene covers a surface area of roughly 60x60 km2 and measures a total of 14 spectral bands at varying resolutions strategically positioned within the VNIR (3 at 15 m), SWIR (6 at 30 m) and TIR (5 at 90 m) spectra (ERSDAC, 2003). Atmospherically corrected ASTER scenes are commercially available from the Japan Space Systems website at a cost of roughly $120 per scene. The release of ASTER reflectance data in the early 2000’s proved to be a major advancement in remote sensing with the capabilities to map out discrete hydrothermal alteration minerals associated with metalogenic processes. The ASTER scanner has 6 bands strategically located in the SWIR between 1.6 and 2.43 µm for the detection of absorption features associated with hydroxyl minerals such as white micas, alunite and kaolinite. These minerals are commonly associated with intense hydrothermal alteration such as the argillic (clays) and phyllic (white micas) alteration


surrounding large Cu-Mo porphyry deposits (Honarmand M., 2011). As such mining companies have been effectively using ASTER data for the past decade to explore for large hydrothermal alteration systems mostly associated with porphyry Cu-Mo and epithermal Au-Ag deposits in relatively arid poorly vegetated regions throughout the world. The overall cost to acquire ASTER data is on the order of 3.3 cents per km2 making it by far the most cost effective geological mapping and property scale targeting tool available for early stage greenfields exploration. Unfortunately, IR radiation only penetrates several centimeters into the ground making the ASTER scanner useless for mapping bedrock alteration in vegetation and overburden covered terrains. Business Model

GlobeTrotters Resource Group Inc. is private prospect generator junior exploration company focused on Cu-Mo porphyry and epithermal Au-Ag exploration in South America with a particular focus on Peru. The company’s exploration strategy was designed around the risk management approach to improving the odds of exploration success. Based on a country scale risk assessment of secondary risks including legal, political, social and financial risks, it was felt that Peru was a relatively low risk country to conduct exploration and develop mines with excellent potential for discovery.

From a geological risk perspective, the Andes region of Peru was selected as a primary region of focus for GlobeTrotters because of its exploration potential for large porphyry Cu deposits. The highly prolific Eocene-Oligocene and Paleocene Cu-Mo porphyry belts of Southern Peru are known to host several of the world’s largest Cu-Mo porphyry deposits including the Cerro Verde Cu-Mo porphyry deposit operated by Freeport-McMoran and the Las Bambas Cu-Mo porphyry skarn deposit presently owned by Glencore. These belts are for the most part brownfields exploration environments where a significant portion of the highly prospective ground is held by major and mid-tier mining companies and land acquisition is expensive and highly competitive.

The team recognized that even though most of the known large porphyry deposits were concentrated in the Paleocene and Eocene-Oligocene belts of Southern Peru, several known porphyry discoveries had also been made throughout the Cretaceous and Miocene belts which extend the entire length of the country. These included the Zafranal Cu-Mo porphyry deposit in the Cretaceous belt in Southern Peru and the Antamina, Toromocho, and La Granja Cu-Mo porphyry deposits located in the Miocene belt in Central and Northern Peru. GlobeTrotter felt that a large portion of these belts were underexplored but still provided the right geological and tectonic setting for the emplacement of economic Cu-Mo porphyry systems.

Based on this risk assessment, the company put together an exploration strategy to explore for Cu-Mo porphyry systems in mineral belts outside of the know Cu-Mo porphyry belts of Southern Peru. The targeting process included the acquisition and processing of 190 ASTER scenes covering the entire Andean regions of Peru. More than 1000 argillic+phyllic alteration anomalies were outlined and integrated with available government 1:100,000 scale geology and known mineral occurrence datasets. These ASTER anomalies were prioritized for field targeting based on their quality, geological and structural setting, geographic location and association with known mineral occurrences. The initial targeting outlined a list of 40 priority targets that were selected for field follow-up. This targeting initiative led to the discovery of two new untested mineralized Cu-Mo porphyry systems in Central Peru. Conclusions

The risk management process can be a very powerful tool for assessing and mitigating risks associated with exploration. At the same time the impacts of these risks can be used to build better exploration strategy to improve overall exploration success. Generally it is an interplay of many risks that often controls overall exploration success but in most geological environments, the key risks directly impacting exploration success are favourable geology and explorability. In brownfields environments, often the geological risks become secondary to political, social and environmental risks as a result of the local population’s prior experience with mining but the financial rewards are substantial enough to justify efforts to mitigate these risks. Outboard of brownfields environments, exploration success is impacted the most by (i) a lack of expertize to recognize favourable geological settings and exploration potential as well as (ii) not taking the financial risks to carry out exploration at an early stage to acquire the necessary geological, geophysical and geochemical datasets to properly evaluate favourable geological environments. This was exactly the case with Falconbridge’s experience work in Labrador in the early 1990’s prior to the discovery of Voisey’s Bay deposit. At the time, Falconbridge geologists did not have a good understanding of the Ni exploration potential of troctolite intrusions and considered Ni geochemical anomalies in troctolites a lower priority. At the same time the lack of financial support at this early stage to conduct regional scale geochemical and geophysical dataset made it extremely difficult for the exploration team to properly evaluate the exploration potential of all of Northern Labrador. Luckily for Anglo America, the Finland government had made the commitment to acquire high resolution airborne radiometrics, magnetics and EM


datasets for all of Finland making the early stage Ni targeting process easier and much more effective.

Unfortunately, most mining companies are not willing to take the financial risks to carry out large scale geochemical and geophysical surveys in most greenfields environment. These surveys are considered too expensive for early stage greenfields exploration and should be provided by government. Fortunately for Anglo American, they were able to acquire from the Finland government one of the most complete countrywide high resolution magnetic, radiometric and EM datasets in the world. This data proved instrumental for understanding the geology and mapping out the Ni exploration potential of Northern Finland.

The most significant breakthroughs, over the past 20 years, in low-cost regional mapping and target generation have come from remote sensing primarily related to the availability of satellite Landsat and ASTER datasets as well as the development of high resolution airborne hyperspectral mapping systems. These datasets are very cost-effective to acquire over large tracks of ground and do an excellent job mapping out large hydrothermal alteration signatures commonly associated with Cu-Mo porphyry and epithermal Au-Ag systems.

Based on the risk model described above for improving exploration success, GlobeTrotters has been able to build an exploration strategy around minimizing geological and exploration risk. The company has done this by (i) exploring for Cu-Mo porphyry systems in the key mineral belts of Peru with similar geological features to the know prolific porphyry belts of Southern Peru and (ii) by utilizing low-cost ASTER data in conjunction with regional geology and known mineral occurrences to develop property scale porphyry targets outside of brownfields environments. This has led to early success with the discovery of two new Cu-Mo porphyry systems in the central part of Peru. References Balch S., Crebbs T.J., King A., Verbiski M., 1998,

Geophysics of the Voisey’s Bay Ni-Cu-Co deposits, - 68th Annual International Meeting, Society of Exploration Geophysics, Expanded Abstracts.

Eckstrand O.R. and Hulbert L.J. 2007, Magmatic Nickel-Copper-Platinum Group Element Deposits, in Goodfellow, W.D., e

I had., Mineral Deposits of Canada: A Synthesis of Major Deposit Types, Disctrict Metallogency, the Evolution of Geological Provinces, and Exploration Methods: Geological Association of Canada, Mineral Deposits Division, Special Publications No. 5, Pages 205-222.

ERSDAC 2003, ASTER Reference Guide Version 1.0, March 2003, ERSDAC (Earth Remote Sensing Data

Analysis Center), http://www.science.aster.ersdac.jspacesystems.or.jp/

ERSDAC 2005, ASTER User’s Guide Version 4.0, July 2005, ERSDAC (Earth Remote Sensing Data Analysis Center) http://www.science.aster.ersdac.jspacesystems.or.jp/

Farnsworth C. H. 1996, International Business; Inco Bid Tops Falconbridge For Big Mine, The New York Times, Published March 8th, 1996.

Hautaniemi H., Kurimo M., Multala J., Levaniemi H. and Vironmaki J. 2005, The “Three In One” Aerogeophysical Concept of GTK in 2004, Aerogeophysics in Finland 1972-2004: Methods, System Characteristics and Applications, Edited by Meri-Liisa Airo, Geological Survey of Finland, Special Paper 39, 2005, Pages 21-74.

Honarmand M., Ranjbar H. and Shahabpour J. 2011, Applications of Spectral Analysis in Mapping Hydrothermal Alteration of the Northwest Part of The Kerman Cenozoic Magmatic Arc, Iran, Journal of Sciences, Islamic Republic of Iran 22(3) 2011: pages 221-238.

HyVista 2014, HyMap, http://www.hyvista.com/ Kerr A. 2003, Voisey’s Bay and the Nickel Potential of

Labrador: A Summary for the Nonspecialist, Current Research (2003), Newfoundland Department of Mines and Energy Geological Survey, Report 03-1, pages 231-239.

King A. 2007, Review of Geophysical Technology for Ni-Cu-PGE deposits, In “Proceedings of Exploration 07: Fifth Decennial International Conference on Mineral Exploration” edited by B. Milkereit, 2007, pages 647-665.

Kreuzer O.P. 2007, Risk Modelling – Increasing the Effectiveness of Our Exploration Investment, Geoconferences (WA) Inc. Kalgoorlie ’07 Conference, 25-27 September, 2007.

Northern Dynasty Minerals Ltd. 2014, Pebble-History, http://www.northerndynastyminerals.com/

Rio Tinto Mongolia 2014, Oyu Tolgoi project, http://www.riotintomongolia.com/

Schodde R. 2013, Long Term Outlook for the Globeal Exploration Industry – Gloom or Doom?, Geological Survey of South Africa GeoForum 2013 Conference, 2-5 July, 2013.

Acknowledgements The authors would like to acknowledge the following for their support in helping to prepare this paper: S. McLean (Transition Metals), J. Foster (Sirius Resources), K. Felon (Vale), R. Turner (Independence Gold), S. Regoci (Garibaldi Resources) and A. Christopher (Teck).

An integrated approach for assessing the mineral potential of the Liptako Metallogenic Province (Iullemmeden Basin, Niger) Nasreddine Bournas (1,*), Djibo-Maïga Abdoul-Wahab (2), Karl Kwan (1), Marta Orta (1), Geoff Plastow (1) , Alex Prikhodko (1) , Shaolin Lu(1) and Keeme Mokubung(1) Corresponding author: [email protected] Summary Airborne geophysical surveys, consisting of Fixed-wing magnetic gradiometer and gamma-ray spectrometer and helicopter-borne VTEM full-waveform time-domain electromagnetic surveys were carried out over the Iullemmeden Basin, in western Niger for mineral exploration and detailed geological mapping. The western portion of the survey area, which belongs to the Liptako Metallogenic Province, hosts several mineral occurrences including base and precious metals, precious stones and special metals in various deposit styles. Geophysical data integration with other available types of information including Landsat TM data and geological maps has been performed using Neural Network (NN) and Enhanced Maximum Likelihood (EML) supervised classification techniques. Combined techniques have proven very useful during the analysis of multidisciplinary data resulting in better understanding of the distribution of mineralization and the results were used for targeting and selecting favorable areas for the exploration of various styles of mineralization within the study area. Results of gamma-ray spectrometer, aeromagnetic and Landsat TM data interpretation and their integration with the known geology suggest the existence of three metallogenic provinces within this area. The first two provinces are associated with tertiary sediments of the Iullemmeden Basin and include potential radioactive mineralizations (Uranium and Thorium). The third province is associated with intrusive and metavolcanic sedimentary formations of the Birimian Liptako basement outcropping partly along the east side of the Niger River and hosting various styles of polymetallic mineralization (Cu, Pb, Zn, Ag), precious metals (gold and diamonds), special metals (Sn, W, Ta, Bi, Li, Ti), iron ore, REE and Ni-PGE mineralization. Introduction A fixed wing airborne gamma-ray spectrometer and magnetic gradiometer survey was carried out over the Iullemmeden sedimentary Basin, located in the western Niger in 2012-2013 covering an area of approximately 73,000 km2 for mineral exploration and detailed geological mapping. The airborne survey was flown along lines

striking in the east-west direction and spaced at 500m apart at constant ground clearance of 100m in average. Three high-sensitivity cesium magnetometers were used to measure the inline and crossline horizontal gradients of the magnetic field, and a set of two NaI (Tl) gamma-ray detectors packs, each of 16L (for downward looking) and 4L (for upward looking) in volume were utilized to measure the gamma-ray radiations. The fixed wing airborne survey was followed by detailed helicopter borne VTEM surveys to collect time-domain EM and magnetic field data over several target areas. In this paper we present the interpretation results of the acquired airborne geophysical data and their integration with the satellite imagery and available geological information. The study is particularly focusing on the western portion of the survey area which is a part of the Liptako Metallogenic Province. Geology and Mineralization The sedimentary series exposed in the Iullemmeden Basin consist mainly of Eocene-Pliocene Continental Terminal and Quaternary formations. The Continental Terminal series which consist of marine and continental sediments host considerable amounts of Phosphate, lignite Coal and iron ore formations. The Precambrian Liptako basement formations outcrop in the western portion of the sedimentary basin and consist mainly of Birimian volcano-sediments and Proterozoic intrusive formations. Since the earliest exploration work carried out in the Liptako basement, these volcano-sediments and intrusive rocks have shown a metallogenic interest for the exploration of various mineral resources (Franconi, 1985); they host several hundred of mineralized occurrences including precious and base metals, precious stones, PGE and REE mineralizations (Figure 1). Interpretation Results

� Geological mapping

The acquired magnetic data were used for mapping geological structures including faults, contacts, mafic

Integrated approach for mineral targeting in the Niger Liptako

metavolcanic formations and various types of intrusive rocks including felsic, mafic and ultramafic formations. The data were also used for delineating potential target areas for the exploration of various mineralization styles including base and precious metals, PGE, ferric metals (Fe, Ni) and iron-ore formations. Figure 2 illustrates the results of magnetic interpretation. The map shows evidence of the presence of numerous NW trending Mezo-Neoproterozoic mafic doleritic dikes intruding the Birimian basement formations and dominating the southwestern portion of the study area. The interpretation also suggests the existence of NE and NW trending faulting systems that seem to affect the doleritic dykes. The NE trending system is suggested to be associated with the lithospheric Guinean-Nubian-Lineament (GLN). Several circular magnetic anomalies suggesting a possible link to kimberlite pipes were identified in the Liptako cratonic area (West African Craton) and along the GNL lineament in the pre-cratonic area within the sediments. Mafic metavolcanics (mV) and mafic-ultramafic intrusives are identified in the western portion of the study area. The metavolcanics occur either as sandwiched between two NE trending faults or as folded formations and affected by NW trending faults. Most of gold deposits known in the Liptako Metallogenic Province are associated with shear-zone hosted tectonic traps within the volcano-sedimentary formations of the Birimian green stone belt.

Joint analysis of radiometric and Landsat TM data was crucial for accurate lithological mapping resulting in a new and more accurate lithological map of the study area.

� Mapping of granite intrusives

Granite intrusives and pegmatite dikes known to host several Tin and Tungsten mineralized zones are delineated with the magnetic and radiometric data; they are associated with magnetic lows and highly radioactive levels. Furthermore, elevated potassium concentration and elevated potassium to thorium ratios coincide with mineralized young intrusives and alkaline rocks. Birimian volcano-sedimentary formations that host several styles of mineralization including copper, gold and tin mineralization are accurately delineated due to their low radioactivity response and moderate magnetic signature.

� Mapping of radioactive minerals The spectrometer data were very helpful for mapping and delineating highly radioactive anomalous zones associated with uranium and thorium mineralization of economic interest. The detected uranium anomalous zones exhibit elevated equivalent uranium concentrations and high uranium to thorium ratios; they are associated with highly altered and mineralized tertiary sandstones and conglomerates.

Figure 1: Simplified geological map of the Iullemmeden bassin showing the location of the main mineralization occurences and the Fixed wing airbone and helicopter VTEM surveys outlines. Geological map after Greigert, 1966.


� Regolith mapping Regoliths of West Africa represent an important source of commodities including iron, bauxite and thorium (Metelka, 2011). The joint use of several data sets including Landsat TM imagery, gamma-ray spectrometer data and digital elevation models has been widely used as a standard approach for regolith mapping (Jaques, 1997). The integrated approach helped detect and delineate several regolith units within the survey area. The terrain shape, aspect and slope were calculated and used to locate highlands and plateaus. Unsupervised classification analysis helped detect thorium anomalous zones, and finally iron-rich regolith landforms were identified from the iron-oxide index and alteration maps derived from the Landsat TM data (Figure 3).

� Mapping of bedrock conductors VTEM data were used to characterize the resistivity distribution of the subsurface and to help detect shallow and deep-seated alteration zones and bedrock conductors suggesting a possible association with metallic mineralization including copper, gold, nickel and other Sulphide mineralization types. Figure 4 depicts the VTEM and magnetic inversion results for the western portion of Line C shown in Figure 1. The CDI section highlights a highly conductive clay horizon on the top of the section and several steeply dipping bedrock

conductors beneath the clay formations. It also depicts a deep-seated conductive horizon attributed to the contact between the basement and overlaying sediments. Data Integration Results Geophysical data integration with available other types of datasets such as Landsat TM imagery, geological and geochemical data has been widely used in mining exploration for targeting and delineation of mineralized

zones based on the relationship that may exist between the various types of mineralization and their geophysical signatures. The approach of integrating multidisciplinary datasets is commonly used for accurate targeting and selecting favorable areas for the exploration of mineralization and alteration (Reford et al., 2004). In this

Figure 2: Magnetic interpretation results superimposed on the magnetic susceptibility depth slice.

Figure 3: Alteration map derived from the Landsat TM data including features of magnetic interpretation.

Figure 4: VTEM interpretation results. The sections are the magnetic susceptibility and EM flow CDI.


study, two different techniques have been used for mineralization targeting; the first is based on the Neural Network analysis and the second, on the Classification and Clustering technique.

� Neural network (NN) analysis

The input data for the NN consisted of 5 different datasets including the magnetic and spectrometric channels and their ratios. The training target was chosen on known mineralized zone including copper, gold and tin mineralization. Areas of high similarity coefficients represent good similarities with known mineralization and therefore, are considered as potential targets for base and precious metallic mineralization (Figure 5).

� Classification and clustering analysis (CCA) Classification is another targeting technique; two different approaches are used: unsupervised and supervised. The supervised classification requires training areas, whereas the unsupervised does not require any training area and is simply based on the statistical analysis of multidisciplinary datasets. Classification techniques were used for targeting radioactive mineralization including uranium and thorium and other style of mineralization associated with intrusives such as Tin and Tungsten. Results of supervised classification using the Erdas-ERmapper Enhanced Maximum Likelihood (EML) technique are depicted in Figure 6. Conclusions and recommendations The analysis of the airborne aeromagnetic, gamma-spectrometric and VTEM data acquired over the Liptako area has provided a new insight in geological mapping and mineral distribution within this area, thus increasing the potential mineral of the province. Geophysical data integration with Landsat TM and geology using neural network and Classifications techniques helped identify potential targets for the exploration of various styles of mineralization including base and precious metals and radioactive raw minerals. Detailed ground and helicopter-borne geophysical surveys including induced polarization, gamma-ray spectrometer and Time Domain EM are highly recommended as a follow-up to the airborne geophysical survey to better define and prioritize the targets for drill-testing.

Figure 5: Neural network simulation results for copper, gold and Tin mineralization targeting superimposed on the magnetic data.

Figure 6: Supervised calssification results for uranium, thorium and Sn mineralization targeting superimposed on the magnetic data.


References

- Franconi A., J. JOO and I. Zibo, 1985, “Plan Mineral du Niger, Tome IV, volumes I et II”.

- Jaques, A.L., P. Welman, A. Whitaker and D. Wyborn, 1997. High-resolution geophysics in modern geological mapping, AGSO J. Aust. Geol. Geophys. 17, 159-173.

- Greigert, J, 1966, Carte géologique du Niger. - Metelka V., 2011. Geophysical and remote sensing

methodologies applied to the analysis of regolith and geology in Burkina Faso, West Africa, Ph.D. thesis, University of Toulouse III and Charles University in Prague.

- Reford S., G. Lipton and H. Ugalde, 2004. Predictive ore deposit targeting using Neural Network analysis, SEG expanded Abstracts: 1198-1201.

Acknowledgments

We are thankful to the Ministry of Mines and Industrial Development of Niger for permission to present these results.

Geologically Constrained Magnetic Inversion at the Great Whale Iron Property, Québec Joël Dubé, Dubé & Desaulniers Geoscience Ltd., 7977 Décarie Drive, Ottawa, Ontario K1C3K3 Taronish Pithawala*, Geosoft Inc., Suite 810 - 207 Queens Quay West, Toronto, Ontario M5J1A7 Summary The Great Whale Iron property, located east of Hudson Bay in Québec comprises three Lake Superior type magnetite iron deposits (Figure 1). One of these deposits, Deposit A, consists of several hundred feet of quartz-magnetite banded iron formation lying within a greater sequence of schistose interbedded pyroclastic and sedimentary rocks. Mineralization consists almost entirely of magnetite. We conducted 3D inversions of publicly available airborne magnetic data in conjunction with known surface and subsurface geological and geochemical data in order to reduce risk in future ore definition drilling. We test the following types of inversions and compare the results: smooth unconstrained susceptibility; iterative reweighting focussed susceptibility; and finally susceptibility constrained with wireframes constructed using drillhole data.

Figure 1: Regional plan map for Great Whale Iron property (from Met-Chem Canada Inc.). Introduction The Great Whale Iron deposits are classified as a Lake Superior type iron formation; specifically the deposits are composed of banded iron that are associated or parent to the Meta-Taconite type. The magnetite iron ore formation (MIF) is mostly a simple, uniform, laminated rock comprising fine to very fine grained quartz and magnetite with little to no hematite content. Gregory (1958) describes the rocks in the area as Precambrian age. The sequence of schistose interbedded pyroclastics and sediments (including the magnetite iron formation) comprise a long north north-east trending belt within surrounding granitic rocks. The

MIF has an average estimated magnetite content of 45-50%. Gregory (1958) hypothesizes that the MIF is a product of primary deposition. Deposit A is heavily folded with a general structure that is essentially an anticline with a homoclinal sequence; the axial plane dips steeply to the east and the fold axis plunge is nearly horizontal (Gregory 1958). Much of the MIF in Deposit A is outcropping and has been digitally mapped (Figure 2) by Met-Chem Canada Inc. for their 2006 43-101 compliant technical report for Niocan Inc. 58 diamond drillholes (DDH) were drilled in 1958 on Deposit A at an average dip of 45 degrees. Average DDH length is ~200m. Holes were drilled in a regular array pattern (Figure 2). Lithology and mineralisation data for each hole is available, however no borehole susceptibility measurements were taken. The data is publicly available on Québec’s Ministère des Ressources naturelles (MRN) website. In October 1998 a helicopter magnetic survey was conducted on the Fagnant Lake property over Deposit A. This data is also available from Québec’s MRN. The magnetic survey lines are spaced 200m apart with the average sensor height being 48m above the terrain (St-Hilaire, 1998).

Figure 2: Geological plan map of Deposit A. MIF in bright orange. Drillhole traces in black (from Met-Chem Canada Inc.). Method Since the MIF is composed primarily of recrystallized magnetite and the hematite content is low to non-existent, we believe that it is possible to use 3D magnetic inversion to model the magnetite rich portion of the formation. 3D inversion models can mitigate the risk and cost of drilling by providing an intuitive 3D subsurface picture that takes into account geological, geochemical, and geophysical observations. We use various inversion techniques provided by Geosoft’s VOXI Earth Modelling service.

Great Whale Iron Property Inversion Modelling

The airborne magnetic data was obtained from Québec’s MRN and input into Geosoft’s VOXI Earth Modelling service. A linear trend was removed from the data to allow the inversion to focus solely on the rock properties within the mesh volume that contribute to the observed response. The data was also anti-aliased and resampled so that there was one data point for each surface cell. An absolute error was assigned to the entire dataset at a value of 5% of the standard deviation. An IGRF correction appropriate for the time and location of the survey was applied as well. We chose a cell size of 50m and a model bottom of -700m. The DEM was obtained from the Canadian Digital Elevation Database. No vertical expansion was applied to the active cell size because we apply geological constraints further into the investigation; this ensures that the geological constraint is not skewed at depth. Drillhole data from the MRN contains lithology and mineralisation information which we used to create a 3D wireframe using Geosoft’s Oasis montaj software. The wireframe model is constructed to coincide with geological mapping of outcrops, DDH logs, and the structural findings summarized by Gregory (1958). When designing the wireframe, we avoid making conjectures about how the MIF is connected at depth or in areas of sparse drilling and instead limit the model to the depth and areal extent of the DDH data (Figure 3). We use this model and the surface mapping to determine whether an inversion result satisfies known geology.

Figure 3: 3D wireframe model of MIF based on DDH logs. DEM in grey; drillhole traces in black. Four susceptibility models are presented in order of increasing complexity: a smooth inversion using no a priori information; a model sharpened using VOXI’s Iterative Reweighting Inversion Focussing (IRIF); an IRIF model constrained simply with a susceptibility upper bound and a shallow depth weighting; and finally a model constrained with a geological wireframe and a susceptibility upper bound.

Model Results The default VOXI susceptibility inversion is a smooth unconstrained model using only the observed magnetic data and any auxiliary information (e.g. error assignment, cell size, etc.) given by the modeller. The Great Whale unconstrained susceptibility inversion result (Figure 4) mathematically satisfies the airborne geophysical observation, but does not correlate well to known geology or the hypothesized maximum susceptibility of the MIF (2.2 SI) (Symons et al., 1983). While the inversion result positively matches the MIF’s overall trend, the distribution of the susceptibility is too smooth given the formation’s sharp contacts. Furthermore, the values suspected for the MIF (~0.8 – 2.2 SI) (Symons et al., 1983) do not outcrop in the regions indicated by the geological plan map. Nevertheless, the smooth unconstrained model gave us confidence that the inversion setup and auxiliary parameters are suitable for the observed data – allowing us to investigate the effects of other techniques on the model. Iterative Reweighting Inversion Focussing (IRIF) is a mathematical inversion technique offered by Geosoft’s VOXI. IRIF refines an otherwise unconstrained model by using the preliminary result as a reweighting factor in a successive iteration. The effect is a geophysical model that is less smooth, and with a concentrated distribution of susceptibility – more closely resembling a realistic Earth model. The automated method is useful when there are no physical constraints the modeller can use with confidence. The Great Whale IRIF susceptibility model matches the overall trend of the MIF and was able to infill volumes near the surface that the unconstrained model left out (Figure 5). These near surface volumes are confirmed by the DDH data. This focussed model also better aligns with the structural orientation of the fold limbs. However, it should be noted that the IRIF model’s susceptibility values are in general much greater than the expected value for the MIF. These highly susceptible volumes are small and placed in deeper parts of the model. Based on the relative success of the IRIF model above, we place two simple but reasonable constraints on a successive IRIF inversion – an upper bound of 2.2 SI, applied to the entire model; and a shallow depth weighting to match outcropping MIF. The result is a 3D distribution of susceptibility that correlates better to surface outcrops, and satisfies known rock properties (Figure 6). Magnetic anomalies not already targeted by drillholes are made apparent; and since the inversion matches what we see in known areas, we have confidence in the result in areas with no a priori information. This model could have been used confidently to target shallow test drillholes in Deposit A. With the additional information gained from drilling, we


can further constrain the inversion to see if the susceptibility continues at depth or spreads out laterally. The final inversion model incorporated the geological wireframe as well as an upper bound constraint of 2.2 SI. The wireframe drawn from DDH data was converted into a parameter reference model for the inversion. The MIF volume was assigned a susceptibility of 1.7 SI, which is the average for magnetite iron ore at four past producing mines (Symons et al., 1983), whereas all cells outside the volume were assigned a susceptibility of 0 SI. A parameter weighting voxel was used in conjunction with the reference model. The objective of the weighting voxel is to tell the inversion how confident we are in the parameter reference model. Inside the volume defined by the DDH data, we assign a high degree of confidence, elsewhere in the model we have very little. The effect on the inversion is that while our value of 1.7 SI is preferred (not fixed) within the wireframe volume, the inversion algorithm is free to operate unencumbered everywhere else. By constraining the inversion in areas where we have a great deal of knowledge, we limit the algorithm to a more accurate property distribution throughout the model – this is particularly important for realistic results in unconstrained regions of the model (where we’re seeking new drilling targets). The geologically constrained Great Whale susceptibility inversion is expectedly the most accurate of the 4 models presented (Figure 7). The parameter reference model successfully placed a magnetic volume with the same structural features of the wireframe close to the surface – matching the outcrop patterns seen in the geological plan map. Since the parameter reference constraint does not fix the susceptibility values, a realistic distribution is found throughout the model. Regions in the wireframe that were hypothesized to be continuous but were left out in the sketch were confirmed by the inversion. In addition, regions not explored by surface geological mapping or by drilling, manifested magnetic anomalies with a structure similar to known areas. Conclusions We note that the traditional, smooth unconstrained susceptibility model does not provide a result that is consistent with even simple geological constructs such as the sharp boundaries of the MIF. When drillhole data or refined structural models are absent, we find that the IRIF models provide a magnetic distribution that is far more consistent with broad-scope geological trends than the aforementioned smooth inversion. Lastly, we find that constraining the inversion in regions where we have a priori information leads to a model that is consistent with

known geology and can better elucidate areas yet to be explored – mitigating the risk and cost of drilling. References Gregory, Alan F., Geology of the iron occurrence, Deposit A, Great Whale Iron Mines Limited near Great Whale Québec, August 1958, (GM 10487). Met-Chem Canada Inc., Technical Report on Great Whale Iron Property for Niocan Inc., Reference No.: 26039, August 2006 St-Hilaire, Camille, Levé Électromagnétique et Magnétique Héliporté, Secteur de la Grande Rivière de la Baleine, Bloc Fagnant, Novembre 1998, (GM 56267). Symons, D.T.A., and Stupavsky, M. Magnetization Characteristics of Algoman Banded Formations and Deposits in Ontario, 1983, (Ontario Geological Survey Open File Report 5447). Acknowledgements The authors would like to thank Elizabeth Baranyi and Darren Andrews of Geosoft Inc. for their assistance with this study.


Figure 4: Unconstrained susceptibility model (clipped spatially) compared to geological wireframe (grey).

Figure 5: Unconstrained IRIF susceptibility model (clipped spatially) compared to geological wireframe (grey).

Figure 6: IRIF susceptibility model with simple upper bound and depth weighting constraints (clipped spatially) compared to geological

wireframe (grey).

Figure 7: Geologically constrained susceptibility model (clipped spatially) compared to geological wireframe (grey).

�� Mathieu Landry*, Raglan Mine, Glencore

Summary

Geophysical methods play an important role in exploration strategies worldwide. The addition ofsuch data can contribute to a better understanding of the geological sector of interest. By carefully applying basic principles, explorers can maximize their chances for success: 1) Seek a rigorous understanding of the physical properties of the rocks under investigation 2) Associate geophysical data with geological data and expertise to better constrain the geological model 3) Use creativity to interpret new data and/or re-interpret historical data in order to provide new insights about the geological processes being studied.

Introduction

The advent of large-scale airborne geophysical surveys and especially the regional mapping of magnetism have undoubtedly helped to improve the quality and consistency of geological mapping worldwide (Nabighian et al., 2005) & (Thomson et al., 2007). This is demonstrated conceptually in Figure 1. By integrating geological expertise with a more quantitative, continuous dataset, it can be argued that the fidelity of interpretation is improved and that identification of significant features for the explorer becomes more reliable.

Fundamentally, the explorer’s problem is geological.The goal is to understand a thermo-chemo-mechanical process that happened epochs to eons ago in order to maximize efficiency for quick and great discoveries. It is important to determine how best to complement the exploration challenge with geophysical data such that the most useful insights can be gained for exploration success.

Improving geological perspective by combining geophysical data

The ultramafic rocks associated with NiS mineralization at Raglan are predominantly composed of serpentinized peridotite, which exhibits a high magnetic intensity (Beard et al., 2009), (Gunn, 1997),(King, 2007). The magnetic signature at Raglan was used to develop and improve the spatial representation of a thermo-mechanical erosion modelby Lesher (2007) analogous to meandering flows,

Figure 2: Simple conceptual representation of geological mapping improvement with the use of magnetism mapping

Figure 1a Schematic representation of outcrops location of 2 hypothetical rock types.

Figure 1b Schematic representation of a hypothesized geological mapping using simple contouring rules.

Figure 1c Total magnetic intensity (TMI) mapping

Figure 1d Schematic theoretical improvement to the geological map based on the integration of TMI data

Geology & geophysics; an inevitable match for exploration success

such as sinuous ‘rilles’, as observed on the moon and other planets (Lesher et al., 2011). As can be seen from Figure 2, this concept can be considered a valid exploration framework for the Raglan Mine property.

An Aerodat frequency domain electromagnetic (FDEM) airborne survey was flown in 1989 over the Raglan property. The magnetic response of the ultramafic rocks of interest causes permeability

Figure 2a Total field magnetic map for the central part of the Raglan Belt between the Zone 2-3 and Boundary complexes (from Osmond and Watts, 1999), imaging the magnetite-rich serpentinized peridotite complexes. The surface expressions of the complexes are outlined in black and appear to be connected down-dip to the north beneath overlying Chukotat Group basalts. The magnetic bodies in the underlying Povungnituk Group to the south are sills (Lesher, 2007).

Figure 2b Meandering lava channel exploration model for the ores in the Raglan Belt (from Green and Dupras, 1999).


effects that disturb the electromagnetic response of FDEM systems (Huang and Fraser, 1998), such that it causes negative conductivity anomalies. By taking advantage of this peculiarity, a re-interpretation to combine the magnetic and electromagnetic components of this survey has allowed production of a new mapping product that enhances the perception of the thermo-mechanical sinuous rille model by Lesher et al. (2011). As shown in Figure 3, the improved interpretation can be used to identify where the intrusion of interest eroded extensive pyrrhotite-rich sedimentary layers of the interpreted proto-basin mapped by the conductivity response.

Considering geological model reviews based on new geophysical insights Some materials exhibit a high target/host contrast in certain physical properties. The relatively high conductivity of Raglan ore makes exploration by electromagnetic techniques, especially inductive techniques, a preferred method (Watts, 1997), (Osmond et al., 2002) and (King, 2007). Some EM techniques thus permit clear direct detection of discrete NiS ore deposits near-surface (defined in this case as <400m). Exploration with such techniques becomes more difficult with increasing depth due to field diffusion laws & associated permeability and skin depth effects. To circumvent this challenge, the explorer might consider the use of audio magnetotelluric (AMT) methods based on the attractive depth penetration capabilities assumed from the tremendously low frequencies used. Unfortunately, the theoretical limits of the AMT depth of maximum sensitivity as a function of frequency and conductivity does not indicate the possibility to detect relatively small, Raglan-type, high-conductivity ore

bodies at great depths (McNeice & Stevens, 2006) and (Chouteau et al., 2007). However, through creative association, the explorer can still extract value from this method for indirect mapping of subsurface structures in order to improve the geological understanding. In the case of Raglan, some features identified from a Titan24 MT survey completed in 2006 revealed geometries displaying diapiric form (Figure 4) that is common in evolving thermodynamic systems under Rayleigh-Taylor instabilities, such as magmatic intrusions (Weinberg, 1994), (van Keken, 1997) and (Diez, 2009). This serves as a basis to consider different and novel

models and may increase the chance of success continuity in a maturing exploration camp.

Figure 4 Potential ‘diapiric’ geometries (?) imaged from 2D inversions of Titan24 MT data at Raglan.

Figure 3 Composite product of an Aerodat FDEM survey using TMI and the co-axial in-phase response of the 935Hz frequency to highlight highly magnetic serpentinized peridotite associated with economic mineralization (in pink with blue halo), interpreted to cut through moderately conductive extensive Po-rich sedimentary layers represented by the green shaded relief ridges.


Conclusions Many different geophysical techniques and configurations are available on the market today. Choosing the right method at the right time and place largely depends on the commodity being investigated, the geological environment complexity and the timing relative to the maturity of the exploration play. By understanding the close relationship of the physical properties of the rocks being investigated, the selection and timing of the geophysical methods being used can be optimized. Furthermore, by having a robust and creative process to integrate geophysical and geological data together, explorers can ensure greater fidelity to the evolving geological model and maximize their chances of success over time as the exploration play matures. References M. N. Nabighian, V. J. S. Grauch, R. O. Hansen, T. R. LaFehr, Y. Li, J. W. Peirce, J. D. Phillips, and M. E. Ruder, 2005, The historical development of the magnetic method in exploration: 75th Anniversary, SEG, Geophysics, Vol. 70, NO. 6 (November-December 2005); P. 33ND–61ND. Thomson, S., Fountain, D., Watts, T., Airborne Geophysics – Evolution and Revolution, 2007, Plenary Session: The Leading Edge, Paper 2, In "Proceedings of Exploration 07: Fifth Decennial International Conference on Mineral Exploration" edited by B. Milkereit, 2007, p. 19-37. Beard, J.S., Frost, B.R., Fryer, P., McCaig, A., Searle, R., Ildefonse, B., Zinin, P., and Sharma, S.K., 2009, Onset and Progression of Serpentinization and Magnetite Formation in Olivine-richTroctolite from IODP Hole U1309D, Journal of Petrology, Volume 50, Number 3, P. 387-403, 2009, doi:10.1093/petrology/egp004 P.J. Gunn, & M.C. Dentith, 1997, Magnetic responses associated with mineral deposits, AGSO Journal of Australian Geology & Geophysics, 17(2), p.145-158 King, A., 2007, Review of Geophysical Technology for Ni-Cu-PGE deposits, Plenary Session: Ore Deposits and Exploration Technology, Paper 45, In "Proceedings of Exploration 07: Fifth Decennial International Conference on Mineral Exploration" edited by B. Milkereit, 2007, p. 647-665

Lesher, C.M., 2007, Ni-Cu-(PGE) Deposits in the Raglan area, Cape Smith Belt, New Quebec, in Goodfellow, W. D., ed., Mineral Deposits of Canada: A Synthesis of Major Deposit-Types, District Metallogeny, the Evolution of Geological Provinces, and Exploration Methods: Special Publication No. 5, Mineral Deposits Division, Geological Association of Canada, p. 351-386. Osmond, R., and Watts, A., 1999, 3D geophysical model of the Raglan Belt, in Lesher, C.M., ed., Komatiitic Peridotite-Hosted Fe-Ni-Cu-(PGE) Sulphide Deposits in the Raglan Area, Cape Smith Belt, New Québec: Laurentian University, Mineral Exploration Research Centre, Guidebook Series, v. 2, p. 185-190. Green and Dupras, 1999. Green, A.H., and Dupras, N., 1999, Exploration Model for Komatiitic Peridotite-Hosted Ni-Cu-(PGE) Mineralization in the Raglan Belt, in Lesher, C.M., ed., Komatiitic Peridotite-Hosted Fe-Ni-Cu-(PGE) Sulphide Deposits in the Raglan Area, Cape Smith Belt, New Québec: Laurentian University, Mineral Exploration Research Centre, Guidebook Series, v. 2, p. 191-199. Williams, D.A., Kerr, R.C., & Lesher, C.M., 2011, Mathematical modeling of thermomechanical erosion beneath Proterozoic komatiitic basaltic sinuous rilles in the Cape Smith Belt, New Québec, Canada, Mineralium Deposita International Journal for Geology, Mineralogy and Geochemistry of Mineral Deposits, ISSN 0026-4598, Volume 46, Number 8, p. 943-958. Huang, H., and Fraser, D.C., 1998, Magnetic permeability and electrical resistivity mapping with a multifrequency airborne EM system, Exploration Geophysics 29, p. 249-253. Watts, A., 1997, Exploring for Nickel in the 90s, or ‘til depth us do part’, Integrated Exploration Case Histories, Paper 132, In “Proceedings of Exploration 97: Fourth Decennial International Conference on Mineral Exploration” edited by A.G. Gubins, 1997, p. 1003–1014. Osmond, R.T., Watts, A.H., Ravenhurst, W.R., Foley, C.P., and Leslie, K. E., 2002, Finding nickel from the B-field at Raglan – ‘To B or not dB’, CSEG Recorder, Focus Article, p. 45-47. McNeice, G. & Stevens, K., 2006, MT Workshop – Laval, 0306, Falconbridge internal material.


Chouteau, M., Boulanger, O., and Giroux, B., 2007, Detection of Deep Conductive Massive Orebodies by Magnetotelluric Surveys, DIVEX Project SC25, Annual Report. Weinberg, R.F., and Podladchikov, Y., 1994, Diapiric ascent of magmas through power law crust and mantle, Journal of Geophysical Research, Vol 99, No, B5, P. 9543-9559. van Keken, P., 1997, Evolution of starting mantle plumes: a comparison between numerical and laboratory models, Elsevier, Earth and Planetary Science Letters 148, p. 1-11. Diez, M., Connor, C.B., Kruse, S.E., Connor, L., and Savov, I.P., 2009, Evidence of small-volume igneous diapirism in the shallow crust of the Colorado Plateau, San Rafael Desert, Utah, Geology, August 2012, v. 40, p. 695-698.

Revitalizing near mine and regional exploration using 3D integrated analogous targeting at the Granduc Property, BC Julie Palich – Caracle Creek International Consulting Inc.* Vlad Kaminski – Caracle Creek International Consulting Inc. Brad Leonard – Castle Resources Inc. (CRI)

Summary

3D integrated analogous targeting, whereby new targets are identified based on their geologic and geophysical similarity to known orebodies should be considered an essential step to exploring in and around historical mining camps. This approach has been used for the exploration of the Granduc Property located in the mountainous terrain of northwestern British Columbia, Canada. Existing geologic and geochemistry data were combined with modern methods of geophysical interpretation using the GoCAD 3D visualization platform to provide a cost effective means for justifying and prioritizing the exploration and development of new targets. 3D inversion of the magnetics data and 2D layered earth imaging (LEI) of the electromagnetics data were undertaken to provide a detailed understanding of the electromagnetic (EM) and magnetic signature with respect to the known geology, mineralization and structure associated with the Granduc deposit. This primary integrated model has subsequently been used for the evaluation and prioritization of regional EM and magnetic targets identified from the 2D data collected in the airborne survey.

Introduction

The historic Granduc Property, located 40 km northwest of Stewart in the steep mountainous terrain of northwestern British Columbia, has been the target of exploration programs since the 1930’s. The Granduc deposit is an aerially extensive Besshi-type copper-silver volcanogenic massive sulphide deposit. Exploration in the 1950’s led to the development and operation of the Granduc Mine from 1968 – 1984, resulting in the extraction of approximately 15.42 million tonnes of copper ore. Mining operations at Granduc were closed due to low metal prices and not due to a lack of ore. Following closure of the mine, limited exploration was undertaken for the next 20 years. Since 2010, the Property has been explored by Castle Resources Inc. (CRI), who have embarked on an aggressive exploration program including over 60,000 metres of diamond drilling and a comprehensive 3D study of available information to enhance exploration in the area.

The focus of the current exploration program is to grow the “footprint” of the deposit as well as target potentially new mineralized areas. Initial review of the historical mapping,

drilling and assay data, and preparation of a compliant resource estimate by CRI has led to the identification of a number of geologically-favourable targets. In 2010, the geologic data were complemented by a helicopter-borne AeroTEMIII time-domain electromagnetic and magnetics survey over the property. Preliminary interpretation of the 2D airborne survey data confirmed the spatial coincidence of EM and magnetic highs over the Granduc deposit.

Geology and Mineralization

Stratified rocks exposed on Granduc mountain and to the north are subdivided into two easily recognizable units, termed the Western and Eastern series, that are separated by the north-northwest striking South Unuk shear zone. The older Western series rocks consist of deformed and foliated, greenschist facies metavolcanic and metasedimentary rocks, including the Granduc Mine series (McGuigan and Marr, 1979), units lying north of the North Leduc glacier and units in the hanging wall of the Granduc fault on Granduc mountain. The younger Eastern series rocks are much less deformed and are mainly volcanic. The boundary between western and eastern series rocks is easily identifiable north of Granduc mountain, however on Granduc mountain itself, the boundary is uncertain. These Eastern series rocks belong to the Lower to Middle Jurassic Hazelton Group. The mapped geological associations are shown in Figure 1.

The principle exploration target on the Property is for volcanogenic massive sulphide (VMS) deposits of the Besshi-type. Besshi VMS sulphide mineralization at Granduc consists predominantly of iron sulphides (pyrite and/or pyrrhotite), with lesser chalcopyrite intimately associated with magnetite; sphalerite and galena may or may not be present. Besshi deposits can contain highly varied and complex sulphide mineralogy, including arsenopyrite, galena, bornite, tetrahedrite-tennantite, cobaltite, stannite and molybdenite. Quartz (chert), tourmaline, carbonate, albite, sericite, chlorite,amphibole can be found as gangue minerals in the deposits.

The principle copper-silver bearing volcanogenic massive sulphide zones on the Property are located in a north-south trending, steeply westward dipping strata of the Late Triassic Western series (Stuhini Group). Two major mineralized zones have been identified through drilling - Granduc deposit (Main Zone) and the North Zone.

3D integrated analogous targeting at the Granduc Property, BC

Geophysical Data Acquisition, Processing and Inversion

A helicopter-borne geophysical survey was undertaken using Aeroquest's AeroTEM III time domain helicopter electromagnetic system deployed in conjunction with a high sensitivity caesium vapour magnetometer (Kahue, Bishop and Mirza 2010). Full-waveform streaming EM data were recorded at 36,000 samples per second. The streaming data comprised the transmitted waveform, and the X component and Z component of the resultant field at the receivers. The streaming EM data along with ancillary data were recorded with AeroDAS acquisition system. The survey over the Granduc block comprised 2376 line-km flown at 75 metre line spacing and in 90º/270º flight direction with an average flight height of 68 m.

An unconstrained magnetic inversion was undertaken using the UBC-GIF Mag3D code. Prior to inversion the total

magnetic intensity (TMI) data were corrected to the International Geomagnetic Reference Field (IGRF) and reduced to accommodate every 10-th station along profile for the final inversion. The deicmation of the original data sets increased the along-profile sampling distance to approximately 20-25 meters.

The inversion was undertaken using a 50 m3 mesh with a padding of 1900 m, resulting in 9,884,100 cells. The inversion was executed allowing for a 7 nT data misfit. Daubechies-2 wavelet compression was used for the sensitivity matrix compression with standard Mag3D parameters for relative threshold and relative reconstruction error. The initial magnetic susceptibility model was set to 10-7 SI units and the reference model was set to zero.

Conductivity Depth Sections (CDS) were generated from the EM data using a layered-earth inversion (LEI) algorithm (Farquharson and Oldenburg, 1993, Ellis 1998) utilized by Condor Consulting Inc. The LEI algorithm modeled the EM data with a 28-layered earth model increasing in thickness from the surface to depth in an approximately logarithmic fashion. For this project the first layer was 5 m thick while the deepest was 232 m thick. A starting model of 1,000 ohm-m (1 mS/m) was used, with a reference model of 10,000 ohm-m (0.1 m S/m). Two runs were performed; one using only the Z off-time channels and a second run where all the channels (Z on and off-time) were processed. The primary data were by a factor of six prior to calculating the inversion model.

Integrated Geological and Geophysical Modelling

The Mira GoCAD platform was used for the integration of the existing geological and geophysical data into a 3D visualization workspace. Table 1 summarizes the geospatial data incorporated into the model. Table 1: Summary of data included in the 3D model. Data Type Information Cultural Claims/Crown grants, zones of

interest Topography Digital Elevation Model (DEM) HistoricWorkings

Granduc mine tunnels, Mined out areas

Geology Regional geology, structural, modeled faults, modeled lithology, mre shells

Drilling Diamond drillhole collars and traces, lithology, assays

Geochemistry Au, Ag, Cu, and Fe assay points isosurfaces for Au, Ag, Cu, Fe, within ore shells

Geophysics 2D TMI, VD1 & Zoff EM contour maps, 3D magnetics inversion, EM LEI profiles

Figure 1: Property geology of the Granduc Mine in northwestern British Columbia, Canada. (McGuigan and Harrison 2010)


Once the available data were imported into the GoCAD model, surfaces were constructed for geology, topography, and structural data. Drilling and drill sections were interpreted and 3D lithology wireframes were constructed for the units of interest. Wireframes were created for the key lithologies. Structural data from the drilling logs were examined and major faults associated with the Main Zone were constructed. Geochemical data from drill assay results were interpolated in 3D to provide interpolation between sample locations and across the lithological units to provide grade estimations for Au, Ag, Cu, and Fe in percent (pct) or parts per million (ppm).

The magnetic inversion results were imported as geospatially referenced 3D grids and isosurfaces were generated for magnetic susceptibilities of 0.002, 0.004, 0.006, 0.008, 0.010, 0.020, 0.030, and 0.050 SI units (representing the magnetic susceptibility data range).

The 2D LEI inversions generated array-based data files that could not be directly imported into GoCAD. As a result the inversion data were converted into a GoCAD compatible, geospatially referenced format using an internally-developed script where data were converted into grid-centred pointsets. The LEI data were presented to a depth extent of 600 m below ground surface. This is deemed considerably deeper than the AeroTEM III system is penetrating (estimate 200- 300 m depth below ground surface depending on the geometry of the target), however a number of the inversion results appear deeper than the effective depth of investigation and were validated by known drilling information. 3D inverse distance gridding was subsequently undertaken using an isotropic search ellipse of 75 m x 75 m x 75 m. Isosurfaces were then generated in 2 mS/m increments for conductivities ranging from 7 – 15 mS/m.

Integrated Interpretation

Joint interpretation of the magnetics data with the geologic information indicates that the main controlling structure is the South Unuk Shear Zone (SUSZ), which is striking in the NNW direction and dividing the stratigraphic units into the Eastern and the Western series. To the west of the SUSZ, the magnetic signature shows a complex sequence of magnetic lineations that map the complex geology and faulting associated with the western series stratigraphy. To the east of the SUSZ, the magnetic signature is more subtle due to the extensive glacial cover, and more uniform stratigraphy of the eastern series units interpreted to be present in this area.

There appear to be several magnetic and EM anomalies associated with the Main and North Zones. Most notably, a magnetic high (greater than 0.04 SI) is present coincident

with Granduc mine exhibiting a clear association with the massive sulphides deposited at depths bound by 500 mAHD and 900 mAHD. The southern extent of this anomaly can be followed underneath the South Leduc glacier.

A conductive anomaly (greater than 0.13 mS/m) also occurs approximately coincident with the oreshells developed for the Granduc Mine based on known drilling and mining information (Figure 2). The conductive anomaly is interpreted as having a shallower extent and dip compared to the modeled oreshells. This feature is also intimately associated with the Granduc Fault on the west side and the B Fault to the east (which effectively is the eastern limit of the Granduc orebody) within the Main Zone. As with the modeled oreshells, this feature has very little extension east of the B Fault. While the majority of the body has a depth extent of 300 m, it is interpreted as extending to greater depths to the north over an area that does not appear to have previously been drill tested. Several smaller conductive features, not previously drill tested, are also present along strike to the south of the Main Zone on the east side of the Granduc Fault within Granduc Mine Series units.

The mineralized zone within the North Zone is characterized by a moderate magnetic susceptibility feature (0.03 – 0.05 SI units) with no apparent coincident EM anomaly. The magnetic anomaly is hosted within two stratigraphic units: a green siliceous phyllite hosting minor calcareous horizons with lenses of epidote and dark grey-green augite-bearing andesite flows and minor tuff. The anomaly continues beyond the known drilling in the northern direction (underneath the glacier). In general the susceptibility of this body is lower compared to the Main Zone (0.04 SI) and the body is rather deep seated (between 750 m horizon and 1000 m horizon), however there is a plume of highly susceptible material extending upward to reach 1700 mAHD. This body has good correlation to mineralization with the upper extent of the mineralized area, but has been modelled as more vertically-dipping than the ore zone at depth, which may be an artifact of the unconstrained inversion process.

From an analogous targeting standpoint, several other comparable magnetic and EM anomalies in the surrounding area beyond the extent of previous drilling and mine workings were evaluated for their mineralization potential. From the integrated interpretation, the highest priority targets were a magnetic anomaly identified along strike to the north of the North Zone anomaly, a coincident magnetic and EM response along strike to the north of the Main Zone, the smaller EM features located to the south of the Main Zone, and a large coincident magnetic and EM feature located in a remote area on the north side of the


North Leduc Glacier, representing a greenfields regional target (Figure 3).

Conclusions

Use of integrated geological and geophysical interpretation of the airborne data collected over the Granduc Property has confirmed the association of known mineralization with both anomalous magnetic and EM features. From a geophysical standpoint, the ability to correlate these anomalies to known geologic and assay data enables validation of the inversion results as they confirm the correct depth weighting used in the 3D inversion. By evaluating other near mine and regional magnetic and electromagnetic anomalies in terms of their analogous characteristics to the geologic/geophysical characteristics of the Granduc Mine, it is possible to provide data-rationalized targeting and improved prioritization for future exploration and drilling programs.

References

Ellis, R. G., 1998, Inversion of airborne electromagnetic data, 68th Ann. Internat. Mtg: Soc. of Expl. Geophys., 2016-2019.

Farquharson, C.G. and Oldenburg, D.W., 1993, Inversion of time-domain EM data for a horizontally layered earth, Geophysical Journal International, Vol. 114, pp 433-441.

Kahue, C., Bishop, M. and Mirza, A., 2010, Report on a Helicopter-Borne AeroTEM System Electromagnetic Magnetic Survey Granduc & Tunnel Project Stewart, B.C., Canada; for Bell Copper Corporation; Aeroquest Airborne Ltd report #10016, December 2010.

McGuigan, P.J. and Harrison, D.J. (2010), Technical report: Granduc property, Northwestern British Columbia, Canada. Prepared for Castle Resources Inc.

McGuigan, P.J. and Marr, J. (1979), Surface geology of the Granduc Mine Area Progress Report, Esso Minerals Canada Ltd. Unpublished company report.

Acknowledgements

The authors would like to thank Castle Resources Inc. for providing permission to publish the results of this case study, and their constructive comments and review of its content.

Figure 2: Main Zone and south extension conductors (>13 mS/m) looking south superimposed on the Granduc Fault (green), B Fault (yellow), interpreted massive iron formation and oreshells.

Figure 3: Regional target viewed from the southeast showing a conductivity isosurface of 13 mS/m (red) and magnetic susceptibility isosurface of 0.03 SI units (green).

Geophysical Processing and Interpretation with geological controls: Examples from Bathurst Mining Camp Peter Tschirhart, Sander Geophysics, Ottawa, Bill Morris*, SGES, McMaster University, Hamilton, and Hernan Ugalde, PGW, Toronto. Summary The primary link between geology and geophysics is petrophysical property measurements. Physical properties are populations. A geological lithology may be represented by a uni- or multi-modal physical property distribution. Source Edge Detection routines applied to geophysical data provides geologists with a basic framework for constructing lithological maps. Edge stacking and data thresholding improves the resolution of source edges. Integrating multi-parameter edge data provides a lithological context. Regional – residual data processing is essential to highlight near surface sources. Apparent susceptibility mapping derived from HFEM surveys discriminates shallower sources and locates regions of enhanced remanent magnetization. Inversion of magnetic anomalies occasionally yields an estimate of the orientation of the remanence vector. Comparison of this vector with an APWP predicted remanence vector provides information on the age of the deposit and its tectonic history. Introduction The Bathurst Mining Camp, (Figure 1) has been an important mining camp for over 50 years, hosting 46 volcanogenic massive sulphide (VMS) deposits including the supergiant 121 Mt Brunswick #12 deposit. Geologically the BMC comprises volcanics, intrusives and sediments which record the Late Ordovician – Silurian development, closure, and subduction of the Tetagouche – Exploits back-arc basin. Geological mapping across the BMC is limited by the less than 1% outcrop. Geophysical exploration techniques have played an important role in the discovery of all the major deposits. Many of the original deposits were found by electromagnetic surveys which located mineralization by the resistivity contrast between the conductive sulphides and the resistive host rock. Recent gravity surveys have sought to find ore bodies based on density contrasts. Aeromagnetic and radiometric surveys have been used to assist the geological mapping. Mine development has provided a large borehole database. Combined these resources provide an excellent resource for assessing a number of our fundamental assumptions regarding the link between geology and geophysics.

Physical Rock Property Database Geophysical surveys record the spatial variation of physical rock properties. Typically these include density, magnetic susceptibility and resistivity. Often physical property fluctuations do not translate directly into lithologies mapped by geologists. Lithologies are based on mineralogical and grain size variations. A physical rock property data base then is an essential element in linking geologic and geophysical interpretations. However, for the rock property database to be useful it must contain reliable, repeatable calibrated data.

Figure 1. Geological map of the Bathurst Mining Camp showing distribution of individual nappe sheets. Sampling points for physical rock property measurements are indicated by dots and borehole identifiers. Physical property measurements are made on individual outcrops, on drill core, or by logging a borehole. This creates two problems; 1) nature of the rock, weathered versus fresh and 2) instrumentation. There are a range of instruments available for measuring each physical property. While tools may appear to be similar details of their operation might differ. For example, magnetic susceptibility meters have varying coil dimension and signal frequency. More critical is when a physical property is measured by different techniques. Density on core is usually based on the ratio of two measurements of weight, while density from core logging is derived from gamma radiation measurements.

Geology, Geophysics, Petrophysics and Models: Bathurst

Physical rock properties are populations. One measurement of a physical property is insufficient. One should always apply statistical methods to investigate the range of values that is representative of a lithology. Is there evidence for more than one population? If more than one suite of measurements is available then comparing results provides some indication of the reliability of the reported average values.

Figure 2. Statistical representation of density variations recorded by three independent surveys of rhyolite samples from the Bathurst Mining Camp. A) Frequency histogram, B) Cumulative frequency / Probability plot, C, D, E) Box and whisker plot of all and sub-categories of rhyolite. Three separate suites of physical rock property measurements have been collected for the Bathurst Mining Camp. Figure 2 presents results for rhyolites. All three collections report a similar average density. The bore hole logging results detect a lower density sub-set not detected by the two core based surveys. When modeling this unit one might have to invoke two possible density scenarios. Cross-plots of density and magnetic susceptibility commonly result in two distinct trends; a paramagnetic group and a ferromagnetic group. This is important when thinking about magnetic modeling since paramagnetic sources only have induced magnetic fields while ferromagnetic sources may also have remanent magnetic components. Source Edge Detection Airborne geophysical surveys provide spatially continuous regional data coverage which directly reflects petrophysical differences and thus the underlying geology. A modern geologic mapping exercise involves fusion of this information with what is typically limited outcrop mapping.

Source Edge Detection (SED) routines provide a quantitative approach to locating the spatial distribution of lateral changes in any physical property. The underlying thesis behind SED is that a physically uniform lithology will be associated with a homogeneous geophysical response. The boundary between two such zones is marked by an inflection in the observed signal. All SED routines apply a transformation which converts the local signal inflections to local maxima. After applying this transformation one then isolates all ridge peaks and assigns them as being representative of a source edge.

Figure 3. Compilation of source edge detection routines applied to magnetics (blue), electromagnetics (red) and gravity gradiometry (green). A) RGB composite, B) CMY composite, C) edge pixel distribution A number of different algorithms to compute the transformation from inflection to ridge have been published. When applied to the same data set the “mapped edges” detected by the different SED algorithms are similar but not identical. It is proposed that the optimum expression of the location of a source edge is defined by an area which has the highest density of solutions in a compilation of all source edge products. Edge stacking can also be extended to include inputs from multiple surveys of the same physical property. For example, it is common to have more than one aeromagnetic survey for an area. Individual surveys may differ in terms of line spacing, flight direction, and flight height. They will differ in terms of noise content. The geologic contacts (edges) are constant so a summation of edges computed using multiple edges from multiple surveys should accentuate valid solutions. Integrating edge detection from multiple physical properties (magnetics, gravity and electromagnetics) highlights boundaries which are associated with varying physical property contrasts respectively density, susceptibility and resistivity (Figure 3).


Regional – residual separation The geophysical signal recorded at any point in a survey represents the sum of the contributions from all sources weighted by their distance from the observation point and the amplitude of the anomalous source. For the purposes of geological mapping one ideally wishes to see only the contribution from near surface sources. Regional – residual separation is a processing step that is applied to all gravity and magnetic data sets. The most common approach employing some form of wavelength filtering assumes that long wavelength anomalies represent deep sources while shorter wavelength correspond to shallow sources. It is possible that some longer wavelength anomalies might be near surface sources. An alternative approach to regional – residual separation for near surface magnetic sources is presented using frequency domain helicopter electromagnetic (HFEM) data. Using a procedure first developed by Fraser (1981) we derive a grid of apparent magnetic susceptibility from the HFEM data. Using the MSHFEM along with known topography and original flight path data we can calculate a magnetic intensity grid by forward modeling (Figure 4). There are two immediate benefits to this approach. First, HFEM systems have a limited effective depth of penetration, within the first hundred meters from surface. Any magnetic methods detected by this method must be located in the near surface. Second, the HFEM derived susceptibility is completely independent of magnetic remanence. In contrast apparent susceptibility computed from the original magnetic intensity data incorporates all magnetic signal sources in its derivation. Cross-plotting of MSHFEM versus MSTMI serves to reveal areas where the observed magnetic field is dominated by magnetic remanence and provides an estimate of the polarity of the remanence contribution. Applying this method to a subset of the Bathurst Mining Camp shows that the majority of the total magnetic field is mapping near surface sources. The procedure locates a region of remanent magnetization. Issues associated with this approach arise in the presence of near surface conductive bodies and anthropogenic sources.

Figure 4 Comparison between measured and modeled magnetic anomalies. Left: residual magnetic anomaly. Right Forward model magnetic anomaly

Magnetic Model Interpretation The magnetic anomaly pattern created by a source body anomaly is controlled by three sets of parameters: 1) the location and depth of the source, 2) the geometry of the source body, and 3) the magnetic properties of the source. Locally enhanced concentrations of paramagnetic minerals produce an increased magnetic susceptibility which in turn produces an induced magnetic field anomaly pattern whose geometry is moderated by the orientation of the present earth’s field at the observation point. With enhanced concentrations of ferromagnetic minerals it is quite possible that some of the observed magnetic signal could be caused by the presence of remanent magnetisation. Magnetic remanence is acquired under a number of very specific conditions. In a ocean margin tectonc setting remanence could be acquired when an igneous magma cools, and during thermochemical processes related to regional metamorphism and deformation. Perhaps most critical in a mining camp setting, remanence may record the hydrothermal event associated with emplacement of the ore body. Just as the Present Earth’s Field has a specific orientation (declination, inclination and intensity) for a specific location on the earth’s surface, a remanence vector records the orientation of the local magnetic field at the time of remanence acquisition. Changes in the orientation of the magnetic vector versus time are provided by paleomagnetic studies as recorded in Apparent Polar Wander Paths. Knowing the location of the observation site and the approximate age of the rocks it is possible compute probable representative declination and inclination values. Locally the orientation of the remanence direction is also controlled by the timing of remanence acquistion relative to the timing of folding of the strata. If remanence acquisition predates folding then the remanent vector is rotated in the same sense and magnitude as the bedding dip and strike. Simple trigonometry requires this vector must lie on a small circle with an axis parallel to bedding strike. When computing a magnetic model of a source one is often unsure of the relative magnitude of the induced and remanent magnetisation components. Fluctuations in the Koenigsberger ratio cause the computation to generate a remanence direction solution which lies somewhere on a great circle between the PEF and the remanence direction. Borehole intersections constrain the geometry of the Armstrong B deposit to be a steeply dipping sheet. Comparison of the model derived remanence vector and the APWP derived remanence vector confirm that this deposit cooled below 350oC at around 430 Ma. And it has not be substantially rotated in any way since this All magnetic model interpretations involve an objective comparison


between the observed and the computer generated magnetic anomaly pattern. The best representation of the geology present in the subsurface is mathematically defined by the optimum match, or the minimum Root Mean Square Error. It is important to recognize that our geophysical data may not be capable of correctly outlining the geometry of finite source bodies.

Figure 5. Orientation of possible remanence directions for New Brunswick during period 410 to 470 Ma. Bedding strike and dip direction indicated by dotted line.

Figure 6. Remanence inversion input and results. (A) The observed magnetic signal of the Armstrong B deposit. (A’) Input geophysical model with reference information. (B) Resultant magnetic signal. (B’) Resultant geophysical model. A starting model of the Armstrong B mineral deposit is constrained by lithologic logs from a number of boreholes. Geologically constrained inversions of magnetic data

incorporate some prior knowledge regarding the lithological, structural and/or petrophysical characteristics of the source target. Inclusion of this information helps focus the computed model towards a more geologically meaningful solution. Unfortunately, in many cases the inversion will have difficulty accurately replicating the geology (Figure 5). The product of all supra-surface magnetic surveys is a 2D grid which attempts to represent the magnetic intensity variations on the observed surface. There are two fundamental limitations. First, observations are taken at incremental points along a line. Point spacing is always less than between line spacing. And it is the line spacing that controls the cell size used in the gridding algorithm. Second, airborne surveys occur at some level above the ground mean that signal smoothing will occur. The deposit is a sub-vertical slab. Physical property measurements show the magnetic signal is driven by pyrrhotite associated with the ore body. A 2004 Fugro Megatem survey flown at mean terrain clearance of 120m data data over the deposit was collected in two separate data blocks. Stitching the two datasets together increased spatial resolution of the magnetic signal. During the inversion computation the estimated width of the magnetic source body is increased from 25m to 163.9m. There are two possible explanations. First, there is a fundamental spatial resolution imposed by the survey. Or, second, the anomaly is describing the orebody and its alteration halo which may also contain enough pyrrhotite to register as being anomalous. Borehole magnetic surveys suggest the first alternative is more likely. Conclusions Rock property measurements are essential. They should be treated as populations. Integration of the output of multiple source edge routines applied to variable data sets provides an improved estimation of geological patterns. Employing apparent susceptibility derived from HFEM can accentuate near surface magnetic sources and isolate regions that are remanently magnetized. Comparing inversion model derived remanence with APWP derived estimates provides both temporal and tectonic information. Acknowledgments (Optional) Funding support for Peter Tschirhart was provided by NRCan TGI4 and a NSERC graduate scholarship. Funding for Hernan Ugalde was provided by NRCan TGI3 and Votorantim Mining. The New Brunswick Department of Mines provided financial and logistical support. None of this would have been possible without guidance provided by Cees Van Staal and Neil Rodgers of the Geological Survey of Canada, and Jim Walker, Reg Wilson and Stuart McCutcheon of NBDM. Pierre Keating. Mark Pilkington and Mike Thomas of the Geological Survey of Canada have been a constant source of support.

�� !��

Summary For 80 years or more, interpreters of potential field observations have "seen" faults, dykes, pipes, horst and graben, folds and shallow and deep basins. The challenge to be faced in this context is to reduce the ambiguity in these interpretations and to build more community confidence in the predicted geology. A more formal process of not only locating these features, but also attributing them with depth, strike, dip, thickness and property contrasts contributes to ambiguity reduction.The advent of much higher resolution observations, in the form of scalar and now vector and tensor gradients, allows for a significant reduction in ambiguity. The 2D form of a body is inherent in tensor gradiometry measures. The data from these newer systems is expanding exponentially. It is now possible join LIDAR, EM and potential field data into a common interpretation workflow that is automated. At the same time, computational geometry engines have evolved that can manage large scale representations of 3D geology, without the approximation of a gridded representation. Highly detailed shapes are computed that are not just thin plates, lines or simple bricks. The rise of implicit radial basis function technology and its companion cokriging, make it much easier and efficient to represent surfaces and arbitrary volumes. These can also be adapted to reflect the variable resolution of the potential fields with distance from the observations. These automated "feature extraction" objects result in 3D geology starting models that are much richer than just a "null half space". The benefit of also using the principles of structural geology in 3D to interpret the field mapping and the geophysics contributions, adds extra constraints.

Introduction Interpretation approaches for gravity and magnetic data continues to evolve. Better and cheaper instrumentation systems lead to compilations up to continental scale with resolution to sample spacing of 50m or less (Nakamura, 2011). These freely available 40 Gigabyte plus grids expose a ‘lack’ of ability to grapple with multi-scale investigations in current tool kits, without compromising the resolution (Figure 1). In recent “future of mineral resources exploration” discussions (Uncover, 2012), the realization has been made that to meet the looming resource shortage, an up scaling of geophysics technology has become a desperately needed

requirement.

Figure 1. Isostatically corrected gravity grid for Australia. Cell size 800m, Lambert Conic Conformal. Note, many primary lineaments, mostly deeply rooted in the mid-to lower-crust. Current blockers a. Sparse Data Despite all the geoscience progress to date, once we move away from the comfort of intensive drilling, or 3D seismic, we must make use of ‘sparse data’ methods, as the only information available is the airborne geophysics, topography maps and some surface geology mapping. The inherent ambiguity involved with both gravity and magnetic surveys is well known. However, these often form the only independent datasets available to illuminate the subspace. Structural geology principles are also one of the only other sources of constraint. b. Body shapes The use of structured grids in 3D geologies creates artificial barriers and impediments as 100,000,000 cells cannot do justice on a continental scale to the task of separating high resolution geology models that accurately reproduce known features at sufficient resolution. Traditional codes use simple cubes. To break the shackles, developments with triangulated surfaces are the new direction. This requires infinitely thin sheets, thin plates, and/or boundary elements or facets, to describe the range of new primitives for geology models. Thus, the multi-scale nature (fractal) of any geological setting needs to be respected in any 3D geology map. The intersections of the feature with the

��

topography, and estimates of the varying dip and thickness along the strike, are the first objectives. Following that are the extents. Ideally, some sequence of placement can also be gleaned. Dyke swarms and fault networks are a particular challenge. c. Computational Geometry Engines Collaborative efforts, often supported by large defense combines, continue the push to construct simpler, more powerful geometry computation engines to aid 3D scientific and engineering computation and visualization of the physical, tangible world. Examples are found in many countries, seeking to gain a competitive edge for their manufacturing sectors, or to solve complex engineering problems. Amongst these are Open Cascade, SOLIDWORKS, VTK and more recently CGAL. This enabling technology requires a huge investment in personnel and resources. Very little of the focus of these efforts are to refine and adapt methods that work for both geology and geophysics. The newer libraries are designed to accept implicit functions at the front end of the rendering pipeline. Geology, due to its multi-scale complexities, and its aliased sampling, requires special purpose mathematical functions for its description. First approximations of the geology map in 3D at various scales are now appearing in Australia, France etc. Systematic use of 3D geology/geophysics mapping that helps exploration is still elusive in the minerals industry. While these geology models are appearing, there is a call for supercomputers to deal with the cubes while doing inversions and geophysical calculations. This is unnecessary. Also, a range of computational geometry problems must be properly solved to enable the new way forward e.g. the clipping of the dykes/ faults with the topography. Methods a. Geology and Implicit Functions Geology bodies can be described by the use of implicit functions. This was first shown by Lajaunie, 1997 and is now becoming ubiquitous (Geomodeller, Leapfrog, GemCom etc). The key findings from France, is that the trend or gradient of the geology has a dominant role to play in realizing constrained 3D renderings. This is simply the measured foliation or strike / dip pairs. This means the implicit functions to be used should be informed by both contacts and gradients and use co-kriging or equivalent radial basis function, to optimize the use of the available ‘sparse data’ effectively. b.Geophysical Response A family of new geophysical responses techniques have been developed (Holstein 2009, Gotze, 1988), that address this emerging need. Ideally these can be represented anywhere in 3D space with adaptive resolution, to suit the

needs of the interpreter. A current development allows for variable properties within a triangle element, and this leads to an even coarser mesh being adequate for the geophysical response (Holstein 2014). c.Isoparametric Bodies Engineers, in the early days of finite elements, realized that discretized 3D geometry that also could serve as a way of carrying the Cauchy stress/strain approximations was a great simplification and serendipitous development. We are facing similar requirements in geophysics. The need is to generate geometry realizations of the geology models that are also close to optimal for computing the geophysical responses. The tendency is to over-specify or over discretize a model, even if the original data does not support that resolution. We advocate an adaptive resolution, fine at the surface, where we have some constraints and much coarser at depth (Figure 2).

Figure 2 3D surface intersecting with topography, faint blue triangles at full resolution, green triangles show the adaptiuve bevaiour with depth. The benefits are immediately obvious - • The model is in balance with the resolution of the

mapping and survey data • The rendered models have better than 90% reduction

in number of triangles, without loss of a balanced precision and resolution.

• Geophysical responses from each of the bodies can be seen in real time.

• Inversion of the ‘geology bodies’ while preserving topology significantly enhances the constraints.

Broadly, triangulated surfaces can be made to represent • Faults • Dykes with variable thickness • Boundaries of closed bodies

.

��

Examples The multi-scale methods known as ‘Worming’ (Hornby, et.al. (1999) & Florio et al(2013), ‘Naudy’ (1971) and ‘Euler’ deconvolution FitzGerald et al (2004) can each be significantly upgraded. The original inspiration comes from working with the ‘worming’ technology and upgrading it to handle full tensor gravity gradiometry (FTG) datasets. There is an immediately obvious improvement that derives from using measured gradients to those calculated. Interpreters who are wanting to work in 3D, find the ‘worms’ interesting, but would prefer properly registered ‘contact’ surfaces in 3D, not upward continued contoured lines. In particular, 2D seismic section interpretation can be greatly assisted by 3D fault networks from gravity. Thus, some of the art of the geophysical investigator for predicting the dip of fault structures from a profile of gravity data McGrath, (1991) immediately resonates with this work. This has been adapted and automated as it fits very neatly into the workflow for worming. The extra work for full tensor gradients, does not require the upwards continuation strategy. There is enough inherent information in the measured tensor to directly invert for dip, as frequently as required along the feature. FitzGerald, et al (2011) ie less ambiguity with tensors. a.Darling Fault Calibration Arguably, the most prominent interruption to Australia's geology is the Darling Fault. Three separate seismic traverses cut this dominant feature, running north south near Perth. The dip of this ‘fault zone’ is around 80 degrees to the west – see figure 3. The dip is estimated by creating gravity profiles over the fault every 100kms, and applying the McGrath algorithm. This calibration serves as a clear and definite target which validates this technology FitzGerald et al (2013). b. Northern Territory Fault network To further demonstrate the potential of this technique, the

Northern Territory of Australia was chosen to define a deep crustal fault network. Examples of this further calibration work are presented. c.Magnetic Data and Dykes A systematic study of dykes helps to understand and solve many geological problems. It helps to recognize Large Igneous Provinces (LIPs), particularly for the Precambrian period. Many LIPs can be linked to regional-scale uplift, continental rifting and breakup, and climatic shifts. In the Paleozoic and Proterozoic, LIPs are typically deeply eroded. They are represented by deep-level plumbing systems consisting of giant dyke swarms, sill provinces and layered intrusions. In the Archaean the most promising LIP candidates are greenstone belts containing komatiites. Detailed study of dykes is therefore considered to be an important tool for Paleo-continental reconstructions. Paleocontinental reconstructions are critical to provide a tectonic context for major ore deposits, the tracing of metallogenic belts between blocks, and identifying new prospective regions for mineral deposits of a wide variety of types. To date, being able to easily capture, in 3D, any representation of a dyke swarm, has been largely intractable. The quest for 3D dyke networks predicted directly from an analysis of magnetic TMI and/or full magnetic tensor data offers hope however. The simultaneous use of Naudy, VTK and implicit functions, significantly simplify this work and also make it possible to claim an upgraded capability. For each dyke, initial estimates of the centre line, linear extents and thickness variations, as well as susceptibility are made and then carried into the 3D geometry representations. The intersection of the predicted dykes with the topography can now also be routinely done, along with an adaptive triangulation with depth, so that the geophysical forward model can be compared to the observed signal. The method has always been applied to profile data of aeromagnetic data. This has been added too, by also now

��

supporting full tensor Gravity (in alpha) and full tensor magnetic observations. The example shown here (Figure 4a & b) derives from the Platreef area in RSA and the original dataset is full tensor magnetics. Figure 4b shows the skeletonized dykes, depth adapted, and the thickness of the dyke estimated at each node. The display is in Paraview, where all the rest are in the geophysics tool. Conclusions A progression towards a viable upgrading of geology and geophysics techniques is occurring over a period of several years. This is not being done in isolation, but rather by leveraging upon several key new technologies in related fields. The quest is for continental scale, high resolution modeling in 3D that actually can take count of faults, dykes and sills, as well as cratons, and basins. Automatic technology to capture the deep 3D crustal fault network on a continental scale has evolved to include gravity derived features with estimated dips and extents. This relies on some older work for determining dips from profile data. Similarly, complex 3D dyke networks can now also be realized in a simple and direct manner. This progression can adds value, by further reducing ambiguity, when large scale gravity or magnetic gradiometer surveys are available. References FitzGerald, D., Reid, A. and McInerny, P., 2004, New discrimination techniques for Euler deconvolution: Computers & Geosciences, 30, 461–469. FitzGerald, D., 2006, Innovative Data Processing Methods for Gradient Airborne Geophysical Datasets: The Leading Edge, 25, No. 1, 87-94. FitzGerald, D.,Milligan, P., 2013, Defining a deep fault network for Australia, using 3D “worming”: :ASEG Extended Abstracts, Melbourne. FitzGerald, D. Holstein, H., Foss, C., 2011, Automatic modelling and inversion for dykes from magnetic tensor gradient profiles – recent progress: SEG Extended

Abstracts, Houston. Florio G. and Fedi, M., 2013, Multiridge Euler deconvolution: Geophysical Prospecting, In Press. Gotze, H, Lahmeyer, B., 1988 Applications of three-dimensional interactive modeling in gravity and magnetics, Geophysics, 53, 1096-1108. Holstein, H., FitzGerald, D.J. Anastasiades, C., 2009, Gravimetric anomalies of uniform thin polygonal sheets, SAGA proceedings. Hornby, P., Boschetti F. and Horowitz, F.G., 1999, Analysis of potential field data in the wavelet domain: Geophysical Journal International, 137, 175-196. Lajaunie, C., Courrioux, G. and Manuel, L., 1997, Foliation fields and 3D cartography in Geology: Mathematical Geology 29, 571–584. McGrath, P. H., 1991, Dip and depth extent of density boundaries using horizontal derivatives of upward continued gravity data: Geophysics, v. 56, no. 10 Nakamura, A., Bacchin, M., Milligan, P.R. and Tracey, R., 2011, Isostatic Residual Gravity Map of Onshore Australia (1st Edition), scale 1:5 000 000: Geoscience Australia, Canberra. Naudy, H., 1971. Geophysics, 36, 717-722. Uncover, 2012, www.science.org.au/policy/documents/uncover-report.pdf

Figure 4a above, shows the dyke swarm in the Platreef area of RSA. The original dataset is full tensor magnetics in this case, though TMI also works well. Figure 4b left, shows the skeletonized dykes, depth adapted, and the thickness of the dyke estimated at each node. This display is using Paraview.

Using implicit geological constrains in 3D EM modelling James Macnae ([email protected]) RMIT University Summary The earliest models used for “interpreting” EM data were simple: the half-space, thin sheet, tabular plate and sphere. All had geological analogues: half-spaces were appropriate for thick sediments, thin sheets modeled glacially deposited clay layers in Canada, and the regolith with transported and in-situ weathered material in Australia. Tabular bodies and spheres in free space were approximate modes for volcanogenic massive sulphides, depending on the shape on the target, itself a function of its sub-horizontal deposition, subsequent history of burial, tilting and folding during its plate-tectonic voyage. More recently, there are a number of 1D, stitched 1D, 2D, plate- and sphere-like 3D and blocky 3D models that can be used in AEM interpretation and inversion. Required geological features of EM models (and current state of art): 1D: a) should allow for discontinuities (unconformities)

between layers (good, but not well handled with blanket smoothness constraints). b) Should allow for gradual change, generally a decrease, of conductivity with depth in each layer as compaction increases (not well handled at present except maybe by multiple layers with smoothness)

2D: a) the appropriate features of 1D + dipping lateral discontinuities (blocky steps not good, finite element best) b) facies changes within layers (handled OK with lateral/spatial constraints) c) surveys non-perpendicular to local strike (sometimes requires a fiddle)

3D: a) the model must necessarily be able to approximate the geometry of the target (very poorly handled in blocky 3D models, to the point where small tabular bodies cannot be modeled at all) b) the model must handle current gathering as well as vortex induction (not possible with existing plate models such as Maxwell) b) the target may well be collocated with a 2D inhomogeneity, commonly a metavolcanic-metasediment boundary (not well modeled with a free-space background)

Current software developments in AEM processing using spectral methods (CDI3D) are designed to overcome some of the current limitations in 3D modelling by separating slowly varying (in a spatial sense) 1D/2D effects from 3D target responses, and fitting appropriate vortex / current gathering models to handle discrete targets and lateral discontinuities in the regolith / overburden.

Introduction In interpreting electromagnetic data we need to incorporate many constraints, and our software should make this easy Data constraints Observed values R(x*,y*,z*,t*) + N(T,t*, x*,y*,z*) where * indicates sampled (e.g. on flight lines with specific x*y*z*, at discrete delay times t*, with non-stationary noise N present at time T). Commonly assume that the expectation value �N� = 0, and make quasi-stationary assumption on T. Fails near e.g. powerlines where �Nz� = U(x-x0)/R2, where Nz is the z component of N, U is u=the unbalanced current in the powerline and R2 = (x-x0)2 + (y-y0)2 + (z-z0)2 for the powerline coordinates under the survey line. �N�increases in the afternoons (a turbulence component affecting all AEM systems including ZTEM, and a sferic component affecting all controlled source systems) Physical constraints AEM induction creates large, diffuse current systems. We get “smooth data” (plus noise). Undersampled, noisy AEM data can at best provide a fuzzy picture of the ground, whose resolution decreases with depth due to its diffusive nature. We need to be aware that physical properties affecting electromagnetics are “complex”: the physics allows for spatially varying conductivity permittivity, & magnetic permeability, these may commonly be anisotropic (almost universal in sediments), and frequency or time (IP, SPM) dependent. Fluid distribution (porosity, hydraulic permeability, salinity) has a major effect on EM response Geological constraints in the absence of “ground truth” Generally, AEM is flown because the geology in imperfectly known. All AEM data is collected over geology, and virtually all is affected by geology, with the possible exception of high-altitude references and marine surveys. “Minimum geology” must allow for 1) unconformities (sharp physical property boundaries) 2) facies changes, resulting in smoothly varying physical property changes 3) mass balance such as horizon thickening on folds 4) fluid pathways controlled by fracturing / rock composition and history 5) weathering and overburden transport structures.

Macnae 3D EM

Vertical Anisotropy Electrical conductivity anisotropy is determined by Geology. In untilted sediments for example, horizontal conductivity is larger than vertical conductivity by a significant factor. AEM, inducing horizontal current flows only in a layered earth, is sensitive only to the horizontal conductivity Many near-surface conductors are clay-rich, and clay tends to have anisotropic conductivity, with the ratio of vertical to horizontal resistivity ρv/ρh in the range of 1 to 25, commonly 2 or 3 (Mousseau, 1967, Greve et al, 2012). When resistivity (or conductivity) is measured or calculated using grounded current sources (such as in ERT), the apparent or inverted resistivity is the geometrical mean √(ρvρh) of the vertical and horizontal values (Edwards et al, 1984, Christensen, 2000). With “typical” clay anisotropies of 2 or 3 to 1, the ERT measured resistivity will be 40% to 70% greater than the true horizontal resistivity ρh which is the appropriate value for EM, where induction creates purely horizontal current flow in a layered earth. [Using the reciprocal relationship of resistivity and conductivity, ERT conductivity over clays will typically be 40% to 70% less

than that measured by any inductive EM system, such as VTEM.] Figure 1 shows a “rare” example where over 10 square km of ground was covered with detailed dipole-dipole resistivity and an HEM survey. The shallow conductivity (estimated from the n=1 spacing of a 50 m dipole-dipole survey is imaged in the upper map of the figure, together with contours of the AEM conductivity which clearly map the same features. When the ratio of the airborne/ground conductivity is taken however (imaged in lower section), high values (red and pink) are indicative of strong vertical anisotropy, showing considerable continuity in the EW direction. 2D, 3D and SPM Recent work in the Musgraves (Western Australia and the Northern Territory of Australia) has collected data from multiple AEM systems over several nickel exploration targets. The data also show effects of topography, regolith and basement discontinuity. Stitched 1D (pseudo 2D) sections show these targets at incorrect depths / dips. Blocky 3D inversions with “realistic pixel sizes for numerical stability and execution time” are unable to fit the data over the target. Simple Maxwell plate models can fit the target(s) at late delay times. The CDI3D spectral inversion code however fits both these targets and regolith inhomogeneity using implicit “geological” constraints in terms of its built in models. Estimated backgrounds provide conductivities (directly related to geological composition and local salinity) are used to constrain the current gathering components of the fit.

Three lines of data, presented as a CDI are shown in Figure 2. Note (a) near surface conductivity, (b) the discrete target on the middle line. The central line shown here was inverted using a number of algorithms by Ley-Cooper et al (2012), and theses results are shown in Fig 6.

Figure 1: Using ground dipole-dipole conductivity (colour, top) and airborne HEM data (contours) to predict anisotropy. While near surface conductors in both ground ressitivity survey data and the HEM survey CDI conductivities are N-S or at 45°, the apparent anisotropy is largely E-W ,

Figure 2: CDI of 3 adjacent lines, showing conductor B on the central line. The target has limted strike-length as its response is not evident 100 m either side. This CDI has fit and eliminated SPM. Red is conductivre, blue resistive.

Macnae 3D EM

The various commercial algorithms did not recognise SPM decays, and fitted them with deep conductive layer as seen in Figs 3 and 6. Figs 5 and 6 show the results of automatic separation of AEM data into a “smooth background” and a ”local anomaly”, with the smooth background going through the CDI process to produce a conductivity-depth section, and the local anomaly being fit (by Spectral method of Annan, Macnae 2013) with a small conductor located in the basement under the near-surface zone.

References Edwards, R., Nobes, D., and Gómez‐Treviño, E. (1984). ”Offshore electrical exploration of sedimentary basins: The effects of anisotropy in horizontally isotropic, layered media.” Geophysics, 49(5), 566–576. Greve, A., M.S. Andersen, R.I. Acworth; Monitoring the transition from preferential to matrix flow in cracking clay soil through changes in electrical anisotropy Geoderma Volumes 179–180, June 2012, Pages 46–52. Ley-Cooper, Y., T. Munday, K. Blundell, J. Gum, C. Sorensen, A. Viezzoli, L. Cox, & G. Wilson, 2012, From 1D to full 3D inversion of AEM data for target definition in the Musgrave Province of South Australia: ASEG 3D EM Workshop, Brisbane, Australia. Macnae, J., 2013, 3D-spectral CDIs... a fast alternative to 3D inversion? AEM14 expanded abstracts. Mousseau, R.J., 1967, Measurement of Electrical Anisotropy of Clay‐like Materials, Journal of Applied Physics (Volume: 38, Issue: 11)

AcknowledgementsYusen Ley-Cooper and coauthor’s for permission to use the Musgrave comparison originally presented at the 3D EM workshop at ASEG, 2012. CD3D for use of CDI3D code.

Figure 3: Sample decays for the VTEM data. An SPM source is expected to have a 1/t decay and smallish amplitude, fitted with a black dashed line.

Figure 5: Autimatically stripped, picked and fitted (heavy fines) result of the residual step response (thin lines) after “sensible” background subtraction. The model is a small, shallowly dipping target that fits the data at all but he earliest two profiles.

Figure 4: (a) Later delay time VTEM data (b) stripped residual (step response) and (c) background CDI section (with SPM effect subrtracted) including location/size of fitted tabular target. The white ovel is in fact a circle showing the effect of vertical exaggeration on the section.

Macnae 3D EM

Figure 4: Mixture of 1D, 2D and 3D processing and inversion applied to VTEM data. Each section has “geological” deficiencies. (From Ley-Cooper et al., 2012). At A, the source is near-surface SPM and not a deep conductor, at B the source is a shallow confined conductor.

KEGS Symposium 2014 “Integrating Geophysics and Geology”

KEGS Symposium 2014 “Integrating Geophysics and Geology” Toronto, Canada - March 1st, 2014.

Choices for effectively incorporating geological constraints into geophysical inversion Peter Lelièvre*1, Angela Carter-McAuslan1, Colin Farquharson1 and Nick Williams2 1Department of Earth Sciences, Memorial University, St. John's, Canada 2High Power Exploration Inc., Melbourne, Australia Summary Geophysical inversion provides the means to unite geophysical and geological data sources. As more constraining geological information is incorporated into an inversion workflow, the inherent non-uniqueness of the inverse problem is reduced. This yields a higher potential to resolve features that are less well-constrained by the geophysical data alone, e.g. deeper features. Geological information comes in many forms: it may relate to lithology, mineralogy, structure, alteration, etc.; it may be quantitative, e.g. measured data, or qualitative, e.g. expert insight; it may be point-located or distributed across finite volumes; and it may be relevant to a single or multiple physical properties. In an inversion, this information translates into many different classes of mathematical measures or constraints and can impact parameterization choice, mesh design, regularization approach and optimization strategy. As such, there is no best generalized methodology for incorporating geological constraints into geophysical inversion; each scenario must be assessed and appropriate choices made. We introduce important types of geological information and explain how each can be turned into mathematical measures or constraints for use in geophysical inversion. We mention three very different styles of inversion: discrete body inversion, volumetric physical property inversion and contact surface geometry inversion. Each has certain advantages and disadvantages when it comes to incorporating geological information; practitioners should choose a formulation based on their a priori geological information and exploration questions. Types of geological information Important examples of geological information that may be available are: • Physical property measurements on rock

samples, or lithology observations combined with petrophysical information

• Physical property trends, e.g. density often increases with depth

• Structural orientations • Expected shapes and aspect ratios of bodies, e.g.

a body should be plate-like or pipe-like • Sharpness of contacts between rock units, e.g.

physical properties may change sharply across a fault or smoothly across an alteration zone

• Locations of contacts between rock units • Physical property relationships • Topology, e.g. relative positions of rock units.

An added consideration is that one should take into account the reliability of any type of information included in an inversion, be it geophysical or geological data. Discrete body inversion Early geophysical inverse problems discretized the earth by assuming a homogeneous background with simple shapes for one or more causative target bodies, e.g. an ellipsoid or dipping plate. Simple bodies can be fully represented with very few parameters, e.g. lateral location and depth, major-axis lengths, orientation angles and physical properties. The inverse solution requires finding the parameters that best fit the geophysical data. The computational requirements for the solution are relatively low. Examples of this type of inversion are presented by Oldenburg and Pratt (2007). Clearly, such a specific parameterization makes many assumptions, but if these assumptions hold then it is a powerful way to incorporate geological information. Information regarding physical properties, orientations, aspect ratios and the expected shape of bodies feed directly into the few parameters that control the model, and the models contain sharp boundaries between rock units. Because of the reduced number of parameters, stochastic sampling approaches can be employed to investigate the reliability of recovered parameters. Volumetric physical property inversion As computational methods and computing power increased, geophysical inversion moved to a more general representation of the Earth. Standard practice is now to discretize the volume of interest in to many tightly packed cells. These may be rectangular prisms in a rectilinear mesh or octree mesh, tetrahedra in an unstructured mesh, or any other option (see Figure 1 for some examples). The geometry of the mesh remains constant during the inversion and the physical property in each cell (or in the case of a joint inversion, multiple properties) are the variables of interest. An objective function, ϕ, is defined with the form

(1) ϕ(m) = ϕd(m) + βϕm(m)


KEGS Symposium 2014 “Integrating Geology and Geophysics” Toronto, Canada - March 1st, 2014.

Figure 1: Some mesh choices for volumetric inversion: (a) “A” shape discretized on the following meshes, (b) rectilinear mesh, (c) quadtree mesh (the 2D equivalent of a 3D octree mesh), (d) unstructured triangular mesh (the 2D equivalent of a 3D tetrahedral mesh). The model vector, m, contains the physical property values in each cell in the mesh. The data misfit term, ϕd, measures the fit of the observed geophysical data to that synthesized for the candidate model, and the tradeoff parameter β is used to achieve the desired level of data fit. Because there are typically many more model parameters than geophysical data observations, a regularization term, ϕm, is required to make the problem mathematically tractable. The regularization can be designed to incorporate as much geological information as possible. Li and Oldenburg (1996) used a regularization function of the form

(2) ϕm(m) = |Ws(m-mref)|2 + |Wmm|2

The first “smallness” term measures the difference between the model and an optional reference model, mref. The second “smoothness” term measures the model differences between adjacent cells in Cartesian directions. Ws and Wm are matrix operators that apply weights across the volume. Additional constraints can be added to the objective function, e.g. to specify the acceptable range of values of a physical property in particular cells. The regularization and constraints are the pathways through which geological information is incorporated. A model that minimizes the objective function, subject to any constraints, will fit the geophysical data (to some desired degree) and honour any geological information incorporated.

We now take a closer look at how each item in our list of geological information can be incorporated into a volumetric inversion. Incorporating physical property information Point-located physical property data can be incorporated using the reference model and weights in equation (2) or through bound constraints added to the optimization problem, e.g. as by Li and Oldenburg (2003) to specify that magnetic susceptibility must be positive:

(3) min(m) ϕ(m) = ϕd(m) + βϕm(m) s.t. m ≥ 0

More generally, the bound constraints can specify different lower and upper bounds for each model parameter. Several studies have shown the dramatic improvements that can result from applying even a small number of physical property bound constraints. For example, see Figure 2, from the work of Williams (2008). Specifying physical property trends Specifying that physical properties should increase or decrease in a particular direction, across a specific region of the model, is a more difficult task. One option is to build the desired trend into a reference model and introduce that reference model only into the smoothness term in equation (2), rather than into the smallness term. Another option is to use linear constraints that specify the physical property difference between two cells:

(4) min(m) ϕ(m) = ϕd(m) + βϕm(m) s.t. Am ≤ b

Figure 2: Recovered models for a synthetic gravity inversion based on a geological scenario of nickel exploration in Western Australia’s Yilgarn Craton (adapted from Williams et al., 2009). The inversion in (a) has no geological information incorporated; the inversion in (b) has physical property bounds placed across the surface and down two drill-holes, indicated in black; (c) shows the true model.



Although linear constraints add some computational difficulty, methods have been developed to solve this optimization problem (see Lelièvre and Oldenburg, 2009; Lelièvre et al., 2009). Specifying orientations and aspect ratios Orientation information may specify the strike, dip and plunge of a feature at a specific location (essentially point-located) or across a larger volume of the model. Aspect ratios, or length scales, may be associated with this information; for example the major-axis lengths of a plate-like, rod-like or ellipsoidal feature. Li and Oldenburg (2000) and Lelièvre and Oldenburg (2009) developed a smoothness regularization approach that measures the smoothness in any direction. This allows one to specify general orientations and length scales throughout the model. Figure 3 shows an example of the improvements that can be obtained by including orientation information in an inversion. Recovering models with sharp discontinuities With sum-of-squares measures in equation (2), the recovered models typically exhibit smooth distributions of physical properties. This may not be consistent with the geological scenario. Farquharson and Oldenburg (1998) developed methods for employing general measures that can yield models with a more blocky character, that is, models with sharper discontinuities. Locations of contacts between rock units 3D geological ore deposit models are commonly created during delineation drilling. The location of contacts may be known at points from down-hole intersections and outcrop mapping. The contacts can be interpolated between boreholes and extrapolated outwards. Contacts may also be interpreted from seismic sections. Physical properties may change sharply across these contacts. To incorporate such a scenario into a volumetric inversion, one can specify appropriate orientations for the contacts at the observed or interpreted locations and set smoothness weights or norms appropriately around them. The choice of the mesh used becomes important for this type of geological information. 3D Earth models typically comprise wireframe surfaces that represent the geological contacts between different rock units. Wireframe surfaces (tessellated surfaces comprising connected triangles) are appropriate for representing geological contacts as they are sufficiently general and flexible that they can be

made to represent arbitrarily complicated geological structures and topography. Rectilinear and octree meshes always produce pixellated representations of the Earth and are incompatible with geological models built on wireframes. Lelièvre and Farquharson (2013) are developing inversion methods on unstructured tetrahedral meshes that can conform exactly to these types of geological models and exactly honour the contact surfaces within. Figure 3 shows an example of incorporating a contact surface from a geological model into an inversion. Incorporating physical property relationships When multiple types of geophysical data have been collected, e.g. gravity and EM data, and there are multiple causative physical properties, e.g. density and conductivity, one may attempt to invert the data simultaneously while coupling the physical property models in some way. The different geophysical datasets may sense the Earth in different ways and a simultaneous joint inversion may help to overcome the resolution limitations of any one particular dataset. To jointly invert, the multiple physical property models must be mathematically coupled. There are many options for doing so, discussed in Lelièvre et al. (2012a). These allow for explicit relationships between physical properties, a more implicit correlation or

Figure 3: 3D recovered models for a synthetic gravity gradiometry inversion based on the Voisey’s Bay deposit (adapted from Lelièvre and Farquharson, 2013): (a) unconstrained inversion result; (b) with physical property bounds inside and outside of a wireframe contact surface (grey); (c) with bounds and a preferred elongation direction specified as guided by the result in (b). The true sulphide body is transparent red.



statistical relationship between properties, and structurally-based coupling. There is no best generalized method for coupling the physical property models in a joint inversion: one must use a coupling method that is consistent with the geological information available. Honouring topological rules While the volumetric physical property style of inversion provides many options for incorporating many different types of geological information, topological rules are far more challenging. Topological rules may, for example, specify that rock unit A may contact unit B but not C, or that A must be below C. Without a direct tie between lithologies and physical properties, one can not incorporate topological rules into the inversion process. To help, lithological inversions have been developed (e.g. Bosch and McGaughey, 2001) where each cell in the model can take one of several allowed lithologies. Petrophysical information is used to translate from a lithological model to a physical property model so that geophysical data can be synthesized. Although the inversion seeks lithology rather than a physical property distribution directly, the parameterization of the volume of interest can be identical. Hence, we discuss lithological inversions as related methods. Lithological inversion methods perform a stochastic sampling and, hence, are far more computationally demanding than methods seeking physical property distributions. Contact surface geometry inversion Another approach for achieving sharp interfaces in a model, and including information regarding the locations of lithological contacts, is to build the interfaces directly into the model parameterization. For example, Fullagar et al. (2000) represents the Earth volume of interest as a set of tightly packed vertical rectangular prisms. Internal contacts divide each prism into homogenous layers and the inversion controls the vertical position of the contacts within each layer. Such a parameterization is able to achieve sharp, clearly defined interfaces between the stacked layers but requires that the assumption of stacked layers is appropriate for the geological scenario. Lelièvre et al. (2012b) are developing inversion methods that work directly with 3D wireframe geological models. The parameters are the locations of the vertices of the triangular surface facets in the wireframe model. In their inversions, the geometry of the contact surfaces are free to change but the physical properties of each unit remain fixed.

Because these methods work directly with a wireframe geological model, any geological information that can be incorporated into such a geological model may also be included in the geophysical inversion. Figure 4 shows a proof-of concept example for a 2D scenario. Conclusion We have discussed many of the methods available for incorporating geological information into geophysical inversions. There may be many options for including a particular piece of geological information into an inversion. Each scenario must be assessed and appropriate choices made. Those choices may impact the model parameterization, mesh design, regularization and optimization used. Practitioners should choose a formulation based on their a priori geological information and exploration questions, rather than try to adapt a formulation to fit a specific problem. Constrained inversion carries with it additional work to get the geological information into the correct format to supply to the inversion, but there are many examples in the literature demonstrating the improvements to subsurface models that can be obtained by unifying geological and geophysical data through constrained inversion.

Figure 4: 2D true and recovered models for a synthetic gravity and traveltime tomography joint inversion based on the Voisey’s Bay deposit (adapted from Lelièvre et al., 2013): (a) the true model; (b) a volumetric joint inversion result; (c) a starting wireframe model, black, built from the result in (b); (d) the wireframe model after performing a surface geometry inversion where the red vertices in (c) were allowed to move. Faint grey lines in (d) show the initial model from (c). White lines in all panels indicate the position of contacts in the true model. Small white dots indicate downhole locations for the tomography survey.



References Bosch, M., and J. McGaughey, 2001, Joint inversion of gravity and magnetic data under lithologic constraints: The Leading Edge, 20, 877-881. Farquharson, C. G., and D. W. Oldenburg, 1998, Non-linear inversions using general measures of data misfit and model structure: Geophysical Journal International, 134, 213-227. Fullagar, P. K., N. A. Hughes, and J. Paine, 2000, Drilling constrained 3D gravity inversion: Exploration Geophysics, 31, 17-23. Lelièvre, P. G., and D. W. Oldenburg, 2009, A comprehensive study of including structural orientation information in geophysical inversions: Geophysical Journal International, 178, 623-637. Lelièvre, P. G., D. W. Oldenburg, and N. C. Williams, 2009, Integrating geologic and geophysical data through advanced constrained inversions: Exploration Geophysics, 40, 334-341. Lelièvre, P. G., C. G. Farquharson, and C. A. Hurich 2012a, Joint inversion of seismic traveltimes and gravity data on unstructured grids with application to mineral exploration: Geophysics, 77, K1-K15. Lelièvre, P. G., P. Zheglova, T. Danek, and C. G. Farquharson, 2012b, Geophysical inversion for contact surfaces: SEG Technical Program Expanded Abstracts 2012, doi: 10.1190/segam2012-0716.1. Lelièvre, P. G., and C. G. Farquharson, 2013, Gradient and smoothness regularization operators for geophysical inversion on unstructured meshes: Geophysical Journal International, 195, 330-341. Lelièvre, P. G., C. G. Farquharson, and C. Tycholiz, 2013, Integrating geological constraints into 3D geophysical inversions using unstructured meshes: Society for Geology Applied to Mineral Deposits (SGA) General Assembly, August 2013, Uppsala, Sweden. Li, Y., and D. W. Oldenburg, 1996, 3-D inversion of magnetic data: Geophysics, 61, 394-408. Li, Y., and D. W. Oldenburg, 2000, Incorporating geological dip information into geophysical inversions: Geophysics, 65, 148-157. Li, Y., and D. W. Oldenburg, 2003, Fast inversion of large-scale magnetic data using wavelet transforms and a logarithmic barrier method: Geophysical Journal International, 152, 251-265. Oldenburg, D. W., and D. A. Pratt, 2007, Geophysical inversion for mineral exploration: a decade of progress in theory and practice: Exploration 07: Fifth Decennial International Conference on Mineral Exploration, 61-95.

Williams, N. C., 2008, Geologically-constrained UBC-GIF gravity and magnetic inversions with examples from the Agnew-Wiluna greenstone belt, Western Australia: PhD Thesis, University of British Columbia. Williams, N. C., P. G. Lelièvre, and D. W. Oldenburg, 2009, Constraining gravity and magnetics inversions for mineral exploration using limited geological data: ASEG 20th International Geophysics Conference Expanded Abstracts.

A Framework for the Integration of Geological and Geophysical Data John McGaughey*, Mira Geoscience Ltd, Montreal Glenn Pears, Mira Geoscience Asia Pacific Pty Ltd, Brisbane Peter Fullagar, Fullagar Geophysics Ltd, Vancouver Summary A flexible and practical framework is proposed for the quantitative integration of geological and geophysical data, based on project work carried out over the past several years across a wide range of geological settings and geophysical methods. Our approach to integrated interpretation may be summarized as adherence to the following principles: 1) inversion is a means for refining geological models, not an end in itself; 2) an understanding of the geological meaning of physical rock properties is fundamental; 3) a starting model can be constructed from geophysical data, if necessary; and 4) performing inversion on a geological model is highly advantageous, not least because it permits adjustment of geometric boundaries as well as physical properties. We illustrate this common-sense integration framework through case studies from Australia and Canada. Introduction The role of geophysics is necessarily evolving as modern exploration addresses the challenge of finding significant new deposits at depth, under weathering or cover, or in complex brownfields settings. In this exploration context targeting will rely more on recognition of ore system signatures in multi-disciplinary 3D model space, and less on recognition of anomalies in data. Interpretation in model space, not data space, implies that each data set plays a role in the construction of the earth model upon which targets will be developed. Geophysics will be required increasingly to define the three-dimensional geological context (lithology, alteration, mineralization, and structure) below and between any existing drill holes. The future interpretation of geophysical data will demand tight integration with geological data. The objective of geophysical interpretation is a single model, acceptable to geologists and geophysicists, rather than a “constrained inversion”. The model of integrated interpretation discussed here requires geologists and geophysicists to work closely together on multi-disciplinary earth model interpretations, as opposed to overlaying independently derived geological and geophysical models. The concept of cross-disciplinary asset teams sharing a jointly interpreted “common earth

model” was pioneered more than 25 years ago in the oil and gas industry, in response to the difficult geoscientific challenge posed by dwindling access to conventional exploration targets. We believe that the minerals industry is now at a similar stage, requiring a similar approach to exploration interpretation. The following statement, made in an oil exploration context, is applicable today to mineral exploration:

The advent of 3D earth modelling computer systems suggests there is potential to transform the work processes in cross-disciplinary asset teams. By sharing common digital 3D representations of the subsurface, the team can iterate between disciplines more easily, rapidly incorporating new information into existing models. Up to now, many cross-disciplinary teams have emphasized the importance of software communication, 3D visualization and data access. From now on, we believe that earth modelling issues will assume greater significance in the business of these teams. Garett et al. (1997).

We have been applying quantitative geophysical analysis to the integrated interpretation problem in mineral exploration for many years, across a range of scales, commodities, and geophysical methods. The principles of our approach are summarized in the following sections, and illustrated in case studies from the Mt. Isa Inlier (Chalke et al., 2012) and Sudbury Basin (Perron et al., 2011). Inversion as a Means for Refining Geological Models Exploration normally progresses via revision of a model which, ideally, is a compilation of all relevant information, shared across all disciplines (McGaughey, 2006). Inversion plays an important role in defining model features which are, or are not, compatible with the geophysical data. Therefore inversion can be regarded as a component in an overarching earth modelling effort. Moreover, like geological modelling, inversion can be viewed as a sequence of recursive steps, not necessarily a single stage process. Thus a conventional “unconstrained” inversion might be completed as a first step. Geological models are comprised of surfaces, mainly litho-stratigraphic contacts and structures, which divide the ground into rock type domains. Geological models are categorical, insofar as each sub-surface domain is assigned

A Framework for the Integration of Geological and Geophysical Data

to a rock type. In petrophysical (or property) models the sub-surface is divided into cells, to which are assigned one or more physical properties. Cell boundaries are often artificial, i.e. bear no relation to geological contacts and structures. Models for geophysical inversion must be petrophysical, and can also be geological. Inverting on geological models delivers flexibility and control not available with pure property models (Fullagar and Pears, 2007). In particular, geological models permit geometry inversion, to alter the shape of boundaries, as well as property inversion. If geophysics alone is able to directly detect and unambiguously delineate mineralization, “unconstrained” inversion of geophysical data in isolation may be effective. However, unconstrained inversion is often of questionable value, especially if applied to potential field data from complex and poorly understood geological settings; non-uniqueness of inversion results is severe in such cases. Interpreting Geological Meaning of Physical Properties Our approach to integrated interpretation is to define the geological objective, tied to an exploration objective, whenever possible. Reconciling the geophysical data with an exploration-driven geological hypothesis both tests and refines the hypothesis. Understanding how geophysical signatures relate to geology requires a project-specific assessment of the rock properties. The relationships are often complicated in the proximal alteration zones of ore systems, where alteration may overprint the structural and formational boundaries. In practice a reliable, comprehensive physical property database is rarely available, with the result that physical properties are deduced via joint interpretation of the geological and geophysical data. In an integrated interpretation framework, the geological model, whether conceptual or empirical, must be attributed with appropriate physical properties to enable reconciliation with geophysical data. In the absence of hard petrophysical data, starting values must be assigned on the basis of published compilations, subjective judgement, and/or initial unconstrained inversions. Starting Model Construction from Geophysical Data The project conceptual model in combination with geological and geophysical data usually gives us what we require to build the initial three-dimensional geological model that will form the foundation of our exploration decisions. Where does the model come from?

Prior to the advent of inversion, geological interpretation of geophysical data was almost universally cast in terms of the geometry of structural and formational boundaries (Figure 1).This is consistent with the reasonable interpretational starting assumption that geological boundaries are the primary control on geophysical data. Yet typically when the same data are inverted, a smoothly-varying physical property continuum is illogically presented in place of the geological domain boundaries. In our view, in the absence of hard data to the contrary, the initial model should comprise homogeneous volumes truncated by structural and formational boundaries. Any physical property discontinuities (the primary controls on the geophysical data) will occur at those geological boundaries. For a mature project, a well-defined model usually exists or can be constructed from geological logs and maps. A process of geophysical forward modelling and inversion can rapidly validate or revise the model. This process usually reveals new information about the relationships between geology, geophysics and rock properties, as well as important interpretational insight into the exploration problem at hand. In contrast, the more common practice of visually overlaying three-dimensional geological models on independently derived unconstrained geophysical inversions is of questionable value because the connection between geological model and inverted rock property distribution is so tenuous.

For a less mature project, a geologically-based model does not necessarily exist. We have often heard that three-dimensional geological modelling tools have limited value

Figure 1: Example of conventional interpretation of magnetic data, in which the primary control on the form of the data is attributed to structural and geological domain boundaries (Siddorn, 2010).


in areas without drilling or other subsurface data, leaving no alternative to unconstrained inversion. This is mistaken: a three-dimensional geological model of significant exploration value can usually be developed from surface geological and geophysical data, and we have often accomplished this through quantitative interpretation of potential field data alone. In such cases, a geologically-based geophysical interpretation will comprise not just one inversion, but many to understand the relationship between geology, geophysics and rock properties. The process begins with modelling three-dimensional boundaries of key domains, be they lithological or petrophysical. This is a highly interpretative step, in effect establishing relationships between geophysics and geology. In the absence of drillhole information to constrain the geological model at depth, shallow structure is projected down-plunge. This process can draw upon mapping and magnetic data interpretation, unconstrained inversion results, rapid forward modelling of geologically plausible structure, or other existing interpretations such as interpreted geological sections, seismic, or magnetic depth-to-source solutions. Implicit geological modelling algorithms enable rapid construction of plausible models from sparse data. A best estimate of rock properties is then assigned to each three-dimensional geological domain. It is easy to over-parameterize the earth and fit geophysical data. It is a far more useful process to identify key geological domains with characteristic physical properties and explain the geophysical response based on the configuration of those interpreted domains. Inversion Applied to Geological Models Once the initial three-dimensional geological model is constructed, we apply inversion to it, adjusting geometric boundaries and physical properties as appropriate until quantitative reconciliation of the geophysical data is achieved. Geological models permit geometry inversion as well as physical property inversion (Fullagar et al., 2008). Thus the interpretational decision can be made as to the degree to which geophysical data is responding to physical property discontinuities at geological domain boundaries versus heterogeneity within or across boundaries. Figure 2 illustrates how models may be parameterized for geometry inversion. Software capable of geometry inversion can also perform physical property inversion, either simultaneously, for example Bosch et al. (2006), or sequentially, e.g. Fullagar et al. (2004). For physical property inversion on a geological model, the rock type attribute provides added

flexibility and control. For example, the properties of entire homogeneous units can be optimized by inverting their respective properties. Homogeneous unit inversion is fast, even for large models, because only a handful of parameters are involved. Therefore, crude initial estimates can be rapidly reconciled with the data. The advantages of inverting on a geological model are summarized below: • a natural driver for integrated interpretation; • simplifies interpretation of physical property models,

respecting geological domains; • fast optimization of homogeneous properties; • greater control, e.g. restriction of changes to particular

geological units; • incorporation of magnetic remanence by formation; • assignment of statistical distributions by rock type; • geometry inversion as well as property inversion; • restriction of changes to particular geological units. Figure 3 illustrates a simple example of an integrated interpretation in which gravity data was inverted to both define the geometry of individual geological units and density heterogeneity within the units.

Figure 2: Parameterization of geological models for geophysical inversion of either geometrical domain boundaries or internal physical property heterogeneity. The illustration is for the programs VPmg and VPem1D (Fullagar et al., 2013).


Conclusions Inversion can be regarded as one component of an earth modelling system, to advance interpretation of the sub-surface. The immediate aim of inversion is to improve a starting model in order to achieve a satisfactory fit to geophysical data. Owing to the limitations imposed by physics, logistics, and experimental error, there are usually an infinite number of models which are acceptable in terms of data fit. Of these, the models which conflict with what is already known about the geology and petrophysics of the area must be rejected. Hence the need for geological and petrophysical constraints. Most 3D geophysical inversion models are purely petrophysical, i.e. comprised of cells attributed with physical properties only. Geological models, on the other hand, are comprised of surfaces which enclose rock type domains. If inversion is performed on a geological model rather than a pure property model, a richer set of options is available, both for inversion style and for constraints. In particular, geometry inversion can be applied to modify geological boundaries, or the variability of a property within a domain can be controlled to conform to a prescribed probability distribution. Moreover, petrophysical constraints which are specific to a rock type, e.g. remanent magnetization parameters, cannot be easily applied to pure property models. The exploration challenge of new discovery at depth, below cover or weathering, or in complex brownfields settings demands the integration of geology and geophysics in a firmly quantitative framework. Approaches are required that directly take into account the exploration objective and conceptual geological model, utilizing tools that permit construction of an earth model consistent with both geological and geophysical data. In practice this means that integration of geology and geophysics is maximized when we combine advanced geological modelling tools, physical property analysis, and geophysical inversion tools that can modify geological boundaries as well as physical property distributions. References Bosch, M., Meza, R., Jimenez, R., and Honig, A., 2006, Joint gravity and magnetic inversion in 3D using Monte Carlo methods: Geophysics, 71, G153-G156. Chalke, T., McGaughey, J., and Perron, G., 2012, 3D software technology for structural interpretation and modelling: Structural Geology and Resources 2012, Kalgoorlie, Western Australia, 16-20.

Fullagar, P.K., Pears, G.A., Hutton, D., and Thompson, A., 2004, 3D Gravity and Aeromagnetic Inversion for MVT Lead-Zinc Exploration at Pillara, Western Australia: Exploration Geophysics, 35, 142-146. Fullagar, P.K., and Pears, G.A., 2007, Towards geologically realistic inversion: Proceedings of Exploration ’07, Fifth Decennial International Conference on Mineral Exploration, Toronto. Fullagar, P.K., Pears, G.A., and McMonnies, B., 2008, Constrained inversion of geological surfaces - pushing the boundaries: The Leading Edge, 27, 98-105. Fullagar, P.K., Pears, G.A., and Reid, J.E., 2013, Hybrid 1D/3D geologically constrained inversion of airborne TEM data, ASEG 23rd International Geophysical Conference and Exhibition, 11-14 August 2013, Melbourne, Australia. Garett, S., Griesbach, S., Johnson, D., Jones, D., Lo, M., Orr, W., and Sword, C., 1997, Earth model synthesis: First Break, 15, 13-20. McGaughey, J., 2006, The Common Earth Model: A Revolution in Mineral Exploration Data Integration: in GIS Applications in the Earth Sciences, Geological Association of Canada Special Publication #44, J.R. Harris (ed.), 567-576. Perron, G., Fullagar, P., Pears, G., Phillips, N., Williston, C., Gerrie, V., and Everest, J., 2011, 3D Litho-Prediction Model from In-Situ Physical Rock Property Logging and Constrained Potential Fields Data Inversion: Presented at the 31st GOCAD Meeting, Nancy, France, April 19. Siddorn, J.P., 2010, The geological interpretation of aeromagnetic data: A geologist’s perspective: KEGS Symposium, March 6, 2010, Toronto.

Figure 3: Integrated interpretation of geological controls (drill hole pierce points) and gravity data to establish both the geometry of inidivual geological units and density heterogeneity within them (Fullagar et al., 2004).

KEGS Symposium 2014 Feedback Sheet

(Please print clearly) Name / Affiliation / Contact Info (optional):

______________________________________________________________________________ Are you a KEGS member? Yes No If No, would you like to receive a member registration form? Yes No (Registration includes annual Breakfast). The monthly KEGS bulletin is available free of charge upon request (membership not required). To

receive the Bulletin, please provide your email address. __________________________________ Please rate the following from 1 (strongly dislike) to 5 (strongly like) and comment: Symposium Topic 1 2 3 4 5 ___________________________

Symposium Venue 1 2 3 4 5 ___________________________

Symposium Time / Date 1 2 3 4 5 ___________________________

Symposium Breakfast 1 2 3 4 5 ___________________________

Symposium Lunch 1 2 3 4 5 ___________________________

Symposium Reception 1 2 3 4 5 ___________________________

Symposium Booklet 1 2 3 4 5 ___________________________

Symposium Online Archives 1 2 3 4 5 ___________________________ KEGS is soliciting topics for the PDAC 2015 Symposium. Please provide a suggestion(s) below: ______________________________________________________________________________

______________________________________________________________________________ Comments (Please comment on today’s Symposium or any other KEGS-related topic.) ______________________________________________________________________________

______________________________________________________________________________

______________________________________________________________________________

______________________________________________________________________________

______________________________________________________________________________

______________________________________________________________________________

______________________________________________________________________________

______________________________________________________________________________

______________________________________________________________________________

kegs pdac symposium 2014 booklet preliminary · geophysical results and the geological knowledge...

Documents