runoff prediction in ungauged basins.pdf
TRANSCRIPT
-
Runoff Prediction in Ungauged BasinsSynthesis across Processes, Places and Scales
Predicting water runoff in the mostly ungauged water catchmentareas of the world is vital to practical applications such as thedesign of drainage infrastructure and flooding defences, for run-off forecasting and for catchment management tasks such aswater allocation and climate impact analysis.This important new book synthesises decades of rigorous
analytical research from around the world, forming a holisticapproach to catchment hydrology, and providing a one-stopresource for hydrologists in both developed and developing coun-tries. It brings together results from individual location-basedstudies with comparative analysis along gradients of climate andlandscape features. Topics include data for runoff regionalisationand the prediction of runoff hydrographs, flow duration curves,f low paths and residence times, annual and seasonal runoff, andf loods.Illustrated with many case studies, and including a f inal chap-
ter on recommendations for researchers and practitioners, thisbook is written by expert international authors involved in theprestigious International Association of Hydrological Sciences(IAHS) Predictions in Ungauged Basins (PUB) initiative. It is akey resource for academic researchers in the fields of hydrology,hydrogeology, ecology, geography, soil science, and environ-mental and civil engineering, and professionals working withwater runoff in ungauged water basins.
Gunter Bloschl is Professor of Hydrology, Director of theCentre for Water Resource Systems, and Head of the Institute ofHydraulic Engineering and Water Resources Management at theVienna University of Technology. He has published extensivelyon subjects related to hydrology and water resources and servedas an editor and associate editor for ten of the best scientificjournals in the field. Professor Blschl has been elected Fellowof the American Geophysical Union and the German Academy ofScience and Engineering, has chaired the International Associ-ation of Hydrological Sciences (IAHS) Predictions in UngaugedBasins (PUB) initiative, and has been elected President of theEuropean Geosciences Union. Recently he has been awarded theprestigious Advanced Grant of the European Research Council(ERC).
Murugesu Sivapalan is Professor of Civil and Environmen-tal Engineering, and of Geography and Geographic InformationScience at the University of Illinois. He was founding chair of theInternational Association of Hydrological Sciences (IAHS) Pre-dictions in Ungauged Basins (PUB) initiative. He has publishedextensively on catchment hydrology in several international jour-nals and is Executive Editor of the European GeosciencesUnions Hydrology and Earth System Sciences journal. ProfessorSivapalan has also received the European Geophysical SocietysJohn Dalton Medal, the International Hydrology Prize of theIAHS, and the Hydrological Sciences Award and the RobertE. Horton Medal of the American Geophysical Union. He wasalso the recipient of the Centenary Medal of the AustralianGovernment and an Honorary Doctorate of the Delft Universityof Technology.
Thorsten Wagener is Professor of Water and EnvironmentalSecurity in the Department of Civil Engineering at the Universityof Bristol. He is a Vice President of the International Associationof Hydrological Sciences, editor of the journal Hydrology andEarth System Sciences, and associate editor of several otherjournals. Dr Wagener has been awarded DAAD (Deutscher Aka-demischer Austauschdienst) fellowships, an IEMSS Early CareerExcellence Prize, Best Paper Awards of the Journal of Environ-mental Modeling and Software, a US EPA Early Career Award,the Walter Huber Civil Engineering Research Prize of the Ameri-can Society of Civil Engineers, an Alexander von HumboldtFoundation Fellowship, and an Education and Public ServiceAward of the Universities Council for Water Resources.
Alberto Viglione is a Research Hydrologist at the ViennaUniversity of Technology. During 20047 he conducted doctoralresearch on Non-supervised statistical methods for the prediction
-
of hydrological variables in ungauged sites at the HydraulicDepartment of the Politecnico of Turin. He has authored orco-authored numerous papers in hydrology, particularly onfloods, from both statistical and process-based perspectives, andon hydrological characterisation of river basins. Dr Viglione hasdeveloped software for regional frequency analysis and rainfallrunoff modelling under the R environment, which is availableonline. He also acts as a reviewer for several prestigious journals,and has been involved in a number of research projects related tohydrology and flood frequency analysis in Italy, Austria andseveral other European countries.
Hubert Savenije is Professor of Hydrology and Head of theWater Resources Section at the Delft University of Technologyand also serves as Editor-in-Chief of Hydrology and Earth SystemSciences and Physics and Chemistry of the Earth. He is President-Elect of the International Association of Hydrological Sciences(IAHS) as well as Chair of the IAHS Predictions in UngaugedBasins (PUB) initiative. Professor Savenije has published widelyin several leading international journals and was Vice Rector atUNESCO-IHE, Institute for Water Education. He is Past-President of Hydrological Sciences of the European GeosciencesUnion (EGU), and Past-President of the International Commis-sion on Water Resources Systems of the IAHS. He has beenawarded, among other things, the Henry Darcy Medal of theEGU and the EGU Batch Award.
-
Runoff Predictionin UngaugedBasinsSynthesis acrossProcesses, Placesand Scales
edited by
GNTER BLSCHLTechnische Universitt Wien, Austria
MURUGESU SIVAPALANUniversity of Illinois, Urbana-Champaign, USA
THORSTEN WAGENERUniversity of Bristol, UK
ALBERTO VIGLIONETechnische Universitt Wien, Austria
HUBERT SAVENIJETechnische Universiteit Delft, the Netherlands
-
cambridge univers ity press
Cambridge, New York, Melbourne, Madrid, Cape Town,Singapore, So Paulo, Delhi, Mexico City
Cambridge University PressThe Edinburgh Building, Cambridge CB2 8RU, UK
Published in the United States of America by Cambridge University Press, New York
www.cambridge.orgInformation on this title: www.cambridge.org/9781107028180
Cambridge University Press 2013
This publication is in copyright. Subject to statutory exceptionand to the provisions of relevant collective licensing agreements,no reproduction of any part may take place withoutthe written permission of Cambridge University Press.
First published 2013
Printed and bound in the United Kingdom by the MPG Books Group
A catalogue record for this publication is available from the British Library
Library of Congress Cataloguing in Publication Data
Runoff prediction in ungauged basins : synthesis across processes, places and scales / edited by Gnter Blschl, TechnischeUniversitt Wien, Austria, Murugesu Sivapalan, University of Illinois, Urbana-Champaign, USA, Thorsten Wagener, University ofBristol, UK, Alberto Viglione, Technische Universitt Wien, Austria, Hubert Savenije, Technische Universiteit Delft, The Netherlands.
pages cmISBN 978-1-107-02818-0 (Hardback)1. Runoff. 2. Rain and rainfall. 3. RunoffMathematical models. 4. Rain and rainfallMathematical models. 5. Hydrology.
I. Blschl, Gnter, 1961 editor of compilation.GB980.R87 2013551.4808dc23 2012036513
ISBN 978-1-107-02818-0 Hardback
Cambridge University Press has no responsibility for the persistence oraccuracy of URLs for external or third-party internet websites referred toin this publication, and does not guarantee that any content on suchwebsites is, or will remain, accurate or appropriate.
-
Contents
List of contributors page ixForeword by Thomas Dunne xvPreface xixAbstract xxii
1 Introduction 11.1 Why we need runoff predictions 11.2 Runoff predictions in ungauged basins
are difficult 31.3 Fragmentation in hydrology 41.4 The Prediction in Ungauged Basins initiative: a
response to the challenge of fragmentation 51.5 What this book aims to achieve: synthesis across
processes, places and scales 61.5.1 Synthesis across processes 71.5.2 Synthesis across places 81.5.3 Synthesis across scales 8
1.6 How to read the book and what to get out of it 9
2 A synthesis framework for runoffprediction in ungauged basins 112.1 Catchments are complex systems 11
2.1.1 Co-evolution of catchmentcharacteristics 11
2.1.2 Signatures: a manifestation ofco-evolution 13
2.2 Comparative hydrology and the Darwinianapproach 152.2.1 Generalisation through comparative
hydrology 152.2.2 Hydrological similarity 182.2.3 Catchment grouping: exploiting the
similarity concept for PUB 202.3 From comparative hydrology to predictions in
ungauged basins 222.3.1 Statistical methods of predictions in
ungauged basins 222.3.2 Process-based methods of predictions
in ungauged basins 232.4 Assessment of predictions in ungauged basins 23
2.4.1 Comparative assessment as a means ofsynthesis 23
2.4.2 Performance measures 252.4.3 Level 1 and Level 2 assessments 26
2.5 Summary of key points 26
3 A data acquisition framework for runoffprediction in ungauged basins 293.1 Why do we need data? 293.2 A hierarchy of data acquisition 30
3.2.1 Assessment based on global data sets 313.2.2 Assessment based on national
hydrological network and national surveys 313.2.3 Assessment based on local field visits
including reading the landscape 323.2.4 Assessment based on dedicated
measurements 343.3 Runoff data 34
3.3.1 What runoff data are needed for PUB? 343.3.2 What runoff data are there? 353.3.3 How valuable are runoff data for PUB? 36
3.4 Meteorological data and water balancecomponents 363.4.1 What meteorological data and water
balance components are needed forPUB? 36
3.4.2 Precipitation 363.4.3 Snow cover data 393.4.4 Potential evaporation 393.4.5 Remotely sensed data for calculating
actual evaporation 403.4.6 Remote sensing of soil moisture and
basin storage 403.5 Catchment characterisation 41
3.5.1 Topography 413.5.2 Land cover and land use 413.5.3 Soils and geology 42
3.6 Data on anthropogenic effects 433.7 Illustrative examples of hierarchical data
acquisition 443.7.1 Understanding process controls on runoff
(Tenderfoot Creek, Montana, USA) 443.7.2 Runoff predictions using rainfallrunoff
models (Chicken Creek, Germany) 473.7.3 Forensic analysis of magnitude and
causes of a flood (Selka Sora, Slovenia) 493.8 Summary of key points 51
4 Process realism: flow paths and storage 534.1 Predictions: right for the right reasons 534.2 Process controls on flow paths and storage 55
-
4.3 Inference of flow paths and storage fromresponse characteristics 574.3.1 Inference from runoff 574.3.2 Inference from tracers 59
4.4 Estimating flow paths and storage in ungaugedbasins 644.4.1 Distributed process-based models 644.4.2 Index methods 644.4.3 Methods based on proxy data 65
4.5 Informing predictions of runoff in ungaugedbasins 664.5.1 Process-based (rainfallrunoff) methods 674.5.2 Statistical methods 674.5.3 Role of field visits, reading the landscape,
photos and other proxy data 684.5.4 Regional interpretation and similarity 68
4.6 Summary of key points 69
5 Prediction of annual runoff in ungaugedbasins 705.1 How much water do we have? 705.2 Annual runoff: processes and similarity 71
5.2.1 Processes 725.2.2 Similarity measures 785.2.3 Catchment grouping 79
5.3 Statistical methods of predicting annual runoffin ungauged basins 835.3.1 Regression methods 835.3.2 Index methods 845.3.3 Geostatistics and proximity methods 885.3.4 Estimation from short records 88
5.4 Process-based methods of predicting annualrunoff in ungauged basins 895.4.1 Derived distribution methods 895.4.2 Continuous models 905.4.3 Proxy data on annual runoff processes 91
5.5 Comparative assessment 925.5.1 Level 1 assessment 925.5.2 Level 2 assessment 96
5.6 Summary of key points 100
6 Prediction of seasonal runoff in ungaugedbasins 1026.1 When do we have water? 1026.2 Seasonal runoff: processes and similarity 104
6.2.1 Processes 1046.2.2 Similarity measures 1116.2.3 Catchment grouping 114
6.3 Statistical methods of predicting seasonalrunoff in ungauged basins 1186.3.1 Regression methods 1186.3.2 Index methods 1186.3.3 Geostatistical and proximity methods 119
6.3.4 Runoff estimation from short records 1216.4 Process-based methods of predicting seasonal
runoff in ungauged basins 1236.4.1 Derived distribution methods 1236.4.2 Continuous models 124
6.5 Comparative assessment 1266.5.1 Level 1 assessment 1276.5.2 Level 2 assessment 129
6.6 Summary of key points 134
7 Prediction of flow duration curves inungauged basins 1357.1 For how long do we have water? 1357.2 Flow duration curves: processes and similarity 137
7.2.1 Processes 1387.2.2 Similarity measures 1417.2.3 Catchment grouping 145
7.3 Statistical methods of predicting flow durationcurves in ungauged basins 1477.3.1 Regression methods 1487.3.2 Index flow methods 1487.3.3 Geostatistical methods 1517.3.4 Estimation from short records 152
7.4 Process-based methods of predicting flowduration curves in ungauged basins 1537.4.1 Derived distribution methods 1537.4.2 Continuous models 154
7.5 Comparative assessment 1567.5.1 Level 1 assessment 1567.5.2 Level 2 assessment 158
7.6 Summary of key points 162
8 Prediction of low flows in ungauged basins 1638.1 How dry will it be? 1638.2 Low flows: processes and similarity 164
8.2.1 Processes 1648.2.2 Similarity measures 1678.2.3 Catchment grouping 170
8.3 Statistical methods of predicting low flows inungauged basins 1728.3.1 Regression methods 1728.3.2 Index low flow methods 1758.3.3 Geostatistical methods 1768.3.4 Estimation from short records 178
8.4 Process-based methods of predicting lowflows in ungauged basins 1798.4.1 Derived distribution methods 1798.4.2 Continuous models 1808.4.3 Proxy data on low flow processes 180
8.5 Comparative assessment 1818.5.1 Level 1 assessment 1828.5.2 Level 2 assessment 184
8.6 Summary of key points 188
vi Contents
-
9 Prediction of floods in ungauged basins 1899.1 How high will the flood be? 1899.2 Floods: processes and similarity 190
9.2.1 Processes 1919.2.2 Similarity measures 1969.2.3 Catchment grouping 200
9.3 Statistical methods of predicting floods inungauged basins 2039.3.1 Regression methods 2039.3.2 Index flood methods 2059.3.3 Geostatistical methods 2089.3.4 Estimation from short records 209
9.4 Process-based methods of predicting floodsin ungauged basins 2119.4.1 Derived distribution methods 2129.4.2 Continuous models 2159.4.3 Proxy data on flood processes 217
9.5 Comparative assessment 2199.5.1 Level 1 assessment 2209.5.2 Level 2 assessment 222
9.6 Summary of key points 225
10 Prediction of runoff hydrographs inungauged basins 22710.1 What are the dynamics of runoff? 22710.2 Runoff dynamics: processes and similarity 228
10.2.1 Processes 22910.2.2 Similarity measures 23310.2.3 Catchment grouping 236
10.3 Statistical methods of predicting runoffhydrographs in ungauged basins 23810.3.1 Regression methods 23810.3.2 Index methods 23810.3.3 Geostatistical methods 239
10.4 Process-based methods of predicting runoffhydrographs in ungauged basins 24010.4.1 Structure of rainfallrunoff models
for ungauged basins 24110.4.2 Parameters of rainfallrunoff
models in ungauged basins:overview 246
10.4.3 A-priori estimation of modelparameters 247
10.4.4 Transfer of calibrated modelparameters from gaugedcatchments 251
10.4.5 Constraining model parametersby dynamic proxy data andrunoff 256
10.5 Comparative assessment 26210.5.1 Level 1 assessment 26310.5.2 Level 2 assessment 266
10.6 Summary of key points 268
11 PUB in practice: case studies 27011.1 Predictions in Ungauged Basins in a societal
context 27011.2 Hydrological insights from long-term runoff
patterns across Krishna Basin, India 27211.3 Predicting mean annual runoff across
Huangshui Basin, China 27711.4 An index approach to mapping annual
runoff in a Siberian catchment, Russia 28011.5 Predicting spatial patterns of inter-annual
runoff variability in the Canadian Prairies 28311.6 Seasonal flow prediction with uncertainty
in South Africa and Lesotho 28911.7 Setting environmental flow targets in
north-east USA 29311.8 Continuous simulation of low flows for
hydropower development in Ontario, Canada 29711.9 Estimating flow duration curves for
hydropower development in central Italy 30011.10 Implementing the EU flood directive in
Austria 30511.11 Revision of Australian Rainfall and Runoff
for improved flood predictions 30911.12 Understanding flow paths for hydrograph
prediction in an Andean catchment, Chile 31311.13 Frequency of runoff occurrence in ephemeral
catchments in France 31711.14 Overcoming data limitations for hydrograph
prediction, Luangwa Basin, Zambia 32111.15 Remotely sensed lake levels to assist runoff
modelling in Ghana 32811.16 Model enhancements for urban runoff
predictions in the south-west USA 33211.17 Runoff predictions to help meet Millennium
Development Goals in Zimbabwe 33711.18 Runoff predictions in support of the National
Water Audit, Australia 34511.19 Distributed runoff predictions in the Mekong
River basin 34911.20 Implementing the EU Water Framework
Directive in Sweden 35311.21 Summary of key points 360
12 Outcomes of synthesis 36112.1 Learning from synthesis 36112.2 Synthesis across processes, places and scales 363
12.2.1 Synthesis across processes 36312.2.2 Synthesis across places 36712.2.3 Synthesis across scales 36912.2.4 Inter-comparison of methods 371
12.3 Synthesis of Newtonian and Darwinianframeworks 37412.3.1 Evidence for co-evolution 374
Contents vii
-
12.3.2 Comparative hydrology and theNewtonianDarwinian synthesis 376
12.3.3 A new unified uncertainty frameworkfor PUB 379
12.4 Synthesis and the science community 38112.4.1 Accumulation of knowledge in the
hydrological sciences 38112.4.2 Role of the community 382
13 Recommendations 38413.1 Advancing runoff predictions in ungauged
basins 38413.1.1 Understanding as the key to better
predictions 38413.1.2 Exploiting runoff signatures and
linking them 38413.1.3 Addressing uncertainty from a process
perspective 38413.1.4 Data availability and predictions 385
13.2 Advancing hydrological science globallyvia PUB 385
13.2.1 Viewing catchments as complexsystems 385
13.2.2 Comparative hydrology to detectco-evolution patterns 385
13.2.3 NewtonianDarwinian synthesis 38513.2.4 The globe is our laboratory 385
13.3 Organising the hydrology community toadvance science and predictions 38513.3.1 Capacity building 38513.3.2 Collaborative endeavour 38613.3.3 Knowledge accumulation 38613.3.4 Hydrology, a global science 386
13.4 Best practice recommendations forpredicting runoff in ungauged basins 386
Appendix: Summary of studies used in thecomparative assessments 388
References 415Index 463
viii Contents
-
Contributors
Ghazi Al-RawasSultan Qaboos University, Department of Civil & ArchitecturalEngineering, College of Engineering, PO Box 33, Al-Khodh,P.C. 123, Muscat, Sultanate of Oman
Vazken AndrassianIrstea, UR Hydrosystmes et Bioprocds, 1 rue Pierre-Gilles deGennes CS 100 30, 92761 Antony Cedex, France
Tianqi AoSichuan University, Department of Hydrology and WaterResources, College of Water Resources and Hydropower, No. 24,Yihuanlu Nanyiduan, Chengdu, Sichuan 610065, China
Stacey A. ArchfieldUS Geological Survey, 10 Bearfoot Road, Northborough, MA01532, USA
Berit ArheimerSwedish Meteorological and Hydrological Institute,Folkborgsvgen 1, 601 76 Norrkping, Sweden
Andrs BrdossyUniversity of Stuttgart, Institute of Hydraulic Engineering,Pfaffenwaldring 61, 70569 Stuttgart, Germany
Trent BiggsSan Diego State University, Department of Geography, 5500Campanile Drive, San Diego, CA 921824493, USA
Gnter BlschlVienna University of Technology, Institute of HydraulicEngineering and Water Resources Management, Karlsplatz13/2222, 1040 Vienna, Austria
Theresa BlumeHelmholtz Centre Potsdam, GFZ German Research Centre forGeosciences, Section 5.4 Hydrology, Telegrafenberg 14473Potsdam, Germany
Marco BorgaUniversity of Padova, Department of Land and AgroforestEnvironments, via dellUniversit 16, 35020, Legnaro (PD), Italy
Helge BormannUniversity of Siegen, Department of Civil Engineering,Paul-Bonatz-Str. 911, 57068 Siegen, Germany
Gianluca BotterUniversit di Padova, Dipartimento IMAGE, Via Loredan 20,35131 Padova, Italy
Tom BrownUniversity of Saskatchewan, Centre for Hydrology, Kirk Hall,117 Science Place, Saskatoon, SK, S7N 5C8, Canada
Donald H. BurnUniversity of Waterloo, Department of Civil and EnvironmentalEngineering, 200 University Avenue West, Waterloo, Ontario,N2L 3G1, Canada
Sean K. CareyMcMaster University, School of Geography & Earth Sciences,General Science Building, Rm 238, 1280 Main Street West,Hamilton, Ontario, L8S 4L8, Canada
Attilio CastellarinUniversity of Bologna, Department DICAM,Viale Risorgimento2, 40136 Bologna, Italy
Francis ChiewCSIRO Land and Water Black Mountain, Christian Laboratory,Clunies Ross Street, GPO Box 1666, ACT 2601, Australia
Franois ColinMontpellier SupAgro, UMR LISAH, 2 Place Pierre Viala, 34060Montpellier Cedex 2, France
Paulin CoulibalyMcMaster University, School of Geography and Earth Sciences,Office GSB 235, 1280 Main Street West, Hamilton, Ontario, L8S4L7, Canada
Armand CrabitINRA, UMR LISAH, 2 Place Pierre Viala, 34060 MontpellierCedex 2, France
Barry CrokeThe Australian National University, Integrated CatchmentAssessment and Management Centre (iCAM) and NationalCenter for Groundwater Research and Training, The FennerSchool of Environment and Society, iCAM, Bldg 48a, LinnaeusWay, Canberra ACT 0200, Australia
Siegfried DemuthUNESCO, Section on Hydrological Systems and Global Change,Division of Water Sciences, Natural Sciences Sector, 1 rueMiollis, 75 732 Paris Cedex 15, France
Qingyun DuanBeijing Normal University, GCESS, 19 Xinjiekouwai, Beijing100875, China
Giuliano Di BaldassarreUNESCO-IHE, Institute for Water Education, Westvest 7, 2601DA Delft, the Netherlands
Thomas DunneUniversity of California-Santa Barbara, Bren School ofEnvironmental Science & Management, Bren Hall 3510, SantaBarbara, CA 931065131, USA
-
Ying FanRutgers University, Department of Earth and Planetary Sciences,Wright Laboratories, 610 Taylor Road, Piscataway,NJ 088548066, USA
Xing FangUniversity of Saskatchewan, Centre for Hydrology, Kirk Hall,117 Science Place, Saskatoon, SK, S7N 5C8, Canada
Boris GartsmanLaboratory for Land Hydrology and Climatology, PacificGeographical Institute FEB RAS, Radio Street 7, Vladivostok690041, Russia
Alexander GelfanRussian Academy of Sciences, Watershed Hydrology Laboratory,Water Problems Institute, 3 Gubkina Str., 119333 Moscow,Russia
Mikhail GeorgievskiState Hydrological Institute, Remote Sensing Methods and GISLab, 23 second line, St Petersburg, 199053, Russia
Nick van de GiesenDelft University of Technology, Water Resources Section,Stevinweg 1, 2628 CN Delft, the Netherlands
David C. GoodrichUSDA-ARS, Southwest Watershed Research Center, 2000E Allen Rd, Tucson, AZ 857191596, USA
Hoshin V. GuptaThe University of Arizona Tucson, Department of Hydrology &Water Resources, Harshbarger Building Room 314, 1133 EastNorth Campus Drive, AZ 857210011, USA
Khaled HaddadUniversity of Western Sydney, Building XB, Kingswood Schoolof Engineering, UWS Locked Bag 1797, Penrith, NSW 2751,Australia
David M. HannahUniversity of Birmingham, School of Geography, Earth andEnvironmental Sciences, Edgbaston, Birmingham,B15 2TT, UK
H. A. P. HapuarachchiBureau of Meteorology, GPO Box 1289, Melbourne, VIC 3001,Australia
Hege HisdalNorwegian Water Resources and Energy Directorate (NVE),Hydrology Department, PO Box 5091, Maj., N-0301 Oslo,Norway
Kamila HlavovSlovak University of Technology, Department of Land and WaterResources Management, Radlinskho 11, 813 68 Bratislava,Slovak Republic
Markus HrachowitzDelft University of Technology, Water Resources Section,Stevinweg 1, 2600 GA Delft, the Netherlands
Denis A. HughesRhodes University, Institute for Water Research, PO Box 94,Grahamstown, 6140, South Africa
Gnter HumerDipl.-Ing. Gnter Humer GmbH, Feld 16, 4682 Geboltskirchen,Austria
Ruud HurkmansUniversity of Bristol, University Road, Bristol, BS8 1SS,UK (Previously at Wageningen University, Hydrologyand Quantitative Water Management Group,Droevendaalsesteeg 3a, 6708 PB Wageningen,the Netherlands)
Vito IacobellisPolitecnico di Bari, Dipartimento di Ingegneria delle Acquee di Chimica, Campus Universitario, Via E. Orabona 4,70125 Bari, Italy
Elena IlyichyovaRussian Academy of Sciences, Siberian Branch, V. B. SochavaInstitute of Geography, Ulan-Batorskaya St. 1, Irkutsk, 664033,Russia
Hiroshi IshidairaUniversity of Yamanashi, Interdisciplinary Graduate School ofMedicine and Engineering, 4311 Takeda, Kofu, Yamanashi4008511, Japan
Graham JewittUniversity of KwaZulu-Natal, School of Agricultural, Earth andEnvironmental Sciences, PBag X01, Scotsville, 3209,South Africa
Shaofeng JiaChinese Academy of Sciences, Institute of Geographical Sciencesand Natural Resource Research, No. 11A Datun Road, Beijing100101, China
Jeffrey R. KennedyUS Geological Survey, 520 N. Park Ave, Suite 221, Tucson,AZ 85719, USA
Anthony S. KiemThe University of Newcastle, Environmental and Climate ChangeResearch Group, School of Environmental and Life Sciences,Faculty of Science and Information Technology, Callaghan,NSW 2308, Australia
Robert KirnbauerVienna University of Technology, Institute of HydraulicEngineering and Water Resources Management, Karlsplatz13/2222, 1040 Vienna, Austria
Thomas R. KjeldsenCentre for Ecology & Hydrology, Maclean Building, CrowmarshGifford, Wallingford, Oxfordshire, OX10 8BB, UK
Jrgen KommaVienna University of Technology, Institute of HydraulicEngineering and Water Resources Management, Karlsplatz13/222, 1040 Vienna, Austria
x List of contributors
-
Leonid M. KorytnyRussian Academy of Sciences, Siberian Branch, V. B. SochavaInstitute of Geography, Ulan-Batorskaya St. 1, Irkutsk, 664033,Russia
Charles N. KrollSUNY College of Environmental Science and Forestry,Environmental Resources Engineering, Syracuse,NY 13210, USA
George KuczeraThe University of Newcastle, Faculty of Engineering and BuiltEnvironment, Engineering A130, University Drive, Callaghan,NSW 2308, Australia
Gregor LaahaUniversity of Natural Resources and Life Sciences, Institute ofApplied Statistics and Computing, Gregor Mendel-Str. 33, 1180Vienna, Austria
Henny A. J. van LanenWageningen University, Hydrology and Quantitative WaterManagement Group, Droevendaalsesteeg 3a, 6708 PBWageningen, the Netherlands
Hjalmar LaudonSwedish University of Agricultural Sciences (SLU), Departmentof Forest Ecology and Management, 901 83 Ume, Sweden
Jens LiebeUnited Nations University, UN-Water Decade Programme onCapacity Development (UNW-PC), UN Campus,Hermann-Ehlers-Str. 10, 53113 Bonn, Germany
Shijun LinGuangdong Research Institute of Water Resources andHydropower, No.116 Tianshou Road, Tianhe district, Guangzhoucity, 510635, China
Gran LindstrmSwedish Meteorological and Hydrological Institute,Folkborgsvgen 1, 601 76 Norrkping, Sweden
Suxia LiuChinese Academy of Sciences, Key Laboratory of Water Cycle &Related Land Surface Processes, Institute of GeographicalSciences and Natural Resources Research, No. A11 Datun Road,Beijing 100101, China
Jun MagomeUniversity of Yamanashi, Interdisciplinary Graduate School ofMedicine and Engineering, 4311 Takeda, Kofu 4008511,Japan
Danny G. MarksUSDA Northwest Watershed Research Center, 800 Park Blvd.,Ste 105, Boise, ID 837127716, USA
Dominic MazvimaviUniversity of the Western Cape, Department of Earth Sciences,Private Bag X17, Bellville 7535, Cape Town, South Africa
Jeffrey J. McDonnellGlobal Institute for Water Security, National Hydrology Research
Centre, University of Saskatchewan, 11 Innovation Boulevard,Saskatoon SK S7N 3H5, Canada
Brian L. McGlynnDuke University, Nicholas School of the Environment,Division of Earth and Ocean Sciences, Box 90328,Durham, NC 27708, USA
Kevin J. McGuireVirginia Polytechnic and State University, VA Water ResourcesResearch Center and Department of Forest Resources andEnvironmental Conservation, 210-B Cheatham Hall (0444),Blacksburg, VA 24061, USA
Neil McIntyreImperial College London, Department of Civil andEnvironmental Engineering, Imperial College Road, London,SW7 2AZ, UK
Thomas A. McMahonThe University of Melbourne, Department of InfrastructureEngineering, Victoria 3010, Australia
Ralf MerzThe Helmholtz Centre for Environmental Research (UFZ),Department Catchment Hydrology, Theodor-Lieser-Strae 4,06120 Halle/Saale, Germany
Robert A. MetcalfeOntario Ministry of Natural Resources, c/o Trent University,DNA Building, 2140 East Bank Drive, Peterborough, Ontario,K9J 7B8, Canada
Alberto MontanariUniversity of Bologna, Department DICAM, Viale Risorgimento2, 40136 Bologna, Italy
David MorrisCentre for Ecology & Hydrology, Maclean Building,Benson Lane, Crowmarsh Gifford, Wallingford,OX10 8BB, UK
Roger MoussaINRA, UMR LISAH, 2 Place Pierre Viala, 34060 MontpellierCedex 2, France
Lakshman NandagiriNational Institute of Technology Karnataka Surathkal,Department of Applied Mechanics and Hydraulics,Srinivasnagar, Mangalore, Karnataka,575025, India
Thomas NesterVienna University of Technology, Institute of HydraulicEngineering and Water Resources Management, Karlsplatz13/2222, 1040 Vienna, Austria
Taha B. M. J. OuardaMasdar Institute of Science and Technology, Masdar City, AbuDhabi, PO Box 54224, United Arab Emirates
Ludovic OudinUniversit Pierre et Marie Curie Paris 6, Boite 123, 4 PlaceJussieu, 75252, Paris Cedex 05, France
List of contributors xi
-
Juraj ParajkaVienna University of Technology, Institute of HydraulicEngineering and Water Resources Management, Karlsplatz13/2222, 1040 Vienna, Austria
Charles S. PearsonNational Institute of Water and Atmospheric Research, PO Box8602, Riccarton, Christchurch 8440, New Zealand
Murray C. PeelThe University of Melbourne, Department of InfrastructureEngineering, Victoria, 3010, Australia
Charles PerrinIrstea, UR Hydrosystmes et Bioprocds, 1 rue Pierre-Gilles deGennes, CS 10030, 92761 Antony Cedex, France
John W. PomeroyUniversity of Saskatchewan, Centre for Hydrology, Kirk Hall,117 Science Place Saskatoon, SK, S7N 5C8, Canada
David A. PostCSIRO Land and Water Black Mountain, Christian Laboratory,Clunies Ross Street, GPO Box 1666, ACT 2601, Australia
Ataur RahmanUniversity of Western Sydney, School of Computing,Engineering and Mathematics, Locked Bag 1797, Penrith NSW2751, Australia
Liliang RenHohai University, State Key Laboratory of Hydrology, WaterResources and Hydraulic Engineering, No. 1 Xikang Road,Nanjing 210098, China
Magdalena RoggerVienna University of Technology, Institute of HydraulicEngineering and Water Resources Management, Karlsplatz13/2222, 1040 Vienna, Austria
Dan RosbjergTechnical University of Denmark, Department of EnvironmentalEngineering, Miljoevej, Building 113, DK-2800 KongensLyngby, Denmark
Jos Luis SalinasVienna University of Technology, Institute of HydraulicEngineering and Water Resources Management, Karlsplatz13/2222, 1040 Vienna, Austria
Jos SamuelMcMaster University, Department of Civil Engineering, 1280Main St. West, Hamilton, Ontario, L8S 4L7, Canada
Eric SauquetIrstea, UR HHLY Hydrology-Hydraulics, 3 bis quai Chauveau CP 220, 69336 Lyon, France
Hubert H. G. SavenijeDelft University of Technology, Water Resources Section,Stevinweg 1, 2628 CN Delft, the Netherlands
Takahiro SayamaPublic Works Research Institute, International Centre for Water
Hazard and Risk Management, 16 Minamihara, Tsukuba,Ibaraki 3058516, Japan
John C. Schaake1A3 Spa Creek Landing, Annapolis, MD21403, USA
Kevin ShookUniversity of Saskatchewan, Centre for Hydrology, Kirk Hall,117 Science Place, Saskatoon, SK, S7N 5C8, Canada
Murugesu SivapalanUniversity of Illinois at Urbana-Champaign, Department of Civiland Environmental Engineering, 2524 Hydrosystems Laboratory,301 N. Mathews Ave., Urbana, IL 6180, USA
Jon Olav SkienJoint Research Centre European Commission, Institute forEnvironment and Sustainability, Land Resource ManagementUnit, Via Fermi 2749, TP 440, I-21027 Ispra (VA), Italy
Chris SoulsbyUniversity of Aberdeen, Northern Rivers Institute, St. MarysKings College, Old Aberdeen, AB24 3UE, UK
Christopher SpenceEnvironment Canada, National Hydrology Research Centre,11 Innovation Boulevard, Saskatoon, Saskatchewan, S7N 3H5,Canada
R. Sri SrikanthanBureau of Meteorology, Water Division, GPO Box 1289,Melbourne 3001, Australia
Tammo S. SteenhuisCornell University, Biological and Environmental Engineering,206 Riley Robb, Ithaca, NY 148535701, USA
Jan SzolgaySlovak University of Technology, Department of Land and WaterResources Management, Radlinskho 11, 813 68 Bratislava,Slovakia
Yasuto TachikawaKyoto University, Department of Civil and Earth ResourcesEngineering, Graduate School of Engineering, C1 Nishikyo-ku,Kyoto 6158540, Japan
Kuniyoshi TakeuchiPublic Works Research Institute, International Centre for WaterHazard and Risk Management, 16 Minamihara, Tsukuba-shi,Ibaraki-ken 3058516, Japan
Lena M. TallaksenUniversity of Oslo, Department of Geosciences, Postboks 1047,Blindern, N-0316 Oslo, Norway
Drthe TetzlaffUniversity of Aberdeen, Northern Rivers Institute, St. MarysKings College, Old Aberdeen, AB24 3UE, UK
Sally E. ThompsonUniversity of California, Department of Civil andEnvironmental Engineering, 760 Davis Hall, Berkeley947201710, USA
xii List of contributors
-
Elena TothUniversity of Bologna, Department DICAM, Viale Risorgimento2, 40136 Bologna, Italy
Peter A. TrochThe University of Arizona, Department of Hydrology and WaterResources, John W. Harshbarger Building, 1133 E JamesE. Rogers Way, Tucson, AZ 85721, USA
Remko UijlenhoetWageningen University, Hydrology and Quantitative WaterManagement Group, Droevendaalsesteeg 3a, 6700 AAWageningen, the Netherlands
Carl L. UnkrichUSDA-ARS, Southwest Watershed Research Center, 2000E Allen Rd, Tucson, AZ 857191596, USA
Alberto ViglioneVienna University of Technology, Institute of HydraulicEngineering and Water Resources Management, Karlsplatz 13/222, 1040 Vienna, Austria
Neil R.VineyCSIRO Land and Water Black Mountain, Christian Laboratory,Clunies Ross Street, GPO Box 1666, ACT 2601, Australia
Richard M. VogelTufts University, Department of Civil and EnvironmentalEngineering, Anderson Hall, 200 College Avenue, Medford, MA02155, USA
Thorsten WagenerUniversity of Bristol, Department of Civil Engineering, QueensBuilding, University Walk, Bristol, BS8 1TR, UK
M. Todd WalterCornell University, Biological and Environmental Engineering,222 Riley Robb, Ithaca, NY 148535701, USA
Guoqiang WangBeijing Normal University, College of Water Sciences,Xinjiekouwai Street 19, Haidian, Beijing, China
Markus WeilerAlbert-Ludwigs-University of Freiburg, Institute of Hydrology,Fahnenbergplatz, 79098 Freiburg, Germany
Rolf WeingartnerUniversity of Bern, Institute of Geography and Oeschger Centrefor Climate Change Research, Hallerstrasse 12, 3012 Bern,Switzerland
Erwin WeinmannMonash University, Department of Civil Engineering, Building60, Victoria 3800, Australia
Hessel WinsemiusDeltares, Inland Water Systems, Rotterdamseweg 185, 2600 MHDelft, the Netherlands
Ross A. WoodsNational Institute of Water and Atmospheric Research (NIWA),PO Box 8602, Riccarton, Christchurch 8440,New Zealand
Dawen YangTsinghua University, Department of Hydraulic Engineering,100084 Beijing 100084, China
Chihiro YoshimuraTokyo Institute of Technology, Department of Civil Engineering,2121-M14, Ookayama, Tokyo 1528552, Japan
Andy YoungWallingford HydroSolutions Ltd, Maclean Building, BensonLane, Crowmarsh Gifford, Wallingford, OX10 8BB, UK
Gordon Young (IAHS President)34 Vincent Avenue, PO Box 878, Niagara on the Lake, Ontario,L0S 1J0, Canada
Erwin ZeheKarlsruhe Institute of Technology, Institute of Water Resourcesand River Basin Management, Kaiserstrae 12, 76129 Karlsruhe,Germany
Yongqiang ZhangCSIRO Land and Water Black Mountain, Christian Laboratory,Clunies Ross Street, GPO Box 1666, ACT 2601, Australia
Maichun C. ZhouSouth China Agricultural University, College of WaterConservancy and Civil Engineering, Wushan Road, TianheDistrict, Guangzhou 510642, China
List of contributors xiii
-
Foreword
Prediction in ungauged basins: context,challenges, opportunities
Society increasingly looks to science for predictions, or atleast explanations, of events, resources and hazards.Examples appear continually in government committeehearings, serious newspapers and television channels, andinternational re-insurance industries, to choose one com-mercial example. This expectation is particularly obviousin the case of water. People are nervous about (forexample) water shortages, crop-threatening droughts,floods, water chemistry and pricing. Hydrologists wouldbe wise to improve the capacity and reliability of theirpredictions to respond to this societal need.Prediction is commonly thought of as a fundamental
capability of science. Observations and understanding,conceptualised as explanatory theories, allow predictions,which then can be tested to refute or increase confidence inthe original theory. In hydrology, the quality of these testsdoes not typically match up with tests in some othersciences that apply laboratory-tested principles and modelsto the environment. We have much to learn from discip-lines, such as atmospheric science, physical oceanographyand astrophysics, that have been successful at applyinglaboratory-tested principles at large scales in complexenvironments. Although it is true that hydrology has tocontend with the interactions of more media (rock, soil,vegetation, engineered structures) than do the disciplinesmentioned, there are still intellectual traditions, organisa-tional approaches, analytical methods, and technologies tobe learned from them in order to test the generality oflandscape-scale hydrological theories.Empirical investigation, including experimentation, is
another fundamental tool of science. Hydrological investi-gations are conducted in a wide variety of environments with diverse climates, topography, soils, land cover andmanipulation by humans. They also occur at a wide rangeof temporal and spatial scales, and involve single or mul-tiple processes. It is difficult to organise the vast amount ofinformation from these studies into coherent theories.Diverse results are then seen as contradictory or at leastleading to so much confusion that prediction is impossible.Yet, it should not be surprising that results from differentlocations differ in magnitude, even in the absence or pres-ence of certain processes. Theories predict such a result,and allow organised interpretation and resolution of differ-ences in measured magnitudes. Yet the bulk of the hydro-logical literature comprises a conceptually disorderedresource of unique descriptions, single-process studies
and methods, but few attempts at organisation throughthe medium of broadly applicable, quantitative theory, oreven conceptually organised descriptive summaries, thatwould facilitate both understanding and prediction. This isthe fragmentation problem, which the Predictions inUngauged Basins (PUB) initiative has reduced to anencouraging degree, as this book illustrates.There is a resilient meta-hypothesis in hydrological sci-
ence that quantitative theories of linked hydrological pro-cesses at landscape scale, implemented and tested in atransparent and rigorous manner, could leverage the exten-sive body of environmental measurements into more reli-able predictions. This is an interesting and challenging, ifstill unproven, idea. Such a development would requireimprovements in the conduct of both modelling and empir-ical investigations a trajectory assessed in this book forthe specific case of runoff predictions in ungauged basins(and by extension, unrecorded conditions in monitoredlocations).The first attempt at promulgating general hydrological
theory, based on fluid mechanics and thermodynamics,was Eaglesons 1970 book Dynamic Hydrology, followedby his suite of papers in Water Resources Research (1978)laying out a statistical dynamic formulation of variouslinked components of a land-surface water budget. Earlierattempts to develop theories of single processes, such asthe work of Darcy, Richards, Horton, Theis, Toth, Penmanand Monteith, had pointed the way, but Eagleson provideda guide for integration. More recent contributions by manypeople have illuminated ways of dealing with the repre-sentation of processes at a wide range of scales, and withthe unwelcome fact that many material properties import-ant in hydrology exhibit large, but crudely measured andpoorly understood spatial variations.The current book is an assessment of progress in com-
bining models with new data-collection and processingtools to make predictions of streamflow, and therefore ofassociated hydrological fluxes such as evaporation andgroundwater storage changes. It embraces the scientificapproach of making a model-based prediction and thentesting it, and recording the errors in an objective manner.Such a strategy measures progress in skill developmentand facilitates judicious assessment of the reliability ofpredictions. This is not common in hydrology, where cali-bration routinely hides uncertainties in process representa-tion, landscape and material properties, and spatial
-
variability of atmospheric events. However, the globalsurvey of prediction methods indicates that tested, integra-tive modelling of biophysically controlled hydrologicalmechanisms is still a minority activity in catchment-scalestreamflow predictions. Most predictions still consist ofsummaries, calibrations and extrapolations of strictlyempirical information.The local, empirically focused approach has obvious
utility for making certain kinds of predictions within theinterpolated range of measurements, although manyhydrologists have emphasised the degree to which uncer-tainties debilitate the use of such predictions, even in thatrange. The approach is even more unreliable when it isapplied to the important domains that society cares about,which lie outside of monitored localities (the ungaugedbasin), outside of the recorded range of possible events,and in conditions of climate and land cover that may notyet exist but are anticipated. For these (true) predictionchallenges, according to the meta-hypothesis, there wouldbe value if we had methods based on a sound scientificplatform of mechanistic understanding and rigoroustesting. We would then know how well we could predict,and we would be able to agree on critical uncertainties, andto focus scientific research and technological innovationon them. But we are unlikely to select either of these foci ifmost hydrological predictions continue to be pragmaticallyad hoc, not rigorously tested and compared, and aimed atgenerating locally acceptable solutions, rather than trans-ferable hydrological understanding. The PUB initiative hasmade progress in overcoming these parochial vices.It is often said that the urgent applicability of hydro-
logical knowledge to human affairs encourages short cutsand discourages exploitation of the best practices of sci-ence. Although this may be understandable in specificapplications under limitations of time and resources, thereis no fundamental reason why the subsidised researchcommunity needs to limit its investigations in the sameway. The research community is free to address the meta-hypothesis that rigorously formulated biophysical processmodels could assimilate new and better landscape meas-urements to yield predictions tested in the same way thatsome other environmental sciences achieve. This wouldnot require that a prediction be correct at the first attempt,and it would allow an incorrect prediction to be explainedby investigating whether the form of a mechanistic repre-sentation or the value of a critical parameter were accurate,rather than calibrating the prediction against measuredoutput and announcing success. The book assesses pro-gress in this direction.Making this suggestion is not to argue against current
hydrological practice. Much of it is required for urgentpolicy and management (remember that Jos Ortega yGasset said, Life cannot wait until the sciences have
explained the universe scientifically.). Nevertheless, giventhe widespread dissatisfaction with hydrologists ability topredict flood discharges or water yields in ungauged basinsor those expected to result from climate change and otheranthropogenic disturbances, surely there is an argument tobe made for some investment in a different, more distinct-ively scientific strategy on the part of some hydrologists.The results could generate a higher-yielding strategy forthe discipline.The global organisation of the PUB initiative has also
fostered an intermediate approach to organising the diver-sity of knowledge and approaches to runoff prediction,even for conditions and scales for which rigorous mechan-istic models are not yet formulated, or at least parame-terised, adequately for reliable predictions. This approachinvolves comparative hydrology comparing results andthe success of prediction methods at representing hydro-logical outcomes for different regions and scales. Thestrategy, referred to in the book as hydrological synthesis,is a formal way of comparing hydrological experience indiverse circumstances (I would say through the interpret-ation of available theory wherever possible), and is awelcome approach to ordering hydrological knowledge.Hydrological synthesis converts the disordered body ofcase study results into a gradually expanding sample ofgeographical range (or parameter space) that can be com-pared and interpolated. It raises questions about howregional differences and the general behaviour of hydro-logical response result from the nature of landscapeatmosphere interactions over time scales ranging fromseconds to landform evolution time. The compilation ofresponses and predictions produces insights about whichfactors are critical controls on hydrological response atvarious scales. It also indicates the progress made in pre-diction at each scale, and the hydrological uncertaintiesrequiring resolution. Synthesis through comparison ofresults and localities also encourages the use of informa-tion on patterns of hydrological landscape features, such astopography, soil and plant communities, to choose predic-tion methods, to group useful information and to applyresults. These patterns are strongly correlated because theyhave evolved interactively, and the associations betweenthem tend to limit the variability of hydrological responseinto clusters and trends, albeit with a still-unwelcome (forprediction) degree of variability. However, the PUB strat-egy has successfully organised knowledge, yielded trans-ferable generalisations, and highlighted hypotheses forfurther investigation, as the book documents.The fluid mechanical and thermodynamic theories of
biophysical mechanisms at various scales needed forimproving predictions are more securely developed andtested, at least at laboratory scale, than is the other part ofthe hydrological prediction problem which is to attach
xvi Foreword
-
these theories to the very complex boundaries and materialproperties of landscape features. The need to choose tem-poral and spatial resolutions for making these connectionsleads to challenging uncertainties about the operations ofthe mechanistic theories themselves. Although the problemis often characterised as a need for better parameterisation(which in hydrological modelling usually means someform of averaging of response to a stimulus), or forhigher-resolution modelling with the same equations, thereis often a need for better formulation in the sense ofrepresenting better how a process works, or even whichprocess is working. For example, the volume and timing ofrunoff into a channel could be calculated as (a) the result ofoverland flow from a long contributing area with variableabstraction of water from the flow and a high surfaceresistance or (b) of unsaturated and then saturated subsur-face flow from a shorter contributing area with differentforms of flow resistance and water storage along the flowpath. One could calibrate either representation againstmeasurements of rainfall and runoff from a catchment.However, extrapolations of resulting predictions based onthe inaccurate formulation to much larger rainstorms,snowmelt, drier initial conditions, timber harvest or otherreasonably likely conditions would not be reliable. Theunreliability could lead directly to misinformation if theprediction were required to predict not simply runoff butalso soil-moisture patterns and evaporation, erosion, waterquality, land management or effective pollution regulation(for example). This problem of improving process formu-lation, as it relates to hydrological prediction at landscapescale, needs to be tackled systematically through fieldmeasurement campaigns, modelling experiments andsyntheses of the kind documented here that search forenvironmental patterns and extend knowledge beyondindividual case studies. The attractiveness of the challengeis that it invites new discoveries, based on new forms ofmeasurement at a still-unsampled range of scales, as wellas improvements in technology and physical and math-ematical technique.The landscape-features side of the problem is also an
essential and attractive research target, but it also is chal-lenging. Critical quantities vary in complicated, irregularand wide-ranging ways, and for some of them there is noagreed-upon measurement method. An extra impedimentarises because the disciplines that have studied these fea-tures have attracted fewer scientists with an appetite for
quantitative, theory-based generalisation. Thus, we haverelatively few high-quality measurements of hydrologicallandscape properties and few quantitative theories of howcoherent patterns and random variations develop in the firstplace, or how they differ from place to place. Techno-logical developments have occurred in the measurementof surface hydrological properties, such as topography andalbedo, and new data-processing methods for analysingand representing patterns are being employed. But themeasurement and useful representation of subsurfacematerial properties and geometry remains a serious impedi-ment to prediction. Modelling the co-evolution of land-scape patterns is beginning to develop, which shouldconstrain the number and types of patterns that need tobe considered for hydrological prediction.In addition, field scientists need to engage with the task of
theory-building in order to make useful measurements ofcritical properties that would encourage coherent progressin hydrological prediction. Field studies need to be designedand reported in a manner consistent with theoretical gener-alisation and hypothesis testing. The task of field studies isnot only to report on exceptional circumstances (althoughthese extend the sampling of geographical conditions), butto gradually extend understanding in a coherent and replic-able manner. Often, reports that are portrayed as defyingconventional wisdom turn out, at least qualitatively, to bephysically reasonable and even predictable when extanttheory is applied to the local circumstance. The strategy ofhydrological synthesis a formal way of comparing hydro-logical experience in diverse circumstances (I would saythrough the interpretation of available theory, whereverpossible) emphasised in this book is a welcome approachto ordering hydrological knowledge and setting an agendafor new discoveries and generalisations.The book expresses an enduring aspiration of the com-
munity that established the International Association ofScientific Hydrology (since renamed as International Asso-ciation of Hydrological Sciences). It re-focuses the goal ofone branch of that community on a distinctively scientificapproach to understanding and utilising hydrology at atime when technological advances in measurement andcomputation are becoming available for creative exploit-ation in the service of humankind. This is an excitingprospect.
Thomas Dunne
Foreword xvii
-
Preface
Sustainable management of river basins requires a varietyof tools that can generate runoff predictions over a range oftime and space scales. The most widely used predictivetools for runoff are essentially data-driven, i.e., they areestimated from gauged data. Unfortunately, in most catch-ments around the world runoff is not gauged. In any givenregion, in any part of the world, only a small fraction of thecatchments possess a stream gauge where runoff is gauged.All other catchments have no stream gauge, and are there-fore ungauged, and yet runoff information is needed almosteverywhere people live for a multitude of managementpurposes.Lack of universal theories or equations applicable dir-
ectly at the catchment scale has led to a plethora of modelsbeing developed and used for predicting runoff. Thesemodels differ markedly in their model concepts and struc-ture, their parameters, and the inputs they use. They alsodiffer in terms of what dominant processes they represent,and the scales at which they make predictions. Mostmodels are developed by people with different disciplinarybackgrounds, while benefiting from local observations,experiences and practices, which are influenced by localclimate conditions and catchment characteristics. Conse-quently, they tend to have unique features not applicable inother places: every hydrological research group around theworld seemingly studies a different object: their localcatchment. The net result has been considerable fragmen-tation, a cacophony, and a dissipation of effort that is notconducive to further advances.The Decade on Predictions in Ungauged Basins (PUB)
launched by the International Association of HydrologicalSciences (IAHS) in 2003 was aimed at achieving majoradvances in the capacity to make predictions in ungaugedbasins, through harnessing improved understanding of cli-matic and landscape controls on hydrological processes.The future vision of PUB was indeed to help make a trans-formation from cacophony to a harmonious melody. Oneof the clear tasks that the PUB initiative set out to achievewas to address the fragmentation of modelling approachesthrough comparative evaluation: Classify model perform-ances in terms of time and space scales, climate, datarequirements and type of application, and explore reasonsfor the model performances in terms of hydrologicalinsights and climatesoilvegetationtopography controls.This book has completed such a comparative evaluation,which is one of its major highlights.
However, PUB also had a higher ambition. It was feltthat focusing on a grand problem such as PUB, whichneeded to draw heavily on new fundamental and theoret-ical advances in hydrology and associated earth systemsciences to address the immediate problem-solving needsof society, had the potential benefit of enabling hydrologyto meet both its scientic and its societal obligations. Inother words, PUB was also seen as the vehicle to advanceand revitalise the science of hydrology. Indeed, over thepast decade, the PUB community has made huge strides inadvancing both predictive capability and fundamentalunderstanding of hydrological processes, working togetherin a concerted and coordinated manner. The PUB effort hashelped to challenge long-held assumptions and questioncommon paradigms, and has increased the constructivedialogue between different sub-disciplines and schools ofthought.So, as PUB comes to an official close, this book repre-
sents one contribution to PUB, which can be seen asanother important step both in the development of PUBand in the growth of hydrology as a holistic earth science.This book does not even pretend to document all of theadvances and contributions that the PUB community hasachieved. These are substantial and should not gounnoticed, and will be documented in some other form.This book is mainly and explicitly focused on a synthesisof runoff predictions in ungauged basins, organisedacross processes, places and scales. This synthesisattempts to place current practice, experience and predic-tion uncertainty of a range of prediction methods in anordered way, so that new insights can be gained not onlyabout the methods themselves but also about how runoffvariability changes across processes, places and scales.We believe we have succeeded in bringing some order tothe disorder, to the extent of at least partially helping totransition from cacophony to a harmonious melody,even if it meant asking new questions that only futureresearch can answer. We believe that this is a positivedevelopment both for predictions and for the science.This is also reflected in the organisation of the book, asone whole, internally self-consistent tome, with onecoherent theme, rather than a collection of chapters thatfocus on several different PUB themes that one couldequally well produce under the circumstances. We hadto make a choice about the organisation of the book inour quest for the harmony; we hope that it made a
-
difference, since there was a clear price we had to pay forachieving this harmony.The book started as a PUB benchmark report, but over
time metamorphosed into a synthesis, mainly building on acomparative assessment of thousands of studies from allaround the world, with predictive uncertainty assessedalong the axes of processes, places and scales, and inter-preted hydrologically. The literature on the current state ofthe art of runoff predictions, in terms of the various signa-tures, was reviewed by over 130 contributing authors, andorganised into several chapters. The painstaking effort ofcomparative assessment was carried out by several ableassistants in Vienna, who toiled hard to do the analyses andgenerate new insights from them, with the support ofseveral of the book contributors as well as the cooperationof scores of the original study authors, who provided thedata sources needed for the assessment. All of the chapterswent through numerous (countless) revisions by the con-tributors themselves, and then by the editors, as part of theoverall synthesis effort.The evolution of both the outlines and the contents of
the book over the past three years or more was a processunto itself, confusing and confounding editors and con-tributors alike as it moved along, crystallising into itspresent shape only in the final few months. In other words,the editorial process was a co-evolutionary process, no lessa complex system than the catchment that is the subject ofour study. In that sense, the blind men and the elephantmetaphor applies to the book just as it does to the PUBinitiative. Some messages that appear in the synthesischapter were not there at the beginning, or were onlyvaguely there, and emerged through the process of writingthe book. In that sense, the synthesis did serve its purpose.However, we are humble enough to admit that the book isnot the end, but only the end of a beginning, and we remainblind men, still blind to the true reality that is catchmenthydrology. Hopefully, there is a broader vision thatchallenges the narrow vision that has helped us to shapethis book.Readers will not fail to notice scores of photographs of
real catchments in colour throughout the book. Our deci-sion to include them is a tribute to the late Vit Kleme,former President of the IAHS, and comes from our deter-mination that catchments henceforth be seen as inimitableobjects that are alive, and not just defined by the tech-niques and abstractions we may use, from time to time, toanalyse them. In a series of papers, Kleme emphasised theprimacy of process understanding over techniques inhydrological research. Techniques are certainly neededfor predictions in ungauged basins, but the focus in thisbook as in much of the PUB initiative has been on thehydrology, with the techniques playing an essential butsupporting role. It is the hydrological interpretation of the
patterns that various analysis techniques and models pro-duced that takes centre stage in this book.The contents of the book, in a sense, reflect the lessons
learned from the diversity offered by natures own experi-ments, as expressed through the thousands of studiessurveyed in this book. Through its comparative perform-ance assessment of methods across processes, places andscales, the book takes an approach to generalisationthrough learning from the differences and similaritiesbetween catchments around the world. It throws light onthe status of PUB at the present moment and can serve as abenchmark against which future progress can be judged.Along the way, the book has also come out with a newscientific framework that can potentially guide futureefforts aimed at improving runoff predictions in ungaugedbasins and at advancing the science of hydrology. Thisproposed new framework has centred on a higher-levelsynthesis of what we can learn from individual places withgeneralised understanding gained from comparing the dif-ferences between places, thus benefiting from legacies ofco-evolution. Maybe this is too esoteric to include in abook on predictions, or maybe it will trigger new ways ofdoing things; only time will tell.This book is aimed at hydrologists, earth and environ-
mental scientists with an interest in water, especially earlycareer scientists, aspiring graduate students and practition-ers, as well as hydrology teachers. Yet it is not a textbook,manual, handbook nor even a monograph. It puts predic-tions in a new light, it makes catchment hydrology morecoherent and exciting, and runoff variability and waterbalance more holistic. Perhaps the book can serve as aready reference for students and new entrants to hydrologywanting to get a sense of the holistic nature of catchmentwater balance, and to imagine new and innovative ways tomake runoff predictions. We hope that, in the long run, itwill change the way hydrology is taught, researched andpractised.The book owes its genesis to the IAHS Predictions in
Ungauged Basin initiative. We are truly grateful to theIAHS for having the wisdom and courage to start such a10-year global effort, and for bringing the hydrologicalcommunity together globally on such a daunting task.We could not have completed this book without the under-lying community support that IAHS managed to pulltogether. Indeed, the synthesis presented in the book isbuilt on the collective experience, inputs and insights of alarge number of researchers around the world, and istherefore truly a grassroots effort, even if the grass rootsare no longer visible. We as editors hope that, in spite ofthe challenges and frustrations faced along the way, it hasbeen a worthwhile effort and that the final product reflectswell on the best intentions and profound wisdom of thehydrology community, and the high ambitions of the PUB
xx Preface
-
initiative. We are grateful to Kuniyoshi Takeuchi forhaving the foresight to launch the initiative and to himand subsequent Presidents (Arthur Askew and GordonYoung), and Secretary-General Pierre Hubert, for provid-ing unconditional support to the PUB initiative generally,and to this book project in particular. We are also gratefulto Jeff McDonnell and John Pomeroy, the other two PUBChairs, for leading PUB during their terms in a way thatthe PUB spirit never wavered, and PUB continued to makeprogress towards its goals.We would like to profusely thank the 130 contributors,
including coordinating contributors, for their efforts atpulling together material for the various chapters, and forputting up with us as we continually edited the material tothe point that, very truly, not one sentence that any authorcontributed remains intact in the book. We thank them allfor having faith in the editors to the end, in spite of theobvious frustrations. Special thanks to Magdalena Rogger,Thomas Nester, Jrgen Komma, Juraj Parajka, Jose LuisSalinas, Emanuele Baratti, Rasmiaditya Silasari, PatrickHogan and Gemma Carr for the invaluable assistance theyrendered towards the comparative assessment exercise,redrawing of the figures, editing, proofreading and overallproject management. Without their efforts, this book
would never have seen the light of day. We would alsolike to acknowledge the financial support of the AustrianAcademy of Sciences through the Predictions in UngaugedBasins project, and the US National Science Foundationthrough the Hydrologic Synthesis project, and the insti-tutional support of our respective employers, especiallyVienna University of Technology and the University ofIllinois, for making available considerable resources toenable us to work together over extended periods. We aregrateful to Thomas Dunne for being willing to write aforeword to the book; the PUB initiative benefited fromToms wisdom during its formative stages, and it is grati-fying that he followed the initiative to its end and was ableto read the book and offer wise thoughts once again topotential readers. Finally, we would like to extend ourthanks to our respective families for putting up with ourlong absences, in both body and spirit, over the three yearsthat it took to complete this book.
G. BlschlM. SivapalanT. WagenerA. Viglione
H. H. G. Savenije(Editors)
Preface xxi
-
Abstract
This book is devoted to predicting runoff in ungaugedbasins (PUB), i.e., predicting runoff at those locationswhere no runoff data are available. It aims at a synthesisof research on predictions of runoff in ungauged basinsacross processes, places and scales as a response to thedilemma of fragmentation in hydrology. It takes a com-parative approach to learning from the differences andsimilarities between catchments around the world. Thebook also provides a comparative performance assess-ment (in the form of blind testing) of methods that arebeing used for predictions in ungauged basins, inter-preted in a hydrologically meaningful way. It thereforethrows light on the status of PUB at the present momentand can serve as a benchmark against which future
progress on PUB can be judged. In so doing, thebook has also come out with a new scientific frameworkthat can guide the advances that are needed to underpinPUB and to advance the science of hydrology as awhole. The synthesis presented in the book is built onthe collective experience of a large number of research-ers around the world inspired by the PUB initiative ofthe International Association of Hydrological Sciences,which makes it truly a community effort. It has providedinsights into the scientific, technical and societal factorsthat contribute to PUB. On the basis of the synthesispresented in this book, recommendations are made onthe predictive, scientific and community aspects of PUBand of hydrology as a whole.
-
Abstract xxiii
-
1 IntroductionContributors: G. Blschl,* M. Sivapalan, T. Wagener,A. Viglione and H. H. G. Savenije
1.1 Why we need runoff predictions
During the February 2007 Zambezi River flood, PauloZucula, the director of Mozambiques National Institutefor Disaster Management, was trying to contain the disas-ter along the river: The evacuated people have been incamps for over a week without proper feeding they areisolated and we cant go there by road, so we have to airliftsome of them and drop food, he said. Some 90 000 peoplewere made homeless by the flood. According to an Oxfamworker, about 1000 people a day were arriving at thecamps even without any shelter being provided. The gov-ernment had learned the lessons of the previous 2001flood, however, during which about 700 people died. Thistime it promptly launched missions by boat and helicopterto evacuate people from affected areas. But they wererapidly running short of food for the people in the 33temporary camps, which also lacked tents, medicine andclean water.In January 2008 a major flood again struck the Zambezi.
This time some 50 000 people in Mozambique were dis-placed by the flood. Property and infrastructure is againbeing wrecked but we are more worried about the people,said Paulo Zucula. The flood in the Zambezi valley was infact worse than the floods of February 2007, and theauthorities were forced to evacuate areas where the victimsof earlier floods had been resettled.What has this disaster got to do with runoff Predictions
in Ungauged Basins (PUB)? A lot! The hydrology of theZambezi valley in Mozambique is strongly affected by thepresence of the Cahora Bassa Dam (see Figure 1.1). Thisdam, designed to release about 1900 m3/s through itsturbines for hydropower generation, has a limited floodrelease capacity. In order to deal with major flooding, ithas to lower its reservoir level substantially before theonset of the flood season each year. It is a perpetualtrade-off between the economic value of hydropower gen-eration, the risk of flood damage and the risk of dam
failure. What makes the operation of the dam even morecomplicated is that large parts of the upstream catchmentof the Zambezi are literally ungauged. Runoff in the mainstream of the Zambezi River may be governed by theupstream Kariba Dam, from where warnings are issuedwhenever they open the flood gates, but the operators haveno knowledge of the inflow from the intermediate catch-ment, of which the Luangwa with its 50 000 km2 is thelargest. The Luangwa is completely ungauged. As a result,operators sometimes have to open the floodgates anddischarge more water than would be necessary, with thebenefit of hindsight. Better runoff predictions in theLuangwa could reduce flood releases, increase hydropowerproduction, improve flood warning, and reduce down-stream damage and suffering.In Chapter 11, a case study by Hessel Winsemius shows
that, even in an ungauged basin such as the Luangwa,much can be done in terms of improved flood predictions.Figure 1.2 is a screen dump of the online model thatWinsemius developed for the Luangwa River basin, com-pletely based on remotely sensed data (mostly precipitationand meteorological information). The model has been
Figure 1.1. The Cahora Bassa Dam, Mozambique, spilling throughone of the eight flood gates.
Runoff Prediction in Ungauged Basins: Synthesis across Processes, Places and Scales, ed. Gnter Blschl, Murugesu Sivapalan, Thorsten Wagener,Alberto Viglione and Hubert Savenije. Published by Cambridge University Press. Cambridge University Press 2013.
* Coordinating contributor
mozkadiogluHighlight
mozkadiogluHighlight
-
operational since 2009 and gives hourly updates of runoffestimates. The runoff is very substantial compared to theaverage outflow of the Cahora Bassa Dam, and so thesepredictions can indeed make a difference between life anddeath, early warning and forced displacement. It shows thedirect relevance of PUB for society.The challenges in the routine management of the Cahora
Bassa Reservoir demonstrate the importance of runoffpredictions for reservoir management, but there are manyother purposes for which runoff predictions are needed.Flood predictions are needed for the design of spillways,culverts, dams, levees, reservoir management, river restor-ation and risk management. Low flow predictions areneeded for determining environmental flows for ecologicalstream health, drought management, river restoration andassessing the dilution of discharges into a stream. Table 1.1illustrates the range of problems for which runoff predic-tions in the context of integrated water resources and risk
management are needed. All of them have direct societalrelevance (Carr et al., 2012). Clearly, runoff predictionsare important to a large part of humanity.Unfortunately, in most catchments around the world,
runoff is not measured. In any given region, in any partof the world, only a fraction of the catchments possess astream gauge where water levels are gauged, which arethen transformed into runoff, i.e., the volume of water perunit time that flows through a cross-section of a stream. Allthe other catchments have no stream gauge, and so areungauged, and yet runoff information is needed almosteverywhere people live for the multitude of purposes out-lined above.The only recourse is therefore to predict runoff in these
catchments or locations using alternative data, or infor-mation or knowledge. How one can predict runoff for theseungauged catchments and how well one can do this are thesubject matter of this book.
Figure 1.2. The online modelthat predicts runoff from theLuangwa River basin into theZambezi, just upstream from CahoraBassa Reservoir.
Table 1.1. Need for runoff predictions in ungauged basins
Hydrological problem Water management purpose
How much water do we have? Water allocation, long-term planning, groundwater rechargeWhen do we have water? Water supply and hydropower production, planning of restoration measuresFor how long do we have water? Ecological purposes, hydropower potential, industrial and domestic
water supply, irrigationHow dry will it be? Environmental flows for ecological stream health, drought management, river restoration,
assessing dilution of effluentsHow high will the flood be? Design of spillways, culverts, dams, dam removal, levees, reservoir management, river restoration,
risk managementWhat are the dynamics of runoff? All of the above plus water quality (sediments, nutrients) predictions
2 Runoff Prediction in Ungauged Basins
-
1.2 Runoff predictions in ungauged basins aredifficult
So, how can one predict runoff at the catchment scale?Unfortunately, there are currently no universal theoriesor equations applicable for predicting runoff at the catch-ment scale. Most of the knowledge we have of processesthat occur within the catchment has been derived at thepoint or laboratory scale (Dooge, 1986; Blschl, 2005b).The equations of flow of water are essentially valid at thelaboratory scale. Similarly, theories of infiltration we cur-rently use are point-scale equations, and overland flow isclearly defined at the hydrodynamic scale, developed inhydraulic laboratories where turbulent processes are verywell researched. The challenge for predictions is to movefrom the well-researched point-scale equations to thecatchment scale, something termed the upscaling problem.One way of addressing the upscaling problem is to dividethe catchment into smaller elements, which are smallenough to apply these point-scale equations, and thenassemble these pieces together to form a model of theentire catchment to make the required runoff predictions.This approach could work, in principle geometrically thecatchment can be easily decomposed into sufficiently uni-form elements. This so-called reductionist approach is thenthe most logical way of building predictive models. Forestimating runoff in ungauged catchments the approachwill then lead to a form of distributed process-based hydro-logical models that solve the governing equations for mass,momentum and energy in a spatially explicit way, drawingon as much laboratory-scale process understanding aspossible. In this book we will call this the Newtonianapproach, as the essence of such models is based onNewtonian physics or mechanics.The Newtonian approach has numerous strengths. First
and foremost, it is based on cause-and-effect relationships.If you change an input or a parameter of the model at somelocation, there is a clearly defined response of the runoff tothis change. This is very important for many applications,in particular for those related to change prediction. Landuse change effects can be directly simulated by these typesof models and, similarly, the approach lends itself naturallyfor climate impact analyses. Second, these models arespatially explicit and have the potential to represent pro-cesses within the catchment in much detail, such as spatialpatterns in the infiltration characteristics, the exact channelshape or the presence of any hydraulic structures. Again,there is considerable benefit in the spatial representation, asany detailed knowledge one may have about the catchmentcan be fully exploited. Third, the underlying equations,such as Darcys law or Mannings equation are known towork at the laboratory scale for a wide range of flow condi-tions, so it should be possible to extrapolate them to a wide
range of hydrological situations, such as under much higherprecipitation. The underlying equations are universal, soshould be applicable everywhere at all times. This isappealing since it will generate generalisable knowledge.Also, there are many examples from sister disciplines, suchas the atmospheric sciences and river hydraulics, wheredistributed models are the universal currency.However, there are three problems with the Newtonian
approach for predicting catchment runoff. When subdivid-ing the catchment into computational units, it is necessaryto characterise the system through which the water flowsfor every single element. In principle, this may appear to bea trivial task, but in practice it turns out to be very difficult.In essence, the medium through which the water flows isunknown. It is difficult to identify the spatial (and depth)distribution of the flow parameters, such as the hydraulicconductivity that describes how easily water movesthrough a medium such as soil or rock. The runoff esti-mated by the models is usually very sensitive to theseparameters, and even a small change will produce a bigchange in runoff. It is not feasible to measure these param-eters everywhere in a catchment, even in a research catch-ment, let alone in routine applications needed in waterresources management, where almost always there arestrict resource and time limitations. Second, even if wewere able to characterise parameters such as hydraulicconductivity and roughness for every pixel within a catch-ment, computational resources currently do not allow us toactually use laboratory-scale computational elements atleast a trillion elements would be needed for a catchment ofpractical interest. Because of this, the elements or buildingblocks of distributed processes-based hydrological modelsare usually much larger, at least tens of metres. This leadsto the problem of quantifying the flow processes withinsuch elements, i.e., how to parameterise the effects ofsubgrid variability. This parameterisation is not very wellunderstood either. Preferential flow phenomena may leadto flow dynamics that are very different from those at thelaboratory scale. Third, many of the processes controllingcatchment runoff are in fact not physical processes butchemical and biological processes. For example, soilchemical processes may strongly affect the infiltrationcharacteristics. Biological activity of earth worms andplants may alter the hydraulic conductivity considerably.Streamaquifer interactions are often controlled by bio-logical activity at their interface and transpiration is, ofcourse, a biologically driven process. So, while the flowprocesses per se are physical phenomena, they are con-trolled by many other processes that cannot be quantifiedby means of Newtonian physics.Because of these issues, distributed process-based
hydrological models often tend to produce biased resultswhen applied to real catchments. To reduce bias in the
Introduction 3
-
runoff predictions, at least some of the model parametersneed to be calibrated to runoff data. However, this is ofcourse not possible in ungauged basins.A number of alternative methods have therefore been
developed and have been the methods of choice in prac-tical applications for a long time. These alternativesinvolve the use of runoff data from gauged catchments ina region, and models for ungauged catchments thatstrongly build on these runoff data. These can be statisticalmodels or simple process models of a conceptual kind,without recourse to Newtonian physics. However, thesemodels are centred on the notion of similarity betweenthe gauged and the ungauged catchments. These types ofmodels acknowledge that, even though there are no runoffdata in the catchment of interest, runoff data do exist inother, similar catchments, and these can be transferred insome way in space to help make runoff predictions in theungauged basins.
1.3 Fragmentation in hydrology
Because distributed process-based hydrological models arenot the only method of making runoff predictions inungauged basins, a plethora of other methods have beendeveloped that are based on the notion of similarity. There isno one standard method of runoff predictions in ungaugedbasins, rather there are literally hundreds of differentmethods. They differ by their model structure, their param-eters, and by the inputs they use. They also differ in whatprocesses they represent. Depending on the environments,the relative role of snow processes, runoff generation pro-cesses and transpiration processes may differ, as may thefactors that control them. Some of the differences betweenthe models are directly related to the differences in climateand catchment characteristics. Also, historically, hydrolo-gists have had less incentive than researchers in other dis-ciplines in the earth sciences to collaborate with colleaguesaround the world, as the land surface is organised intoseparate river basins, and there is little water exchangeacross them. Unlike meteorology, for example, a singlecatchment can be studied with much success in isolation.As hydrologists we do not have a single object of study as,say, a physicist who studies the structure of a particularatom. All physicists around the world may study the hydro-gen atom and the models they come up with relate to thesame common object one hydrogen atom. In contrast,every hydrological research group around the world isstudying a different object, i.e., a different catchment withdifferent response characteristics. This is a fundamentaldifference that hydrology must face up to.All of these factors, collectively, have contributed to the
fragmentation of hydrology at various levels.
Processes: Different processes in hydrology have oftenbeen dealt with separately, and therefore hydrologists haveoften looked at flow characteristics at different time scalesin an independent way. The annual water yield is usuallystudied independently of knowledge of low flows of thecatchment of interest; floods are often studied independ-ently of knowledge of the seasonal flow patterns withincatchments; and flow duration curves are studied separ-ately. Is there a deeper connection between these pro-cesses? What is needed is a simultaneous treatment ofthese processes at different time scales.
Places: As each research group has tended to analysetheir own catchments, over the years tremendous under-standing of runoff processes has been developed for indi-vidual places, but transferring this to other places has beenhard. Models are often tailor-made to a particular catchmentand it is hard to reason why a particular model structure ormodel parameters should be preferred over others. Differentschools of thought have developed their own favouritemethods for different environments and purposes, e.g., stat-istical versus causal methods or physically based versusconceptual models. Generalising the findings of how wellthe models work and why has been notoriously difficult.
Scales: Research has been performed over a huge rangeof scales, and connecting them has caused tremendousdifficulties. This is known as the scale problem in hydrol-ogy (Blschl and Sivapalan, 1995). When upscalinglaboratory-scale infiltration equations to the catchmentscale, assumptions need to be made about the naturalhydrological variability and how it is organised (Blschl,2001). Similarly, routing equations at a plot scale maydiffer from those at the hillslope scale. This situation hasbeen exacerbated by the spectrum of disciplines involved,including engineers, geologists, soil scientists and meteor-ologists, each of them with different worldviews of at whatscales processes should be conceptualised.Current textbooks on hydrology propagate the same
fragmented vision of hydrology, organised by process,and written in the form of recipes, e.g., ten different for-mulas for estimating infiltration, potential evaporation, andso on. The situation is literally analogous to a cacophonyof noises not a harmonious melody (Sivapalan, 1997;Sivapalan et al., 2003b). This fragmentation can be bestillustrated by the famous Indian legend of the six blindmen and the elephant. By touching different parts of thebody of an elephant, these blind men are trying to figureout for themselves what an elephant may look like, buthave no other way of seeing it. Each of them tries tomake inferences about the elephant by touching one bodypart of the elephant: it seems like a wall to the blind manthat touches the side of the elephant, a spear to the one whofeels the tusk, a snake to the one who handles the trunk, atree to the one who feels the leg, a fan to the one who
4 Runoff Prediction in Ungauged Basins
-
touches the ear, and a rope to the one who touches the tail(Figure 1.3). The experiences and interpretations of the sixblind men are different, which makes it difficult for them tocome up with a collective understanding and agree on thetrue nature of the beast they are trying to visualise.A verse of John Godfrey Saxes (181687) version of thisfamous Indian legend clearly brings out this confusion:
And so these men of IndostanDisputed loud and long,Each in his own opinionExceeding stiff and strong,Though each was partly in the right,And all were in the wrong!
Perhaps catchment hydrology is in a similar state. Like thesix blind men, hydrologists too are often partly in the rightbut they continue to fail to grasp the holistic picture of thecatchment, the object of their study. There is a clear andurgent need to develop a unified vision of hydrology at thecatchment scale that overcomes the limitations that arisefrom their narrow perspectives. What is needed is a syn-thesis that helps to broaden their perspectives, and to gobeyond what is perceived by an individual researcher orgroup.
1.4 The Prediction in Ungauged Basins initiative:a response to the challenge of fragmentation
About a decade ago, a new global initiative was launchedby the hydrological community, under the aegis of theInternational Association of Hydrological Sciences(IAHS). Called Prediction in Ungauged Basins (PUB),one of the motivations of this grassroots initiative was toovercome the fragmentation in catchment hydrology (Siva-palan et al., 2003b; SSG, 2003). The idea was to bring thescientific community together to use PUB to advance the
collective understanding in hydrology, just as the six blindmen in the Indian legend might have joined forces toenlighten themselves about the elephant to seek wisdomfrom other sources. The PUB initiative has been guided bya number of overarching principles to help reach its goals(Figure 1.4). First and foremost, the initiative was aboutreal hydrological processes in real catchments, embracinga multitude of processes, places and scales. If real progresswas to be made in overcoming the fragmentation, a diversepopulation of real catchments in different regions and atdifferent scales had to be examined. A diverse range ofprocesses and a diverse range of data and approacheshad to be brought together, all focusing on the commonscience problem of predictions in ungauged basins. Bymaking different catchments and methods comparable,the aim was to synthesise existing knowledge and createnew knowledge, and in this way help improve the predict-ive abilities and reduce uncertainty. Comparability ofdiverse places, methods and applica