Runoff Prediction in Ungauged Basins Synthesis across Processes,Places and Scales Predictingwaterrunoffinthemostlyungaugedwatercatchment Gu¨nter Blo¨schl is Professor of Hydrology, Director of the areas of the world is vital to practical applications such as the CentreforWaterResourceSystems,andHeadoftheInstituteof designofdrainageinfrastructureandfloodingdefences,forrun- HydraulicEngineeringandWaterResourcesManagementatthe off forecasting and for catchment management tasks such as ViennaUniversity ofTechnology.Hehaspublishedextensively waterallocationandclimateimpactanalysis. onsubjectsrelatedtohydrologyandwaterresourcesandserved This important new book synthesises decades of rigorous as an editor and associate editor for ten of the best scientific analytical research from around the world, forming a holistic journals in the field. Professor Blöschl has been elected Fellow approach to catchment hydrology, and providing a one-stop oftheAmericanGeophysicalUnionandtheGermanAcademyof resourceforhydrologistsinbothdevelopedanddevelopingcoun- Science and Engineering, has chaired the International Associ- tries. It brings together results from individual location-based ationofHydrologicalSciences(IAHS)PredictionsinUngauged studieswithcomparativeanalysisalonggradientsofclimateand Basins (PUB) initiative, and has been elected President of the landscapefeatures.Topicsincludedataforrunoffregionalisation EuropeanGeosciencesUnion.Recentlyhehasbeenawardedthe and the prediction of runoff hydrographs, flow duration curves, prestigious Advanced Grant of the European Research Council flowpathsandresidencetimes,annualandseasonalrunoff,and (ERC). floods. Murugesu Sivapalan isProfessorofCivilandEnvironmen- Illustratedwithmanycasestudies,andincludingafinalchap- tal Engineering, and of Geography and Geographic Information ter on recommendations for researchers and practitioners, this ScienceattheUniversityofIllinois.Hewasfoundingchairofthe book is written by expert international authors involved in the International Association of Hydrological Sciences (IAHS) Pre- prestigious International Association of Hydrological Sciences dictions inUngauged Basins (PUB) initiative. He has published (IAHS) Predictions in Ungauged Basins (PUB) initiative. It is a extensivelyoncatchmenthydrologyinseveralinternationaljour- keyresourceforacademicresearchersinthefieldsofhydrology, nals and is Executive Editor of the European Geosciences hydrogeology, ecology, geography, soil science, and environ- Union’sHydrologyandEarthSystemSciencesjournal.Professor mental and civil engineering, and professionals working with SivapalanhasalsoreceivedtheEuropeanGeophysicalSociety’s waterrunoffinungaugedwaterbasins. John Dalton Medal, the International Hydrology Prize of the IAHS, and the Hydrological Sciences Award and the Robert E. Horton Medal of the American Geophysical Union. He was also the recipient of the Centenary Medal of the Australian Government and an Honorary Doctorate of the Delft University ofTechnology. ThorstenWagenerisProfessorofWaterandEnvironmental SecurityintheDepartmentofCivilEngineeringattheUniversity ofBristol.HeisaVicePresidentoftheInternationalAssociation of Hydrological Sciences, editor of the journal Hydrology and Earth System Sciences, and associate editor of several other journals.DrWagenerhasbeenawardedDAAD(DeutscherAka- demischerAustauschdienst)fellowships,anIEMSSEarlyCareer ExcellencePrize,BestPaperAwardsoftheJournalofEnviron- mentalModelingandSoftware,aUSEPAEarlyCareerAward, theWalterHuberCivilEngineeringResearchPrizeoftheAmeri- can Society of Civil Engineers, an Alexander von Humboldt Foundation Fellowship, and an Education and Public Service AwardoftheUniversitiesCouncilforWaterResources. Alberto Viglione is a Research Hydrologist at the Vienna UniversityofTechnology.During2004–7heconducteddoctoral researchon‘Non-supervisedstatisticalmethodsfortheprediction of hydrological variables in ungauged sites’ at the Hydraulic Department of the Politecnico of Turin. He has authored or co-authored numerous papers in hydrology, particularly on floods,from bothstatistical andprocess-based perspectives,and onhydrologicalcharacterisationofriverbasins.DrViglionehas developedsoftwareforregionalfrequencyanalysisandrainfall– runoff modelling under the R environment, which is available online.Healsoactsasareviewerforseveralprestigiousjournals, andhasbeeninvolvedinanumberofresearchprojectsrelatedto hydrology and flood frequency analysis in Italy, Austria and severalotherEuropeancountries. Hubert Savenije is Professor of Hydrology andHead of the Water Resources Section at the Delft University of Technology andalsoservesasEditor-in-ChiefofHydrologyandEarthSystem SciencesandPhysicsandChemistryoftheEarth.HeisPresident- Elect of the International Association of Hydrological Sciences (IAHS) as well as Chair of the IAHS Predictions in Ungauged Basins(PUB)initiative.ProfessorSavenijehaspublishedwidely in several leading international journals and was Vice Rector at UNESCO-IHE, Institute for Water Education. He is Past- PresidentofHydrologicalSciencesoftheEuropeanGeosciences Union (EGU), and Past-President of the International Commis- sion on Water Resources Systems of the IAHS. He has been awarded, among other things, the Henry Darcy Medal of the EGUandtheEGUBatchAward. Runoff Prediction in Ungauged Basins Synthesis across Processes, Places and Scales edited by GÜNTER BLÖSCHL TechnischeUniversitätWien,Austria MURUGESU SIVAPALAN UniversityofIllinois,Urbana-Champaign,USA THORSTEN WAGENER UniversityofBristol,UK ALBERTO VIGLIONE TechnischeUniversitätWien,Austria HUBERT SAVENIJE TechnischeUniversiteitDelft,theNetherlands cambridge university press Cambridge,NewYork,Melbourne,Madrid,CapeTown, Singapore,SãoPaulo,Delhi,MexicoCity CambridgeUniversityPress TheEdinburghBuilding,CambridgeCB28RU,UK PublishedintheUnitedStatesofAmericabyCambridgeUniversityPress,NewYork www.cambridge.org Informationonthistitle:www.cambridge.org/9781107028180 ©CambridgeUniversityPress2013 Thispublicationisincopyright.Subjecttostatutoryexception andtotheprovisionsofrelevantcollectivelicensingagreements, noreproductionofanypartmaytakeplacewithout thewrittenpermissionofCambridgeUniversityPress. Firstpublished2013 PrintedandboundintheUnitedKingdombytheMPGBooksGroup AcataloguerecordforthispublicationisavailablefromtheBritishLibrary LibraryofCongressCataloguinginPublicationData Runoffpredictioninungaugedbasins:synthesisacrossprocesses,placesandscales/editedbyGünterBlöschl,Technische UniversitätWien,Austria,MurugesuSivapalan,UniversityofIllinois,Urbana-Champaign,USA,ThorstenWagener,Universityof Bristol,UK,AlbertoViglione,TechnischeUniversitätWien,Austria,HubertSavenije,TechnischeUniversiteitDelft,TheNetherlands. pages cm ISBN978-1-107-02818-0(Hardback) 1. Runoff. 2. Rainandrainfall. 3. Runoff–Mathematicalmodels. 4. Rainandrainfall–Mathematicalmodels. 5. Hydrology. I. Blöschl,Günter,1961–editorofcompilation. GB980.R872013 551.4808–dc23 2012036513 ISBN978-1-107-02818-0Hardback CambridgeUniversityPresshasnoresponsibilityforthepersistenceor accuracyofURLsforexternalorthird-partyinternetwebsitesreferredto inthispublication,anddoesnotguaranteethatanycontentonsuch websitesis,orwillremain,accurateorappropriate. Contents List of contributors page ix 3 Adata acquisition framework for runoff Forewordby Thomas Dunne xv prediction in ungaugedbasins 29 Preface xix 3.1 Whydoweneeddata? 29 Abstract xxii 3.2 Ahierarchyofdataacquisition 30 3.2.1 Assessmentbasedonglobaldatasets 31 1 Introduction 1 3.2.2 Assessmentbasedonnational 1.1 Whyweneedrunoffpredictions 1 hydrologicalnetworkandnationalsurveys 31 1.2 Runoffpredictionsinungaugedbasins 3.2.3 Assessmentbasedonlocalfieldvisits aredifficult 3 includingreadingthelandscape 32 1.3 Fragmentationinhydrology 4 3.2.4 Assessmentbasedondedicated 1.4 ThePredictioninUngaugedBasinsinitiative:a measurements 34 responsetothechallengeoffragmentation 5 3.3 Runoffdata 34 1.5 Whatthisbookaimstoachieve:synthesisacross 3.3.1 WhatrunoffdataareneededforPUB? 34 processes,placesandscales 6 3.3.2 Whatrunoffdataarethere? 35 1.5.1 Synthesisacrossprocesses 7 3.3.3 HowvaluablearerunoffdataforPUB? 36 1.5.2 Synthesisacrossplaces 8 3.4 Meteorologicaldataandwaterbalance 1.5.3 Synthesisacrossscales 8 components 36 1.6 Howtoreadthebookandwhattogetoutofit 9 3.4.1 Whatmeteorologicaldataandwater balancecomponentsareneededfor 2 Asynthesis framework for runoff PUB? 36 prediction inungauged basins 11 3.4.2 Precipitation 36 2.1 Catchmentsarecomplexsystems 11 3.4.3 Snowcoverdata 39 2.1.1 Co-evolutionofcatchment 3.4.4 Potentialevaporation 39 characteristics 11 3.4.5 Remotelysenseddataforcalculating 2.1.2 Signatures:amanifestationof actualevaporation 40 co-evolution 13 3.4.6 Remotesensingofsoilmoistureand 2.2 ComparativehydrologyandtheDarwinian basinstorage 40 approach 15 3.5 Catchmentcharacterisation 41 2.2.1 Generalisationthroughcomparative 3.5.1 Topography 41 hydrology 15 3.5.2 Landcoverandlanduse 41 2.2.2 Hydrologicalsimilarity 18 3.5.3 Soilsandgeology 42 2.2.3 Catchmentgrouping:exploitingthe 3.6 Dataonanthropogeniceffects 43 similarityconceptforPUB 20 3.7 Illustrativeexamplesofhierarchicaldata 2.3 Fromcomparativehydrologytopredictionsin acquisition 44 ungaugedbasins 22 3.7.1 Understandingprocesscontrolsonrunoff 2.3.1 Statisticalmethodsofpredictionsin (TenderfootCreek,Montana,USA) 44 ungaugedbasins 22 3.7.2 Runoffpredictionsusingrainfall–runoff 2.3.2 Process-basedmethodsofpredictions models(ChickenCreek,Germany) 47 inungaugedbasins 23 3.7.3 Forensicanalysisofmagnitudeand 2.4 Assessmentofpredictionsinungaugedbasins 23 causesofaflood(SelškaSora,Slovenia) 49 2.4.1 Comparativeassessmentasameansof 3.8 Summaryofkeypoints 51 synthesis 23 2.4.2 Performancemeasures 25 4 Process realism: flow paths and storage 53 2.4.3 Level1andLevel2assessments 26 4.1 Predictions:rightfortherightreasons 53 2.5 Summaryofkeypoints 26 4.2 Processcontrolsonflowpathsandstorage 55 vi Contents 4.3 Inferenceofflowpathsandstoragefrom 6.3.4 Runoffestimationfromshortrecords 121 responsecharacteristics 57 6.4 Process-basedmethodsofpredictingseasonal 4.3.1 Inferencefromrunoff 57 runoffinungaugedbasins 123 4.3.2 Inferencefromtracers 59 6.4.1 Deriveddistributionmethods 123 4.4 Estimatingflowpathsandstorageinungauged 6.4.2 Continuousmodels 124 basins 64 6.5 Comparativeassessment 126 4.4.1 Distributedprocess-basedmodels 64 6.5.1 Level1assessment 127 4.4.2 Indexmethods 64 6.5.2 Level2assessment 129 4.4.3 Methodsbasedonproxydata 65 6.6 Summaryofkeypoints 134 4.5 Informingpredictionsofrunoffinungauged basins 66 7 Prediction of flow duration curves in 4.5.1 Process-based(rainfall–runoff)methods 67 ungauged basins 135 4.5.2 Statisticalmethods 67 7.1 Forhowlongdowehavewater? 135 4.5.3 Roleoffieldvisits,readingthelandscape, 7.2 Flowdurationcurves:processesandsimilarity 137 photosandotherproxydata 68 7.2.1 Processes 138 4.5.4 Regionalinterpretationandsimilarity 68 7.2.2 Similaritymeasures 141 4.6 Summaryofkeypoints 69 7.2.3 Catchmentgrouping 145 7.3 Statisticalmethodsofpredictingflowduration 5 Prediction ofannual runoff in ungauged curvesinungaugedbasins 147 basins 70 7.3.1 Regressionmethods 148 5.1 Howmuchwaterdowehave? 70 7.3.2 Indexflowmethods 148 5.2 Annualrunoff:processesandsimilarity 71 7.3.3 Geostatisticalmethods 151 5.2.1 Processes 72 7.3.4 Estimationfromshortrecords 152 5.2.2 Similaritymeasures 78 7.4 Process-basedmethodsofpredictingflow 5.2.3 Catchmentgrouping 79 durationcurvesinungaugedbasins 153 5.3 Statisticalmethodsofpredictingannualrunoff 7.4.1 Deriveddistributionmethods 153 inungaugedbasins 83 7.4.2 Continuousmodels 154 5.3.1 Regressionmethods 83 7.5 Comparativeassessment 156 5.3.2 Indexmethods 84 7.5.1 Level1assessment 156 5.3.3 Geostatisticsandproximitymethods 88 7.5.2 Level2assessment 158 5.3.4 Estimationfromshortrecords 88 7.6 Summaryofkeypoints 162 5.4 Process-basedmethodsofpredictingannual runoffinungaugedbasins 89 8 Prediction of low flows inungauged basins 163 5.4.1 Deriveddistributionmethods 89 8.1 Howdrywillitbe? 163 5.4.2 Continuousmodels 90 8.2 Lowflows:processesandsimilarity 164 5.4.3 Proxydataonannualrunoffprocesses 91 8.2.1 Processes 164 5.5 Comparativeassessment 92 8.2.2 Similaritymeasures 167 5.5.1 Level1assessment 92 8.2.3 Catchmentgrouping 170 5.5.2 Level2assessment 96 8.3 Statisticalmethodsofpredictinglowflowsin 5.6 Summaryofkeypoints 100 ungaugedbasins 172 8.3.1 Regressionmethods 172 6 Prediction ofseasonal runoff inungauged 8.3.2 Indexlowflowmethods 175 basins 102 8.3.3 Geostatisticalmethods 176 6.1 Whendowehavewater? 102 8.3.4 Estimationfromshortrecords 178 6.2 Seasonalrunoff:processesandsimilarity 104 8.4 Process-basedmethodsofpredictinglow 6.2.1 Processes 104 flowsinungaugedbasins 179 6.2.2 Similaritymeasures 111 8.4.1 Deriveddistributionmethods 179 6.2.3 Catchmentgrouping 114 8.4.2 Continuousmodels 180 6.3 Statisticalmethodsofpredictingseasonal 8.4.3 Proxydataonlowflowprocesses 180 runoffinungaugedbasins 118 8.5 Comparativeassessment 181 6.3.1 Regressionmethods 118 8.5.1 Level1assessment 182 6.3.2 Indexmethods 118 8.5.2 Level2assessment 184 6.3.3 Geostatisticalandproximitymethods 119 8.6 Summaryofkeypoints 188 Contents vii 9 Prediction of floods in ungauged basins 189 11 PUB inpractice: casestudies 270 9.1 Howhighwillthefloodbe? 189 11.1 PredictionsinUngaugedBasinsinasocietal 9.2 Floods:processesandsimilarity 190 context 270 9.2.1 Processes 191 11.2 Hydrologicalinsightsfromlong-termrunoff 9.2.2 Similaritymeasures 196 patternsacrossKrishnaBasin,India 272 9.2.3 Catchmentgrouping 200 11.3 Predictingmeanannualrunoffacross 9.3 Statisticalmethodsofpredictingfloodsin HuangshuiBasin,China 277 ungaugedbasins 203 11.4 Anindexapproachtomappingannual 9.3.1 Regressionmethods 203 runoffinaSiberiancatchment,Russia 280 9.3.2 Indexfloodmethods 205 11.5 Predictingspatialpatternsofinter-annual 9.3.3 Geostatisticalmethods 208 runoffvariabilityintheCanadianPrairies 283 9.3.4 Estimationfromshortrecords 209 11.6 Seasonalflowpredictionwithuncertainty 9.4 Process-basedmethodsofpredictingfloods inSouthAfricaandLesotho 289 inungaugedbasins 211 11.7 Settingenvironmentalflowtargetsin 9.4.1 Deriveddistributionmethods 212 north-eastUSA 293 9.4.2 Continuousmodels 215 11.8 Continuoussimulationoflowflowsfor 9.4.3 Proxydataonfloodprocesses 217 hydropowerdevelopmentinOntario,Canada 297 9.5 Comparativeassessment 219 11.9 Estimatingflowdurationcurvesfor 9.5.1 Level1assessment 220 hydropowerdevelopmentincentralItaly 300 9.5.2 Level2assessment 222 11.10 ImplementingtheEUflooddirectivein 9.6 Summaryofkeypoints 225 Austria 305 11.11 RevisionofAustralianRainfallandRunoff 10 Prediction of runoff hydrographsin forimprovedfloodpredictions 309 ungauged basins 227 11.12 Understandingflowpathsforhydrograph 10.1 Whatarethedynamicsofrunoff? 227 predictioninanAndeancatchment,Chile 313 10.2 Runoffdynamics:processesandsimilarity 228 11.13 Frequencyofrunoffoccurrenceinephemeral 10.2.1 Processes 229 catchmentsinFrance 317 10.2.2 Similaritymeasures 233 11.14 Overcomingdatalimitationsforhydrograph 10.2.3 Catchmentgrouping 236 prediction,LuangwaBasin,Zambia 321 10.3 Statisticalmethodsofpredictingrunoff 11.15 Remotelysensedlakelevelstoassistrunoff hydrographsinungaugedbasins 238 modellinginGhana 328 10.3.1 Regressionmethods 238 11.16 Modelenhancementsforurbanrunoff 10.3.2 Indexmethods 238 predictionsinthesouth-westUSA 332 10.3.3 Geostatisticalmethods 239 11.17 RunoffpredictionstohelpmeetMillennium 10.4 Process-basedmethodsofpredictingrunoff DevelopmentGoalsinZimbabwe 337 hydrographsinungaugedbasins 240 11.18 RunoffpredictionsinsupportoftheNational 10.4.1 Structureofrainfall–runoffmodels WaterAudit,Australia 345 forungaugedbasins 241 11.19 DistributedrunoffpredictionsintheMekong 10.4.2 Parametersofrainfall–runoff Riverbasin 349 modelsinungaugedbasins: 11.20 ImplementingtheEUWaterFramework overview 246 DirectiveinSweden 353 10.4.3 A-prioriestimationofmodel 11.21 Summaryofkeypoints 360 parameters 247 10.4.4 Transferofcalibratedmodel 12 Outcomes ofsynthesis 361 parametersfromgauged 12.1 Learningfromsynthesis 361 catchments 251 12.2 Synthesisacrossprocesses,placesandscales 363 10.4.5 Constrainingmodelparameters 12.2.1 Synthesisacrossprocesses 363 bydynamicproxydataand 12.2.2 Synthesisacrossplaces 367 runoff 256 12.2.3 Synthesisacrossscales 369 10.5 Comparativeassessment 262 12.2.4 Inter-comparisonofmethods 371 10.5.1 Level1assessment 263 12.3 SynthesisofNewtonianandDarwinian 10.5.2 Level2assessment 266 frameworks 374 10.6 Summaryofkeypoints 268 12.3.1 Evidenceforco-evolution 374 viii Contents 12.3.2 Comparativehydrologyandthe 13.2.1 Viewingcatchmentsascomplex Newtonian–Darwiniansynthesis 376 systems 385 12.3.3 Anewunifieduncertaintyframework 13.2.2 Comparativehydrologytodetect forPUB 379 co-evolutionpatterns 385 12.4 Synthesisandthesciencecommunity 381 13.2.3 Newtonian–Darwiniansynthesis 385 12.4.1 Accumulationofknowledgeinthe 13.2.4 Theglobeisourlaboratory 385 hydrologicalsciences 381 13.3 Organisingthehydrologycommunityto 12.4.2 Roleofthecommunity 382 advancescienceandpredictions 385 13.3.1 Capacitybuilding 385 13 Recommendations 384 13.3.2 Collaborativeendeavour 386 13.1 Advancingrunoffpredictionsinungauged 13.3.3 Knowledgeaccumulation 386 basins 384 13.3.4 Hydrology,aglobalscience 386 13.1.1 Understandingasthekeytobetter 13.4 Bestpracticerecommendationsfor predictions 384 predictingrunoffinungaugedbasins 386 13.1.2 Exploitingrunoffsignaturesand linkingthem 384 Appendix: Summary ofstudiesused inthe 13.1.3 Addressinguncertaintyfromaprocess comparative assessments 388 perspective 384 References 415 13.1.4 Dataavailabilityandpredictions 385 Index 463 13.2 Advancinghydrologicalscienceglobally viaPUB 385