climate change and irrigation demand uncertainty and adaptation

Therefore, understanding the regional impacts of climate change on irrigation demand for crop production is important for water-shed managers and agricultural producers to understand for

Studies

jo u r n a l h o m e p a g e :w w w e l s e v i e r c o m / l o c a t e / e j r h

Sean A Woznickia, A Pouyan Nejadhashemia,∗,

Masoud Parsinejadb

a Department of Biosystems and Agricultural Engineering, Michigan State University, East Lansing, MI,

United States

b Department of Irrigation and Reclamation Engineering, University of Tehran, Karaj, Iran

a r t i c l e i n f o

Article history:

Received 8 January 2014

Received in revised form 12 December 2014

Accepted 17 December 2014

Available online xxx

Keywords:

Irrigation demand

Climate change

SWAT

Crop yield

Adaptation

Uncertainty

a b s t r a c t

Study region:The Kalamazoo River Watershed, southwest Michigan, USA.

Study focus:Climate change is projected to have significant impacts on agri-cultural production Therefore, understanding the regional impacts of climate change on irrigation demand for crop production is important for water-shed managers and agricultural producers to understand for effective water resources management In this study, the Soil and Water Assessment Tool was used to assess the impact of climate change on corn and soybean irrigation demand in the Kalamazoo River Watershed Bias-corrected statistically down-scaled climate change data from ten global climate models and four emissions scenarios were used in SWAT to develop projections of irrigation demand and yields for 2020–2039 and 2060–2079 Six adaptation scenarios were devel-oped to shift the planting dates (planting earlier and later in the growing season) to take advantage of periods with greater rainfall or lower temperature increases.

New hydrological insights for the region:Uncertainty in irrigation demand was found to increase moving from 2020–2039 to 2060–2079, with demand generally decreasing moving further into the future for corn and soybean A shift in timing of peak irrigation demand and increases in temperature lead

to corn yield reductions However, soybean yield increased under these con-ditions Finally, the adaptation strategy of planting earlier increased irrigation demand and water available for transpiration, while delaying planting resulted

in demand decreases for both crops.

©2014TheAuthors.PublishedbyElsevierB.V.Thisisanopen

accessarticleundertheCCBY-NC-NDlicense (http://creativecommons.org/licenses/by-nc-nd/4.0/)

∗ Corresponding author at: Department of Biosystems & Agricultural Engineering, Farrall Agriculture Engineering Hall, 524

S Shaw Lane, Room 225, Michigan State University, East Lansing, MI 48824, United States Tel.: +1 517 432 7653;

fax: +1 517 432 2892.

E-mail address: pouyan@msu.edu (A.P Nejadhashemi).

http://dx.doi.org/10.1016/j.ejrh.2014.12.003

2214-5818/© 2014 The Authors Published by Elsevier B.V This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

xxx.e2 S.A Woznicki et al / Journal of Hydrology: Regional Studies xxx (2015) xxx.e1–xxx.e18

1 Introduction

Populationgrowthandlandusechangeduetoagriculturalexpansionanddeforestationhave sig-nificantlyincreasedpressureonglobalfreshwaterresources(Nejadhashemietal.,2012).Asclimate changebecomesmoreprevalentglobally,thefutureavailabilityoffreshwaterforhuman consump-tion,agriculturalproduction,and manufacturingbecomesmoreuncertain.Bytheendofthe21st century,theprojectedrangeofglobaltemperatureincreasesrelativeto1980–1999isbetween1.1 and6.4◦C,dependingongreenhousegasemissions(Solomonetal.,2007).Meanwhile,themagnitude

ofprojectedprecipitationchangesvarieswidelydependingongeographicregionandspatialscale

impacts.Onaday-to-daybasis,moreheavyprecipitationeventsarepredicted,eveninsomeregions wheremeanrainfallislikelytodecrease(Solomonetal.,2007).Thesechangesareexpectedtoexert additionalpressureonagriculturalproduction,whileatmosphericCO2hasthepotentialtoimprove photosynthesisbyupto30%(LongandOrt,2010)

Anumberofstudieshaveattemptedtounderstandtheeffectsofclimatechangeonwaterusein agricultureintheformofchangesinnetirrigationrequirements,demand,andcropwateruse.Thisis importantbecausetheagriculturalindustryisthelargestuseroffreshwater–waterwithdrawalsfor irrigationaccountfor70%ofallwateruseglobally(Fischeretal.,2007).However,amajorityofthese studieswereperformedonlargespatialscales(e.g.global,continental,orregional)atlowresolution (e.g.monthlytemperatureandprecipitationdata),withafewwatershed-scaleexceptions(Gondim

Span-ishwatershedusingdynamicallydownscaleddailyclimatedataandtheSoilandWaterAssessment Tool(SWAT).Dependingonthecrop(amongcorn,apples,andalfalfa),irrigationrequirementswere projectedtoincreaseby40–250%bytheendofthe21stcentury,attributeddirectlytodecreasesin growingseasonwateravailability,increasesinevapotranspiration,andchangesincropphenology.In

aBrazilianwatershed,Gondimetal.(2012)determinedthatirrigationwaterdemandwillincreaseby 8–9%bythemid-21stcenturyusingmonthlyclimateprojections.Demandincreaseswereattributed

toprojectedrainfalldecreases(11–18%)andevapotranspirationincreases(6.5–8%).Harmsenetal

cumula-tiveprecipitation,referenceevapotranspiration(ET0),andtheresultingprecipitationdeficitbetween thetwoinPuertoRico.PrecipitationdeficitinFebruaryincreasedupto90mmdependingon emis-sionsscenarioandlocation,whileinSeptemberprecipitationexcessincreasedbyupto277mm.Xiong

loweremissionsscenarios andincreaseby44%and36%in the2020sand 2040s,respectivelyfor higheremissionsscenarios.Thomas(2008)examinedirrigationdemandacrossChinaforthe2030s, projectingbothincreasesanddecreasesindemanddependingontheregion.Forexample,decreasing evapotranspirationinsomelocationsresultindemanddecreasesby100mm,whiledrierregionsin westernChinacanexpectincreasesinirrigationdemand.Globally,Fischeretal.(2007)projectedthat

by2080two-thirdsofirrigationrequirementincreasescanbeattributedtowarmingandchangesin precipitationpatterns,whiletheremainingthirdofincreasesareduetoextendedgrowingseasons

intemperatureandsub-tropicalregions.Overall,globalnetirrigationrequirementsincreasebyover 50%indevelopingregionsand16%indevelopedregions.Schaldachetal.(2012)examinedchanges

inirrigationwaterrequirementsin2050inNorthAfricanandEuropebasedonclimateandlanduse change.Dependingontheregion,climatemodel,andsocioeconomicconditions,demandmaydecrease

byupto30%orincreasebyupto264%,indicatinglargeuncertaintiesinprediction.Forexample,in Zimbabwecornirrigationwaterrequirementswereprojectedtoincreaseby66%bythe2050sand 99%bythe2090s,althoughtherewasconsiderableuncertaintyintheseprojections(Nkomozepiand

2090swereprojectedtobe2.4%and7.9%,respectively(ChungandNkomozepi,2012)

Previousstudiesthatdeterminedtheimpactsofclimatechangeonirrigationdemandexamined

asmallnumberofmodelsandscenariosandthereforecouldnotcapturethewiderangeofclimate modelpredictions.Theproposedstudyisuniqueinthattenglobalclimatemodels(GCMs)combined withfouremissionsscenariosareusedtoconsiderthewiderangeofuncertaintyinclimatemodels andtheirimpactonirrigationdemand.UsingalimitednumberofGCMsresultsinalimitedviewof

plausiblefutureclimate,unreliableprojectionsofthefuture,andconsequentlypoordecision-making andadaptation(Goharietal.,2013).Inaddition,moststudiesconsiderlargespatialscales(e.g.global

orregional)withlowresolution,limitingapplicabilityforproducersinneedofadaptationstrategies

Inthisstudy,theclimatologicaldataandirrigationdemandisexaminedatafinetemporalresolution (daily),whilemoststudiesconsidermonthlyorannualtimescales.Thegoalofthisstudyisto under-standhowirrigationdemandisimpactedbyclimatechangeforcornandsoybean,twocommoncrops grownintheKalamazooRiverWatershed,Michigan.Thespecificobjectivesofthisstudyareto:(1) determinetheimpactsofclimatechangeonkeyelementsinwaterbalance;(2)understandthefuture changeinirrigationdemand;(3)mapthespatialvariabilityinirrigationdemand;and(4)examine possibleadaptationstrategiesbasedonshiftingplantingdate

2 Materials and methods

2.1 Studyarea

TheKalamazooRiverWatershed,hydrologicunitcode(HUC)04050003,wasthesubjectofthis study(Fig.1).LocatedinsouthwestMichigan,thewatershedareaisapproximately4844km2,with theKalamazooRiverdrainingintoLakeMichigan.Elevationrangesfrom176mto386mabovesea level,while50%ofthewatershedareaislessthan270m

Thewatershedisprimarilyagriculturalandforested.About46%ofthelandisdedicatedto agricul-ture,primarilycorn,soybean,andpasture.Deciduousforestandforestedwetlandscover35%ofthe watershed,whileurbanlandoccupies14%oftheland.Theremaininglanduseiscomprisedofvarious vegetablecrops.Soilswithinthewatershedarepredominantlysandy.Sandyloam,loam,sand,and loamysandmakeup32%,24%,11%,and9%,respectively.HydrologicsoilgroupsAandBare domi-nant,covering35%and49%ofthewatershed,respectively.Meanwhile,Dsoilsonlyoccuron3%ofthe watershedarea

2.2 SWATmodeldescription

TheSoil and WaterAssessment Tool(SWAT), developed bytheUnited States Departmentof Agriculture-AgriculturalResearchService(USDA-ARS,)simulatestheimpactofvaryingtopography, soils,landuse,andmanagementpracticesonhydrology,waterquality,andoverlongtimeperiods

thatiscapableofsimulationonadailytime-step.SWATdelineatesawatershedintosubwatersheds basedontopographycharacteristics.Subwatershedsarefurtherdiscretizedintohydrologicresponse units(HRUs)basedonhomogeneouslanduse,soiltype,andslopecharacteristics.Calculationsare gen-erallycompetedattheHRUlevelandaggregatedtothesubwatershedandwatershedscales.There aremultiplecomponentssimulatedbytheSWATmodel,includingweather,hydrology,soilerosion andsedimenttransport,nutrientcyclingandtransport,plantgrowth,andlandmanagementpractices

OverlandhydrologyinSWATisbasedonthewaterbalance(Neitschetal.,2011).Surfacerunoffis calculatedusingtheSCScurvenumberequation(Mockus,1972),whichisbasedonrainfall,surface storage,interception,infiltrationpriortorunoff,andaretentionparameterbasedonsoils,landuse management,slope,andsoilwatercontent.Multiplepathwaysofwaterinthesoilaresimulated, includingplantuptake,evaporation,percolationintoshallowanddeepaquifers,andlateralflowfor streamflowcontribution.Maincomponentsofthelandphaseofthehydrologiccyclearechangein soilwatercontent,precipitation,surfacerunoff,evapotranspiration,percolationoutoftherootzone, andreturnflowfromgroundwaterintotherootzone

PotentialevapotranspirationissimulatedinSWATusingthePenman–Monteithmethod(Monteith,

1965),whichaccountsforenergynecessarytosustainevaporation,surfaceresistance,aerodynamic resistance,andstrengthofwatervaporremovalmechanisms.Actualevapotranspirationiscalculated

byaccountingforevaporationofrainfallinterceptedbytheplantcanopy,potentialtranspirationbased

onplantgrowthunderidealconditions,andsoilwaterevaporation

xxx.e4 S.A Woznicki et al / Journal of Hydrology: Regional Studies xxx (2015) xxx.e1–xxx.e18

Fig 1. Kalamazoo River Watershed; (a) land use, (b) weather station and streamflow gauging station locations.

CropgrowthinSWATissimulatedbyaccumulationofheatunits(differencebetweendailymean temperatureandbasetemperatureforcropgrowth)throughoutthegrowingseason(Neitschetal.,

2011).Maturityisreachedwhenthenumberofaccumulatedheatunitsequalspotentialheatunits ActualPotentialgrowthissimulatedusingleafareadevelopment,lightinterception,andbiomass productionassumingspecies-specificradiationuseefficiency.Canopycover,height,androot devel-opmentarealsosimulatedbasedonheatunits.Water,nitrogen,andphosphorusuptakebytheplant

aresimulatedbasedonsoilwatercontent,fractionofnitrogeninplantbiomass,andfractionof phos-phorusinplantbiomass,respectively.SWATsimulatescropgrowthconstraintssuchaswaterstress, temperaturestressandnutrientstress(Neitschetal.,2011).Waterstressisbasedontherelationship betweenactualandpotentialplanttranspiration.Temperaturestressismodeledexponentiallywith dailyaverageairtemperatureandoptimaltemperaturesforplantgrowth,wheredivergencefrom optimalconditionsresultsingrowth-reducingstress

IrrigationapplicationinSWATcanbemanuallyscheduledorautomaticallyappliedbasedonsoil waterdeficitbelowfieldcapacity(Neitschetal.,2011).Watercanbewithdrawnfromareach,reservoir, shallowaquifer,deepaquifer,orsourceoutsidethewatershed.Limitscanbeplacedonwithdrawal basedonminimumstreamflow,fractionoftotalstreamflowthatcanbewithdrawn,oramaximum irrigationamountthatcannotbeexceededinoneday.Otherparametersforconsiderationarean irrigationefficiencyfactor(accountingforconveyanceandevaporativelosses)andasurfacerunoff ratio(fractionofirrigatedwaterthatisconvertedtorunoff)

2.3 Datacollection

2.3.1 Physiographicdata

SWATmodelsetupandparameterizationrequirestheuseofmultiplespatialdatasets.Topography datawasobtainedfromtheUnitedStatesGeologicalSurveyNationalElevationDatasetintheformofa

10mresolutiondigitalelevationmodel,whichwasusedinwatersheddelineation.The2011Cropland DataLayer(30mresolution)developedbytheUnitedStatesDepartmentofAgriculture-National Agri-culturalStatisticsService(USDA-NASS)wasselectedforlanduse/landcoverrepresentation.Spatial andtabularsoildatawasobtainedfromtheUSDANaturalResourcesConservationServiceSoilSurvey GeographicDatabaseata1:24,000resolution.Commonagriculturalpracticesandrotationswithin thewatershedwereimplementedforcorn,soybean,andwinterwheatlanduses

2.3.2 Climatedata

Historicalclimatedatafrom1980to1999wasobtainedfromtheNationalClimaticDataCenter Dailyprecipitation,maximumtemperature,andminimumtemperaturevariableswereavailablefor nineprecipitationandtemperaturegaugeswithinthewatershed

ProjectedfutureclimatescenarioswereobtainedfromtheUnitedStatesGeologicalSurvey(USGS) GeoDataPortal(http://cida.usgs.gov/climate/gdp/).Thescenariosareasuiteofbias-corrected statis-ticallydownscaleddailyprecipitation,maximumtemperature,andminimumtemperaturedatawith

a1/8◦spatialresolution,availablefrom1960to2100(Hayhoeetal.,2013).Basedonthisresolution,

34gridcellswereusedinSWAT.Datafromtencoupledatmosphere–oceangeneralcirculation mod-els(AOGCMs)drivenbyfourIntergovernmentalPanelonClimateChange(IPCC)SpecialReporton EmissionsScenarios(SRES)storylines(Nakicenovicetal.,2000)wereused.AOGCMsincludedwere thefollowing:ParallelClimateModel(PCM),CommunityClimateSystemModelversion3(CCSM3), GeophysicalFluidDynamicsLaboratory(GFDL)2.0,GFDL2.1,HadleyCentreCoupledModelversion

3(HADCM3),BergenClimateModelversion2(BCM2),CoupledGlobalClimateModelT47 (CGCM3-T47),CGCM3-T63,CentreNationaldeRecherchesMeteorologiques(CNRM),ECHAM5(developedby theMaxPlanckInstitute),andECHO(developedbyMeteorologicalInstituteoftheUniversityofBonn (Germany),InstituteofKMA(Korea),andModelandDataGroup)

EachmodelwasforcedwithatleasttwoSRESstorylinesdevelopedusingprojectionsofpopulation change,demographics,technology,andsocio-economicfactorstoestimategreenhousegasemissions

higher(A1FI),mid-high(A2),mid(A1B),andlower(B1).Usingmultipleclimateprojectionsand sto-rylinesallowsforconsiderationofvariousparametric,structural,andforcinguncertainties(Knutti

etal.,2010).Theresultingnumberofscenarios(combinationofGCMsandSRESstorylines)was35 AtmosphericCO2concentrationswereobtainedfromtheBERNcarboncyclemodel,whichvarybased

onyearandSRESstoryline(IPCC,2001).CO2concentrationsforSRESstorylinesandtimeslicesused arepresentedinTable1

Twofuturetimesliceswereselectedforanalysis:2020–2039and2060–2079.SWATmodel pro-jectionsofirrigationdemandforthesetimesliceswerecomparedtothe1980–1999modelcontrol

xxx.e6 S.A Woznicki et al / Journal of Hydrology: Regional Studies xxx (2015) xxx.e1–xxx.e18

Table 1

Atmospheric CO 2 concentrations.

(ppm)

A1FI (ppm)

A2 (ppm)

B1 (ppm)

Adapted from IPCC (2001).

period.ByanalyzingthechangesbetweenthecontrolandfuturetimeslicesacrossallmodelsandSRES storylines(herebyreferredtoasa“scenario”),anensembleofprojectedimpactsofclimatechange

onirrigationdemandisdeveloped.Monthlyprecipitationandtemperatureforthecontrolandfuture timeslicesarepresentedinFigs.2and3,respectively

2.4 Calibrationandvalidation

SWATmodelcalibrationandvalidationofstreamflowandcropyieldwereperformedtoensure thatthemodelreplicatesobservedphysicalbehaviorbeforeextendingthestudytounderstandthe impactsofchangeonwaterquantityandcropgrowthinthewatershed.Dailystreamflowcalibration andvalidationoccurredatthreeUSGSgaugingstationsinthewatershed:04105000(BattleCreekat BattleCreek,MI),04105500(KalamazooRivernearBattleCreek,MI),and04106000(KalamazooRiver

atComstock,MI).TheUSGSstationno.04106000wascalibratedfrom1985to1992andvalidatedfrom

1993to1999.Theremainingstations(04105000and04105500)werecalibratedfrom1980to1989 andvalidatedfrom1990to1999.Calibrationandvalidationwascompletedusingobservedclimate datafromnineweatherstationsdisplayedinFig.1).Theperiodsofcalibrationandvalidationwere selectedbasedonthetimeperiodofavailabledata

Threestatisticalcriteriawereusedtoevaluategoodness-of-fitbetweenthesimulatedandobserved streamflowdata:Nash–Sutcliffeefficiency(NSE),rootmeansquareerror-observationsstandard devi-ationratio(RSR),andpercentbias(PBIAS).Rangingfromnegativeinfinitytoone(withavalueofone indicatingperfectfit),NSEdeterminestherelativemagnitudeoftheresidualvariancecomparedto theobserveddatavariance(Moriasietal.,2007).RSRstandardizesrootmeansquareerrorusingthe

Fig 2. Kalamazoo River Watershed monthly total precipitation for model control period (1980–1999) versus (a) 2020–2039 and (b) 2060–2079 across all climate scenarios.

Data from Hayhoe et al (2013).

Fig 3.Kalamazoo River Watershed mean daily maximum and minimum temperature for model control period (1980–1999) versus (a) 2020–2039 and (b) 2060–2079 across all climate scenarios).

Data from Hayhoe et al (2013).

standarddeviationofobservations,withvaluesrangingfromzerotoalargepositivenumber(where

alowernumberindicatesbettermodelperformance).PBIASmeasuresthetendencyofthesimulated datatoover-predictorunder-predictcomparedtotheobserveddata,wheretheoptimalvalueiszero andsmallpositiveandnegativepercentagesareacceptable

Basedonstreamflowgoodness-of-fitrecommendationsbyMoriasietal.(2007)foramonthly time-step,asatisfactoryNSEisgreaterthan0.50,satisfactoryRSRislessthanorequalto0.70,andsatisfactory PBIASliesbetween−25%and+25%

Annual crop yield calibration was performed for the two major crops in the Kalama-zoo River Watershed: corn and soybean Observed crop yields from the control period (1980–1999)were obtainedat thecounty level from theNational AgriculturalStatistics Service

Water-shedwascalculatedbasedontheportionofeachcountyinwhichthewatershedresides.County-level cropyieldsarereportedbyNASSinbushels/ac,whileSWATestimatescropyieldinmetrictons/ha with20%moisturecontentatharvest(Srinivasanetal.,2010).PBIASwasusedastheevaluationcriteria followingapreviousstudyoncropyieldcalibrationbySrinivasanetal.(2010)

2.5 Irrigationdemand

Determinationofirrigationdemandbyagriculturalcropsisbasedonthesoil–waterbalance(NEH,

1993),presentedinEq.(1)

WhereIRRisirrigationdemand(mm),Pisprecipitation(mm),Dpisdeeppercolationfromthecrop rootzone(mm),ROissurfacerunoff(mm),GWisgroundwatercontributiontothecroprootzone (mm),andSWisthechangeinsoilwaterinthecroprootzone(mm).OndaysinwhichIRRwas negative(usuallyduetoaprecipitationevent)thenumberwasforcedtozero

2.6 Adaptationstrategies

Plantingdateswereshiftedinordertoexaminepossibleclimatechangeadaptation.Forthecontrol scenario,thetotalirrigationdemandbycornandsoybeanwascalculatedforthegrowingseason

xxx.e8 S.A Woznicki et al / Journal of Hydrology: Regional Studies xxx (2015) xxx.e1–xxx.e18

Table 2

Daily streamflow calibration and validation results.

a http://waterdata.usgs.gov/usa/nwis/uv?04105000.

b http://waterdata.usgs.gov/usa/nwis/uv?04105500.

c http://waterdata.usgs.gov/usa/nwis/uv?04106000.

d Nash–Sutcliffe efficiency.

e Percent bias.

f Root mean square error-observations standard deviation ratio.

(May–September).Meanwhile,plantingdatewasshiftedby±10,±20,and±30daysasanadaptation

toprojectedfutureclimates.Thisallowsforbeneficialchangesincropwateruseorharvestyield

byavoidingtemperaturestressduringkeycropdevelopmenttimesorutilizinggreaterprecipitation volumes.Finally,thetotalirrigationdemandandyieldswerecomparedagainstthecontrolscenarios forcornandsoybean

3 Results and discussion

3.1 Calibrationandvalidation

3.1.1 Streamflowcalibration

Dailystreamflowcalibrationandvalidationweresatisfactoryforalllocationsaccordingto guide-linesdevelopedbyMoriasietal.(2007).Goodness-of-fitresultsfor alllocationsarepresentedin

station(04105000)ispresentedinFig.4

Fig 4. Observed vs SWAT simulated streamflow for USGS station 04105000.

Fig 5.Observed vs SWAT simulated (a) corn yield and (b) soybean yield.

3.1.2 Cropyieldcalibration

Atime-seriescomparisonofobservedandsimulatedannualcornandsoybeanyieldsfortheentire KalamazooRiverWatershedarepresentedinFig.5aandb,respectively.Thearea-weightedaverage PBIASvaluesforcornandsoybeanatthewatershedlevelwere1.9%and9.4%,respectively.Lowpositive PBIASvaluesindicatethattheSWATmodelslightlyunder-predictedyields,particularlyforsoybean, thoughthevaluesarewithinsatisfactorylimits.Poorcalibrationinsomeyearsforsoybean(e.g.1992)

isduetothewiderangeofyieldsobservedatthecountylevel,whiletheKalamazooRiverWatershed liesintencounties.Thismakesit difficulttoachieveanaccuratesimulatedvaluewhenmultiple locationsinthewatershedhaveverydifferentyields

3.2 Keyelementsinwaterbalance

Thedynamicsofthewaterbalanceareprojectedtobealteredduetoclimatechange,although someuncertaintyisapparentregardingthemagnitudeofthesechanges,butthetrendsamongwater quantityvariablesarerelativelyconsistent.Fisher’sleastsignificantdifference(LSD)wasperformedto determinesignificantdifferencesbetweenwaterbalancevariablesacrosstimeslices.Averageannual waterquantityvariablesunderallscenariosfortheKalamazooRiverWatershedarepresentedinFig.6 Averageannualprecipitationisgenerallyprojectedtoincreasemovingfromthecontrolperiodto 2060–2079,asindicatedbyincreasesinthescenariomeans.However,someclimatescenariosproject annualprecipitationdecreasesupto60mm.Almost80%ofthescenariospredictannualprecipitation increasesby2020–2039,whileby2060–207990%ofscenariospredictincreases

Almostallvariablesresultedinstatisticallysignificantchangesmovingfromthecontrolperiod

to2020–2039or2060–2079.Percolationtrendsarecorrelatedwithprecipitation,asmostscenarios projectionsresultinpercolationincreasesfollowinggreaterprecipitationvolumes.Finally,decreases

inspringsnowmeltoccurinalmostallscenariosbecausemorewinterprecipitationfallsintheform

ofrainduetowarmertemperatures.Changesinsurfacerunoffwerestatisticallyinsignificant,with

anequalnumberofscenariosprojectingslightincreasesordecreases.Thesechangesdependonthe magnitudeoftheprecipitationchangeandresultingET,soilwaterholdingcapacity,andpercolation

xxx.e10 S.A Woznicki et al / Journal of Hydrology: Regional Studies xxx (2015) xxx.e1–xxx.e18

Fig 6.Watershed-wide water balance, where bars represent means and error bars represent maximum and minimum among all climate scenarios.

AverageannualETdecreasesslightlyfor80%ofscenariosin2020–2039and2060–2079,likely becauseatmosphericCO2concentrationsincrease.IncreasingCO2levelscausepartialclosureofplant stomatabecausetheydonotneedtoopenaswidelytoobtainnecessaryCO2 forphotosynthesis

3.3 Irrigationdemand

Cornandsoybeanirrigationdemandwereprojectedtochangeinresponsetoalteredfutureclimate, withchangesvaryingdependingonthemonth.Asobservedwiththewaterbalance,monthlydemand becomesmoreuncertainmovingfurtherintothefuturewithawiderrangeofpossibilitiesdepending

ontheclimatemodelandSRESstoryline.Monthlyarea-weightedirrigationdemandfor cornand soybeanacrossalltimeslicesarepresentedinFig.7aandb,respectively

CornirrigationdemandisgreatestinJuneandJuly,withlowestdemandsoccurringinMayand September(Fig.7a).AveragedemandacrossallscenariosandscenariosincreasesinMay,June,and September,whiledecreasesintheaveragesoccurforJulyandAugust.Thesechangesaremore pro-nouncedin2060–2079thanin2020–2039,whichreflectsthegreatertemperatureincreasesandmore pronouncedprecipitationchangesprojectedinthelate21stcentury(Figs.3and4).Peakirrigation demandshiftsfromoccurringgenerallyinJulyforthe1980–1999controltimeslicetomore com-monlyoccurringinJune.Thisshiftismoreapparentmovingfurtherintothefuture:demandisslightly greaterinJunefor2020–2039,butJulyaveragedemandfor2060–2079isabout20mmlessthanin June.GreateruncertaintyfurtherintothefutureisapparentforJune,July,andAugust,withagreater rangebetweenthefirstandthirdquartiles(boxes)andwhiskers/outliers.Theincreaseinuncertainty

isreflectedinthecoefficientofvariation (CV)acrossallscenarios foreachmonth/timeslice For example,CVinJulyincreasesfrom4%(1980–1999)to12%(2060–2079),whileAugustCVsmove from7%to21%inthesametimeperiod.Thesechangesindicatetheclimatescenarios’uncertainty

inprojectingfutureprecipitation,particularlyinJulyandAugust.Meanwhile,thefutureuncertainty increasesaresmallerinMay(CVfrom5%to11%)andSeptember(CVfrom6%to11%)movingfrom 1980–1999to2060–2079

Averageseasonaltotalsforcornirrigationdemandandcropyieldacrossallscenariosandtimeslices arepresentedinTable3.Changesinirrigationdemand,whethertheyareincreasesordecreases,vary

bymodelandSRESstoryline.By2020–2039,thewatershed-wideaveragechangeincornirrigation demandovertheentiregrowingseasonisbetween+20mmand−30mm.Movingto2060–2079,this changeisbetween+19mmand−41mmfrom1980to1999.Mostscenariosgenerallyprojectdecreases

intotalseasonalirrigationdemand(57%and71%predictdecreasesby2020–2039and2060–2079, respectively).Consequently,97%ofscenariospredictyielddecreasesby2020–2039,whileallofthem