GroundwaterinfluencesonsoilmoistureandsurfaceevaporationqXiChen1,QiHu*
ClimateandBio-AtmosphericSciencesGroup,SchoolofNaturalResourceSciences,UniversityofNebraska-Lincoln,
237L.W.ChaseHall,Lincoln,NE68583-0728,USA
Received11September2003;revised6April2004;accepted16April2004
Abstract
Soilhydrologicalprocessesplayanimportantroleinland-atmospheresystem.Inmostclimatemodels,theseprocessesaredescribedbysoilmoisturevariationsinthefirst2mofsoilresultingfromprecipitation,evaporation,andtranspiration.Groundwatereffectsonsoilmoisturevariationsandsurfaceevaporationareeitherneglectedornotexplicitlytreated.Althoughgroundwatermayhaveasmalleffectonsoilmoistureinareaswithadeepgroundwatertable,groundwatercanactasasoilwatersourceandhavesubstantialeffectsinareaswherethewatertableisnearorwithinamodel’ssoilcolumn.Howgroundwateraffectssoilmoisture,itsverticaldistribution,aswellasthesurfacewaterfluxaretheissuesaddressedinthisstudy.Asoilhydrologicalmodelwasdevelopedtoincludegroundwatereffectsbyallowingwaterexchangebetweentheunsaturatedzoneandgroundwater.Themodelusesaverticallyvaryingsaturationhydraulicconductivity,andisevaluatedusingobservationsatonestationintheNebraskaSandHills.Modelresultsshowitsabilitytodescribetherolesofgroundwaterinmaintainingtheobservedsoilmoisture,especiallyindeeplayers.Inaddition,comparisonsshowthatthesoilmoisturecontentinthefirstmeterofthesoilcolumnfromthemodelwithgroundwateris21%greaterthanthatfromamodelwithoutgroundwater.Highsoilmoisturecontentintherootzoneresultsinincreasedevapotranspiration(ET).TheaverageETinthreeperiodsfrom1998to2000is7–21%higherwhengroundwaterisconsideredinthemodel.Becauseofthegroundwatereffects,spatialvariationsinthegroundwatertablecancreateanadditionalspatialvariabilityofsoilmoistureandsurfacewaterflux.Thisadditionalvariabilitycouldbeimportantindevelopmentofstormsinregionswhosedomainhasalargeportionwithhighgroundwatertable.q2004ElsevierB.V.Allrightsreserved.
Keywords:Groundwater;Soilmoisture;Surfaceevaporation;Model;NebraskaSandHills
1.Introduction
Temporalandspatialvariationsofsoilmoisturearereceivingincreasedattentioninclimatestudies
AgriculturalResearchDivision,UniversityofNebraska-LincolnContributionNumber15988.
*Correspondingauthor.Tel.:þ1-402-472-6642;fax:þ1-402-472-6614.
E-mailaddress:qhu2@unl.edu(Q.Hu).1 Presentaddress:LaboratoryofWaterResourcesDevelopment,- 1 -HehaiUniversity,Nanjing,China.
0022-1694/$-seefrontmatterq2004ElsevierB.V.Allrightsreserved.doi:10.1016/j.jhydrol.2004.04.019
qbecausesoilmoistureisanessentialelementinprocessesthatdrivelandsurfacewaterandenergyfluxes,whichaffectecosystemdynamicsandbiogeo-chemicalcyclesintheland-atmospheresystem.Soilmoistureintheunsaturatedzonechangesasaresultofprecipitationrechargeandwaterexchangewithboththeatmosphereandgroundwater.Moststudiesofwaterexchangebetweentheunsaturatedzoneandtheatmospherehavefocusedonunderstandingsoilmoisturevariationsandtheireffectsonatmos-phericboundarylayerprocessesaffectingweather
http://www.paper.edu.cn
286X.Chen,Q.Hu/JournalofHydrology297(2004)285–300
andclimate(e.g.Manabe,1969;Ookouchietal.,1984;Dickinson,1984;PielkeandAvissar,1990;Pielkeetal.,1991).Inthosestudies,groundwatereffectsonsoilmoisturevariationhavebeenneglected,however.Inthemeantime,severallandsurfaceandsoilhydrologicalmodels(LSM)havebeendevelopedandusedinmesoscaleatmosphericmodelsandgeneralcirculationmodels(GCM).SomeofthoseLSMhavesophisticatedverticalexchangeprocessesofmoistureaswellastemperaturebetweenthesoil,thebiosphere,andtheatmosphere(Dickinson,1984;Dickinsonetal.,1986;Sellersetal.,1986;PanandMahrt,1987;Xueetal.,1991;ChenandDudhia,2001),whileothersusesimplifiedrepresentationsofverticalexchangephy-sicsthatincorporatetheeffectsofspatialheterogen-eityintopography,soil,andvegetationonsoilmoisturevariationandrelatedhydrologicalprocesses(e.g.EntekhabiandEagleson,1989;FamigliettiandWood,1990,1994;Woodetal.,1992;Schaakeetal.,1996;Stieglitzetal.,1997;Kosteretal.,2000).InmostoftheLSM,thesoilcolumnisdividedintonumberoflayerstodescribe,invariousdegreesandindifferentways,verticalsoilmoistureexchangeprocesses,suchasinfiltrationandpercolation,surfaceandundergroundrunoff,andeffectoftherootdensityprofileonsoilmoisture(Acs,1994).Theselayerscomprisearootzone,usually1mthick,andadeepsoilzonebeneathit.Thethicknessofthelayerbeneaththerootzonevaries;itis1mintheOSUmodel(PanandMahrt,1987)andtheLSMusedinthePennState-NCARMM5(ChenandDudhia,2001).Itvariesbetween1and2mintheSSiBmodel(Xueetal.,1991),andisfixedat10mintheBATSscheme(Dickinsonetal.,1986).ManyoftheseLSM,particularlytheonesbasedontheTOPMODEL(BevenandKirkby,1979),arefocusedontheeffectsofspatialheterogeneityofsoilmoistureonthesurfacehydrologyandtheatmosphere(e.g.EntekhabiandEagleson,1989;FamigliettiandWood,1994;Stieglitzetal.,1997;Kosteretal.,2000).Thesemodelsemployverticallayersthatextendtothedepthofthegroundwatertableanddo,infact,accountfortheimpactofdistributionofthegroundwatertableonspatialheterogeneityofsoilmoistureintheupperlayers.However,fewofthesemodelsexplicitlyaccountfortheeffectsofgroundwateronsoil moistureandsurfaceevaporation.OtherLSM,whichextendovergreaterverticaldepths,treat- 1 -thedeepestsoillayeronlybypermittinggravity
drainageacrossthelayer’slowerboundary.Theydonotpermitgroundwaterinputtothedeepestlayerofthemodelsoilcolumn.
Inmanyshallowgroundwaterregions,suchaswetlandsandlowlandsinrivervalleys,ahighgroundwatertableandsignificanthydraulicgradientsbetweenthesaturatedzoneandtherootzoneleadtocontinuoussupplyofgroundwatertotherootzone.Inthoseregions,theroleofgroundwaterinvariationsoftherootzonesoilmoisturebecomesessential.OneareawheregroundwaterservesasamajorsourceofsoilmoistureistheSandHillsinwest-centralNebraska.TheSandHills(,50,000km2)haveathicksurfacelayeroffinesand.Throughthissandylayer,soilwaterpercolatestothegroundwaterataspeedasfastas3.4mday21(Bleed,1998),leavinglittletimeforevaporationtoconsumethesoilwater.Thisuniquegeologicalsettinghelpscreatealargegroundwaterreserveinthearea:theOgallalaaquifer.IntheSandHills,groundwatereffectsdeterminethevariationofsoilmoistureaswellassurfaceevapor-ationandstreamflow(Bleed,1998;Bentall,1998).Indeed,areaswithsuchsignificantgroundwatereffectsonthesoilmoistureandsurfacewatercomprisearelativelysmallfractionoftheentirelandsurface.Yetthegroundwatertabledistributioninthoseareascreatesanadditionalspatialheterogeneity,similartothatcreatedbyvariationsintopography,surfacevegetation,andsoilproperties,andarecriticallyimportantforregionalprocessesinfluencingspatialvariationsofsoilmoisture,evapotranspiration(ET),andprecipitationandfloods(Woodetal.,1992).Thelackofunderstandingoftheeffectofgroundwateronsoilmoisturepromptedthisstudy.Inthispaper,weevaluateanddescribetheinfluenceofgroundwateronrootzonesoilmoistureusingbothobservationaldataandmodelinganalysis.Asoilhydrologicalmodelthatincludesthegroundwatereffectonsoilmoistureisdescribedinthenextsection.Inadditiontoincludinggroundwater,thismodelusesaverticallyvaryingsaturatedsoilhydraulicconduc-tivity,assuggestedinBeven(1984)andElsenbeeretal.(1992),toconsiderthedecreaseofsoilpermeabilitywithdepth.InSection3,effectsongroundwaterintwoextremelydryandwetweatherscenariosarepresented.InSection3.2.1,themodelisvalidatedusingobservations,includinggroundwater
http://www.paper.edu.cn
X.Chen,Q.Hu/JournalofHydrology297(2004)285–300287
tablevariation,atGudmundsenintheNebraskaSandHills,andisthenusedtodescribesoilmoisturevariationsandhowtheyareaffectedbygroundwater.Themodeldataarefurtherusedtoanalyzetheeffectsofgroundwateronlocalwatercycle.GroundwatereffectsatasmallregionalscaleareexaminedinSection4.Resultsshowthatspatialvariationsofthegroundwatertableandgroundwatereffectsonsoilmoisturecanmoreaccuratelydescribethetotalsurfacemoisturefluxinthearea,suggestingtheimportanceofincludinggroundwaterinLSMtoimprovedescriptionofregionalwatercycles.AsummaryisgiveninSection5.
thewatereventuallyreachesthesaturatedzonetorechargethegroundwater.Concurrentlywiththeserechargingprocesses,evaporationandtranspirationaretakingplaceatthecanopyandthegroundsurface,resultinginsoilwaterloss.
SoilmoisturevariationinthemodelisdescribedbytheRichard’sequation:
›u››u›K¼DþþFðt;uÞ;›t›z›z›z
ð1Þ
2.Asoilhydrologicalmodel
Thestructureofthesoilhydrologicalmodelused
inthisstudyisshowninFig.1.Themodelhasasurfacelayerofvegetationcanopyandfoursoillayers.Thethicknessesofthesoillayersfromshallowtodeepsoilwerespecifiedas0.1,0.15,0.25,and0.5m.Thesoilcolumncontainsarootzonewhosemoisturevariationisinfluencedbygroundwater.Thesourcesofsoilwaterareprecipitationandground-water.Invegetatedareas,aportionoftheprecipi-tationisinterceptedbythecanopy,andtherestfallstothegroundandinfiltratesintothesoilandfurtherpercolatesdowntodeepersoillayers.Someof
whereuissoilmoisturecontentinm3m23,tistime,ztheverticalcoordinate,Fðt;uÞthesourceandsinktermaccountingforprecipitation,evaporation,andsurfacerunoff,Ktheunsaturatedhydraulicconduc-tivity,andDthesoilwaterdiffusivity.BothKandDarefunctionsofuandarecomputedfromKðuÞ¼Ksðu=usÞ2bþ3andDðuÞ¼KðuÞð›C=›uÞ;whereCissoilwatertensionfunctionandCðuÞ¼Cs=ðu=usÞbinwhichbisacurve-fittingparameter.Intheseexpressions,Ks;Cs;andbarefunctionsofsoiltypesfollowingCosbyetal.(1984).BothKandDarehighlynon-linearfunctionsofsoilmoisture.
IntegratingEq.(1)throughthesoillayersundertheassumptionofverticallyhomogeneoussoilhydraulicpropertieswithineachlayeryields
›u1›u¼2D2K1þPd2R2Edir2ET1;d1›t›z1
ð2Þ
Fig.1.Aschematicofthemultiplayersoilhydrologicalmodel.
- 1 -
http://www.paper.edu.cn
288X.Chen,Q.Hu/JournalofHydrology297(2004)285–300
d›u22›t¼D›u›z2D›u
þK2K1›z21
22ET2;ð3Þd›u33›t¼D›u›z2D›u
þK2K32E2›z32T3;ð4Þ
and
d›u44›t¼D›u›zþK2K33
4:ð5Þ
Intheabove,thesubscript,i¼1;2,3,and4,isthesoillayerindex(seeFig.1),dithethicknessofithsoillayer,Pdtheprecipitationfallingontheground,Rthesurfacerunoff,Kitheverticalunsaturatedsoilhydraulicconductivity.InEq.(5)K4isthegravita-tionalpercolationorsubsurfacerunoff,andinEq.(2),Ediristheevaporationfromthetopsoilsurface,andETiinEqs.(2)–(4)arethetranspirationbyvegetationthroughroots.Asimplelinearmethod(MahfoufandNoilhan,1991)Edir¼ð12sfÞbEp
isusedtocalculateEdir;whereb¼ðu12uwÞ=ðuref2uwÞ;inwhichurefanduwarethefieldcapacityandwiltingpoint,respectively.Ontherightsideoftheaboveequation,EpisthepotentialevaporationcalculatedbyaPenman-basedenergybalanceapproachthatincludesastability-dependentaerodynamicresistance(MahrtandEk,1984),andsfisthefractionofsurfacevegetationcover.TheevaporationinEqs.(2)–(4)iscalculatedfrom
ET¼sfEpBcb12ðWc=SÞ0:5c
whereBcisafunctionofcanopyresistance,Wctheinterceptedcanopywaterandcalculatedaccordingtothebudgetforinterceptedcanopywater,andSthemaximumcanopycapacity.Thissystemiscurrentlyusedinthesoil-hydrologymoduleofMM5.ForadditionaldetailsconcerningeachterminEqs.(2)–(5),thereaderisreferredtoChenetal.(1996)andChenandDudhia(2001).
ToobtainEq.(5),animportantassumptionwasmade,i.e.thehydraulicdiffusivityinthelayerbeneaththefourthlayerofthemodelissettozero,orequivalently,thehydraulicgradientbetweenthemodel’sdeepestlayerandthegroundwatertableis negligible.Asaconsequence,thesoilwaterfluxacrossthelowerboundaryofthemodel’ssoilcolumn
- 1 -isonlythegravitationalpercolationorsubsurface
runoff;nowaterfluxintothecolumnisallowed.Thisassumptionmaybevalidforareaswheregroundwatertableisdeepandfarfromthelowerboundaryofthesoilcolumn.However,inareasofshallowground-water,thewatertablemaybehighenoughtocreateasubstantialhydraulicgradientbetweenthedeepestsoillayerandthewatertable,or,evenextendintothemodelsoillayers.Inthoseareas,Eq.(5)becomesinvalid.
Soilmoisturevariationsinshallowgroundwaterareasbehaveverydifferentlyfromthoseinareaswithdeepgroundwatertable.ThesedifferencescanbeseeninFig.2showingcomparisonsofobservedsoilmoistureatGudmundsenandAinsworthintheSandHills.GudmundsenisatthecenteroftheSandHills(Fig.3)andhasanaveragegroundwatertableat1.5mduring1989–1992and1.22mduring1999–2000.AinsworthisattheperipheryoftheSandHillsandhasanaveragegroundwatertableabout9mbelowthesurface.Inaddition,becauseAinsworth’suppersoillayershavehighclaycontent(Gudmundsen’shavemostlyfinesand),watercontentintheupperlayersatAinsworthisoftenhigherthanthatatGudmundsen,particularlyduringthewetperiodfromAprilthroughJune.Atthedeeperlayers,however,thesoilmoisturecontentismuchsmalleratAinsworththanatGudmundsen.ThehighmoisturecontentindeepsoillayersinGudmundsenislargelyattributedtotheinfluenceofgroundwater.SuchinfluenceistrivialatAinsworthbecauseofitsdeepgroundwatertable.Lackingagroundwatersource,thesoilmoisturecontentinthedeeplayersissmallerthanintheupperlayersinthewetseasonandlargerindryperiods(Fig.2a).Incontrast,atGudmundsen,soilmoistureindeeplayersisaffectedbygroundwater,andtheobservedsoilmoisturecontentinthoselayersisalwayslargerthanintheshallowlayers(Fig.2b).Thelargewatercontentinthedeepsoillayersmaintainsanupwardverticalsoilmoisturegradientinsoilandcontributestosoilmoistureatshallowlayersandtoevaporation.Clearly,thevariationinsoilmoistureatGudmundsencanbedescribedaccuratelyonlywhenthegroundwatereffectisconsidered.
Amethodtoincludethegroundwatereffectonsoilmoistureinshallowlayersandrootzoneistouseanon-zerohydraulicdiffusivitybetweenthedeepestmodelsoillayerandthegroundwatertable.
http://www.paper.edu.cn
X.Chen,Q.Hu/JournalofHydrology297(2004)285–300289
ThenEq.(5)becomes
›u4›u›ud4¼D2DþK2K4:›t›z3›z43
ð6Þ
ThetermDð›u=›zÞ4isdeterminedusingthesoilmoisturedifferencebetweenthesaturatedzone,whichcouldextendintodeepmodellayers,andthenext
unsaturatedsoillayerandthedistancebetweengroundwatertableandthemid-pointoftheaffectedlayer,Zg(seeFig.1).Eq.(6)alsoallowsthegroundwatertabletovarywithtime.
Inadditiontoincludingthegroundwatereffect,wealsouseaverticallyvaryingsaturationsoilhydraulicconductivity,Ks:Beven(1984)and
Fig.3.TheNebraskaSandHillsandthelocationsofGudmundsenandAinsworthstations.AreaAisusedintheregionalanalysisdescribedin - 1 -Section4.
http://www.paper.edu.cn
290X.Chen,Q.Hu/JournalofHydrology297(2004)285–300
Elsenbeeretal.(1992)showedthatinthenaturalenvironmentwiththepresenceofgroundwater,hydraulicconductivitydecreasesexponentiallywithdepth:KsðzÞ¼K0e2fz:
ð7Þ
InEq.(7),K0isthehydraulicconductivityatthesurfaceandfisthee-foldingdepth.ThisvaryingKsalsohasbeenusedinmostTOPMODEL-basedsoil-hydrologymodels(e.g.FamigliettiandWood,1994;Stieglitzetal.,1997;ChenandKumar,2002).WithEq.(7),thenewmodel,consistingofEqs.(1)–(4)and(6),describesexchangesofgroundwaterinthesaturatedzoneandsoilmoistureoftheunsaturatedlayersintherootzone.
3.Groundwatereffectonsoilmoisture3.1.Sensitivityanalysis
Toelucidategroundwatereffectonsoilmoisture,weintegratedthenewmodelinanofflinesetting(uncoupledtotheatmosphericmodel).Specifically,soilmoisturevariationwasdeterminedforloamysandsoilwithdifferentgroundwatertabledepthsintwohypotheticalextremeweathercases.Thefirstcaseconsistsof3dayswithcontinuousrainataconstantrateof1.27mmh21andnoET.ThesecondcasehasthreeconsecutivedrydayswithtotaldailyETof6.78mm,whichisthemeanAprildailyevaporationatGudmundsen(calculatedusingthe1998–2000dataofanautomatedweatherstationatGudmundsen).ThediurnalvariationofETisfrom0.0mmh21at00:00localtimeto0.44mmh21at14:00localtime.
Table1
SoilandvegetationrelatedmodelparametersSoil
LoamysandVegetationGrass
Soilmoistureinthetop1moftheunsaturatedzonewasdeterminedfromintegrationofthemodeloverfourlayerswithdepthsof0.1,0.15,0.25,and0.5mfromtoptobottom.Theinitialsoilmoistureinthesefourlayerswassettobe0.022,0.06,0.182,and0.399m3m23,respectively,againbasedonobser-vations.TheparametersdescribingthesoilhydraulicpropertiesandradiationpropertiesofagrasscoveratthesitearelistedinTable1.
Thegroundwatereffectisidentifiedfromcompari-sonsofresultsfromthenewmodeltoresultsfromtheoriginalmodel(1)–(5),whichdoesnotconsidergroundwater.Fig.4showsthesoilmoisturevariationsintherainycaseusingaconstantgroundwatertableatthecenterofthefourthlayer.Becauseofthegroundwater,thefourthlayerinthenewmodelremainssaturated(solidline).Inthethirdlayer,asmallerdifferenceisobservedbetweentheresultsfromthetwomodels.Thedifferenceinsoilmoisturebegantoincreaseattheendofthefirst2days.Thisdelaywascausedbyboththemodels’responsetimetoinitialsoilmoistureconditionandthegraduallystrengtheningeffectofthesoilmoistureinthefourthlayeronthethirdlayer.Inthemodelwithoutgroundwater,thedryinginthefourthlayerwaslargeduetosubsurfacerunoff.Thedryinginthefourthlayerexertedasimilareffectonsoilmoistureinthethirdlayer.Thetotalsoilmoistureinthemodelwithoutgroundwaterismuchsmallerthanthemodelwithgroundwater.Thesedifferencesdepictthegroundwatereffectonsoilmoistureinthisrainycase.InthepanelsofFig.4wealsoplottedtheresultsfromamodelwithoutgroundwaterbutusinganexponentiallydecreasingsaturationsoilhydraulicconductivityðKsÞ:Inthiscalculation,weusedane-foldingdepthof1.65m21,whichwasestimated
us(m3m23)0.421
Albedo0.19
Cs(m)0.036
Z0(m)0.08
Ks(ms21)1.41£1025Rcmin40.0
uf(m3m23)0.283
Rgl100.0
uw(m3m23)0.028
hs
36.35
b4.26––
us;volumetricwatercontentatsaturation;Cs;saturationsoilsuction;Ks;hydraulicconductivityatsaturation;uf;fieldcapacity;uw;wiltingpoint;b;anexponentinthefunctionthatrelatessoilwaterpotentialandwatercontent;Z0;roughnesslengthinmeter;Rcmin;minimalstomatal
21todoubleitsminimumvalue;hs;aparameterusedincalculatingF2:(Details resistanceinsm;Rgl;thevisiblesolarfluxforwhichF1isabout- 1 -oftheseparametersaregiveninChenandDudhia(2001)).
http://www.paper.edu.cn
X.Chen,Q.Hu/JournalofHydrology297(2004)285–300291
Fig.4.Temporalvariationsofsoilmoistureofdifferentlayersinthreeexperiments.
fromcomparisonsofmodelcalculatedsoilmoistureinthefourlayersagainstobservationsatGudmund-sen.Comparisonswiththeresultsfromtheoriginalmodel,whichusesconstantKs;indicatethataKsvaryingwithdepthcanreducethesubsurfacerunoff(percolation)andincreasesoilmoistureindeeplayers.
Withaloweredgroundwatertable,groundwatereffectonsoilmoistureweakens.Thiseffectisshownbythenewmodel’sresultsinFigs.5and6.Fig.5showsthatalthoughthemoistureinputincreasesintherainycase,theamountofsoilmoisturedecreaseswhenthegroundwatertable,Zg;increases.Attheendofthethreerainydays,thetotalsoilmoisturewasreducedbyabout19%whenthedepthofthegroundwatertabledroppedfromthecenterofthefourthlayerto1mbelowit.Similareffectalsois - 1 -showninFig.6forthedrycase,wherein
thegroundwaterhasamuchgreatereffectonsoil
moisture.Attheendofthe3days,thetotalsoilmoisturewasreducedby25%inresponsetothesamechangeofthegroundwatertable.Thesechangesinsoilmoisturecausedbygroundwaterwillfurtheraffectthesurfaceevaporationandsoilwaterexchangewiththeatmosphere.3.2.Casestudies
Aftershowingthegroundwatereffectonsoilmoisture,weapplythenewmodeltorealcasesandevaluatethemagnitudeandimportanceofground-watereffectsintheSandHillsofNebraska.ThemodelisvalidatedagainstGudmundsen’sobservationsandthenusedtoquantifythegroundwatereffectsonsoilmoistureandevaporation.
3.2.1.Modelvalidation
Invalidatingthemodel,weusedittodescribevariationsofsoilmoistureatGudmundsenforthreeperiodsfrom1998to2000whensufficientatmos-phericandgroundwatertabledatawereavailable,andcomparedthemodelresultsagainstinsituobser-vationsofsoilmoisture.
TheclimateandweatherdatausedtodrivethemodelarefromanautomatedweatherstationatGudmundsen.Amongthedataarehourlyaveragedairtemperature,relativehumidity,windspeedanddirectionat3mabovethesurface,globalsolarradiation,andhourlyprecipitation.Soilmoisturealsowasmeasuredatthestationatfourdepths,0.10,0.25,0.50,and1.0m,whichrepresenttheprobedepthsofthelayers0–0.127m(0–500),0.127–0.381m(5–1500),0.381–0.762m(15–3000),0.762–1.22m(30–4800),respectively.BecausecontinuousmeasurementsoftheseatmosphericandsoilmoisturedatawereavailableonlyintheperiodsofMay5–November30,1998,April1–August14,1999,and April20–July23,2000,themodelwasvalidatedthosethreeperiods.
- 1 -forhttp://www.paper.edu.cn
ThedepthtogroundwatertableatGudmundsen
wasmeasuredatanobservationwell,operatedbytheNebraskaConservationandSurveyDivision(CSD).Theobservationhadamonthlyschedulefrom1989to1992,butwaschangedafter1992toaneeds-basedmeasurement.Becausemonthlyobservationsofthegroundwatertabledepthwerenotavailableineverymonthinthethreevalidationperiods,weusedthefollowingmethodtoderivethemonthlyaveragedepthtogroundwatertableforthemonths.WeusedtheonlyavailablemeasurementsmadeinApril1998,June1999,andMay2000asthe‘anchor’pointsforvariationsofmonthlygroundwatertabledepthineachperiod.Theyare1.07,1.31,and1.26m,respectively.Fortheothermonthsineachperiod,thegroundwatertabledepthswerecalculatedusingthevalueoftheanchormonth(e.g.Aprilforthe1998validationperiod)andthemonthlygroundwatertablechangeratecomputedfromtheobservedmeanmonthlyvariationin1989–1992(Fig.7).Additionally,becausethegroundwatertablevariesslowlyandisusuallyconsideredwelldescribedbydataofmonthly
resolution(Dunne,1978),weusedaconstantgroundwatertabledepthforindividualmonthsinourcalculations.
Theseatmosphericandgroundwatertabledatawereusedtodrivethenewmodel,whosesoilcolumnwasdiscretizedintofourlayersfromthesurfaceto1.22mbeneathittomatchtheprobelayersoftheobservation.Themodellayersfromtoptobottomwere0.127,0.253,0.38,and0.46m,respectively.Modelinitialconditionswerespecifiedusingtheobservationstakenatthebeginningofeachvalidationperiod.OtherparametersusedinmodelvalidationsincludedthosedescribingthepropertiesofloamysandsoilandgrasscoveratGudmundsen.ValuesofthoseparametersarethesameasthoseinTable1.AverticallyvaryingKswasusedwithacalibratede-foldingdepthf¼1:65m21:
ModelresultsforthethreeperiodsareshowninFigs.8–10.Alsoplottedinthesefiguresare,forcomparisonpurposes,observedsoilmoistureandthesoilmoisturecalculatedfromthemodel(1)–(5)withoutthegroundwater.Comparisonsoftheresultsfromthetwomodelsandwiththeobservationsindicatethatgroundwaterhasminorinfluenceonsoilmoistureinthefirstandsecondlayers,whichcomprisethetop0.38minthesoilcolumn.(Thedifferencebetweenthemodeledandobservedsoilmoistureinthetoptwolayersissometimeslarge,resultingfromconsiderablevariationsofthesoilmoisturecausedbybothprecipitationandevapor-ation.)Groundwatereffectonsoilmoistureishoweverquitesubstantialinthethirdandfourthlayersfrom0.38to1.22m.Withoutthegroundwater,thecalculatedsoilmoistureinthosedeeplayersismuchlower.Thisunrealisticdryness,ascomparedtotheobservation,isparticularlysevereinspringandearlysummer(ApriltoJune)whenthegroundwatertableishighestintheyearand,hence,theground-water effectonsoilmoistureismostsignificant.Thedrynessalsoissevereinwetperiodswhen
- 1 -http://www.paper.edu.cn
thesubsurfacerunoffdrainsthedeeplayersmore
effectivelythanindryperiodswithlesssoilwateravailablefordepletion.Withthegroundwater,thecalculatedsoilmoistureinthedeeplayersisclosetotheobservedandyieldstotalsoilmoistureinthesoilcolumnnearlyidenticaltotheobserved(Fig.8e).Thesecomparisonsdemonstratethatthegroundwaterisessentialtomaintainingrealisticsoilmoisturecontentindeepsoillayersandsoilmoistureprofile.Theyindicatethatthegroundwateranditsseasonalvariationplayanimportantroleinsoilmoistureanditsprofilevariationsand,thus,shouldbetakenintoaccountinsoilhydrologymodelsinordertocorrectlydescribesoilmoistureandrelatedsurfacehydrologi-calprocesses,especiallyinareaswheretheground-watertableishigh.
3.2.2.Groundwatereffectsonsoilmoistureandevaporation
AfterverifyingthattheimprovedmodelcapturedthesoilmoisturevariationatGudmundsen,weusedthemodeldatatoevaluatethegroundwaterinfluenceonthesoilwaterbudgetintherootzone.TheresultsaresummarizedinTable2.Amongthesebudgetcomponentsaretotalevapotranspiration,E;surfacerunoff,R;drainagetogroundwater,Rgð¼K4ÞinEq.(6),andgroundwaterloss,LG,discussedbelow.Eisthesumofthedirectevaporationfromthetoplayer,Edir;evaporationofinterceptedrainwateroncanopy,Ec;andtranspirationviacanopyandrootsinthesoillayers,Et:DetailsofparameterizationsforthesecomponentsweredescribedinSection2andinChenetal.(1996)andChenandDudhia(2001).Table2showsthatbecauseofthemoisturesupplyfromthegroundwatertheaverageEfromthemodelwithgroundwaterisgreaterby21,8,and7%foreachofthethreeperiods,respectively,thanthatfromthemodelwithoutgroundwater.Thesurfacerunoff,R;iscalculatedfromR¼Pd2Imax;whereImaxisthemaximuminfiltration.BecauseofthehighinfiltrationrateintheSandHills,PdseldomexceedsImaxsoRisminimal.Incalculation,thevalueofImaxwassetequaltothehydraulicconductivityatsaturation(Table1).
Thetotalgroundwaterlossinformofsupplytosoilmoisture(LG)dependsonthesoiltype,precipitation,potentialevaporation,anddepthofthegroundwatertable.Forafixedsoiltype,thelosscanbecomputed
http://www.paper.edu.cn
294X.Chen,Q.Hu/JournalofHydrology297(2004)285–300
Fig.8.Simulatedandobservedvariationsofsoilmoisture(unit:m3m23)fromMay5throughNovember30,1998.
from:
›CLG¼2KðCÞ
›z
›u¼2D
›z4
4
:ð8Þ
Thetotallossineachoftheperiodscalculatedusing
Eq.(8)isshowninTable2.Betweenthetwoperiods - 1 -in1998and2000,theaveragedepthofgroundwater
tableisaboutthesame.However,moreprecipitation
in1998correspondstolargerEandmoreLG.Ourcalculationsfurthershowthat58%oftheEin1998wasfromthegroundwaterwhereasonly32%wasfromthegroundwaterin2000.Ontheotherhand,betweenthetwoperiodsin1999and2000,thetotalrainfallissimilarbutthedifferenceinthedepthto
http://www.paper.edu.cn
thegroundwatertableislarge.ThisdifferencehasresultedinalargedifferenceinLG.Specifically,whenthegroundwatertablesankto1.32min1999only15%ofEwasfromthegroundwatervs.32%in2000withahigherwatertableat1.17m.
Precipitationrechargetothegroundwater,Rg;wascalculatedasthegravitationalpercolationthroughthesoilcolumn,K4inEq.(6).Theresults - 1 -inTable2showthatthispercolationconsumes18–30%ofthetotalprecipitationatGudmundsen.
Thispercentagerangeofrainfallrechargetogroundwateriscomparabletotheannualrechargeamountof20–30%ofannualprecipitationintheSandHills(NebraskaNaturalResourcesCommis-sion,1986).Apparently,thislargerechargeofrainfallhasplayedakeyroleinmaintainingtherichgroundwaterintheNebraskaSandHillsandadjacentsemiaridregions.Thisrechargewouldbe
http://www.paper.edu.cn
296X.Chen,Q.Hu/JournalofHydrology297(2004)285–300
Fig.10.Simulatedandobservedvariationsofsoilmoisture(unit:m3m23)fromApril20throughJuly23,2000.
significantlyunderestimatedinthemodelwithoutthegroundwater.Anexplanationforthedifferenceinthisrechargebetweenthetwomodelsisthattherootzoneisdrierinthemodelwithoutgroundwater(seethesoilmoisturevaluesinTable2)becauseoflackofgroundwatertransfertothatzone.Moreinfiltratedwateriskeptinthezonetochargethesoilwatercapacityandsupportevaporationandtranspiration.Inthemodelwithgroundwater, - 1 -thesoiliswetter(seeTable2)becauseof
groundwatertransferand,hence,moreinfiltrated
waterispassingthroughthatzonetochargethegroundwater.
4.Groundwatereffectsonregionalsoilmoistureandevaporation
Becauseofthegroundwatereffectsonsoilmoistureandthelocalwatercycle,spatialvariations
http://www.paper.edu.cn
X.Chen,Q.Hu/JournalofHydrology297(2004)285–300
Table2
SimulatedandobservedhydrologicalcomponentsSimulationperiods
Classification
Pd(mm)R(mm)E(mm)Rg(mm)LG(mm)Soilmoisture(m3m23)
StartingEndingMean
1998(May5–November30)Nogroundwater
WithgroundwaterObservation
1999(April1–August14)Nogroundwater
WithgroundwaterObservation
2000(April20–July23)Nogroundwater
WithgroundwaterObservation
Disthemeandepthtogroundwater.
404404404274274274245245245
0.690.690.000.004.684.68
468588384416343370
4.80129.77.6848.244.5148.15
0.2510.2510.2510.2700.2700.2700.2180.2180.218
0.1620.2710.2780.1400.1350.1450.1470.1680.172
297
D(m)
340.261.05117.9
0.1541.160.2091.160.2091.160.1781.320.2411.320.2401.320.1701.170.2171.170.2151.17
inthegroundwatertabledepthcanresultinspatialheterogeneityinsoilmoistureand,subsequently,surfacemoisturefluxacrossaregion.Itisimportanttoevaluatethesignificanceofsuchheterogeneityanditseffectonregionalevaporation.Weexaminedtheregionalsoilmoisturevariationsandcomparedtheregionalsurfacemoisturefluxcalculatedfromthetwomodelsina72km2SandHillsarea(areaAinFig.3).Becauseofdatalimitationsthisevaluationisonlyfor1998.Theobserved1998annualaveragegroundwatertableintheareawasobtainedfromspatialinterp-olationofnearly700observationwellsintheSandHills(sixofthemareinareaA;datafromtheNebraskaCSD:http://csd.unl.edu/csd/metadata/topog.html).Theaveragedepthofthegroundwatertableinthestudyareawas13m,andthedepthvariedfrom0minwetmeadowsofinterdunalvalleysto57munderlargesanddunes.DetailsoftheobservedspatialdistributionofgroundwatertabledepthareshowninFig.11,inwhichthelightshadingindicatesshallowgroundwatertableinmeadowsandwetlands,andthedarkshadingindicatesduneswithadeepwatertable.
Insimulatingspatialvariationsofthesoilmoisture,wedividedtheareaintogrids,andeachgridhasasize300m£300m.Thesamesandysoilandgrasscoverparameterswereusedateachgridcell.Atitscenter,theobservedgroundwatertable,rainfallandotheratmosphericvariablesobservedatGudmundsenwereusedtocalculatesoilmoisture,evaporation,andtranspirationusingboththetwo - 1 -modelswithandwithoutthegroundwater.Model
simulationsstartedonMay7,1998usingthe
observationtakenat01:00localtimeastheinitialconditionandendedat00:00localtimeonSeptember30,1998.
Fig.11.Observedspatialvariationofgroundwatertabledepth(m).Scalesalongtheabscissaandordinateareinmetersfromthesouthwestcorneroftheregion.
http://www.paper.edu.cn
298X.Chen,Q.Hu/JournalofHydrology297(2004)285–300
Becausethesoiltypeandlandcoverarethesameatallthegridcells,themodelwithoutgroundwaterproduceduniformsoilmoistureandsurfaceevapor-ationacrosstheentirearea.Incontrast,themodelwiththegroundwaterdescribedconsiderablespatialheterogeneityofthesevariables.Forexample,Fig.12showsthetotalsoilmoisturecontentsimulatedbythemodelwithgroundwaterinawetconditiononJune14,1998,afterarainevent,andFig.13showsthesoilmoistureinadryconditiononSeptember13,1998,after19consecutivedrydays.Inbothresults,theinterdunalvalleysarerichinsoilmoisture,andthemoisturecontentdecreasesawayfromthevalleysfollowingchangesinthegroundwatertable,showingthegroundwatereffectonthesoilmoisture.Withthecapacitytorepresentthesespatialvariationsinsoilmoistureandtheircontributiontosurfacemoistureflux,themodelwithgroundwatergivesaverydifferentareaaveragedE:
Fig.12.SimulateddailysoilmoisturedistributiononJune14,1998afterarainevent.
- 1 -Fig.13.SimulateddailysoilmoisturedistributiononSeptember13,1998after19consecutivedrydays.
Forinstance,onJune14,1998,theobservedaveragepotentialevaporationwas0.225mmh21;theaveragedEcalculated1bythemodelwithgroundwateris0.182mmh2whereasitisonly0.153mmh21fromthemodelwithoutgroundwater.Thisdifferenceismoresignificantindryperiods.OnSeptember13,1998,theobservedpotentialevaporationwas0.272mmh21andtheaveragedEfromthemodelwithgroundwateris0.0735mmh21,whichismorethandoubleof0.0318mmh21fromthemodelwithoutgroundwater.Thisdifferenceshowstheeffectofspatialvariationingroundwatertableonregionalsoilmoistureandsurfaceevaporation.
5.Summaryandconcludingremarks
Waterflowintherootzoneispredominantly
verticalandcanbedescribedasone-dimensionalin
http://www.paper.edu.cn
X.Chen,Q.Hu/JournalofHydrology297(2004)285–300299
calculatingsoilmoisture.Usually,thisone-dimen-sionalflowisdownward,drivenbyboththegravita-tionalforceandadownwardgradientofsoilmoisture.Thisdownwardmovement,alongwithmoisturedepletionprocesses(e.g.evaporationandtranspira-tion),hasbeenaccountedforinLSMthatareusedinbothGCMandregionalatmosphericmodels.Inadditiontoprecipitation,theothersourceofsoilmoistureintherootzoneisthegroundwater.Indeed,theeffectofgroundwateronsoilmoistureintherootzone,andhencesurfaceevaporation,isdependentonthegroundwatertabledepth.Theeffectcanbesignificantinareaswherethegroundwatertableisnearthesurface.Suchareasincludewetlands,low-landsinriverbasins,andareassimilartotheinterdunalvalleysintheSandHillsofNebraska.Becauseoftheeffectofgroundwater,theverticalgradientofthesoilmoisturecontentisupwardinthoseareas,creatingauniquesoilhydrologicalenvironment.Spatialvariationsofsoilmoisturecreatedbytheheterogeneityofthegroundwatertableinthoseareascanaddanextraspatialvariabilityofthesurfacewaterfluxestoimpactbothlocalandregionatmosphericmoisturedistribution.Thus,correctlyrepresentingthisspatialvariabilityassoci-atedwithgroundwaterisimportantnotonlyforimprovingatmosphericmodelstoaccuratelydescribethehydrologicalprocessesinatmospherebutalsoforimprovingourunderstandingofclimateprocessesrelatedtoandinfluencedbythegroundwatervariations.
Thisstudyintroducedamethodtodescribethegroundwatereffectinlandsurfacemodelswhichcanbeusedinregionalaswellasglobalatmosphericmodels.Afterincludingthismethodinarevisedsoil-hydrologymodelofthePennState-NCARMM5,weusedthenewmodelanddemonstratedthegroundwatereffectsonsoilmoistureandsurfaceevaporation,usingtheNebraskaSandHillsasanexample.Comparisonsoftheresultsfromthenewmodelwithgroundwaterandtheoriginalonewithoutgroundwatershowedthatthenewmodeldescribedrootzonemoisturecontentanditsvariationmuchmoreaccuratelythanthemodelwithoutthegroundwater.Ouranalysesalsoshowedthatinthetop1mofthesoil,thetotalsoilmoisturecalculated usingthemodelwithgroundwatercouldbeasmuchas21%morethanthatinthemodel
- 1 -withoutthegroundwater.Moresoilmoistureinthe
rootzoneresultedinincreasedtotalevaporation.Theaveragesurfaceevaporationwas7–21%higherwhenthegroundwatereffectwasconsidered,closertotheobservedvalue.Extendedcalculationsofasmallareaof5.8£12.4km2nearGudmundsenintheSandHillsyieldedresultsillustratingthatthespatialheterogeneityofthegroundwatertablecannotonlycreateanadditionalspatialvariationsatamagnitudesimilartothatbylandcoverandtopo-graphyvariations,butalsodescribeanareaaverageevaporationnearlydoubledtheamountfromamodelwithoutgroundwater.
Byshowingthestronginfluencesofgroundwateronsoilmoistureandsurfaceevaporation,thisstudyencouragesthatthemethoddescribedinSection3beusedinLSMtoaccountforthegroundwatereffectsinhydrologicalprocessesinbothsoilsandatmosphere.Althoughregionswithgroundwatertabledepthhighenoughtohavesignificanteffectsonsoilmoistureandsurfaceevaporationisasmallfractioninhemisphericandgloballand,thoseregionscanbeasubstantialpartofmodeldomainsinregionalandmesoscalemodels.Thesemodelshaveoftenbeenusedinstudyingregionallandprocesseffectsonseverestormdevelop-ment,suchastheSandHillseffectonseverestormandtornadodevelopmentinNebraskaandthestormandflooddevelopmentinsoutheasternUnitedStates(e.g.the1993floodcasestudies)wheregroundwatertablealsoishigh.Becausethesurfacehydrologicalprocesseshaveplayedanessentialroleparticularlyinenhancingstorms,thegroundwateranditsspatialvariationcouldbeimportantindevelopmentofthoseprocesses.Theinfluenceofthegroundwaterhasbeenneglectedintheexistingstudies,however,applyingLSMwithoutgroundwaterandshouldbeincludedinfuturestudies.
Acknowledgements
TheauthorswishtothankDrsX.H.Chen,C.Rowe,andM.Andersonforusefuldiscussionsduringthiswork,andDrK.HubbardoftheHighPlainsRegionalClimateCenter(HPRCC)forhelpfuldiscussionsontheHPRCCdatausedinthisstudy.Thanksalsogotothetwoanonymousreviewerswhosecommentshelpedimprovetheclarityofthismanuscript.
http://www.paper.edu.cn
300X.Chen,Q.Hu/JournalofHydrology297(2004)285–300
ThisworkwassupportedbyNOAAcontractNA06GP0226throughtheUniversityofNebraska-LincolnandbyUSDACooperativeResearchProjectNEB-40-008.
References
Acs,F.,1994.Acoupledsoil–vegetationscheme:description,
parameters,validation,andsensitivitystudies.J.Appl.Meteorol.33,268–284.
Bentall,R.,1998.In:Bleed,A.,Flowerday,C.(Eds.),AnAtlasofthe
SandHills,ResourceAtlas,No.5a,ConservationandSurveyDivision,UniversityofNebraska,Lincoln,NE,pp.93–114.Beven,K.J.,1984.Infiltrationintoaclassofverticallynonuniform
soils.Hydrol.Sci.J.29,425–434.
Beven,K.J.,Kirkby,M.J.,1979.Aphysicallybased,variable
contributingareamodelofbasinhydrology.Hydrol.Sci.Bull.24,43–69.
Bleed,A.,1998.In:Bleed,A.,Flowerday,C.(Eds.),AnAtlasofthe
SandHills,ResourceAtlas,No.5a,ConservationandSurveyDivision,UniversityofNebraska,Lincoln,NE,pp.67–92.Chen,F.,Dudhia,J.,2001.Couplinganadvancedlandsurface-hydrologymodelwiththePennState-NCARMM5modelingsystem.PartI:Modelimplementationandsensitivity.Mon.Weath.Rev.129,569–585.
Chen,J.,Kumar,P.,2002.Roleofterrestrialhydrologicmemoryin
modulatingENSOimpactsinNorthAmerica.J.Climate15,3569–3585.
Chen,F.,Mitchell,K.,Xue,Y.,Pan,H.,Koren,V.,Duan,Q.Y.,Ek,
M.,Betts,A.,1996.Modelingofland-surfaceevaporationbyfourschemesandcomparisonwithFIFEobservations.J.Geophys.Res.101,7251–7268.
Cosby,B.J.,Hornberger,G.M.,Clapp,R.B.,Ginn,T.R.,1984.A
statisticalexplorationoftherelationshipsofsoilmoisturecharacteristicstothephysicalproprietiesofsoil.WaterResour.Res.20,682–690.
Dickinson,R.E.,1984.Modellingevapotranspirationforthree
dimensionalglobalclimatemodels.In:Hansen,J.E.,Takahashi,T.(Eds.),ClimateProcessesandClimateSensitivity,Geophys.Monogr.Ser.,Vol.29.AGU,Washington,DC,pp.58–72.Dickinson,R.E.,Henderson-Sellers,A.,Kennedy,P.J.,Wilson,
M.F.,1986.Biosphere-atmospheretransferscheme(BATS)fortheNCARcommunityclimatemodel,NCARTech.NoteTN-275þSTR,69pp.,Boulder,CO.
Dunne,T.,1978.Fieldstudiesofhillslopeflowprocesses.In:
Kirkby,M.J.,(Ed.),HillslopeHydrology,Wiley,Chichester,pp.227–293.
Elsenbeer,H.,Cassel,K.,Castro,J.,1992.Spatialanalysisofsoil
hydraulicconductivityinatropicalrainforestcatchment.WaterResour.Res.28,3201–3214.
Entekhabi,D.,Eagleson,P.S.,1989.Landsurfacehydrology
parameterizationforatmosphericgeneralcirculationmodels
- 1 -
includingsubgridscalespatialvariability.J.Climate2,816–831.
Famiglietti,J.S.,Wood,E.F.,1990.Evapotranspirationandrunofffromlargelandareas:landsurfacehydrologyforatmosphericgeneralcirculationmodels.In:Wood,E.F.,(Ed.),LandSurface–AtmosphericInteractionsforClimateModels:Obser-vation,Models,andAnalyses,Kluwer,TheHague.
Famiglietti,J.S.,Wood,E.F.,1994.Multiscalemodelingofspatiallyvariablewaterandenergybalanceprocesses.WaterResour.Res.30,3601–3678.
Koster,R.D.,Suarez,M.J.,Ducharne,A.,Stieglitz,M.,Kumar,P.,2000.Acatchment-basedapproachtomodelinglandsurfaceprocessesinageneralcirculationmodel.1.Modelstructure.J.Geophys.Res.105,24809–24822.
Mahfouf,J.F.,Noilhan,J.,1991.Comparativestudyofvariousformationfrombaresoilusinginsitudata.J.Appl.Meteorol.30,1354–1365.
Mahrt,L.,Ek,M.,1984.Theinfluenceofatmosphericstabilityonpotentialevaporation.J.Climate8,2039–2057.
Manabe,S.,1969.Climateandtheoceancirculation.I.TheatmosphericcirculationandthehydrologyoftheEarth’ssurface.Mon.Weath.Rev.97,739–774.
NebraskaNaturalResourcesCommission,1986.Policyissuestudyonintegratedmanagementofsurfacewaterandgroundwater.StateWaterPlanningandReviewProcess,45pp.
Ookouchi,Y.,Segal,M.,Kessler,R.C.,Pielke,R.A.,1984.Evaluationofsoilmoistureeffectsonthegenerationandmodificationofmesoscalecirculations.Mon.Weath.Rev.112,2281–2291.
Pan,H.L.,Mahrt,L.,1987.Interactionbetweensoilhydrologyandboundary-layerdevelopment.Boundary-LayerMeteorol.38,185–201.
Pielke,R.A.Sr.,Avissar,R.,1990.Influenceoflandscapestructureonlocalandregionalclimate.LandscapeEcol.4,133–155.Pielke,R.A.Sr.,Dalu,G.,Snook,J.S.,Lee,T.J.,Kittel,T.G.F.,1991.Nonlinearinfluenceofmescoscalelanduseonweatherandclimate.J.Climate4,1053–1069.
Schaake,J.C.,Koren,V.I.,Duan,Q.Y.,Mitchell,K.,Chen,F.,1996.Simplewaterbalancemodelforestimatingrunoffatdifferentspatialandtemporalscales.J.Geophys.Res.101,7461–7475.Sellers,P.J.,Mintz,Y.,Sud,Y.C.,Dalcher,A.,1986.Asimplebiospheremodel(SiB)forusewithingeneralcirculationmodels.J.Atmos.Sci.43,505–531.
Stieglitz,M.,Rind,D.,Famiglietti,J.S.,Rosenzweig,C.,1997.Anefficientapproachtomodelingthetopographiccontrolofsurfacehydrologyforregionalandglobalclimatemodeling.J.Climate10,118–137.
Wood,E.F.,Lettenmaier,D.P.,Zartarian,V.G.,1992.Aland-surfacehydrologyparameterizationwithsubgridvariabilityforgeneralcirculationmodels.J.Geophys.Res.97,2717–2728.Xue,Y.,Sellers,P.J.,Kinter,J.L.,Shukla,J.,1991.Asimplifiedbiospheremodelforglobalclimatestudies.J.Climate4,346–364.
因篇幅问题不能全部显示，请点此查看更多更全内容