CLIMATE CHANGE, HYDROLOGY, AND WATER RESOURCES

Peter H. Gleick
PacificInstituteforStudies
in Development,
Environment, andSecurity
Berkeley,California

Abstract. Growingatmospheric concentrations of carbon soil moisture,changelake levels,and affect waterquality.

dioxide and other trace gases are leading to climatic Such changesraise the possibilityof environmentaland
changeswith importantimplicationsfor the hydrologic socioeconomicdislocations,and they have important
balanceand water resources.These "greenhousegases" implicationsfor future water resourcesplanning and
are expectedto alter the radiativebalanceof the atmos- management.This paper reviews state-of-the-artresearch
phere, causingincreasesin temperatureand changesin into theimplicationsof climaticchangesfor thehydrologic
many other climatic variables. Recent hydrological cycle and for water resourcesand discussesthe implica-
researchstronglysuggests that this so-called"greenhouse tions of such changesfor future water planning and
effect" will alter the timing and magnitudeof runoff and management.

INTRODUCTION A BRIEF HISTORY OF THE DEVELOPMENT OF THE

CLIMATE AND WATER LITERATURE

Concern over global climatic changes caused by

growingatmospheric concentrations
of carbondioxideand Much of the earliest literature on climatic change
other trace gases has increasedin recent years as our focusedon the ability of society to alter local climatic
understanding of atmospheric
dynamicsandglobalclimate conditions,ratherthan on the impactsof climaticchanges
systemshas improved. These gases,principallycarbon on society[seeStudyof Critical EnvironmentalProblems
dioxide(CO2),nitrousoxide(N20), methane (CH4),and (SCEP), 1970; Studyof Man's Impact on Climate (SMIC),
chlorofluorocarbons (CFCs), alter the heat balanceof the 1971]. How, for example, might large reservoirsaffect
Earth by retaining long-wave radiation that would downwind temperaturesor precipitationpatterns? How
otherwisebe lost throughthe Earth's atmosphereto space. might large-scale land use modifications alter local
This effect, known colloquiallyas the greenhouse effect, climaticconditions?Even whenattentionfinally turnedto
gainedwidespread publicattentionin 1988aftera seriesof anthropogenicclimatic changes, such as greenhouse
unusuallywarm yearswere attributedto risingconcentra- warming,very little of the early work consideredpossible
tionsof greenhouse gases[Hansen,1988]. impactson waterresources; instead,mostresearchfocused
Despite recent improvementsin our understanding of on the implicationsof higher temperaturesand altered
atmosphericdynamicsand large-scaleclimatic processes precipitationpatternsfor agriculturalproductionand the
the climatic effects of greenhousegases are still only level of the oceans. This lack of attention to water
partially understood.One of the most importantconse- resourceswas due to three factors: an early focus on
quencesof futurechangesin climatewill be alterationsin large-scalecomputersimulationsof the climate, uncer-
regionalhydrologiccyclesand subsequent effectson the tainty about how future precipitationpatternswould be
quantityandqualityof regionalwaterresources.Yet these affectedby climaticchanges,and doubtsaboutappropriate
consequences are poorly understood. hydrologicalmethodsfor evaluatingfuture(and uncertain)
Recent hydrologicalresearch strongly suggeststhat climaticchanges.
plausibleclimatic changescausedby increasesin atmos- One of the few early attemptsto evaluate the interac-
pheric trace gas concentrations will alter the timing and tions among hydrology, water supply, and climatic
magnitudeof runoff and soil moisture,changelake levels conditions was the National Research Council Panel on
and groundwateravailability, and affect water quality. Water and Climate [1977, p. 3] study Climate, Climatic
Sucha scenarioraisesthe possibilityof dramaticenviron- Change, and Water Supply. At that time, the authors
mentalandsocioeconomic dislocationsandhaswidespread concluded,
implications for future water resourcesplanning and Future water shortagesmay be exacerbatedby climatic
management. change,but unfortunatelythe climatologist'scurrentforecast

Copyright1989 by the AmericanGeophysical

Union. Reviewsof Geophysics,
27, 3 / August1989
pages329-344
8755-1209/89/89RG-00765 $05.00 Papernumber89RG00765
329 ß
330 ß Gleick'. CLIMATE CHANGE AND HYDROLOGY 27, 3 / REVIEWSOF GEOPHYSICS

ability is insufficientto aid the water-resourceplanner or Much remainsto be done. Fundamentalquestionshave

hydrologicdesigner.To be usefulto water-resource planning, yet to be answeredabouthow greenhouse warmingwill
climatic-changeforecastswould need to be specificby area alter regional precipitation patterns and how water
and be accurateover the 50 to 100 year design-lifeof the
availability and quality will be affected. Very few
water-resource system.... There is no evidencethat sucha
forecast ability either exists or will appear within the watershedshave been studiedin detail using appropriate
immediate future. models or methods. Little work has been done on the
interactionsamong climate, vegetation responses,and
Despiteproblemsin accuratelyforecasting regionaleffects, waterresources.The role of management in mitigatingthe
the National Research Council Panel on Water and impactson water resourceshas been inadequatelyas-
Climate [1977, p. 19] also concluded,"them are many sessed. Questionsremain about the implications of
usefulmeasuresthatcouldbe implementednow thatwould climatic changefor sharedinternationalrivers. And the
help to mitigate the undesirableeffects of future water role of international water law and water treaties in
shortages." resolving climate-induceddisputes is still unresolved.
Not until 1985, when the Scientific Committee on Theseissuesare of greatimportanceto societyif critical
Problems of the Environment (SCOPE) published its impactsare to be identifiedand if concertedeffortsare to
report "Climate Impact Assessment"[Kates et al., 1985], be madeto reducetheconsequences of climaticchange.
were water resourcesconsideredin detail by the climate
impactscommunity. Even then,majorreviewscontinued
to ignore water resources(see, for example,Bolin et al., METHODS OF ANALYSIS
[1986]). By thistime,however,thehydrologiccommunity
had begun to raise importantquestionsaboutassessment Novalcyet al. [1985] and Beran [1986] distinguish
techniquesand critical water resourceissues. How might betweenhydrologyand water resources,where hydrology
ecosystemsand resource managementbe affected by is the scienceof the hydrologicregime, and water re-
greenhousewarming? What typesof hydrologicmodels sources refersto the quantityandqualityof wateravailable
are availablefor climate impactassessment,and which are at a particulartime andplace. Thushydrologists focuson
appropriatefor regionalstudies? theeffectsof climaticchangeson precipitation,runoff,soil
Researchersin the 1980s also began to explore the moisture,and the statisticsof water availability,whereas
regional hydrologic effects of greenhousewarming waterresourceplannersare moreinterestedin the effect of
throughderailedcasestudies,usinga varietyof methods. climatic changes on municipal water supply, hydro-
In 1985 the World MeteorologicalOrganization(WMO) electricityproduction,reservoirdesignand management,
publisheda review of the effectsof climatic variationson reliableyield, andirrigationrequirements.
water resourcesystemsthat includedspecificrecommen- Ideally, thereare three stepsto evaluatingthe implica-
dations for testing and validating different assessment tionsof climaticchangesfor water: (1) Developquantita-
methods[Klemeg,1985]. This was followedin 1987 by a tive scenariosof changesin major climaticvariables,such
summaryof the sensitivityof water resourcesystemsto as temperature,precipitation,and evapotranspiration; (2)
both future climatic change and existing climatic simulatethe hydrologiccyclefor a basinof interest,using
variability [WMO, 1987]. Also in 1987, the International the scenariosdevelopedin step 1; and (3) assessthe
Associationof HydrologicalSciences(IAHS) devoteda implications of the hydrologicvariationsidentifiedin step
sessionof the XIXth General Assemblyof the Interna- 2 for the performanceof suchwater resourcesystemsas
tional Union of Geodesy and Geophysics(IUGG) to dams,aqueducts, reservoirs,groundwaterrechargebasins,
climateand water resourcesand publisheda collectionof and so on. These stepsare indicatedschematicallyin
articleson the subject[Solomonet al., 1987]. In 1988 the Figure 1. Hydrologists,water-resourceplanners, and
Australian division of Atmospheric Research held a engineers haveimportantrolesto play in eachstep.
meetingon climatic changeand publisheda book with a Accuratequantitativeestimatesof changesin major
sectionevaluatingthe implicationsof greenhouse warming long-term climatic variables such as air temperature,
for Australianhydrologyand water resources[Pearman, precipitation, evapotranspiration,
andvegetationtypesand
1988]. Most recently, the AmericanAssociationfor the distributionsare needed in order to provide accurate
Advancementof Sciences(AAAS) conveneda panel to forecasts of wateravailabilityandquality. Unlessreliable
write a book on the implicationsof climatic changeand climate forecastscan be producedby climatologists,the
variabilityfor waterresourcesin the United States[AAAS, predictivevalue of all hydrologicassessments are neces-
1989]. Under the auspicesof the AAAS an attemptwas sarilylimited.
madeto studythe entirescopeof the problem,from basic Armedwith suchquantitativeclimatological estimates,
physical, climatological,and hydrologicissuesto ques- hydrologistscan use simulationmodelsof watershedsto
tions of political and economic allocation of water evaluatethe impactsof climatic changeson runoff, soil
resources under new climatic conditions. moisture,snowmeltrates,andotherhydrologicvariables.
27, 3 /REVIEWS OF GEOPHYSICS Gleick: CLIMATE CHANGE AND HYDROLOGY ß 331

Runoff and Soil Moisture Scenarios Temperature and Precipitation Scenarios

GCM Data GCM Data
Paleoclimate Data Paleoclimate Data
Recent Analogue Data Recent Analogue Data
Hypothetical Data Hypothetical Data

(Daily, monthly, annual T,P)

Hydrologic Mode s

(Dail monthly, annual runoff or soil m sture)

Reservoir Operation
Irrigation Supply
- Water Management Models Hydroelectric production
Urban Water Supply

Figure 1. Methods for evaluatingthe hydrologic and water climatologicalvariables to drive hydrologicmodels, and the
resourceimplicationsof climatic changes. Climate change hydrologicoutputis usedas inputto watermanagement models.
scenarioscan be generatedusing general circulation models Alternatively,hydrologicscenariosof runoff and soil moisture
(GCMs), paleoclimaticdata, recent climate analoguedata, or can be generatedand useddirectlyto drive watermanagement
hypotheticalclimate data. Two types of scenarioscan be models. Each technique has advantagesand limitations as
generated: hydrologic or climatological. Climatological described in the text.
scenariosuse changesin temperature,precipitation,or other

Finally, water managerscan use the new hydrologic cannot"do the experiment" of alteringglobal climate in a
estimatesto evaluatethe performanceand designof their laboratory,we mustattemptto model climateand climatic
systemsunderdifferent climatic conditions. Thus flood changes,an imprecisealternativebecauseof the complex-
controlor reservoiroperationcanbe studied,hydroelectric ity of the global climate system. Much of the effort of
outputoptimized,emergencyfloodplainmanagement plans trying to understandclimatehas focusedon the develop-
revised,anddroughtplanningreviewed. ment of computer models of the many intricate and
To assessthe implicationsof the greenhouseeffect for intertwinedphenomenathat make up the climate. The
water resources,regional-scaledetails of future changes most complex of these models, typically referred to as
are needed for temperature,precipitation,evaporation, "generalcirculationmodels" or "global climate models"
wind speed,and otherhydroclimatological variables. Our (GCMs), are detailed, time-dependent,three-dimensional
ability to predictthesedetails,however,is limited, and we numericalsimulationsthat include atmosphericmotions,
must therefore resort to the use of climate scenarios. Such heat exchanges,and importantland-ocean-iceinteractions
scenariosshouldbe internallyconsistentpicturesof future [Manabe, 1969a, b; Schlesingerand Gates, 1980;Manabe
conditions,and they can be constructedby a numberof and Stouffer, 1980; Manabe and Wetheraid, 1980;
methods,includingthe direct use of generalcirculation Wetheraid and Manabe, 1981; Ramanathan, 1981;
models, paleoclimatic reconstructions,recent historical Manabe et al., 1981; Hansen et al., 1983, 1984, 1988;
climate analogues, or purely hypothetical climatic Washingtonand Meehl, 1983, 1984; Wilsonand Mitchell,
scenarios[WMO, 1987]. 1987; Mearns et al., 1989; Schneideret al., 1989].
Generalcirculationmodelscurrentlyprovide the best
informationon the responseof the atmosphereto increas-
Direct Use of General Circulation Models ing concentrations of greenhousegases. In theory,GCM
There are diverseways in which globalclimatemay be estimatesof changes in hydrologic variables, such as
affected by human actions. But we are unable to see runoff,couldbe useddirectlyto estimatechangesin water
clearly either the directionof changesin climate or the resources. The U.S. EnvironmentalProtectionAgency
nature of the subsequentsocietal impacts. Becausewe [1984] took this approachto estimatepossibleimpactsof
332 ß Gleick: CLIMATECHANGEAND HYDROLOGY 27, 3 / REVIEWS
OF GEOPHYSICS

greenhouse runoff,and soil everywhere

warmingon precipitation, [seeManabe,1969a].Runoffoccurs
whenthe
moistureoverlargeareasof theUnitedStates.Usingthe soil moisture exceeds this capacity, and the rate of
GoddardInstitutefor SpaceStudies(GISS) model,they evaporation
is determinedasa simplefunctionof thesoil
notethat althoughsignificanthydrologicchanges
canbe moistureand the potentialevaporation
rate [Manabeand
expected undera doubledCO2 climate,the hydrologic Wetheraid, 1985]. Efforts are being made to improve
outputfromGCMswasnotsufficiently to define GCM hydrology[Dickinson,1984], such as through
accurate
futureconditionsin particularwatersheds. improvements
in vegetationparameterizations
and the
Manabe and Wetheraid [1986] used the Geophysical behaviorof soils,but until suchimprovements occur,those
FlUid DynamicsLaboratory(GFDL) generalcirculation interested in the implicationsof climaticchangesfor water
model to look at soil moisturein the midcontinentalregion resourcesmust look for other methods of evaluating
of the United Statesandconcludedthat significantdrying hydrologicimpacts.
may occurif the concentration
of CO2 in the Earth's Changesin climatologicalvariables estimatedby
atmosphereis doubled.Althoughthe resultsfrom the GCMs, suchas changesin temperature
and precipitation,
variousGCMs are not in completeagreementon the issue are consideredto be more reliable than GCM-predicted
of midcontinental summerdryness[e.g., Schlesinger and changesin runoffor soilmoisture.As a result,thereis a
Mitchell, 1985, 1987;Mitchell and Warrilow, 1987], more trendtowardusingtemperatureandprecipitation estimates
recentstudiesare beginningto confirmsignificantsoil fora doubled-CO
2 environment
asinputstomoredetailed
moisturereductions[Wilsonand Mitchell, 1987]. regional
models.Thisapproach
is described
below.
One other feature of GCM-computed hydrologic
changes
merits
special
attention: andearlier Paleoclimate
thedecrease Analogues
disappearance
of wintersnowpack. Although estimates
of Variationsof climaticand hydrologicconditionsfrom
changes
in precipitation
patternsvary considerably
from onegeologicalperiodto anotherandfromonemillennium
one GCM to another,all GCMs showshortersnowfalland to anotherare clearlyevidentin the geologicrecord. By
snowmelt seasonsdue to increasesin average tempera- lookingat thesevariationsand tryingto identifyperiods
tures.The changein snowconditions, in turn,leadsto the that may be similarto futuregreenhouse conditions,we
earlierand significantincreaseof evaporationfrom bare canexplorewheresignificant
changes
in wateravailability
soils,whichis responsible for muchof the summersoil may appear.
moisture effects discussed earlier. Similar alterations of Paleoclimaticanaloguesare reconstructions
of informa-
snowfall and snowmelt conditions have now been tion on precipitation,temperature,
evaporation,or other
identifiedin regionalmodel studies,discussed in detail climaticvariablesfrom a varietyof long-termrecordssuch
below. astreerings,pollendeposits, vegetation or fossiltypes,the
GCM-generated hydrologic datasufferfromtwo major appearance or disappearance of civilizations,lacustrine
limitations. First, the spatialresolutionof GCMs is too (lakesediment) deposits,
shorelineterraces, tracesof dunes
coarseto provide hydrologicinformationon a scale and other morphologicalfeatures,and chemicalisotope
typicallyof interestto hydrologists. (GCM resolution is ratiosin ice cores.Figures2 and3 showtwo suchclimate
unlikelyto dramaticallyimprovefor manyyearsbecause reconstructions.
of theextremecostof high-speed computertime(a factor Nicholsonand Flohn [1980] exploredwateravailability
of 2 increasein resolutionrequiresapproximately a factor in Africa usinga varietyof paleoclimaticrecords. They
of 8 increasein computertime[Somemille, 1987]). With a suggest thatpartsof the Sahelregionweredrier 18,000
typical model resolution of4.5ø latitudeby7.5ø longitude years before present(B.P.), becamemore moistin the
and nine verticallayersin the atmosphere, computing1 period10,000-4500yearsB.P.,andthenbecamedrierup
year of weatherat 30-minintervalstakes10 hoursof to the present. Thesechangescan then be relatedto
computer timeona CrayXMP computer, oneof thefastest possible driving forces in the global atmospheric
in theworld.)Presentresolutions
areusually
from4ø to circulation.
7.5ølatitude
by5øto 10ølongitude, ofhundreds The useof pollenin coresfrom lakesandmarshysites
gridareas
of thousands
of squarekilometers. Yet hydrologists
are alsocan providevery long climaterecords,as shownby
interested in climatic events that occur on the scale of tens Adamand West[1983], who analyzedpollenin coresfrom
or hundreds of squarekilometers,
a scaleseveralordersof Clear Lake in northern California to estimate both
magnitude finerthancurrentGCM resolution. temperatureand precipitationduring the past 130,000
Second,hydrologic parameterizations
in GCMsarevery years. Velichko[1984] also used "paleofloristics"to
simpleand oftendo not providethe detailedinformation reconstruct
pastclimaticconditionsacrossthe areaof the
necessaryfor waterresourceplanning[WMO, 1987]. The Soviet Union during the late Pleistoceneinterglaciation.
GCM soil moisturebudget, for example, is typically Thesedata are now being usedby Soviet climatologists
computed by theso-called"bucketmethod,"in whichthe seekingclimateanalogues for assessingtheimplicationsof
field capacityof the soil is assumedto be uniform greenhousewarmingon soilmoisture andrunoff.
27, 3 /REVIEWS OF GEOPHYSICS Gleick: CLIMATE CHANGE AND HYDROLOGY ß 333
30

Figure 2. The 400-yearrunoff recordin the

Colorado River Basin reconstructedfrom tree
rings(plottedas a 10-yearmovingaverage)
[from Stocktonand Jacoby,1976]. Note the
anomalouslyhighrunoffduringtheearlytwenti-
ethcentury.Thisshortperiodwasusedto deter-
mine allocations for the 1922 Colorado River
compact.

tD c,D ½D D- t'.- 0D CO C• O•

ppmv
Figure 3. Atmosphericcarbon dioxide con- 400
centrationin partsper million by volume (ppmv)
for thepast160,000years,reconstructed from the
Vostok
icecore,
Antarctica
[from
Barnola
etal., 350
1987]. NotethehighCO2 concentration
during
themostrecentinterglacial,
the low CO2 con-
centrationduring the most recent glacial maxi-
mum,therisein CO2 at thestartof thecurrent
300
P
re•s
triall
a••Vel,
aCia'
Max
imum
Prev
iou
s,i•e
rg,
ac
ia,
interglacial(18,000 years B.P.), and the recent 250
dramaticincrease
in CO2 concentration
dueto
industrial activities.
200

150
0 20 40 60 80 100 120 140 160

Thousand Years Before Present (yr BP)

Morgan [1985] compiled a continuousrecord of Perhapsthe most ambitiouseffort to use paleoclimatic

temperaturefrom the third century A.D. calculatedfrom datato evaluatefutureclimaticimpactsis the attemptof
oxygen isotopeice core data from the Antarctic. Aris- Soviet climatologiststo develop information on past
tarain et al. [1986] measured the deuterium content of periodsthat are, in theory,comparableto future periods
snowlayersin Antarcticato generatea temperaturerecord expectedunderconditionsof greenhouse warming[Budyko
from 1850 to the present, and they note that several and Izrael, 1987]. In this work the Holoceneoptimum
thousandyearsof recordmay be accessiblein partsof the (about6000 to 5000 yearsB.P.) is considered
an analogue
Antarctic. Indeed, the more recent Vostok core [Barnola for a IøC warming; the lastinterglacial
(about125,000
et al., 1987] extendsthe recordof atmospheric
CO2 yearsB.P.) is consideredan analoguefor a 2ø to 2.5øC
concentration back 160,000years(Figure3). warming;andthe Plioceneclimaticoptimum(about3 to 4
In an early climatereconstruction,Snyderand Langbein millionyearsB.P.)corresponds
to a warmingof 3ø to4øC.
[1962] evaluatedthe role of changesin temperatureand Althoughthereare limits to the parallelsthat canbe drawn
precipitationin alteringthe level of a Pleistocenelake in betweenthepastperiodsandfutureconditions, thistypeof
Spring Valley, Nevada. Similarly, Goodfriendet al. activity can provide valuable insights into climate
[1986] exploredthe paleoclimaticevidencefor climatic dynamics.
changesin the areaof the JordanRiver Basinand the Dead Techniques are alsoavailablefor extendingthe current
Sea, anotherterminal lake sensitiveto changesin inflow instrumental record back several centuries [Stockton,
and evaporation.They identifiedlargefluctuationsin the 1977]. Such reconstructionscan provide valuable
level of the Dead Sea in the late Pleistoceneperiod up information on both past climatic conditions and the
through4300 yearsB.P. vulnerability of our water resource system to future
334 - Gleick: CLIMATE CHANGE AND HYDROLOGY 27, 3 / REVIEWSOF GEOPHYSICS

changes. In a striking example, Stocktonand Jacoby On a far larger scale,Bradley et al. [1987] compiled
[1976] usedtree tings to extendthe runoff recordin the precipitationdata from the mid 19th century,northern
ColoradoRiver basinback morethan400 years(Figure2). hemisphere landareasin an effort to look for trends.In the
This kind of study has direct water managementand last 30 to 40 years, significantincreasesin mid-latitude
policy implications. For example, the original 1922 precipitation (35ø to 50øN)anddecreases in low-latitude
Colorado River water allocation was based on the precipitation
(5ø to 35øN)havebeenobserved.While
hydrologicrecord availableat the time: about 30 years these trends are consistentwith GCM projectionsof
from the late 1890sto the early 1920s. In 1976,whenthe precipitation changes associated withincreasingconcentra-
historical record was reconstructed back to the middle tions of trace gases, they cannot be unambiguously
1500s, the period from 1890 to the 1920s stoodout as a attributedto this cause. Such analyseshelp to define
time of abnormallyhigh runoff (Figure2). The 400-year naturallarge-scalehydrologicalvariabilityand are useful
record now shows that more water was allocated to users in defining the longer-termperspectivefor identifying
thanis likely to be availableon a long-termaveragebasis. anthropogenic perturbations.
If the long-term record had been available in 1922, the Finally, recentclimatic anomaliescan provideinsights
overallocationmightnot haveoccurred. into the vulnerability of the present society to future
Several things limit the usefulnessof paleoclimatic climatic changes. Glantz and Ausubel[1984], Changnon
scenariosfor evaluating the impacts of future climate [1987a], Gleick [1988], Morrisette [1988], Cohen [1988],
changeson waterresources.First, the fartherwe go back and Karl and Reibsame[1989] all discussregionalwater
in time, the moredifficult it becomesto recoverhydrologic resourceimpacts that are a function of past climatic
data. Thus the evidencefor developingsuchscenariosis variability. Thesecasestudiesprovideinformationon how
limitedin extentand scope. Second,the causesof climatic to avoid environmentallyand economicallycosily climate
shiftsover geologictime may differ considerably from the impactsin the future.
anthropogenicchanges now anticipated. And third, The greatestadvantageto usingrecentclimateanalogies
because these past changes typically predate human or case scenariosis that they are based on human ex-
activity, we have no evidenceof how they might affect perience. Perhapsthe greatestdrawbacksto this approach
society.Becauseof thislast limitation,attentionhaslately are thatanticipatedgreenhouse climaticchangesmay have
turned to the use of more recent climatic extremes within a causeunlikethosethatled to pastclimaticvariabilityand
the instrumental record. that the magnitudeof the anticipatedchangesdue to the
greenhouseeffect are larger than most historicalnatural
RecentClimateAnalogues variations. Thus suchanalysesare often criticizedon the
Modem instrumentaldata can be used to develop groundsthat the recentpast is an unreliableguide to the
climatescenarios.Examplesfrom the literatureincludea future.
reviewandadjustmentof instrumental temperature records
from central Canada from 1768 up to 1910 [Ball and
Kingsley,1984]; a reconstruction of temperature for North Hypothetical ClimateScenarios
Americafrom 1602 to 1961 usinghigh-qualitytree ring Many hydrologistsusepurelyhypotheticalscenarios to
records[Fritts and Lough, 1985]; a reconstruction of the assess the behavior of watersheds in relation to future
long-termprecipitationrecordat Padova,Italy, from 1725 climatic conditions. Suchscenarios (e.g.,+2ø,+3ø,+4øC;
to 1981 [Camuffo, 1984]; and the developmentof an increasesor decreasesof 10% or 20% precipitationor
annual rainfall index for southern California from 1769 to evapotranspiration) permit the testingof wide rangesof
1834 using agriculturalrecordsfrom the Spanishperiod hydrologicvulnerabilities. Althoughthesescenariosare
[Rowntree,1985]. These types of analysescan greatly the easiestto develop,they are not particularlyrealistic,
extendtheusefulhistoricalrecordin a region. and they often lack intemal consistency.Care shouldbe
Long-termrecordscan thenbe analyzedfor examplesof taken,therefore,in interpretingtheresults.
climaticvariabilitythatoffer insightsintothe vulnerability Table 1 liststhe rangeof hypotheticalscenariosusedin
of watersystems.Wigleyet al. [1984] lookedat precipita- a variety of studies. The valueschosentypicallyreflect
tion records in England and Wales for the period bestestimatesof changesin importantclimatic variables,
1766-1980 to evaluatewhethertherehadbeena changein althoughextremevaluesareoccasionally chosento explore
the frequencyof wet and dry rainfall extremes. Jones where a systemmight fail to perform as expectedor
[1984] then used these data to reconstructriver flow to designed.Thusthepracticeof usinghypothetical tempera-
study the effects of climate on water availability. tureincreases of 1ø, 2ø, 3ø, or 4øCreflectstheconsensus
Similarly, Palutikof [1987] used scenariosconstructed thatgreenhouse warmingwill producetemperature risesin
from the instrumentalrecordto explorepossibleimpacts this range for an equivalentdoubling of atmospheric
on UnitedKingdomwaterresources in theearlyyearsof a carbondioxide. Greateruncertaintyaboutthe magnitude
climaticchange. andeventhe directionof regionalprecipitation changesis
27, 3 /REVIEWS OF GEOPHYSICS Gleick: CLIMATE CHANGE AND HYDROLOGY ß 335

TABLE 1. RegionalStudiesUsingHypotheticalClimate Scenarios

Scenario

Evapo-
Study Temperature transpiration Precipitation

Stocktonand Boggess[1979] _+2øC + lO%

ß
N•mec and Schaake [1982] +iøC, +3øC _+10%,25%
Revelle and Waggoner [1983] +2øC,+4øC -10%
Flashcka et al. [ 1987] _2øC ß
+10%, 25%
Gleick [1986b, 1987a, b] +2øC,+4øC _+0%, 10%, 20%
Fitzgerald and Walsh [1987] _+5%,10%, 15% +5%, 10%, 15%, 20%
ß
Schaake [1989] _+10% + 10%, 20%
ß

reflected in the choice of both increases and decreases in (1) thosemodelsthatrely primarily on empiricalor statisti-
monthlyor annualaveragerainfall. cal techniquesfor evaluatingthe hydrologiccharacteristics
of a regionor for extendingthe existinghydrologicrecord,
suchas the modelsof Schwarz[1977], Revelle and Wag-
COMBINING CLIMATE SCENARIOS WITH REGION- goner [1983], and Stocktonand Boggess[1979], and (2)
AL HYDROLOGIC MODELS thosetechniquesthat are physicallybasedmathematical
descriptionsof hydrologicphenomena, the so-calleddeter-
Once scenariosof climate change are developed, ministicor conceptualmodels[N•mecand Schaake,1982;
hydrologicmodelscan be used to estimateimpactson Gleick, 1986a, b, 1987a, b; Mather and Feddema, 1986;
water resources. If accurate estimates of future water Cohen, 1986; Flaschka et al., 1987; Bultot et al., 1988].
availability are to be calculated,regional hydrologic Stochastichydrologicmodelswere initially introduced
evaluationsneed to incorporatethe complexitiesof to analyzereservoirdesignand operation. Thesemodels
snowfall and snowmelt,topography,soil characteristics, providethe meansfor estimating,amongotherthings,the
natural and artificial storage,and monthly or seasonal probabilityof sequences of future dry or wet yearsgiven
variations. past hydrologicdata, and they are often easierto design
The conceptof usinghydrologicmodelsfor assessing and manipulatethan deterministicmodels. But since
regionalimpactsof climaticchangeshasseveralattractive stochastichydrologic techniques assume that future
characteristics.First, diversemodelingtechniquesexist. hydrologicbehaviorwill look statisticallylike the past,
This permits flexibility in identifying and choosingthe theyare of limited usein evaluatingthe effectsof climatic
mostappropriateapproachto take to evaluateany specific changes,which may alter the underlyingdistributionsand
region. Second,hydrologicmodelscanbe chosento fit the physicalrelationships amonghydrologicvariables.
characteristicsof the availabledata. Third, regional-scale For thatreason,considerable attentionhasbeengivento
modelsrequire far fewer computerresourcesand are far the wide range of deterministic, physically based
easierto manipulateand modify than generalcirculation hydrologicmodels. Many deterministicmodelshavebeen
models. Fourth,regionalmodelscan be usedto evaluate developed to analyze different types of hydrologic
the sensitivityof specificwatersheds to bothhypothetical phenomena.The modelsvary in their ability to represent
changesin climateandto changespredictedby large-scale or reproducesmall- and large-scalefeaturesof watersheds,
GCMs or climate analogues. And finally, methodsthat from narrowly focused models that study short time
incorporateboth detailed regional characteristicsand period, site-specificcharacteristicsto general models
outputfrom GCMs will be well situatedto takeadvantage capableof incorporating waterbalancesin a largeregion.
of the continuingimprovements in theresolution,regional Eachtypeof modelhasstrengths and limitations,depend-
geography, andhydrologyof globalclimatemodels. ing on modeldesign,datarequirements, and the objectives
Many typesof hydrologicsimulationmodelshavebeen of theanalyst.
developedin recent years to help hydrologistsstudy Becauseof thesevariedstrengths andlimitationsa setof
ecosystems, to aid in the engineeringdesignof structures, criteria for using regional models to evaluate the
andto studythe response of watershedsto differenttypes hydrologicimpacts of climatic changesmust be con-
of perturbations.Many classificationschemeshavebeen sidered. Gleick [1986b], expandingon the work of
used to discriminateamong these models, including Schaakeand Kaczmarek [1979], developeda set of six
physical and mathematical, continuous and discrete, importanttechnicalfactorsto be considered whenselecting
dynamic and static, descriptiveand conceptual,and andusinga regionalhydrologic modelto studytheimpacts
stochasticand deterministicmodels [Viessmanet al., 1977; of changesin climateon regionalwaterresources:(1) the
Linsleyet al., 1982]. For thepurposeof climateimpactas- inherentaccuracyof the model; (2) the degree to which
sessmentan importantdistinctioncan be madebetween modelparametersdependupon the climatic conditionsfor
336 ß Gleick: CLIMATE CHANGE AND HYDROLOGY 27, 3 / REVIEWSOF GEOPHYSICS

which the model is calibrated;(3) the availability of input
data, including comparative historical data; (4) the
accuracyof the input data; (5) model flexibility, easeof
use,and adaptabilityto diversehydrologicconditions; and
(6) compatibilitywith existinggeneralcirculationmodels.
magnitudeof flows. For assessment techniquesto be
applicableto many diverse watershedsthe modeling
techniques mustbe flexible in their approach.As noted
earlier, additionalcomplexity is no guaranteeof model
accuracy;it also reducesmodel adaptabilityand easeof
use.

The InherentAccuracyof the Model

Models vary in their ability to reproduceexisting Compatibility
WithLarge-Scale
ClimateModels
hydrologic conditions in a watershed. The World The desire for compatibilitywith general circulation
Meteorological Organization [1975] studieda series of models arises from our desire to understand how climatic
simulationmodelsin order to evaluatetheir strengthsand changeswill affect water resources. Since general
weaknesses.They concludedthat many modelsperform circulationmodelsnowprovidethe mostdetailedinforma-
well for humidbasinsbut that explicit accountingmodels tion on future changesin climate, the ability to link
suchas water balancemodelswere distinctlysuperiorin regionalhydrologicmodelswith outputfrom GCMs will
semiaridand arid basins. Perhapseven more significant improveour abilityto understand the regionalimpactsof
was the observationthat simpler models showedbetter globalclimaticchanges.AlthoughGCMs do notprovide
resultsthancomplexmodelswhenthe qualityof inputdata sufficient detail on regional impacts to be used for
is poor. purposesof prediction,they do provide an internally
consistentdescriptionof plausible patternsof climatic
InitialModelCalibrationandChanging Conditions change. By linking GCMs and regional hydrologic
When first calibratinga model,assumptions are made models,we candevelopmethodsthatenhancethe abilities
aboutinitial valuesof inputparameters.If necessary, input of both. In thisway, estimatesof climaticchangescanbe
parametersare varied until the fit betweenpredictedand incorporatedinto models of watershedsto evaluate
observeddata is consideredsatisfactory.The dependence possiblealterationsin water availabilityand vulnerable
of a model on its initial calibrationhas particularsignifi- pointsof existingwater supplies.And this techniquewill
cance for those circumstances where a climatic change become even more important as GCMs improve their
significantlyaffectsthe underlyingpredefinedparameters resolutionandtheirhydrologicparameterizations.
of a model. Under theseconditions,suchas a changein Models designedto evaluatethe impactsof climatic
the extentof vegetativecover,the initial calibrationof the changes onrunoffmustbe ableto (1) reproducereasonably
model is no guaranteethat the model can be usedaccu- well the historicalstreamflowrecord and (2) simulatethe
rately for evaluating conditions following a climatic streamflow under climatic conditions that are different
change. from the conditions for which the model has been
calibrated. There are extensive discussions in the literature
The Availabilityof Input Data, IncludingComparative about the ways in which hydrologic models can be
Historical Data validated [Chow, 1964; Viessman et al., 1977; WMO,
Historicaldata must be availableboth to calibrateany 1975; Klemeg, 1985]. Almost all of these, however, deal
givenmodelandto evaluatetheaccuracyandapplicability with validatingmodelsunder stationaryclimatic condi-
of a modelto any givenwatershed.Unlessa modelcanbe tions. When the goal is to investigateeffectsof climatic
validated using actual historical records, its value forchanges,additionaleffort must be made to extend the
estimating the effects of changesin climate must be validationto conditionsof nonstationaryclimates. Strictly
questioned. speaking,this type of validationis not possibleuntil the
climaticchangesactuallyoccur and the "experimentis
The Accuracyof InputData done." There are, however, teststhat can be applied to
The qualityof the inputdatais alsoimportantfor model provideat leastsomemeasureof confidencethattheinitial
validationand for the initial choiceof modelparameters. modelcalibrationis still valid. Among thesetestsis the
Hydrologic measurementsare frequently inaccurate, differentialsplit sampletest, which is appliedto a model
leadingto complicationsin achievinga good fit during wheninitial conditionsare to be changed.In thiscase,two
modelcalibration.Adjustinginputparameters to meetbad periods with different historicalconditionsare chosen,
datawill resultin biasesin subsequentmodelruns. suchas a periodof highaverageprecipitationanda period
of low averageprecipitation.For example,if the climatic
ModelFlexibility,Easeof Use,andAdaptability changeto be modeledis a transitionto a warmer,wetter
Modeling techniquesdesignedfor one hydrologic scenario,the modelshouldbe calibratedon a dry, cooldata
regimemay not be directlyapplicableto differentwater- set and then validated for the other extreme. A variation
sheds,becauseof variationsin the timing, nature, and on this testis to calibratethe modelon averageconditions
27, 3/REVIEWS OF GEOPHYSICS Gleick: CLIMATECHANGE AND HYDROLOGY ß 337

and verify that the relativeerrorsduringdry periodsand circulationmodels [Cohen, 1986; Gleick, 1986b, 1987a, b;
wet periodsareof similardistribution andmagnitude. Flaschka et al., 1987; Bultot et al., 1988; U.S. Environ-
What is the basisfor credibilityof a regionalhydrologic mentalProtectionAgency,1988]. Table 2 lists someof
model?Klemel [1985,p. 9] summarizes, thosestudiesthat usedgeneralcirculationmodelscenarios
of climatechangeas inputto regionalhydrologicmodels.
A hydrologicalsimulationmodelmustdemonstrate, beforeit
Cohen [1986] evaluatedthe implicationsof general
is usedoperationally,how well it canperformthekind of task
for which it is intended. Since no simulation model is circulationmodel temperatureand precipitationscenarios
for water levels in the Great Lakes. The net basin water
intendedmerely to showhow well it fits the datausedfor its
development,performancecharacteristics derived from the supplyof the Great Lakes was predictedto decline in
calibration data set are insufficient evidence for its satisfac- responseto greenhousewarming over a wide range of
tory performance. The data used for model validation climatescenarios,includingsomewith large increasesin
(verification)must representa situationsimilar to that in
precipitation,althoughthe overallresultswere sensitiveto
which the model will be used, as it is recognizedthat data
from the actual situation to be simulated are not available. changesin wind speed and other assumptionsabout
variablesthat may alter lake evaporationrates. Using
Differentmethodshavebeenusedin the pastfew years temperatureand precipitationdata from two GCMs (the
to evaluateclimate-inducedchangesin water availability GeophysicalFluid DynamicsLaboratorymodel and the
(see Dooge [1986], Betart [1986], Gleick [1986b], and GoddardInstitute for Space Studies model), net basin
Changnon [1986b] for summaries of methods and supplydecreased by between15% and30%.
approaches),including both conceptualand stochastic In a studyof the SacramentoBasinin California,Gleick
models. In one of the earliest comprehensivestudies, [1986a, b, 1987a, b] identified hydrologicimpactsthat
Stocktonand Boggess[1979] usedthe empiricalrelation- were robust and consistentover a wide range of both
shipsderivedby Langbeinet al. [1949] to predictchanges hypothetical and GCM-generated climate change
in runoff for the 18 water resourceregionsof the coter- scenarios. These include large decreasesin summersoil
minous United States. Using only hypotheticalannual moisture levels, decreases in summer runoff volumes,
averagechanges,they concludedthat warmer and drier major shifts in the timing of average monthly runoff
shiftsin climatewould be the mostproblematicfor water throughoutthe year, and large increasesin winter runoff
availability. Revelle and Waggoner[1983] also usedthe volumes. Using eight differentgeneralcirculationmodel
empirical relationshipsof Langbein et al. to evaluate temperatureand precipitationscenarios,including both
hypotheticalclimate changescenariosfor the Colorado increasesand decreasesin averageprecipitation,summer
River Basin. They concludedthat annualrunoff in this runoff decreasedby between30% and 68%, winter runoff
increasedby 16% to 81%, and summer soil moisture
regionis especiallysensitiveto changesin temperature.In
one of the first regional studies,N•mec and Schaake decreased by 14% to 36% (Figures4a-4c). The principal
[1982] useda moredetaileddeterministicapproachto look mechanismdriving thesehydrologiceffectsis a dramatic
at the impactof hypotheticaltemperature changein snowfalland snowmeltconditions. Becauseof
andprecipitation
scenarios in an arid and a humid watershed. higher temperaturesa greater fraction of annual and
More recently,a numberof studieshave offered new seasonalprecipitationfalls as rain, rather than as snow.
insights into hydrologic vulnerabilitiesto greenhouse This hasthe effect of reducingtotal annualsnowpack.In
warming. These studies involve more accurate and addition,the precipitationthat is receivedas snowbegins
comprehensiveregional water balance models and to meltearlierin the springandmeltsfaster,leadingto less
incorporateclimate scenariosdeveloped from general spring and summer snowmelt runoff and decreasesin

TABLE 2. Hydrologic StudiesUsing GCM Output

Study GISS GFDL NCAR OSU

Gleick [1987b] X X X
Cohen [1986] X X
Mather and Feddema [1986] X X
Sandersonand Wong [1987] X
Bultot et al. [1988] X
U.S. Environmental Protection Agency
[1989] X X X

Different GCM runs were used by different researchers.Thus the resultsof each
studymust be reviewedindividually. Pleaserefer to individualstudiesfor detailsof
methodologies and basin-specificassumptions.Abbreviationsare definedas follows:
GISS: Goddard Institute for Space Studies,New York; GFDL: GeophysicalFluid
DynamicsLaboratory,Princeton, New Jersey;NCAR: NationalCenterfor Atmospheric
Research,Boulder, Colorado;andOSU: OregonStateUniversity,Corvallis.
338 ß Gleick: CLIMATE CHANGE AND HYDROLOGY 27, 3 / REVIEWSOF GEOPHYSICS

Percent Change (Summer Runoff) Figure 4. Changesin seasonalityof runoff and

lOO
soil moisture in the Sacramento Basin, Califor-
80 nia, due to climatic change[from Gleick, 1987b].
60 (a) The percentchangein averagesummer(June,
40
July, and August) runoff between the model
historicalrun ("base case") and eight GCM-
20 derived doubled carbon dioxide scenarios is
N.I.
0 shown. Note that all scenarios show a decrease

-20 in summerrunoff. (b) The sameplot is shown

for averagewinter runoff (December,January,
-40 -30
-40 andFebruary).Note thatall scenarios showlarge
-60 -48
-48 -53 increasesin winter runoff. (c) The percent
-80 change in average summer (June, July, and
- 100
I August)soil moisturebetweenthe basecaseand
T Only T,P(r) T,P(a) eight GCM-derived scenariosis also illustrated.
All scenariosshow large decreasesin summer
soil moisture. These changesare driven by an
Percent Change (Winter Runoff)
lOO increasein the ratio of rain to snowin winter,by
81
80-
a faster and earlier spring snowmelt, and by
66
increased summer evapotranspiration[Gleick,
60
1987a, b]. Abbreviationsare defined as follows:
38
40 33 NCAR: National Center for AtmosphericRe-
26
20
16 search; GFDL: GeophysicalFluid Dynamics
Laboratory;GISS: GoddardInstitutefor Space
0
Studies;T only: temperaturechangesonly; T,
-20 - P(r): temperature andrelative(percent)changein
-40 precipitation;T, P(a): temperatureand absolute
-60
changein precipitation;and N.I.: NCAR relative
precipitationrun not included.
-80
I I
-100
T Only T,P(r) T,P(a)

Percent Change (Summer Soil Moisture)

100

8O

60

4O

20
N.I.
0

-20
I NCAR
-20
-4O - -28
-33 -31 -36
-60 • GFDL
-80

- 100
[ I I • GISS
T Only T,P(r) T,P(a)

summer soil moisture. This effect has now been noted in [1985]. Second, in watersheds with a seasonal snowfall
otherregions[Shiklamanov,1987;Bultotet al., 1988; U.S. and snowmeltpatternthe seasonaldistributionof runoff
EnvironmentalProtection Agency, 1988; Martinec and and soil moistureis more sensitiveto temperaturethanto
Rango, 1989] and has been identified in GCM results precipitation[Gleick, 1987b; Bultot et al., 1988; U.S.
[Manabe and Wetheraid, 1986; Wilson and Mitchell, Environmental Protection Agency, 1988]. In these
1987]. watersheds,higher temperaturesreducethe ratio of snow
Two other resultsare particularlynoteworthy. First, to rain during the winter, hasten the onset of spring
annualrunoff appearsto be more sensitiveto changesin snowmelt, and increase the rate of snowmelt runoff, as
precipitationthan to changesin temperature[Gleick, described above.
1987b; Flaschka et al., 1987; Karl and Reibsame, 1989], The results of these regional studies suggestthat
an effect describedtheoreticallyby Wigley and Jones physicallybasedmodelsare able to provideconsiderable
27, 3/REVIEWS OF GEOPHYSICS Gleick: CLIMATECHANGE AND HYDROLOGY ß 339

informationon the regionalhydrologiceffectsof climatic evaluatedreservoirreliability under conditionsof altered

changes,despiteuncertaintiesaboutmanyregionaldetails runofffor a humidand an add watershed.They concluded
of the future climate. Such informationhas important that increasedstoragemay be neededto offset the effects
ramificationsfor long-rangewater resourceplanning,for of climatic changes,if traditionaldefinitionsof reliable
agriculturalwater developmentand conservation,and for yieldsare to be maintained. More recently,Dracup and
industrial water use over the next several decades. Kendall [1989] evaluatedhow changesin runoff due to
greenhouse warmingmay alter reservoirreliability or the
need for reservoircapacity. In their analysis,reservoirs
IMPACT ON WATER RESOURCES AND WATER designedand operatedunder currentclimatic conditions
MANAGEMENT will be severelystressedby climaticchanges.Of particu-
lar concern is that the volume of storage needed to
Ultimately, unless climatological changes can be maintainexistingwateryieldsmay changesignificantlyas
converted into estimates of how the availability of greenhouse warmingprogresses.
freshwaterresourcesmay change, water managersand
plannerswill be unwilling or unableto implementnew Water Quality
management policiesor to planfor new conditions. Coutant[ 1981] identifiedrisingtemperatureandreduc-
Freshwater resourcesare critical for both ecological tionsin streamflowdue to climaticchangesto be of pri-
servicesand human development. Among the critical mary importantfor water quality. Existingcomputer
issues are the quality of drinking water, supply for simulations of wastedegradation and oxygendynamics
industrial activities, agricultural water use, sewage canbe usedto quantifythe implications for waterquality
treatment,navigation,hydroelectricityproduction,and a of greenhouse
warming. Henderson-Sellers
[1987] devel-
wide range of environmentalservices,such as fisheries, opeda simplebox diffusionmodelto investigatethe ef-
waterfowl habitat, and wetlandspreservation. Table 3 fectsof increasing
atmospheric
CO2 uponreservoir
water
[fromNovakyet al., 1985] presentsthe sensitivityof water quality and suggestedthat increasedevaporativelosses
managementproblemsto climatic variability and change. from reservoirscouldlead to enhanced nutrientloading
andacceleratedeutrophication.
TABLE 3. The Sensitivityof Water Resourcesto Climatic
Variability Jacoby [1989] and the U.S. EnvironmentalProtection
Agency[1988] addressedthe issue of using existing
Sensitivityto ClimaticEvents
WaterResource
Management Within- Multi-
modelsfor identifyinghow to maintainsatisfactory
water
Methods& Techniques Year Annual Year Century qualityunderconditionsof changingclimate. In par-
ticular,if climaticchangesreducethe freshwateravailable
Protectionagainstfloods X X
River training X X X for dilutingwastewater,recharging
aquifers,or repelling
Drainage X X X salinityin estuaries,expensivewater quality protection
Water quality X X X X
Wastewater renovation X X measuresmay be required. Williams [1988] looked at the
Water supply X X X effectsof climaticchangeson salt waterpenetrationinto
River canalization (dams) X X
Storagereservoirs X X X the SanFranciscoBay and the Sacramento/San Joaquin
Groundwater utilization X X X Delta. He wamedthata greenhouse warmingmayrequire
Water transfers X X X
greaterreleasesof freshwaterfrom theupstream reservoirs
Soil moisturemanagement X
Erosion control X (so-called"carriagewater") during springand summer
low-flow periodsto flush salt water back towardthe San
FromNovaIcyet al. [1985]
FranciscoBay to protectwatersupplyintakesandnatural
While little research has been done in these areas, some ecosystems in the delta. The increased needfor carriage
recent findingsare describedbelow. Many other water water could threaten the amount of freshwater that can be
resourceproblemswill also arise, includingimpactson made available for southern California.
groundwaterwithdrawal and rechargeand effects on
islands dependenton shallow freshwater lenses and EnergyProduction
rainfall. Far moreresearchinto thesequestionsis needed. Hydroelectricityproductionis vulnerableto existing
climatic variability. Reductionsin annual runoff, in-
ReservoirReliability creasedinterannualvariability of drought episodes,and
An early paper on reservoir reliability by Schwarz more pronouncedseasonalityof flows would all affect
[1977] identified a range of likely changesand offered a hydroelectricproduction. In regions where significant
framework for identifying where the failure of existing fractionsof electricityare generatedwith water(seeFigure
operatingmethodsmight occur. N•mec and Schaake 5), a climatic change that reduceswater availability or
[1982] lookedat a setof hypotheticalclimatescenariosand changesthe seasonality of runoff may lead to the needfor
340 ß Gleick: CLIMATECHANGE AND HYDROLOGY 27, 3 / REVIEWSOF GEOPHYSICS

Figure 5. Regions of the United States dependenton climatic changesthat reduce the amountof water availablefor
hydroelectricity
for between20% and 100%of electricitysupply hydroelectricitygeneration.
[from Gleick, 1989]. Theseregionsare particularlysensitiveto

1'4

. , .
.

.o 12

2:

KEY TO REGIONS

1 NEW ENGLAND 7 UPPER MISSISSIPPI 12 TEXAS-GULF 17 PACIFIC NORTHWEST

2 MID-ATLANTIC 8 LOWER MISSISSIPPI 13 RIO GRANDE 18 CALIFORNIA
3 SOUTH ATLANTIC GULF 9 SOURIS-RED-RAINY 14 UPPER COLORADO 19 ALASKA
4 GREAT LAKES 10 MISSOURI 15 LOWER COLORADO 20 HAWAII
5 OHIO 11 ARKANSAS-WHITE-RED 16 GREAT BASIN 21 CARIBBEAN
6 TENNESSEE

expensivereplacementpower [Gleick, 1988, 1989;Miller, both rising evapotranspimtiondue to rising temperatures

1989]. In California,duringthe 1976-1977 drought,$500
million worthof fossilfuelswereconsumed to replacelost
hydroelectricity
supplies[Gleick,1988].
Miller [1989] examinedin more detail the potential
impact of altered water supplieson the electricpower
industry in the United States. She identified seasonal
reductionsin water availabilityas an importantthreatto
reliablehydroelectricitysupply. Anothermajorproblem
facing the energy industryis the availabilityof cooling
waterfor thermalpowerplants.Fortunately, theconsump-
tive useof waterfor thermoelectriccoolingis onlyaround
2% of the total consumptive use, and a wide varietyof
alternativecooling technologiescan be adoptedshould
futurewatershortages emerge.
IrrigationDemands decreaseby nearly 30%, unless efforts are made to
Hargreaves[1981] discussed the implications
of higher improvethe efficiencyof water use or to developnew
temperature
for agriculturalwaterdemand. He pointedto supplies.
anddecreasing precipitation
ascriticalfactorsfor identify-
ing climatic zones for agriculture. He also noted that
irrigationmaybecomeincreasingly importantto ensurethe
adequacyof watersuppliesfor cropproduction.
Peterson and Keller [1989] extended this work and
evaluatedthe implicationsof changesin temperature and
precipitation
for thenetirrigationrequirement, thedepthof
irrigationneededto maximizeproductionat a particular
location,and the futureirrigatedpercentof cultivatedland.
They concluded that a warmer climate will have its
greatestimpactin dry regions(suchas the westernUnited
States),whereirrigatorswill be hard-pressedto maintain
currentlevelsof irrigation. For their scenarioof a 3øC
temperatureincreaseand a 10% decreasein precipitation,
cultivated areas in the western United States could
27, 3 /REVIEWS OF GEOPHYSICS
Environmental Services
Coutant[1981] raisedthe questionof adversebiological
responses in estuariesdue to alteredfreshwaterflows and
increasedsalinityandrecommended the useof mathemati-
cal and biophysicalmodelsto studyestuaries,suchas the
Chesapeake Bay, the HudsonRiver, or the DelawareBay.
Josselynand Callaway [1988] look at the implicationsof
decreased freshwaterflows and changesin salinityfor the
rich ecosystemof the San FranciscoBay-San Joaquin
Delta regionand identifiedsomeadverseconsequences of
greenhousewarming for phytoplanktonproductivityand
fish breedingandsurvival.
Cooper [1989] noted that impacts on recreation,
wildlife, and estheticsare often the early warning signs
that somethingis wrong with our environment. He
identified a wide range of climatic changesthat would
adversely affect environmental services offered by
freshwaterresourcesand proposedactive measuresto
anticipateandalleviateanydetrimental consequences.
CONCLUSIONS
Global climatic changescausedby an increasedcon-
centrationof atmospheric tracegaseswill haveimportant
and diverseimplications for freshwater resources.While
we cannotseeclearlythe directionor magnitudeof many
importantchanges, research in thepastfew yearshashigh-
lightedimportantregionalvulnerabilities andhasimproved
our understanding of appropriatemethodsfor addressing
hydrologicandwaterresourcequestions.
In thelongrun,large-scale generalcirculationmodelsof
theclimatemay be ableto give usvaluableinformationon
detailed,regionalimpactson watersupplies.But waiting
until sucha capabilityis availablemeanswaitinguntilcli-
mate impactsunambiguously begin to appear. For that
reason,hydrologists and water plannersare relyingon a
varietyof othermethodsto increaseour understanding of
climatic vulnerabilities,including reviewing the paleo-
climateand morerecentinstrumentalrecordsand usingre-
gionalhydrologicmodelsto explorea wide rangeof cli-
matechangescenarios.Thesetechniques haveidentifieda
numberof importantproblemsthatmaysoonfaceus.
Some of the regional resultsdescribedhere support
recentsuggestions thatmid-latitudesummersoil moisture
reductions may occurin manyregionsof the world. The
principalphysicalmechanisms involved(the decreasein
snowas a fractionof total winterprecipitation,an earlier
and fasterdisappearance of winter snowpackbecauseof
higher average temperatures, and a more severe
evapotranspiration demand during the warmer summer
months)are both physicallyplausibleand consistent with
the hydrologicmechanisms that lead to regionalsummer
soil moisturedrying in the GCMs. While other,counter-
vailinghydrometeorologic featuresmay well exist,suchas
Gleick: CLIMATE CHANGE AND HYDROLOGY ß 341
cloudcover/evapotranspiration
feedbacks,the consistency
of the soil moisture and runoff results observed here must
be considereda first warning of possible important
changesin regionalwater availability. As more informa-
tion on theseotherfactorsdevelops,it can be incorporated
into hydrologicmodelsto providemore derailedregional
assessments.
Ultimately,if realisticestimatesof changesin regional
water availability are to be calculated, a number of
advances are needed. In order to be valuable to water
resourceplanners,regional hydrologicassessments must
include(1) a focuson shorttime scalessuchas daysor
weeks,ratherthan annualor even monthlyaverages;(2)
the ability to incorporateinto regionalstudiesthe increas-
ingly detailedassessments of changesproducedby GCMs;
(3) the ability to produceinformationon hydrologically
important variables, such as changes in runoff and
availablesoil moisture;and (4) the ability to incorporate
snowfalland snowmelt,vegetationchanges,topography,
soil characteristics,
naturalandartificialstorage,aridother
regionalcomplexities.
Finally, the long constructiontimes and subsequent
lifetimes of reservoirs, dams, and water transfer facilities
meanthat planningshouldbegin todayfor changesthat
may not becomeevidentfor years. Yet changesin water
resourcesmanagementand planning will only come if
thoseresponsiblefor our water systemscan be convinced
thatthe problemof climaticchangeis sufficientlyreal and
pressingto require their attention. Much work remainsto
be done.
ACKNOWLEDGMENTS. GarrisonSpositowasthe editorin
chargeof thispaper. He expresses
gratitudeto J. Dracup,K. M.
Loague,and R. T. Wetheraidfor assistance in evaluatingits
technicalcontentand to W. Duffield for servingas a cross-
disciplinaryreferee.
REFERENCES
Adam, D. P., and G. J. West, Temperatureand precipitation
estimatesthrough the last glacial cycle from Clear Lake,
California,pollen data,Science,219, 168-170, 1983.
American Associationfor the Advancementof Sciences,Climate
and Water: Climatic Variability, Climate Change,and the
PlanningandManagement of U.S.WaterResources, editedby
P. Waggonerand R. R. Revelle,JohnWiley, New York, in
press,1989.
Aristarain, A. J., J. Jouzel, and M. Pourchet, Past Antarctic
peninsulaclimate (1850-1980) deduced from an ice core
isotoperecord,Clim. Change,8, 69-90, 1986.
Ball, T. F., andR. A. Kingsley,Instrumental
temperature
records
at twositesin centralCanada:1768to 1910,Clim.Change,6,
39-56, 1984.
Barnola, J. M., D. Raynaud,Y. S. Korotkevich,and C. Lorius,
Vostokice coreprovides160,000-yearrecordof atmospheric
C02,Nature,329,408-418,1987.
Beran,M., The waterresourceimpactof futureclimatechange
andvariability,in Effectsof Changesin Stratospheric
Ozone
and Global Climate,vol. 1, Overview,editedby J. G. Tires,
pp.299-330, UnitedStatesEnvironmental ProrationAgency,
Washington,D.C., 1986.
342 ß Gleick: CLIMATE CHANGE AND HYDROLOGY 27, 3 / REVIEWSOF GEOPHYSICS

Bolin, B., R.. Doos, J. Jager, and R. A. Warrick (Eds.), The Gleick, P. H., Methods for evaluatingthe regional hydrologic
greenhouseeffect, climatic change,and ecosystems, SCOPE impactsof global climatic changes,J. Hydrol., 88, 99-116,
Rep.,29, 541 pp. 1986. 1986b.
Bradley, R. S., H. F. Diaz, J. K. Eischeid,P. D. Jones,P.M. Gleick, P. H., The developmentand testingof a water balance
Kelly, and C. M. Goodess,Precipitationfluctuationsover model for climate impacts assessment: Modeling the
northernhemisphereland areassince the mid-19th century, Sacramento Basin, Water Resour. Res., 23, 1049-1061,
Science,237, 171-175, 1987. 1987a.
Budyko,M. I., and Y. A. Izrael (Eds.),Anthropogenic Climatic Gleick, P. H., Regionalhydrologicconsequences
of increasesin
Change(in Russian),406 pp., Gidrometeoizdat,
Leningrad, atmospheric
CO2 andothertracegases,Clirn.Change,10,
1987. 137-161, 1987b.
Bultot, F., A. Coppens, G. L. Dupriez, D. Gellens, and F. Gleick,P. H., Climatechangeand California:Past,present,and
Meulenberghs, Repercussions of a CO2doubling onthewater future vulnerabilities,in Societal Responsesto Regional
cycle andon the waterbalance:A casestudyfor Belgium,J. ClimaticChange: Forecastingby Analogy,editedby M. H.
Hydrol., 99, 319-347, 1988. Glantz,pp. 307-327, WestviewPress,Boulder,Colo., 1988.
Camuffo, D., Analysis of a seriesof precipitationat Padova, Gleick, P. H., Vulnerabilitiesof water supply, in Climate and
Italy, Clim. Change,6, 57-78, 1984. Water: Climatic Variability, Climate Change, and the
Changnon,S. A., Jr., Climatefluctuationsandrecord-high levels PlanningandManagement of U.S.WaterResources,editedby
of Lake Michigan, Bull. Am. Meteorol. Soc., 68(11), P. Waggonerand R. R. Revelle, JohnWiley, New York, in
1394-1402, 1987a. press,1989.
Changnon,S. A., Jr., An assessment of climatechange,water Goodfriend,G. A., M. Magarits,and I. Carmi, A high standof
resources,andpolicy research,Water Int., 12, 67-76, 1987b. theDeadSeaat the endof theNeolithicperiod: Paleoclimatic
Chow, V. T. (Ed.), Handbookof AppliedHydrology,McGraw- and archeologicalimplications,Clirn. Change,9, 349-356,
Hill, New York, 1964. 1986.
Cohen,S. J.,Impacts
of CO2-induced
climaticchange
onwater Hansen,J. E., The greenhouse
effect: Impacton currentglobal
resourcesin the Great Lakes Basin, Clim. Change, 8, temperatureand regionalheat waves, statementto the United
135-153, 1986. States Senate, Comm. on Energy and Natur. Resour.,
Cohen,S. J., GreatLakeslevelsand climatechange: Impacts, Washington,D.C., June23, 1988.
responses,and futures, in Societal Responsesto Regional Hansen,J., G. Russell,D. Rind, P. Stone, A. Lacis, S. Lebedeff,
ClimaticChange: Forecastingby Analogy,editedby M. H. R. Ruedy,and L. Travis, Efficient three-dimensional
global
Glantz,pp. 143-168, WestviewPress,Boulder,Colo., 1988. models for climate studies: Models I and 1I, Mon. Weather
Cooper,C. F., Recreation,esthetics,and wildlife, in Climate and Rev., 111(4), 609-662, 1983.
Water: Climatic Variability, Climate Change, and the Hansen,J. E., A. Lacis,D. Rind,G. Russell,P. Stone,I. Fung,R.
PlanningandManagement of U.S.WaterResources,editedby Ruedy, and J. Lerner, Climatic sensitivity: Analysisof
P. Waggonerand R. R. Revelle, JohnWiley, New York, in feedback mechanisms, in Climate Processes and Climate
press,1989. Sensitivity,Geophys.Monogr.Ser., vol. 29, editedby J. E.
Coutant,C. C., Foreseeable
effectsof CO2-induced
climatic HansenandT. Takahashi,pp. 130-163, AGU, Washington,
change: Freshwaterconcerns,Environ. Conserv.,8, 285-297, D.C., 1984.
1981. Hansen,J.,I. Fung,A. Lacis,D. Rind,S. Lebedeff,R. Ruedy,G.
Dickinson, R. E., Modeling evapotranspiration
for three- Russell,andP. Stone,Globalclimatechanges as forecastby
dimensionalglobal climatemodels,in ClimateProcessesand GoddardInstitutefor SpaceStudiesthree-dimensional
model,
ClimateSensitivity,
Geophys.
Monogr.Ser.,vol. 29, editedby J. Geophys.Res.,93, 9341-9364, 1988.
J. E. HansenandT. Takahashi,
pp. 58-72, AGU, Washington, Hargreaves,G. H., Water requirementsandman-inducedclimate
D.C., 1984. change,Proc. Am. Soc.Civ. Eng., 107(IR3), 247-255, 1981.
Dooge,J.C. I., Effects
of CO2increases
onhydrology
andwater Henderson-Sellers,
B., The impact of increasingatmospheric
resources,paperpresented at Symposium
onCO2 andOther carbondioxideconcentrations
uponreservoirwater quality,
Greenhouse Gases: Climatic and AssociatedImpacts,Eur. The Influenceof ClimateChangeandClimaticVariabilityon
Econ.Community,Brussels,November4, 1986. the HydrologicRegimeandWaterResources, editedby S. I.
Dracup, J. A., and D. R. Kendall, Floods and droughts,in Solomon, M. Beran, and W. Hogg, IAHS Publ., 168,
Climateand Water: ClimaticVariability,ClimateChange, 571-576, 1987.
and the Planningand Managementof U.S. Water Resources, Jacoby, H., Water quality, in Climate and Water: Climatic
editedby P. WaggonerandR. R. Revelle, JohnWiley, New Variability,ClimateChange,and the PlanningandManage-
York, in press,1989. mentof U.S. WaterResources, editedby P. WaggonerandR.
Fitzgerald,B. M., andM. A. Walsh,Significance of greenhouse R. Revelle, JohnWiley, New York, in press,1989.
changesto irrigationwater suppliesin New SouthWalesmA Jones,P. D., Riverflowreconstruction
from precipitation
data,J.
casestudyof theSevernValley, report,16 pp., Dep. of Water Climatol., 4, 171-186, 1984.
Resour.,N. Sydney,N.S.W., Aust., 1987. Josselyn,M., and J. Callaway, Ecologicaleffects of global
Flaschka,I. M., C. W. Stockton,andW. R. Boggess,Climatic climatechange: Wetlandresources of San Francisco Bay,
variation and surface water resources in the Great Basin report, Romberg Tiburon Cent. for Environ. Stud., San
region,WaterResour.Bull., 23(1), 47-57, 1987. FranciscoState Univ., San Francisco,Calif., 1988.
Fritts,H. C., andJ. M. Lough,An estimateof averageannual Karl, T., andW. Reibsame,The impactof decadalfluctuations
in
temperaturevariationsfor North America, 1602 to 1961, Clim. meanprecipitation andtemperature on runoff: A sensitivity
Change,7, 203-224, 1985. studyovertheUnitedStates,Clim. Change,in press,1989.
Glantz, M. H., and J. H. Ausubel,The OgallalaAquifer and Kates,R. W., J. H. Ausubel,and M. Berberian(Eds.), Climate
carbon dioxide: Comparisonand convergence,Environ. impactassessment, SCOPERep.,27, 625 pp., 1985.
Conservation, 11, 123-131, 1984. Klemeg,V., Sensitivityof water resourcesystemsto climate
Gleick, P. H., Regionalwater availabilityand globalclimatic variations,Worm Clim. Appl. ProgrammeWCP-98, World
change: The hydrologic consequences of increasesin Meteorol.Organ.,Geneva,1985.
atmospheric CO2 and other trace gases,Ph.D. thesis, Langbein,W. B., et al., Annualrunoff in the United States,U.S.
ERG-DS-86-1,688pp., EnergyandResour.Group,Univ. of Geol. Surv. Circ., 5, 1949.
Calif., Berkeley,1986a. Linsley,R. K., Jr.,M. A. Kohler,andJ. L. H. Pauhlus,
Hydrol-
27, 3 /REVIEWS OF GEOPHYSICS Gleick: CLIMATE CHANGE AND HYDROLOGY ß 343

gyfor Engineers,McGraw-Hill, New York, 1982. Pearman, G. I. (Ed.), Greenhouse: Planning for Climate
Manabe, S., Climate and the ocean circulation, I, The atmos- Change, 752 pp., Division of Atmospheric Research,
pheric circulationand the hydrologyof the Earth's surface, Commonwealth
Scientificand IndustrialResearchOrganiza-
Mon. WeatherRev., 97(11), 739-774, 1969a. tion (CSIRO), CSIRO Publications,East Melbourne, Aust.,
Manabe, S., Climate and the ocean circulation, II, The atmos- 1988.
pheric circulationand the effect of heat transferby ocean Peterson,D. F., and A. A. Keller, Irrigated agriculture,in
currents,Mon. WeatherRev., 97(11), 775-805, 1969b. Climateand Water: ClimaticVariability,ClimateChange,
Manabe, S., and R. J. Stouffer, Sensitivityof global climate and the Planningand Managementof U.S. Water Resources,
modelto anincrease of CO2 concentration in theatmosphere, editedby P. Waggonerand R. R. Revelle, JohnWiley, New
J. Geophys.Res.,85, 5529-5554, 1980. York, in press,1989.
Manabe, S., and R. T. Wetheraid, On the distribution of climate Ramanathan,V., The role of ocean-atmosphere
interactionsin
changeresultingfrom an increase
in CO2-content
of the the CO2climate
problem,
J. Atmos.
Sci.,38, 918-930,1981.
atmosphere,
J. Atmos.Sci., 37, 99-118, 1980. Revelle, R. R., and P. E. Waggoner, Effects of a carbon
Manabe,S., andR. T. Wetheraid,
CO2 andhydrology,
Adv. dioxide-inducedclimatic change on water suppliesin the
Geophys.,28A,131-157, 1985. western United States, in Changing Climate, edited by
Manabe, S., and R. T. Wetheraid, Reduction in summer soil National Academy of Sciences, pp. 419-432, National
wetness induced by an increase in atmosphericcarbon AcademyPress,Washington,D.C., 1983.
dioxide, Science,232, 626-628, 1986. Rowntree,L. B., A crop-basedrainfall chronologyfor pre-
Manabe,S., R. T. Wetherald,andR. J. Stouffer,Summerdryness instrumental
recordin southernCalifornia,Clim. Change,7,
dueto an increase
of atmospheric
CO2 concentration,
Clim. 327-342, 1985.
Change,3, 347-386, 1981. Sanderson,M., and L. Wong, Climatic changeand Great Lakes
Martinec, J., and A. Rango, Effects of climate change on water levels,The Influenceof Climate Changeand Climatic
snowmeltrunoff patterns,Remote Sensing and Large-Scale Variability on the Hydrologic Regime and Water Resources,
Global Processes,Proceedingsof the IAHS Third Interna- editedby S. I. Solomon,M. Beran,andW. Hogg,IAHS Publ.,
tional Assembly, Baltimore, Maryland, IAHS Publ., 186, 168, 477-487, 1987.
1989. Schaake, J. C., From climate to flow, in Climate and Water:
Mather, J. R., and J. Feddema, Hydrologic consequencesof ClimaticVariability,ClimateChange,and the Planningand
increases
in tracegasesandCO2 in theatmosphere,
in Effects Managementof U.S. WaterResources, editedby P. Waggoner
of Changesin StratosphericOzoneand Global Climate, vol. andR. R. Revelle,JohnWiley, New York, in press,1989.
3, pp. 251-271, U.S. EnvironmentalProtectionAgency, Schaake,J. C., and Z. Kaczrnarek,Climatevariabilityand the
Washington,D.C., 1986. design and operationof water resource systems,World
Meams, L., P. H. Gleick, and S. H. Schneider, Climate forecast- Climate Conference,Overview Pap. 12, 23 pp., World
ing in Climate and Water: Climatic Variability, Climate Meteorol. Organ., Geneva, 1979.
Change, and the Planning and Managementof U.S. Water Schlesinger, M. E., and W. L. Gates,The Januaryand July
Resources,edited by P. Waggonerand R. R. Revelle, John performanceof the OSU two-level atmosphericgeneral
Wiley, New York, in press,1989. circulationmodel, J. Atmos.Sci., 37, 1914-1943, 1980.
Miller, K. A., Water, electricity,and institutionalinnovation,in Schlesinger,M. E., and J. F. B. Mitchell, Model projectionsof
Climate and Water: Climatic Variability, Climate Change, the equilibriumclimaticresponseto increasedcarbondioxide,
and the Planningand Managementof U.S. Water Resources, Projectingthe ClimaticEffectsof IncreasingCarbonDioxide,
editedby P. Waggonerand R. R. Revelle,JohnWiley, New edited by M. C. MacCrackenand F. M. Luther, Rep.
York, in press,1989. DOE-0237,pp. 81-148, U.S. Dep. of Energy,Washington, D.
Mitchell, J. F. B., and D. A. Warrilow, Summer drynessin C., 1985.
northern mid-latitudesdue to increased CO2, Nature,330, Schlesinger,M. E., and J. F. B. Mitchell, Climate model
238-240, 1987. simulations of the equilibriumclimaticresponseto increased
Morgan, V. I., An oxygenisotope-climaterecordfrom the Law carbondioxideRev. Geophys.,25, 760-798, 1987.
Dome,Antarctica,Clim. Change,7, 415-426, 1985. Schneider,S. H., P. H. Gleick, and L. Mearns, Prospectsfor
Morrisette,P.M., The rising level of the Great Salt Lake: An climatechange,in Climateand Water: ClimaticVariability,
analogueof societaladjustmentto climate change,in Societal ClimateChange,and the Planningand Managementof U.S.
Responses to Regional Climatic Change: Forecastingby Water Resources,editedby P. Waggonerand R. R. Revelle,
Analogy, edited by M. H. Glantz, pp. 169-196, Westview JohnWiley, New York, in press,1989.
Press,Boulder, Colo., 1988. Schwarz,H. E., Climatic changeand water supply: How
National ResearchCouncil Panel on Water and Climate (Ed.), sensitiveis thenortheast?,in Climate,ClimaticChange,and
Climate, Climatic Change, and Water Supply, 132 pp., Water Supply,editedby National ResearchCouncil Panelon
NationalAcademyof Sciences, Washington,D.C., 1977. Water and Climate, pp. 111-120, National Academyof
NCmec,J., andJ. Schaake,Sensitivityof waterresourcesystems Sciences,Washington,D.C., 1977.
to climatevariation,Hydrol. Sci.J., 27(3), 327-343, 1982. Shiklamanov,I., Effects of climate changeson Soviet rivers,
Nicholson, S. E., and H. Flohn, African environmental and paperpresentedat the InternationalSymposiumof the XIXth
climatic changesand the generalatmosphericcirculationin General Assembly,Int. Union of Geod. and Geophys.
late Pleistoceneand Holocene,Clim. Change, 2, 313-348, (IUGG), Vancouver,B.C., Can.,Aug. 9-22, 1987.
1980. Snyder, C. T., and W. B. Langbein,The PleistoceneLake in
Novaky, B., C. Pachner,K. Szesztay,and D. Miller, Water Spring Valley, Nevada, J. Geophys.Res., 67, 2385-2394,
resources,Climate Impacts Assessment,edited by R. W. 1962.
Kates, J. H. Ausubel,and M. Berberian,SCOPE Rep., 27, Solomon,S. I., M. Beran,andW. Hogg (Eds.),The Influenceof
187-214, 1985. ClimateChangeand Climaticvariabilityon the Hydrologic
Palutikof, J.P., Some possibleimpacts of greenhousegas Regimeand Water Resources, IAHS Publ., 168, 640 pp.,
inducedclimate changeon water resourcesin Englandand 1987.
Wales, The Influence of Climate Change and Climatic Somerville,R. C. J., The predictabilityof weatherand climate,
Variability on the HydrologicRegime and Water Resources, Clim. Change,11,239-246, 1987.
editedby S. I. Solomon,M. Beran,andW. Hogg,IAHS Publ., Stockton,C. W., Interpretation
of pastclimaticvariabilityfrom
168, 585-596, 1987. paleoenvironmental
indicators,
in Climate,ClimaticChange,
344 ß Gleick: CLIMATE CHANGE AND HYDROLOGY 27, 3 / REVIEWSOF GEOPHYSICS

and Water Supply,editedby NationalResearchCouncilPanel Washington, W. M., andG. A. Meehl, Seasonalcycleexperiment

on Water and Climate, pp. 34-46, National Academy of ontheclimatesensitivity dueto a doubling of CO2 withan
Sciences,Washington,D.C., 1977. atmosphericgeneralcirculationmodel coupledto a simple
Stockton,C. W., and W. R. Boggess,Geohydrological implica- mixed-layeroceanmodel, J. Geophys.Res., 89, 9475-9503,
tions of climate change on water resourcedevelopment, 1984.
report,U.S. Army CoastalEng. Res. Cent., Fort Belvoir, Va., Wetheraid, R. T., and S. Manabe, Influence of seasonal variation
1979. uponthe sensitivityof a modelclimate,J. Geophys.Res.,86,
Stockton,C. W., and G. C. Jacoby,Jr., Long-termsurface-water 1194-1204, 1981.
supply and streamflowtrendsin the upper ColoradoRiver Wigley, T. M. L., and P. D. Jones,Influencesof precipitation
Basin basedon tree-ring analyses,Lake Powell Res. Proj. changes anddirectCO2 effectson streamflow, Nature,314,
Bull. 18, 70 pp., Univ. of Ariz., Tucson,1976. 149-152, 1985.
Studyof Critical EnvironmentalProblems,Man's Impact on the Wigley, T. M. L., J. M. Lough,andP. D. Jones,Spatialpatterns
Global Environment:Assessment and Recommendations for of precipitation in England and Wales and a revised
Action,319 pp., MIT Press,Cambridge,Mass.,1970. homogeneousEngland and Wales precipitationseries, J.
Study of Man's Impact on Climate, Inadvertent Climate Climatol., 4, 1-25, 1984.
Modification,308 pp.,MIT Press,Cambridge, Mass.,1971. Williams,P. B., The impactsof climatechangeon the salinityof
U.S. Environmental Protection Agency, Potential climatic San Francisco Bay, report 452 for U.S. Environmental
impactsof increasing atmospheric CO2 with emphasis on Protection Agency, Philip Williams and Associates,San
water availabilityand hydrologyin the United States,report, Francisco,Calif., 1988.
Strategic studies Staff, Office of Policy Anal., Office of Wilson,C. A., andJ. F. B. Mitchell,A doubled CO2 climate
Policy,Plann.andEval., Washington,D.C., 1984. sensitivityexperimentwith a globalclimatemodelincludinga
U.S. EnvironmentalProtectionAgency,The potentialeffectsof simpleocean,J. Geophys.Res.,92, 13,315-13,343, 1987.
global climate changeon the United States,draft report to World Meteorological Organization, Intercomparison of
Congress,Office of Policy, Plann.and Eval., Washington,D. conceptualmodelsused in operationalhydrologicalforecast-
C., Oct. 1988. ing, Oper.Hydrol. Rep. 7, Geneva,1975.
Velichko, A. A., Late Pleistocene spatial paleoclimatic World Meteorological Organization, Water resources and
reconstructions,in Late Quaternary Environmentsof the climatic change: Sensitivityof water-resources systemsto
Soviet Union, edited by A. A. Velichko, pp. 261-285, climatechangeand variability,Rep. WCAP-4, WMO/TD 247,
Universityof MinnesotaPress,Minneapolis,1984. 50 pp., 1987.
Viessman, W., Jr., J. W. Knapp, (3. L. Lewis, and T. E.
Harbaugh,Introductionto Hydrology, 2nd ed. Harper and
Row, New York, 1977.
Washington,W. M., andG. A. Meehl, Generalcirculationmodel
experimentson the climatic effects due to a doublingand P. H. Gleick, PacificInstitutefor Studiesin Development,
quadruplingof carbon dioxide concentration, J. Geophys. Environment, and Security,1681 ShattuckAvenue,SuiteH,
Res., 88, 6600-6610, 1983. Berkeley,
CA 94709.