3Climate Models and Their Projections of Future Ch

31BookID182985_ChapID3_Proof#1-10/10/2009AbstractThischapterdescribesglobalclimatemodelsandtheiroutput.Thecur-rentapproachesforanalyzingtheirsimulations,characterizingtherangeoflikelyfutureoutcomes,andmakingprojectionsrelevantforimpactanalysisaredescribed,specificallyreferringtothelatestassessmentreportoftheIntergovernmentalPanelonClimateChange.Weprovideasummaryoffutureprojectionsofaveragetem-peratureandprecipitationchangesatcontinentalscales,togetherwithabroadbrushpictureofthelikelychangesinindicesofextremes,characterizingbothtemperatureandprecipitationevents.Ananalysisofchangesingrowingseasonlengthisalsopresentedasanexampleofclimatemodeloutputanalysisdirectlyrelevanttostud-iesofclimatechangeimpactsonfoodsecurity.3.1WhereDoClimateChangeProjectionsComefrom?Humansareconductinganunprecedented,deliberateyetuncontrolledexperimentusingourplanetasitssubject.Human-inducedemissionsofgreenhousegasesandotherpollutants,togetherwithchangesinlanduse,likedeforestation,arealteringourclimatesystempropertiesinwaysthatarealreadydetectable(Hegerletal.2007).Theexperimentiscontinuing,withfutureemissionsprojectedtosteadilyraisethegreenhousegasconcentrationsintheatmosphere.Thisisbecausegreen-housegaseslikeCO2,unlikeothergasspecieshavealonglifemeasuredindecadesandcenturies,sothatemissionsovertheyearsaccumulateandincreasinglyalterthenaturalstateofthesystem.Chapter3ClimateModelsandTheirProjectionsofFutureChangesClaudiaTebaldiandRetoKnuttiD.LobellandM.Burke(eds.),ClimateChangeandFoodSecurity,AdvancesinGlobalChangeResearch37,DOI10.1007/978-90-481-2953-9_3,©SpringerScience+BusinessMedia,B.V.200C.Tebaldi()ClimateCentral,Princeton,NJ,USAemail:[e-mail]hericandClimateScience,ETH(SwissFederalInstituteofTechnology),Zurich,Zwitzerland132C.TebaldiandR.KnuttiBookID182985_ChapID3_Proof#1-10/10/2009BookID182985_ChapID3_Proof#1-10/10/2009Becausewearechangingthenaturalclimatestatelikeneverbefore,itwouldbeunreliabletosimplyextrapolatecurrenttrendsintothefutureinordertopredictwhatwewillexperienceasaresult.Thisisparticularlytruewhenwefocusonregionalchanges,whicharemostimportantfordevisingadaptationmeasures.Theinteractionsandreactionsofthesystemaretoocomplicatedtobeapproximatedbystatisticalmodels.Infact,aswewillsee,theyareoftencomplicatedenoughtopresentachallengeevenforprocess-based,dynamicalclimatemodels.Rather,climatescientistsusenumericalmodelstoconstructsurrogatesoftherealsystem,inordertoperformacontrolled,andreplicable,versionoftheexperiment.Inthisfashiontheycantestdifferentassumptionsinfutureanthropogenicemissionsandotherparametersregulatingtheclimatesystem,spanawiderangeofuncertaintyatleastwithregardtotheknownunknowns,andthusofferarangeofclimatechangescenariosattemptingtospanasubstantialportionoftherelevantuncertainties.Thereexistsahierarchyofclimatemodels,fromsimpleenergy-balancemodelsthatcanonlyapproximatethetrajectoryofglobalmeantemperaturetomodelsofintermediatecomplexity(Claussenetal.2002)thatcanonlyresolveverylargeregions,toglobalcoupledmodels,whicharethesubjectofthischapter.Theseextremelycomplexcomputermodels,alsocalledatmosphere–oceangeneralcircu-lationmodels(GCMs),dividethesurfaceoftheEarth,thedepthsoftheoceansandthelayersoftheatmosphereintogridboxes.TheseGCMsdescribetheevolu-tionofahostofclimatevariablesateachgridboxandforvarioustimesteps(betweenafewminutesandanhour)bysolvingdifferentialequationsderivedfromwell-establishedphysicallaws,suchasconservationofenergyandangularmomentum.Inthetypicalclimatechangeexperimentthesimulationstartsfromconditionsrepresentativeoftheclimateofpre-industrialtimes(around1850),andisper-formedbylettingthesystemevolveaccordingtothelawsofphysics,undisturbed(i.e.notprescribinganyobservedchanges),exceptforso-calledexternalforcingstothesystem.Someoftheseexternalforcingsoccurnaturally,likechangesinsolarirradiance(the11-yearsolarcycleforexample)orvolcaniceruptions,whichmaybeenergeticenoughtospewlargequantitiesofaerosolsinthestratosphere.Thevolcanicdustactsasareflectivecloud,partiallyshieldingthesurfaceoftheEarthfromincomingradiationandthushavingashort-livedcoolingeffectontheorderofafewyears.Particularlyimportantinclimatechangeexperimentsareincreasingatmosphericconcentrationsofgreenhousegases,whichareanotherformofexternalforcing,butanthropogenicratherthannatural.Theyareimposedaccordingtostandardscenariosagreeduponbythescientificcommunity,reflectinghypothesesaboutthefutureevolutionofsocio-economic,technologicalandpoliticalfactors.Climatemodelsimulatedchangesarethereforetermedprojectionsratherthanpredictions,becausetheyareusuallyconditionalontheassumedstorylineorscenario.Thesystemrespondstotheseprotractedanthropogenicforcingsbyalteringitsbehaviorinatrend-likefashion,ratherthanbycyclicalorepisodicchangeswhicharetypicallytheresultofnaturaldisturbances.ThesechangescanbeassessedbyanalyzingtheoutputofaGCMexperimentwhichistypicallyatleasttwoandahalfcenturies333ClimateModelsandTheirProjectionsofFutureChangesBookID182985_ChapID3_Proof#1-10/10/2009long,producingsimulationsofclimatefr