China Economic Review xxx (xxxx) xxx–xxx ContentslistsavailableatScienceDirect China Economic Review journal homepage: www.elsevier.com/locate/chieco Climate change impacts on China's agriculture: The responses from market and trade Wei Xiea, Jikun Huanga,⁎, Jinxia Wanga, Qi Cuia,c,⁎⁎, Ricky Robertsonb, Kevin Chenb aChinaCenterforAgriculturalPolicy,SchoolofAdvancedAgriculturalSciences,PekingUniversity,Beijing100871,China bInternationalFoodPolicyResearchInstitute,Washington,DC20005-3915,USA cBeijingKeyLabofStudyonSci-TechStrategyforUrbanGreenDevelopment,SchoolofEconomicsandResourceManagement,BeijingNormal University,Beijing100875,China ARTICLE INFO ABSTRACT Keywords: China'sfoodsecurityhasbeenfacingseveralchallenges,whicharelikelytobeworseneddueto Climatechange climatechange.Thepurposeofthispaperistoprovideanevidenceontheimpactsofclimate Foodsecurity changeonChina'sagriculture,withparticularattentiontothemarketandtraderesponses.Using Market projectedcropyieldchangesforChinaandits'maintradingpartnersunderchangingclimate,we Trade employ an agricultural partial equilibrium model (CAPSiM)and a linked national and global China equilibriummodel(CAPSiM-GTAP)toassesstheimpactsonfoodproduction,price,tradeand self-sufficiencyofChina.Ourresultsshowthatclimatechangewillhavesignificanteffectson crop production though with large differences among crops. Under the worst climate change scenarioRCP8.5,wheatyieldinChinaisprojectedtodeclineby9.4%by2050,whichisthe biggestyieldreductionamongthecrops.However,themarketcanalsorespondtotheclimate change,asfarmerscanchangeinputsinresponsetoreducedyieldsandrisingprices.Asaresult, production losses for most crops are dampened. For example, wheat production loss under RCP8.5reducestoonly4.3%duetomarketresponse.Theadverseimpactsoncropproduction willbefurtherreducedafteraccountingforthetraderesponseasfarmersadjustproductionto muchhigherpricesinthemoreseverelyaffectedcountries.Thepaperconcludesthatweneedto learnmorefromfarmerswhooptimizetheirproductiondecisionsinresponsetothemarketand tradesignalsduring climate change. Amajor policyimplication is that policymakers need to mainstreamthemarketandtraderesponsesintonationalplansforclimateadaptation. 1. Introduction China'sagricultureisexpectedtofacechallengesinthefuturemainlyduetorisingfooddemandandconstraintsoflandandwater resources.AlthoughChinahaslargelyensureditsfoodsecurityinthepast40years,ithasincreasinglyreliedoninternationalmarkets toensureitsfoodsupplysince2004(Ali,Huang,Wang,&Xie,2017;FAO,2017).Withincreasingpopulation,higherincomeand constraintsofresources,thepressureonChina'sfoodsecurityisgoingtoincreaseinthefuture.Huang,Wei,Cui,andXie(2017) predictedthatChina'soverallfoodself-sufficiencyislikelytofallfrom94.5%in2015toaround91%by2025. ClimatechangewilllikelyaggravatethechallengesChinafacesonitsfoodsecurityinthefuture.China'sannualaveragetem- peraturehasbeenrisingsignificantlyoverthepastsixdecadesandthewarmingtrendwillcontinueunderthefutureprojections(Cui, ⁎Correspondingauthor. ⁎⁎Correspondingauthor. E-mailaddresses:[email protected](J.Huang),[email protected](Q.Cui). https://doi.org/10.1016/j.chieco.2018.11.007 Received15December2017;Receivedinrevisedform21November2018;Accepted26November2018 1043-951X/ © 2018 Elsevier Inc. All rights reserved. Please cite this article as: Xie, W., China Economic Review, https://doi.org/10.1016/j.chieco.2018.11.007 W.Xieetal. China Economic Review xxx (xxxx) xxx–xxx Xie,&Liu,2018;Liang&Yan,2016;Meehl,2007;Nakicenovicetal.,2000).Itisgenerallyacceptedthatthemechanismofclimate changeaffectingChina'sagricultureismainlythroughrisingtemperatureandincreasingfluctuationinprecipitation(Wuetal.,2014; EditionCommitteeofChina'sNationalAssessmentReportonClimateChange,2015). TheimpactsofclimatechangeonChina'sagriculturehavebeenwidelystudiedintheliteraturethroughbiophysicalmodels(e.g., Li & Geng, 2013; Wang, Huang, & Yang, 2014; Xiong, Matthews, Holman, Lin, & Xu, 2007; Tao, Hayashi, Zhang, Sakamoto, & Yokozawa,2008;Xiongetal.,2009;Piaoetal.,2010).Forexample,Linetal.(2005)foundthatthenegativeimpactsofclimate changeonwheatyieldinChinacouldreachupto5.6–18.5%underA2scenario1by2020s.Similarly,Taoetal.(2008)suggestedthat ifthetemperatureincreasesby1°C,riceyieldwoulddeclineby6.1–18.6%evenafterconsideringtheadaptationmeasures.Xiong, Conway,Lin,andHolman(2009)predictedamoderatedecreaseinriceyieldintherangeof4.9–8.6%in2050s.Meanwhile,some otherstudiesalsoprovidetheevidenceonpositiveimpactsofclimatechangeonsomeofthecrops.Linetal.(2005),forexample, concludedthatirrigatedmaizeyieldwouldincreaseslightlyby2020underB2scenario1. Amajorlimitationofthebiophysicalmodelsforassessingclimatechangeimpactisthattheytendtooverestimateadverseimpacts ofclimatechangeonagriculture,astheyfailtoaccountfortheunderlyingbufferingcapabilityofeconomicsystem,whichthelater attainsthroughadjustmentsinproductioninputsandstructure.Forexample,Wang,Huang,andYang(2009)usedageneralequi- librium economic model to assess the climate change impacts on agriculture in China and found that the percentage decrease in productionofrice,wheat,andmaizein2030wouldbelowerthantheyieldchangespredictedbybiophysicalcropmodelers.Using the global general equilibrium model (AGLINK), Zhai, Lin, and Byambadorj (2009) also found that climate change would cause China'stotalcropproductiontodecreaseonlyslightly(0.2–0.5%)in2080.Someglobalstudiesonclimatechangedoexplicitlycover Chinawhileaccountingforendogenousresponseofmarkets(Calzadillaetal.,2013;Nelson,Valin,etal.,2014;Parry,Rosenzweig, Iglesias,Livermore,&Fischer,2004;Zhaietal.,2009;Zhouetal.,2017).Nevertheless,thesestudieseitherlackempiricallybased dataonyieldshocksformaincropsinChinaordonotapplythedetailednationaleconomicmodelforChinathatcanreflectChina's agriculturemarketaccurately. InternationaltradeisalsoanotherimportantfactoraffectingChina'sfoodmarketbutfewstudieshaveconsideredtheroleoftrade whileassessingclimatechangeimpactsonChina.Around2004,Chinaturnedanetimporterofagricultureproductsformpreviously anetexporter,somuchsothatin2016 > 80milliontons(Mt)ofsoybeanwasimported.Atthesametime,Chinabecametheworld's largestimporterofrice(Huangetal.,2017).Thereareseveralimportantglobalstudiesontheroleofinternationaltradeinclimate changeonagriculture.Forexample,ReillyandHohmann(1993)madethefirstattempttodiscusstheroleofinternationaltradein assessingclimatechangeimpacts.Later,BaldosandHertel(2015)exploredthepotentialforamorefreelyfunctioningglobaltrading systemtomaintainimprovedfoodsecurityinthelongrun(i.e.by2050).Morerecently,Brownetal.(2017)suggestedthatglobal tradewouldcontinuetoplayacentralroleinassuringthatglobalfoodsystemadaptstoachangingclimateinthatitislikelyto facilitate the movement of food from areas of surplus to areas of deficit. However, there is no China-focused study that assesses climatechangeimpactsonChina'sagriculturewhileaccountingfortheroleofinternationaltrade. TheoverallpurposeofthispaperistoprovideanupdatedandmorereliableevidenceontheimpactsofclimatechangeonChina's production,prices,tradeandself-sufficiencyofmajorcrops,withparticularfocusonthemarketandtraderesponses.Ourstudyaims togivesomeperspectivetothestudiesthat(i)focusonlyontheimpactsofclimatechangeonnationalfoodmarkets(ii)usesingle regionmodeland(iii)failtoconsiderthepricetransmissionfromtherestofworld.Ourstudyexaminestheclimatechangeimpacts onmajorcropstowards2050undertheworstclimatechangescenario(measuredwithrepresentativeconcentrationpathway,i.e., RCP8.5)andthebestclimatechangescenarioRCP2.6.2Toachievethispurpose,weusetheeconometricallyestimatedprojected changesintheyieldsofmajorcropsinChina,whilewederivetheprojectedcropyieldchangesforChina'smaintradingpartnersfrom aprocess-basedbiophysicalmethod.Next,weemployawidely-usedagriculturalpartialequilibriummodel(ChinaAgriculturePolicy Simulation Model, CAPSiM) of China to assess the climate change impacts on agriculture, thus considering the domestic market responses.Thenweusethelinkednationalandglobalequilibriummodel(CAPSiM-GTAP)toassesstheclimatechangeimpactson agriculture,whereinweconsiderboththemarketandtraderesponses.Thelinkedmodelapproacheffectivelytransmitstheeffectsof foreigncountries'climateshocksonagriculturetoChinaviatrade,whileallowingustouseamorepreciseanddetailednational economicmodel. Ourresultsshowthattheeffectsofclimatechangeoncropproductionaresignificantbuthavelargevariationsamong crops. Undertheworstclimatechangescenarioi.e.,RCP8.5,amongallcropsinChina,wheatyieldisprojectedtoexperiencethelargest decreaseof9.4%by2050.Aftertakingintoaccountthemarketresponse,productionlossesformostcroparedampened(e.g.wheat production loss reduces to only 4.3%) because of the growers' response to changes in agricultural prices under climate change. Moreover,ifweconsidertheimpactsofclimatechangefromtherestoftheworld,whichaffectChina'stradeandthereforedomestic production,theseverityofclimatechangeimpactsonChina'sagriculturalproductionwillbefurtherreduced,e.g.toaround4%for wheat.Thestudyconcludesthatweneedtolearnmorefromfarmerswhorespondtochangingclimateaccordingtothemarketand tradesignals,andfurthermainstreamtheselessonsintonationaladaptationdevelopmentplan. Therestofthepaperisorganizedasfollows:Section2introducesdatasourcesforyieldchangesunderdifferentclimatechange 1A2andB2scenariosrepresentdifferentcarbonemissionpathwaysandcorrespondinglydifferenttemperatureincreaseinthefuture. 2RCP2.6 and RCP8.5 are named after a possible range of radiative forcing values in the year 2100 relative to pre-industrial values (+2.6 and+8.5W/m2,respectively).WecaneasilyseethatRCP2.6andRCP8.5representlowandhighcarbonemissionpathwaysandcorrespondingly lowandhightemperatureincrease,respectively,inthefuture(IPCC,2014). 2 W.Xieetal. China Economic Review xxx (xxxx) xxx–xxx Table1 Thecropproduction(Mt)andclimatechangeimpactsoncropyieldofChinaunderRCP2.6andRCP8.5(%). Production(2012) 2030 2050 RCP2.6 RCP8.5 RCP2.6 RCP8.5 Wheat 121.02 −2.28 −3.39 −4.83 −9.39 Rice 204.24 −0.56 −0.78 −1.34 −2.60 Maize 205.61 0.33 −0.01 0.25 0.31 Soybean 13.60 0.26 0.08 0.31 0.42 Cotton 6.84 0.73 1.76 1.74 4.24 Rapeseed 14.00 −0.17 0.18 0.03 0.61 Peanut 16.69 −0.20 −0.20 −0.37 −0.20 Sugarbeet 11.74 −0.35 −0.14 −0.65 −0.38 Note:Thebaseyearis2012. Source:Theproductionin2012comesfromtheCAPSiMdatabase;theyieldchangecomesfromWang,2016. scenarios for China and its main trade partners. Section 3 describes the simulation methodology, baseline scenario and climate changescenarios.Section4presentsandanalyzestheresultsforclimatechangeimpactsonChina'sagricultureandtheroleofmarket andtrade.Section5concludesthestudywithpolicyimplications. 2. ClimatechangeshocksforbiophysicalyieldsofcropsinChinaandtherestofworld 2.1. ClimatechangeshocksforbiophysicalyieldsofcropsinChina Inthisstudywecoverrice,wheat,maize,soybean,cotton,rapeseed,peanutandsugarbeetastheyarethemajorcropsproduced inChina.WebeganwithextractingthechangingtrendsoftemperatureandprecipitationforChinafromthedownscalesimulationof LiangandYan(2016),basedontheRCPscenariosofIntergovernmentalPanelonClimateChangeFifthAssessmentReport(IPCC AR5)(IPCC,2014).BothRCP8.5andRCP2.6scenariosaremodeledinthisstudyastheworstandbestclimatechangescenario, respectively.InLiangandYan(2016),severalglobalcirculationmodels(GCM),providedbyCMIP5,areappliedtoprojectmonthly temperatureandprecipitationduring2010–2100ineachprovinceofChinawithbaseyearof1980–2010.Thenweestimateboth annualaverageandstandarddeviationoftemperatureandprecipitationduringeachcrop'sgrowthseason.Theprojectionsshowthat comparedto2012,theannualaveragetemperatureandprecipitationduringgrowingseasonofeachcropwillincreasesignificantly in2020–2050,whilethestandarddeviationofannualprecipitationwillincreasesignificantlyforeachcrop(seeAppendixFig.1). Thisshowsbothtemperatureandprecipitationwillincrease,butthelatterwillhavemoreannualfluctuationduringcropgrowing seasonsofthefutureyears. WeobtainthechangesinannualcropyieldunderclimatechangeinChinafromauniqueeconometricestimationofWang(2016). ThestudyusedChina'sprovincialpaneldatatoestimateclimatechangeimpactsontheyieldsofdifferentcropsintermsofchangesin annualtemperature,precipitationandtheirstandarddeviationsduringthegrowthseasonofmajorcropproducingprovinces,while controllingfordifferencesinagricultureinputsandtechnologyprogress(AppendixTable2).Thestudyfinallyillustratedanonlinear correlationbetweenclimatevariablesandcropyield,andextrapolatedtheannualchangesofChina'scropyieldsunderIPCC'sfour RCPscenariosfortheperiod2010–2050(Table1).3 ThephysicalimpactsofclimatechangeoncropyieldsinChinavaryconsiderablyamongcropsandareshowninTable1.Wheat, rice,peanut, and sugarbeet are projectedto experience yield reductionsunder both RCP 8.5and RCP 2.6 scenarios, with wheat expectedtobearthehighestyieldloss.Specifically,wheatyieldwoulddeclinesignificantlyin2050i.e.,by4.83%underRCP2.6and 9.39%underRCP8.5.Nexttowheat,riceyieldwouldhavemoderateyieldreductionin2050of1.34%underRCP2.6and2.60% underRCP8.5.Duetothechangingclimate,theyieldsofpeanutandsugarbeetareprojectedtodroponlymarginally.Othercrops, includingcotton,rapeseed,soybeanandmaizemayseepositiveyieldimpactsfromclimatechange.Amongthesecrops,cottonwill havethemostsignificantincreaseinyieldduetoclimatechange,followedbyrapeseed,soybean,andmaize.UnderRCP8.5,cotton yieldisprojectedtoincreaseby1.74%in2030and4.24%in2050.Comparedwithcotton,thepositiveimpactsofclimatechangeon soybeanandmaizeyieldarerathersmallsuchthattheiryieldswouldincreaseby<0.5%in2050underRCP8.5.Toconcordwith cropsectorsinCAPSiMmodel,weestimateoilseedyieldchangeastheaverageofchangesinrapeseedandpeanutyieldsweightedby harvestareain2015. 3Theinterestedreaderscancontactthecorrespondingauthoraboutthedatasource,detailestimationmethodandresults. 3 W.Xieetal. China Economic Review xxx (xxxx) xxx–xxx Table2 Theannualimpactsofclimatechangeoncropyieldintherestoftheworldto2050(%). RCP2.6 RCP8.5 Rice Wheat Maize Soybean Rice Wheat Maize Soybean Australia&NewZealand −0.02 −0.27 −0.12 −0.09 −0.19 −0.40 −0.28 −0.28 Japan 0.15 −0.09 −0.04 0.01 0.16 0.09 −0.18 −0.01 Korea −0.01 0.23 −0.35 −0.03 −0.06 0.24 −0.66 −0.07 Indonesia −0.01 0.00 −0.17 −0.08 0.00 0.00 −0.33 −0.13 Malaysia −0.04 0.00 −0.17 0.12 −0.08 0.00 −0.42 −0.05 Philippine −0.03 0.00 −0.18 −0.11 −0.05 0.00 −0.38 −0.24 Thailand −0.05 0.05 −0.38 −0.23 −0.14 −0.00 −0.79 −0.31 Vietnam −0.09 0.03 −0.22 −0.12 −0.18 0.03 −0.57 −0.26 Canada 0.00 0.05 0.15 0.17 0.00 −0.03 0.05 0.14 USA −0.09 0.06 −0.22 −0.03 −0.27 0.07 −0.63 −0.21 Argentina −0.03 −0.10 −0.05 −0.01 −0.12 −0.15 −0.25 −0.15 Brazil −0.07 −0.04 −0.14 −0.13 −0.18 −0.18 −0.41 −0.34 EU_28 0.07 0.13 −0.03 0.08 0.04 0.06 −0.20 0.00 RestofWorld −0.07 0.05 −0.13 −0.15 −0.20 −0.01 −0.37 −0.37 Note:Thebaseyearis2012. Source:SimulationresultsfromIFPRI. 2.2. ClimatechangeshocksforbiophysicalyieldsofcropsintherestofWorld Theestimatesonclimatechangeimpactsoncropyieldsforothercountriesarebasedonthebiophysicalsimulationsofaprocess- basedcropmodelbytheInternationalFoodPolicyResearchInstitute(IFPRI).Annualyieldchangesofmajorcrops,i.e.,wheat,maize, rice,andsoybean,inresponsetoclimatechangearelistedinTable2fortheworld's14regions/countries.Theyieldchangesofthese crops are estimated using the Decision Support System for Agrotechnology Transfer (DSSAT) model and IPCC RCP scenarios for 2011–2050withthebaseyearof2010.AsshowninTable2,climatechangeisprojectedtocausedifferentyieldchangesamong majorcropsinothercountriesunderRCP2.6andRCP8.5.Mostcountrieswouldseeareductionintheirannualaveragecropyields inresponsetoclimatechange.NoticethatUSA,ArgentinaandBrazil,whoarealsothemainexportersofmaizeandsoybean,would haveseriousyieldlossesby2050.Particularly,theannualaveragedecreaseinmaizeyieldwillbe>0.4%forUSAandBrazil,and 0.2%forArgentinaunderRCP8.5.Whilethesoybeanyieldwillfallby>0.3%forBrazil,and>0.15%forUSAandArgentinaunder RCP8.5. Thesupplyofthesecrops,alsoregardedasthemajoragriculturalcommoditiesimportedbyChina,willbesignificantly threatened by climate change, ensuing a major global hike in their respective prices. Furthermore, we can also find that Canada wouldbenefitfromclimatechangeinbothmaizeandsoybeanyieldby2050(increaseby0.05%perannum(p.a.)formaizeand 0.14%p.a.forsoybeanunderRCP8.5).Moreover,mostcountriesarefoundtohavenegativeimpactsofclimatechangeonriceand wheatyieldsby2050.AustraliaandNewZealandareprojectedtoexperiencethebiggestdecreaseinwheatyield(−0.40%p.a.), whileUSAhavethebiggestdecreaseinriceyield(−0.27%p.a.)underRCP8.5.Atthesametime,JapanandEuropeanUnionwould benefitfromclimatechangeintermsofbothriceandwheatyieldsunderRCP8.5. HerewewanttonotethatusingclimateshocksforChinaandtherestofworldfromdissimilarsourcesinthisstudycanresultin some inconsistencies. In fact, our motive is to include the effects of adaptation by farmers to reflect the real impacts of climate change,whichwedothroughoureconometricestimationforChina.AsChinaisourmainstudyregion,ourpriorityistomakesure thattheresultsofChinahavehighaccuracy.However,duetotheunavailabilityofdataforalltheothercountries,itisimpossibleto dotheeconometricestimationfortherestofworld,forwhomweusethesimulationresultsfromabiophysicalmodel.Additionally,a comparison between our econometric results and the biophysical simulation results for China could reveal if there are large dis- crepanciesbetweenbothsetsofyieldshocks.Here,wefindthatthetwomethodshavesimilarresultsfortheimpacts(forexample, under RCP8.5 in 2050, the econometric results for rice and soybean are −2.6% and 0.42% respectively; while the biophysical simulationresultsforriceandsoybeanare−3.2%and−0.87%respectively).Wethinkthatthecropyieldlossesforothercountries without consideration of adaptation might be slightly overestimated, so the results of our study might also be somewhat over- estimatedinoursimulationsofeconomicmodels. 3. Simulationmethodologyandscenarios 3.1. SimulationModel In order to consider the domestic market responses to climate change impacts on China's agriculture, we have used a widely recognizedagriculturalpartialequilibriummodel(ChinaAgriculturePolicySimulationModel,CAPSiM).Themodelwasdeveloped attheChinaCenterforAgriculturePolicy(CCAP)inthemid-1990sasapartialequilibriummodelforanalyzingpoliciesaffecting 4 W.Xieetal. China Economic Review xxx (xxxx) xxx–xxx China'sagriculturalproduction,consumption,prices,andtrade(Huang&Li,2003;Li&Huang,2004).Sincethen,CAPSiMhasbeen periodicallyupdatedandexpanded,whiletherecentversionsoftheCAPSiMaredesignedtotrackchangesintradeliberalization, urbanization,andclimatechange(Yang,Huang,Rozelle,&Martin,2012;Huangetal.,2017).InCAPSiM,thecropssectorsaremore disaggregatedandaccountfor>90%ofChina'sagriculturaloutput.Themodelcovers21agriculturalcommodities:includingrice, wheat,maize,othercoarsegrain,sweetpotato,potato,soybean,edibleoilcrops,cotton,vegetables,fruits,othercrops,aswellassix livestock products and three fishery sectors. The accompanying database of CAPSiM has been updated to 2015 according to the officialstatisticsfromChina'sNationalBureauofStatisticsandNationalCustoms.CAPSiMcaninvestigatetheclimatechangeimpacts onChina'sagriculturetorevealtheresponsefromlocalmarkets,withtheassumptionthatclimatechangeeffectsfromothercountries donottranscendtoChinaviatrade. Thentoconsiderboththemarketandtraderesponsessimultaneously,wealsousedthelinkednationalandglobalequilibrium model(CAPSiM-GTAP)toassesstheclimatechangeimpactsonagriculture.GTAP(GlobalTradeAnalysisProject)modelisawell- recognized multi-country, multi-sector computable general equilibrium model, and is often used for international trade analysis (Hertel,1997).GTAPmodelhastheadvantageofsimulatingglobalpricechangesofagriculturalcommoditiesinresponsetoclimate change. However, in contrast to the China module of GTAP model, the CAPSiM model also has the following advantages: first, CAPSiMisa partial equilibriummodel ofChina'sfoodmarket presenting the supplyanddemandin volumetric(quantity) terms. Whereas,GTAPmodelisageneralequilibriummodelusingdollarvaluesforsupplyanddemandrelationships.Forfoodmarkets, quantityimpactsareveryimportantforcapturingtheeffectsofclimatechangeoranyothershocks.Thisisoneofthereasonsthat researchersusuallyrelyonpartialequilibriummodelstoprojectthequantitylevelresultsforthefuture(forexample,FAO-OECD AgriculturalOutlook;USDAAgriculturalProjections).Second,mostofthekeyparametersofCAPSiMmodelarederivedfromthe empiricallybasedstudiesconductedbyCCAP,incontrasttogeneralizedparametersusedinGTAPmodel.Third,thebasedataof CAPSiMhasbeenupdatedtomorerecentyear(2015)reflectingChina'smarketstructuremoreprecisely,whilethelatestdatabaseof GTAPmodelisbasedonmarketconditionsin2011.Moreover,theCAPSiMbasedprojectionsonfuturefoodmarketforChinaarealso widelyacceptedinChina.We,therefore,havehigherconfidenceinCAPSiMresultsincomparisontoresultsfromtheChinamoduleof GTAPmodel.Overall,alinkedmodelbetweenCAPSiMandGTAPoffersthebestofbothindividualmodelssuchthatwecantransmit theeffectsofothercountries' climateshockstoChinaviatrade,while simultaneouslyusingamorepreciseanddetailed national economicmodel.Finally,tomapthesectorsbetweentheCAPSiMandGTAPmodel,theGTAPversion9databaseisaggregatedinto 15regionsand25sectors(AppendixTable1). FollowingHorridgeandZhai(2005),weestablishedalinkagemodulebetweenCAPSiMandGTAPmodeltoevaluatetheclimate change impacts while considering both the responses of market and trade concurrently. The key idea of CAPSiM-GTAP linking method,asproposedbyHorridgeandZhai(2005),istotransmittheglobalpricechangesfromGTAPmodelintothenationalmodel throughtrade.Specifically,inCAPSiM,theglobaldemandpriceforChina'sfoodexportandtheglobalsupplypriceforChina'sfood importareexogenousandareupdatedusingtheprojectionofOECD-FAOagriculturaloutlook(OECD/FAO,2018)underthebaseline scenario.Underourproposedfirstscenario(climatechangescenarioconsideringonlydomesticmarketresponseunderRCP2.6and RCP8.5usingCAPSiM),aswedonotconsidertheglobalpricechangescausedbyclimatechangeinothercountries,wekeepthe globaldemandpriceforChina'sfoodexportandtheglobalsupplypriceforChina'sfoodimportsameasthebaselinescenario(only useCAPSiMmodelasgiveninAppendixTable3).Underourproposedsecondscenario(climatechangescenariosconsideringboth thedomesticmarketresponseandthetraderesponseusinglinkedCAPSiM–GTAPmodel),weproceededinthreesteps:1)weassume thatclimatechangeonlyaffectsChinaandthereforeweonlyshockChina'scropyieldsandkeepthecropyieldsforallothercountries unchangedinGTAPmodel.Ideally,ifthestructureforChina'seconomyinbothCAPSiMandGTAPmodelweresimilartoeachother, wewouldexpecttohavethesameresultsfromthissimulationasinthefirstscenario.However,asChinaisrepresenteddifferentlyin bothmodels,weanticipatethatourCAPSiMmodelcanbetterreflectChina'sfoodmarketthantheChinamoduleinGTAPmodel.2) We assume that climate change strikes all over the world, so we shock all countries' crop yields in GTAP model. 3) We take the differenceofglobalfoodpricesbetweenstep1and2(step2-step1)asakintotheimpactsofclimatechangeinothercountrieson China'sfoodmarket.Thus,weincorporatethedifferenceinglobalpricebetweenthetwostepsintoCAPSiMtoreflecttheimpactsof climatechangeinothercountriesonChina'sfoodmarketthroughtrade(ofcourse,underthesecondscenario,weshockboththecrop yieldsandtheglobalfoodprices–theglobaldemandpriceforChina'sfoodexportandtheglobalsupplypriceforChina'sfoodimport– inCAPSiM)(seeAppendixTable3). 3.2. Baselinescenario ForanalyzingtheimpactsofclimatechangeonChina'sagriculture,weestablishabaselinescenariotowards2050forbothGTAP model and CAPSiM. The GTAP baseline is constructed by recursively updating the database such that given GDP targets are met throughgivenexogenousestimatesoffactorendowmentsi.e.skilledlabor,unskilledlabor,capital,naturalresources,andpopulation. The procedure and the exogenous macro assumptions are discussed in details in Hertel (1997) and Walmsley, Dimaranan, and McDougall(2006).ForthebaselineinCAPSiM,severalkeyassumptionsareusedforthebaselinescenarioconcerningGDPgrowth, populationgrowth,urbanizationrate,urbanandruralhouseholds'incomegrowth,andagriculturaltechnologyadvancement(forin depthdiscussionseeHuangetal.,2017). 5 W.Xieetal. China Economic Review xxx (xxxx) xxx–xxx In the baseline projection, China's agricultural production will continuously increase in the future, with a simultaneous and significantriseintheimbalancebetweenagriculturalproductionanddemand.Demandforfeedgrainswillgrowfasterthantheir domesticproduction,leadingtodecliningself-sufficiencyrates.4By2050,domesticproductionofriceandwheatwillalmostmeet China'sdomesticdemand,bothreachinghighself-sufficiencyratesofover95%.However,formaize,whichexperiencedover-supply inrecentyearsmainlyduetopolicyinterventionsinChina,thedemandwillincreasesignificantlyincomingdecadesduetorising demandbylivestockproduction.IfChinadoesnotimplementtariffratequota(TRQ)5inthefuture,China'smaizeimportisprojected tosurpass40milliontonsby2050,leadingtoaself-sufficiencyrateof<85%.Similartomaize,soybeanimportisprojectedtocross 100milliontonsin2050,resultinginaself-sufficiencyrateof<10%forChina.Demandforsugarandedibleoilswillbesignificantly higher than their respective domestic productions, leading to decreasing self-sufficiency ratios for both commodities. In contrast, domestic production of vegetables and fruits is projected to increase in pace with domestic demand, ensuring almost full self- sufficiencyinthefuture. China's livestock supply-demand balance mostly depends on policies governing feed grain trade and grassland development. CAPSiM projection shows that aquatic products will almost keep supply-demand balance requiring minimal import. However, li- vestockself-sufficiencymayundergosignificantchangesduetomanyuncertaintiessurroundingitsdemandandsupply.IfChinawere toremovetheimportlimitationsonfeedgrainandthusmakewayfordomesticlivestockproductionfedbycheapimportedfeed grain, pork and poultry could retain high self-sufficiency rates. In contrast, livestock imports in China will significantly increase mainlyduetomaizeimportlimitation(e.g.TRQ)andinadequategrasslanddevelopment.Inthelattersituation,CAPSiMprojections showthatinadditiontoconsiderableporkandpoultryimports,Chinawillimportlargequantitiesofbeef,muttonanddairyby2050, andwillhaveself-sufficiencyratesrangingover70–80%acrossdifferentlivestockproducts. 3.3. Climatechangescenarios In the CAPSiM settings, percent change of crop yield is a linear function of the percentage change of crop price, input prices (includingfertilizer,land,andlabor),aswellasotherfactors(suchasclimatechangeconditions).Thus,climatechangeimpactson cropyieldsdiscussedinsection2aretransmittedintothecropproductionmoduleintheCAPSiMthroughshiftingthecropyield changes.Meanwhile,cropyieldchangesaresimulatedinGTAPmodelastheshifttototalfactorproductivityofthecropsectors.In RosonandMensbrugghe(2010),variationsinagriculturalyieldaremodeledaschangesinmultifactorproductivityforagricultural activities,sothatoutputvolumesvarydespiteusingthesamemixofproductionfactors(theyusedtheENVISAGEmodel—ageneral equilibriumeconomicmodel).InNelson,Mensbrugghe,etal.(2014),forthegeneralequilibriumeconomicmodels,theyieldshocks of climate change are implemented as shifts in the land efficiency parameters of the sectoral production functions; while for the partialequilibriummodels,theshockswereintroducedasadditiveshiftersinayieldorsupplyequation.Robinson,vanMeijl,Valin, and Willenbockel (2014) also discussed the incorporation of yield shocks into general/partial equilibrium models. It can thus be concludedthatregardlessofthemodeltypei.e.generalorpartialequilibrium,somestudieschosetoshockTFP;whiletheothers shocklandefficiency.Inourstudy,fortheCAPSiM,theshocksareintroducedasadditiveshiftersincropyield;forGTAPmodel,crop yieldchangesaresimulatedaschangesintotalfactorproductivity(TFP)ofthesecropsectors.Becauseweusedthelinkedmodel,we kepttheshockmethodsconsistentbetweenCAPSiMandtheChinamoduleinGTAP. WeconstructedtwoseparateclimatechangescenariostosimulatetheimpactsofclimatechangeonChina'sfoodsupply,prices, tradeandself-sufficiency, andexamine the marketand traderesponses. 1) Climatechangescenarioswith consideringmarket re- sponse(usingCAPSiM)underRCP2.6andRCP8.5;2)climatechangescenariowithconsideringboththemarketresponseandthe impactsonrestoftheworld(ROW)(usinglinkedCAPSiM-GTAPmodel).Comparingchangesinbiophysicalcropyieldswithchanges incropproductionestimatedusingCAPSiMonlycouldrevealtheresponseofdomesticmarketinbufferingclimatechangeimpacts, becauseCAPSiMmodelkeepsfoodimportandexportpricesunchanged.ThelinkedCAPSiM–GTAPmodel,ontheotherhand,allows thefoodimportandexportpricestochangewithchangesinglobalfoodprices,whichareprojectedbytheGTAPmodel.Acom- parisonbetweentheresultsfromCAPSiMandCAPSiM-GTAPmodelcouldrevealtheresponseofglobaltradeinbufferingclimate changeimpacts(AppendixTable3). 4. SimulatedresultsforclimatechangeimpactsonChina'sagriculture ThefollowingsectiondescribessimulatedresultsforclimatechangeimpactsonChina'sagriculturalproduction,prices,andtrade basedontheCAPSiMandthelinkedCAPSiM-GTAPsimulationsin2015–2050.ComparingtheCAPSiMresultswiththebiophysical impactsofclimatechangecanrevealtheresponseofdomesticmarketinbufferingclimatechangeimpacts.Thentheassessmenton climatechangeimpactsconsideringtheresponseofglobaltradewillbediscussedbasedonthesimulationresultsfromthelinked CAPSiM-GTAPmodel.Tothisend,percentagechangesindicatedinthetextrefertothedifferenceofagriculturalproduction,prices andtradewithoutandwithclimatechange. 4Theself-sufficiencyrateisdefinedastheratioofdomesticfoodproductiontofoodsupply(productionplusnetimport) 5Themaizeimportquotaissetat7.2Mt.in2017,anda65%tariffwillbeimposedontheimportedmaizebeyondthequota. 6 W.Xieetal. China Economic Review xxx (xxxx) xxx–xxx Table3 TheimpactsofclimatechangeoncropproductioninChinaunderRCP2.6and8.5(%). RCP2.6 RCP8.5 CAPSiM-GTAP CAPSiM CAPSiM-GTAP 2050 2030 2050 2030 2050 2030 2050 Rice −0.27 −0.46 −0.28 −0.55 −0.30 −0.67 −0.21 −0.22 Wheat −0.92 −1.61 −0.97 −2.21 −1.95 −4.28 −1.92 −4.03 Maize 0.24 0.20 0.40 3.58 −0.11 −0.64 1.01 1.93 Soybean 0.38 0.29 1.48 2.98 0.00 −1.47 4.26 16.75 Cotton −0.48 0.74 −0.10 2.35 2.07 3.57 2.49 9.30 Oilseed 0.06 0.17 0.11 0.23 0.19 −0.10 0.41 0.72 Sugar −0.11 −0.21 −0.15 −0.45 −0.32 −0.66 −0.53 −1.53 Note:Thebaseyearis2012. Source:CAPSiMandCAPSiM-GTAPsimulations. 4.1. ClimatechangeimpactsonChina'sagriculturalproduction Climate change will have varying impacts on China's future crop production. From the CAPSiM simulations, rice, wheat, and sugarwillhaveproductionlossesduetoclimatechangebothundertheRCP2.6andRCP8.5(Row1–2,Table3),whereinwheatis projectedtohavethehighestproductionreductionby2050(−1.61%underRCP2.6and−4.28%underRCP8.5).Noticethatthe climate change impacts on production of these crops are less than the yield losses estimated by the econometric model. Wheat productionlossin2050underRCP8.5(4.28%)islessthanhalftheyieldlossduetoclimatechange(9.39%).Thisindicatesthatthe domesticmarketevidentlyplaysanimportantroleindampeningclimatechangeimpacts.Whentheclimatechangehitscroppro- duction,thefarmersimprovetheirproductionpracticesinlightoftheirpreviousexperienceundersimilarsituations,whichatleast partially reduces the production losses caused by climate change. Farmers are likely to increase frequency and strength of field management,suchasirrigation,weeding,adoptingdrought-resistantvarieties,amongothers.Theseresultssignifytheimportantrole thatthedomesticmarketcanplayinbufferingclimatechangeimpacts. Moreinterestingly,somecropswithpositiveyieldchangeswillenduphavingproductionreduction(Table3).Forexample,by 2050maizewillhaveslightyieldincreaseunderRCP8.5(0.31%,Table1),however,itsproductionisprojectedtodecreaseunder RCP8.5(−0.64%,Table3).Themechanismatactionisthatriceandwheataremostlydomesticallyproducedandtheiryields,in contrasttomaize,aremoreseriouslyaffectedbyclimatechangeinChina.Keepinginmindtheimportanceofriceandwheat,the farmerswouldincreasetheirproductionbynotonlyimprovingfieldmanagement,butalsobytakingagriculturalinputs(e.g.,land andlabor)awayfromthepositivelyaffectedcrops(likemaize).Asaresult,thepositiveimpactsofclimatechangeonmaizeyield wouldbeoffsetbydeclininginputsoflandandlabor,andevenrendermaizeoutputtodecline.Moreover,bothsoybeanandoilseed cropshaveslightlypositiveoutputimpactsduetoclimatechange,exceptin2050underRCP8.5(Table3).Similartomaize,the substitutioneffectsbetweencropswouldoffsettheslightyieldincreaseforsoybeanandoilseedcropsbroughtbyclimatechange.In 20 ent 15 Physical impacts CAPSiM CAPSiM-GTAP er diff 10 der %) e unos ( 5 n changscenari 0 o -5 cti u od -10 Pr -15 Wheat Rice Maize Soybean Fig.1.Comparisonofcropphysicalimpactsandproductionchanges(%)intheCAPSiMandCAPSiM-GTAPmodelin2050underRCP8.5(thebase yearis2012). 7 W.Xieetal. China Economic Review xxx (xxxx) xxx–xxx Table4 TheimpactsofclimatechangeoncroppriceinChinaunderRCP2.6and8.5(%). RCP2.6 RCP8.5 CAPSiM CAPSiM-GTAP CAPSiM CAPSiM-GTAP 2030 2050 2030 2050 2030 2050 2030 2050 Rice 1.75 2.92 1.52 2.91 1.61 4.55 2.81 7.41 Wheat 3.85 6.83 3.71 7.17 5.49 15.47 8.56 22.91 Maize 0.16 0.31 0.23 6.35 0.00 0.23 2.95 8.86 Soybean −0.02 −0.02 1.64 6.73 −0.03 0.10 7.06 30.27 Cotton 0.28 −0.27 1.02 3.82 −0.58 −0.77 0.54 10.42 Oilseed −0.05 −0.15 0.11 0.42 −0.27 0.10 0.56 3.00 Sugar 0.24 0.32 0.24 0.58 0.23 0.51 0.81 2.06 Note:Thebaseyearis2012. Source:CAPSiMandCAPSiM-GTAPsimulations addition,cottonproductionwouldbenefitfromclimatechangeby2050byrelativelylowermarginsunderbothRCP2.6(0.74%)and RCP8.5(3.57%).Thisisbecausecottonyieldincreasesaremuchlargeby2050(1.74%forRCP2.6and4.24%forRCP8.5)although partlyoffsetbythesubstitutioneffects. Further,theclimatechangeimpactsinothercountrieswillcausecrossborderrippleeffectsandwillfurthersoftentheimpactsof climatechangeonChina'sagriculture(Fig.1).Forexample,whilesoybeanproductionin2050willdecreaseslightlyunderRCP8.5 (−1.47%)intheCAPSiMresults,thesameisprojectedtosignificantlyincrease(16.75%)intheCAPSiM-GTAPlinkagemodel.This effectcouldbeattributedtooppositeimpactsofclimatechangeonsoybeanyieldsinChinaandtheothercountries.Whilesoybean yieldisprojectedtoincreaseslightlyinChina(Table1),theyieldsformainexporters,suchasBrazil,Argentina,andUSA,areall projectedtodecreasesignificantly(Table2).Soybeanoutputreductionintheaforementionedglobalexporterswouldcausesevere shortageintheglobalmarket,whichwillfurtherincentivizethefarmersinChinatoimprovesoybeanproduction.Consequently, China'ssoybeanproductionwouldexpandintheCAPSiM-GTAPresults.Asimilareffectofinternationaltradecanalsobefoundon maize production, which has a slight decrease in the CAPSiM results in 2050 under RCP8.5 (−0.64%) but a slight increase in CAPSiM-GTAPresults(1.93%)(Fig.1).AlthoughbothriceandwheatoutputswoulddeclineinCAPSiMsimulationunderRCP8.5, theoutputreductionsintheCAPSiM-GTAPresultsarelowerthanthoseintheCAPSiMresults(Fig.1).Forexample,wheatoutput wouldreduceby4.28%intheCAPSiMresultsin2050underRCP8.5,andby4.03%intheCAPSiM-GTAPresults.Thissetofresults showsthatafterweconsidertheroleofinternationaltradeinclimatechangeassessment,thenegativeimpactsofclimatechangeon China'sagriculturewillbefurtherreduced,atleastpartially. 4.2. ClimatechangeimpactsonChina'sagriculturalprices Thepricesofthenegativelyaffectedcropsunderclimatechangewouldincreasesindomesticmarketby2030and2050both underRCP2.6andRCP8.5.InCAPSiMsimulation,themarketclearingmechanismdictatesthatwhenclimatechangecausesyield Table5 ImpactsofclimatechangeoncropnetimportunderRCP2.6and8.5(%). RCP2.6 RCP8.5 CAPSiM CAPSiM-GTAP CAPSiM CAPSiM-GTAP 2030 2050 2030 2050 2030 2050 2030 2050 Rice 9.86 22.16 16.31 61.88 15.35 43.93 1.66 −50.37 Wheat 10.86 20.33 13.22 43.11 22.87 56.81 16.69 30.87 Maize −1.11 −0.47 −1.50 −10.87 0.64 1.69 −4.74 −1.70 Soybean −0.06 −0.05 −0.99 −3.31 −0.02 0.28 −3.75 −13.71 Cotton 0.60 −0.62 0.12 −1.97 −2.60 −3.00 −3.12 −7.81 Oilseed −0.11 −0.35 −0.15 0.80 −0.38 0.46 −0.62 0.14 Note:Thebaseyearis2012. Source:CAPSiMandCAPSiM-GTAPsimulations 8 W.Xieetal. China Economic Review xxx (xxxx) xxx–xxx Table6 Impactsofclimatechangeoncropself-sufficientrateunderRCP2.6and8.5(absolutepercentchange). RCP2.6 RCP8.5 CAPSiM CAPSiM-GTAP CAPSiM CAPSiM-GTAP 2030 2050 2030 2050 2030 2050 2030 2050 Rice −0.05 −0.11 −0.08 −0.29 −0.08 −0.21 −0.01 0.23 Wheat −0.24 −0.48 −0.28 −1.00 −0.50 −1.37 −0.37 −0.78 Maize 0.16 0.13 0.23 2.85 −0.09 −0.46 0.70 0.71 Soybean 0.06 0.04 0.31 0.77 0.00 −0.21 1.04 4.02 Cotton −0.27 0.34 −0.05 1.07 1.15 1.63 1.38 4.24 Oilseed 0.02 0.07 0.04 −0.08 0.08 −0.08 0.15 0.08 Sugar −0.06 −0.13 −0.08 −0.26 −0.18 −0.40 −0.27 −0.83 Note:Thebaseyearis2012. Source:CAPSiMandCAPSiM-GTAPsimulations reduction, domestic production of the crops will decrease, and consequently, the inadequate domestic supply will raise the local prices.Rice,wheat,andsugarwouldhavetheirlocalpricetoincreasebyhighestmarginsinresponsetoyieldreductioncausedby climatechange.Forexample,wheatwouldhavethelargestpriceincreasein2050ofaround6.83%underRCP2.6and15.47%under RCP8.5(Table4),becauseitwouldexperiencetheworstyielddamage.Ricewillseeamoderatepricehikeby2050of2.92%under RCP2.6and4.55%underRCP8.5.Moreover,thedomesticpricesofallothercropswillalsoincreasesin2050underRCP8.5except forcotton.Consistentwithitspositiveyieldshock,cottonwouldhaveareductioninlocalpriceof0.58%in2030and0.77%in2050 underRCP8.5.However,whilemaizewillexperienceslightyieldincreaseunderbothRCP2.6andRCP8.6,itsdomesticpricefor Chinawillincreasemarginally,mainlyduetothesubstitutioneffectsmentionedinsection4.1. AscomparedtotheCAPSiMresults,thedomesticpricesforallthecropswillincreasebymuchhighermarginsifweconsiderthe responseofinternationaltradeusinglinkedCAPSiM-GTAPmodel(Table4).Climatechangehavesignificantimpactsnotonlyoncrop prices in China, but also on the crop prices in other countries. The global prices would increase sharply for the crops with high negativeyieldchangesduetoclimatechangesuchthatChinawillbeunabletoimportthesecropsatthenewpricelevels.Asaresult, the reduced supply will lead to a sharp rise in China's domestic crop prices. Our results show that domestic prices of wheat and soybeanwouldfurtherincreasegreatlyinthelinkedCAPSiM-GTAPresults,mainlybecauseChina'smaintradingpartnerswillsuffer moresevereyieldreductionforthesecrops. 4.3. ClimatechangeimpactsonChina'sagriculturaltradeandself-sufficiency Inadditiontocropproductionandprices,climatechangewillalsosignificantlyaffectChina'stradeintheseagriculturalcom- modities.IntheCAPSiMresults,thecropswithnegativeyieldshocks,especiallyriceandwheat,willseeincreaseintheirnetimports in 2030 and 2050 (Table 5). Wheat is projected to have the most significant increase in net import in 2050 both under RCP 2.6 (20.33%)andRCP8.5(56.81%).Comparedwitharound4%productionreductionofwheatin2050underRCP8.5,theseemingly largepercentageincrease(56.81%)inwheatnetimportisnotactuallylargeinvolumetermsaswheatimporthasverysmallsharein China'stotalwheatdemand.Othercrops,includingcotton,oilseed,andsoybean,areexpectedtohaveslightreductionsintheirnet importsinresponsetoclimatechange,astheirrespectiveyieldswouldincreaseslightlyinChina. Ontheotherhand,netimportsofthecropsintheCAPSiM-GTAPresultsdifferfromthoseintheCAPSiMresults.ThoughChina's domesticpricesofcropswouldriseduetoclimatechange,globalcroppriceswouldalsoincreaseduetoreducedproductioninseveral majorproducingcountries.IftheglobalcroppricesincreasemorethantheincreaseinChina'scropprices,Chinawouldinevitably reduceitsnetimportsofthecrops.Forexample,in2050underRCP8.5, China'snetimportofwheatisprojectedtoincreaseby 56.81%intheCAPSiMresults,buttheincreaseisreducedto30.87%intheCAPSiM-GTAPresults(Table5).Similartowheat,other cropsalsohavelowernetimportsinthelinkedCAPSiM-GTAPresults,e.g.,China'snetimportforsoybeanwillfallby13.71%(>10 Mt)in2050underRCP8.5ascomparedto0.28%increaseofnetimportforsoybeanintheCAPSiMresults. ThoughclimatechangewouldthreatenChina'sself-sufficiencyinmanyagriculturalcommodities,thecropself-sufficiencyrates willincreasewhenconsideringtheclimateshocksinothercountries.Comparedtothebaselinescenario,cropsexperiencingnegative yieldshockswillhavedecreasingself-sufficiencyratesintheCAPSiMresults(Table6).Amongthesecrops,wheathasthelargest decreaseinself-sufficiencyratein2050(by0.48percentagepointsunderRCP2.6and1.37percentagepointsunderRCP8.5),which isconsistentwiththefactthatwheathappenstobethecropwiththemostsignificantoutputreductionandnetimportincrease. UnderRCP8.5scenario,allothercropswillhavelowerself-sufficiencyratesby2050comparedto2010,exceptforcotton,which benefitsmostfromclimatechange.Theoverallself-sufficiencyrateofmajorcereals6in2050woulddecreaseby0.21percentage pointsunderRCP2.6,and0.65percentagepointsunderRCP8.5.Ontheotherhand,intheCAPSiM-GTAPresults,allcropswould have higher self-sufficiency rates as compared to the corresponding numbers in the CAPSiM results. For example, soybean's self- 6Majorcerealsincluderice,wheat,andmaize. 9 W.Xieetal. China Economic Review xxx (xxxx) xxx–xxx sufficiencyratewouldincreaseby0.46percentagepoints(0.77–0.31)in2050underRCP2.6and2.98percentagepoints(4.02–1.04) underRCP8.5whenconsideringtheclimateshocksinothercountries.Theseresultsfurthershowthatwhenconsideringtheclimate shocksinothercountries,China'sagriculturalself-sufficiencywillincrease. 5. Conclusionsandpolicyimplications Agriculture,animportantsectorinChina,ismandatedtofeedover1.3billionpeopleofthecountryandprovideimportantinputs formanyindustries.Suchprospect,however,islikelytobethreatenedbytheyielddamagescausedbyclimatechange.Theprevious studiesonclimatechangeeffectsonagricultureinChinadidnotaccountforthebufferingcapabilityoflocalmarketandinternational trade.Tofillthisgapintheliterature,weassessclimatechangeimpactsonChina'sagricultureandresponsesfrommarketandtrade usinganagriculturalpartialequilibriummodel,CAPSiM,anditslinkagemodelwithGTAPmodel(CAPSiM-GTAP).Inthispaper,the climatechangeimpactsareexaminedduring2020–2050underRCP2.6andRCP8.5scenarios.Ourresultsshowthatclimatechange wouldhavesignificanteffectsonagricultureproductionofChinabutwithlargevariationsamongcrops.Undertheworstclimate changescenario,i.e.,RCP8.5,wheatproductionisprojectedtodeclinebyaround9.4%by2050,thebiggestproductionreduction amongthecrops.Theresultsalsosuggestsomeevidenceoftheadaptationcapabilityofmarketresponsetoclimatechangewherein farmersintensifyagronomicinputs,improvefieldmanagementandadjustproductionstructure.Whenweaddthemarketresponseto the mix, production loss for wheat under RCP8.5 reduces to only 4.3%. Global agricultural trade provides additional adaptation capability to climate change damage for China, where the country can further avoid crop production losses and raising its self- sufficiencyof important food crops,at leastpartly. Whenconsideringboth domestic market andinternationaltrade responsessi- multaneously,wheatproductionlossunderRCP8.5wouldreducefurthertoaround4%. Ourresultshaveimportantpolicyimplicationsfornationaladaptationplans.First,theadaptationpoliciesshouldprioritizethe cropsbasedontheseverityofproductionlosses.Specifically,theinvestmentsinadaptationmeasuresshouldbechanneledtomore negativelyaffectedcropsandtotheonesthatplaymorevitalroleinnationalfoodsecurity.Secondly,thepoliciesfacilitatingmarket integrationandfreetradewouldhelptobufferclimatechangeimpacts.Ingeneral,whenclimatechangestrikes,farmersintrinsically increaseagronomicinputs(labor,irrigation,pesticide,andothers)toadapttoclimatechange,becausetheyexpecthighpricesin lightoftheirpreviousexperienceofhighpriceduetoclimatechange.Moreso,ifthedomesticmarketandinternationaltradearefree of distortions and barriers, wherein the price will increase to some reasonable extent in times of climate change. Then in the subsequentcropseason,farmerswillincreaseinputsashighastheycantopreventproductionlossesbasedontheirexperiencewith the price increase during previous climate changes. On the contrary, if the market is cluttered with interventions or the trade is restricted,farmerscannotexperiencegeneralpricechangeduetopreviousclimatechanges,andwhenclimatecontinuestochange, theymaynotincreaseinputstothatextent.Thus,inadditiontothehardmeasuresforadaptation(suchasinvestmentinirrigation system),thesoftmeasures(e.g.reducingmarketintervention,reducingimporttariffsandimportquotasorothertradebarriers)are recommendedinordertoreduceproductiondamagescausedbyclimatechange.Thirdly,tooptimizeadaptationplans,weneedto learn more from farmers who respond to changing climate according to the market and trade signals, and then improve and mainstreamthepracticesadoptedbyfarmersintonationaladaptationdevelopmentplans.Thefarmer'sadaptationmeasurescarried bythemselvesaremuchimportantinadaptingtoclimatechange,includingincreasingnumberofirrigationsandotherfieldman- agementmeasures.Theonlythingweneedtodoistokeepmarketsfreeandremovetradebarriers. Acknowledgements TheauthorsacknowledgetheirrespectivefinancialsupportsfromMinistryofScienceandTechnology,China(2012CB955700), NationalNaturalSciencesFoundationofChina(71503243;71873009;71333013),NSFC-CGIAR(71161140351),AustralianCentre forInternationalAgriculturalResearch(ADP/2010/070),andNationalSocialScienceFundofChina(16AJL009). 10