Wethankthereviewerfortakingthetimetoreviewourmanuscriptandprovideconstructivecomments.Inrevisingourmanuscript,wenotedthatourmodelsimulationshadusedafixedpre-industrialnitrogendepositionrate.Inourresubmission,wereflectedthatitwouldmakemoresensetoshowresultsfromLPJ-GUESSwiththenitrogencycleswitchedoff.Thiswasbecausetheprincipalaimofourpaperwastoexplorethesensitivityofthecarboncycleto‘expressions’ofElNinoandwemightexpectthatthissensitivitywouldbegreatestusingtheC-onlyversionofLPJ-GUESSascarbonuptakeisnotlimitedbynutrientavailability(whichmaydeclinewithwateravailabilityindryyears,whennitrogenimmobilisationratesincrease).Nevertheless,asoneofourmainregionsofinterestwasthetropics,wewouldnotexpectastronglimitationbynitrogen(Vitouseketal.1984)andasaresult,wedonotanticipateastrongsensitivityinourresultstoourchoiceofbiogeochemicalcycle.ToassuretheEditor/Reviewerofthisinsensitivitywehaveshowntheresultsofbothcycles(N-cycleon/off)below(Fig1).WealsousedthisopportunitytoupdatethemodelcomparisonagainstthemorerecentTRENDYv7runs.Overall,wefoundthatLPJ-GUESSisclosetotheTRENDYv7ensemblemeanandsimulationsaremostlywithinthemodelrange(i.e.acrossTRENDYmodels)whenweswitchthenitrogencycleoff.ThespatialdistributionofthesummedcompositeGPPanomalies(seefig.2)showsthatLPJ-GUESSpicksupthemainanomaliesassociatedwithEPElNinoeventsandremainswithintheTRENDYmodels’range.Finally,LPJ-GUESShasastrongnegativebiasinAustralia.Asourresultsshow,Australiadoesnotmakealargecontributiontolong-termchangesinanyofthecarbonfluxesandpools.

Fig.1.MonthlycompositeanomaliesduringtheElNinodeveloping(y0)anddecaying(y1)yearingrossprimaryproduction(GPP;greenlines)andterrestrialecosystem

respiration(TER;sumofautotrophicandheterotrophicrespiration;redlines)forallCPandEPElNinoeventslistedinappendixtableA1averagedovertheglobe,thetropics(23°S–23°N)andAustralia.ThedottedlinesshowtheTRENDYv7composite,thesolidlinesaretheindividualLPJ-GUESSrunwhereweswitchofthenitrogencycle,thedashedlinesshowthemodelrunswithdynamicnitrogencycling(compareWangetal.,2018).ρandR2showthecorrelationcoefficientsandR2valuesbetweentheLPJ-GUESSandtheTRENDYensemblemean.TheshadedareashowsthemodelspreadoftheindividualTRENDYmodels.

Fig.2:Compositeanomaliesingrossprimaryproduction(GPP)summedoverthetheElNinodevelopinganddecayingyearforallCPandEPElNiñoeventslistedintab.B1fortheindividualTRENDYmodels,theTRENDYcompositeandtheindividualLPJ–GUESSrun(compareWangetal.,2018).

Belowweaddressthereviewer’scommentspointbypoint.Weaddourrepliesinitalicsandhighlightsuggestedmodificationsinthemanuscriptinred.

Referee#2Inthismanuscript,Teckentrupetal.usedLPJ-GUESSforcedbymanipulatedclimatedatasetstostudytheinfluencesoftwoexpressionsofElNino(CPandEP)ontheterrestrialcarboncycle.TheauthorssuggestedthattheexpressionsofElNinoonlyinfluenceinterannualvariabilityofNBP(e.g.CPcausedlargerIAVinNBPthanEPattheglobalscale)butnotthelong-termchangeinNBP.TheyconcludedthattherelativefrequencyofCPandEPisnotcriticalinmodelsasCP/EPdidnotyielddetectablechangesinlong-termNBP.Thesciencequestionisinteresting,thestoryiswelltoldandthereisnomajorflawinthemethod.Thatbeingsaid,thereareafewquestionsthatpuzzledmeafterreadingthemanuscript,whichIhopetheauthorscouldclarifyabitbeforeIcouldsupportit.Wethankthereviewerfortheirassessmentandtheacknowledgementofourcontributions.

1. Oneofthenovelpointspresentedisthat“impact(ofCPandEP)onlongertimescalesisnotwellunderstood”.ElNino,eitherCPorEP,isknowntodominatetheinterannualvariabilityofterrestrialcarboncycling.ItisnotclearlystatedintheIntroductionwhywewouldexpectaninfluenceofCP/EPElNinoatlongertimescalesinthefirstplace.Inanotherword,woulditbeasurprisethatCP/EPElNinoexertnochangeonlong-termNBP,aswealreadyknownthatElNinoinfluencesIAVratherthanlong-termvariabilityofthecarboncycle.PerhapstherelativemorefrequentCPoccurrencesinthefuturecouldbeanissuelongtermbutthecurrentmodelsmaynotincludepropermechanisms(i.e.shiftinspeciescomposition,acclimations)toanswerthequestion.WeagreethatElNinostudieshavemostlyfocusedoninterannualtimescales.However,inarecentstudy,Parketal.2020foundthatdecadalvariabilityinENSOinfluencesthelongtermterrestrialglobalcarboncycle.Further,asnotedbythereviewer,ashiftinElNinopatternscouldaltercumulativenetbiomeproduction,whichmayaltercompetitivepatternsofplantspecies,bothofwhichmayinfluencethecarbonstoredinvegetationandsoil.Similarly,interannualvariabilityinprecipitationpatternsinducedbydifferenttypesofElNinomightchangevegetationdynamicsinsemiaridareas/savannaecosystems.Asaresult,wedothinkthefocusofourstudyiswarranted.OurresultsimplythattheexpressionofElNinodidnotleadtoanyofthechangesdescribedinearlierstudies.Weagreewiththereviewerthatitispossiblethatthismayinpartrelatetomissingmechanismsthatwouldcapturespeciescompositionchanges.Yetwethinkthisisunlikelytobetheexplanation,giventhatElNinoeventsareveryshort-livedandspatiallyvariablewhichlikelypreventsadirectshiftinvegetationinmostbiomesduetochangesinmeteorology.Whilstthissummaryofourfindingsagreeswiththereviewer’spointabove,westillhadtodotheexploratoryworktodeterminewhetherthiswasinfacttrue.Wehaveamendedthemotivationtextintheintroductiontomoreclearlycapturetheseissues:

‘AshiftinElNinopatternscouldchangecumulativenetbiomeproduction,whichmayaltercompetitivepatternsofplantspecies,bothofwhichmayinfluencethecarbonstoredinvegetationandsoil.Similarly,interannualvariabilityinprecipitationpatternsinducedbydifferenttypesofElNinomightchangevegetationdynamicsinsemiaridareas/savannaecosystems(cf.ScheiterandHiggins,2009;Whitleyetal.,2017).’

2. ThestudyisaimedatstudyingthesensitivityoftheterrestrialcarboncycletoCP/EPElNino.AndtheauthordidsobyreplacingtheclimateanomaliesduringCPtoEPandviceversa.CPisreportedtocauselargerglobalIAVthanEP.Yes–thatiscorrect.Myconcernsis:(usingglobalsimulationasanexample)isthislargersensitivityoftheterrestrialcarboncycletoCPisduetothechangesintheinherentclimatesensitivityofcarbonduringCP/EP,oristhissimplycausedbythegenerallylargerclimateanomaliesduringCP(Fig.B5).Iwouldassumethereasonisthelatter,astheinherentclimatesensitivityofcarboncycleisessentiallypredefinedbythemodel(inthiscaseLPJ-GUESS)structure,sowhatweseehere(IAVofNBPinCP>EP)isperhapsjustbecausetheIAVofclimateinCP>EP.Weagreewiththereviewerthatthemodelsimulationssuggestthattheresultslargelyfollowtheclimateanomalies.EmployingaDGVMprovidestheopportunitytoexplorepossiblelageffects,suchaschangesinfiredynamics,andchangesvegetationcomposition.TheselagprocessesthatmightbecapturedbyaDGVMallowedustoexplorethereviewer’squestionaboutthe‘inherentclimatesensitivityofcarbonduringCP/EP’.Infact,wefoundthattheperturbationsforcedonthevegetationweretoosmalltocausesignificantcarryovereffects,andconcludethereforethatclimateanomalieswerethekeycontrolfortheobservedchanges.Aswiththereviewer’spreviouspointabout‘long-termresponses’,wedonotthinktheanswerisself-evidentandfeltitwasimportanttoexplorethemodelsensitivities.

3. missedchanceonthespatialandphenologyofcarbonfluxes.WhileIhave

doubtsaboutthereporteddifferencebetweenCPandEPatinterannualorlongertimescales,Ifeeltheirdifferenceisperhapsmorepronouncedatseasonalscalesandspatial,whenCPandEPshowapparentcontrastingtemporalpatterns(e.g.Fig1).AswasalsonotedbyChyleketal.2018,thetimedelayofCO2riseafterSSTincreaseisoneofthepronounceddifferences,andthedifferenceisonlyaround3months.Focusingonlongertimescalesmighteasilyjustaveragedouttheseimportantcharacteristics.IthinktheauthorshavedoneanicejobindemonstratingthespatialdifferenceofcarbonsinksunderCP/EP,andtheseresultsperhapsworthmorehighlights.Weagreewiththereviewerthatchangesarelikelymoresignificantatseasonaltimescales.Asthereviewernotes,ourresultsshowthateveniftherearestrongimpactsonshortertimescales,theseeffectsdisappearondecadaltimescaleswhich

isakeyconclusioninourpaper.Wethereforechoosenottofurtherexploreshortertimescalesgiventhisisalreadysomethingthathasbeenreportedintheliterature.

4. Withthat,IwouldalsosayitmaybeastretchtosayCP/EPisnotcriticalinfuturemodels,astheirmajordifferenceislikelytobeclearerseasonallyandspatially(e.g.differentcarbonsinkdistribution,phenologyofcarbonuptake).WeagreewiththereviewerthatthedifferentexpressionsofElNinoaffecttheterrestrialcarboncycleoninterannualtimescalesandthatindividualeventspotentiallyhavelargeimpactonspecificregions.However,ourresultssuggestthattheperturbationslinkedtotheexpressionofElNinomightbetoosmalltotriggerchangesinvegetationdynamicsthatlastlongerthanaseasonorayear.Nevertheless,wehaveadjustedthefuturedirections:‘BasedonthisanalysiswesuggestthatourmodelsensitivitywouldlikelybesimilartothatdisplayedbytheotherTRENDYmodels,althoughwewouldanticipatesubtleregionaldifferences,particularinthetropicsifanalternativeDGVMhadbeenused’andtheconclusions:‘OurresultsthereforesuggestthattheimpactofdifferentexpressionsofElNinoonthecarboncycleonlongtimescalesislikelytobesmall.’

5. L11.Pleasespecifywhatkindoflongertimescaleeffect(i.e.decadalmean,decadalvariationortrend?)Weused‘longertimescaleeffect’todescribetheeffectaclimatewithonlyCPElNinooronlyEPElNinoeventsmighthaveonterrestrialvegetationafter45years.Wedonotanalysedecadalmean,variationortrendbutratherassesstheeffectbycomparingthefinalyearofthetwodifferentscenariostothatofthecontrolrun(wherebothexpressionsofElNinooccur).

6. L84andL104.IfCRU-NCEPv7covers1901-2016,whynotconsiderthe2015/2016ElNinointheanalysis.Wechosetheyear2013asthelastyearofourexperimentrunbecauseitisENSO-neutral.

7. L84.BysayingCRU,didyoumeanCRUNCEP.Thankyou,yesandwechangedthetextaccordingly.

8. L119-120.IamnotsureIunderstandhowtochoosethereplacementsforCPandEPcorrectly.WhythereisaneedtoresampleclimateanomaliesusingONIandhowdowelocatetheCPthatisusedtoreplaceaEP(inthesame10-yearwindowshowninFig1?).

WeusetheapproachaccordingtoYuandKim,2009.TheyusetheONIindextoidentifyElNinoeventswhichcomprisebothCPandEPElNinoevents.Basedonfourdifferentindices,theythenfurtherdifferentiatebetweenCPandEPElNinoevents.Wechangedthedescriptioninthe‘IdentificationofElNinoevents’sectionto:‘TheyfirstclassifiedElNinoeventsbasedontheOceanicNinoIndex(ONI)whichcomprisebothCPandEPElNinoevents.Basedonfourindices,theythenfurtherdifferentiatebetweenCPandEPElNinoevents.’WealsousetheONIindexasaguidanceforthereplacementoftheindividualElNinoevents.WereplacedanEPElNinoeventwithaCPtype(andviceversa),whenbotheventsstart,endandpeakatthesimilartimesintheyearaccordingtotheONIindexandhavesimilarmagnitudesintheONIindex(seemethods).Weupdatedthemethodssection:‘WeusedtheONIindextodefinethestart,endandstrengthoftheindividualElNinoeventsandresampledtheclimateanomaliesbasedontheONI.WereplacedanomaliesintheclimateforcingassociatedwithElNinoeventsaccordingtothebestfitindurationandamplitudeinONI,i.e.eventsthatstartandendatasimilartimeintheyearandhaveasimilartimingandmagnitudeofthepeakinONI.’

9. L210.DoesLPJ-GUESShaveacomponenttosimulatespeciescomposition?Thanksforpointingoutthispotentialconfusion.LPJ-GUESSrepresentsvegetationinformofplantfunctionaltypes,notindividualspecies.Wereplaced‘speciescomposition’with‘vegetationcomposition’.

10. B1-B4:Unitofcarbonfluxesinsupplementaryfigures.Perm2?Thankyouforpointingthisout,wechangeditaccordinglyandupdatedthefigures.

Scheiter,S.andHiggins,S.I.(2009),ImpactsofclimatechangeonthevegetationofAfrica:anadaptivedynamicvegetationmodellingapproach.GlobalChangeBiology,15:2224-2246.doi:10.1111/j.1365-2486.2008.01838.xYu,J.-Y.andKim,S.T.:IdentifyingthetypesofmajorElNiñoeventssince1870,InternationalJournalofClimatology,33,21052112,465https://doi.org/10.1002/joc.3575,https://rmets.onlinelibrary.wiley.com/doi/abs/10.1002/joc.3575,2013.Vitousek,P.M.Litterfall,nutrientcycling,andnutrientlimitationintropicalforests.Ecology65,285–298(1984).Whitley,R.,Beringer,J.,Hutley,L.B.,Abramowitz,G.,DeKauwe,M.G.,Evans,B.,Haverd,V.,Li,L.,Moore,C.,Ryu,Y.,Scheiter,S.,Schymanski,S.J.,Smith,B.,Wang,Y.-P.,Williams,M.,andYu,Q.:Challengesandopportunitiesinlandsurfacemodellingofsavannaecosystems,Biogeosciences,14,4711–4732,https://doi.org/10.5194/bg-14-4711-2017,2017.