Responses to reviewer 2 - tc.copernicus.org

Responsestoreviewer#2Reviewer2haspointedoutsomeimportantdetailsofthemanuscriptthatneedtobeclarifiedintherevisedversion.Theyhavealsomadeanumberofeditorial/technicalsuggestionsforimprovingthemanuscript.Weappreciatetheirinputandfirstrespondtothe5‘General’commentsandthenthe41‘Technical’commentsthatweremade.Includedattheendoftheresponseletterarethe3newfiguresthatweproposeaddingtothesupplementaryinformationoftherevisedmanuscript.Generalcomments:1.Itdoesn’treallymakesensetouseuncalibrated14CagesintheGeologicalSetting.IsuggestthattheexistingradiocarbondatesfromKellyandBennike(1992)shouldbere-calibratedusingMarine20andthesamedeltaRasthenewmarinecores.

Ononehandweagreethatitwouldbehelpfultoprovidere-calibrateddatesfortheoriginaldatapresentedinKellyandBennike(1992).WecandothisusingtheverylargedRuncertaintywehaveadoptedinthismanuscript,makingthepublisheddatesdirectlycomparabletoourresult-thesecanbeincludedinTable1.However,intheIntroduction/Backgroundweprefertocontinueusingtheraw14CagesfortheirsamplesandreferthereadertoTable1toseetheequivalentcalibrateddateusingMarine20andourdRuncertaintyenvelope.Usingtheraw14CagesinthemaintextensuresthattheIntroduction/backgrounddoesnotbecomeoutdatedwhennewconstraintsondRfortheregionemerge.IntheDiscussionwecanreportthecalibratedagerangesforeasiercomparisonwithourdata(whenneeded).OurdRof300+/-300isdesignedtoprovidearobustestimatethatcoverstherangesinpublishedliteratureandisusefulfordatingthelithostratigraphicboundariesinourrecords.

2)Theresultsectionisamixtureofdescriptionsandinterpretations.Exampleline:287-288,299-300,308-309,317,324,336-338.Isuggesttoclearlydividetheresultsectionintotwoseparatesub-sections:descriptionfollowedbyinterpretation.Thiswillallowthereadertoassessthedataandfollowthelogicintheinterpretations.

Wecanre-organizethissectionoftheresults,endingeachoftheLithologicUnitdescriptionswithsomeofthebasicinterpretationsofthedepositionalenvironment.ThebroaderinterpretationofthefaciessuccessionwillremainatthestartoftheDiscussion.

3)Figure5offersagreatsummaryofthemostimportantdata.ItwouldbereallynicetocomplimentthefigurewiththeCTscansfromthesuppl.materialorthehigh-resolutionpicturefromtheXRFscanner.ItisreallyapitythattheCTscansarehiddeninthesuppl.material.

Weagree,andseeingthecitationmetricsforthearticle,itisclearthattheimagesinthesupplementarymaterialarenotbeingwidelyviewed.WewillintroduceanewfigureinthemanuscriptthatshowtheCT-image,lithologicunitsandlocationofradiocarbondatesforeachcore.

4)Thewaytheageoftheindividualunitshasbeenconstraineddifferfrommoststudiesasitusesthemin.andmax.agesfromeachunittodefinetheagerange.However,astheradiocarbondatesarenotalwaysplacedoptimallyattheboundariesbetweenunitsthismakesitdifficulttocomparetheagerangesoftheunitsbetweenthedifferentcores.I

suggestthatanage-depthmodelforeachsedimentcoreisproduced.Thiswouldmakeitpossibletodeterminetheageattheboundaries(withanuncertainty)andalsoallowforafiguretobemadewheretheproxydata(fromfigure5)isplottedonanagescale.Thisisstandardprocedureanditwouldmakeagreatsupplementtothediscussionsection5.2-5.4.

Althoughwepresentalargenumberofnew14Cdates,wedonothavetheabilitytogeneratemeaningful,continuousagemodelsforeachofthecores.Thevastmajorityofsamplesweinvestigateddidnotcontainenoughforaminiferatoobtainradiocarbondates,sowehavenotbeenabletodateeachcoreattheresolution,orexactdepth,thatwouldwewant.Atthesametime,itisonlyLU5/4andLU1bwherewedonothavedatedsampleswithinafewcentimetersofthelithologicboundary.

Weexpectimprovedagemodelstoemergefromfuturework.Thiscouldcomefrom1)additionalradiocarbondates,2)Improvedconstraintsonthelocalreservoireffect(i.epairedsamples,210Pb,137Cs,tephra),and3)Improvedstratigraphiccorrelationbetweencores,andpotentiallyindependentagecontrol,fromgeneratingpaleosecularvariationrecords–aswasdonebyReillyetal.,2019forPetermannGlacier.Thesemeasurementsareinprogress,andmayallowustostackmoredatesfromthedifferentcoresontoamasterchronology.Therefore,wefeelthatwhileusingthisreallybroaddR,andinlightofnewdatathatshouldsoonemerge,itisnottherighttimetostartgeneratingdetailed,conventionalagemodelsforeachcore.

Westronglyfeelthattheapproachwehavetakenhere,isarobustwaytodatethemajoreventsandenvironmentalchangesthatoccurredthroughtheHolocene.BecausewehaveappliedsuchalargedR,improvedconstraintsondRand/orapplyingabaysianmodelingtechnique,willnarrowtheuncertaintyrangesoftheunitboundaries.Furthermore,sincewepresentourdataasmaximumconstraintsfortheonsetofthedifferentunits(usingtheyoungestdatefromtheunderlyingunit),theywillremaincorrect.However,werecognizethatpresentingmoreconventionalage-depthmodelsissomewhatnecessarytoevaluatesomeoftheindividualages,andelucidatethereasoningbehind‘accepting’or‘rejecting’someofthereturneddates.Inparticular,thereviewerhaslateraskedquestionsregardingwhywerejectedsomedatesfrom10-GCand8-PC.Assuggested,wehavecompiledage-depthfiguresforeachcore(ReviewFig.1),andcanincludethisinthesupplementaryinformation.Whilethesearenotagemodelsperse,theydoillustratethestratigraphicorderingofdatesineachcore,andhowtheyalignwithourproposedunitboundaryages.

Indoingthis,therearetwoimportantinsightswehavedrawnthatwerealizemustbeexplainedbetterintherevisedmanuscript.ThefirstistoexplainwhytheyoungagereturnedfromtheLU3/LU2boundaryin10GCisnotusedtodatethisboundary(comment21byReviewer2below).ThereasonsfornotusingthisdateforthebaseofLU2are1)olderagesthatareinstratigraphicorderarefoundintheothercoresforLU2,2)Thereisstrongevidenceforaperiodoferosion/non-depositionacrosstheboundaryat10GContheoutersill(ReviewFig.2tobeincludedinsupplementarymaterial),and3)Thissamplein10GCcomprised4cmofmaterial(50-54cm),withsomeofthiscomingfromtheoverlyingLU2andsomefrombelowinLU3.Itisnotpossibletoidentifyhowmanyofthedatedspecimenscamefromeachunit.

Thesecondpointwewillmorefullyillustrateanddiscussiswhyweidentifiedthelowestdatein8-PC(nearthebaseofLU5)asanoutlier.Therearetworeasonsforthisthatwewillhighlightintherevisedmanuscript.1)Theagedifferencebetweenthelowesttwodatesin8-PCis1680years,buttheyareonlyseparatedby24.5cm,suggestingthatthelowestmostageistooold,ordepositionwasnotcontinuous,and2)theolder,lowestmostagewasobtainedfromasamplecontainingmixedbenthicforaminfera,whichingeneralseemedtobemorepronetoreturningoldagescomparedtomono-specificsamplesofC.neoteretis.MoreclearlyrecognizingthatsedimentationduringLU5wasnotcontinuousisimportant.Thelaminatedunitclearlycontainsnumerouserosionalzonesofunknownduration,whicharelogicallymorefrequentnearthebaseoftheunit,inwhatweinterpretasagroundingzoneproximalsetting.Wehaveaddedsymbolsindicatingthisonourcompilationof14Cdates(Figure8inoriginalmanuscript),andincludeanadditionaldetailedinterpretedimagecontainingexamplesfrom7-PC,8-PCand9-PCtothesupplementarymaterial(ReviewFigure3).

5)Itisclearthatthemostchallengingunittointerpretisthediamictonunit3.Theunitdiffersfrommostotherunitswhicharelaminated.Itonlyresemblesunit6thelowermostunitwhichisinterpretedassubglacialtill.However,theauthorspreferanalternativeexplanationwhereunit3representsmassiveIRDdepositionduringaperiodwheretheicefrontismostretracted.Theyalsodiscussotherpossibilitiesbutfindthemlesslikely.Iamnotcompletelyconvincedbutagreethatitisdifficulttheinterpretationofunit3isnotstraightforward.Iwonderifunit4insteadcouldrepresenttheperiodwhereRGismostretractedandthatunit3representsthephasewhereitbeginstoreadvancesendingicebergs(IRD)intoSherardOsbornFjordagain.Ifcorrect,theonsetofreadvanceisc.6calkaBPwhichcoincideswiththegeneralcoolingtrendintheAgassizicecorerecord.

Weshouldstartbyclarifyingthatwedonotarguefor‘massiveIRDdeposition’duringLU3–whichimplieshighratesofIRDinput–butratherslowandsustaineddepositionofIRDintheabsenceofsignificantmeltwaterderivedsedimentsthatcontributedtodepositionofLU5,4,2and1.Wethinkthisisclearinthemanuscriptandwantedtoclarifythishere,asweareunsureifthereviewermeant‘rapidlydeposited’whentheysay‘massiveIRDdeposition’.IsitpossiblethatLU4representsthemostretractedphaseofRyderGlacier,andLU3anadvance?Wedonotbelieveso.Onebasicreasonistheoverallfaciessuccession.IfLU3representedare-advance,itbecomeshardtounderstandhowthiscantransitionintopotentiallyseasonallyopenwaterconditions,andgenerallyamelioratedclimateconditions,duringthedepositionofLU2.OurinterpretationisalsoconsistentwithwhathasbeendescribedbyKellyandBennike(1992)basedonmappinganddatingofraisedshorelinesandmorraines.Inparticular,aswestateintheGeologic,oceanographicandglaciologicsetting:“...peatdepositsoverwhichtheicemarginadvancedprovideanageof5100±13014CaBP(Station41),whileatSteensbyGlacier,reworkedmarinemacrofossilsinlateralmorainesyieldanageof4870±8014CaBP(Station34;KellyandBennike,1992).”.CalibratedusingIntcal20andMarine20respectively,theseprovideagesof5830±170calaBP(forthepeat)and4560±410calaBP(forthereworkedmolluscsontopofthelateralmorraines).Thesearemoreconsistentwithourinterpretationthatthere-advanceoccurredduringLU3/LU2.ObviouslythisrequiressomeclarificationintheDiscussionofthemanuscriptwherewewillmorecloselytietheearlierfindingsofKellyandBennike

(1992)toourresults.ProvidingupdatedcalibratedagesinTable1willhelpwiththiscomparison.Importantly,inonesensetherevieweriscorrect,wedonotknowwhenthere-advancebegan.Intherevisedmanuscript,wewillpointoutthattheinitialre-advancelikelybeganduringLU3,butcritically,amarinebasedglacierandicetongueweredefinitelyestablishedbytheonsetofLU2.Thisremainsconsistentwithourinterpretationofthefaciessuccession.

Technicalcomments:1.Line20:ChangetoGreenlandIceSheet.

Thishasbeencorrected.2.Figure1:AddGl.forglacierafterHumboldt,Petermannetc.AlsoaddIceSheetafterGreenland.

Thishasbeencorrected.3.Line43:ChangetoGreenlandIceSheet.

Thishasbeencorrected.4.Line61:ChangetoLastGlacialMaximum.

Thishasbeencorrected.5.Line66:ChangetoMöller.

Thishasbeencorrectedinthemaintextandreferencelist.6.Line75:AddglaciersafterPetermann.

Thishasbeenadded.

7.Line75:ChangetoNioghalvfjerdsfjordGlacier.Thishasbeenadded.

8.Line110:ChangetonorthGreenland.Thishasbeencorrected.

9.Line118:9390+-90dateisnotintable1Thiswasanoversightandthedatehasnowbeenincludedinthetable.

10.Table1:CombinewithTableandcalibratetheoldageswithMarine20.

Wecanincludere-calibratedMarine20datesinthetableusingthesamedRasweapplytomakecomparisonsmorestraightforward.

11.Line127:DeletecalaBPafter>9.5.

Thishasbeencorrected.12.Line129:Marktheice-dammedlakeonthemap.

Toourknowledgethelimitsandextentoftheproposedicedammedlakehavenotbeenmappedout.Forthisreason,wehavenotportrayeditsextentonthefigure.ItisnotclearwhetheritwouldhaveextendedfromSherardOsbornFjordacrosstoVictoriaFjord,oroccupiedamorerestrictedpartofWulffland–whichwouldultimatelydependontheglacier(s)configurationandrelativesealevelatthetimeofitsexistence.

13.Line131:Arethedatedshellsreworkedintothemoraine?Goodquestion.Infactwehadmadeaslighterrorinreportingthestationnamesandmixedupstation36and40.Theproblemisthataclearstationdescriptionisnotavailableforsomeinstances,orhardtointerpretfromthesummaryworkbeKellyandBennike,1992.Wehavegonebacktotheoriginalpublicationsdescribingthedatedmaterial(KellyandBennike,1985,BennikeandKelly,1987),andthesealsodonotprovideaverygoodgeologicalcontextforthesamples.WehaveaddedmoredetailedsamplelocaldescriptionsinTable1,andcorrectedourpreviouserror.Inshort,atstation36,inwesternWarmingLand,thedatedshellswerefrommarinesiltsthatwereyoungerthantheWarmingLandStademorrainesandreturnedandageof8210±12014CaBP(not6480±10014CaBPthatweoriginallywrote).Theageof6480±10014CaBPcamefromshellsfromamarinesiltinfrontofRyderGlacier(Station40)andfromwhatweunderstand,constrainsthetimingforiceretreattowardsthemodernposition.Bennike,O.&Kelly,M.1987:Radiocarbondatingofsamplescollectedduringthe1984expeditiontoNorthGreenland.Rapp.Grønlandsgeol.Unders.135,8-10.Kelly,M.&Bennike,O.1985:QuaternarygeologyofpartsofcentralandwesternNorthGreenland:apreliminaryaccount.Rapp.Grønlandsgeol.Unders.126,111-116.

14.Line136:ChangetoRyderGlacier.

Thishasbeencorrected.

15.Table2:Couldbemovedtosuppl.material.Table2containsthemeta-dataforthecoringstations.Weprefertoleavethetableinthemainpaper,asitisanimportantresourceforfuturestudies.Thosewhowanttolocatethecoresgeographicallyetc,shouldnotneedtodigthroughthesupplementarymaterial.

16.Figure4:Isnotshowingmuchandcouldbemovedtosuppl.material.Onceagain,weprefertokeepthisinthemainmanuscript.Weagreethatamoredetailedassessmentofthesubbottomdatashouldbeundertakeninfuturestudies.However,inthecontextofthismanuscript,thisfigureclearlyillustratesthegeneralthicknessofsedimentsontopofacousticbasement,andhowfarthecorespenetratedintothissedimentarycover.

17.Line167-168:Changelithifiedtocompacted.ThiswasalsosuggestedbyReviewer1,andhasbeenchangedto‘consolidated’.

18.Table3:13Cismissingforsample26.

Wehaveadded‘N/A’tothetable,asthesamplewastoosmalltoprovidea13Cmeasurement.

19.Line268:DeleteglacialafterHolocene.Thishasbeendeleted.

20.Line271:Changethroughtoand.

Thishasbeenchanged.

21.Figure5:On10-GCthe2450dateseemstobewithinunit3butitismarkedasunit2infigure8?

PleaseseeourdetailedresponsetoGeneralcomment4.Essentially,thisdatewasobtainedfroma4-cmthicksamplethatincludedsedimentsfromLU3andLU2,andwebelievethatahiatusexistsbetweentheseunitsatthisspecificstation.

On7-PCthedate7090seemstobeanoutlierbutitisnotmarkedwithred.

Tworeasonsforthis:1)Thetwoagesatthisdepthoverlapat1-sigma.ObvisoulytheywillnotifthedRisreduced,butfornowthiswasabasiccriteriaweappliedtoidentifyoutliers.2)Thetwodatesfromthatintervalarefromaplankticandabenthicsample,andwecannotbesurethatdifferentdRvalueswouldnotresolvetheapparentoffset.Therefore,wehavenotdiscardedeitherdateatthistime.

Also,whatisthesquarenextto7090representing?

Thiswasfromanearlierversionofthefigureandusedtodifferentiatewhetheritwasfromaplankticorbenthicforamsample.Wedidnotmeantocarrythisconventionovertothepublishedmanuscript,andhavemadeallsamplelocationscircles.

Whyisthelastdateincore8-PC/GCanoutlier?PleaseseeourdetailedresponsetoGeneralcomment4.WehaveadjustedourinterpretationandnowusethisagetodatethebaseofLU5,andprovideevidencethatsedimentationontheinnersillduringLU5wasdiscontinuous.22.Figure8:Idon’tunderstandwhythe14Cdatesinthisplothaveanormaldistribution?Alsoseegeneralcomment3.

Inthisfigurewehaveshownthelikelihooddistribution,meanand1-sigmarangeforeachsamplewhencalibratedusingMarine20andadRof300+/-300years.Theyarenotallexactlynormallydistributed,buttheydonohavetheskewnessthatwouldariseifweweremodelingtheagesandshowedaposteriordistribution.Wecanclarifythisinthefigurecaption.

23.Table4:Notimportantandcanbeomittediftheagedepthmodelsassuggestedingeneralcomment4willbemade.

Inlinewithourresponsetogeneralcomment4,weprovidetheseconventional‘age-depth’modelsinthesupplementarymaterial.However,westronglybelievethattheapproachofdatingthelithoistratigraphicboundariesmakesmostsenseatthistime.Anyconventionalagemodelswepublishatthispointwillmostcertainlybere-visedinthenearfuture.AlthoughthiswillreducetheuncertaintyintheagesfortheLU’stheywillverylikelytostaywithinthereportedagerangewehavedefinedbecauseoftheverylargedRwehaveused.

24.Line393:ChangetoNorthernHemisphere.

Thishasbeenchanged.25.Figure9:Reallyniceillustration–aretheradiocarbonagesfromEllesmerere-calibrated?

Yes,theywererecalibratedusingIntcal20.Weshouldhaveindicatedthisinthefigurecaptionandwilldosointherevisedms.

26.Line419:ChangetoMöller.

Thishasbeenchanged.

27.Line423:AddcalkaBPafter12.5.

Thishasbeenchanged.28.Line430:DeleteonelinFullford.

Thishasbeencorrected.29.Line435:ChangetoGrIS. Thishasbeenchanged.30.Figure10.Again,areallygreatillustration.CouldyouaddthelocationsoftheWarmingLandandKapFulfordStadesonthefigure?

Thesehavebeenaddedtothefigure.31.Line466:Itisstatedthat…LU3rangefrom6.3to3.9calkaBP.However,theupperpartofunit3in10-GCis245014CaBP.Whyisthisdateomittedinthesummary?

PleaseseeourdetailedresponsetoGeneralcomment4.32.Line508-510:Temperatureswerenot2.5-4ºCwarmeruntil6.2-6ka.TheywerestillhighbutthepeakwarmthoccurredinthebeginningoftheEarlyHoloceneandwasinsolationdriven.

Wehaveadjustedthissentencetoread:“andpeaklatesummerairtemperaturesinferredfromδ18OofchironomidsinSecretandDeltasølakesthatwere>2oCwarmerthenpresentuntil6.2calkaBP(Axfordetal.,2019;Lasheretal.,2017)”asthisiswhattheyreportintheirwork(seeFig.6inAxfordetal.,2019).

33.Line517:ZekollarimodelssuggestthatatleastpartoftheHansTausenicecapsurvivedtheHTM.

Wewereawareofthis,andhadstatedthatthesoutherndomeoftheicecaphaddisappeared.However,wecanmakethismoreclearbystatin::“Thistimingforglacieradvanceisconsistentwithcoolingseeninlakebasedtemperaturereconstructionsaround4calkaBP(Lasheretal.,2017)andtheoldestestimatedage(3.5to4.0calkaBP)foriceatthebaseofthesoutherndomeofHansTausenicecap,whichhaddisappearedduringtheMiddleHolocene-althoughnorthernpartsoftheicecaphadsurvived(MadsenandThorsteinsson,2001;Landviketal.,2001;Zekollarietal.,2017).”

34.Line520:Changeto:MiddleHolocene.

Thishasbeenchanged.35.Line521:Changeto:GrIS.

Thishasbeenchanged.Figure11:Again,agreatillustration.Maybeconsiderchangingthewhitecolorofthemodernicelimittored.ItwouldalsobegreattogettheKapFuldfordandWarmingLandmorainesonthemaps.

Thesehavebeenaddedtothefigure.36.Line540:or245014CaBP?SeecommentLine466.

WearenowclearinthemanuscriptthatthisageisnotusedtodatethebaseofLU2,becauseweinterpretthatthetransitionfromLU3toLU2at10-GCisnon-conformable.

37.Line564:ChangetoFunderetal.,2011.

Thishasbeencorrected.38.Line585:northGreenland.

Thishasbeencorrected.39.Line600:Søndergaardetal(2020)havepublishedapaperinClimateofthePastonthedeglaciationonInglefieldLand,SmithSoundandnaresStraitthatwouldfitintothediscussion.

Wehaveaddedthisreferencetosection5.3intheDiscussion(5.3MiddleHoloceneinlandretreatandcollapseofRyder’sicetongue)

40.Line601:CanthedifferencesinfjordphysiographyplayaroleinthedifferenttimingofretreatbetweenPetermannandRyderglaciers?SherardOsbornFjordisdeeperandpotentiallymoresusceptibletodynamiciceretreatcomparedtotheshallowerPetermannfjord.

Justbelowthisweargue(lines610-625)thatthephysiographyofSherardOsbornfjordisconducivetoglacierandiceshelfstability.Westatethatthereislittleevidenceforcollapseorsurgeevents,intheformofIRDpulses.Giventheargumentswehavealreadylaidout,wedonotfeeltheneedtobacktrackandsaythephysiographyofthefjordcouldhavebeenakeyfactorinRyder’sretreatfromthecoast.

41.Line650:Changeto:LateHolocene.

Thishasbeencorrected.