Abrupt glacial climate shifts controlled by ice sheet changes, Nature

Abrupt glacial climate shifts controlled by ice sheet changes, Nature
of 18
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  LETTER  doi:10.1038/nature13592 Abrupt glacial climate shifts controlled by icesheet changes Xu Zhang 1 , Gerrit Lohmann 1,2 , Gregor Knorr 1,3 & Conor Purcell 1,3 DuringglacialperiodsoftheLatePleistocene,anabundanceofproxy datademonstratestheexistenceoflargeandrepeatedmillennial-scalewarmingepisodes,knownasDansgaard–Oeschger(DO)events 1 .Thisubiquitous feature of rapid glacial climate change can be extendedbackasfaras800,000yearsbeforepresent( BP )intheicecorerecord 2 ,andhasdrawnbroadattentionwithinthescienceandpolicy-making communitiesalike 3 .Manystudieshavebeendedicatedtoinvestigat-ingtheunderlyingcausesofthesechanges,butnocoherentmechanismhasyetbeenidentified 3–15 .Hereweshow,byusingacomprehensivefullycoupledmodel 16 ,thatgradualchangesintheheightoftheNorth-ernHemisphereicesheets(NHISs)canalterthecoupledatmosphere–oceansystemandcauserapidglacialclimateshiftscloselyresembling DOevents.Thesimulatedglobalclimateresponses—includingabruptwarmingintheNorthAtlantic,anorthwardshiftofthetropicalrain-belts,andSouthernHemispherecoolingrelatedtothebipolarseesaw—aregenerallyconsistentwithempiricalevidence 1,3,17 .Asaresultofthecoexistence of two glacial ocean circulation states at intermediateheights of the ice sheets, minor changes in the height of the NHISsandtheamountofatmosphericCO 2 cantriggertherapidclimatetran-sitionsviaalocalpositiveatmosphere–ocean–sea-icefeedbackintheNorthAtlantic.Ourresults,althoughbasedonasinglemodel,thusprovideacoherentconceptforunderstandingtherecordedmillennial-scalevariabilityandabruptclimatechangesinthecoupledatmosphere–oceansystem,aswellastheirlinkagestothevolumeoftheintermediateice sheets during glacials. AcommonexplanationforDOcyclesinvolveschangesintheAtlanticmeridional overturning circulation 5,6 (AMOC), perhaps triggered by freshwaterforcing  3–6 .However,therepresentationoffreshwatersrcin,timingandmagnitude—usedtoforceabruptoceancirculationchanges— variesbetweenmodels,andpalaeoceanographicevidencefortheseoceancirculation changes and their relationship to freshwater perturbationremainselusive(Methods).Nevertheless,onefundamentalcharacter-isticofDOevents,possiblyofferingacluetotheirorigin,isthatalmostall events occurred during glacial periods when global ice volume was varying at intermediate levels 1,2,18 , suggesting a potential relationshipbetween the intermediate ice sheets and the existence of millennial-scaleclimatevariability  7 .Totestthishypothesis,weuseafullycoupledEarthSystemModel(COSMOS) 16 toassesshowchangesintheheightoftheNHISsaffectglobalclimateduringglacialperiods,andfindthatthegreaterheightof theNHISsforcesa relativelystrongerAMOC.Asdiscussed later, a nonlinear behaviour of the AMOC exists under theintermediate height of the NHISs, equivalent to a decrease in sea levelof  , 60m relative to today (Methods).Toinvestigatethecharacteristicsofthisabruptclimateshift,wecon-ductatransientsimulation(ISTran45)bygraduallyincreasingtheheightoftheNHISsfromthatoftheintermediateicesheets,atarateofdecreaseof1.8cmyr 2 1 equivalentsealevel(ESL)(Fig.1a–f).Inresponsetoalinearincrease in ice sheets, surface air temperature (SAT) in the northernNorth Atlanticis characterized by a gradual warmingrather thanby atwo-step strengthening process demonstrated by the AMOC strength(Fig.1a–c).Initially,overthefirst70years,SATinthenorthernNorthAtlantic gradually warms up by 4 u C, coincident with an increase inAMOCstrengthof3sverdrups(Sv,where1Svisequaltoavolumeflow rateof10 6 m 3 s 2 1 ).OncetheSATwarminghaspassedathermalthresh-old(about 2 2.5 u Cinourmodel),theweakAMOCincreasesabruptly by  , 10Sv within half a century, attaininga strong overturningmode,alongwithafurtherwarmingof  , 3 u CinthenorthernNorthAtlantic(Figs 1b, c and 2a, b).The SAT in the northern North Atlantic is characterized by strong  variabilityduringtheweakAMOCmode(Fig.1c,modelyears 2 150to0), also accompanied by variability in sea-ice concentration (SIC) andsubsurfacetemperature(SubST)(Figs1c–eand2c).Infact,theSAT,SICandSubSTvariabilitiesofthewarmclimatestate(strongAMOCmode)lie within the range of high variability of the cold climate state (weak AMOCmode)(Fig.1c–eandExtendedDataFigs3,4).WeattributethishighvariabilityassociatedwiththeweakAMOCmodetotheinterplay betweensea-icechangeandsubsurfacewarminginthenorthernNorthAtlantic.Indeed,thecoldclimate(weakAMOCmode)supportssea-iceformationandthebuild-upofastronghaloclineinthenorthernNorthAtlantic. However, under the intermediate height of the ice sheets, theaccumulatedwarmsubsurfacewatermasscangraduallydestabilizethewatercolumn,triggeringthereleaseofsubsurfaceheatandresumingawarm sea surface temporally  19,20 (Fig. 1c–e). Once the gradual surfacewarminginthenorthernNorthAtlanticpassesthethermalthreshold,the reduced SIC and the weakened stratification undermine the inter-actionbetweenthesea-icechangeandsubsurfacewarming,loweringtheinternalclimatevariability(forexampleSAT,SICandSubST)(Figs1c–eand 2c,d).Thus, the surface warmingacts toincreaseair–seaheat flux byreducingSIC,promotingafastattainmentofvigorousdeepconvec-tion.ThisfurtherwarmsthenorthernNorthAtlanticandcompletestheabrupttransitiontoawarmclimatestate(Fig.1b–d).Inparallel,thehigh variabilityinthenorthernNorthAtlanticundergoesashift,alongwiththemodetransitiontotheNordicSea,causingafluctuationofSICandresultinginlargetemperaturevariationsthere(Fig.2c,dandExtendedData Fig. 4).TheabrupttransitionfromweaktostrongAMOCmodesiscontrolledbyanonlinearresponseoftheglacialoceantogradualvariationsintheNHISheightbytwomechanisms:first,changesintheexportofseaicetotheNorthAtlantic,andsecond,adjustmentsofthesubpolarandsub-tropicalgyresystems,bothofwhichoccurinresponsetochangesinthewind field forced by variations in the NHIS height, especially the Lau-rentideIceSheet(LIS;ExtendedDataFigs2and5).Ourmodelsimula-tionsshowthatvariationsintheheightoftheLISleadtochangesinthegyrecirculationandsea-icecoveragebyshiftingtheNorthernwesterlies.AssociatedwithagradualincreaseoftheNHISs,themaximumwesterly windstressshiftsnorthwardsto , 57 u N.Consequently,thezonalwindstressovertheSouthLabradorSeadecreases,effectivelyweakeningtheexport of sea ice to the northeastern North Atlantic. This reduces theSIC and results in surface warming over the region, permitting openocean convection andtheformationof NorthAtlantic Deep Water.Inparallel,a strengthenedwind-stress curl accelerates the NorthAtlanticgyresystems,encouragingboththenorthwardsadvectionofheatviathe 1 Alfred Wegener Institute Helmholtz Centre for Polar and Marine Research, Bussestrasse24, D-27570 Bremerhaven, Germany.  2 MARUM-Center for Marine Environmental Sciences, University Bremen,Leobener Strasse, D-28359 Bremen, Germany.  3 School of Earth and Ocean Sciences, Cardiff University, Cardiff CF103AT, UK. 0 0 M O N T H 2 0 1 4 | V O L 0 0 0 | N AT U R E | 1  NorthAtlanticCurrentandverticalmixinginthesubpolarNorthAtlantic,the latteracting toincrease the ventilationofsubsurface warmwaters,resultinginlarge-scaleheatlossandfurtherconvectionintheopenocean.Togainadeeperunderstandingontheeffectofvariabilityinice-sheetheightonmillennial-scaleglacialclimateandthegoverningmechanismsofsuchchange,thestabilityoftheAMOCwithrespecttointermediateNHIS heights was further tested in COSMOS (Fig. 3 and Methods).Afterthehysteresis,branch1(pointsaandbinFig.3a,b)demonstratestheabruptincreaseinAMOC,branches2(pointsb–einFig.3a,b)and3(pointseandfinFig.3a,b)representthewarmglacialclimatestateandrapid Northern Hemisphere cooling, respectively, and branch 4 (pointsf–hinFig.3a,b)representsthecoldglacialclimatestate.Indeed,theasso-ciatedchangeinAMOCafterthishysteresiscurveisremarkablysimilartothegeneralshapeoftheDOcycle 1 (Fig.3b).Mostnotably,theAMOChysteresiscurvedemonstratesaglacialoceancharacterizedbyaclassicalbistableregime,correspondingtochangesinsea-levelofupto , 18masindicated by proxy data 21 (Fig. 3a). Within the bistable range, two dis-tinctglacialoceanmodeswithstrongandweakAMOC(notan‘off’mode)coexist under identical boundary conditions and are characterized by spatial patterns, for example abrupt warming in the North Atlantic, a 0.450.40    H  s   f 0–4–8–12–16    N   E   A   t   l  a  n   t   i  c   S   A   T   (   º   C   )    N   E   A   t   l  a  n   t   i  c   S  u   b   S   T   (   º   C   ) –1000100200Model year–16.0–15.5–15.0–14.5  S  o ut  h  er n O c e an S A T  (  º   C  )   75553515 NE A t  l   ant  i   c S I   C  (   % )   2015105 A M O C i  n d  ex (   S v  )   Model year0200400600–15.5–15.0–14.5–14.0  S  o ut  h  er n O c e an S A T  (  º   C  )   80604020 NE A t  l   ant  i   c S I   C  (   % )   2015105 A M O C i  n d  ex (   S v  )   205195    C   O    2    (  p .  p .  m .   ) 1850–4–8–12–16    N   E   A   t   l  a  n   t   i  c   S   A   T   (   º   C   ) 8765    N   E   A   t   l  a  n   t   i  c   S  u   b   S   T   (   º   C   ) abcdefghi jk l Figure 1  |  Transient simulations with gradually increasing NHIS height (ISTran45) and CO 2 level (TrGHG04). a  ,  g  , Transient forcing; b , h ,AMOCindex(AMOCmaximumintheNorthAtlantic north of 45 u N;  c ,  i , SAT index;  d ,  j , SICindex;  e ,  k  , SubST index (at 100 – 1,000m waterdepth) of the northern North Atlantic (average in56–65 u N,5–30 u W); f  , l ,SATindexintheSouthernOcean (60–76 u S zonal mean).  a  – f  , transientexperiment ISTran45;  g  – l , transient experimentTrGHG04. Bold lines show the 30-year running mean of the srcinal data (grey lines). Shadingsindicate one standard deviation of the indices in30-year windows. The vertical purple, blue andreddotted lines represent the starting points for thetransient simulations, abrupt AMOC transitionsand cooling in the Southern Hemisphere,respectively. Negative model years indicate thecontrol simulation of NHIS_0.4 (Extended DataTable 1). The mode transition in ISTran45 occursin about model year 60 (blue dashed line in  a  ).Thistransitionshowsthatandecreaseof  , 1.1minESL (corresponding to a 1.2% increase in NHISheight) can trigger the rapid glacial climate changewhen the NHISs are close to its height threshold.Hsf, height scaling factor (Methods). RESEARCH LETTER 2 | N AT U R E | V O L 0 0 0 | 0 0 M O N T H 2 0 1 4  northwardshiftofthetropicalrainbelts,andSouthernHemispherecool-ingrelatedtothebipolarseesaw,consistentwiththeexistingproxydatafor interstadial and stadial conditions, respectively  17 (Figs 1a–f and 3candExtendedDataFig.6).Inparticular,theweakmodesassociatedwiththeAMOChysteresisremaincharacterizedbyactiveverticalconvectionin the North Atlantic (Fig. 2a, b and Extended Data Fig. 2), consistentwithproxydatasuggestingthatmajorchangesinoceancirculationareidentified only for the DO–Heinrich events, not for other stadials asdepicted here 22 .Giventhesameice-sheetconfiguration,thestrongAMOCmode,incomparisonwiththeweakcirculation,ischaracterizedbyareducedSICin the northern North Atlantic, allowing heat loss to the atmosphereand enhancing the AMOC by the formation of North Atlantic DeepWater(Fig.4a).Inparallel,thewarmerSATassociatedwiththisreducedSICactstoenhancethelocalatmosphericlow-pressuresystem(Fig.4b),whichinturnbooststheGulfStreamwatermassandheattransportby the altered wind system over the North Atlantic ocean (Fig. 4c). Thisprocessrepresentsanatmosphere–ocean–sea-icepositivefeedbackmech-anismstimulatingfurthermeltingofseaiceandthusheatfluxfromtheocean, which acts to maintain convection in the North Atlantic.Our results demonstrate that, in combination with the positiveatmosphere–ocean–sea-icefeedback,onlyaslightvariationintheheightof the intermediate ice sheets (that is, a change of less than 2m in sealevel) is capable of instigating the abrupt AMOC mode transitions by an adjustment of the heat distribution and SIC in the North Atlantic(Figs1c–eand3).Althoughthehydrologicalcompensationduetotheglobalicevolumechangesisnotconsideredhere,salinificationordilu-tionintheNorthAtlanticshowsasimilareffecttoincreasingordecreas-ing the NHISs on triggering the mode shifts of the AMOC (ExtendedDataFig.7).However,weattributedtheexistenceofthebistableregimetothepositivelocalatmosphere–oceanfeedbackmechanismratherthanthehydrologicalbalanceintheAtlanticbasinassuggestedbynumerousother studies (see, for example, refs 3, 4, 6).ItisknownthatsomelargeDOwarmings,afterHeinrichevents,areaccompaniedbylargemillennial-scaleincreasesinatmosphericcarbondioxide( , 20p.p.m.) 23 ,withanimpactonaglobalscale 24 .Wehavethere-fore performed a transient experiment (TrGHG04) in which a linearincreaseinatmosphericCO 2 concentrationrangingfrom185to205p.p.m.isimposedover500years(ataconstantice-sheetsizeof40%maximumglaciallevel;Methods).Inthissimulation,asimilaryetevenmoreabruptNorthernHemispherewarmingissimulated(Fig.1g–l).Thisisduetoareleaseofadditionalsubsurfaceheat(resultingfromaweakenedAMOCthatexistsinresponsetotheincreasingCO 2 level) 25 totheatmosphere,thereaftertriggeringanabruptwarmingof  , 7 u CinthenorthernNorthAtlanticwithin20years.ThetimingandmagnitudeofthiswarmingareconsistentwithproxydatarevealingHeinrich–DOeventstobesimul-taneouswithincreasedglobalCO 2 (refs1,23,24),indicatingthatincreasesinatmosphericCO 2 alonecanalsotrigger 13 andevenamplifytheabruptNorthernHemispherewarmingundertheintermediateheightsoftheNHISs.The model results discussed here demonstrate for the first timethattheglacialoceanischaracterizedbyahysteresisbehaviourwithrespectto changes in the NHISs, corroborating a previous hypothesis that themillennial-scale variability during glacial periods is dependent on theexistenceofcontinentalicesheets 7 .Thebistableoceanregimeaccount-ingfortheglacialabruptclimateswitchesisrelatedtotheintermediate volumesoftheNHISs,consistentwiththefundamentalcharacteristicof themillennial-scalevariabilityintherecords 1,2,17,18 .Ourresultstherefore  40o 48o 56o 64o 72o 01,0002,0003,0004,0005,00030° S30° N45° N60° N0°   80° N72° N64° N56° N48° N40° N45° W30° W15° W0°15° E80° N72° N64° N56° N48° N40° N45° W30° W15° W0°15° E 01234567 a bc d –5–4–3–2–102468101214161718192021252831 Sv   01234567 ºC    D  e  p   t   h   (  m   ) 01,0002,0003,0004,0005,00030° S30° N45° N60° N0° Figure 2  |  Ocean circulation and internal SAT variability under cold andwarmclimatestatesinexperimentISTran45.  Coldandwarmclimatedefinedas model years 2 200 to 0 and 110–250 of ISTran45 (Fig. 1a and Methods). a  ,  c , AMOC ( a  ) and standard deviation of SAT ( c ) in the cold climate. b ,  d , AMOC ( b ) and standard deviation of SAT ( d ) for the warm climate.AMOC plots indicate zonally integrated meridional transport in the AtlanticBasin from 30 u S. Note the increased strength and northward shift of thedeep-water convection sites in  b  compared with that in  a  . The black rectanglein c indicatesthe area usedtocalculatetheNorthAtlanticSAT, SICandSubSTindices in Fig. 1. LETTER RESEARCH 0 0 M O N T H 2 0 1 4 | V O L 0 0 0 | N AT U R E | 3  provideanovelandbroadframeworkforunderstandingtheoccurrenceof rapid climate changes during glacials, indicating that only minorchangesinNHISs,atmosphericcarbondioxideand/orhydrologicalbal-anceintheNorthAtlanticmayinduceabruptoceancirculationchangeat the intermediate heights of the ice sheets.Thetimingofmillennial-scalefluctuationsinsealevelduringstadialsandinterstadialsisheavilydebated 21,26 .TheAMOChysteresispresentedheresuggeststhatatransitionfromaweak(orstrong)tostrong(orweak)AMOCmodeiscompatiblewithadecrease(orincrease)insealevelatthe end of stadials (or interstadials). Furthermore, our study indicatesakeytempo-spatialchangeininternalclimatevariabilityduringrapidglacialclimatechanges,providinganewdynamicframeworktoexplainthe recorded sea surface warming and its increased fluctuation during ice-rafting or at the end of stadials in the northern North Atlantic 27,28 .ThischaracteristiccanpotentiallyreconcilethelargeuncertaintyofseasurfacereconstructionsinthenorthernNorthAtlanticandNordicSeaduring glacials 27–29 . Nevertheless, new high-resolution records using improvedagemodelsarerequiredfortestingthesehypothesesfurther.Havingadditionallytestedtheresponseoftheclimatesystemtochang-ingorbitalconfigurations,wefoundnosignificantresponseinAMOCinourmodel(Methods).However,theicesheetsinourstudyareprescribedaccordingtotheLastGlacialMaximumconfigurationandarethusdecou-pledfromchangesinexternalforcingandtheinternalatmosphere–oceansystem.Furtherstudieswithsimilarclimatemodels,preferablyinclud-ing dynamic interactive ice sheet components, should be able to testwhether rapid changes inglacialclimate canbetriggered byvariability in ice sheets related to internal feedbacks within atmosphere–ocean–cryosphere systems 15,30 and to weak external forcing  9 . Online Content  Methods, along with any additional Extended Data display itemsandSourceData,areavailableintheonlineversionofthepaper;referencesuniqueto these sections appear only in the online paper. Received 17 January; accepted 10 June 2014.Published online 13 August 2014. 1. Dansgaard,W. etal. Evidenceforgeneralinstabilityofpastclimatefroma250-kyrice-core record. Nature  364, 218–220 (1993). 0510152025 a bc –8–7–6–4–3–2–1–0.5– 90° N60° N30° N30° S60° S90° S0°120° W120° E60° W60° E180°180° °C 01234Model year (kyr)Hsf051015                                                                                                                                                                                                                                                                                                                                                                                                                                                                      A                                                                                                                                                                                                                                                                                                                                                                                                                                                                      M                                                                                                                                                                                                                                                                                                                                                                                                                                                                               O                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          C                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 i                                                                                                                                                                                                                                                                                                                                               n                                                                                                                                                                                                                                                                                                                                                                                                                                                                    d                                                                                                                                                                                                                                                                                                                                                      e                                                                                                                                                                                                                                                                                                                                                 x                                                                                                                                                                                                                                                                                                                                                                                                                                                                                (                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              S                                                                                                                                                                                                                                                                                                                                                   v                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 )                                                                                                                              ahgbcdef ~18 m ESLLGM NHISPI NHIS 12341                                                                                                                                                                                                                                                                                                                                                                                                                                                                       H                                                                                                                                                                                                                                                                                                                                                s                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 f 0.40.35 abcdefgh Ice-sheet height                                                                                                                                                                                                                                                                                                                                                                                                                                                                      A                                                                                                                                                                                                                                                                                                                                                                                                                                                                      M                                                                                                                                                                                                                                                                                                                                                                                                                                                                               O                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          C                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 i                                                                                                                                                                                                                                                                                                                                               n                                                                                                                                                                                                                                                                                                                                                                                                                                                                    d                                                                                                                                                                                                                                                                                                                                                      e                                                                                                                                                                                                                                                                                                                                                 x                                                                                                                                                                                                                                                                                                                                                                                                                                                                                (                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              S                                                                                                                                                                                                                                                                                                                                                   v                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 )                                                                                                                               4 4 3424 0° Figure 3  |  AMOC hysteresis with respect to changes in NHIS height and itsrelationship to recorded abrupt climate variability. a  , Response of glacialocean to variations in the NHIS heights.  b , AMOC index for the experimentsassociated with the AMOC hysteresis (Methods).  c , Simulated annual meanSAT anomaly between the strong and weak AMOC modes with reconstructedtemperaturechanges(circles;Methods).Thestrong(orweak)modein c andinFig. 4 is an ensemble mean model results for points b-e (or f–h, a) of thehysteresis curve in  a  . Numbers in  b  correspond to the hysteresis branchesdefined in  a  . The AMOC strength in the pre-industrial (PI) control run is , 16Sv in our model 16 . The low AMOC indices corresponding to the weak modes do not represent an ‘off’ AMOC state (as indicated by the freshwater-hosing-induced Heinrich mode 6,16 ), but rather a weakened AMOC (Fig. 2 andMethods). Crosses in  a   represent the experiments performed to analyse theresponse of the glacial ocean to NHIS change; circles indicate the simulationsrelatedtothehysteresisbehaviourofglacialocean(Methods).TheNHISheightdifference between the Last Glacial Maximum (LGM) and the present-day levelisequivalenttoachangeinsealevelof  , 92m,indicatinganESLdifferencebetween points b and f of about 18m. abc 80° N70° N60° N50° N40° N60° W40° W 20° W0°6004005000200100–100–200–400–500–60080° N70° N60° N50° N40° N30° N60° W40° W20° W0°84621–1–2–4–6–8060° N50° N40° N30° N20° N80° W65° W50° W35° W20° W5° W84621–1–2–4–6–80mºCSv Figure 4  |  Annual mean anomaly maps between strong and weak AMOCmodes. a  , Mixed-layer depth anomaly (shaded) and absolute values for 15%(dashedlines)and90%(solidlines)seaiceconcentrationinthestrong(redline)and weak (blue line) modes.  b ,  c , Anomalies of SAT (shaded) and sea-levelpressure (contour) ( b ) and anomaly of barotropic horizontal stream function(shaded) and its absolute value in the strong AMOC mode (contour) ( c ). RESEARCH LETTER 4 | N AT U R E | V O L 0 0 0 | 0 0 M O N T H 2 0 1 4  2. Barker,S. etal. 800,000yearsofabruptclimatevariability. Science 334, 347–351(2011).3. Masson-Delmotte, V.  et al.  in  Climate Change 2013: The Physical Science Basis.Contribution of WorkingGroup I to the Fifth Assessment Report of theIntergovernmental Panel on Climate Change  (eds Stocker, T. F.  et al .) 383–464(Cambridge Univ. Press, 2013).4. Ganopolski, A. & Rahmstorf, S. Rapid changes of glacial climate simulated in acoupled climate model.  Nature  409,  153–158 (2001).5. Clark,P.U.,Pisias, N.G.,Stocker,T.F.&Weaver,A. J.Theroleofthethermohalinecirculation in abrupt climate change.  Nature  415,  863–869 (2002).6. Rahmstorf,S.Oceancirculationandclimateduringthepast120,000years. Nature 419, 207–214 (2002).7. Wunsch, C. Abrupt climate change: an alternative view.  Quat. Res.  65,  191–203(2006).8. Knorr,G.&Lohmann,G.RapidtransitionsintheAtlanticthermohalinecirculationtriggeredbyglobalwarmingandmeltwaterduringthelastdeglaciation. Geochem.Geophys. Geosyst.  8,  Q12006 (2007).9. Dima, M. & Lohmann, G. Conceptualmodelfor millennial climate variability: apossible combined solar–thermohaline circulation srcin for the , 1,500-yearcycle.  Clim. Dyn.  32,  301–311 (2008).10. Kwasniok, F. & Lohmann, G. Deriving dynamical models from paleoclimaticrecords: application to glacial millennial-scale climatevariability.  Phys. Rev. E   80, 066104 (2009).11. Arzel, O., ColindeVerdie`re, A. & England, M. H.The role ofoceanic heattransportandwind stress forcing in abrupt millennial-scale climate transitions.  J. Clim. 23, 2233–2256 (2010).12. Li, C., Battisti, D. S. & Bitz, C. M. Can North Atlantic sea ice anomalies account forDansgaard–Oeschger climate signals?  J. Clim.  23,  5457–5475 (2010).13. Banderas, R., Alvarez-Solas, J. & Montoya,M. Role of CO 2  and Southern Oceanwinds in glacial abrupt climate change.  Clim. Past   8, 1011–1021 (2012).14. Dokken,T. M., Nisancioglu, K. H.,Li,C., Battisti, D. S. & Kissel, C. Dansgaard–Oeschger cycles: interactions between ocean and sea ice intrinsic to the Nordicseas.  Paleoceanography  28,  491–502 (2013).15. Petersen,S. V., Schrag, D. P. & Clark, P. U. A new mechanism for Dansgaard–Oeschger cycles.  Paleoceanography   28,  24–30 (2013).16. Zhang, X., Lohmann, G., Knorr, G. & Xu, X. Different ocean states and transientcharacteristics in Last Glacial Maximum simulations and implications fordeglaciation.  Clim. Past   9,  2319–2333 (2013).17. Voelker, A.H.L.Globaldistributionofcentennial-scale recordsforMarineIsotopeStage (MIS) 3: a database.  Quat. Sci. Rev.  21,  1185–1212 (2002).18. Schulz,M., Berger, W. H., Sarnthein, M. & Grootes, P. M. Amplitude variations of1470-yearclimateoscillationsduringthelast100,000yearslinkedtofluctuationsof continental ice mass.  Geophys. Res. Lett.  26,  3385–3388 (1999).19. Rasmussen,T.L.&Thomsen,E.TheroleoftheNorthAtlanticDriftinthemillennialtimescale glacialclimate fluctuations. Palaeogeogr. Palaeoclimatol. Palaeoecol. 210, 101–116 (2004).20. Kim, J.-H.  et al.  Pronounced subsurface cooling of North Atlantic waters offNorthwestAfricaduring Dansgaard–Oeschgerinterstadials.  Earth Planet.Sci.Lett. 339–340,  95–102 (2012).21. Siddall,M.,Rohling,E.J.,Thompson,W.G.&Waelbroeck,C.MarineIsotopeStage3sea level fluctuations:data synthesis and newoutlook.  Rev. Geophys.  46, RG000226 (2008).22. Peck, V. L., Hall,I. R., Zahn, R. & Scourse, J. D. Progressive reduction in NEAtlantic intermediate water ventilation prior to Heinrich events: response toNW European ice sheet instabilities?  Geochem. Geophys. Geosyst.  8,  Q01N10(2007).23. Ahn,J.&Brook,E.J.AtmosphericCO 2 andclimateonmillennialtimescalesduringthe last glacial period.  Science  83,  83–85 (2008).24. Barker,S.&Knorr,G.AntarcticclimatesignatureintheGreenlandicecorerecord. Proc. Natl Acad. Sci. USA 104,  17278–17282 (2007).25. Gregory,J. M.  et al.  A model intercomparison of changes in the Atlanticthermohaline circulation in response to increasing atmospheric CO 2 concentration.  Geophys. Res. Lett.  32,  L12703 (2005).26. Arz,H.W.,Lamy,F.,Ganopolski,A.,Nowaczyk,N.&Pa¨tzold,J.DominantNorthernHemisphereclimatecontrolovermillennial-scaleglacialsea-levelvariability. Quat.Sci. Rev.  26,  312–321 (2007).27. Peck, V. L., Hall,I. R., Zahn, R. & Elderfield, H. Millennial-scale surface andsubsurfacepaleothermometry from the northeast Atlantic, 55–8ka BP . Paleoceanography   23,  1–11 (2008).28. Jonkers, L.  et al. A reconstruction of sea surface warming in the northernNorth Atlantic during MIS3 ice-rafting events.  Quat. Sci. Rev.  29,  1791–1800(2010).29. Waelbroeck,C. etal. Constraintsonthemagnitudeandpatternsofoceancoolingatthe Last Glacial Maximum.  Nature Geosci.  2,  127–132 (2009).30. Macayeal,D.R.Binge/purgeoscillationsoftheLaurentideIceSheetasacauseofthe North Atlantic’s Heinrich events. Paleoceanography  8,  775–784 (1993). Acknowledgements  We thank the colleagues in the Paleoclimate Dynamics group attheAlfredWegenerInstituteHelmholtzCentreforPolarandMarineResearch(AWI)inBremerhavenforusefuldiscussions,andI.Hallforcommentsonapreviousdraftofthemanuscript. G.K. acknowledges helpful discussion with S. Barker. We thank the AWIcolleagues who provide technical support on graphics and keep the AWIsupercomputer running. This study is promoted by Helmholtz funding through thePolar Regions and Coasts in the Changing Earth System (PACES) programme of theAWI. Funding by ‘Helmholtz Climate Initiative REKLIM’ (Regional Climate Change), ajointresearchprojectoftheHelmholtzAssociationofGermanresearchcentres(HGF),is gratefullyacknowledged (G.K.). Author Contributions  X.Z., G.L. and G.K. developed the research and designed theexperiments.X.Z. conducted the modelsimulations and analysed the data. X.Z., G.L.andG.K.interpretedtheresults.X.Z.ledthewrite-upofthemanuscriptwithsignificantcontribution by C.P. and the other authors. All authors discussed the results andcontributed to the preparation of the manuscript. Author Information  Reprints and permissions information is available Theauthors declare no competing financial interests.Readersarewelcometocommentontheonlineversionofthepaper.Correspondenceand requests for materials should beaddressed to X.Z. ( LETTER RESEARCH 0 0 M O N T H 2 0 1 4 | V O L 0 0 0 | N AT U R E | 5
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks