J. Meteor. Res.  2014, Vol. 28 Issue (5): 803-819   PDF    
http://dx.doi.org/10.1007/s13351-014-4006-6
The Chinese Meteorological Society
0

Article Information

LIU Wei, LIU Zhengyu. 2014.
Assessing the Stability of the Atlantic Meridional Overturning Circulation of the Past, Present, and Future
J. Meteor. Res., 28(5): 803-819
http://dx.doi.org/10.1007/s13351-014-4006-6

Article History

Received January 16, 2014;
in final form March 11, 2014
Assessing the Stability of the Atlantic Meridional Overturning Circulation of the Past, Present, and Future
LIU Wei1 , LIU Zhengyu2,3    
1 Scripps Institution of Oceanography, University of California, San Diego, La Jolla, CA 92093, USA;
2 Laboratory of Climate, Ocean, and Atmosphere Studies, Peking University, Beijing 100081, China;
3 Center for Climatic Research, University of Wisconsin-Madison, Madison, WI 53706, USA
ABSTRACT:This paper is a review of the recent developmenTof researches on the stability of the Atlantic meridional overturning circulation (AMOC). In particular, we will review recent studies that attempt to best assess the stability of the AMOC in the past, present, and future by using a stability indicator related to the freshwater transport by the AMOC. These studies further illustrate a potentially systematic bias in the state-of-the-art atmosphere-ocean general circulation models (AOGCMs), in which the AMOCs seem to be over-stabilized relative to that in the real world. This common model bias in the AMOC stability is contributed, partly, to a common tropical bias associated with the double intertropical convergence zone (ITCZ) in most state-of-the-art AOGCMs, casting doubts on future projection of abrupt climate changes in these climate models.
KeywordsAMOC     stability indicator     freshwater transports     ITCZ     deglaciation    
1. Introduction

The Atlantic meridional overturning circulation(AMOC)consists of a lower limb of denser, coldersouthward return flow at depth, and an upper limbof returning flow of near-surface warm waters(Fig. 1). Presently, it is estimated that the AMOC hasa mean magnitude of 18 ± 1. 0 Sv(1 Sv = 106m3s-1), based on direcTobserv ations at 26. 5°N by theRAPID/MOCHA array between April 2004 and April2009(Cunningham et al., 2007; Kanzow et al., 2007; Johns et al., 2011). The AMOC affects the climateby its significant northward oceanic heat transport, about 1. 3 PW(1 PW = 1015W)over 24°-26°N(Hall and Bryden, 1982; Trenberth and Caron, 2001; Johns et al., 2011). A collapse of the AMOC would inducesignificant climate impacts, such as a bipolar see-sawresponse with a pronounced cooling over the North Atlantic and warming in the Southern Hemisphere(Drijfhout et al., 2011; Shakun et al., 2012), the shifTofthe intertropical convergence zone(ITCZ)as well asthe Atlantic rainfall(Vellinga and Wood, 2002), themodulation of North Atlantic sink for CO2(Schuster and Watson, 2007), and variations in marine ecosystems(Schmittner, 2005). Generally, a collapse of theAMOC has often been associated with the multipleequilibria of the AMOC, and has been used to indicatethe observed abrupt changes in past climates(e. g., Clark et al., 2002). In the Holocene(around last 10000years), the absence oFrapid climate changes seems tosuggest that the AMOC has remained in a mono-stableregime(Grootes et al., 1993). Nevertheless, the futureclimate response to global warming may induce an intensified hydrological cycle, which may destabilize theAMOC(Broecker et al., 1985; Rahmstorf, 2002) and cause potential abrupt climate changes. Therefore, itis important to investigate the AMOC stability under various climates. Herein, one issue arises, i. e., forthe real climate system of the present, the stability ofthe AMOC cannot be assessed with sensitivity experiments, but such a diagnostic indicator is highlyneeded. Besides the assessment, such an indicatoralso enables a stability intercomparison among climatemodels, which then can be used for selecting somevalidated models to predict the future changes of theAMOC. In this study, we will review the studies so farof the AMOC stability in the past, present, and futureclimates.

Fig. 1. A simplified schematic of the AMOC showingboth the overturning and gyre recirculation components. Warm water flows northward in the upper ocean(red), gives up heat to the atmosphere(atmospheric flow gainingheat represented by the changing color of broad arrows), sinks, and returns as a deep cold flow(blue). Latitude ofthe 26. 5°N AMOC observations is indicated. Note thatthe actual flow is more complex. For example, see Boweret al. (2009; their Fig. 1)for the intermediate depth circulation in the vicinity of the Grand Banks and Biastochet al. (2008; their Fig. 2)for the mid-depth circulationaround South Africa, showing the importance of eddies intransferring heat and salt from the Indian Ocean to theAtlantic Ocean. [From Srokosz et al., 2012]

The rest of the paper is structured as follows. Section 2 describes some insights on the AMOC multipleequilibria from the paleoclimate perspective. Fromthese insights, substantial theoretical and modelingstudies of the AMOC stability are reviewed in Section3, especially on the topic of the AMOC stability indicator. Then, based on the indicator, the stability ofthe AMOC under the present and future climates is assessed in Sections 4 and 5, respectively. In particular, a critical issue is posed that, due to a common double ITCZ problem, most state-of-the-art atmosphereocean general circulation models(AOGCMs)may exhibit a systematic bias in simulating the AMOC stability. This may distort the estimation of current AMOCstability or prediction of the AMOC behavior in futureclimate. Concluding remarks and further discussionare given in Section 6. 2. The AMOC stability in past climates

Past climate records have shown clear evidenceof various types of abrupt climate changes on millennial timescales, notably the Dansgaard/Oeschger(D/O)cycles(Dansgaard et al., 1993) and Heinrichevents(see reviews oF Rahmstorf, 2002; Clark et al., 2002, 2007). These cycles are characterized by opposite temperature responses in the Northern and Southern Hemispheres in a so-called bipolar "see-saw" response(e. g., Stocker and Johnsen, 2003), as indicatedin ice cores in Greenl and and Antarctic. These abruptchanges have been speculated to link with abruptchanges in the AMOC and its associated heat transport(e. g., Ganopolski and Rahmstorf, 2001; Liu et al., 2009). In one hypothesis, the abrupt D/O events areproposed to be caused by the multiple equilibria orbi-stability of the AMOC, i. e., the switch of interstadial and stadial modes in the D/O cycles is inducedby transitions between different equilibrium states ofthe AMOC(Broecker et al., 1985). This poinTof viewis consistent with the reconstruction of the North Atlantic Deep Water(NADW)production. As shown inSarnthein et al. (1994), the NADW production is reduced from the interstadial mode to the stadial modein a D/O event, or interrupted in a Heinrich event, which represents a weaker or collapsed AMOC herein(Fig. 2).

Fig. 2. (Left panel)Simulated D/O and Heinrich events. (a)Forcing, (b)Atlantic overturning, (c)Atlantic salinity(S)at 60°N, (d)air temperature in the northern North Atlantic sector(60°-70°N), and (e)temperature over Antarctica(temperature values are given as the difference from the present-day climate, ΔT)(from Ganopolski and Rahmstorf, 2001). (Right panel)Data-model comparison for several benchmark time series. (A)June insolation at 60°N(purple)(Berger, 1978) and atmospheric CO2 concentration(green)(Joos and Spahni, 2008), ppmv means parts per million byvolume. (B)Sea level from the reconstruction(gray)(Peltier, 2004) and model(meters of equiv alent global sea level(ESL)for meltwater). (C)Freshwater fluxes(FWF)in the model. (D)Pa/Th ratio at Bermuda(GGC5 core)as a proxyfor the AMOC strength(McManus et al., 2004), and model maximum AMOC transport(below 500 m). (E)Greenl and surface air temperature(SA T)based on Greenl and Ice Sheet Project 2(GISP2)δ18O reconstruction with boreholetemperature calibration(Cuffey and Clow, 1997) and in the model(model offset by -3‰). (F)Antarctic surface airtemperature based on Dome C δ18O reconstruction(Jouzel et al., 2007) and in the model. (G)Sea surface temperature(SST)from the Iberian Margin from reconstructions(Waelbroeck et al., 1998; Bard et al., 2000) and model. (H)SSTfrom the Cariaco Basin from reconstruction(Lea et al., 2003) and model(model offset by 4‰). (I)Rainfall in CariacoBasin from reconstruction(Peterson et al., 2000) and model. In(B)to(I), gray is used for the reconstruction, and red and blue for experiments DGL-A and DGL-B, respectively. The five circles on DGL-A in(D)represent the glacial state(GLA; 19 ka), H1(17 ka), PreBA(14. 7 ka), Recovery(REC; 14. 5 ka), and BA(14. 35 ka). All model variables are annualmeans with a 20-yr running average. Overall, model simulations, especially DGL-A, are in good agreement with theproxy records, especially outside the tropical Atlantic. BP means before present. [From Liu et al., 2009]

Despite a strong body of evidence that associatesthe AMOC with the abrupt changes(e. g., Broecker et al., 1985; Clark et al., 2002; McManus et al., 2004), the trigger mechanism of these abrupt changes of theAMOC has remained yet unclear. One c and idate isthe meltwater pulse. For instance, one major meltwater pulse, the Meltwater Pulse 1A(MWP-1A), asregarded, triggers the transition from the Heinrichevent 1(H1)to Bø lling-Allerø d(BA). However, thisc and idate is challenged by uncertainties in the leadlag relationship between the meltwater pulse and theAMOC change. Due to a poor chronology of the reconstructions, the meltwater history and even the locationof meltwater pulses(such as MWP-1A)have remainedcontroversial(Clark et al., 1996; Peltier, 2005; Stanford et al., 2006; Deschamps et al., 2012). As a result, it is di°cult to establish a precise chronological orderbetween the meltwater pulses and the AMOC changes. If an abrupt AMOC change does not follow a meltwater discharge within a self-adjustment scale(severalhundred years), the AMOC change can be viewed asa response of a bi-stable AMOC to a smooth changeof freshwater forcing(e. g., Ganopolski and Rahmstorf, 2001; Knorr and Lohmann, 2003; Weaver et al., 2003). Otherwise, the AMOC change is likely forced by theabrupt change of meltwater forcing due to the instability of ice sheet(e. g., MacAyeal, 1993), instead ofrelated to the mono-stable AMOC itself(e. g., Liu et al., 2009). Because currenTobserv ations are not sufficient to distinguish the lead-lag time between abruptchanges in the AMOC and corresponding meltwaterpulses unambiguously, many abrupt climate eventsare still considered most likely to be caused by a bistable AMOC. This suggests that the instability ofthe AMOC could play a key role in past rapid climatetransitions. Meanwhile, such changes in the AMOCare of great importance to either presenTor future climates, considering that the potential for an AMOCcollapse is a key uncertainty in future climate projections. Thereby we will continue to review the AMOCmultiple equilibria as well as its stability indicator. 3. The AMOC multiple equilibria and stability indictor

The nonlinear nature of the AMOC results in theexistence of its multiple equilibria. One pioneeringstudy was from Stommel(1961)who used a simpletwo-box model to propose a positive salinity advection feedback in destabilizing the AMOC and leadingto a bi-stable circulation. The Stomnel box modelwas further examined(Mu et al., 2004) and extendedto an inter-hemisphere box model(e. g., Rooth, 1982; Welander, 1986; Rahmstorf, 1996, see Fig. 3). Lateron, the existence of the AMOC multiple equilibria hasbeen demonstrated in ocean general circulation models(OGCMs), ranging from a three dimensional model ofa single, flat basin(Bryan, 1986)in which a conceptof "halocline catastrophe" was proposed, to zonallyaveraged global models(e. g., Marotzke et al., 1988; Stocker and Wright, 1991a, b; Hughes and Weaver, 1994) and global OGCMs(e. g., Marotzke and Willebr and, 1991; Power and Kleeman, 1993; Weaver et al., 1993). Furthermore, the AMOC multiple equilibria were explored in two types of more advancedmodels, the earth system models of intermediate complexity(EMICs) and atmosphere-ocean general circulation models(AOGCMs). Rahmstorf et al. (2005)compared 11 EMICs and found that all the modelssimulate AMOCs with significant multiple equilibria(Fig. 4). The bi-stability of the AMOC in EMICswas further demonstrated in the limiTof low mixing(Hofmann and Rahmstorf, 2009). In contrast to therobust bi-stability of EMICs, there has been little evidence of the AMOC multiple equilibria in state-of-theart AOGCMs. For example, in AOGCMs from Coupled Model Intercomparison Project phase 3(CMIP3)(Stouffer et al., 2006), most AMOCs exhibit monostable behaviors. The circulations recover to theiroriginal conveyor states after a termination of freshwater perturbation(Fig. 5)

①So far, two exceptions can be found among current AOGCMs: the GFDL R30 and FAMOUS models; see Manabe and Stouffer(1988) and Hawkins et al. (2011), respectively.

Fig. 3. (a)A simple 4-box model of cross-hemisphericthermohaline flow. NADW forms in box 2; its outflow towards box 1 is controlled by the density difference betweenboxes 2 and 1. Salinities in the boxes are determined bythe flow and the surface freshwater fluxes entering boxes 1, 2, and 3. Only two of these three fluxes are independent, since their sum must vanish in a steady state. Therefore, the surface freshwater fluxes are portrayed as two atmospheric vapor transports F1 and F2. (b)The three flowregimes(solid)of the box model. The dashed line is an unconditionally unstable solution, and S is the saddle-nodebifurcation point. [From Rahmstorf, 1996]
Fig. 4. Hysteresis curves found in the model intercomparison from(a)coupled models with 3-D global ocean models, and (b)those with simplified ocean models(zonally averaged or, in case of the MIT_UWash model, rectangular basins). Curves were slightly smoothed to remove the effect of short-term variability. Circles show the present-day climate stateof each model. [From Rahmstorf et al., 2005]
Fig. 5. Time series of the AMOC intensity evolution in the 1. 0-Sv water-hosing experiments for the CMIP/PMIPmodels. [From Stouffer et al., 2006]

There is so far no consistent explanation why theAMOC tends to be bi-stable in intermediate models, but not in AOGCMs. The systematic lack of multiple equilibria in AOGCMs, however, seems to indicatethat certain factors in common are inclined to overstabilize the AMOC in these models. Great endeavorshave been put in seeking such factors in several aspects. From the ocean aspect, some studies suggestedthat increasing oceanic diapycnal diffusivity caninhibit the AMOC multiple equilibria by generatinga more diffusive and linear circulation(Manabe and Stouffer, 1999), whereas others proposed that strongdiffusivity would enhance the multiple equilibria viaa strong oceanic upwelling(Prange et al., 2003; Nof et al., 2007). From the atmosphere aspect, stronginternal atmospheric variability was found to set upa stochastic forcing in generating the bimodality ofthe AMOC(e. g., Cessi, 1994; Timmermann et al., 2003), which was further supported by results fromseveral intermediate models. The wind stress feedbackwas argued to stabilize the AMOC in modern climate(Mikolajewicz, 1996; Schiller et al., 1997)but to destabilize the AMOC under glacial climate(Arzel et al., 2008). From the coupling aspect, the AMOC multiple equilibria were suggested to be suppressed by astrong ocean-atmosphere coupling and associated precipitation response(Yin et al., 2006; Yin and Stouffer, 2007). In brief, great divergences exist in the abovemodel-based arguments, which provide a strong motivation to assess the AMOC stability against the realworld.

To evaluate the stability of the AMOC offline ina complex climate model, and more importantly inthe real world, one has to formulate a diagnostic indicator of the AMOC stability. By using a freshwater budget(discussed later) and treating the Atlantic and Arctic basin as a united "box"(Fig. 3a), Rahmstorf(1996)first proposed the freshwater export bythe AMOC across the southern boundary(approximately 34°S)of the Atlantic, abbreviated as FOT, asa diagnostic indicator of the AMOC stability. Thisindicator can be used to evaluate the AMOC stability because it captures the salinity-advection positivefeedback(Stommel, 1961)that is critical for the bistability of the AMOC. Physically, consider an AMOCstate with a freshwater export. An initial pulse of perturbation freshwater flux in the North Atlantic weakens the AMOC and therefore reduces the exporT offreshwater. This will lead to an accumulation of freshwater that further enhances the initial pulse of freshwater perturbation, and eventually result in a collapseof the AMOC. In practice, the freshwater transportby the AMOC is calculated as the freshwater transport by the zonal mean overturning circulation.

This indicator was later adopted by Weber et al. (2007), who found that all the models in the Paleoclimate Modelling Intercomparison Project(PMIP)except ECBilt/CLIO exhibit a freshwater import acrossthe southern border of the Atlantic basin, which correctly suggested a mono-stable AMOC in these models(Stouffer et al., 2006). However, the only exception, ECBilt/CLIO, though with a freshwater export, failed to indicate a bi-stable AMOC, since its AMOCwas still apt to recover in a pulse hosing experiment(de Vries and Weber, 2005). Therefore, FOT does notappear to be an accurate stability indicator for theAMOC, at least in some EMICs or AOGCMs. One probable cause is that FOT includes the Arctic freshwater budget and therefore does not represent the realnet freshwater forcing exerted on the AMOC.

In a later study, Dijkstra(2007)proposed an alternative indicator Σ, which is defined as the freshwater transport convergence by the AMOC for theAtlantic basin, i. e., the net freshwater transport between the southern and northern boundaries(approximately 34°S and 60°N). This indicator was subsequently demonstrated largely valid in an OGCM coupled with an energy-balance atmosphere model(Huisman et al., 2010). Nevertheless, one concern of thedefinition of Σis that the northern boundary of theAtlantic basin is placed at 60°N. This definition excludes the GIN Seas(Greenl and, Icel and, and Norwegian Seas)region, a major region for the NADWformation(e. g., Schiller et al., 1997; Holl and et al., 2007; Renold et al., 2010). As related to the status ofBering Strait(e. g., Hu et al., 2008, 2012), the freshwater transport through this region is expected to havea significant effecTon the AMOC stability(e. g., Holl and et al., 2001; Komuro and Hasumi, 2005; Oka and Hasumi, 2006; Rennermalm et al., 2006), so this indicator may not correctly indicate the AMOC stabilityin some AOGCMs(Liu and Liu, 2013)due to the lackof the freshwater transport via the GIN Seas.

Liu and Liu(2013)proposed a refined indicator, ΔMov, which includes the GIN Seas region and is defined as the net freshwater transport betweenthe southern and northern boundaries(approximately34°S and 80°N). Same as FOT and Σ, ΔMovis derived from a decomposition of freshwater transport and a basin-integrated freshwater budget. In the Atlantic, the meridional freshwater transport can be divided into two parts: the meridional overturning part(Mov)that is thought to be associated with the AMOC and the azimuthally asymmetric part(Maz)that isassumed to be associated with the wind-driven gyrecirculation:

In Eqs. (1) and (2), S0 is reference salinity(unit:psu); the overbar and brackets denote the zonal integration and zonal averaging along a latitude circle, respectively; v is the velocity normal to the section and s is salinity; v' and s'are deviations from theiralong section means. Here, it is worth noting thatMov and Mazare just simply geometric decompositions for representing the horizontal and overturningcontributions to the freshwater transport, which donot serve as a neat division between the contributionsfrom different physical processes such as the NADWformation and the gyre circulation. Thus, by neglecting the effects of diffusion, the freshwater budgeToverthe Atlantic basin(roughly between 34°S and 80°N)can be approximately estimated as a balance betweenthe net evaporation(Enet) and the freshwater transport through the southern and northern boundaries, i. e., where Enet = E - P - R - M +Br, i. e., the sum ofevaporation E, precipitation -P, runoff -R, sea icemelting -M, and brine rejection Br due to sea icemelting. The subscripts S and N denote southern and northern boundary, respectively. MazS and MazNareassociated with the gyre circulation, whilst MovS and MovNare associated with the overturning circulation. ΔMaz = MazS - MazN and ΔMov = MovS - MovNare the convergence due to gyre and overturning, respectively. Again, as mentioned before, the latter(ΔMov)is defined as the stability indicator of theAMOC since it denotes a basin-scale salinity advectionfeedback proposed by Stommel(1961). Particularly, ifthe active AMOC leads to a net freshwater divergence(ΔMov < 0), this indicator potentially implies a multiple equilibria behavior of the AMOC. 4. The AMOC stability in the modern climate

To assess the bi-stability of the AMOC of thepresent day, it is essential to examine ΔMov and thefreshwater budget in the real world. A net evaporation(Enet > 0)currently exists in the Atlantic(e. g., Schmitt et al., 1989), which is primarily compensatedby the oceanic freshwater import. At the southernboarder(around 34°S), the gyre circulation induces afreshwater import(MazS > 0)in that the Brazil Current transports saltier water southward in the westernboundary of the gyre, whilst the interior flow, especially the Benguela Current, transports fresher water northward(Fig. 6b). On the other h and, theAMOC exports freshwater southward(MovS < 0)dueto the salinity stratification at nearly 34°S(Rahmstorf, 1996). Particularly, the surface and thermocline water(< 500 m)at 34°S are saltier than NADW(1000-3500 m)underneath, such that the upper/lowerlimb of the AMOC transports saltier/fresher waternorthward/southward to generate a freshwater export(Fig. 6a). Based on available instrumental observ ations, MovShas been estimated ranging from -0. 34 to-0. 1 Sv(Weijer et al., 1999; Huisman et al., 2010; Bryden et al., 2011; Hawkins et al., 2011; Garzoli et al., 2012). For example, Weijer et al. (1999) estimated MovSof -0. 2 Sv using the "best estimate" solutionof an inversion from Holfort(1994). Huisman etal. (2010)suggested that MovS ≈ -0. 1 Sv basedon a dataset from Gouretski and Koltermann(2004). Bryden et al. (2011) estimated MovS ≈ -0. 34 to-0. 1 Sv based on two transatlantic hydrographiccruises along 24°S in 1983 and 2009 and two different methods. Garzoli et al. (2012)reported severalestimations, i. e., a mean value of MovS = -0. 16 Svfrom the expendable bathythermograph(XBT)datacollected along 27 sections at nominally 35°S duringthe period 2002-2011; values of MovSas -0. 15 and -0. 14 Sv for the cruises conducted during 1993 and 2003; and MovSof -0. 11 Sv from the Argo climatological section. Besides, based on oceanic reanalysis data, Hawkins et al. (2011)estimated that MovSis mostlywithin the range of -0. 2 to -0. 1 Sv. In summary, aforesaid observ ational values of MovSare between -0. 34 and -0. 1 Sv, which suggests a bi-stable AMOC in themodern climate if the transport indicator FOT(hereFOT = MovS)is employed.

Fig. 6. (a)Atlantic overturning stream function(contoured in Sv)for the "present-day" equilibrium of theglobal model, superimposed on a plot of Atlantic zonalmean salinity, and (b)zonal section of meridional flow(contoured in cm s-1) and salinity(color)at 30°S in theglobal model. [From Rahmstorf, 1996]

At the northern boundary, freshwater transportsentering the Atlantic are composed of three components, one via the Fram Strait, one via the westernBarents Sea, and the other via the Canadian ArcticArchipelago(refer to Serreze et al., 2006 for more details). For each component, currenTobserv ations arelimited to the total freshwater transport that includesboth AMOC(MovN) and gyre(MazN)contributions. Individual contribution, especially MovN, is unknownfrom observation so far. Herein, Liu et al. (2013)resorted to a relationship as diagnosed from a climatemodel that MovNis about 80% of the total freshwaterimport from the Arctic, such that MovN was estimatedas an imporTof approximately -0. 15 Sv. Therefore, bycombining the observ ational values of MovS and MovN, ΔMov was estimated ranging from -0. 2 to +0. 05 Svover current Atlantic, which indicates an AMOC closeto neutral but with a tendency towards bi-stable under modern climate.

In short, in spite of significant uncertainties, available evidence from both paleo- and modern observ ations suggests that the AMOC in the real world islikely to be bi-stable. This sets a target to test byusing climate models. Nevertheless, climate models, especially the state-of-the-art AOGCMs(see Weber et al., 2007 for the PMIP models; see Weaver et al., 2012 for the CMIP phase 5(CMIP5)models), exhibiTopposite results to observ ations. The AMOC freshwatertransports in the models mostly appear as an import(MovS > 0)across the southern boundary, which thenleads to a freshwater convergence over the Atlantic(ΔMov > 0)(Liu et al., 2014a). As such, it may offer a partial explanation why these AOGCMs tend tosimulate a mono-stable AMOC for the present. Particularly, as compared with the observ ation, AOGCMsgenerally exhibit a systematic bias in the AMOC freshwater transports at the southern boundary(MovS). The exporTof MovSin the observation is biased as animport in AOGCMs(Weber et al., 2007), which thenresults in a systematic distortion of the AMOC stability. Further analyses show that the distortion of freshwater transport in AOGCMs originates mainly fromthe salinity bias in the models. In comparison withthe observ ation, most AOGCMs simulate much freshening surface and thermocline waters and a slightlysaltier NADW around 34°S, which then results in afreshwater import by the AMOC across the southernborder of the Atlantic(Fig. 7). Meanwhile, they alsosimulate a somewhat stronger freshwater import MovNfrom the Arctic. The biased imports, especially fromthe south, by the AMOC lead to a freshwater convergence across the Atlantic basin, which causes a monostable AMOC in these AOGCMs.

Fig. 7. (a)The zonal mean salinity along 34°S across the Atlantic in the IPCC AR4 models and CCSM3 T31. Theobserved salinity for WOA/P datasets(Levitus et al., 1998; Steele et al., 2001)is also shown in black curve(from Liu et al., 2013). (b)The zonal-mean salinity at the southern border of the Atlantic basin as a function of depth for thecontrol state(0 ka)for the PMIP2 simulations. (c)As in(b)but for the PMIP1. 5-type simulations. In(b) and (c), theobserved Levitus salinity profiles are marked by squares. [From Weber et al., 2007]

Further analyses show that this systematic biasin salinity, and in turn, in freshwater transport, ispartly caused by the notorious tropical bias associated with the double ITCZ in AOGCMs(Liu et al., 2014a). In the tropical Atlantic( and eastern Pacific), AOGCMs generally suffer from a common bias in theannual mean climate, which is characterized by a double ITCZ straddling across the equator and an excessive cold tongue pEnetrating westward along the equator(e. g., Mechoso et al., 1995; Davey et al., 2002; Lin, 2007). This bias leads to excessive rainfall and therefore a negative sea surface salinity(SSS)bias in theSouth Atlantic, freshening the inflow across roughly34°S via the upper branch of the AMOC. As a result, the southward freshwater export MovSis reducedgreatly or even reversed to a northward import, whichinduces a freshwater convergence(ΔMov > 0)over theAtlantic in most AOGCMs and implies a mono-stableAMOC.

To eliminate the potential bias in the AMOC stability in AOGCMs, the first remedy is to correct thesurface climate bias. However, the fix of the doubleITCZ problem is beyond the reach of current climatemodel developers. Therefore, as a practical approachso far, Liu et al. (2014a)employed a global flux adjustment method(e. g., Manabe and Stouffer, 1988; Yin and Stouffer, 2007)to reduce the model climatebias of the NCAR CCSM3 T31, despite the fact thatthis method is known to have undesirable side effectson climate models(e. g., Marotzke and Stone, 1995; Neelin and Dijkstra, 1995). In the paper, they arguedthat the flux adjustment method is a useful first step, considering the impossibility of fixing the model biasin current stage and the primary goal of the AMOCstability only.

Figure 8 shows the results from Liu et al. (2014a). Similar to 7 models without flux adjustment in the 4thAssessment ReporTof the Intergovernmental Panel onClimate Change(IPCC AR4), the NCAR CCSM3 T31(Collins et al., 2006; Yeager et al., 2006)exhibits a surface freshening bias at 34°S, which then leads to aweak freshwater exporTof MovS = -0. 013 Sv. Accordingly, a freshwater convergence of ΔMov = 0. 114Sv is generated over the Atlantic basin, implying amono-stable AMOC. This mono-stable AMOC wasthen confirmed explicitly in a freshwater hosing experiment. On the other h and, after adopting a globalflux adjustment, the ITCZ in CCSM3 remains to thenorth of the equator and the upper ocean salinity biasis greatly reduced at 34°S. This results in a strongfreshwater exporTof MovS = -0. 185 Sv and in turn afreshwater divergence of ΔMov = -0. 113 Sv across theAtlantic. Such a negative value of ΔMovindicates abi-stable AMOC and is validated by a subsequent hosing experiment. Nevertheless, it merits attention that, besides the tropical bias related to the double ITCZ, salinity biases in other regions might also contributeto the distortion in the freshwater transports acrossthe Atlantic and thus the AMOC stability. Seen fromFig. 8c, a simply correction of the tropical bias(TRS)by restoring the SSS and SST in the tropical Atlanticbetween 15°S and 15°N can only correct about halfof the distortion in ΔMov, as compared with a parallel globally restoring experiment(GRS). This furtherindicates that the tropical bias related to the doubleITCZ plays a major role but far from the whole storyin distorting the AMOC stability in current AOGCMs.

Fig. 8. Time evolutions of the decadal mean(a, b)AMOCstrength and (c)AMOC freshwater transports. (a)CCSM3T31 CTL run(black) and the hosing experiment CTL-H(gray). (b)CCSM3 T31 CTL run(< year 1000), GRSrun(year 1000-1900), ADJ run(> year 1900)(black)inthe transient period, and the hosing experiment ADJ-H(gray), with the vertical gray dashed lines representing thechange time between CTL to GRS and from GRS to ADJ(AMOC strength is defined as the maximum streamfunction value below 500 m within the Atlantic basin; the 100-yr hosing period is shaded in light gray). (c)Evolution ofthe AMOC freshwater transport at the southern boundary(MovS; blue solid), northern boundary(MovN; red solid), and the divergence indicator(ΔMov; black solid)in thetransition period from CTL(< year 1000)to GRS(year1000{1900), and ADJ(> year 1900). The freshwater transports of the tropical restoring run(TRS; year 1000{1300)are also shown in dashed lines(MovS in blue, MovN in red, and ΔMov in black). Here, CTL denotes the present daycontrol run of the NCAR CCSM3 T31, GRS denotes therun with global restoring SST and SSS, TRS denotes therun with tropical(15°S-15°N)restoring SST and SSS, and ADJ denotes the run with global heat and virtual salt °uxadjustment. [From Liu et al., 2013]
5. The AMOC stability in the future climate

Studies related to the IPCC AR4(e. g., Meehl et al., 2007) and CMIP5 model results(Weaver et al., 2012)show that the AMOC in the historical simulations matches more closely to observ ations than thatin the CMIP3(Cheng et al., 2013). Also, similar tothe CMIP3, all the CMIP5 models predict a weakening of the AMOC by 22%-30% in response to theincrease of atmospheric CO2in the 21st century; thisweakening, however, will develop gradually with little chance of abrupt collapse(also see Delworth et al., 2008). Moreover, the global warming is most unlikelyto result in an AMOC collapse beyond the end of the21st century. Particularly, Weaver et al. (2012)explored the AMOC behavior under anthropogenic radiative forcing, greenhouse gas, and aerosol emissionin various scenarios of the representative concentrationpathways(RCPs; detailed in Moss et al., 2010). Theyadopted FOT as a predictor of the transient, radiatively forced behavior of the AMOC and discoveredthat 40% of the CMIP5 models were in a bi-stableregime of the AMOC during the RCP integrations. Inthe strongest forcing scenario, RCP8. 5, two CMIP5models(the NCAR CCSM4 and Bern3D)eventuallyrealized a slow shutdown of the AMOC. A furtheranalysis(Jahn and Holl and, 2013)showed that theAMOC collapse in the NCAR CCSM4 was caused bysuch a process: an enhanced Arctic sea-ice meltingincreased the liquid freshwater imports from the Arctic, freshened the surface layer in the northern NorthAtlantic, and finally shut down the deep convectionthere, the NADW formation, and the AMOC. It is interesting to find that almost all the CMIP5 AOGCMsab and on the flux adjustment; thus potentially, theyhave a tropical bias due to the double ITCZ, and inturn, a bias in the AMOC stability. Provided thatsuch a bias in the AMOC stability is corrected inthe CMIP5 models, it remains unclear how the modelAMOC will behave in future RCP scenarios. 6. Conclusions and discussion

In this paper, we have reviewed the history ofresearch on the AMOC stability and its significanceto climate change in the past, present, and future. Ofparticular interest is the question if the AMOC willremain stable or change abruptly in the near future. To predict future abrupt changes of the AMOC, it isessential to build a state-of-the-art climate model withcredible AMOC stability. Nevertheless, underst and ing and /or evaluating the AMOC stability in state-of-theart AOGCMs is challenging. Recent results suggestthat state-of-the-art AOGCMs may exhibit a systematic bias in the AMOC stability. This systematic bias, if true, is likely to distort future climate projectionsof abrupt climate change significantly. As a result, some approaches to correct this bias, such as a globalflux adjustment, are needed prior to the conductionof climate simulations and predictions.

It should be pointed out that, in spite of significant progress, many important issues on the stabilityof the AMOC remains open. First, previous studies on the AMOC stability indicator are based on anactive AMOC in equilibrium and may not be applicable for an evolving AMOC(Hawkins et al., 2011). Meanwhile, paleo-data analysis suggests that the pastAMOC has never maintained a perfect equilibrium(e. g., Severinghaus and Brook, 1999; McManus et al., 2004). As such, a generalized stability indicator isproposed(Sijp, 2012; Sijp et al., 2012; Liu et al., 2013). It can be defined as L = ∂ov/∂ov, whereL denotes the indicator, and ovare the AMOCstrength and the AMOC-induced freshwater transportconvergence in the equilibrium state(Liu et al., 2013). In contrast to ΔMov, L is more generalized in showing how the Atlantic freshwater transport modulatesas the AMOC transits from one equilibrium to another. It does not require a divergence-free freshwatertransport in the Atlantic for a collapsed AMOC and therefore manages to correctly monitor the AMOCstability through a slow evolution, such as the lastdeglaciation(Liu et al., 2014b).

Second, the diagnostic indicator, either FOT orΔMov, is based on a hypothesis derived from the boxmodel oF Rahmstorf(1996), i. e., a zero net freshwatertransport(FOT = 0 or ΔMov = 0)is induced by a collapsed AMOC due to the absence of mass transport. This hypothesis is usually taken for granted withoutvalidation(e. g., Hawkins et al., 2011; Weaver et al., 2012; Liu et al., 2013). Recently, Liu and Liu(2014)found that this hypothesis can still be achieved in anAOGCM, but by a compensation of non-zero mass and freshwater transports across MovS and MovN. Therefore, one should be cautious in interpreting theAMOC stability in terms of the freshwater transport.

Third, by definition, ΔMovindicates a basin-scalefreshwater feedback associated with the NADW cell and is valid only when the AMOC collapses to a veryweak NADW cell(e. g., Liu and Liu, 2013). However, some studies(e. g., Gregory et al., 2003; Saenko et al., 2003; Sijp and England, 2006; Sijp et al., 2012)showed that, for a bi-stable AMOC, the collapsed circulation appears as an Antarctic intermediate water(AAIW)reverse cell whose non-linear behaviors suppress the NADW formation and governs the collapsedstate. The AAIW reverse cell has a strong effecTonthe Atlantic freshwater budget, which may render theindicator ΔMovno longer valid.

Finally, one uncertainty on the AMOC stabilityis Agulhas leak age, a transporTof warm and saltyIndian Ocean waters into the Atlantic Ocean(Gordon et al., 1992; De Ruijter et al., 1999; Lutjeharms, 2006). Paleo-proxy records show substantial glacialto-interglacial variations in the amounTof Agulhasleak age(e. g., Biastoch et al., 2008, 2009; Beal et al., 2011; de Deckker et al., 2012), which is accompaniedby modulations of the stratification at 34°S and thusthe AMOC stability. However, many coarse resolution climate models fail to correctly simulate Agulhasleak age due to a poor resolving nonlinear dynamics(inertial mechanisms and ring formation)associatedwith Agulhas leak age. Thus, a correct simulation ofAgulhas leak age using high-resolution climate models(e. g., Biastoch et al., 2008; Tsugawa and Hasumi, 2010)is needed in the future for assessing the stabilityof the AMOC.

References
[1] Arzel, O., M. H. England, and W. P. Sijp, 2008: Reduced stability of the Atlantic meridional overturning circulation due to wind stress feedback during glacial times. J. Climate, 21, 6260-6282.
[2] Bard, E., F. Rostek, J. L. Turon, et al., 2000: Hydrological impact of Heinrich events in the subtropical Northeast Atlantic. Science, 289, 1321-1324.
[3] Beal, L. M., W. P. M. De Ruijter, A. Biastoch, et al., 2011: On the role of the Aguhlas system in ocean circulation and climate. Nature, 472, 429-436.
[4] Berger, A., 1978: Long-term variations of daily insolation and quaternary climatic changes. J. Atmos. Sci.,35, 2362-2367.
[5] Biastoch, A., C. W. Boning, and J. R. E. Lutjeharms, 2008: Agulhas leakage dynamics affects decadal variability in Atlantic overturning circulation. Nature,456, 489-492.
[6] —-, —-, F. U. Schwarzkopf, et al., 2009: Increase in Agulhas leakage due to poleward shift of Southern Hemisphere westerlies. Nature, 462, 495-499.
[7] Bower, A. S., M. S. Lozier, S. F. Gary, et al., 2009: Interior pathways of the North Atlantic meridional overturning circulation. Nature, 459, 243-247.
[8] Broecker, W. S., D. Peteet, and D. Rind, 1985: Does the ocean-atmosphere system have more than one stable mode of operation? Nature, 315, 21-26.
[9] Bryan, F., 1986: High-latitude salinity effects and interhemispheric thermohaline circulations. Nature, 323,301-304.
[10] Bryden, H., B. King, and G. McCarthy, 2011: South Atlantic overturning circulation at 24 S. J. Mar. Res.,69, 39-56.
[11] Clark, P. U., R. Alley, L. D. Keigwin, et al., 1996: Origin of the first global meltwater pulse following the last glacial maximum. Paleoceanography, 11, 563-577.
[12] —-, N. G. Pisias, T. F. Stocher, et al., 2002: The role of the thermohaline circulation in abrupt climate change. Nature, 451, 863-869.
[13] —-, S. Hostetler, N. Pisias, et al., 2007: Mechanisms for a -7-kyr climate and sea-level oscillation during marine isotope stage 3. Ocean Circulation: Mechanisms and Impacts. Schmittner, A., J. Chiang, and S. Hemming, Eds., American Geophysical Union, Geophysical Monograph 173, Washington D. C.,209-246.
[14] Cessi, P., 1994: A simple box model of stochastically forced thermohaline flow. J. Phys. Oceanogr., 24,1911-1920.
[15] Cheng, W., J. C. H. Chiang, and D. Zhang, 2013: Atlantic meridional overturning circulation (AMOC) in CMIP5 Models: RCP and historical simulations. J. Climate, 26, 7187-7197, doi: 10.1175/JCLI-D-12-00496.1.
[16] Collins, W. D., C. M. Bitz, M. L. Blackmon, et al., 2006: The community climate system model version3 (CCSM3). J. Climate, 19, 2122-2143.
[17] Cuffey, K. M., and G. D. Clow, 1997: Temperature, accumulation, and ice sheet elevation in central Greenland through the last deglacial transition. J. Geophys. Res., 102, 26383-26396.
[18] Cunningham, S. A., T. Kanzow, D. Rayner, et al., 2007: Temporal variability of the Atlantic meridional overturning circulation at 26. 5 N. Science, 317, 935-938.
[19] Dansgaard, W., S. J. Johnsen, H. B. Clausen, et al., 1993: Evidence for general instability of past climate from a 250-kyr ice-core record. Nature, 364, 218-220.
[20] Davey, M. K., M. Huddleston, K. Sperber, et al., 2002: STOIC: A study of coupled model climatology and variability in tropical ocean regions. Climate Dyn.,18, 403-420.
[] de Deckker, P., M. Moros, K. Perner, et al., 2012: Influence of the tropics and southern westerlies on glacial interhemispheric asymmetry. Nat. Geosci., 5, 266-269.
[21] De Ruijter, W. P. M., A. Biastoch, S. S. Drijfhout, et al., 1999: Indian-Atlantic inter-ocean exchange: Dynamics, estimation, and impact. J. Geophys. Res.,104, 20885-20910.
[21] de Vries, P., and S. L. Weber, 2005: The Atlantic freshwater budget as a diagnostic for the existence of a stable shut down of the meridional overturning circulation. Geophys. Res. Lett., 32, L09606, doi:10.1029/2004GL021450.
[22] Delworth, T. L., et al., 2008: The potential for abrupt change in the Atlantic meridional overturning circulation in abrupt climate change: Final report, synthesis & assessment product 3. 4, CSSP, Reston, VA, U.S. Geological Survey, 258-359.
[23] Deschamps, P., N. Durand, E. Bard, et al., 2012: Icesheet collapse and sea-level rise at the Bolling warming 14600 years ago. Nature, 483, 559-564.
[24] Dijkstra, H. A., 2007: Characterization of the multiple equilibria regime in a global ocean model. Tellus,59A, 695-705.
[25] Drijfhout, S. S., S. L. Weber, and E. van der Swaluw, 2011: The stability of the MOC as diagnosed from model projections for pre-industrial, present and future climates. Climate Dyn., 37, 1575-1586.
[26] Ganopolski, A., and S. Rahmstorf, 2001: Rapid changes of glacial climate simulated in a coupled climate model. Nature, 409, 153-158.
[27] Garzoli, S., M. Baringer, S. Dong, et al., 2012: South Atlantic meridional fluxes. Deep-Sea Res. I, 71,21-32, doi: 10.1016/j.dsr.2012.09.003.
[28] Gordon, A. L., R. F. Weiss, W. M. Smethie, et al., 1992: Thermocline and intermediate water communication between the South Atlantic and Indian Oceans. J. Geophys. Res., 97, 7223-7240.
[29] Gouretski, V. V., and K. P. Koltermann, 2004: WOCE Global Hydrographic Climatology, Berichte des Bundesambtes für Seeschifffahrt und Hydrographie. Tech. Rep. Alfred-Wegener-Institut für Polarund Meeresforschung, 35, 52 pp.
[30] Gregory, J. M., O. A. Saenko, and A. J. Weaver, 2003: The role of the Atlantic freshwater balance in the hysteresis of the meridional overturning circulation. Climate Dyn., 21, 707-717, doi: 10.1007/s00382-003-0359-8.
[31] Grootes, P. M., M. Stuiver, J. W. C. White, et al., 1993: Comparison of oxygen isotope records from the GISP2 and GRIP Greenland ice cores. Nature,366, 552-554.
[32] Hall, M. M., and H. L. Bryden, 1982: Direct estimates and mechanisms of ocean heat transport. Deep-Sea Res., 29, 339-359.
[33] Hawkins, E., R. S. Smith, L. C. Allison, et al., 2011: Bistability of the Atlantic overturning circulation in a global climate model and links to ocean freshwater transport. Geophys. Res. Lett., 38, L10605, doi:10.1029/2011GL047208.
[34] Hofmann, M., and S. Rahmstorf, 2009: On the stability of the Atlantic meridional overturning circulation. Proc. Natl. Acad. Sci. USA, 106, 20584-20589.
[35] Holfort, J., 1994: Groβraumige Zirkulation und meridionale Transporte im Sudatlantik (Large-scale circulation and meridional transports in the South Atlantic), Ph. D. dissertation, Institut für Meerskunde Kiel, 96 pp.
[36] Holland, M. M., C. M. Bitz, M. Eby, et al., 2001: The role of ice-ocean interactions in the variability of the North Atlantic thermohaline circulation. J. Climate, 14, 656-675.
[37] —-, J. Finnis, A. P. Barrett, et al., 2007: Projected changes in Arctic Ocean freshwater budgets. J. Geophys. Res., 112, G04S55, doi:10.1029/2006JG000354.
[38] Hu, A., B. L. Otto-Bliesner, G. A. Meehl, et al., 2008: Response of thermohaline circulation to freshwater forcing under present day and LGM conditions. J. Climate, 21, 2239-2258.
[39] —-, G. A. Meehl, W. Q. Han, et al., 2012: Role of the Bering Strait on the hysteresis of the ocean conveyor belt circulation and glacial climate stability. Proc. Natl. Acad. Sci. USA, 109, doi:10.1073/pnas.1116014109.
[40] Hughes, T. M., and A. J. Weaver, 1994: Multiple equilibria of an asymmetric two-basin ocean model. J. Phys. Oceanogr., 24, 619-637.
[41] Huisman, S. E., M. den Toom, H. A. Dijkstra, et al., 2010: An indicator of the multiple equilibria regime of the Atlantic meridional overturning circulation. J. Phys. Oceanogr., 40, 551-567.
[42] Jahn, A., and M. M. Holland, 2013: Implications of Arctic sea ice changes for North Atlantic deep convection and the meridional overturning circulation in phys. Res. Lett.,40, 1206-1211.
[43] Johns, W. E., M. O. Baringer, L. M. Beal, et al., 2011: Continuous, array-based estimates of Atlantic Ocean heat transport at 26. 5 N. J. Climate, 24, 2429-2449.
[44] Joos, F., and R. Spahni, 2008: Rates of change in natural and anthropogenic radiative forcing over the past 20000 years. Proc. Natl. Acad. Sci. USA,105, 1425-1430.
[45] Jouzel, J., V. Masson-Delmotte, O. Cattani, et al., 2007: Orbital and millennial Antarctic climate variability over the past 800000 years. Science, 317, 793-796.
[46] Kanzows, T., S. A. Cunningham, D. Rayner, et al., 2007: Observed flow compensation associated with the MOC at 26. 5 N in the Atlantic. Science, 317, 938-941.
[47] Knorr, G., and G. Lohmann, 2003: Southern Ocean origin for the resumption of the Atlantic thermohaline circulation during deglaciation. Nature, 424, 532-536.
[48] Komuro, Y., and H. Hasumi, 2005: Intensification of the Atlantic deep circulation by the Canadian Archipelago throughflow. J. Phys. Oceanogr., 35,775-789.
[49] Lea, D. W., D. K. Pak, L. C. Peterson, et al., 2003: Synchroneity of tropical and high-latitude Atlantic temperatures over the last glacial termination. Science, 301, 1361-1364.
[50] Levitus, S., D. Johnson, J. Antonov, et al., 1998: World Ocean Database 1998. Volume I: Introduction. 346 pp.
[51] Lin, J.-L., 2007: The double-ITCZ problem in IPCC AR4 Coupled GCMs: Ocean-atmosphere feedback analysis. J. Climate, 20, 4497-4525.
[52] Liu, W., and Z. Liu, 2013: A diagnostic indicator of the stability of the Atlantic meridional overturning circulation in CCSM3. J. Climate, 26, 1926-1938.
[53] —-, Z. Y. Liu, and A. X. Hu, 2013: The stability of an evolving Atlantic meridional overturning circulation. Geophys. Res. Lett., 40, 1562-1568.
[54] —-, and Z. Liu, 2014: A note on the stability indicator of Atlantic meridional overturning circulation. J. Climate, 27, 969-975.
[55] —-, Z. Y. Liu, and E. C. Brady, 2014a: Why is the AMOC monostable in coupled general circulation models? J. Climate, 27, 2427-2443, doi:10.1175/JCLI-D-13-00264.1.
[56] —-, Z. Liu, J. Cheng, et al., 2014b: On the stability of the Atlantic meridional overturning circulation during the last deglaciation. Climate Dyn., doi:10.1007/s00382-014-2153-1, in press.
[57] Liu, Z., B. L. Otto-Bliesner, F. He, et al., 2009: Transient simulation of last deglaciation with a new mechanism for Bølling-Allerød warming. Science, 325,310-314.
[58] Lutjeharms, J. R. E., 2006: The Agulhas Current.
[59] MacAyeal, D. R., 1993: Binge/Purge oscillations of the Laurentide Ice Sheet as a cause of the North Atlantic’s Heinrich events. Paleoceanography, 8, 775-784.
[60] Manabe, S., and R. J. Stouffer, 1988: Two stable equilibria of a coupled ocean-atmosphere model. J. Climate, 1, 841-866.
[61] —-, and —-, 1999: Are two modes of thermohaline circulation stable? Tellus A, 51A, 400-411.
[62] Marotzke, J., P. Welander, and J. Willebrand, 1988: Instability and multiple steady states in a meridionalplane model of the thermohaline circulation. Tellus,40A, 162-172.
[63] —-, and J. Willebrand, 1991: Multiple equilibria of the global thermohaline circulation. J. Phys. Oceanogr.,21, 1372-1385.
[64] —-, and P. H. Stone, 1995: Atmospheric transports, the thermohaline circulation, and flux adjustments in a simple coupled model. J. Phys. Oceanogr., 25,1350-1364.
[65] McManus, J., R. Francois, J. Gherardi, et al., 2004: Collapse and rapid resumption of Atlantic meridional circulation linked to deglacial climate changes. Nature, 428, 834-837.
[66] Mechoso, C. R., A. W. Robertson, N. Barth, et al., 1995: The seasonal cycle over the tropical Pacific in coupled ocean-atmosphere general circulation models. Mon. Wea. Rev., 123, 2825-2838.
[67] Meehl, G., T. F. Stocker, W. D. Collins, et al., 2007: Global Climate Projections. Climate Change 2007: The physical Science Basis. Contribution of WGI to the 4th Assessment of the IPCC. Solomon et al., Eds. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA, 996 pp.
[68] Mikolajewicz, U., 1996: A meltwater induced collapse of the thermohaline circulation and its influence on the oceanic distribution of δ14 C and δ18 O. Max-PlanckInstitue fur Meteorologie Rep., 189, 25 pp.
[69] Moss, R. H., J. A. Edmonds, K. A. Hibbard, et al., 2010: The next generation of scenarios for climate change research and assessment. Nature, 463, 747-756.
[70] Mu, M., L. Sun, and H. A. Dijkstra, 2004: The sensitivity and stability of thermohaline circulation of ocean to finite amplitude perturbations. J. Phys. Oceanogr.,34, 2305-2315.
[71] Neelin, J. D., and H. A. Dijkstra, 1995: Oceanatmosphere interaction and the tropical climatology. Part I: The dangers of flux correction. J. Climate,8, 1325-1342.
[72] Nof, D., S. Gorder, and A. de Boer, 2007: Does the Atlantic meridional overturning cell really have more than one stable steady state? Deep-Sea Res. I, 54,2005-2021.
[73] Oka, A., and H. Hasumi, 2006: Effects of model resolution on salt transport through northern highlatitude passages and Atlantic meridional overturning circulation. Ocean Modell., 13, 126-147, doi:10.1016/j.ocemod.2005.12.004.
[74] Peltier, W. R., 2004: Global glacial isostasy and the surface of the ice-age earth: The ICE-5G (VM2) model and GRACE. Annu. Rev. Earth Planet. Sci., 32,111-149.
[75] —-, 2005: On the hemispheric origins of meltwater pulse1a. Quat. Sci. Rev., 24, 1655-1671.
[76] Peterson, L. C., G. H. Haug, K. A. Hughen, et, 2000: Rapid changes in the hydrologic cycle of the tropical Atlantic during the last glacial. Science, 290(5498),1947-1951.
[77] Power, S., and R. Kleeman, 1993: Multiple equilibria in a global ocean general circulation model. J. Phys. Oceanogr., 23, 1670-1681.
[78] Prange, M., G. Lohmann, and A. Paul, 2003: Influence of vertical mixing on the thermohaline hysteresis: Analysis of an OGCM. J. Phys. Oceanogr., 33,1707-1721.
[79] Rahmstorf, S., 1996: On the freshwater forcing and transport of the Atlantic thermohaline circulation. Climate Dyn., 12, 799-811.
[80] —-, 2002: Ocean circulation and climate changes during the past 120,000 years. Nature, 419, 207-214.
[81] —-, M. Crucifix, A. Ganopolski, et al., 2005: Thermohaline circulation hysteresis: A model intercomparison. Geophys. Res. Lett., 32, L23605, doi:10.1029/2005GL023655.
[82] Rennermalm, A. K., E. F. Wood, S. J. Déry, et al., 2006: Sensitivity of the thermohaline circulation to Arctic Ocean runoff. Geophys. Res. Lett., 33, doi:10.1029/2006GL026124.
[83] Renold, M., C. C. Raible, M. Yoshimori, et al., 2010: Simulated resumption of the North Atlantic meridional overturning circulation-slow basin-wide advection and abrupt local convection. Quat. Sci. Rev., 29,101-112, doi: 10.1016/j.quascirev.2009.11.005.
[83] Rooth, C., 1982: Hydrology and ocean circulation. Prog. -149.
[84] Saenko, O. A., A. J. Weaver, and J. M. Gregory, 2003: On the link between the two modes of the ocean thermohaline circulation and the formation of global-scale water masses. J. Climate, 16, 2797-2801.
[85] Sarnthein, M, K. Winn, S. Jung, et al., 1994: Changes in East Atlantic deepwater circulaton over the last30,000 years: Eight time slice reconstructions. Paleoceanography, 9, 209-267.
[86] Schiller, A., U. Mikolajewicz, and R. Voss, 1997: The stability of the North Atlantic thermohaline circulation in a coupled ocean-atmosphere general circulation model. Climate Dyn., 13, 325-347.
[87] Schmitt, R. W., P. S. Bogden, and C. E. Dorman, 1989: Evaporation minus precipitation and density fluxes for the North Atlantic. J. Phys. Oceanogr., 19,1208-1221.
[88] Schmittner, A., 2005: Decline of the marine ecosystem caused by a reduction in the Atlantic overturning circulation. Nature, 434, 628-633.
[89] Schuster, U., and A. J. Watson, 2007: A variable and decreasing sink for atmospheric CO2 in the North Atlantic. J. Geophys. Res., 112, C11006, doi: 10.1029/ 2006JC003941.
[90] Serreze, M. C., A. P. Barrett, A. G. Slater, et al., 2006: The large-scale freshwater cycle of the Arctic. J. Geophys. Res., 111, C11010, doi: 10.1029/2005JC003424.
[91] Severinghaus, J. P., and E. J. Brook, 1999: Abrupt climate change at the end of the last glacial period inferred from trapped air in polar ice. Science, 286,930-934.
[92] Shakun J. D., P. U. Clark, F. He, et al., 2012: Global warming preceded by increasing CO2 during the last deglaciation. Nature, 484, 49-54, doi: 10.1038/nature10915.
[93] Sijp, W. P., 2012: Characterising meridional overturning bistability using a minimal set of state variables. Climate Dyn., 39, 2127-2142.
[94] —-, and M. H. England, 2006: Sensitivity of the Atlantic thermohaline circulation and its stability to basinscale variations in vertical mixing. J. Climate, 19,5467-5478.
[95] —-, M. H. England, and J. M. Gregory, 2012: Precise calculations of the existence of multiple AMOC equilibria in coupled climate models. Part I: Equilibrium States. J. Climate, 25, 282-298.
[96] Srokosz, M., M. Baringer, H. Bryden, et al., 2012: Past, present, and future changes in the Atlantic meridional overturning circulation. Bull. Amer. Meteor. Soc., 93, 1663-1676.
[97] Stanford, J. D., E. J. Rohling, S. E. Hunter, et al., 2006: Timing of meltwater pulse 1a and climate responses to meltwater injections. Paleoceanography,21, PA4103, doi: 10/1029/2006PA001340.
[98] Steele, M., R. Morfley, and W. Ermold, 2001: PHC: A global ocean hydrography with a high quality Arctic Ocean. J. Climate, 14, 2079-2087.
[99] Stocker, T. F., and D. G. Wright, 1991a: Rapid transitions of the ocean’s deep circulation induced by changes in surface water fluxes. Nature, 351, 729-732.
[100] —-, and —-, 1991b: A zonally averaged ocean model for the thermohaline circulation. Part II: Interocean circulation in the Pacific-Atlantic basin system. J. Phys. Oceanogr., 21, 1725-1739.
[101] —-, and S. J. Johnsen, 2003: A minimum thermodynamic model for the bipolar seesaw. Paleoceanography, 18, 1087, doi: 10.1029/2003PA000920, 4.
[102] Stommel, H., 1961: Thermohaline convection with two stable regimes of flow. Tellus, 2, 244-230.
[103] Stouffer, R. J., J. Yin, J. M. Gregory, et al., 2006: Investigating the causes of the response of the thermohaline circulation to past and future climate changes. J. Climate, 19, 1365-1387.
[104] Timmermann, A, H. Gildor, M. Schulz, et al., 2003: Coherent resonant millennial-scale climate oscillations triggered by massive meltwater pulses. J. Climate,16, 2569-2585.
[105] Trenberth, K. E., and J. M. Caron, 2001: Estimates of meridional atmosphere and ocean heat transports. J. Climate, 14, 3433-3443.
[106] Tsugawa, M., and H. Hasumi, 2010: Generation and growth mechanism of a Natal Pulse. J. Phys. Oceanogr., 40, 1597-1612.
[107] Vellinga, M., and R. Wood, 2002: Global climatic impacts of a collapse of the Atlantic thermohaline circulation. Climatic Change, 54, 251-267.
[108] Waelbroeck, C., L. Labeyrie, J.-C. Duplessy, et al., 1998: Improving past sea surface temperature estimates based on planktonic fossil faunas. Paleoceanography,13, 272-283.
[109] Weaver, A. J., J. Marotzke, P. F. Cummins, et al., 1993: Stability and variability of the thermohaline circulation. J. Phys. Oceanogr., 23, 39-60.
[110] —-, O. A. Saenko, P. U. Clark, et al., 2003: Meltwater pulse 1A from Antarctica as a trigger of the BøllingAllerød warm interval. Science, 299, 1709-1713.
[111] —-, J. Sedláček, M. Eby, et al., 2012: Stability of the Atlantic meridional overturning circulation: A model intercomparison. Geophys. Res. Lett., 39, L20709, doi: 10.1029/2012GL053763.
[112] Weber, S. L., S. S. Drijfhout, A. Abe-Ouchi, et al., 2007: The modern and glacial overturning circulation in the Atlantic Ocean in PMIP coupled model simulations. Clim. Past, 3, 51-64.
[113] Weijer, W., W. P. M. De Ruijter, H. A. Dijkstra, et al., 1999: Impact of interbasin exchange on the Atlantic overturning circulation. J. Phys. Oceanogr., 29,2266-2284.
[114] Welander, P., 1986: Thermohaline effects in the ocean circulation and related simple models. Large-Scale Transport Processes in the Oceans and Atmosphere. Willebrand, J., and D. L. T. Anderson Eds., D. Reidel Publishing Company, Dordrecht, The Netherlands, 163-200.
[115] Yeager, S. G., C. A. Shields, W. G. Large, et al., 2006: The low-resolution CCSM3. J. Climate, 19, 2545-2566.
[116] Yin, J. J., M. E. Schlesinger, N. G. Andronova, et al., 2006: Is a shutdown of the thermohaline circulation irreversible? J. Geophys. Res., 111, D12104, doi:10.1029/2005JD006562.
[117] —-, and R. J. Stouffer, 2007: Comparison of the stability of the Atlantic thermohaline circulation in two coupled atmosphere-ocean general circulation models. J. Climate, 20, 4293-4315.