Editing IPCC:AR6/WGI/Chapter-9 (section)

==== 9.2.3.1 Atlantic Meridional Overturning Circulation ====

<div id="h3-7-siblings" class="h3-siblings"></div>

Atlantic Meridional Overturning Circulation (AMOC) is the main overturning current system in the South and North Atlantic oceans. It transports warm upper-ocean water northwards, and cold, deep water southwards, as part of the global ocean circulation system ( [[IPCC:Wg1:Chapter:Chapter-2#2.3.3.4.1|Section 2.3.3.4.1]] ). Changes in AMOC influence global ocean heat content (OHC) and transport ( [[#9.2.2.1|Section 9.2.2.1]] ); global ocean anthropogenic carbon uptake changes and climate sensitivity (Cross-Chapter Box 5.3); and dynamical sea level change ( [[#9.2.4|Section 9.2.4]] ). Since AR5/SROCC, confidence in modelled and reconstructed AMOC has decreased due to new observations and model disagreement. Confidence levels have been revisited in modelled AMOC evolution during the 20th century, the magnitude of 21st-century AMOC decline, and the possibility of an abrupt collapse before 2100.

The AR5 ( [[#Flato--2013|Flato et al., 2013]] ) found that the mean AMOC strength in CMIP5 models ranges from 15 to 30 Sv for the historical period. The multi-model mean overturning at 26°N in CMIP5 and CMIP6 is comparable to the RAPID array measurements ( [[#Reintges--2017|Reintges et al., 2017]] ), but the inter-model spread in CMIP6 is as large (10–31 Sv) as in CMIP5 ( [[IPCC:Wg1:Chapter:Chapter-3#3.5.4|Section 3.5.4]] ; [[#Weijer--2020|Weijer et al., 2020]] ). Biases in simulations of the present-day AMOC and associated deep convection in the subpolar gyre and Nordic Seas were large in CMIP5 models, with many models exhibiting ocean convection that is too deep, over too large an area, too far south, and occurring too frequently ( [[#9.2.1.3|Section 9.2.1.3]] and Figure 9.5; [[#Heuzé--2017|Heuzé, 2017]] ) related to biases in sea ice extent, overflows, and freshwater forcing ( [[#Deshayes--2014|Deshayes et al., 2014]] ; H. [[#Wang--2015|]] [[#Wang--2015|Wang et al., 2015]] ). As a result, the AMOC in CMIP5 was nearly always too shallow, with too weak a temperature contrast between the northward and southward flowing branches. Deep convection errors are still large in CMIP6, and the shallow bias in AMOC persists ( [[#Weijer--2020|Weijer et al., 2020]] ; [[#Heuzé--2021|Heuzé, 2021]] ). Since AR5, there is emerging evidence that enhancing horizontal resolution can reduce long-standing climate model biases in AMOC strength, where the magnitude and profile of northward heat transport at 26°N become more comparable to observations ( [[#Chassignet--2020|Chassignet et al., 2020]] ; [[#Roberts--2020|Roberts et al., 2020]] ). The sensitivity of the AMOC to ocean resolution, however, is model-dependent and can be positive as well as negative ( [[#Roberts--2020|Roberts et al., 2020]] ). An increase in AMOC strength at 26°N, with higher resolution in the ocean component, has been associated with too strong (deep) convection in the subpolar gyre and too deep winter mixed layers ( [[#Jackson--2020|]] [[#Jackson--2020|L.C. Jackson et al., 2020]] ), which occurs in most CMIP6 models that are unable to overflow deep water formed in the Nordic Seas across the Greenland–Iceland–Scotland Ridge. Models with a correct AMOC strength may do so by compensating a lack of deep-water outflow from the Nordic Seas through too much deep convection and deep-water formation in the Labrador and Irminger Seas ( [[#Heuzé--2021|Heuzé, 2021]] ).

Models and paleoreconstructions have often assumed a close relation between the AMOC and deep convection in the Labrador Sea; the Labrador Sea convection variability has been interpreted as connecting to AMOC variability. Observational studies have been inconclusive on whether this relation exists ( [[#Buckley--2016|Buckley and Marshall, 2016]] ). New insight from observed overturning in the eastern and western subpolar gyre in the North Atlantic in OSNAP ( [[#Lozier--2019|Lozier et al., 2019]] ; [[#Petit--2020|Petit et al., 2020]] ) reveals that 15.6 ± 3.1Sv takes place north of the OSNAP array between Greenland and Scotland, with only 2.1 ± 0.9 Sv of overturning occurring across the Labrador Sea, as found with the OSNAP 53°N array spanning the mouth, calling into question the validity of the Labrador Sea convection–AMOC link ( [[#Lozier--2019|Lozier et al., 2019]] ). Although these results are derived from only the first 21 months of data from monitoring since 2014, hydrographic observations during 1990–1997 previously found small overturning (1–2 Sv) in the Labrador Sea ( [[#Pickart--2007|Pickart and Spall, 2007]] ). However, previous estimates of Labrador Sea Water formation (obtained with different techniques) suggest larger overturning ( [[#Haine--2008|Haine et al., 2008]] ). Part of this controversy could be explained if a large fraction of newly formed Labrador Sea Water is not exported from the Labrador Sea. The OSNAP observations are supported by previous hydrographic measurements in showing strong east–west symmetry in isopycnal slope in the Labrador Sea in periods of both strong and weak convection; this implies compensating northward and southward transport above and below the potential density surface that separates the upper and lower overturning limbs ( [[#Lozier--2019|Lozier et al., 2019]] ), despite large deep convection variability ( [[#Yashayaev--2007|Yashayaev, 2007]] ; [[#Yashayaev--2016|Yashayaev and Loder, 2016]] ). New observations of deep winter mixing in the Irminger Basin ( [[#de%20Jong--2018|de Jong et al., 2018]] ; [[#Josey--2019|Josey et al., 2019]] ) support the assertion that the Irminger Sea, in addition to the Nordic Seas ( [[#Chafik--2019|Chafik and Rossby, 2019]] ), are the main sources of overturning in the eastern subpolar gyre, consistent with OSNAP ( [[#Petit--2020|Petit et al., 2020]] ). It is unclear to what extent models are in disagreement with this view of overturning in the subpolar gyre, as a direct comparison with OSNAP of model analyses partitioning the overturning into a western and eastern part is mostly lacking, with a notable exception ( [[#Menary--2020a|Menary et al., 2020a]] ). Other results give rise to considerable uncertainty over veracity of the models in simulating the overturning partitioning between east and west and the role of various drivers of AMOC variability, including: the analysis of water mass formation in CMIP6 models ( [[#Heuzé--2021|Heuzé, 2021]] ); the analysis between Labrador Sea Water formation and AMOC in a suite of ocean-only models ( [[#Danabasoglu--2014|Danabasoglu et al., 2014]] ); and the fact that when the OSNAP observing system design was tested in an eddy-permitting ocean model comparable amounts of overturning in the western and eastern subpolar gyre were found ( [[#Susan%20Lozier--2017|Susan Lozier et al., 2017]] ). Disagreement between models and OSNAP observations may decrease in higher-resolution models ( [[#Menary--2020a|Menary et al., 2020a]] ). In summary, multiple lines of evidence provide ''medium agreement'' between models and observations on drivers of change and variability in the AMOC and, in particular, the role of Labrador Sea deep convection in constituting AMOC variability.

The AMOC is a potential driver of Atlantic Multi-decadal Variability (AMV), but there is new evidence that anthropogenic aerosol changes have contributed to observed AMV changes, and that underestimation of the magnitude and duration of AMV changes in CMIP5 is tempered in CMIP6 ( [[IPCC:Wg1:Chapter:Chapter-3#3.7.7|Section 3.7.7]] and Annex IV.2.7). Comparison of observed AMOC variability at the RAPID section with modelled variability reveals that CMIP5 models appear to largely underestimate the interannual and decadal time scale variability ( [[#Roberts--2014|Roberts et al., 2014]] ; [[#Yan--2018|Yan et al., 2018]] ), and similar results are found when analysing CMIP6 models ( [[IPCC:Wg1:Chapter:Chapter-3#3.5.4.1|Section 3.5.4.1]] ). By underestimating the multi-decadal AMOC–AMV link and other low-frequency AMOC variability, climate models also underestimate internal variability in subpolar SSTs that feed back on the North Atlantic Oscillation (NAO). This causes the NAO to lack variability on multi-decadal time scales ( [[#Kim--2018|Kim et al., 2018]] ). Despite the role of the AMOC in generating AMV through subsurface temperatures in antiphase with SST and downward heat fluxes into the ocean that anticorrelate with SSTs (R. [[#Zhang--2019|]] [[#Zhang--2019|]] [[#Zhang--2019|Zhang et al., 2019]] ), it is generally accepted that AMOC forcing of SST variability exists alongside stochastic wind forcing and external forcing by aerosols ( [[#Bellomo--2018|Bellomo et al., 2018]] ; [[#Haustein--2019|Haustein et al., 2019]] ; [[#O’Reilly--2019|O’Reilly et al., 2019]] ; [[#Wills--2019|Wills et al., 2019]] ).

The SROCC ( [[#Collins--2019|Collins et al., 2019]] ) assessed that in situ observations (2004–2017) and sea surface temperature reconstructions indicate that AMOC has weakened relative to 1850–1900 ( ''medium confidence'' ). However, SROCC also assessed that there is insufficient data to quantify the magnitude of the weakening, or to properly attribute it to anthropogenic forcing, due to the limited length of the observational record. Here, this assessment is adjusted to ''low confidence'' in the weakening (as also discussed in Sections 2.3.3.4.1 and 3.5.4.1). The CMIP5 multi-model mean showed no 20th century trend in AMOC ( [[#Cheng--2013|Cheng et al., 2013]] ). The CMIP6 multi-model mean slightly opposes the reconstructed decline due to a strong increase in the 1940–1985 period ( [[#Menary--2020b|Menary et al., 2020b]] ; [[#Weijer--2020|Weijer et al., 2020]] ), thought to be in response to aerosol forcing ( [[IPCC:Wg1:Chapter:Chapter-3#3.5.4.1|Section 3.5.4.1]] ), followed by a smaller decline since the 1990s. Also, agreement between different proxy-based reconstructions is weak in many details ( [[#Moffa-Sánchez--2019|Moffa-Sánchez et al., 2019]] ) and questions can be raised regarding various proxies used in reconstructions ( [[IPCC:Wg1:Chapter:Chapter-2#2.3.3.4.1|Section 2.3.3.4.1]] ). For instance, SST-based proxies can be influenced by atmospheric and other processes acting on different time scales ( [[#Moffa-Sánchez--2019|Moffa-Sánchez et al., 2019]] ; [[#Jackson--2020|Jackson and Wood, 2020]] ). In addition, many proxies are indirect and based on AMOC-related processes assumed to be similar to those found in models, such as the link between AMOC and Labrador Sea convection, which has been questioned recently (see above). In addition, the subpolar gyre from which many AMOC proxies are taken may vary independently of AMOC, with similar patterns in SST and OHC driven by wind variability ( [[#Williams--2014|Williams et al., 2014]] ; [[#Piecuch--2017|Piecuch et al., 2017]] ). Finally, a new dynamic reconstruction of the Atlantic inflow to the Nordic Seas suggests no slowdown over the past 70 to 100 years ( [[#Rossby--2020|Rossby et al., 2020]] ), in contrast to a new compilation of proxy reconstructions which suggests that AMOC is presently in its weakest state in the last millennium ( [[#Caesar--2021|Caesar et al., 2021]] ), reinforcing the evidence that motivated the previous SROCC assessment. [[IPCC:Wg1:Chapter:Chapter-3#3.5.4.1|Section 3.5.4.1]] also questions the veracity of the models’ forced AMOC response during the 20th century. Given the large discrepancy between modelled and reconstructed AMOC in the 20th century, and the uncertainty over the realism of the 20th century modelled AMOC response ( [[IPCC:Wg1:Chapter:Chapter-3#3.5.4.1|Section 3.5.4.1]] ), we have ''low confidence'' in both.

The strength of AMOC has been measured directly since 2004 using the RAPID Array ( [[IPCC:Wg1:Chapter:Chapter-2#2.3.3.4.1|Section 2.3.3.4.1]] ; [[#Smeed--2018|Smeed et al., 2018]] ). RAPID-based estimates show a large amount of variability compared to CMIP models ( [[#Roberts--2014|Roberts et al., 2014]] ). Observed changes since 2004 are too short for the evaluation of a long-term trend given the decadal scale internal variability ( [[IPCC:Wg1:Chapter:Chapter-2#2.3.3.4.1|Section 2.3.3.4.1]] ). Nevertheless, [[#Smeed--2018|Smeed et al. (2018)]] argue that, between 2007 and 2011, AMOC shifted to a state of reduced overturning – decreasing from 18.8 Sv between 2004 and 2008 to 16.1 Sv after 2008. A shift in AMOC strength of this magnitude is not captured by CMIP5 and CMIP6 models, which generally underestimate interannual to decadal AMOC variability ( [[IPCC:Wg1:Chapter:Chapter-3#3.5.4.1|Section 3.5.4.1]] ). Additional evidence since SROCC also raises the inconsistency between the RAPID weakening in the 3000–5000 m depth range and the relative constancy of deep overflows from the Arctic ( [[#Østerhus--2019|Østerhus et al., 2019]] ), implying that the recent decrease in AMOC at 26.5°N ( [[#Smeed--2018|Smeed et al., 2018]] ) is not caused by overflow weakening or reduced overturning in the Nordic Seas, although the weakening occurred almost exclusively in the 3000–5000 m depth range associated with a reduction of Lower NADW ( [[#9.2.2.3|Section 9.2.2.3]] ). It is unclear what causes a weakening of the deepest limb of AMOC at 26.5°N, if the main sources for this flow farther north remain constant. Various estimates of AMOC and associated heat transport suggest an increase since the 1940s with a subsequent decrease since the 1990s ( [[IPCC:Wg1:Chapter:Chapter-2#2.3.3.4.1|Section 2.3.3.4.1]] ), supported by ocean reanalysis ( [[#Jackson--2019|Jackson et al., 2019]] ), forced ocean model simulations ( [[#Robson--2012|Robson et al., 2012]] ; [[#Danabasoglu--2016|Danabasoglu et al., 2016]] ) and CMIP6 simulations ( [[#Menary--2020a|Menary et al., 2020a]] ). This suggests that the observed AMOC-shift between 2007 and 2011 may be part of a longer-term decrease ( ''medium confidence'' ), which has been attributed to be part of multiannual variability ( [[#Rhein--2019|Rhein et al., 2019]] ).

The SROCC ( [[#Collins--2019|Collins et al., 2019]] ) found that AMOC will ''very likely'' weaken over the 21st century. In CMIP6 projections, the modelled decline starting in the 1990s continues in all future projections, almost independent of the forcing scenario until about 2060, after which low-emissions scenarios show stabilization, while high-emissions scenarios continue to exhibit AMOC decline (Figure 9.10; [[#Menary--2020b|Menary et al., 2020b]] ; [[#Weijer--2020|Weijer et al., 2020]] ). Despite differences in overall AMOC strength, location and latitude of deep convection, sea ice and SST bias and representation of deep overflows, the model projections are qualitatively similar. This agreement suggests that AMOC decline may be governed by large-scale constraints independent of the details of the models. In theoretical models of the thermohaline circulation, the circulation strength is proportional to a density or pressure difference between the subpolar North Atlantic and subtropical South Atlantic ( [[#Kuhlbrodt--2007|Kuhlbrodt et al., 2007]] ; [[#Weijer--2019|Weijer et al., 2019]] ). In all models, the north-south pressure gradient decreases in the 21st century, as subpolar waters warm faster than subtropical waters, and an enhanced hydrological cycle drives freshening at subpolar latitudes, while subtropical latitudes feature more evaporation and salinification ( [[#9.2.1|Section 9.2.1]] ). As a result, surface waters at subpolar latitudes become more buoyant and more stable, so that deep water formation driving the AMOC declines ( [[#9.2.1.3|Section 9.2.1.3]] ). Projected AMOC decline by 2100 ranges from 24 [4 to 46] % in SSP1-2.6 to 39 [17–55] % in SSP5-8.5 ( ''medium confidence'' ) ( [[IPCC:Wg1:Chapter:Chapter-4#4.3.2.3|Section 4.3.2.3]] ). Note that these ranges are based on ensemble means of individual models, largely smoothing out internal variability. If single realizations are considered, the ranges become wider, especially by lowering the low end of the range ( [[IPCC:Wg1:Chapter:Chapter-4#4.3.2.3|Section 4.3.2.3]] ). In summary, it is ''very likely'' that AMOC will decline in the 21st century, but there is ''low confidence'' in the model’s projected timing and magnitude. In addition, freshwater from the melting of the Greenland Ice Sheet (Sections 9.4.1.3 and 9.4.1.4) could further enhance the future weakening of AMOC in the 21st century ( [[#Collins--2019|Collins et al., 2019]] ; [[#Golledge--2019|Golledge et al., 2019]] ).

<div id="_idContainer028" class="Basic-Text-Frame _idGenObjectStyleOverride-1"></div>

[[File:59bfd54f25dac32fb13e065da36f3646 IPCC_AR6_WGI_Figure_9_10.png]]

'''Figure 9.10''' '''|''' '''Atlantic Meridional Overturning Circulation (AMOC) strength in simulations and sensitivity to resolution and forcing.''' '''(Top left)''' AMOC magnitude (units: Sverdrup (Sv) = 10 <sup>9</sup> kg <sup></sup> s <sup>–1</sup> ) in Paleoclimate Modelling Intercomparison Project (PMIP) experiments. '''(Top right)''' Time series of AMOC from Coupled Model Intercomparison Project Phase 5 and 6 (CMIP5 and CMIP6) based on ( [[#Menary--2020b|Menary et al., 2020b]] ). '''(Bottom left)''' Percent change in AMOC strength per year at different resolutions over the 1950–2050 period with colours for model families ( [[#Roberts--2020|Roberts et al., 2020]] ). '''(Bottom right)''' A compilation of percentage changes in the simulated AMOC after applying an additional freshwater flux in the subpolar North Atlantic at the surface for a limited time ( [[#de%20Vries--2005|de Vries and Weber, 2005]] ; [[#Stouffer--2006|Stouffer et al., 2006]] ; [[#Yin--2007|Yin and Stouffer, 2007]] ; [[#Jackson--2013|Jackson, 2013]] ; [[#Liu--2013|Liu and Liu, 2013]] ; [[#Jackson--2018|Jackson and Wood, 2018]] ; [[#Haskins--2019|Haskins et al., 2019]] ). Symbols indicate whether the AMOC recovers within 200 years (circles), is starting to recover (upwards arrow), or does not recover within 200 years (downwards arrow). Symbol size indicates rate of freshwater input. Further details on data sources and processing are available in the chapter data table (Table 9.SM.9).

Both AR5 ( [[#Collins--2013|Collins et al., 2013]] ) and SROCC ( [[#Collins--2019|Collins et al., 2019]] ) assessed that an abrupt collapse of AMOC before 2100 was ''very unlikely'' , but SROCC added that, by 2300, an AMOC collapse was ''as likely asnot'' for high-emissions scenarios. The SROCC also assessed that model bias may considerably affect the sensitivity of the modelled AMOC to freshwater forcing. Tuning towards stability and model biases ( [[#Valdes--2011|Valdes, 2011]] ; [[#Liu--2017|Liu et al., 2017]] ; [[#Mecking--2017|Mecking et al., 2017]] ; [[#Weijer--2019|Weijer et al., 2019]] ) provides CMIP models a tendency toward unrealistic stability ( ''medium confidence'' ). By correcting for existing salinity biases, [[#Liu--2017|Liu et al. (2017)]] demonstrated that AMOC behaviour may change dramatically on centennial to millennial time scales, and that the probability of a collapsed state increases. None of the CMIP6 models features an abrupt AMOC collapse in the 21st century, but they neglect meltwater release from the Greenland Ice Sheet. Also, a recent process study reveals that a collapse of AMOC can be induced, even by small-amplitude changes in freshwater forcing ( [[#Lohmann--2021|Lohmann and Ditlevsen, 2021]] ). As a result, we change the assessment of an abrupt collapse before 2100 to ''medium confidence'' that it will not occur ''.''

<div id="9.2.3.2" class="h3-container"></div>

<span id="southern-ocean"></span>