Anthropogenic aerosol forcing of the Atlantic meridional overturning circulation and the associated mechanisms in CMIP6 models

Uncertainty in the strength of the Atlantic Meridional Overturning Circulation (AMOC) is analyzed in the Coupled Model Intercomparison Project Phase 3 (CMIP3) and Phase 5 (CMIP5) projections for the twenty-first century; and the different sources of uncertainty (scenario, internal and model) are quantified. Although the uncertainty in future projections of the AMOC index at 30°N is larger in CMIP5 than in CMIP3, the signal-to-noise ratio is comparable during the second half of the century and even larger in CMIP5 during the first half. This is due to a stronger AMOC reduction in CMIP5. At lead times longer than a few decades, model uncertainty dominates uncertainty in future projections of AMOC strength in both the CMIP3 and CMIP5 model ensembles. Internal variability significantly contributes only during the first few decades, while scenario uncertainty is relatively small at all lead times. Model uncertainty in future changes in AMOC strength arises mostly from uncertainty in density, as uncertainty arising from wind stress (Ekman transport) is negligible. Finally, the uncertainty in changes in the density originates mostly from the simulation of salinity, rather than temperature. High-latitude freshwater flux and the subpolar gyre projections were also analyzed, because these quantities are thought to play an important role for the future AMOC changes. The freshwater input in high latitudes is projected to increase and the subpolar gyre is projected to weaken. Both the freshening and the gyre weakening likely influence the AMOC by causing anomalous salinity advection into the regions of deep water formation. While the high model uncertainty in both parameters may explain the uncertainty in the AMOC projection, deeper insight into the mechanisms for AMOC is required to reach a more quantitative conclusion.

There is a controversy about the origin of the recent decadal Atlantic Meridional Overturning Circulation (AMOC) slowing observed at 26.5°N and concurrent sea surface temperature cooling in the central and eastern mid-latitude North Atlantic. We investigate decadal AMOC slowing events simulated in a multi-millennial preindustrial control integration of the Kiel Climate Model (KCM), providing an estimate of internal AMOC variability. Preindustrial control integrations of 15 models participating in the Coupled Model Intercomparison Project phase 5 also are investigated, as well as historical simulations with them providing estimates of AMOC variability during 1856–2005. It is shown that the recent decadal AMOC decline is still within the range of the models’ internal AMOC variability and thus could be of natural origin. In this case, the decline would represent an extreme realization of internal variability provided the climate models yield realistic levels of AMOC variability. The model results suggest that internal decadal AMOC variability is large, requiring multi-decadal observational records to detect an anthropogenic AMOC signal with high confidence. When analyzing the strongest decadal AMOC slowing events in the KCM, which have amplitudes similar to or larger than the recently observed decadal AMOC decline, the following composite picture emerges: a very strong decadal AMOC decline is preceded by a decadal rise in atmospheric surface pressure over large parts of the mid-latitude North Atlantic. The change in low-level atmospheric circulation drives reduced oceanic heat loss over and diminished upper-ocean salt content in the Labrador Sea. In response, oceanic deep convection and subsequently the AMOC and northward oceanic heat transport weaken, and anomalously cold sea surface temperatures develop in the central and eastern mid-latitude North Atlantic.

This study investigates how variations of the Atlantic meridional overturning circulation (AMOC) affect sea surface temperature (SST) within the simulations of the coupled model intercomparison project phase 5. In particular, we explore whether the SST response is interhemispheric in nature, specifically as reflected in the Atlantic SST Dipole index, or whether the response is localized more in the North Atlantic Ocean. In the absence of direct observational data, this Dipole index has been proposed to approximate AMOC variations over the duration of the instrumental temperature record. We find that typically, on timescales between decadal and centennial, the SST Dipole index correlates with the AMOC with coefficients ranging from 0.2 to 0.7, typically with a 0–6 year lag, and thus explains less than half of the AMOC variance. In just two models this value slightly exceeds 50 %. Even for the models with the highest correspondence between the AMOC and the Dipole index, the correlation between the two variables is controlled mainly by SST variations in the North Atlantic, not the South Atlantic, both for the model control and historical simulations. Consequently, in nearly all models, the North Atlantic SST provides a better indicator of AMOC variations than the Atlantic Dipole. Thus, on decadal to centennial timescales AMOC variability affects mainly the North Atlantic Ocean, with the sensitivity of the North Atlantic SST between 40 and 60°N, given by the multi-model average, of about 0.3 °C per 1 Sv of AMOC change, explaining roughly one third of the SST variance.

<p>The Arctic and North Atlantic Ocean play a fundamental role in Earth’s water cycle, distribution of energy (i.e. heat), and the formation of cold, dense waters. Through the Atlantic meridional overturning circulation (AMOC), heat is transported to the high-latitudes. Classically, the climate impact of AMOC variations has been investigated through hosing experiments, where anomalous freshwater is artificially added or removed from the North Atlantic to modulate deep water formation. However, such a protocol introduces artificial changes in the subpolar area, possibly masking the effect of the AMOC modulation. Here, we develope a protocol where AMOC intensity is modulated remotely through the teleconnection of the tropical Indian Ocean (TIO), so as to investigate more robustly the impact of the AMOC on climate. Warming in the TIO has recently been shown to strengthen the Walker circulation in the Atlantic through the propagation of Kelvin and Rossby waves, increasing and stabilizing the AMOC on longer timescales. Using the latest coupled-model from Insitut Pierre Simon Laplace (IPSL-CM6), we have designed a three-member ensemble experiment nudging the surface temperatures of the TIO by -2°C, +1°C, and +2°C for 100 years. The objectives are to better quantify the timescales of AMOC variability outside the use of hosing experiments and the TIO-AMOC relationship.  In each ensemble member, there are two distinct features compared to the control run. The initial changes in AMOC (≤20 years) are largely atmospherically driven, while on longer timescales is largely driven by the TIO teleconnection to the tropical Atlantic. In the northern North Atlantic, changes in sensible heat fluxes range from 15 to 20 W m<sup>-2 </sup>in all three members compared to the control run, larger than the natural variability. On the longer timescales, AMOC variability is strongly influenced from anomalies in the tropical Atlantic Ocean. The TIO teleconnection supports decreased precipitation in the tropical Atlantic Ocean during warming (opposite during TIO cooling) events, as well as positive salinity anomalies and negative temperature anomalies. Using lagged correlations, there are the strongest correlations on scales within one year and a delayed response of 30 years (in the -2°C ensembles). In comparing the last 20 years, nudging the TIO induces a 3.3 Sv response per 1°C change. In summary, we have designed an experiment to investigate the AMOC variability without directly changing the North Atlantic through hosing, making way for a more unbiased approach to analysing the AMOC variability in climate models.</p>

The dynamics of the Atlantic meridional overturning circulation (AMOC) vary considerably among different climate models; for example, some models show clear peaks in their power spectra while others do not. To elucidate these model differences, transfer functions are used to estimate the frequency domain relationship between surface forcing fields, including sea surface temperature, salinity, and wind stress, and the resulting AMOC response. These are estimated from the outputs of the Coupled Model Intercomparison Project phase 5 (CMIP5) and phase 3 (CMIP3) control runs for eight different models, with a specific focus on Geophysical Fluid Dynamics Laboratory Climate Model, version 2.1 (GFDL CM2.1), and the Community Climate System Model, version 4 (CCSM4), which exhibit rather different spectral behavior. The transfer functions show very little agreement among models for any of the pairs of variables considered, suggesting the existence of systematic model errors and that considerable uncertainty in the simulation of AMOC in current climate models remains. However, a robust feature of the frequency domain analysis is that models with spectral peaks in their AMOC correspond to those in which AMOC variability is more strongly excited by high-latitude surface perturbations that have periods corresponding to the frequency of the spectral peaks. This explains why different models exhibit such different AMOC variability. These differences would not be evident without using a method that explicitly computes the frequency dependence rather than a priori assuming a particular functional form. Finally, transfer functions are used to evaluate two proposed physical mechanisms for model differences in AMOC variability: differences in Labrador Sea stratification and excitation by westward-propagating subsurface Rossby waves.

Atlantic Meridional Overturning Circulation (AMOC) projections are uncertain due to both model errors, as well as internal climate variability. An AMOC slowdown projected by many climate models is likely to have considerable effects on many aspects of global and North Atlantic climate. Previous studies to make probabilistic AMOC projections have broken new ground. However, they do not drift-correct or cross-validate the projections, and do not fully account for internal variability. Furthermore, they consider a limited subset of models, and ignore the skill of models at representing the temporal North Atlantic dynamics. We improve on previous work by applying Bayesian Model Averaging to weight 13 Coupled Model Intercomparison Project phase 5 models by their skill at modeling the AMOC strength, and its temporal dynamics, as approximated by the northern North-Atlantic temperature-based AMOC Index. We make drift-corrected projections accounting for structural model errors, and for the internal variability. Cross-validation experiments give approximately correct empirical coverage probabilities, which validates our method. Our results present more evidence that AMOC likely already started slowing down. While weighting considerably moderates and sharpens our projections, our results are at low end of previously published estimates. We project mean AMOC changes between periods 1960–1999 and 2060–2099 of −4.0 Sv and −6.8 Sv for RCP4.5 and RCP8.5 emissions scenarios respectively. The corresponding average 90% credible intervals for our weighted experiments are [−7.2, −1.2] and [−10.5, −3.7] Sv respectively for the two scenarios.

North Atlantic simulations in Coordinated Ocean-ice Reference Experiments phase II (CORE-II). Part II: Inter-annual to decadal variability

Ocean heat transport (OHT) plays a key role in climate and its variability. Here, we identify modes of low-frequency North Atlantic OHT variability by applying a low-frequency component analysis (LFCA) to output from three global climate models. The first low-frequency component (LFC), computed using this method, is an index of OHT variability that maximizes the ratio of low-frequency variance (occurring at decadal and longer timescales) to total variance. Lead-lag regressions of atmospheric and ocean variables onto the LFC timeseries illuminate the dominant mechanisms controlling low-frequency OHT variability. Anomalous northwesterly winds from eastern North America over the North Atlantic act to increase upper ocean density in the Labrador Sea region, enhancing deep convection, which later increases OHT via changes in the strength of the Atlantic Meridional Overturning Circulation (AMOC). The strengthened AMOC carries warm, salty water into the subpolar gyre, reducing deep convection and weakening AMOC and OHT. This mechanism, where changes in AMOC and OHT are driven primarily by changes in Labrador Sea deep convection, holds not only in models where the climatological (i.e., time-mean) deep convection is concentrated in the Labrador Sea, but also in models where the climatological deep convection is concentrated in the Greenland-Iceland-Norwegian (GIN) Seas or the Irminger and Iceland Basins. These results suggest that despite recent observational evidence suggesting that the Labrador Sea plays a minor role in driving the climatological AMOC, the Labrador Sea may still play an important role in driving low-frequency variability in AMOC and OHT.

The Atlantic Meridional Overturning Circulation (AMOC) is crucial for global ocean carbon and heat uptake, and controls the climate around the North Atlantic. Despite its importance, quantifying the AMOC’s past changes and assessing its vulnerability to climate change remains highly uncertain. Understanding past AMOC changes has relied on proxies, most notably sea surface temperature anomalies over the subpolar North Atlantic. Here, we use 24 Earth System Models from Coupled Model Intercomparison Project Phase 6 (CMIP6) to demonstrate that these sea surface temperature anomalies cannot robustly reconstruct the AMOC on annual, decadal or centennial timescales. Instead, we find that the net air-sea heat flux anomaly between the Arctic and any given latitude between 26.5°N and 50°N in the North Atlantic is tightly linked to the AMOC anomaly at that latitude on decadal and centennial timescales. On these timescales, the air-sea heat flux proxy works through the conservation of energy, in which energy transferred laterally into the region is typically released to the atmosphere through surface heat fluxes. On annual timescales, however, air-sea heat flux anomalies are modulated more so by atmospheric variability and less by AMOC anomalies. Based on the here identified relationship and observation-based estimates of the past air-sea heat flux in the North Atlantic from reanalysis products, we show the decadal averaged AMOC at 26.5°N has not weakened from 1963 to 2017 although substantial variability exists. Furthermore, we find no decline in the AMOC at any other latitude, though the decadal variability appears distinct between subtropical and subpolar latitudes. This result aligns with previous work that has shown the lack of meridional coherence in the AMOC, and the presence of distinct overturning cells.

Observations over the last 40 years show that the Atlantic Ocean salinity pattern has amplified, likely in response to changes in the atmospheric branch of the global water cycle. Observational estimates of oceanic meridional freshwater transport (FWT) at 26.5° N indicate a large increase over the last few decades, during an apparent decrease in the Atlantic Meridional Overturning Circulation (AMOC). However, there is limited observation based information at other latitudes. The relative importance of changing FWT divergence in these trends remains uncertain. Ten models from the Coupled Model Intercomparison Project Phase 5 are analysed for AMOC, FWT, water cycle, and salinity changes over 1950–2100. Over this timescale, strong trends in the water cycle and oceanic freshwater transports emerge, a part of anthropogenic climate change. Results show that as the water cycle amplifies with warming, FWT strengthens (more southward freshwater transport) throughout the Atlantic sector over the 21st century. FWT strengthens in the North Atlantic subtropical region in spite of declining AMOC, as the long-term trend is dominated by salinity change. The AMOC decline also induces a southward shift of the Inter-Tropical Convergence Zone and a dipole pattern of precipitation change over the tropical region. The consequent decrease in freshwater input north of the equator together with increasing net evaporation lead to strong salinification of the North Atlantic sub-tropical region, enhancing net northward salt transport. This opposes the influence of further AMOC weakening and results in intensifying southward freshwater transports across the entire Atlantic.

The Atlantic Meridional Overturning Circulation (AMOC) has profound impacts on various climate phenomena. Using both observations and simulations from the Coupled Model Intercomparison Project Phase 3 and 5, here we show that most models underestimate the amplitude of low‐frequency AMOC variability. We further show that stronger low‐frequency AMOC variability leads to stronger linkages between the AMOC and key variables associated with the Atlantic multidecadal variability (AMV), and between the subpolar AMV signal and northern hemisphere surface air temperature. Low‐frequency extratropical northern hemisphere surface air temperature variability might increase with the amplitude of low‐frequency AMOC variability. Atlantic decadal predictability is much higher in models with stronger low‐frequency AMOC variability and much lower in models with weaker or without AMOC variability. Our results suggest that simulating realistic low‐frequency AMOC variability is very important, both for simulating realistic linkages between AMOC and AMV‐related variables and for achieving substantially higher Atlantic decadal predictability.

The Atlantic Meridional Overturning Circulation (AMOC) is crucial for global ocean carbon and heat uptake, and controls the climate around the North Atlantic. Despite its importance, quantifying the AMOC’s past changes and assessing its vulnerability to climate change remains highly uncertain. Understanding past AMOC changes has relied on proxies, most notably sea surface temperature anomalies over the subpolar North Atlantic. Here, we use 24 Earth System Models from the Coupled Model Intercomparison Project Phase 6 (CMIP6) to demonstrate that these temperature anomalies cannot robustly reconstruct the AMOC. Instead, we find that air-sea heat flux anomalies north of any given latitude in the North Atlantic between 26.5°N and 50°N are tightly linked to the AMOC anomaly at that latitude on decadal and centennial timescales. On these timescales, air-sea heat flux anomalies are strongly linked to AMOC-driven northward heat flux anomalies through the conservation of energy. On annual timescales, however, air-sea heat flux anomalies are mostly altered by atmospheric variability and less by AMOC anomalies. Based on the here identified relationship and observation-based estimates of the past air-sea heat flux in the North Atlantic from reanalysis products, the decadal averaged AMOC at 26.5°N has not weakened from 1963 to 2017 although substantial variability exists at all latitudes.

A potential future slowdown or acceleration of the Atlantic Meridional Overturning Circulation (AMOC) would have profound impacts on global and regional climate. Recent studies have shown that AMOC responds, among many other processes, to anthropogenic changes in tropical Indian ocean (TIO) temperature. However, internal unforced co-variations between these two basins are largely unexplored as of yet. Here, we use the ERSST v5, HadISST v1, and COBE v2 gridded observational products for the period 1870-2014, as well as dedicated simulations with coupled climate models, and show that internal changes in sea surface temperature gradients between the Indian and Atlantic Ocean (SSTgrad) can drive teleconnections that influence internal variations of North Atlantic climate and AMOC.We separate the unforced observed component (i.e., internal signal) from the forced signal following the residuals method presented by Smith et al. (2019). In the absence of direct AMOC observation we estimate AMOC variability from an SST index (SSTAMOC; Caesar et al., 2018). We find a robust observed relationship between the unforced tropical SSTgrad and SSTAMOC when TIO leads by ~25 years. This time-lag is in line with a recently described mechanism of anomalous tropical Atlantic rainfall patterns that originate from TIO warming and cause anomalously saline tropical Atlantic surface water which slowly propagate northward into the subpolar North Atlantic, ultimately altering oceanic deep convection and AMOC (Hu and Fedorov, 2019; Ferster et al. 2021). Our study now suggests that it is the tropical SSTgrad that drives those AMOC changes, with a limited role for the western tropical Pacific. Pre-industrial control simulations with the IPSL-CM6A-LR model confirm this relationship, indicating a time lag of ~25 years between SSTgrad and SSTAMOC variations. These simulations also confirm that the SSTAMOC is representative of unforced AMOC variations when SSTAMOC leads by 5 years. This work therefore indicates that an unforced pathway between tropical ocean temperature and AMOC exists with a ~20 year lag, which opens the potential for using SSTgrad as precursor to predict future AMOC changes. Caesar, L., Rahmstorf, S., Robinson, A., Feulner, G., & Saba, V. (2018). Observed fingerprint of a weakening Atlantic Ocean overturning circulation. Nature, 556(7700), 191-196.Ferster, B. S., Fedorov, A. V., Mignot, J., & Guilyardi, E. (2021). Sensitivity of the Atlantic meridional overturning circulation and climate to tropical Indian Ocean warming. Climate Dynamics, 1-19.Hu, S., & Fedorov, A. V. (2019). Indian Ocean warming can strengthen the Atlantic meridional overturning circulation. Nature climate change, 9(10), 747-751.Smith, D. M., Eade, R., Scaife, A. A., Caron, L. P., Danabasoglu, G., DelSole, T. M., ... & Yang, X. (2019). Robust skill of decadal climate predictions. Npj Climate and Atmospheric Science, 2(1), 1-10.

Received 22 May 2013; revised 4 September 2013; accepted 8 September 2013; published 23 October 2013. [1] In this paper, simulated variability of the Atlantic Multidecadal Oscillation (AMO) and the Atlantic Meridional Overturning Circulation (AMOC) and their relationship has been investigated. For the first time, climate models of the Coupled Model Intercomparison Project phase 5 (CMIP5) provided to the Intergovernmental Panel on Climate Change Fifth Assessment Report (IPCC-AR5) in historical simulations have been used for this purpose. The models show the most energetic variability on the multidecadal timescale band both with respect to the AMO and AMOC, but with a large model spread in both amplitude and frequency. The relationship between the AMO and AMOC in most of the models resembles the delayed advective oscillation proposed for the AMOC on multidecadal timescales. A speed up (slow down) of the AMOC is in favor of generating a warm (cold) phase of the AMO by the anomalous northward (southward) heat transport in the upper ocean, which reversely leads to a weakening (strengthening) of the AMOC through changes in the meridional density gradient after a delayed time of ocean adjustment. This suggests that on multidecadal timescales the AMO and AMOC are related and interact with each other.

The Atlantic Meridional Overturning Circulation (AMOC) is projected to weaken in the future due to increasing greenhouse gas concentrations, but it is still debated whether anthropogenic climate change can induce an irreversible collapse or “tipping” of the AMOC. Meltwater from the Greenland ice sheet has often been invoked as a key mechanism for a potential AMOC tipping, but it is not explicitly represented in most state-of-the-art (CMIP6) climate models, adding further uncertainty to assessing the likelihood of irreversible AMOC change.Here, we perform ensemble simulations with the CMIP6 model EC-Earth3 to assess the effects of future Greenland ice sheet melt and to probe AMOC reversibility with and without Greenland meltwater. To this end, we force EC-Earth3 with a strong global warming scenario (SSP5-8.5) and a high-end Greenland meltwater estimate from the coupled climate–ice sheet model CESM2-CISM2 until 2300.We find that, as expected, the addition of Greenland meltwater significantly exacerbates the greenhouse gas-induced AMOC weakening especially after the 21st century, with differences mostly attributable to the Arctic Ocean. However, we find no indication of an abrupt AMOC weakening. We then branch off idealized reversibility experiments in which the meltwater and/or greenhouse gas forcings are reversed. Although the AMOC recovery is slow (around two centuries), meltwater-driven additional AMOC weakening in EC-Earth3 appears to be reversible. Regardless of the added meltwater, the AMOC also recovers in an idealized CO2 ramp-down experiment, even overshooting its present-day strength. While our modeling results show little support for an irreversible AMOC change due to future Greenland ice sheet melt, they do underline the importance of representing meltwater in future projections, including overshoot pathways.