Description: Future changes in the Atlantic meridional overturning circulation (MOC) will result from processes both internal and external to the climate system. Here, using the CMIP3 database and simulations with the Kiel Climate Model (KCM), three aspects of modelbased projections of the Atlantic MOC will be discussed: First, while most climate models predict a weakening of the North Atlantic meridional overturning circulation (MOC) during the twenty-first century, large uncertainty exists. Quantification of the different sources of uncertainty – external, internal and model – indicates model error is the largest component, internal variability is significant during the first decades, while scenario uncertainty is almost negligible. The different contributions to model uncertainty – wind and density, salinity versus temperature – will be also discussed. Second, individual studies suggest that multidecadal changes in the MOC are strongly related to large-scale salinity anomalies and therefore to changes in the surface freshwater fluxes and freshwater transport. Here, the general relationship between the MOC and freshwater budget of the Northern Hemisphere is analyzed for the twentieth and twenty-first centuries. Global warming leads to an implified hydrological cycle, which affects the vertical salinity and temperature profiles. The meridional changes in the oceanatmosphere interaction diminish the meridional oceanic density contrast. In the North Atlantic sinking regions, these changes are strongly related to salinity anomalies at the surface. We find in the multi-model mean a strong freshwater export from the Arctic into the northern part of the North Atlantic, stressing the importance of a realistic representation of the hydrological cycle in the models. Third, experiments with KCM indicate that ocean-sea ice-atmosphere interaction in the Southern Ocean could give rise to significant centennial scale changes in the MOC. The model simulates an internal mode of variability on a multi-centennial time-scale set by the interaction between sea-ice cover, open ocean convection in the Weddell Sea and the global MOC. The multi-centennial mechanism is similar to what has been reported with a standalone ocean model forced with mixed boundary conditions and was thought to be spurious

Description: Climate models are known to suffer from various biases and uncertainties. In the subpolar North Atlantic typical biases among models from the Coupled Model Intercomparison Project phase 6 (CMIP6) are found in the mean surface temperature and salinity, and in the mean sea ice concentration, which can affect the air-sea interaction.In this study, we are investigating the diversity of CMIP6 models with respect to their response of the Atlantic Meridional Overturning Circulation (AMOC) to the North Atlantic Oscillation (NAO) in pre-industrial control experiments. This response is sensitive to the mean state of the North Atlantic. We focus on two categories of models: Models that are rather warm-salty versus models that are rather cold-fresh within the subpolar gyre of the North Atlantic. Warm-salty models tend to have a lower sea ice cover in the Labrador Sea (LS) and larger LS heat loss during a positive NAO, compared to cold-fresh models. They also have a weaker stratification in the LS. Sub-surface density changes 1 to 3 years after the NAO are larger in the warm-salty models and establish a zonal density gradient that can cause a stronger delayed AMOC response via the thermal wind balance.These findings stress the need for improvement of the North Atlantic mean state in climate models. Uncertain mean states might further contribute to the uncertainty in AMOC future projections.

Description: An important feature of ocean circulation in the Atlantic is the cross-equatorial and northward transport of water masses at the surface and southward transport at the bottom of the ocean by the Atlantic Meridional Overturning Circulation (AMOC). However, the link between the AMOC and tropical Atlantic variability remain poorly understood. This is partly due lack of long-term observations of the AMOC, with the longest direct measurements available since 2004. Here we construct a dynamic sea-level proxy of the AMOC variability during the twentieth century, which is strongly correlated with the AMOC index during the observational period from 2005-2019 (r=0.50; p=1.48×10-9). This sea-level proxy exhibits a 10-15 year periodicity similar to the pan-Atlantic Decadal Oscillation (ADO) – the north-south bands of alternate anomalies in surface-ocean temperatures with the maximum variance over the tropical Atlantic, and winds from colder bands to the warmer. The sea level-derived proxy leads the ADO pattern by several years, through the interactions of overturning and gyre circulations with Kelvin wave anomalies that propagate from the North Atlantic to the low latitudes and by the thermocline feedback in the Atlantic cold tongue region. The peak of the sea surface temperature variability in the tropical Atlantic in turn drives inter-hemispheric atmospheric teleconnections represented by negative North Atlantic Oscillation phase over the North Atlantic. These findings imply that, rather than a passive role postulated by the prevailing thermodynamic paradigm, AMOC-related ocean circulation plays an active role in ADO variability.