A model-based assessment of the climate impacts of the observed AMOC weakening and their sensitivity to model resolution

Current observations of the Atlantic Meridional Overturning Circulation (AMOC) from the RAPID array show a long-term weakening of nearly 2 Sv since 2004, which would be expected to have produced noticeable and widespread climate impacts. However, these impacts are challenging to isolate in  observations because they are confounded by concurrent global warming signals that also induce long-term trends. To study the impacts associated with persistent AMOC weakening, studies typically rely on long runs forced with freshwater perturbations. The highly idealized nature of these experiments, together with the primary use of low resolution models, limits their applicability to AMOC-related impacts over the recent historical period. 

 

Here, we propose an alternative approach based on the analysis of a large ensemble of control simulations, in which the confounding anthropogenic trends are avoided. We use a total of 14 global coupled simulations from the HighResMIP exercise and the EERIE project. Eight of these simulations were performed with eddy-rich ocean configurations (with a horizontal resolution of about 8 km in mid-latitudes), while the remaining simulations represent the low-resolution counterparts of six of the former. In these runs, we first select 19-year periods in which the AMOC trends are comparable in magnitude to that observed by the RAPID array for 2005-2023 and then produce the associated composites describing the concomitant trends in sea level pressure, surface atmospheric temperature, and precipitation. We compare the composites across resolutions to determine whether and how resolving mesoscale eddy interactions enable different climate impacts. We also repeat the analyses for the few cases in which the simulated trends are at least 50 % stronger than in RAPID, to learn about the potential future changes to come if the observed weakening trend intensifies.