AMOC weakening across latitude and time in CMIP6 future scenarios

Abstract. The Atlantic Meridional Overturning Circulation (AMOC) is projected to weaken in the latest generation of climate models used in the Coupled Model Intercomparison Project, Phase 6 (CMIP6). There is uncertainty to the extent of weakening predicted, associated with model differences within the CMIP6 ensemble. The focus of this study is the latitudinal coherence of AMOC decline and how it differs between models and future forcing scenarios. AMOC weakening is presented as a function of latitude and time between 1900 and 2100 for depth-space and density-space. Weakening originates at high latitudes (45–47° N) in depth-space and 51–61° N in density-space) and propagates southwards. As density-space weakening originates further north than depth-space, density-space is better suited for early warning signals. For density-space, the 25 % weakening signal has an 11-year lag between 45° N and 26° N and in depth-space there is a 9-year lag. However, density-space has an additional 8-year lag between 57° N and 45° N giving a total lag of 19 years. As the AMOC weakens, the southward propagation of weakening (south of 40° N) slows with speeds ranging from 6.29 degrees per year to 0.40 degrees per year, dependent on model and scenario. Propagation speeds in depth-space and density-space are similar. In the strongest future forcing scenario, the propagation speed at 5 % weakening is 2.78 degrees per year and at 30 % weakening is 1.45 degrees per year. Also, the speed of southward propagation slows more intensely in less extreme climate change scenarios than in more extreme ones. As the AMOC weakens, the propagation speed of weakening is found to slow for some models while it remains constant for others. The CMIP6 models are categorised by whether the propagation speed is constant, declines to a constant, declines and increases again or continuously declines. The speed of propagation does not appear to be related to AMOC strength.