Air-sea interaction amplifies the AMOC response to global warming

Oluwayemi A. Garuba1, Wilbert Weijer2 and Philip J Rasch1, (1)Pacific Northwest National Laboratory, Richland, WA, United States, (2)Los Alamos National Laboratory, Los Alamos, NM, United States
Abstract:
The response of the Atlantic Meridional overturning circulation (AMOC) to an increase in CO2 is a dynamic interplay between anomalous surface fluxes arising from atmosphere-ocean or ocean-ice interactions, and changes in the ocean circulation itself. In particular, northward advection of heat and salt from the warm and salty subtropical North Atlantic provide powerful feedbacks that can stabilize or destabilize the AMOC. Understanding the relative roles played by atmospheric feedbacks and advective feedbacks in the ocean is important to project the future of the AMOC in our continually warming Earth system. In this study, we aim to isolate the roles of the advective feedbacks from surface flux feedbacks by performing and analyzing 4xCO2 experiments with the Community Earth System Model (CESM). We use a tracer decomposition technique to distinguish a ‘surface-driven’ component of salinity and temperature anomalies, forced by surface fluxes, from a ‘circulation-driven’ component, forced by circulation changes. In a partially coupled simulation, we then remove this circulation-driven component from air-sea interactions, so that the atmosphere does not ‘see’ the temperature changes brought about by the changes in circulation. We find that in the fully coupled case, the AMOC shuts down in response to a fourfold increase in CO2. In contrast, in the partially-coupled simulation, a weaker but still active AMOC is maintained. Our analysis shows that the (positive) salt advection feedback is partially responsible for the slowdown in both simulations. However, in the fully coupled simulation, the (stabilizing) temperature advection feedback is suppressed, as a cooling of the subpolar North Atlantic, in response to reduced heat advection from the subtropics, is quickly counteracted by ocean heat uptake. In the partially-coupled simulation, this cold anomaly is maintained, and suffices to keep the AMOC going. We find no evidence for feedbacks between ocean circulation response and sea ice melt.