Greenhouse Gas-Forced Changes in the Atlantic Meridional Overturning Circulation and Related Worldwide Sea-Level Change
To investigate the sensitivity of the AMOC to heat flux perturbation in more detail, with more models, in this paper we propose and demonstrate two further FAFMIP experiments. These modify the heat flux forcing in the North Atlantic only. In this work, we use both original and the new FAFMIP experiments to address the following questions: (1) How does the AMOC respond to perturbations to fluxes of heat, freshwater, and momentum? (2) To what extent is the AMOC sensitive to heat flux forcing in the North Atlantic versus elsewhere? (3) Is the spread of AMOC weakening across different models related to the spread of the heat input into the North Atlantic or to interior processes? (4) How does a weakened AMOC affect the global distribution of ocean heat content change? (5) How does a weakened AMOC affect patterns of regional dynamic sea-level change?
Using two new heat-flux-forced experiments, we find that the AMOC weakening is mainly caused by and linearly related to the North Atlantic heat flux perturbation, and further weakened by a positive coupled heat flux feedback. The quantitative relationships are model-dependent, but few models show significant AMOC change due to freshwater or momentum forcing, or to heat flux forcing outside the North Atlantic. AMOC decline causes warming at the South Atlantic-Southern Ocean interface. It does not strongly affect the global-mean vertical distribution of ΔOHC, which is dominated by the Southern Ocean. AMOC decline strongly affects ocean dynamic sea level in the North Atlantic, with smaller effects in the Southern Ocean and North Pacific. The ensemble-mean ocean dynamic sea level and ocean heat content anomaly patterns are mostly attributable to the heat added by the flux perturbation, with smaller effects from ocean heat and salinity redistribution. The ensemble spread, on the other hand, is largely due to redistribution, with pronounced disagreement among the AOGCM
Differences in the fluxes of heat, freshwater, and momentum have been invoked to explain the diverse AMOC responses forcing noted in the literature. Here, we demonstrate that even when consistent forcing is applied, our large ensemble of 15 AOGCMs produces inconsistent AMOC responses. For freshwater and momentum, most AMOC changes are small or insignificant. For heat, the ensemble shows a large spread of AMOC weakening. These results suggest that differences in ocean model structure or control state give each AOGCM a unique sensitivity to forcing.
Differences in the distributions of temperature and salinity (i.e. model state biases) may be a cause of the diverse responses. While it is understood that ocean buoyancy contrasts underpin the AMOC, it is unclear exactly which buoyancy contrasts are relevant; whether between the NADW and the low latitude thermocline waters or Southern Ocean intermediate waters. Temperature and salinity play different roles in setting buoyancy contrasts that support the AMOC, and the relative importance of the two tracers is likely to differ across models. Further, details of the surface balance between evaporation and precipitation are also likely to be key determinants of the overturning. It is possible that the buoyancy contrasts that drive the AMOC are specific to each model, and so perturbations to those contrasts may produce diverse responses. The FAFMIP common forcing framework will be useful to probe these diverse responses in future investigation.