We explore the consequences of allowing ocean mixing to vary with climate by testing several physical parameterizations of diapycnal diffusivity. While microstructure and tracer measurements have shown that diapycnal diffusivities in the tropical thermocline are typically of O(10-5 m2/s), tropical cyclones provide an isolated blast of intense mixing. We estimate the temporal and spatial mean value of mixing provided by tropical cyclones around the globe, and apply it in a coupled atmosphere-ocean model. We find that the magnitude of the oceans' heat flux and the latitudes at which the ocean makes an important contribution to the total meridional transport of heat increase in warmer climates.
In additional experiments, we allow mixing to vary inversely with stratification. This holds the energy available to mixing constant with climate, rather than the diffusivity itself. We find that while the abyssal circulation can be strongly modulated by such a parameterization, the heat flux is less strongly affected.
Finally, we combine these parameterizations to test a more realisitic dependence of mixing on climate. The diapycnal diffusivity varies inversely with stratification, but the amplitude is modified in the upper tropical oceans based on the depth to which tropical cyclones directly mix (generally the upper most 200 meters). We find that the cumulative effects of tropical cyclone mixing could make a modest contribution to the oceans' heat flux in the present climate, and could make a more substantial one in a warmer one.