Are two modes of thermohaline circulation stable?

The standard version of the coupled ocean–atmosphere model developed at the Geophysical Fluid Dynamics Laboratory (GFDL) of NOAA has at least two stable equilibria. One has arealistic and active thermohaline circulation (THC) with sinking regions in the northern NorthAtlantic Ocean. The other has a reverse THC with extremely weak upwelling in the North Atlantic and sinking in the Circumpolar Ocean of the Southern Hemisphere. Although the model has the seasonal variation of insolation, the structure of these two stable equilibria are very similar to those of a previous GFDL model without the seasonal variation. It is noted that the inactive mode of the reverse THC mentioned above is not a stable equilibrium for another version of the same coupled model which has a large coefficient of vertical subgrid scale diffusion. Although the reverse THC cell was produced in the Atlantic Ocean by a massive discharge of freshwater, it began to transform back to the original direct THC as soon as the freshwater discharge was terminated. It appears that there is a critical value of diffusivity, above which two stable equilibria do not exist. Based upon paleoceanographic evidence, we suggest that the stable state of the reverse THC mentioned above did not prevail during the cold periods of Younger Dryas event which occurred during the last deglacial period. Instead, it is likely that the THC weakened temporarily, but reintensified before it reached the state of the reverse THC with no deep water formation in the North Atlantic Ocean.