Size Matters: Why is there overturning in the Atlantic but not in the Pacific?

The model proposed by Gnanadesikan (1999) is extended for two basins, one twice as wide as the other, connected by a re-entrant channel. Sinking occurs in the northern hemisphere of one basin. They are coupled at the northern edge of the channel by a geostrophic exchange flow caused by the difference in depth of the thermocline at the eastern boundaries of each of the basins. Two states, one with sinking in the narrower basin, and one with sinking in the wider basin, are compared. The transport per unit width in the former case is found to be the larger of the two. In an MITgcm run with the same configuration as the theoretical model, northern sinking is found to be stable in the narrower basin, but not in the wider basin, despite zonally uniform surface forcing for salinity, temperature and wind. A geostrophic exchange flow comparable to that in the theoretical model is found in the MITgcm. We hypothesize that this exchange flow may be important in moving salt from the Pacific to the Atlantic, and therefore may set deep-water formation to occur in the Atlantic.