An Isopycnic Model Study of the North Atlantic. Part II: Interdecadal Variability of the Subtropical Gyre

In a companion paper, a spinup integration of the North Atlantic Ocean with the Miami isopycnic-coordinate model was presented. The wintertime mixed layer in the central North Atlantic was subject to relatively little change in salinity or depth but cooled markedly, most probably because of heat loss associated with a partial surface relaxation to climatological sea surface temperatures in a region in which the Gulf Stream was too far to the north. This mixed layer cooling caused the isopycnic layers in the model ventilated subtropical gyre to rise and, surprisingly, to warm. While the experiment was not an attempt to simulate changes in the real Atlantic Ocean, it nevertheless appears from observations that, in recent decades, the mixed layer in nature has undergone a change similar to that exhibited by the model mixed layer. Since it is expected that changes in the ventilated subtropical gyre will be governed largely by changes in the mixed layer in the central North Atlantic, from where the ventilating water masses are subducted, one might therefore anticipate similarities between dine changes in the ventilated regions of the gyres in the model and the real world, even though the cause of the mixed layer changes in the real world may have been different from that in the model. The present paper shows that this is indeed so. In particular, the model behavior closely parallels observed changes in the ventilated subtropical gyre reported by Levitus, in a study of differences between two pentads. The degree of similarity between the model and the observations, including in particular warming of the isopycnic surfaces, leads to a proposal that the changes Levitus observed were caused largely by the subduction of water masses from a cooler mixed layer. Historical changes in the characteristics of the warm North Atlantic Central Water may also be explained by this mechanism. Changes in the wind stress or Ekman pumping fields do not necessarily need to be invoked. Overall, the model provides a framework in which observations from a number of different sources can be understood in a coherent fashion and allows new insights to be gained into the interdecadal variability of the Atlantic Ocean.