Abstract
A reduced-gravity, primitive equation, upper-ocean GCM is used to study subduction pathways in the Atlantic subtropical and tropical gyres. In order to compare the different responses in the pathways to strong and weak wind stress forcings, Hellerman and Rosenstein (HR) and da Silva (DSV) climatological annual-mean and monthly wind stress forcings are used to force the model. It is shown that subtropical–tropical communication is dependent on both the strength and structure of the wind forcing. A comparison between the two experiments shows two results for the North Atlantic: 1) the full communication window between the subtropical and tropical gyres is similar in width despite the difference in the intensity of the winds and 2) the interior exchange window width is substantially larger in the weak forcing experiment (DSV) than the strong forcing experiment (HR), accompanied by a larger transport as well. The South Atlantic exhibits a similar communication between the subtropics and Tropics in both cases. The annual-mean of the seasonally varying forcing also supports these results. A two-layer ventilated thermocline model is developed with a zonally varying, even though idealized, wind stress in the North Atlantic, which includes the upward Ekman pumping region absent from the classical ventilated thermocline model. The model shows that the communication window for subduction pathways is a function of the zonal gradient of the Ekman pumping velocity, not the Ekman pumping itself, at outcrop lines and at the boundary between the subtropical and tropical gyres. This solution is validated using three additional GCM experiments. It is shown that the communication windows are primarily explained by the ventilated thermocline model without considering the buoyancy effects. From the GCM experiments, the interior exchange window, which is a part of the communication window and cannot be explained by the ventilated thermocline model, is widened by two factors: 1) eliminating part of the positive Ekman pumping region in the eastern North Atlantic and 2) weakening the Ekman pumping over the whole region. The implications of these results suggest that changes in the wind forcing on the order of the difference in the wind products used here can have a significant effect on the attributes of the communication window and, hence, the thermocline structure at lower latitudes.
Corresponding author address: Dr. Tomoko Inui, International Arctic Research Center, University of Alaska, 930 Koyukuk Dr., Fairbanks, AK 99775-7335. Email: tomoko@iarc.uaf.edu