Subduction in an eddy-resolving state estimate of the northeast Atlantic Ocean

Abstract

Relatively little is known about the role of eddies in controlling subduction in the eastern half of the subtropical gyre. Here, a new tool to study the eastern North Atlantic Ocean is created by combining a regional, eddy-resolving numerical model with observations to produce a state estimate of the ocean circulation. The estimate is a synthesis of a variety of in-situ observations from the Subduction Experiment, TOPEX/POSEIDON altimetry, and the MIT General Circulation Model. A novel aspect of this work is the search for an initial eddy field and eddy-scale open boundary conditions by the use of an adjoint model. The adjoint model for this region of the ocean is stable and yields useful information despite concerns about the chaotic nature of eddy-resolving models. The method is successful because the dynamics are only weakly nonlinear in the eastern region of the subtropical gyre. Therefore, no fundamental obstacle exists to constraining the model to both the large scale circulation and the eddy scale in this region of the ocean. Individual eddy trajectories can also be determined. The state estimate is consistent with observations, self-consistent with the equations of motion, and it explicitly resolves eddy-scale motions with a 1/6⁰ grid. Therefore, subduction rates, volume budgets, and buoyancy budgets are readily diagnosed in a physically interpretable context. Estimates of eddy subduction for the eastern subtropical gyre of the North Atlantic are larger than previously calculated from parameterizations in coarse-resolution models. Eddies contribute up to 40 m/yr of subduction locally. Furthermore, eddy subduction rates have typical magnitudes of 15% of the total subduction rate. To evaluate the net effect of eddies on an individual density class,

(cont.) volume budgets are diagnosed. Eddies contribute as much as 1 Sv to diapycnal flux, and hence subduction, in the density range 25.5 < [sigma] < 26.5. Eddies have a integrated impact which is sizable relative to the 2.5 Sv of diapycnal flux by the mean circulation. A combination of Eulerian and isopycnal maps suggest that the North Equatorial Current and the Azores Current are the geographical centers of eddy subduction. The findings of this thesis imply that the inability to resolve or accurately parameterize eddy subduction in climate models would lead to an accumulation of error in the structure of the main thermocline, even in the eastern subtropical gyre, which is a region of comparatively weak eddy motions.