DBE Synthesis Supplement: Deep Circulation in the Tropical and Subtropical South Atlantic

Florida State University, Tallahassee FL

Investigators

Abstract

0220407 Speer In the course of the Deep Basin Experiment under the World Ocean Circulation Experiment (WOCE) a new hydrographic and tracer climatology of the tropical and subtropical South Atlantic was constructed and a re-analysis of hydrographic sections was carried out. The meridional density gradients observed in hydrographic sections west of the Mid-Atlantic Ridge (in the Brazil Basin) indicate a strikingly regular pattern of persistent zonal flows coherent over large spatial scales, consistent with the isopycnal distributions of several tracers in the new climatology. This pattern of alternating zonal flows is also apparent in the output from a state-of-the-art numerical model. While the observations (as well as the model) indicate that the notion of a mean deep circulation is sensible they also show that it is not possible to determine the time-averaged geostrophic flows from synoptic observations by simply determining a reference velocity (barotropic adjustment). The difference between the synoptic and the time-averaged geostrophic zonal velocity fields in the Brazil Basin is baroclinic and varies both in time and space. The excellent agreement between the hydrographically observed and the numerical zonal flows in the South Atlantic suggests that it possible to use the numerical results to separate the observations into long-term mean and low-frequency time-varying components. The current analysis will be extended and applied it to a larger region of the South Atlantic with the overriding goal to quantitatively describe the deep circulation in the tropical and subtropical South Atlantic during the WOCE period. The first step will consist of analyzing the spatio-temporal structure of the time-varying model flow to guide the separation of the time-varying from the mean geostrophic flows determined from the hydrographic sections. The fact that the model appears to simulate the mean zonal flows in the Brazil Basin realistically does not imply that the time-varying flows are as well represented, of course. While it is not possible to directly estimate the time-varying flow component from available hydrographic sections analysis of individual profiles as well as differences between two hydrographic sections will allow the PI's to compare important aspects of the time-variable flow field, such as spatial scales and baroclinic structure to the model output. Furthermore, the results of a recent analysis of the time-varying flows derived from ?oat data (Hogg et al. , 2002) will be used for comparison. In addition to revealing new aspects of the deep circulation (with implications for climate, tracer dispersal, biogeography, etc.), the results of the proposed project will significantly improve the analysis of synoptic hydrographic sections with important implications for inverse models and, therefore, for determining the mean circulation from the WOCE data sets. Direct comparison of numerical-model output and oceanic data is an intrinsic part of the proposed project and will reveal strengths and, more importantly, weaknesses of current numerical models. These insights will be useful for improving the models.