A sigma-coordinate primitive equation model for studying the circulation in the South Atlantic Part II: Meridional transports and seasonal variability

The mean and seasonal variability of the circulation and meridional heat transport in the South Atlantic are investigated using a set of numerical experiments. The primitive equation model uses a topography-following (sigma) coordinate. The model domain is limited to the South Atlantic basin. Artificial boundaries at Drake Passage, between Brazil and Angola, and between South Africa and Antarctica are treated as open boundaries. Finally, recent and self-consistent estimates of seasonal fluxes are used to define a model-dependent atmospheric forcing. Quasi-diagnostic simulations forced by constant climatological winds are first conducted to determine the sensitivity of model solutions to bottom topography smoothing, and to diagnose meridional fluxes from a mass field that is relaxed to the annual climatology of Levitus (1982). Model results show good agreement with known climatological circulation features in this basin, especially in the Confluence Region, where coarse resolution models usually give smooth structures. Sensitivity studies show that the more detailed features of the circulation are influenced by the model bathymetry. The model simulates a meridional circulation whose upper branch (the return flow that balances the southward flow of North Atlantic Deep Water) is composed of Intermediate (IW) and Thermocline (TW) Waters. The transport of IW is found to be predominant, and the value of meridional heat transport consequently falls within the low estimates. We notice that the meridional heat balance is sensitive to the position of the Confluence. When this region occurs too far south, the amount of IW contributing to the return flow of the overturning cell is reduced. Prognostic simulations forced by seasonal winds and heat fluxes are studied to quantify the impact of wind forcing on the circulation in the South Atlantic. Particular attention is focused on meridional transports at 30°S. Analysis of the mean annual circulation confirms that the upper branch of the meridional circulation is predominantly composed of IW (9 Sv), rather than TW (5.3 Sv). The mean transport of the lower branch is 16 Sv, in agreement with recent estimates by Schlitzer (1996). The annual meridional heat transport (0.29 PW) is still within the low estimates, but agrees well with other estimates that give a dominant role to IW (Rintoul, 1991). Original results also concern the variability of the upper branch of the meridional circulation. It is shown that the wind creates seasonal variability in the Subtropical Gyre, which has a marked impact on the water mass balance in the South Atlantic. In winter, a large convergence to the north of the Subtropical Gyre (27°S) reduces the northward flow of IW, whereas stronger Ekman pumping favors an equatorward transport of TW. In summer, this convergence disappears and a larger transport of IW is allowed. Thus a more complex scheme is proposed for the meridional circulation, in which local wind forcing in the South Atlantic Basin has a significant role in preconditioning the surface waters of the global overturning cell.