Abyssal flow and forcing: Observations and numerical models of the Argentine Basin

The circulation of the southwestern South Atlantic is studied using observations and models.Data collected in 1988-1989 in the Argentine Basin show cooling and freshening on abyssal isopycnals relative to historical data. This was accompanied by a reduction in density of the abyssal water. Maps of abyssal flow suggest that the main features of the basin include a crescent shaped western boundary current and recirculation, with interior anticy-clonic flow. The watermass change may be traced to anomalous mid-gyre deep convection in the Weddell Sea.An idealized topography isopycnal channel model (Spall, 1994) is used to study the dynamics of the abyssal flow. The model generates an active wave field which maintains the mean circulation. This comprises a western boundary current, broad northward flow on the eastern boundary, and an interior anticyclone. On the eastern boundary, a northward flowing jet is driven by eddy flux of eddy potential vorticity across mean potential vorticity contours (Rhines and Holland, 1979). Tracer distributions depend on both the eddy driven mean flow field, and the time varying flow. Tracer advected in mean flow and diffused with a diffusivity coefficient based on the eddy kinetic energy reproduces the mixing effects of the waves well.The model response to inflow variability depends on forcing frequency, and resonance with natural topographic waves. Forcing variability tends to partition energy into subharmonic frequencies. Results from a model with realistic topography are similar in regions where the flow is rapid and nonlinear. On the western boundary, eddy fluxes drive a recirculation cell which is seen in observations. Where mean flow and wave activity are small, flow is along topographic contours. In contrast to the ideal model, variations in inflow forcing are quickly scattered and damped in the realistic topography model.The research here reflects the awareness that abyssal watermass properties and circulation are not static. Both circulation and watermass properties are shown in model and observations to vary on decadal time scales, and this variability is critical to defining the dispersion of a tracer substance. The modeling work in particular shows the importance of accurate topographic representation in global and regional circulation modeling