Abyssal Mixing in the Brazil Basin*
Additional Document Info
One of the major objectives of the Deep Basin Experiment, a component of the World Ocean Circulation Experiment, was to quantify the intensity and spatial distribution of deep vertical mixing within the Brazil Basin. In this study, basin-averaged estimates of deep vertical mixing rates are calculated using two independent methodologies and datasets: 1) vertical fluxes are derived from large-scale temperature and density budgets using direct measurements of deep flow through passages connecting the Brazil Basin to surrounding basins and a comprehensive hydrographic dataset within the basin interior and 2) vertical mixing rates are estimated from finescale bathymetry and hydrographic data using a functional relationship between turbulent dissipation and bathymetric roughness, deduced from localized measurements of ocean microstructure obtained during the Deep Basin Experiment. The space-time mean estimates of vertical mixing diffusivities across representative surfaces within the Antarctic Bottom Water layer fell in the range <()over bar> similar to 1- 5(x 10(-4) m(2) s(-1)) and were indistinguishable k from each other within the estimation uncertainties. The mixing rates inferred from potential temperature budgets update, and are consistent with, earlier estimates that were based on less data. Mixing rates inferred from budgets bounded by neutral surfaces are not significantly different from the former. This implies that lateral eddy fluxes along isopycnals are not important in the potential temperature budgets, at least within the large estimation uncertainties. Unresolved processes, such as cabbeling and low frequency variability, which complicate inference of mixing from large-scale budgets, have been considered. The agreement between diffusivity estimates based on a modeled relationship between bathymetric roughness and turbulent dissipation, with those inferred from large-scale budgets, provides independent confirmation that the mixing rates have been accurately quantified.