R/V Knorr Cruise 161-6

7 April to 16 May 2000

Brazil Basin Tracer Release Experiment

James R. Ledwell, Norbert Cortes, Terence Donoghue, Brian Guest,
Peter Lemmond, and Cynthia Sellers

Printable version (PDF)


INTRODUCTION

The purpose of the Brazil Basin Tracer Release Experiment is to measure diapycnal (across isopycnal) mixing and epipycnal (along-isopycnal) mixing and stirring in the deep ocean. This cruise is the fourth in the overall experiment. In the first cruise in early 1996, 110 kg of sulfur hexafluoride (SF6) were released on an isopycnal surface near 4000 meters depth in the eastern part of the basin on the flanks of the Mid-Atlantic Ridge (MAR). The location of the release was near 21.7 S, 18.4 W. The release site was over a zonal valley that leads to the MAR and is about 5000 m deep. The isopycnal surface of the release was defined as the surface on which the potential density anomaly, referenced to 4000 dbar pressure, was 45.9408 kg/m3. The release streaks and results of initial sampling in 1996 are described in Polzin et al. [1997].

Surveys of the tracer patch were made in 1997 and 1998, with the results summarized in Ledwell et al. [2000]. The tracer was found to mix with a diapycnal diffusivity of 3 cm2/s near the target density surface, and with an epipycnal diffusivity of 100 m2/s. The diapycnal diffusivity appears to increase toward the bottom. There also appears to be a flow of about 1 cm/s toward the east in the region beneath the target isopycnal surface. This deep eastward flow had carried about 30% of the tracer east of the release site as of 1998, while the rest of the tracer was in a patch that had drifted west and a little south at a mean speed of about 0.3 cm/s. Tracer concentrations in the east tend to increase with depth until a thick layer of uniform tracer and weak stratification is found in the deep water. This layer can be hundreds of meters thick in the valleys and deep depressions. The concentration profiles in the west tend to be more Gaussian in shape, and centered near the target density surface.

Measurements were made of the dissipation rate of turbulent kinetic energy and temperature variance with the High Resolution Profiler in 1996 and 1997. Results from this instrument are reported in Polzin et al. [1997] and in Ledwell et al. [2000]. They indicate mixing which is enhanced at all depths in the eastern part of the basin where the bathymetry is rough and which increases strongly with depth. The diffusivities, adjusted for certain sampling biases, agreed with the tracer result at the target density surface, and increase to around 10 cm2/s near the tops of the bathymetric ridges. Over smooth bathymetry that characterizes the western half of the basin the diffusivity inferred from the dissipation measurements is very low, on the order of 0.1 cm2/s. The hypothesis for this pattern of mixing is that the tides running over the rough bathymetry generate internal waves which propagate upward into the water column where the associated shear intensifies, inducing turbulent mixing events.

Overall, the mixing in the eastern part of the basin appears to be strong enough to provide the heat flux necessary to balance the heat budget for the abyssal water of the Brazil Basin. Hogg et al. [1982] and Morris et al. [1997] infer from the net rate of Antarctic Bottom Water entering the basin in various temperature and density classes that the diapycnal diffusivity averaged over the whole basin is 2 to 4 cm2/s. The tracer release experiment has found values this large at the target density surface in the eastern part of the basin. The dissipation data suggest that the lower diffusivity in the western part may well be compensated by higher values of diffusivity and of heat flux where the isopycnal surfaces come near the bottom in the east.