Estimating eddy diffusivities in the Southern Ocean

Geostrophic eddies are a key mechanism for mixing tracers such as heat or carbon across the Southern Ocean. There is an ongoing debate on the magnitude of eddy mixing in the Southern Ocean. Estimates based on the dispersion of Lagrangian floats have found that eddy mixing is largest in the core of the Antarctic Circumpolar Current, whereas estimates based on satellite-derived surface geostrophic velocities showed a suppression in the core of the Antarctic Circumpolar Current and enhancement on its northern flank. In this study we use a three dimensional velocity field from a state estimate of the Southern Ocean to compare estimates of eddy mixing from tracer release experiments and Lagrangian floats. We find the the two estimates converge if enough floats are deployed and the tracer is sampled at sufficient resolution. Eddy mixing is indeed suppressed in the core of the Antarctic Circumpolar Current from the surface down to ~2000m, while it is enhanced on its flanks and at depth. Hence the discrepancies found in the literature are likely the result of poor Lagrangian statistics, rather than any fundamental difference in the various approaches. These results are then discussed in the context of the Diapycnal and Isopycnal Mixing Experiment in the Southern Ocean (DIMES). A major goal of DIMES is to estimate eddy mixing through releases of an anthropogenic tracer and 150 isopycnal RAFOS floats. Based on our numerical results, we will discuss how to analyze the data in order to obtain statistically robust estimates of eddy mixing. In particular we will focus on diagnostics for the lateral and vertical structure of eddy mixing which represent a key uncertainty in parameterizations of the effects of eddies in coarse-resolution ocean models used in climate studies.