Estimating spatially varying eddy diffusion coefficients

Geostrophic eddies influence the transport of heat, carbon and other climatically important tracers in the global ocean. The rate of eddy mixing is typically quantified through an eddy diffusivity. In coarse resolution ocean models, which are not capable of resolving these eddies, the mixing done by geostrophic eddies has to be parameterised. Yet in their quest to resolve many topical climate problems, the models' credibility is challenged by their extreme sensitivity to the representation of these parameterisations. One of the main challenges in developing eddy parameterisations is to understand the spatial and temporal variability of eddy diffusion coefficients and their relationship to variables available from the coarse-resolution ocean model.

In this talk we will use satellite-derive geostrophic velocity fields to directly simulate the evolution of passive tracers. From these simulations we calculate eddy diffusivities and show how these results relate to properties available from coarse-resolution ocean models. To achieve this, we will use mixing length theory, modified to account for the suppression of eddy mixing by mean flows and wave propagation, to determine the relevant mixing lengths. These results show that a) mixing lengths are proportional to the size of eddies, and not proportional to the Rossby deformation radius as previously thought, b) eddies move at a phase speed determined by linear Rossby wave theory and c) once the actual size of eddies is known, eddy diffusion coefficients can be accurately predicted from mean flow properties.