Laboratory Studies of Stirring by Small-Scale Geostrophic Motions

Laboratory experiments were used to study lateral stirring by small-scale geostrophic motions formed by the adjustment of diapycnal mixing events. Experiments build on analytical and numerical studies by Sundermeyer et al. (2005) and Sundermeyer and Lelong (2005), which predicted the amount of lateral dispersion caused by vortical mode stirring. In the present study, vertically oscillated mesh grids produced localized mixed patches that adjusted to form stable, geostrophically balanced eddies within a rotating, linearly stratified fluid. The geostrophic adjustment dynamics of isolated mixed patches were examined first. These showed that measured velocities and radial displacements associated with each lens depended largely on the extent of each adjustment, i.e. the size of the predicted Rossby radius, R, relative to the initial horizontal length scale of the mixed patch, L. Experiments showed that simple geostrophic velocity scaling using initial mixed patch parameters overestimated the observed velocity by 1-2 orders of magnitude for 1 „T R/L „T 10, consistent with published analytical solutions for the same problem (McWilliams, 1988). A second group of experiments quantified the cumulative effect of many such events as it relates to the parameter dependence of mixing event frequency, ƒÖƒz and R, predicted by Sundermeyer et al. (2005). The linear dependence on ƒÖ was confirmed. The observed diffusivity, however, trended lower with R in a similar way as the observed R/L dependence of the single eddy velocity. Using the analytically predicted velocity in Sundermeyer et al.'s (2005) dispersion model for R/L „d 1, rather than simple geostrophic velocity scaling, more closely predicted the resultant diffusivity observed in the multi-eddy stirring experiments. These results suggest an approach for extending the weakly non-linear geostrophic scaling of Sundermeyer et al. (2005) to a more general non-geostrophic form.