ACC Subduction by Mesoscales

The mesoscale contribution to Subduction in the Southern Ocean was recently studied by Salle'e and Rintoul (2011, SR) using the following mesoscale model. The adiabatic A-regime was modeled with the GM stream function, the diabatic D-regime was modeled with tapering functions,the D-A interface was taken to be at the mixed layer depth and the mesoscale diffusivity was either a constant or given by a 2D model. Since the resulting subductions were an order of magnitude smaller than the data of ±200 m/yr (Mazloff et al., 2010), SR showed that if instead of the above model-dependent mesoscale diffusivities, they employed the ones by Salle'e et al. (2008) from surface drifter observations, the subductions compared significantly better with the data. On those grounds, SR suggested a tenfold increase of the diffusivity. In this work, we suggest that since the mesoscale diffusivity is but one component of a muc large mesoscale parameterization, one should first assess the latter's overall performance followed by the assessment of the predicted ACC subduction. We employ the mesoscale model formulated31 in Canuto et al. (2018; 2019, that includes recent theoretical and observational advances and that was assessed against a variety of data including the output of 17 other OGCMs. The ACC diffusivities compare well with drifter data by Salle'e et al. (2008) and the ACC subduction rates are in agreement with the data.