Using satellite‐derived ice concentration to represent Antarctic coastal polynyas in ocean climate models

Abstract

The focus of this paper is on the representation of Antarctic coastal polynyas in global ice‐ocean general circulation models (OGCMs), in particular their local, regional, and high‐frequency behavior. This is verified with the aid of daily ice concentration derived from satellite passive microwave data using the “NASA Team 2” (NT2) and the bootstrap (BS) algorithms. Large systematic regional and temporal discrepancies arise, some of which are related to the type of convection parameterization used in the model. An attempt is made to improve the fresh‐water flux associated with melting and freezing in Antarctic coastal polynyas by ingesting (assimilating) satellite ice concentration where it comes to determining the thermodynamics of the open‐water fraction of a model grid cell. Since the NT2 coastal open‐water fraction (polynyas) tends to be less extensive than the simulated one in the decisive season and region, assimilating NT2 coastal ice concentration yields overall reduced net freezing rates, smaller formation rates of Antarctic Bottom Water, and a stronger southward flow of North Atlantic Deep Water across 30°S. Enhanced net freezing rates occur regionally when NT2 coastal ice concentration is assimilated, concomitant with a more realistic ice thickness distribution and accumulation of High‐Salinity Shelf Water. Assimilating BS rather than NT2 coastal ice concentration, the differences to the non‐assimilated simulation are generally smaller and of opposite sign. This suggests that the model reproduces coastal ice concentration in closer agreement with the BS data than with the NT2 data, while more realistic features emerge when NT2 data are assimilated.