Structure of the Subpolar Gyre in the Australian-Antarctic Basin Derived From Argo Floats

The climatological structure of the subpolar cyclonic circulation off East Antarctica is delineated with Argo float data from the past decade. Up to 40% of the float profiles in the seasonal ice zone have been without satellite positioning. We refined their position data as following the bathymetry to get appropriate positions in the continental margin. The error of the terrain-following interpolation was estimated by using positioned data to be 23 +/- 27 (78 +/- 70) km for 90 (390) day period. Profiles with the under-ice period shorter than 360 days are adopted. The float trajectories reveal the extent of the subpolar gyre adjoined to the westward Antarctic Slope Current to its south and the southernmost eastward jet of the Antarctic Circumpolar Current along 4,000 m isobath to its north. The subpolar circulation in the Australian-Antarctic Basin comprises of a series of quasi-barotropic subgyre circulations, which are bounded by bathymetric spurs in the continental slope. The temperature field reveals shoreward excursions of Circumpolar Deep Water associated with the subgyres, effectively supplying heat to the continental shelves. An along-slope temperature variation up to 1 degrees C in 27.7-27.8 kg/m(3) sigma(theta) indicates an active cross-slope exchange within the layer. Provided the velocity field and the water mass structure, the subsurface water mass exchange is likely accomplished by a combination of topographically controlled mean flow and the eddy transports. Our findings suggest that the bathymetry primarily determines the structure of the subpolar gyre. Plain Language Summary Subpolar gyres in the Southern Ocean are clockwise circulations that separate cold shelf water from warm offshore water. These circulations are essential for the Antarctic climate since they control heat supply onto the continental shelves. In this study, the physical structure of the subpolar gyre off East Antarctica was investigated using autonomous profiling float data with a