Summary: | In polar regions, ocean-atmosphere interactions are strongly influenced by sea ice and its thickness. Since satellite passive microwave observations became available in the 1970s, significant progress has been made in the study of snow depth and sea ice concentration and extent in these regions. Estimating sea-ice thickness (SIT), instead, turned out to be considerably more difficult. We present a new empirical algorithm to estimate SIT in the Ross and Weddell Seas from Special Sensor Microwave/Imager brightness temperatures. This algorithm combines brightness temperature polarization difference and ratio values to obtain SIT for seasonal ice up to a thickness of about 90 cm during freezing conditions. A series of filters accounts for open water, new ice, and snow on sea ice. Our SIT estimates are consistent with colocated visual ship-based SIT observations made according to the Antarctic Sea Ice Processes and Climate project, showing linear correlation values between 0.73 and 0.96 and root-mean-square-error values between 14 and 24 cm. The seasonal development of the region average SIT derived with our approach agrees with the corresponding values derived from U.S. National Ice Center ice charts. Comparison with colocated polynya distribution maps suggests that the algorithm could be optimized for its performance with regard to SIT values around 50 cm and that a closer investigation of the snow impact on the SIT retrieval is required.
|