A Sea Ice Retrieval Algorithm Based on Microwave Emissivities

Abstract:

We will present results from a sea ice concentration and type retrieval algorithm using microwave radiances from AMSR-2 in machine learning models. The algorithm has four notable features: (1) the use of a 11 GHz channel, (2) satellite brightness temperatures that have been calibrated in an absolute sense, (3) algorithm parameterization in terms of surface emissivity rather than brightness temperature, and (4) detection of thin sea ice that is less than 30 cm thick in addition to the first year and multiyear sea ice classifications. The resulting sea ice concentrations and sea ice types are compared with those retrieved by the National Snow and Ice Data Center and Canadian Ice Service. One novel feature of this approach is that it uses estimates of sea ice emissivities that are correct in an absolute sense as the algorithm basis. We believe that this is the first time that absolute measurements of sea ice emissivity have been derived from satellite observations.