A NEURAL-NETWORK ALGORITHM FOR SEA-ICE EDGE CLASSIFICATION

The NASA Scatterometer (NSCAT), launched in August 1996, is designed t o measure wind vectors over ice-free oceans. To prevent contamination of the wind measurements, by the presence of sea ice, algorithms based on neural network technology have been developed to classify ice-free ocean surfaces. Neural networks trained using polarized alone and pol arized plus multi-azimuth ''look'' Ku-band backscatter are described. Algorithm skill in locating the sea ice edge around Antarctica is expe rimentally evaluated using backscatter data from the Seasat-A Satellit e Scatterometer that operated in 1978, Comparisons between the algorit hms demonstrate a slight advantage of combined polarization and multi- look over using co-polarized backscatter alone. Classification skill i s evaluated by comparisons with surface truth (sea ice maps), subjecti ve ice classification, and independent over lapping scatterometer meas urements (consecutive revolutions).