A Simplified Wind Vector Algorithm for Satellite Scatterometers

Abstract

We derive a simplified algorithm for retrieving wind vectors from microwave scatterometer observations. The azimuthal dependence of the sea surface radar cross section is modeled by a double-cosine function, rather than the traditional second-order cosine expansion. The algorithm is tested using the aircraft scatterometer observations obtained during the AAFE-RADSCAT Experiment in 1975 and 1976. There is little difference between the performance of the simplified algorithm and the more involved least-squares searching algorithm. The AAFE-RADSCAT aircraft observations are sampled so as to simulate the three-look satellite scatterometer NSCAT that will be launched on the Japanese Advanced Earth Observing Satellite (ADEOS). The results of the retrievals indicate that it is probably not possible to uniquely determine the wind direction using just NSCAT observations because of a 180° ambiguity. However, if forecast models can predict the wind direction to an accuracy of ±90°, thereby eliminating the ambiguity, then the results indicate that NSCAT can determine the wind direction to an accuracy of about ±20°. Better performance is obtained when the three observations are the same polarization (either v-pol or h-pol), as compared to using a mix of v-pol and h-pol observations.