Neural network wind retrieval from ERS-1 scatterometer data

This paper presents a neural network methodology to retrieve wind vectors f rom ERS-1 scatterometer data. First, a neural network (NN-INVERSE) computes the most probable wind vectors. Probabilities for the estimated wind direc tion are given. At least 75% of the most probable wind directions are consi stent with European Centre for Medium-Range Weather Forecasts winds (at +/- 20 degrees). Then the remaining ambiguities are resolved by an adapted PRES CAT method that uses the probabilities provided by NN-INVERSE. Several stat istical tests are presented to evaluate the skill of the method. The good p erformance is mainly due to the use of a spatial context and to the probabi listic approach adopted to estimate the wind direction. Comparisons with ot her methods are also presented. The good performance of the neural network method suggests that a self-consistent wind retrieval from ERS-1 scatterome ter is possible.