Maximum-likelihood estimation of surface and rain backscattering parameters

An algorithm is developed for estimating simultaneously the rain and surface backscattering parameters from measurements obtained by a spaceborne single-frequency radar at off-nadir incidence. The authors first construct a simple stochastic rain model from which the joint conditional probability density function for the backscattering parameters is derived. This probability density function is updated in range by incorporating the radar data collected at previous ranges. The retrieval algorithm uses this probability density function to estimate the rain parameters, and a minimum-likelihood approach to estimate the surface backscattering coefficient. The preliminary simulation results using the TRMM radar parameters show that the algorithm can estimate the rain attenuation coefficient to better than 25% and the surface backscattering coefficient to better than 2 dB when the rain is uniform in range. The retrieval accuracy can be further improved for non-uniform rain because of lesser deterministic ambiguities existed in the radar measurements