The Robust Rain-Rate Retrieval Algorithm

The robust rain-rate retrieval (RRR) algorithm produces accurate (error less than 15%) rain-rate estimates even in the presence of very noisy rain-echo signals. Many active and passive algorithms have been studied for measur- ing rain rates from space. Some algorithms require a good initial-guess rain-rate profile. Many rain-profiling algorithms use the ratios of received powers in adjacent (or nearby) range cells to eliminate the poorly known radar calibration factor. Because the signals for the individual range cells are noisy, the power ratios are very noisy, resulting in erroneous rain-rate estimates. Moreover, most algorithms estimate the rain rate separately for each range cell, rather than combining information from many cells jo compute a complete rain-rate profile. The RRR algorithm uses a simple but powerful technique to estimate rain rates accurately: bind the estima- tors with guess intervals. The initial-guess rain-rate profile derived from the guess intervals need not resemble in any way the true rain-rate profile. Utilizing radar properties of rain and concepts from existing algorithms further refines the estimator.