Abstract
This note describes a method to improve the current Tropical Rainfall Measuring Mission (TRMM) passive microwave rain rate estimates by adopting a modified probability density function (PDF) of rain rates in Bayesian retrieval algorithms. Unlike the original database that was generated from cloud model simulations mostly for more "eventful" cloud/rain cases, the modified PDF includes the probability of zero rain rates derived from actual observations. Based on analysis of one year TRMM data, this simple remedy lowers the passive microwave rain rate estimates by 3% and reduces the discrepancy between radiometer and radar estimates by 6%.
