Bayesian fusion of TRMM passive and active measurements

Abstract

A Bayesian method was adopted to combine the instantaneous measurements of the Tropical Rainfall Measuring Mission's (TRMM) radar and radiometer. The method makes multiple estimates of the rain-rate profile using the radar reflectivities assuming various plausible values for the drop size distribution (DSD) shape parameters, then selects those parameter values which produce estimates that are most consistent with the passive observations. The resulting estimates are expressed directly in terms of the DSD parameters, thus allowing one to calculate any rain-related quantity, such as rain rate profile, precipitating liquid water profile, etc. The Bayesian approach also allows one to calculate the "error bar" associated with each estimate.