An optimal estimation methodology is utilized in a combined radar-radiometer retrieval of precipitation profiles. Taking advantage of the unique information content offered by the TRMM Microwave Imager (TMI) and Precipitation Radar (PR), profiles are derived that minimize the differences between single-instrument retrievals, in accordance with the uncertainty that comes from noise in the observations and radiative transfer models used to simulate microwave brightness temperatures. The algorithm retrieves a parameter related to the raindrop size distribution (DSD) and the cloud liquid water path in each column.

This retrieval is performed over the global oceans for a one-year time period, providing a unique view of global rain DSDs and cloud water-rain water relationships. For example, the results highlight the differences between eastern and western Pacific tropical convection and seasonal differences in DSD near Darwin, Australia. Relationships between high-frequency microwave (e.g., 85 GHz) scattering indices and rain DSD are also examined, which have potential utility for retrieval algorithms over land.