A feasible method for merging the TRMM microwave imager and precipitation radar data

Abstract

The TRMM satellite has been operating for more than 14 years, achieving enormous cloud and precipitation measurements through its onboard multiple sensors including both passive and active instruments. However, due to the distinct sampling geometry and spatial resolution of these sensors, notable obstacle exists in using the TRMM multi-sensor data, especially for those from the TRMM microwave imager (TMI) and precipitation radar (PR). In this study, based upon the fact that most contiguous TMI pixels are inherently overlapping (especially at the lower frequencies), the horizontal distribution of microwave brightness temperature could be assumed to form a locally continuous field, which is formulized as conicoid by using the moving surface fitting (MSF) method. According to this principle, the TMI measurements are spatially collocated to match the PR measurements. The precipitation profiles measured by the PR and microwave brightness temperatures observed by the TMI are thus readily merged, supplying great convenience for in-depth data analysis. The general accuracy of the MSF method is evaluated through simulations and the quantitative validation is performed by comparing the mean, standard deviation, and frequency distribution between the original data and the merged data. The statistical results show a maximum bias of the merged TMI brightness temperature less than 1.5K and a relative difference less than 0.9%, attesting the reasonableness of the MSF method for rearranging passive microwave measurements at a finer resolution. Demonstrations of typical applications are present and suggested that such a data-process method could be used with considerable confidence in many relevant applications.