Passive Microwave Precipitation Retrievals Using TMI during the Baiu Period of 1998. Part I: Algorithm Description and Validation

Abstract

The Baiu front is a subtropical convergence zone that is formed over east Asia in early summer (hereinafter referred to as the Baiu period). In this study, an overocean precipitation retrieval algorithm is developed to retrieve precipitation for the Baiu period from brightness temperatures (TBs) supplied by the Tropical Rainfall Measuring Mission (TRMM) Microwave Imager (TMI). The basic idea of the algorithm is to find the optimal precipitation that gives radiative transfer model (RTM)-calculated, field-of-view?averaged TBs that fit best with the TMI TBs at 10.7, 19.7, and 85.5 GHz with vertical polarization. For the RTM calculation, spatial precipitation inhomogeneity and freezing-level height are estimated from TMI TBs. The optimal precipitation with 10-km resolution is obtained by solving the gradient equation of a cost function that is a weighted sum of squares of TB differences between the TMI observation and the RTM calculation. Precipitation retrieved by this algorithm was validated using TRMM precipitation radar (PR) data from the western part of Japan during June?July of 1998. The results indicate the following. Mesoscale (?100 km) structures of precipitation disturbances were retrieved successfully with the algorithm. However, there were discrepancies in position and strength of individual rain cells between the precipitation retrievals and PR data. Precipitation retrieved by the algorithm agreed well with PR data within the precipitation range of 1?25 mm h?1, irrespective of precipitation type. Experimental algorithms were applied to some cases during this period to examine the effect of improvements made to the algorithm, as compared with the authors? previous work. The results show that use of TBs at 10.7 GHz largely improved heavy precipitation retrievals, and that correction using estimated spatial precipitation inhomogeneity alleviated underestimation of heavy precipitation caused by beam-filling error. It was also found that estimating freezing-level height slightly reduced precipitation retrieval errors.