Performance of Satellite Rainfall Estimation Algorithms during TOGA COARE

Abstract

Over 50 satellite rainfall algorithms were evaluated for a 5° square region in the equatorial western Pacific Ocean during TOGA COARE, November 1992–February 1993. These satellite algorithms used GMS VIS/IR, AVHRR, and SSM/I data to estimate rainfall on both instantaneous and monthly timescales. Validation data came from two calibrated shipboard Doppler radars measuring rainfall every 10 min. There was large variation among algorithms in the magnitude of the satellite-estimated rainfall, but the patterns of rainfall were similar among algorithm types. Compared to the radar observations, most of the satellite algorithms overestimated the amount of rain falling in the region, typically by about 30%. Patterns of monthly observed rainfall were well represented by the satellite algorithms, with correlation coefficients with the observations ranging from 0.86 to 0.90 for algorithms using geostationary data and 0.69 to 0.86 for AVHRR and SSM/I algorithms when validated on a 0.5° grid. Patterns of instantaneous rain rates were also well analyzed, with correlation coefficients with the radar observations of 0.43–0.58 for the geostationary algorithms and 0.60–0.78 for SSM/I algorithms. Two case studies are presented to demonstrate the capability of one IR algorithm and three microwave algorithms to estimate instantaneous rainfall rates in the Tropics. The three microwave algorithms differed in their estimates of rain area but all showed greater ability than the IR algorithm to reproduce the spatial pattern of rainfall.