GLOBAL PRECIPITATION ESTIMATES BASED ON A TECHNIQUE FOR COMBINING SATELLITE-BASED ESTIMATES, RAIN-GAUGE ANALYSIS, AND NWP MODEL PRECIPITATION INFORMATION

The ''satellite-gauge-model'' (SGM) technique is described for combini ng precipitation estimates from microwave satellite data, infrared sat ellite data, rain gauge analyses, and numerical weather prediction mod els into improved estimates of global precipitation. Throughout, month ly estimates on a 2.5 degrees X 2.5 degrees lat-long grid are employed . First, a multisatellite product is developed using a combination of low-orbit microwave and geosynchronous-orbit infrared data in the lati tude range 40 degrees N-40 degrees S (the adjusted geosynchronous prec ipitation index) and low-orbit microwave data alone at higher latitude s. Then the rain gauge analysis is brought in, weighting each field by its inverse relative error variance to produce a nearly global, obser vationally based precipitation estimate. To produce a complete global estimate, the numerical model results are used to fill data voids in t he combined satellite-gauge estimate. Our sequential approach to combi ning estimates allows a user to select the multisatellite estimate, th e satellite-gauge estimate, or the full SGM estimate (observationally based estimates plus the model information). The primary limitation in the method is imperfections in the estimation of relative error for t he individual fields. The SGM results for one year of data (July 1987 to June 1988) show important differences from the individual estimates , including model estimates as well as climatological estimates. In ge neral, the SGM results are drier in the subtropics than the model and climatological results, reflecting the relatively dry microwave estima tes that dominate the SGM in oceanic regions.