Use of Satellite-derived Rainfall for Improving Tropical Forecasts

Abstract

Sparsity of conventional data over tropical oceans makes it difficult to analyze well the moisture and divergence fields, and therefore the diabatic forcing of the tropical atmosphere is not well predicted in numerical models. A nudging procedure to improve the precipitation forecast in the National Meteorological Center (NMC) Medium Range Forecast Model (MRF) is developed. The convective parameterization scheme is modified to adjust the predicted rainfall amounts toward the observations in this method. In the absence of conventional data, the rainfall estimates from the satellite measures of the outward-going longwave radiation are utilized as the observed precipitation. Several forecasts from the MRF are presented to show the improvements in intensity and location of the intertropical convergence zone and tropical disturbances with the application of the nudging procedure. Additionally, spurious cyclone and excessive rainfall that were predicted without this procedure either failed to form or their intensifies were considerably reduced. Results from incorporation of the modified convective scheme in the global data-assimilation system within the NMC forecast model are also discussed. The analysis, the subsequent 72-h forecast circulation, and the rainfall amounts are improved with the use of this scheme.