BEAMFILLING ERRORS IN PASSIVE MICROWAVE RAINFALL RETRIEVALS

There are currently large numbers of rainfall retrieval algorithms bas ed upon passive microwave radiances. Most of these algorithms are phys ically based in that they use explicit physical assumptions to derive relationships between brightness temperatures (Tb's) and rainfall. If these assumptions involve observable quantities, then the physical bas is of the algorithms can be extended to determine fundamental uncertai nties in the retrieved precipitation fields. In this paper this proces s begins by examining the largest uncertainty in many of the physical models-the homogenous rainfall assumption. Four months of Tropical Oce ans Global Atmosphere Coupled Ocean-Atmosphere Response Experiment shi pborne radar data is used to describe the horizontal characteristics o f win. The vertical hydrometeor structures needed to simulate the upwe lling Tb are taken from a dynamic al cloud model. Radiative transfer c omputations were performed using a fully three-dimensional Monte Carlo solution in order to test all aspects of the beamfilling problem. Res ults show that biases as well as random errors depend upon the assumed vertical structure of hydrometeors, the manner in which inhomogeneity is modeled in the retrieval, and the manner in which the radiative tr ansfer problem is handled. Unlike previous works. the goal of this pap er is not to determine a mean beamfilling correction or a vertical hyd rometeor profile that should be applied to specific retrieval algorith ms. Rather, it is to explore the impact of inhomogeneous rainfall upon the predicted brightness temperatures so that these relations may eve ntually be used to develop a physically based error model for microwav e precipitation retrievals. Because the predicted Tb's depend upon ass umed cloud vertical structures, the paper offers a procedure to accoun t for the uncertainty introduced by rainfall inhomogeneity rather than a general result. The impact of inhomogeneous rainfall upon specific algorithms must still be investigated within the context of that speci fic algorithm.