Combined radar-radiometer rainfall retrieval for TRMM using structure function-based optimization.

摘要

The Tropical Rainfall Measuring Mission (TRMM) satellite, a joint US-Japanese mission to explore tropical rainfall and its effects on the earth's energy budget, general circulation, and climate, represents the first dual deployment of a precipitation radar and passive microwave radiometer on an earth-viewing satellite. While both radiometers and radars have been used independently to retrieve rainfall, previous attempts to combine the two types of measurements have mostly maintained independent treatment of the data, as these instruments are based on different physical principles.; A method has been developed which objectively combines the measurements within a unified radiative transfer equation (RTE) model, by treating the radar as a radiation source within the RTE model's source function. The retrieval method uses a physical inversion scheme which searches a cloud-radiation database of profiles, produced by a 3-D non-hydrostatic cloud model hurricane simulation, to find first guess hydrometeor profiles whose brightness temperatures (TBs) and reflectivities (Zs) best match the radiometer and radar measurements. The first guess profiles, expressed in terms of structure functions, are modified within a steepest descent optimization scheme, using the structure function coefficients as the optimization control variables. The solution hydrometeor profiles are then subjected to a set of gravity fallout equations to calculate the surface rainfall rates. Precipitation type is also diagnosed from an algorithm using TB thresholds and radar-derived bright band information.; Measurements from aircraft prototype instruments deployed for the TOGA-COARE field experiment are used in the evaluation of the algorithm. A statistical analysis of the results found that the optimization improved the agreement between the measured and first guess TBs and Zs by {dollar}{lcub}sim{rcub}{dollar}20%. Also the determination of the precipitation type was consistent with observations and theoretical considerations. The combination of TB and Z measurements reduce weaknesses inherent to single instrument retrievals, by reduction of the vertical non-uniqueness problems inherent to passive-only retrievals, and reduction of the sensitivity to microphysical assumptions inherent to radar-only retrievals, indicating the potential of combined retrieval, namely that it is greater than the sum of its parts.