An Estimation-Based Precipitation Retrieval Algorithm for Attenuating Radars

Abstract

A new method for retrieving rainfall profiles from a spaceborne radar is introduced. As a result of the frequencies necessary in spaceborne radar applications, attenuation by both rainfall and liquid cloud particles is nonnegligible and must be accurately accounted for before quantitative rainfall estimates can be made. The proposed method is based on the minimization of a cost function that allows one to account for attenuation at each level directly in the iteration process. In addition, the algorithm does not invoke the Rayleigh approximation and is, therefore, applicable at wavelengths characteristic of spaceborne radars. The method is flexible with regard to the parameters to be retrieved and is well-suited for the addition of measurements from other sensors, such as a passive microwave radiometer, to constrain the retrieval. Preliminary results, using simplified assumptions of drop size distribution and particle shape, illustrate the utility of the algorithm provided the attenuation is not severe. At the frequency of the Tropical Rainfall Measuring Mission (TRMM) precipitation radar (14 GHz), synthetic retrievals are accurate to within 20% for rain rates up to 40 mm h?1. On the other hand, at 94 GHz, the frequency of the CloudSat cloud profiling radar, attenuation effects are too severe at rain rates greater than 1.5 mm h?1, suggesting the need for additional information to constrain the retrieval. Such information might come in the form of a path-integrated attenuation (PIA) derived from surface echo measurements or, alternatively, a precipitation water path (PWP) estimate from a passive microwave radiometer. Addition of a simple PWP constraint yields improvements in the retrieved rainfall profiles from both instruments when attenuation is severe. At 94 GHz, in particular, it is found that accurate quantitative rainfall estimates can be made provided the near-surface rain rate does not exceed 10 mm h?1.