Rainfall Estimation from Polarimetric Radar Measurements: Composite Algorithms Immune to Variability in Raindrop Shape–Size Relation

Abstract

Polarization diversity radar measurements such as reflectivity factor, differential reflectivity, and differential propagation phase are extensively used in rainfall estimation. Algorithms to estimate rainfall from polarimetric radar measurements are based on a model for the raindrop shape as a function of drop diameter. Most of the current algorithms use an equilibrium shape–size model for raindrops. Variation of the prevailing mean raindrop shapes from an assumed model has a direct impact on the accuracy of radar rainfall estimates. This paper develops composite algorithms to estimate rainfall from polarimetric radar data without an a priori assumption about the specific form of mean raindrop shape–size model such as equilibrium shape model. The accuracy of rainfall estimates is evaluated in the presence of random measurement errors as well as systematic bias errors. The composite algorithms, independent of a prespecified raindrop shape model, were applied to radar parameters simulated from disdrometer data collected over 3 months, and the corresponding rainfall estimates were found to be in good agreement with disdrometer estimates. The composite algorithms were also tested with Colorado State University CHILL radar observations of the 28 July 1997 Fort Collins (Colorado) flood event. The storm total precipitation estimates based on the composite algorithms developed in this paper were in much better agreement with rain gauge estimates in comparison with conventional algorithms.