Real-Time Correction of Spatially Nonuniform Bias in Radar Rainfall Data Using Rain Gauge Measurements

Abstract

A procedure for real-time correction of spatially nonuniform bias in radar rainfall data using rain gauge measurements is described. Developed to complement the existing gauge-based bias correction procedures used in the National Weather Service (NWS), the proposed procedure is a generalized local bias estimator that may be used under varying conditions of rain gauge network density and types of rainfall. To arrive at the procedure, the correction problem is formulated as a space?time estimation of radar and bin-averaged gauge rainfall from radar rainfall data and rain gauge measurements, respectively, at all hours up to and including the current hour. The estimation problem is then solved suboptimally via a variant of exponential smoothing. To evaluate the procedure, parameter estimation and true validation were performed using hourly radar-rainfall and rain gauge data from the Arkansas?Red Basin River Forecast Center (ABRFC) area. The results indicate that the proposed procedure is generally superior to mean field bias correction, and that the improvement is particularly significant in the cool season.