Effect of Radar-Rainfall Errors on Rainfall-Runoff Modeling

Abstract

Recent years have witnessed significant advances in development of operational radar-rainfall products. These products are desirable for several hydrologic applications such as flood forecasting and rainfall-runoff modeling. It is recognized that radar-rainfall estimates are associated with unknown uncertainties. The nature of these uncertainties and their impact on the prediction accuracy of hydrologic models is not fully understood. The present study presents an analysis of uncertainties of operational radar-rainfall products and how they propagate into rainfall-runoff models. The study uses NWS Multi-sensor Precipitation Estimator (MPE) radar-rainfall products over the Goodwin Creek experimental watershed. Surface rainfall observations from a dense rain gauge network in the watershed are used to analyze error characteristics of radar products. MPE radar data are used as input to a fully distributed hydrologic model (GSSHA) to simulate runoff response during 11 storms recorded in 2002. The study focuses on effect of three different radar error characteristics: systematic error (bias), random error, and temporal and spatial correlations of radar the error filed.