Assessment of Estimation Bias around Rainfall Gage Sites for Five Years of WSR-88D Data over the Mississippi River Basin

摘要

Statistical analyses were performed on five years of WSR-88D weather radar-rainfall estimates that were produced for the Global-Continental scale International Project (GCIP) over the entire Mississippi River basin covering the period 1996-2000. The GCIP data were adjusted using an altered Z-R relation to improve performance at the time scale of individual storms. The accuracy of radar-rainfall estimates were analyzed considering a number of factors including: number of overlapping radar(s), distance from nearest radar to gage, elevation of the gage, and the geographic location of the radar. Additionally, estimation of the bias between radar and gage-rainfall in areas without gages was assessed. Results show that the dominant factor affecting the accuracy of radar rainfall estimates is the number of overlapping radars, and the estimation of the bias in radar pixels without gage data can be approximated with a non-overlapping gage with varying levels of accuracy.