Rainfall Measurement Accuracy in FACE: A Comparison of Gage and Radar Rainfalls

Abstract

The agreement of radar with raingage rainfall measurements during the second phase of the Florida Area Cumulus Experiment (FACE-2) is examined. FACE-2 rainfall was measured in a 1.3 ? 104 km2 target area using 111 nearly uniformly distributed gages at an average density of 117 km2 per gage and using a WSR-57 radar adjusted daily with a dense raingage network distributed over 500 km2. The radar versus gage agreement is studied in order to evaluate the accuracy of the unadjusted radar measurements and the effectiveness of the adjustment technique in improving the radar measurements. Implicit in the comparison is an assumption that the gage rainfalls are the relative standard of accuracy. Before gage adjustment of the radar rainfalls, mean differences between radar and gage target area rainfalls are slightly positive (radar ? 1.10 ? gage) on dry days but become considerably negative (radar ? 0.70 ? gage) on wet days. Following gage adjustment, the mean agreement generally is much improved. However, the day-to-day variation of the differences is not diminished after the adjustment. This is attributed to misrepresentations of the gage-to-radar rainfall ratio in the dense raingage network which are caused by the spatial variation of that ratio within the overall target area.