Rainfall Estimation by the WSR-88D for Heavy Rainfall Events

Abstract

Storms that produce extreme flooding present a special challenge for the WSR-88D rainfall algorithms. The authors assess the utility of weather radar in the investigation of extreme rain-producing storms through both climatological analyses of long-term radar datasets and case studies of storm events. Climatological analyses are presented for long records of WSR-88D volume scan reflectivity observations, for hourly radar rainfall accumulations products (WSR-88D and WSR-57D), and for radar?rain gauge intercomparisons. These analyses provide a context for interpreting case study assessments of WSR-88D rainfall estimates. Case studies are presented of five storms that produced extreme floods in the United States. Events include 1) the orographically enhanced Rapidan storm in the Blue Ridge region of Virginia, which resulted in more than 600 mm of rain during a 6-h period on 27 June 1995; 2) the southeast Texas storms of 16?17 October 1994 in which approximately 750 mm of rain fell during a 6-h time period; 3) the Dallas, Texas, hailstorm of 5 May 1995, which resulted in 16 flash flood deaths during a period of several hours and property damage exceeding $1 billion; 4) the Chicago, Illinois, storms of 17 July 1996 during which the 24-h rainfall record for Illinois was shattered; and 5) Hurricane Fran, which resulted in unprecedented flooding in North Carolina and Virginia during September of 1996. For each event, analyses revolve around volume scan WSR-88D reflectivity observations. The climatological analysis presented, in conjunction with the case studies analyzed, illustrate the significance of 1) brightband contamination, 2) tilt selection, 3) hail, 4) radar calibration, and 5) Z?R relationships for quantitative rainfall estimates by the WSR-88D.