Effects of radar and raingage representations of precipitation on the flood modeling of the remnants of Tropical Storm Henri in the Red Clay Creek Watershed
Date
2012
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
University of Delaware
Abstract
Tropical Storm Henri produced flooding in the Red Clay Creek Watershed that surpassed the 500-year flood event. For this research, the effect of different precipitation data sets on the modeling of the flood hydrographs was investigated. The modeling was performed with the United States Army Corps of Engineers HEC-HMS 3.5 modeling package using both raingage data as well as radar data. Raingage data were acquired from the National Weather Service's (NWS) First-Order gage network. The radar data from the Fort Dix (NJ) radar was converted to precipitation using the National Weather Service's Convective and Tropical Z-R relationships as well as a calibrated Z-R relationship was derived from radar-gage pairs from the convective radar scans. Point observations also were extracted from the radar data at the density of the Delaware Environmental Observing System gage network. Differences in the simulated hydrographs range from the sparse NWS First-Order network, with peak flows underestimated by 70% of the observed flow and timing of the peak lagging the observed by over three hours to the NWS radar data using the Tropical Z-R relationship where peak flow was simulated to within 1% of the observed flow with the timing of the peak matching exactly. Differences between the simulations and the observed values for both Wooddale, DE and Kennett Square, PA were also both compared. While the simulation based off of the Tropical Z-R radar data produced an exaggerated peak value at Kennett Square, the simulation produced a near exact hydrograph for Wooddale. At both Kennett Square and Wooddale for both data sets based off of the Tropical Z-R data, the average differences between the observed and simulated flow were the lowest. From these simulated hydrographs, the simulations using the Tropical Z-R data of which there are both radar and extracted gage values, produced results comparable to the observed data.