Doppler Radar-Derived Rainfall Data Monitoring to Support Surface Water Modeling of TMDL

A Calibrated High-Resolution Precipitation Database (CHPD) represents an enhanced tool for modelers developing Total Maximum Daily Loads (TMDL) or using water resource related surface water (SW), ground water (GW) or integrated SW-GW models. With software support from the National Oceanic and Atmospheric Administration (NOAA), funding support by the Florida Department of Environmental Protection (FDEP) and the Environmental Protection Agency, and academic support from Florida State University (FSU), Florida is the first State to have successfully completed the construction of a practical CHPD. Precipitation input for a watershed model typically has relied on weather data obtained from rain gauges that are sparsely distributed in a watershed. A popular method for computing precipitation is the Thiessen method. This method assigns an area called a Thiessen polygon around each gauge. The polygon is an area in which every interior point is closer to the particular gauge than to any other. In effect, the precipitation in the entire polygon is assumed to be uniform and equal to the gauge value. This method does not provide information about the rainfall distribution between gauges. In addition, the method cannot give the true rainfall distribution for isolated or fast moving storms that may cover one side of a street, but not the other, and change intensity and coverage along its passage. Hourly Doppler radar-derived rainfall (NEXRAD/WSR-88) estimates possess superior temporal and spatial resolution compared to gauges and can be applied to watershed modeling as part of TMDL projects. The Doppler-derived precipitation estimates are produced hourly on a 4-kilometer by 4-kilometer grid covering the nation. Although hourly radar-derived rainfall data provide excellent spatial and temporal coverage, gauges provide more accurate point values if the data are carefully quality controlled. Thus, it is appropriate to optimally combine the two data sources so that the strengths of each are realized. This optimal combination has been completed to produce a historical precipitation database for Florida. Correlation coefficients have been calculated between the resulting hourly optimally combined radar estimates versus coincident independent hourly gauge data. The results generally arc quite good, but indicate that errors may be greater during warm season high precipitation scenarios. This aspect has been corrected in the CHPD based on research at FSU. This paper introduces the eleven years of historical precipitation data in CHPD released through FDEP's computer world-wide-web server and the CHPD applications by using three watershed models - WAMView, Mike-SHE, and WASH. The WAMView application assesses the potential benefits of CHPD data in the Black Creek Basin, which is a small (1,253 Km(2)) watershed in North Florida. The Mike-SHE application investigates the comparative diagnostic advantages of using fully distributed CHPD- or Thiessen (gauge)-derived rainfall rates in the Black Creek Basin of modest topography during both convective and synoptic conditions. The WASH application is the FDEP's first TMDL modeling project using the CHPD data.