Accessibility and Utilization of WSR-88D Radar Precipitation Data for Natural Resource Modeling Applications

The National Weather Service (NWS) operates approximately 160 WSR-88D radar-precipitation stations as part of a Next Generation Radar (NEXRAD) program that began implementation in 1992. Among other products, these radar sites provide spatial rainfall estimates, at approximately 4 km2 resolution (Stage 1, Level 3 data), with nominal coverage of 96% of the coterminous United States. Effective coverage is much less than this in a given radar domain depending upon storm type and topography. As the original intent of this network was to support operational objectives of the Departments of Defense, Transportation and Commerce, the production of these data have been optimized for detection and mitigation of severe weather events that might result in flooding, destruction of property and loss of life. The primary hydrologic application has been river and flood forecast modeling by 13 NWS River Forecast Centers (RFC). As each RFC is responsible for a large river drainage, data processing and quality control of these data are geared toward optimization over a relatively large spatial domain (>100,000 km2). Use of these data for other hydrologic and natural resource applications is hampered by a lack of tools for data access and manipulation. NWRC has modified decoding and geo-referencing programs to facilitate utilization of these data for other research and management applications. Stage 1, Level 3 Digital Precipitation Array (DPA) files were obtained for the Boise, Idaho radar location (CBX) for the period of January 1998 to December 2000. Nine rain-gauge locations in the Reynolds Creek Experimental Watershed and Snake River Birds of Prey National Conservation Area, south of Boise, were georeferenced relative to the CBX Hydrologic Rainfall Analysis Project (HRAP) grid. NEXRAD estimates of total cumulative rainfall at these sites averaged only 20% of that measured by the local gauge network. This underestimate was attributed in the most part to truncation of low intensity rainfall events by the precipitation detection function (pdf) rather than to mis-calibration of the ZR relationship. At this time, these data are unsuitable as inputs for long-term water balance modeling but may be useful in extreme-event or flood-modeling applications. New tools to extract and manipulate specific subsets of Stage 1, Level 2 radar data may improve our ability to use radar reflectance data for a broader number of applications than are currently supported.