J1.3 Evaluation of Urban Hydrologic Prediction Accuracy for Real-time Forecasting Using Radar

Radar measurement characteristics have important consequences on the predictability of flood hazards in urban watersheds. Site-specific hydrologic prediction is made possible by highresolution geospatial data and physics-based distributed modeling. Within the urban context, basin response is typically less than one-hour and basin sizes less than several hundred square kilometers. Forecast locations within such basins could be as small as 10 km and grid cell resolutions on the order of 100 meters. Two main issues affecting predictability at these time and space scales are accuracy of quantitative precipitation inputs and the hydraulic characteristics used by the model simulate runoff processes. Operational system simulation experiments (OSSEs) help identify performance characteristics of the distributed flood forecasting system. Through event reconstruction, existing system performance is assessed using input from the NWS WSR-88D (NEXRAD) to create model forecasts for past events. A new network of radars is being developed as part of an Engineering Research Center testbed, called TEXRAD. Considerable insight into system improvements, adaptations, and applicable watershed scales is expected from the OSSE. Once the new radar network is in place, comparisons of flood forecast accuracy can be made between existing and developmental systems.