High-Resolution Flash Flood Forecasting for the City of Fort Worth

Flash flooding is a serious problem in urban areas; the large percentage of precipitation running off due to greater impervious areas puts lives and properties at greater risks. DFW is the largest inland population center and one of the fastest growing urban areas in the U.S. To aid timely life-saving decision making in such an area, it is necessary to have a flash flood forecasting system that can produce water-level information at a very high resolution both in space and time. For this, accurate quantitative precipitation information (QPI) at high spatiotemporal resolution is a requisite. In this presentation, we describe a prototype flash flood forecasting system under development for the City of Fort Worth that utilizes very high-resolution radar data from a network of the CASA (Collaborative Adaptive Sensing of the Atmosphere) radars.

To take full advantage of the high-resolution QPI, it is necessary to operate hydrologic and hydraulic models at a commensurate scale. In this work, the radar rainfall data from the CASA radars are ingested into the National Weather Service (NWS) Hydrology Laboratory's Distributed Hydrologic Model (HL-RDHM) to provide a city-wide view of flooding threat at a very high resolution. Up to now, this model has been used operationally on a ~4 km grid forced by hourly rainfall. This is the first application where HL-RDHM is implemented at a spatiotemporal resolution of ~500 meters and less than 5 minutes. Since the impervious area plays an important role in rainfall-runoff modeling in urban areas, fractional impervious area maps are produced on grids of ~250 m, ~500 m and other resolutions. In this talk, we present the results of scale sensitivity analysis and comparisons with streamflow observations from 14 pressure transducer-based sensors located throughout the City.