Research on urban flash floods risk warning based on hydrological model simulation*
The urban hydrological model is introduced into the urban flash floods forecasts and warning issue. The land use and land cover (LULC) types data and gridded urban drainage network capacity being parameterized in the model, and forced by the radar rainfall estimates (QPEs and QPFs), the model simulated the urban surface hydrological response and hydraulic processes. The hydraulic model is based on the shallow water equation, which the Alternating-Direction-Implicit (ADI) is used to solve the differential equation by tow steps in the x direction and y direction, respectively. This solution comprises the back water effects in simulation and indirectly intimates the multiple flow direction methods, which recalls the surface water dispersion in turbulence or diffusional effects. The case study demonstrates the online and offline running of the urban hydrological model purposed for the flash floods warning, partially for the model validation. The online hydrological model took the case study on the “7.21” thunderstorm in the year of 2012 occurred in Beijing, in which the radar quantitative-rainfall estimates (QPEs) were forced on the model for reproducing the gridded inundation mappings. The model simulation results resembled the flash floods scenarios of the waterloggings and water inundations on the “7.21” thunderstorm day. The offline model simulation addressed the measuring the rainfall intensity threshold for the ranked risks of the storm producing flash floods, especially the rainfall intensity thresholds for floods-susceptible places (FSPs). Therefore, the hydrological model simulation deduced the 1h, 3h and 6h cumulative rainfalls thresholds inducing water inundation depths over 0.2m, 0.5m, 0.8m and 1.2m in more than 49 FSPs, which had been the rainfall intensity thresholds of the flash floods warning in the blue, yellow, orange and red signals, respectively. The rainfall intensity thresholds in risk warning ranks can be referred by the forecasters for flexible flash floods warning issuances in convenience. The research initiates a possible service mode for the effectiveness improvement on the urban flash floods warning and forecasts, efficiently combining the technologies of the Auto-nowcasting system and the hydrological model, especially forming the backbones for the convenient warning and forecasting services operations, easiness in data manipulation and rational computation resources allocation.