Application of WRF for Uncertainty Analysis of Flood Inundation in DianBao River Basin

He-Yu Zhang

  This study was to determine the uncertainty of flood inundation based on hourly typhoon rainfall forecasting by the Weather Research and Forecasting (WRF) ensemble model. The WRF ensemble model consists of between 10 and 12 members of ensemble forecasting each typhoon events. Physiographic Drainage-Inundation model (PHD model) is used to assess the potential for skillful flood forecasting based on precipitation inputs derived from WRF ensemble. Each ensemble forecasting assumed that members of ensemble was normal distribution. Confidence interval of normal distribution was used for consider the uncertainty of flood inundation. The six typhoon events that occurred in the DianBao river basin were used for testing and evaluation. The results were presented in flood inundation map, and combine four maps with a Combine-Flood-Inundation Map (CFI Map). The CFI map includes that (1) Observation, (2) ensemble mean, (3) ensemble mean plus standard deviation, and (4) ensemble mean minus standard deviation of flood inundation. The results show a striking effect of precipitation inputs on flooding simulation. Overall, the 99.7% confidence interval of CFI map could be expressed uncertainty of flood inundation.

Key words: WRF ensemble model, Physiographic Drainage-Inundation model, uncertainty of flood inundation, Combine-Flood-Inundation Map