[A-HW20] Hydrology & Water Environment

[AHW20-05] A review of ensemble flood forecasting for operational warning

*Megumi Watanabe¹, Shunsuke Ito², Wenchao Ma¹, Dai Yamazaki¹ (1.Institute of Industrial Science, The University of Tokyo, Japan, 2.School of Engineering, The University of Tokyo, Japan)

Keywords:ensemble, flood, forecast, probabilistic, communication

There is growing international interest in the potential of ensemble flood forecast to improve operational warning. Ensemble flood forecasting systems promise two significant advantages over conventional deterministic forecasting techniques depending on water level observations of upstream rivers. First, ensemble flood forecasting systems have a long lead time for effective evacuation. Second, ensemble flood forecasting systems can also provide quantitative probability to estimate the uncertain forecasts. The introduction of an ensemble flood forecasting system has been under consideration in Japan. We need a comprehensive and integrated assessment based on multiple information from ensemble flood forecasting systems. This study surveyed ways to deliver detailed information in operational ensemble flood forecast systems outside of Japan. We summarized effective visualizations and improvements of the forecast systems by user feedbacks. We found that forecasts with multiple variables are delivered to public administration units with low spatial resolution while forecasts with fewer variables are delivered to each local point or each river segment with high spatial resolution. We also found that user feedbacks have improved the forecasts systems through discussions, although ensemble forecasts are complicated, and they have different characters from local forecasts. Communication with potential users concerning the forecasts is necessary for an effective operation.