Abstract
Flood prediction is one of the basic soft countermeasures for safe evacuation and disaster risk reduction of water-related disasters. The current flood forecast predicts how river water levels will change in the next few hours at river sections where water levels are observed. In recent years, flood predictions with a distributed model have been also practically applied due to detailed topographical data and quantitative radar rainfall information. However, the following two problems still remain in the current flood forecasting. One is that river sections other than observation points are out of prediction target, and the other problem is that the situation of flood inundation caused by overtopping or torrential rain is out of the scope of real-time flood predictions. In this study, we report on the recent development of flood predictions including river flow discharge, water level and flood inundation even along small rivers in river basins. Specifically, we will report the results carried out for the Chikusa River basin (covered area 777.6 km2) of Hyogo prefecture and describe the future prospects including the application of the model at the national level.
