A large landslide caused by Typhoon 0514 formed a landslide dam on the Mimi River, Miyazaki Prefecture, Japan, during the night of September 6, 2005. Fortunately, the dam collapsed shortly afterwards and the resulting flood did not cause damage downstream because the flood discharge was small. We estimated the discharge rate downstream from the landslide dam using the discharge rate from hydroelectric dams and the time that the landslide dam collapsed based on the noise that local inhabitants heard. We also studied communications between disaster|prevention personnel regarding the landslide dam. Because they could not see the landslide dam as it formed during the night, and could not predict the flow when the dam burst, local inhabitants could not be suitably evacuated. In this study, we calculated the peak discharge rate when the landslide dam burst using a new simulation model. We found that the peak discharge rate in this model was greatly affected by the longitudinal form of a landslide dam, and had a tendency to increase when the longitudinal length was short. To predict the peak discharge rate rapidly when a landslide dam forms, a system that determines the longitudinal form of the landslide dam and the inflow rate is needed.
Key wordsFTyphoon 0514, landslide dam, peak discharge rate, communication