Abstract
Statistical estimation of present hydrological frequency analysis has significant uncertainty caused by the shortage of extreme hydrological data. We proposed a new theoretical framework of hydrological frequency analysis including confidence interval based on probability limit method test. By using this confidence interval, reliable statistical estimation can be possible under limited extreme hydrological data and we can evaluate hydrological phenonmenon which was considered unexpected conventionaly with reasonable return period. In this research, runoff analysis is performed at Kusaki dam basin by using rainfall time intensity extended to confidence limit value of probable total rainfall as input. Through this runoff analysis, the fluctuation characteristics of probable flood peak discharge is shown. Also, the relationship between shape of probable rainfall intensity and probable flood peak discharge is clarified.
