NEW HYDROLOGICAL FREQUENCY ANALYSIS INTRODUCING CONFIDENCE INTERVAL OF PROBABILITY DISTRIBUTION MODELS BASED ON PROBABILITY LIMIT METHOD TEST

Abstract

 The general purpose of hydrological statistics is to decide probable hydrological variables corresponding to plan scale by using the most fitted probability distiribution for extreme hydrological data. Conventional hydrological statistics focus mainly on the estimation of the probability distribution that fits extreme hydrological data. However, observation data of unprecedented heavy rain deviate greately from the adopted probability distribution. Therefore, the method for evaluating such heavy rainfall has not been established in hydrological statistics. In this research, we aim to show a theoretical framework to evaluate unprecedented heavy rainfall by introducing confidence interval of probability distribution models.