Modeling of Extreme Rainfall Recurrence Using Conditional Probability Approach

Abstract

Forecasts for the occurrence of extreme rainfall are related to time lapse estimates until the next extreme rain event. The accuracy of these estimates depends on the choice of interevent time distributions of extreme rain events. The aim of this research is to estimate the probability of time recurrence of extreme rainfall. In this research, we observed the time between extreme rain events distributed by Weibull and Pareto. Let T be the time interval for the next extreme rain and it is assumed that there has been no extreme rain at time interval (0, t 0]. To analyze the occurrence of the next extreme rain event, the concept of conditional probability P(Δt 0|t 0) is used. P(Δt 0|t 0) is the probability of extreme rain occurrence during the time interval Δt 0. The results show that by using three models, we found that the next extreme rain event is not depending on the time interval Δt 0 since the last extreme rain event. Other results show that for case studies of extremerain events in South Sulawesi it was found that the time of occurrence of extreme rainfall was influenced by parameter value for each model.