Study the Role of Outlying Data in Hydrological Frequency Analysis using L-moments in Halilroud Basin

The exact time of hydrological events occurrence is not predictable, but it is possible to investigate previous extreme events and estimate probability of future occurrences in different return periods. Having a long time series of hydrological variables is important parameter in frequency analysis and outlying data has an important role in quintiles prediction. In present study, we investigate the effect of outlying data in rainfall frequency analysis using L-moment method in Halilroud basin. First, all of recorded maximum annual rainfall was used in frequency analysis and then outlaying data were identified and removed form time series. In this stage, frequency analysis was done without these data. Results showed that outlying data don t have any effect on regional distribution function, but affect at-site distribution function and consequent estimated quintiles in different return periods.