Regional Frequency Analysis of Maximum Daily Rainfall at the Southern Coasts of Caspian Sea by Fuzzy Clustering

The regional frequency analysis is one of the methods that estimates maximum daily rainfall in regions where no data is available. Cluster analysis is used to identify hydrologically homogeneous regions in regional studies. In this study، fuzzy cluster analysis was applied to identify hydrologically homogeneous regions at the southern coast of Caspian Sea in the Golestan، Mazandaran and Gilan provinces. The homogeneity of each region identified by cluster analysis was checked by Hosking-Wallis heterogeneity measures. Then، suggestions were proposed for adjusting the homogeneity of the regions. The best regional statistical distribution was selected based on the goodness of fit test. The generalized logistic distribution was chosen as the best regional statistical distribution with different parameters for each homogenous region in this study. Finally، parameters of maximum daily rainfall were estimated by the L-moment method.