Recognition of Homogeneous Regions Using K-means Partitional Cluster Analysis Method Based on Second-order L-moment(Case Study: Roodbar Dam and Bakhtiari Dam Basins)

A group of sites with sufficient homogeneity in producing flood mechanisms form a homogeneous region for regional flood frequency analysis. Cluster analysis is the generic name of a variety of multivariate statistical procedures that are used to classify given data into similar groups or clusters. This procedure is an efficient method for regional flood frequency analysis. In this article, cluster analysis of annual maximum flow data of 34 gauging stations located in Lorestan province in Roodbar Dam and Bakhtiari Dam basins for 2 to 6 clusters has been performed using R programming language. K-means algorithm that is a type of partitional clustering has been used for cluster analysis. Merged groups of sites have been formed and homogeneous groups of Roodbar and Bakhtiari dam basins has then been identified by performing homogeneity test.