Hierarchical Cluster Approach for Regionalization of Peninsular Malaysia based on the Precipitation Amount

Abstract

The rainfall series from 59 homogeneous stations in Peninsular Malaysia are used in this study. The series are indicated to be multifarious, with unpredictable fluctuations from year to year and from region to region. Therefore it is more appropriate to observe all stations on spatial scales corresponding to clusters instead of the individual series. Single site occurrence analysis is inadequate for regional planning. Hence, it is important to determine regional rainfall frequencies. Hierarchical cluster analysis with 11 different similarity measures and 7 different linkagesare used in this study to find homogeneous rainfall groups. Validity indices are evaluated to determine the best number of clusters in a rainfall data set. The study elucidatesthree homogeneous rainfall regionsnamely A, B and C within Peninsular Malaysia, which correspond to areas with different natural features and rain pattern. Based on this research, a hierarchical clustering method is suggested in order to group individual rainfall series into a limited set of spatial clusters.