Regionalizing precipitation in Iran using GPCC gridded data via multivariate analysis and L-moment methods

Abstract

This study presents a spatial clustering of precipitation in Iran based on gridded data from the Precipitation Climatology Centre from 1951 to 2007. After standardizing the data, two multivariate methods of factor analysis and cluster analysis were used to derive primary homogeneous regions. Then, two statistical tests based on the concepts of L-moments were applied to examine the homogeneity of clusters and the discordancy of their member(s). Next, the geographical similarities of different parts of Iran were considered to move heterogeneous points to other clusters. These two steps were repeated to derive clusters that passed both tests. Finally, for each final cluster (region), the best probability distribution function was determined for frequency analysis. The results showed that, in terms of precipitation, Iran can be categorized into eight main probabilistic homogenous regions, and the Wakeby probability distribution function was chosen for half of them.