Spatial-temporal analysis of seasonal meteorological drought

Meteorological drought is one of the most important natural hazards which frequently occur in any climate. In this study, for regionalization of similar regions from drought situation point of view, a clustering approach was used. For this purpose, 25 years rainfall data of 120 synoptic meteorological stations across Iran were retrieved from IRIMO and examined by routine quality control tests. The drought severity was evaluated using standardized precipitation index (SPI) in a three months window during different seasons of the year. The calculated SPI values were interpolated in Arc GIS environment using Kriging method. The interpolated values formed a matrix consisting 16203 grid points during the 25 years study period. The obtained values were clustered by means of K-means method in different seasons of each year based on SPI values in different seasons. Results indicated that the optimum number of clusters are 8, 10, 9 and 6 for spring, summer, autumn and winter respectively. Besides it was concluded that the type of precipitation systems arriving in Iran during different seasons can affect the clusters direction and shape.