Spatial Clustering of Iran Based on Occurrence of Long Term Meteorological Drought

Identifying of homogeneous regions from viewpoint of drought occurrence over time¡ provide understanding the extent of regional droughts in addition to recognition the impact of different effective precipitation systems in different regions. On the basis of this concept¡ the long term 25 years rainfall data belonging to 120 synoptic meteorological stations which distributed throughout Iran country are used in this study. Standardized precipitation index (SPI) is applied to calculate drought intensity¡ in 12¡ 24¡ 48 time months window and the Kriging method used for drought zoning. Spatial network with 10 kilometers pixel size is considered and interpolated SPIs values is extracted for this network using Arc GIS environment and used to prepare a matrix. The extracted data in matrix form clustered for in different scenarios i.e. annual¡ biennial and quadrennial¡ based on K-means method and optimal numbers of clusters calculated considering silhouette plot values. The results show Iran country classified to 12¡ 9 and clusters considering annual¡ biennial and quadrennial long term meteorological drought and prolonging drought time window caused reduction in numbers of clusters. Moreover¡ each of the three Northern provinces that located in southern Caspian Sea rim¡ classified in separate clusters and southeastern Iran that followed the monsoon precipitation regime¡ constitute a separate cluster too.