Estimation of Seasonal and Annual Precipitation Using Geostatistical Methods (Case Study: Southern Alborz of Semnan Province)

To perform regional analysis and rainfall estimation of the rain gauges of the river basin code of 21 located in Semnan province, multivariate geostatistical approaches were used. Elevation was introduced as a covariant and effective factor in rainfall. Thirty eight rain gauges were selected for a 30 year period. The geostatistical analysis extension module of ArcGIS 9.3 was used for analysis and development of seasonal and annual prediction map. GS+ was used for understanding the details of variogram models and statistical fitting parameters. Average and reduced mean square errors of cross validation indices were considered for prediction of rainfall. The results showed that Cokriging methods with the assumption of elevation as a covariant factor and variance reduction gives accurate rainfall estimation which is mainly due to the high correlation between elevation and rainfall. Although the seasonal correlation is not stable, the value of correlation is not less than 0.6. Therefore, as the coefficient of correlation increases the accuracy of estimation will be enhanced.