Using Statistical Model for Seasonal Rainfall Forecasting Based on Synoptic Patterns of Atmospheric Upper Levels

Statistical modeling has been used for seasonal rainfall forecasting based on synoptical patterns of the atmospheric upper levels in Khorasan province - northeast of Iran. The data of 37 rainfall stations were obtained from Iranian Meteorological Organization and the first stage was filling the gaps estimating and missing data using statistical methods. At the second stage, the RUN-TEST homogeneity procedure were done to find out if the rainfall data are randomly collected. Mean local time series of rainfall have been calculated by Arc GIS software. In order to forecast the seasonal rainfall in the period of Dec ember to May, the relations between rainfall and atmospheric upper level parameters at the difference time intervals were used as inputs of statistical model. Results show that the statistical modeling can successfully predict amount of the rainfall. Root mean square error obtained by stepwise and backward models were 50.4 and 47.3 millimeter respectively.