climatic regionalization of caspian region using multivariate statistical methods

the weaknesses of the traditional climatic classification methods in presenting the climatic realities of caspian region and showing the new methods' merits based on multivariate statistical methods, and geographical information system in presenting the climatic regions of caspian, are among the most important ideas of conducting this study. to do so, we used the data of synoptic stations of north and north east of the country in a period of 20 years. to analyze 19 climatic elements in “factor analysis” method at first the map of each studied climate of the region was traced by “surfer software” on the bases of pixel in a 15*15 km measure. and also by “minimum curvature” method, data of each map become interpolated. after that, data of each map were changed to digital data, which a matrix in a 266*19 measure was formed of those digital data. by analyzing the matrix in “factor analysis” method, it was certificated that 98% data variance of it is of the precipitation and temperature that data consist: 50% of the primary factor and 48% of the secondary factor. quantities of precipitation and temperature were analyzed by “agglomerative hierarchical cluster analyzes” in “ward” method that the caspian region was divided to four regions in the aftermath moderate and humid, moderate and semi humid, cold and low precipitation, and semi cold and low precipitation. these minor climatic regions are greatly influenced by some climatic factors, as the remoteness and farness from the sea and the elevation increase. climatic varieties in the eastern parts, especially in the limits of golestan province, are more evident.