Ali mohammad Khorshiddoust; Ali asghar Shirzad
Volume 18, Issue 49 , November 2014, , Pages 101-118
Abstract
In this research we used multivariable statistical methods (cluster and discriminative analysis) with the purpose of the recognition of spatio-temporal differences of precipitation in similar areas. We used monthly precipitation of 35 synoptic, climatic, and rain-gauging station data records of Northern ...
Read More
In this research we used multivariable statistical methods (cluster and discriminative analysis) with the purpose of the recognition of spatio-temporal differences of precipitation in similar areas. We used monthly precipitation of 35 synoptic, climatic, and rain-gauging station data records of Northern Iran including three provinces of Golestan, Guilan, and Mazandaran for 1995-2007 periods. For grouping and homogenizing the stations, we initially applied Ward cluster analysis method. Then we used discriminative analysis and Wilk’s Lambda for finding out the validity of cluster analysis calculations. Results obtained from cluster analysis with Euclid interval method indicated that 4 major clusters can be drawn according to the amount and the location of the precipitation in the study area. Discriminate analysis showed that 82.3% of the clusters in our analysis were valid and about 17.7% were incorrect. The Wilk’s Lambda method also proved the differences between the means.
