Abstract

Principal Component Analysis (PCA) is an optimum mathematical method to decrease variables into some limited components in order to justify the highest variance of primary variables. In this study some statistical characteristics of annual precipitation of Zanjan city including sum of annual precipitation, number of rainy days, extreme daily precipitation in a year, the ratio of extreme precipitation to the sum of annual precipitation and some characteristics such as Standard Deviation (SD), Skewness (Sk), Kurtosis (Ku), Absolute Mean Deviation (AMD) and Mean Absolute Interannual Variability (MAIV) were was calculated from monthly precipitation for each year, and were introduced principal component analysis technique. The results show that 95% percent of annual precipitation variations can be explained through 4 components. The first component which indicates the highest data variance (42.6%), represents annual precipitation and absolute variability indices including SD, AMD and MAIV. The second component represents the shape of frequency distribution indices (Sk, Ku), the third component represents extreme precipitations and finally the fourth component represents the number of rainy days. The analysis of the trend of components scores show that first and fourth components scores have a significant decreasing and increasing trend, respectively. Round a lines show a precipitation decrease during the period under study from one hand and having uniform temporal distribution on the other hand.