Discuss on reference crop evapotranspiration in arid zone based on principal component analysis

Considering that significant correlation exists among weather factors influencing crop evaportranspiration, if using all the weather factors to found crop evaportranspiration model, there will be all information overlapping. Principle component analysis method could solve this kind of problem. The result of PCA of weather factors influencing crop evaportranspiration in Aksu region shows that: the first four principle´s accumulative contribution rate has reached 94.295%, it could meet the demand. Applying this four principal components to build regression equation and compared with multiple linear regression equation which using all the weather factors as variables, it is found that: all the regression coefficients pass t test with significant correlation. Multiple linear regression equation, the fitting is a bit better than the principal component analysis method, but regression coefficient b₀, b₂, b₃ and b₄ don´t pass t test. The coefficient significance level is not as high as principle component analysis method, and b₄ don´t pass t test. The coefficient significance level is not as high as PCA method.