干旱区生态环境知识资源中心

To analyse the correlations among the five factors of temperature, humidity, precipitation, sunshine and wind speed in the desert, the meteorological data of Hangjin Banner for a total of 52 years from 1959 to 2010 are used as the experimental data. Through correlation analysis and regression analysis, a regression equation is established between any factor as the dependent variable and the other four factors as independent variables. The test results show that each coefficient in the equation passes the 95% significance test. Among them, the regression equation has the best fitting degree when humidity and temperature are used as dependent variables, which are 0.520 and 0.514, respectively. Using the data from 2011-2016 of Hangjin Banner to test the regression equations of humidity and temperature, it is found that the model has better prediction ability. Therefore, it is feasible to apply the regression equation to the analysis and prediction of desert meteorological data.