نوع مقاله : ترویجی
نویسندگان
1 استاد دانشگاه تبریز
2 دانشجوی دکترای دانشگاه تبریز
چکیده
کلیدواژهها
موضوعات
عنوان مقاله [English]
نویسندگان [English]
Due to the complexity of the patterns and mechanisms of precipitation and inappropriate distribution of measuring stations in mountains areas, offering a good model is required for each area. In this study, Data of annual rainfall of 218 synoptic and climatology stations and rain gauge (2000-2005) used that prediction and modeling spatial variation of precipitation was done using multiple regression and Geostatistical analysis in the western slopes of central Zagros Mountain. 20 statin were used to test and evaluation of methods. In order to evaluating the performance of each the models and determining the accuracy of theirs, two statistical indices (MAE and RMSE) were used. The obtained results from evaluation error criteria showed that the universal kriging with the lowest error rate was introduced as the best modeling method than other methods of Geostatistical and regression. Although Geostatistical methods obtained better results than the multiple regression model, therefore the multiple regression method estimated the dependent variable according to the correlation of the independent variables, can obtain better results for modeling higher levels that are lack of measurement stations.
کلیدواژهها [English]