团雾成因的GWR和MGWR建模及其差异性分析

GWR and MGWR modeling of the cause of mass fog and their difference

选取山东省团雾多发点周围半径4 km区域内的地势高程差、河流长度以及高速公路长度等自然条件和地理条件为解释变量,选取2018年山东省团雾多发点团雾发生次数为因变量,利用GWR和MGWR模型研究山东省内高速公路团雾产生可能的诱因,分析变量的空间异质性并对比两种模型的差异。研究结果表明:团雾是由多个诱因共同作用产生的,且不同解释变量拥有不同的贡献度和作用尺度;GWR默认所有的解释变量具有相同的空间尺度,而MGWR能够将不同解释变量对于因变量的影响尺度反映出来,其回归的结果更为可靠。

This study selected terrain elevation difference,river length and the length of the highway and other natural conditions and geographical conditions of 4 km radius area of frequent mass fog sites in Shandong Province as the explained variable,selected mass fog times in 2018 as the dependent variable,and studied the cause for mass fog using the GWR and MGWR modes.The spatial heterogeneity of variables was analyzed and the differences between the two models were compared.It is concluded fog is caused by multiple inducements,and different explanatory variables have different contribution degrees and spatial scales.GWR assumes that all explanatory variables have the same spatial scale,while MGWR can reflect the impact scale of different explanatory variables on dependent variables,and its regression results are more reliable.