新疆地区降水量与海拔的空间变系数模型

Analysis of the spatially varying-coefficient regression model on precipitation and altitude in Xinjiang region

空间变系数回归模型是将变系数回归模型的回归系数转化为包含数据空间位置特征的回归系数函数的模型.以新疆地区降水量和海拔高度为指标,选取1991-2000年气象观测站的数据,运用空间变系数回归模型对新疆各区域的海拔高度和降水量之间的空间关系加以研究.结果表明:新疆各区域的海拔高度与降水量之间呈现显著的空间非平稳性特征,即海拔高度对降水量的影响随着研究区域地理位置的不同而不同,是地理位置的函数,并且得到了具体的影响系数函数.

Spatially variable coefficient regression model is the regression coefficient of the variable coefficient regression model into a regression coefficient function containing the spatial position of the data.Using the precipitation and altitude in Xinjiang as an indicator,the data of meteorological observation stations from 1991 to 2000 were selected and the spatial relationship between altitude and precipitation in different regions of Xinjiang was studied by using the variable coefficient regression model.The results show that there is a significant spatial non-stationarity between altitude and precipitation in Xinjiang.The influence of altitude on precipitation varies with the geographical location of the study area,which is a function of geographical location,and the specific coefficient of influence is obtained.