河南省经济发展的空间相关性及尺度效应

Spatial Autocorrelations and Scale Effects of Economy Development in Henan,China

采用河南省126个县级空间单元、18个市级空间单元,利用改革开放以来8个年份的人均GDP数据,基于ArcGIS和GeoDa软件9种空间权重矩阵构建方式,研究了河南省经济发展全局Moran′s I指数的时空变化,以及县级和市级空间相关性的尺度效应。结果表明:1)河南省县级、市级经济发展具有显著的空间相关性,尤其是在1990年之后,县级、市级Moran′s I指数分别为0.5、0.4左右。2)各种空间权重矩阵构建方式下得到的县级、市级Moran′s I指数,随时间的总体变化趋势基本相同。另外,除了ArcGIS距离平方倒数方式得到的值均过小外,其他权重矩阵方式得到的县级、市级Moran′s I值差异相对较小。欧氏距离权重矩阵更适合于描述经济发展的空间相关性。3)空间相关性的尺度效应似乎有一个临界值,当超过该临界值以后,小尺度(县级)比大尺度(市级)的Moran′s I值大,之前则相反。所以,以县级尺度为基础对河南省经济发展的空间战略进行布局更好,能较好地描述经济发展的空间溢出效应。

Based on 8-year per capita GDP of 126 county level and 18 city level units of Henan Province,China,the global Moran′s Iis calculated using GeoDa and ArcGIS with 9different spatial weight matrix,the spatial correlation and its spatialtemporal change of economic development of Henan Province are studied,and the scale effects of Moran′s Ibetween county scale and city scale are explored.The results show that：1）The county scale and city scale economic development all present significant spatial correlation,especially after 1990,the Moran′s Iof county scale and city scale of Henan is about 0.5and 0.4,respectively.2）The time trends of all Moran′s Icalculated by nine spatial weight matrices are the same.In addition,except the inverse distance square spatial weight matrix of ArcGIS,all other Moran′s I of county scale and city scale are similar,and the Moran′s I with Euclidean distance spatial weight matrix is the biggest.It looks like the Euclidean distance spatial weights matrix construction method is more suitable for describing the spatial autocorrelation of economic development.3）It seems that there is a threshold for the scale effect of spatial autocorrelation.After the value of spatial correlation arrives at the threshold,the Moran′s I of smaller scale（i.e.,county level）are bigger than larger scale（i.e.,city level）,and before it arrives at the threshold,the Moran′s I of larger scale are bigger than smaller scale.So,when government of Henan Province makes spatial economic planning strategically,it is better that based on the county scale than on the city scale,because it can reflect the spatial spillover effect of economic development more quickly.