基于ESDA的京津冀地区草地变化空间分异

论文以县域为基本研究空间单元,借助GIS软件与Geoda空间统计分析软件,利用ESDA方法,对京津冀地区县域草地资源数量变化的总体与局域空间差异进行了分析。结果表明:①1980-2000年京津冀地区草地动态变化系数的Global Moran’s I为0.184 4,α = 0.05时呈显著的空间正相关,草地变化呈现出较明显的空间集聚特征;②时段Ⅰ(1980-1995年)到时段Ⅱ(1995-2000年)的全局Moran指数有所增加,表明草地变化在空间分布上集聚的趋势在增强;③各县域Local Moran’s I的Grid图反映了京津冀地区草地变化在局域空间上的均质性,尤其是高值区域与低值区域的集聚特征十分显著;④LISA集聚4种类型中,正相关的"H-H"类型区、"L-L"类型区占主体部分,二者约占总数的70%以上,其中,在时段Ⅱ,"H-H"关联的县域主要聚集在西北部山区,而中部的平原区域则表现出明显的"L-L"关联特征。

Spatial Difference of Regional Grass Changes based on ESDA at County Level in Beijing-Tianjin-Hebei Area

Grassland serves as the basic resources and condition for the survival of human.It is meaningful for us to protect grassland and restore the areas that are seriously damaged gradually as well as refund the natural ecological grassland.It can not only make the land ecological service get a valid guarantee,but also play an important role in the land ecosystem equilibrium and the formation of the regional pattern for ecological security.Strengthening the research on grassland change at the county level about its characteristics,rules,spatial patterns etc.,have important sense to guiding the protection of grassland at the county level and realizing the sustainable development of social economy.In this paper,based on global and local spatial autocorrelation analyses of exploratory spatial data,the spatial disparities about grassland change at the county level in Beijing-Tianjin-Hebei Area are discussed by using GIS and Geoda software.The conclusions are as follows: 1) During 1980-2000,the global spatial autocorrelation of forest land changes is significant.Global Moran’s I is the significant positive spatial correlation because it is 0.1844.The spatial clustering phenomenon about the changes of grassland in Beijing-Tianjin-Hebei Area appears on the whole.2) There is an obviously temporal increase of Moran’s I value from 1980-1995 to 1995-2000.That is,there was a dramatic increase about grassland change’s spatial clustering in Beijing-Tianjin-Hebei Area.3) The extent of grassland change is almost the same in some region by analyzing the grid figure of Local Moran’s I.Especially,the characteristic of spatial clustering about regional high value and low value is significant.4) The counties of the positive spatial correlation in local indicators of spatial association are in the majority.The regions with the "high-high" correlation are mainly located in the north-west hilly area during 1995-2000.However,the regions with the "low-low" correlation were distributed in middle area during 1995-2000.