长江流域土地利用结构及其空间自相关分析

LAND USE STRUCTURE AND ITS SPATIAL AUTOCORRELATION ANALYSIS IN THE YANGTZE RIVER BASIN

按照中国科学院资源环境分类系统,将WESTDC_Land_Cover_Products2.0数据集重分类为耕地、林地、草地、水域、城乡建设用地和其它未利用土地等6类,借助GIS平台研究了长江流域土地利用数量结构特点及其空间自相关特征。结果表明,长江流域的主要土地利用类型为林地、耕地和草地,三者占流域总面积的90%以上,而水域、城乡建设用地和其它未利用土地所占比例较小。各省市之间土地利用结构差异较大,与一定的空间格局相联系,并形成了土地利用结构特征不同的4个大的区域。各土地利用类型比重属性值相似的省市在长江流域全局空间内聚集分布、空间邻接,呈极显著的全局空间正自相关。其中耕地、林地、草地、城乡建设用地和其它未利用土地分布特征的空间正自相关极显著,水域分布特征的空间正自相关显著,但各种土地利用类型空间分布的重心、格局、空间自相关构成及回归机制不同。各种土地利用类型空间分布的局部空间自相关为高高(H-H)聚集或低低(L-L)聚集,高值为低值所包围或低值为高值所包围的空间孤立点(Spatial Outlier)不显著。研究表明各种土地利用类型在长江流域的空间分布具有明显的聚集性和区域特征。

Based on the WESTDC_Land_Cover_Products 2.0dataset which was reclassified as cropland,forestland,grassland,water,urban &rural built-up land and other unused land in terms of CAS resource and environment classification system,the land use quantitative structure and its spatial autocorrelation characteristic in the Yangtze River Basin were studied using GIS technique.The results showed that forestland,grassland and cropland were the dominating land use types which accounted for more than 90percent of the basin's total area,while the area of water,urban &rural built-up land and other unused land were extremely small.Different provinces and municipalities in the basin shared sharply different land use structures which could be divided into four zones in connection with a certain land use spatial pattern.The provinces and municipalities with similar occupancy rate concerning all kinds of land use types globally clustered and adjoined in the basin,spatially autocorrelating significantly.But distribution centroid,pattern,spatial autocorrelation structure and regression characteristic were different in different land use types.Local spatial autocorrelation for spatial distribution about all land use types in the basin resulted from H-H clustering or L-L clustering,whereas spatial outlier was negligible.As a whole,the research demonstrated that the spatial distribution of all land use types were distinctly of attributes of spatial clustering and regionality in the basin.