基于遥感与GIS中国陆地景观结构分析  

Metrics of terrestrial landscape structure in China based on remote sensing and GIS.

作　　者：王新明[1] 占玉林[1] 王长耀[1] 牛铮[1] 

机构地区：[1]中国科学院遥感应用研究所遥感科学国家重点实验室,北京100101

出　　处：《浙江大学学报（农业与生命科学版）》2006年第4期460-466,共7页Journal of Zhejiang University:Agriculture and Life Sciences

基　　金：国家自然科学基金资助项目(40571117);中国科学院知识创新工程重要方向项目(KZCX3-SW-338)资助

摘　　要：利用遥感和GIS技术,首先根据生态区划计算出描述中国陆地景观结构的十三种景观结构指数,包括斑块密度、最大斑块指数、边界密度、面积加权斑块形状指数、面积加权斑块分维度、平均最近斑块距离、最近斑块距离的标准差、最近斑块距离的相异系数、蔓延度、分布邻接指数、结合度、香农多样性指数、辛普森多样性指数.在相关分析的基础上,使用等级聚合与因子分析方法对最近斑块距离相异系数之外的十二种指数进行分析,最后采用因子分析结果中的前三个公因子分析中国陆地景观结构,三个公因子的方差累积贡献率达88.8%,分别定义为斑块多样性因子、斑块形状复杂性因子、斑块分布均匀性因子.结果表明,在人类活动影响强烈的生态区,这三种公因子取值都比较小,表明在人类活动的影响下,斑块类型多样性减少、斑块形状复杂性降低、斑块分布趋于均匀.This paper illustrates an example of describing the terrestrial landscape structure in China by means of remote sensing and GIS. According to the ecological regions of China, 13 most-commonly-used landscape structure metrics were calculated from the MODIS land cover / land use product provided by NASA/EOS. Since there being much redundant information between these metrics, the methods of correlation analysis, hierarchical clustering and factor analysis were used to avoid the redundancy. The first three factors from the factor analysis were related to patch diversity, complexity of patch shape, evenness of distribution respectively and were adopted to analyze the terrestrial landscape structure in China. At last it comes to the conclusion that the ecological regions more affected by mankind show less values for the three factors, meaning fewer patch attribute types, simpler patch shapes and patches being more evenly distributed.

关 键 词：遥感 GIS 景观生态学 等级聚合分析 因子分析 

分 类 号：S127[农业科学—农业基础科学] TP722.4[自动化与计算机技术—检测技术与自动化装置]