泾河流域景观指数的粒度效应分析  

Effects on Changing Grain Size of Landscape Indices in Jinghe Watershed

作　　者：杨丽[1] 甄霖[1] 谢高地[1] 陈操操[1] 

机构地区：[1]中国科学院地理科学与资源研究所

出　　处：《资源科学》2007年第2期183-187,共5页Resources Science

基　　金：国家重点基础研究发展规划项目:"西部典型区域森林植被对农业生态环境的调控机理"(编号:2002CB111506);国家自然科学基金项目:"北方小麦-玉米农田生态系统服务功能研究"(编号:30570303)

摘　　要：本文以泾河流域1∶10万比例尺下的景观类型图为研究对象,在前人的研究基础上选择了7个景观指数,并对其粒度效应进行了分析,结果表明,各景观指数均具有明显的粒度效应,不同景观指数随粒度增加表现出了不同的变化趋势,根据各景观指数的粒度效应关系,可以将上述7个景观指数分为四类。第一类指数随粒度的增加呈单调递减,具有比较明确的粒度效应关系,这类指数主要有平均形状指数和聚集度指数;第二类指数随粒度增加也呈下降趋势,但在下降的过程中出现了不明显的尺度转折点,这类指数主要有Simpson多样性指数和Shannon均匀度指数;第三类指数随粒度增加总体上呈上升趋势,并具有明显的尺度转折点,这类指数在本研究中只有最大斑块指数一种;第四类指数随着粒度增加,粒度效应关系比较复杂,粒度效应曲线近似于“n”型,并具有明显的尺度转折点,这类指数有斑块密度指数和面积加权平均分维数。结果还表明,对泾河流域1∶10万比例尺下的景观类型图进行景观指数计算的适宜粒度范围为30m^40m。Landscape indices are widely used in the study of landscape pattern change to describe its components and spatial distribution. And most of analyses on landscape pattern are based on grid data, which will lead to the grain-effect problem. That is, analysis results based on landscape indices will vary according to grain sizes. Although many researchers has studied the grain-effect of different landscape indices, but the study about how to choose the suitable grain for vector data at different scale is scarce. For this purpose, this paper present further study based on a study of analysis of landscape pattern change carried in Jinghe Watershed, which belongs to Yellow River basin of China. Here the landscape map of Jinghe watershed at 1 ： 100 000 scale are converted to grid data of 18 grain sizes from 30m to 200 m and then use Fragstats software to get seven landscape indices used in this paper to analysis Jinghe Watershed＇s landscape pattern changes with grain size increasing. And results show that all of landscape indices used in this paper have obvious grain-effect with grain size changing. According to the grain-effect relationship, seven indices are divided into four types. The mean shape index and contagion index are included in the first type. Two indices of the first type decrease rapidly with increasing grain size, showing an obvious scaling relation （correlation coefficient are 0.9855 and 0.9718 respectively）. Indices of the second type include Simpson diversity index and Shannon evenness index. These indices also decrease with increasing grain size, but they have uneonspieuous scale inflexions during decreasing （correlation coefficient are O. 9723 and 0.9633 respectively）. The third type indices increase with the changing grain size, and it has obvious scale inflexions. Only the largest patch index belongs to this type in our study （correlation coefficient is 0. 8406 ）. Indices of the fourth type include patch density index and area weighted mean fractal dimension. Indices of the fourth t

关 键 词：泾河流域 景观指数 粒度效应 密度指数 

分 类 号：P901[天文地球—自然地理学]