摘要
1研究方法区域化变量是普通随机变量在区域内确定位置上的特定取值,它是随机变量,是与位置有关的、特别是与距离有关的函数。区域化变量考虑系统属性在所有分离距离上任意两样本间的差异,并将此差异用其方差来表示,就是变异函数(也叫半方差函数)。由于景观格局具有强烈的空间特性,因此,可以用变异函数来衡量其在空间上的变化,其公式如下:r(^h)=2N1(h)Ni(=h1)[Z(xi)-Z(xi+h)]2(1)在式(1)中,h是两样本点的空间分隔距离,Z(xi)和Z(xi+h)分别是样本在空间位置xi和xi+h上的观测值(i=1,2……N(h)),N(h)是分隔距离为h时的样本总对数。
[Objective] To discuss the influence of the spatial heterogeneity of different green land types on their corresponding ecological efficiency and functions. [Method] ETM+ Images are utilized to classify urban green land into three categories, i.e. park green land, affiliated green land and other green lands; the principles of geostatistics are adopted to analyze their spatial heterogeneity. [Result] Spatial heterogeneity of different green land types differs greatly, which follows the scale law of 'affiliated green land < other green lands < park green land'. [Conclusion] The ecological functions and process of different green land types are different in scale; and the shapes of semi-variance functions are closely related to the spatial range, size and variety of ground objects, which however differ slightly in different directions, look similar or tend to isotropy.
基金
广东省科技厅项目(2005B70101080)~~