景观指数的粒度变化效应

本文以延河流域的1:250000和1:500000土地利用图为对象,以景观格局分析程序Fragstats3.3为分析工具,探讨了不同比例尺条件下景观指数随粒度增加的变化特征.研究结果表明,随着粒度值由25m到400m的逐渐增加,除斑块丰富度外的景观指数均具有明显的尺度效应,其中聚集度和集合度没有尺度转折点,其他指数具有明显或不明显的尺度转折点.对比分析1:250000和1:500000土地利用格局指数的计算结果可以发现:1)尺度转折点不是一个值,而是一个相对较小的区间;2)尺度转折点与研究图件的比例尺有关,比例尺越大,所发生的第一次尺度转折点的粒度就越小;3)第一尺度域是选择适宜粒度的较好取值范围.对延河流域1:250000土地利用图进行景观指数计算的适宜粒度范围是70～90m,1:500000土地利用图的适宜粒度范围是90～120m.

THE EFFECTS OF GRAIN CHANGE ON LANDSCAPE INDICES

Raster data is the most commonly used format in current landscape pattern analysis, and the appropriate grid size (grain) is the key point to calculate the landscape indices. In this paper, the effects of grid size on landscape indices are discussed based on different map scales. The data are two land use maps of Yanhe basin (the scale is 1∶250 000 and 1∶500 000 respectively), and Fragstats3.3 was used as the landscape pattern analysis tool. From the research results, It was found that all the landscape indices have obvious scale effects except patch richness density. Contagion and aggregation index have not the scale inflexion in the grain range, while other indices such as patch density, Shannon's diversity index, mean fractal dimension index, etc. have a scale inflexions with the grain increased from 25m to 400m. The information of diversity, complexity and geometrical characteristics of landscape can be lost rapidly due to the scale inflexions by swallowing patches and modifying patch edges. Comparing the landscape indices values with different map scales (1∶250 000 and 1∶ 500 000 ), it can be inferred: 1) scale inflexions are not a single point, but within a certain range; 2) scale inflexions have relations with the map scales, and larger one will lead to smaller grain of the first scale inflexion; 3) the appropriate grain size should be chosen in the first scale domain. Based on the above, the appropriate grain extents are identified to calculate landscape indices for the land use maps of Yanhe basin, which is from 70m to 90m for 1∶250 000 scale, and from 90m to 120m for 1∶500 000 scale respectively.