论文部分内容阅读
本文旨在筛选出一系列适合描述哈尼梯田景观格局的指数。我们从面积与边缘、形状、核心面积、隔离度/临近度、聚集度、多样性6个方面选取了47个景观指数,基于ArcGIS平台,将土地利用图矢量数据转化成栅格数据,栅格大小由30m×30m,以5m为间隔,增加至175m×175m,共计30幅土地利用图。运用SPSS软件,通过Pearson相关分析,47个景观格局指数分为21个小组,组内指数的两两相关性都在90%以上。每组选取代表性的指数,进行主成分分析,得出前5个主成分的方差变量能够占总方差变量的93%。选取旋转后各分量中具有最大相关性的5个景观结构指数来表达主成分的5个因子分别是分散指数(Splitting index[SPLIT]),斑块面积分布方差指数(Patch area distribution[AREA_CV]),香农多样性指数(Shannon’s diversity index[SHDI]),几何最近邻距离加权平均值指数(Euclidean nearest neighbor distance distribution[ENN_AM]),总核心面积指数(Total core area[TCA])。考虑到哈尼梯田的实际景观特征并结合前人的研究成果,我们增加平均分维数(Patch fractal dimension distribution[FRAC_MN])作为第6个指数。由于景观指数具有尺度效应,我们通过景观格局指数与尺度变化关系图,计算得出红河哈尼梯田景观格局指数的尺度域为40–45m之间。通过分维分析法,识别出哈尼梯田的特征格局在40–45m之间,即研究哈尼梯田的景观格局的最适宜栅格大小为40–45m之间。通过研究,我们认为以上6个景观指数能够反映哈尼梯田的景观格局特征,同时也可以运用这6个指数来描述类似哈尼梯田的景观格局。
This paper aims to screen out a series of indices that are suitable for describing the landscape pattern of Hani terraces. We selected 47 landscape indices from six aspects of area and edge, shape, core area, isolation / proximity, degree of aggregation, and diversity. Based on the ArcGIS platform, we converted land use map vector data into raster data, Size from 30m × 30m, with 5m intervals, increased to 175m × 175m, a total of 30 land use map. By using SPSS software and Pearson correlation analysis, the 47 landscape pattern indices are divided into 21 groups, and the correlation between groups of indices is over 90%. The representative indexes of each group were selected and the principal component analysis was conducted. The variance variables of the first five principal components can account for 93% of the total variance variables. The five factors that express the principal components using the five landscape structure indices with the highest correlation among the components after rotation are Splitting index (SPLIT), Patch area distribution (AREA_CV) , Shannon’s diversity index (SHDI), Euclidean nearest neighbor distance distribution (ENN_AM) and Total core area (TCA). Considering the actual landscape characteristics of the Hani terraces and combining with the previous research results, we increased the Patch fractal dimension distribution (FRAC_MN) as the sixth index. Due to the scale effect of the landscape index, we calculated the landscape scale index of the Hani terraces in Honghe from 40-45m by the relationship between the landscape pattern index and the scale. Through the fractal analysis method, the characteristic pattern of the Hani terraces is identified between 40-45m, that is, the most suitable grid size for studying the landscape pattern of the Hani terraces is between 40-45m. Through the research, we believe that the above six landscape indices can reflect the landscape pattern characteristics of Hani terraces. At the same time, these six indices can also be used to describe the landscape pattern similar to Hani terraces.