The Research On Granularity Effect Of Corralations Among Landscape Pattern Metrics

Landscape pattern metrics showed irregular variation with spatial grain size changing, namely scale effect. In addition, numbers of landscape pattern metrics were not independent. The previous results showed that the outcomes of landscape patterns analysis in different scales were different, thus, researching the scale effect of landscape pattern metrics was required. It was also necessary to research the correlation among the metrics, when we analyzed landscape patterns with some different metrics, which asked for independent. However, the study, which combined two sides, was relatively rare. It could be concluded that the correlations should have scale effect from the scale effect of landscape pattern metrics. As a result, it possessed important theory guiding meaning for researching landscape pattern metrics and landscape pattern analysis that we studied the correlation of forest landscape pattern metrics in grains.By applying ArcGIS, Fragstats, Spss and Excel, forest landscape types were classified and landscape pattern metrics were calculated on West Dongting Lake region. On the basis of analyzing the forest patterns and the scale effects of landscape pattern metrics, this paper studied principally the correlation of landscape pattern metrics and the scale effect of the correlation as the spatial grain size changing, the main contents and results were as follows:(1) 27 landscape pattern metrics formed four types in landscape-level were selected as research indicators. Forest landscape spatial pattern of west Dongting lake region was analyzed in 30m grain. The result indicated that the types of non-forest lands, broad-leaved forests, economic forests and pine forests were the main landscape types and important parts of landscape pattern regulation.(2) The scale effect of selected landscape metrics were studied in 18 grains, from 30m grain to 200m grain. The result showed that:all landscape metrics except diversity metrics were affected by changing grain. The grain effect of metrics were different, the change of total area is irregular and relatively smooth; the change of shape Index Ratio coefficient of variation, largest patch index, interspersion & juxtaposition index, landscape division index and splitting index were smooth and regular. (3) By calculating landscape pattern metrics in each grain on 15 landscapes, the scale effect of the correlation coefficient was studied, and the results showed that:the scale effect of the correlation coefficient could classify into two categories, one is quantitatively changed and another is qualitatively changed. This explained that the correlations of the same landscape pattern metrics were different when grains were different. It was the reason for the scale effect of the correlation that there was scale effect of landscape pattern metrics; however, the trend of two scale effect was not invariably linked. The factors, which affected the scale effect of the correlation among landscape pattern metrics, were respectively the landscape patterns itself, the ecological significance of the metrics, the calculation formula of the metrics and the spatial grainThe phenomenon that the correlation among landscape patterns metrics had scale effect, which was pointed out in this paper, was universal. However, many factors affected the metrics, so it requires further research whether the results, which concluded in study range on West Dongting Lake region, could scale up or down.