三峡库区30 a间土地利用景观特征的粒度效应

该研究是为了测度景观指数在时空上随粒度变化的特征；总结相同景观在不同粒度下的景观特征差异。利用不同的粒度大小，将景观类型图进行了栅格化，并选取景观指数，在类型与整体景观的水平上，进行了景观指数的粒度效应测度。并采用相关性分析的方法，分析了各景观指数与粒度相关性的差异。研究表明：粒度效应能有力地论证三峡库区近30 a景观格局变化的特征；不同的景观指数随粒度也表现出不同的敏感程度，较敏感的是斑块密度、有效网格大小、正归化形状指数、聚合指数。同一类型景观在不同的景观指数测度上表现不同的变化规律，随粒度变化较为敏感的是耕地（坡度＜25°）、林地、草地和城镇建设用地；多数指数呈现出随粒度变化的拐点，尤其在尺度400 m；整体水平上，斑块密度和景观形状指数受粒度变化的影响较大；同一指标与粒度的相关性在类型水平和时间跨度上大多呈现出较高的一致性。此次研究很好反映了景观格局指数在不同空间分辨率（粒度）上的变化规律。

Effects of spatial grain size on landscape pattern of land use in Three Gorges Reservoir Area in 30 years

This paper is purpose to measure landscape indexes traits with changes of grain sizes, and summarize differences of landscape traits in different grain sizes. In this paper, five remote sensing images of Three Gorges Reservoir Area in 1975, 1987, 1995, 2000 and 2005 were taken as data for present study. The images were interpreted into draws of landscape pattern. In the ArcView3.3 environment, the vector draws were transformed into shatter draws according to different spatial grain size. In the same time, landscape indexes were chosen and the effects between spatial grain size and landscape indexes were measured at the scale of pattern and landscape. Then the correlation differences between landscape indexes and spatial grain size were analyzed with the method of correlation analysis. The results showed that, the traits of changes of landscape patterns were proved in 30 years. The effective sensitiveness between different landscape indexes and spatial grain size was different, such as patch density (PD), effective mesh size (MESH), standardized trait index (NSLI), aggregation index (AI), and same landscape pattern in the process of measuring was taken on different change rules, such as cultivated land (slope<25°), forestry land, grassland, urban constructive land. There were some turning points in the changes of effect, especially at 400 m. At the scale of landscape, patch density and land shape index were largely impacted by the spatial grain size. The correlation between the same index and grain shows highly consistency at the scale of pattern and time. Above all, this research can reflect the change rules of landscape indexes in different spatial grain size.