基于GIS/RS的滨海湿地景观格局多尺度分析——以辽宁省大洼县为例

以辽宁省大洼县滨海湿地为研究对象，基于GIS技术，采用2000、2009年两个时相ETM／TM影像数据，选取了斑块类型水平和景观水平上有代表性的27个景观指数来定量研究滨海湿地景观格局的多尺度特征，分析了两个时间尺度上8种空间粒度的滨海湿地景观指数的尺度规律与景观格局的时空分异。研究表明：大洼滨海湿地的斑块密度（PD）、斑块数（NP）和蔓延度（CONTAG）等指数与空间粒度有很明确的函数关系（幂／对数函数等），可预测性强；最大斑块指数（LPI）、最大形状指数（LSI）等总体呈增大趋势，但无明显函数关系；而Shannon多样性、均匀性指数（SHDI、SHEI）等则无空间粒度变化效应。2000～2009年，斑块类型百分比（PLAND）、景观分维数（FRAC_MN）等指数在增大，Shannon多样性（SHDI）等指数减小。说明了人类活动对大洼滨海湿地景观格局的干扰明显．景观破碎度增加，总体湿地景观结构向简单化方向发展。2000、2009年两个时相的重要景观指数斑块数（CA）、边界密度（ED）、聚集度（AI）等与空间粒度呈相似的函数关系，具有较好的预测性。

Multi-scale Analysis of Coastal Wetland Landscape Pattern Based on GIS/RS A Case Study of Dawa County in Liaoning Province

Taking coastal wetland in Dawa County, Liaoning Province for example, multi-scale features of the landscape pattern were studied quantitatively on the basis of GIS, by resorting to image data of two temporal in 2000 and 2009 respectively, ETM/TM, and picking 27 typical landscape indexes due to patch class level and landscape level. Thus, both scale regulation of the landscape index of the coastal wetland and tem-poral-spatial differences of the landscape pattern of 8 types of spatial grain in 2 temporal scales were analyzed. The research result indicates that accurate functions which give fairly good predictabilities, such as power / logarithmic function, exist between spatial grain and indexes like patch density of Dawa coastal wetland, number of patches, contagion index, etc. , while both Shannon＇ s diversity index and Shannon＇ s evenness index do not witness any difference brought by the changes in the spatial grain. Besides, from 2000 to 2009, there were increases in indexes like per-cent of landscape as well as mean fractal dimension of patch, whereas decline occurred in indexes such as SHDI, which demonstrates that activi-ties of human being disturb landscape structure of Dawa coastal wetland significantly-the overall landscape structure of wetland tends to be simp-ler with the growth in landscape fragmentation. Finally, similar function relations does spatial grain share with crucial landscape indexes of 2 tem-poral in 2000 and 2009, number of patches, edge density and aggregation index, etc., that the predictabilities turn out to be relatively well.