中国地表覆盖异质性参数提取与分析

Extraction and Analysis of Land Cover Heterogeneity over China

地表异质性广泛存在于陆地表面各个尺度,是地表参数遥感反演不确定性的主要来源之一。基于高分辨率地表分类参考图,提取出低分辨率混合像元的端元数量和边界长度指标来描述地表异质性。然后以中国地区为例,使用全国30 m空间分辨率Global Land 30地表分类数据集提取出1 km尺度像元的描述混合结构和破碎程度的异质性指标。并基于提取出的异质性指标分析了中国区域在1 km尺度上非均质地表地物类型的组合特征、斑块特征和不同生态群系内部异质性特征。发现山地和生态交错区是主要的高异质性区域,稀树草原生物群系内部异质性最大(平均边界长度为7 426 m),其次依次为森林(4 323 m)、耕地/草地(3 160 m)和灌丛(1 779 m)。

Spatial heterogeneity exists in land surface at every scale, and it is one of key factors to bring uncertainty to land parameter retrieval from remote-sensed data. This paper proposed a methodology to use the boundary length among different land cover types to characterize and quantify land surface heterogeneity based on high-resolution land cover images. Then the heterogeneity feature at 1 kilometer scale in China was extracted from “GlobalLand30” land cover datasets with the spatial resolution of 30 m. The mixed structure, degree of fragmentation and intra-heterogeneity of eight main vegetation biomes from MODIS land cover product over heterogeneous surface in china were analyzed. Mountain area and ecotone are more heterogeneous than other regions. Savanna biome （average boundary length is 7 426 meters） is the most heterogeneous zone followed by forest, grass/crop and shrub biome with average boundary length of 4 323, 3 160, 1 779 meters, respectively.