土地覆被类型空间格局与地形因子的定量关联及其应用

以全国1∶25万土地覆被数据、全国公里网格DEM数据和县级行政区划数据为基础,首先,在国家尺度上宏观分析地形因子对土地覆被类型空间分布格局的影响,土地覆被类型空间分布宏观格局受海拔高度、坡度、地表起伏度的影响明显,而与坡向的关系并不显著;其次,以县级行政单元为统计分析样本,利用多元线性回归分析方法,在国家和区域两个尺度建立土地覆被类型面积占比与主要地形因子之间的定量模型,在6类土地覆被类型中,除草地以外,其他5类土地覆被类型的面积占比均与地形因子之间呈显著相关关系,显著性由高到低的排列顺序依次为森林、农田、荒漠、聚落和湿地水体;最后,以相关性最好的森林的空间分布为例,说明了将所建立的模型应用于栅格单元上某种土地覆被类型面积占比估算的可能性,虽然估算结果与实际情况存在差异,但总体趋势基本保持一致,特别是土地覆被类型面积占比大的区域。

Relationship between Land Cover Distribution Pattern and Terrain Factors and Its Application

National land cover data of China at scale of 1 to 250,000,Digital Elevation Model (DEM) at resolution of 1 km by 1 km and county-level administrative divisions' data were used as basic data to explore the relationship between spatial distribution patterns of land cover and terrain factors.Firstly,impact of the terrain factors on spatial distribution patterns of land cover,which included elevation,slope,aspect and relief degree of land surface,was analyzed macroscopically at nation-wide.Macro distribution pattern of land cover was obviously influenced by altitude,slope and surface undulation.It was almost unaffected by aspect.Secondly,county-level administrative units were used as samples for statistical analysis and multiple linear regression analysis as method to establish quantitative models between area ratio of land cover and main topographical factors at national and provincial scales respectively.Regression analysis results showed that area proportion of five (except grassland) of six categories of land cover was significantly correlated with main terrain factors,which included altitude,slope,aspect,land surface undulation and surface cutting depth.Significant order from high to low was forest,farmland,desert,settlement and wetland & water body.Finally,methodology that the models were applied to estimate area proportion of land cover on grid cell scale was explored with forest as an example,which had the closest relationship with main topographical factors among 6 kinds of land cover.The quantitative models between area proportion of land cover and terrain factors based on county-level statistical unit (macro-scale) might absolutely be used to estimate area proportion of land cover at grid cell (micro-scale).Although there existed some differences between estimated results and actual situation,the overall trend between the two remained the same,especially in the areas where there was high area proportion of land cover types.