基于SRTM DEM的地形起伏度对天水市黄土滑坡的影响分析

以天水市辖区30 m分辨率的SRTM DEM数据为基础，首先利用GIS空间分析模块中邻域统计分析法，依次选取2×2、3×3、4×4、…、28×28、29×29、30×30共29个矩形邻域分析窗口，运用均值变点分析法进行最佳计算单元的计算分析。结果表明：12×12的网格大小（12.96×10⁴ m²）为曲线突变的拐点，作为曲线由陡变缓的阈值，确定12×12的网格大小（12.96×10⁴ m²）为研究区地形起伏度计算的最佳单元，生成研究区地形起伏度分级图。进而采用确定性系数法（CF）、证据权法、信息量法和条件概率法对研究区发育的黄土滑坡灾害与地形起伏度之间的相关关系进行了定量分析。研究结果表明，区内地形起伏度影响因子层中地形起伏度60~120 m的分级因子层为黄土滑坡灾害的易发因子层，而其中60~90 m区间的值最大，为黄土滑坡灾害的优势因子区间，对滑坡的响应最敏感。

Analyzing the impact of relief amplitude to loess landslides based on SRTM DEM in Tianshui Prefecture

Based on the SRTM DEM in the resolution of 30 m in municipal districts of Tianshui Prefecture, spatial neighborhood analysis was done with the window of 29 different sizes (from 2×2, 3×3,…, to 30×30). Then, mean change-point analysis was done to select the best statistical unit (the size of window). The results show that the curve appeared a inflection point at the grid unit of 12×12 mesh size (12.96×10⁴ m²), which can be taken as a threshold and be used as the optimal analytical unit. Hereby, a grade diagram of relief amplitude can be completed in the study area with the mesh size of 12×12. The correlation of loess landslides with relief amplitude were quantitatively calculated with certainty factor method, weights-of-evidence method, information method and conditional probability method. The result shows that the relief amplitude of 60~120 m is the sensitive relief, in which, 60~90 m is the relief with most landslides and is most sensitive to landslides.