基于条件随机模拟的DEM误差分析——以董志塬水土流失等级划分为例

DEM Error Analysis Based on Conditional Stochastic Simulation

DEM是对地球表面的模拟和模型化表达,DEM不可避免的含有误差,且DEM误差具有空间可变性和相关性。常用的DEM误差估算模型为中误差(RMSE),但RMSE为全局变量,无法反映误差的空间性。为了克服RMSE的缺陷,本文采用条件随机模拟实现了DEM误差曲面模拟。通过董志塬水土流失等级划分表明,DEM误差在平坦区域严重影响坡度精度,且坡度最大误差变程大于高程最大误差变程,DEM误差被放大;使用概率模型和模糊度模型分析表明,大部分网格点水土流失等级划分均受到DEM误差影响;条件随机模型的使用可以让DEM用户更加准确的分析和评价DEM误差对最终决策的影响。

A digital elevation model（DEM） is a representation of terrain elevation as a function of geographic location.DEM inevitably contains errors,which are spatial variable and correlated.The usual error model is Root Mean Square Error（RMSE）,whereas it is a global variable,and can not show the DEM error of each grid.In order to overcome the deficiency of RMSE,Dongzhi Tableland was employed as a test region,and an error surface about 6.25km2 was constructed based on the conditional stochastic simulation model（CSS）.Water and soil loss level determination was taken as an example to validate the importance of CSS.The results indicated that the uncertainties in derived slope of the Dongzhi Tableland tend to be found in flatter areas;the slope error range is bigger than DEM error range,which indicates that the DEM error is magnified;probabilistic and fuzzy model shows that almost all grids are influenced by the DEM error in water and soil loss level determination.As a tool for DEM error analysis,conditional stochastic model can improve the user assessment of DEM error.