基于DEM的数字地形分析

该文在对数字高程模型（DEM）数据来源及结构、数字地形分析及其应用、基于DEM的地形分析中的不确定性和误差分析的基础上，以日本东北地区岩手县早池峰山为研究对象，美国MicroImage公司开发的TNTmips地理信息系统为工具，日本国土地理院发行的“数字地图25000”为基础数据，研究基于DEM的数字地形特征提取与分析方法，以及DEM精度对地形特征的影响. 研究结果表明:①以DEM为基础可提取多种地形特征，如坡度、坡向、坡面形态、流域边界、水流路径等，这些特征在地理信息系统的支持下均可用图形和属性数据来表示；②DEM水平精度对基于DEM提取的数字地形特征影响表现为:低精度的DEM将导致研究区平均坡度变小、坡度标准差变大；同时，DEM精度对不同坡度区域表现为不同的影响，其结果按坡度大致可划分为3种不同类型，即0°～10°、10°～35°以及大于35°； DEM精度对坡向的影响除平坡外变化较小，其中平坡面积随DEM精度的降低而增大；低精度的DEM将导致水文地形信息受损，这将严重影响流域水文模型参数的确立及水文过程模拟分析的精度. 

Digital terrain analysis based on DEM

The digital elevation model (DEM), an important source of information, is usually used to express the topographic surface in three dimensions and imitates the essential natural geography. Now, DEM has been applied in physical geography, hydrology, ecology and biology and when connected with topographic attribute extraction, in watershed and flow path analyses. The study has analyzed the digital elevation data sources and their structure, the arithmetic of terrain attribute extraction from DEM and its application as well as DEM’s error and uncertainty algorithm. The Hayachinesan mountain area (in Northeastern Japan) was chosen as the research site and the focus was on terrain analysis and the impacts of DEM’s resolution on topographic attributes using TNTmips GIS software (developed by MicroImage Inc., USA) and “digital map 25000”(published by the Geographical Survey Institute of Japan in 1998). The results show that: ① DEM is a very effective tool for terrain analysis; many terrain attributes can be derived such as slope, aspect, slope type, watershed, standard flow path and so on; these attributes can be displayed with both image and attribute databases with the help of GIS; ②DEM’s resolution has a great influence on terrain attributes; the following details are shown: (a) DEM resolution has a significant effect on slope estimation: the average slope becomes smaller and the standard deviation larger when DEM resolution changes from fine to coarse and the different impacts of DEM’s resolution on different slope ranges can be classified into three gradient classes: 0-10 degree (underestimated slope), 10-35 degree (overestimated slope) and more than 35 degree (little impact on slope estimation); (b) DEM resolution has little effect on aspect estimation, but flat areas become larger when DEM resolution changes from fine to coarse; and (c) the quantity of hydrologic topography information declines as DEM resolution decreases.