文献类型：期刊文献

中文题名：基于RS和GIS的土地利用格局模拟与分析——以河北省沽源县丰元店乡为例

英文题名：Simulation and analysis of land use patterns based on RS and GIS:a case study for Fengyuandian town,Guyuan county,Hebei province

作者：张永民[1]

第一作者：张永民

机构：[1]河南财经学院资源与环境科学系

第一机构：河南财经政法大学资源与环境学院

年份：2007

卷号：30

期号：6

起止页码：981-987

中文期刊名：干旱区地理

外文期刊名：Arid Land Geography

收录：CSTPCD;;北大核心:【北大核心2004】；CSCD:【CSCD2011_2012】；

基金：国土资源部合作项目"典型县级耕地资源分布与生态退耕遥感监测"

语种：中文

中文关键词：土地利用格局;空间模拟;Logistic回归模型;丰元店乡

外文关键词：land use patterns; spatial simulation; logistic regression models; Fengyuandian town

摘要：基于由高分辨率的遥感和数字高程模型数据源得到的土地利用、地形、道路和居民点等空间数据,运用Logistic回归模型对河北省沽源县丰元店乡的旱地、有林地、其它林地、天然草地、改良与人工草地、未利用地,以及其它土地共7种土地利用类型的空间分布格局进行了模拟与分析(模拟网格为50m×50m),并使用ROC方法对所有回归模型的拟合优度进行了检验。结果表明,旱地、有林地、改良与人工草地,以及未利用地和其它土地的回归模型的拟合优度较高(ROC值分别是0.868,0.795,0.830,0.790和0.930),而其它林地和天然草地的回归模型的拟合优度相对较低(ROC值分别是0.711和0.719)。研究结果揭示了地形、道路和居民点分布等因素对研究地区土地利用格局形成与演变的重要决定作用,这为进一步研究丰元店乡未来的土地利用动态变化情景奠定了基础,同时也可以为该乡的土地利用管理提供科学依据。
To understand the relations between land uses and their spatial determinants, a case study is selected for Fengyuandian town located in the southeastern part of Guyuan county, Hebei province on the basis of its land use and DEM data in this study. Such understanding is important for the development of comprehensive land use change models. Thus independent binary logistic regression models were constructed for investigating spatial patterns of land use types of dry land, forest land, other forest, natural grassland, modified and human-made grassland, un-used land, and other land with a spatial unit of 50 m × 50 m cells. In addition, all these land use models were validated by a ROC method. The ROC can be used to compare a map of the actual land use distribution to its modeled probability map. The range of ROC values is between 0.5 and 1.0. The greater the ROC is, the more accurate the modeled probability map will be. The results indicate that, as to the fitting degree, dry land （ ROC = 0.868 ）, forest land （ROC=0.795 ） , modified and human-made grassland （ ROC = 0.830 ）, unused land （ ROC = 0.790 ）, and other land （ ROC = 0.930） models have the high goodness of the land use patterns simulation, and the rest of another two land use （other forest and natural grassland） models are of the relatively low fitting degrees. It is argued that these types of the analysis can provide the valuable information for modeling future land use change sce-narios that need to consider local and regional conditions of actual land use. On the other hand, the probability maps of the land use types obtained from this study can also support the local government on making land management decisions, such as the conversion of the unsustainable dry land to the forest land or grass land.