3个时期达里诺尔国家级湿地自然保护区土地利用动态研究

受气候暖干化和人类活动影响，蒙古高原干旱区内陆湿地土地利用变化显著，及时掌握湿地土地利用变化，可为湿地自然保护区土地资源管理提供参考依据。[方法]本研究以1994、2005和2016年达里诺尔国家级自然保护区Landsat-TM/OLI影像为数据源，采用决策树分类、转移矩阵和动态度等方法对其进行动态分析。[结果]结果表明：研究区土地利用类型分类总体精度较高，主要以草地、林地、湖泊为主，沼泽、沙地、滩地、裸地所占比例较小。1994—2005年238.93 km2地物发生变化，占总面积18.81%，其中，林地和草地分别增加了19.25%和10.94%；沙地和裸地分别减少了66.63%和47.19%。2005—2016年，191.72 km2地物发生变化，占总面积15.09%，其中，沙地和滩地分别增加了19.73%和12.64%；裸地、沼泽和林地分别减少了59.51%、21.36%和15.72%。[结论]研究区地物覆盖类型变化明显，1994—2016年草地、林地、滩地面积增加，湖泊、沼泽、裸地面积减少。土地利用类型主要转移的方向为湖泊类型转化为滩地和沼泽，沼泽和裸地类型转化为草地，总体上湿地功能具有退化趋势。

Dalinur National Wetland Nature Reserve in 3 Periods: Land Use Dynamics

The land use change of inland wetland in arid area of Mongolian plateau is remarkably influenced by warm-dry climate and human activities, so grasping the land use change in wetland in time could provide reference for the management of land resources in wetland nature reserve. Based on the data of Landsat-TM/ OLI remote sensing image of Dalinur National Nature Reserve in 1994, 2005 and 2016, the authors analyzed dynamically the land use change by using decision tree classification, transfer matrix and the dynamic degree. The results showed that: the accuracy of classification of land use types was overall higher, and the land use types in the research area mainly included grassland, forests and lakes, and the proportion of marsh, sandy, mudflat and bare place was small. From 1994 to 2005, 238.93 km2 features changed, which accounted for 18.81% of the total area, among them, the forest and the grassland increased by 19.25% and 10.94% , respectively, the sandy land and the bare land decreased by 66.63% and 47.19%, respectively. From 2005 to 2016, 191.72 km2 features changed, accounting for 15.09% of the total area, among them, the sandy and the The land use change of inland wetland in arid area of Mongolian plateau is remarkably influenced by warm-dry climate and human activities, so grasping the land use change in wetland in time could provide reference for the management of land resources in wetland nature reserve. Based on the data of Landsat-TM/ OLI remote sensing image of Dalinur National Nature Reserve in 1994, 2005 and 2016, the authors analyzed dynamically the land use change by using decision tree classification, transfer matrix and the dynamic degree. The results showed that: the accuracy of classification of land use types was overall higher, and the land use types in the research area mainly included grassland, forests and lakes, and the proportion of marsh, sandy, mudflat and bare place was small. From 1994 to 2005, 238.93 km2 features changed, which accounted for 18.81% of the total area, among them, the forest and the grassland increased by 19.25% and 10.94% , respectively, the sandy land and the bare land decreased by 66.63% and 47.19%, respectively. From 2005 to 2016, 191.72 km2 features changed, accounting for 15.09% of the total area, among them, the sandy and the