Spatial–Temporal and Driving Factors of Land Use/Cover Change in Mongolia from 1990 to 2021

作者：Junming Hao, Qingrun Lin, Tonghua Wu, Jie Chen, Wangping Li, Xiaodong Wu, Guojie Hu, Yune La · 发表于：Remote Sensing · 年份：2023 · DOI：10.3390/rs15071813 · 被引用次数：30 · 研究领域：Rangeland Management and Livestock Ecology、Plant Water Relations and Carbon Dynamics、Remote Sensing in Agriculture

During the past several decades, desertification and land degradation have become more and more serious in Mongolia. The drivers of land use/cover change (LUCC), such as population dynamics and climate change, are increasingly important to local sustainability studies. They can only be properly analyzed at small scales that capture the socio-economic conditions. Several studies have been carried out to examine the pattern of LUCC in Mongolia, but they have been focused on changes in single land types at a local scale. Although some of them were carried out at the national scale, the data interval is more than 10 years. A small-scale and year-by-year dataset of LUCC in Mongolia is thus needed for comprehensive analyses. We obtained year-by-year land use/cover changes in Mongolia from 1990 to 2021 using Landsat TM/OLI data. First, we established a random forest (RF) model. Then, in order to improve the classification accuracy of the misclassification of cropland, grassland, and built and barren areas, the classification and regression trees model (CART) was introduced for post-processing. The results show that 17.6% of the land surface has changed at least once among the six land categories from 1990 to 2021. While the area of barren land has significantly increased, the grassland and forest areas have exhibited a decreasing trend in the past 32 years. The other land types do not show promising changes. To determine the driving factors of LUCC, we applied an RF feature importan...