Assessing Land Degradation and Restoration in Eastern China Grasslands from 1985 to 2018 Using Multitemporal Landsat Data
Caixia Liu,Huabing Huang,John M. Melack,Ye Tian,Jinxiong Jiang,Xiao Fu,Zhiguo Cao,Shaohua Wang
DOI: https://doi.org/10.1109/jstars.2024.3483992
IF: 4.715
2024-11-01
IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Abstract:The grassland ecosystems of Xilingol, China, characteristically part of the vast Eurasian steppe, are currently facing two challenges: natural variations and anthropogenic stress, which are leading to significant degradation. This article harnesses a sequence of high-resolution (30 m) land cover and greenness trend maps derived from multiyear Landsat imagery to describe these ecologically critical shifts over a landscape spanning more than 200 000 km2. By leveraging random forest models complemented with phenological patterns, we streamlined the generation of land cover maps, securing overall accuracies upwards of 94% across eight categorical classifications, as substantiated by rigorous validation. Between 1985 and 2000, there were significant changes in the landscape, such as an increase in farmland of about 4.0 × 103 km2, mostly at the expense of natural grasslands and wetlands. Throughout the study period, an ongoing trend is the noticeable shrinkage of water bodies with the biggest reduction of wetlands reported between 1995 and 2015. Open-pit mining regions began to increase with the start of the 21st century, and from 1985 to the present, urbanization drove the growth of impervious surfaces. These maps offer powerful visual representations of major land use changes, capturing the expansion of surface mining, the retreat of wetland areas, and the growth of urban areas. Therefore, our findings compose an essential part in the documentation and comprehension of the details of wetland reduction, cropland intensification, surface water decline, and rapid urban growth, providing crucial information to conservationists and policymakers working toward sustainable ecosystem management.
imaging science & photographic technology,remote sensing,engineering, electrical & electronic,geography, physical