This is the authors' abstract. We don't add key points for this paper.
This study evaluates the degradation of Inner Mongolian grasslands from 1980 to 2020 by analysing changes in grassland formations. A random forest-based machine learning classification model was developed using the Vegetation Map of China (1:1,000,000, 1980 s), the Vegetation Map of Inner Mongolia (1:250,000, 2009), and multi-source datasets. Spatial distribution maps of grassland formations for four periods (1981–2020) were simulated and validated through field surveys and independent datasets. The model achieved a validation set accuracy of 71.9 % for temperate grasslands in the 1980 s and 68.1 % in the 2000 s. Formation boundaries were the primary areas of classification errors; excluding the 2 km buffer zones significantly enhanced the classification performance. Field surveys showed a 76.3 % agreement between observed dominant or subdominant species and the simulation results. An analysis of temporal distribution revealed a continuous expansion of the Stipa formation, mainly at the expense of the Leymus chinensis and shrub formations. Shrub formations expanded during the 2000 s but declined in the 2010 s; climate change is the primary driving factor behind these changes. This study provides insights into grassland degradation from the perspective of formations, offering valuable data for grassland management and ecological conservation in Inner Mongolian.
Answers are grounded in this paper and cite their sources.
No takes yet. Share an insight, caveat, or question.
Zhang et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: