Spatiotemporal Dynamics and Driving Mechanism of Aboveground Biomass Across Three Alpine Grasslands in Central Asia over the Past 20 Years Using Three Algorithms

作者：Xu Wang, Yansong Li, Yanming Gong, Yanyan Liu, Jin Zhao, Kaihui Li · 发表于：Remote Sensing · 年份：2025 · DOI：10.3390/rs17030538 · 被引用次数：7 · 研究领域：Rangeland Management and Livestock Ecology、Tree-ring climate responses、Fire effects on ecosystems

Aboveground biomass (AGB) is a sensitive indicator of grassland resource quality and ecological degradation. However, accurately estimating AGB at large scales to reveal long-term trends remains challenging. Here, single-factor parametric models, multi-factor parametric models, and multi-factor non-parametric models (Random Forest) were developed for three grassland types (alpine meadow, alpine grassland, and swampy meadow) in the Bayanbuluk Grassland using MODIS satellite data and environmental factors, including climate and topography. A 10-fold cross-validation method was employed to assess the accuracy and stability of these models, and an AGB remote sensing inversion model was established to estimate the AGB of the Bayanbuluk Grassland from 2005 to 2024. Moreover, the BEAST mutation test, Theil–Sen median trend analysis, and Mann–Kendall test were used to analyse the temporal trends of AGB, identify the years of mutation points, and explore the changes in AGB across the entire study period (2005–2024) and at 5-year intervals, considering the influence of climatic factors. The results indicated that the machine learning (RF) model outperformed both multi-factor parametric and single-factor parametric models, with specific improvements in R2 and RMSE across all grassland types. For instance, the RF model achieved an R2 of 0.802 in alpine grasslands, outperforming the multi-factor parametric model with an R2 of 0.531. The overall spatial distribution of AGB exhibited hetero...