Combined active-passive remote sensing inversion of grassland aboveground biomass on the Qinghai-Tibet plateaus and exploration of its driving mechanisms with optical-radar-meteorological systems

The grasslands of the Qinghai–Tibet Plateau constitute the world's highest and most extensive high-elevation grassland ecosystem. Aboveground biomass (AGB) is a key indicator of grassland ecosystem structure and function and is highly sensitive to climate variability. Existing studies primarily focus on improving AGB inversion accuracy, whereas the underlying relationships between AGB and climatic drivers remain insufficiently explored. We employed Sentinel-1/2 imagery from 2018 to 2024 to evaluate the performance of GBDT, RF, and SVM in estimating AGB for three grassland types on the Qinghai–Tibet Plateau: alpine grassland, alpine meadow, and temperate grassland. We subsequently applied shapley additive explanations (SHAP) to identify AGB-sensitive features, constructed phenological curves, and assessed significant AGB trends using the Mann–Kendall test and Theil–Sen median estimator. Furthermore, optical metrics, SAR features, and ten meteorological variables were integrated into a partial least squares path modeling (PLSPM) framework to elucidate the driving mechanisms linking environmental systems and grassland AGB. The results indicate that: Bai and Cotrufo (2022) (1) Incorporating SAR features improved AGB estimation across all three grassland types, reducing RMSE by 3.55–8.10 g/m 2 and increasing R 2 by 0.05–0.08. GBDT exhibited the strongest inversion performance, with average RMSE values 4.105–16.357 g/m 2 lower and R 2 values 0.036–0.169 higher than those of SVM and RF; Balzter et al. (2007) (2) the AGB of the three grassland types ranked from high to low as alpine meadow, temperate grassland, and alpine grassland, all showing significant declines; Bernardino et al. (2024) (3) phenological curves derived from the atmospherically resistant vegetation indices and excess green index revealed that grassland growing season length first shortened and then lengthened; Bueso et al. (2023) (4) PLSPM results demonstrate that a negative statistical pathway between the regional meteorological system and grassland productivity. Temperature, precipitation, and soil moisture were associated with negative path coefficients for AGB, whereas wind speed showed a positive path coefficient. Optical features respond most sensitively to changes in the AGB, whereas SAR backscatter features have relatively limited explanatory power. These three types of systems could provide an effective framework for monitoring grassland productivity on the Qinghai–Tibet Plateau.