Grassland ecosystems in arid regions increasingly experience resilience loss due to intensifying climatic variability. However, the limited interpretability of conventional machine learning models constrains our understanding of underlying ecological drivers. This study constructs an integrative framework that combines temporal autocorrelation (TAC) metrics with explainable machine learning, employing Random Forest and SHAP (SHapley Additive exPlanations) analysis. Time series of satellite-derived vegetation indices from MODIS (2001–2023), particularly the kernel Normalized Difference Vegetation Index (KNDVI), support the generation of TAC and its trend-based derivative δTAC. The framework assesses ecosystem resilience across seven representative grassland types in Xinjiang, capturing diverse responses to climate variability and vegetation dynamics. Results reveal pronounced spatial heterogeneity: resilience declines in radiation-stressed arid zones, while hydrothermally stable regions maintain stronger recovery capacity. Key drivers include temperature variability and vegetation dynamics, with divergent effects among grassland types. Meadow and Typical Steppe exhibit higher resilience under stable hydrothermal regimes, whereas desert and alpine systems show greater sensitivity to warming and climatic fluctuations. This framework enhances diagnostic transparency and ecological insight, offering a spatially explicit, data-driven tool for resilience monitoring. The findings support the formulation of targeted adaptation strategies and sustainable grassland management in response to ongoing climate change.

Grassland ecosystems in arid regions face mounting stress from intensified climate variability and anthropogenic disturbance. Despite the predictive capabilities of machine learning models, their lack of interpretability challenges the transparency of resilience drivers. This study integrates temporal autocorrelation (TAC) metrics with explainable machine learning (ML) to assess grassland resilience dynamics in an arid ecosystem from 2001 to 2023. Results reveal spatial divergence, with reduced resilience in radiation-dominated arid zones and stronger recovery in hydrothermally stable areas. The model identifies temperature variability and vegetation activity as dominant contributors to resilience trends, exhibiting marked heterogeneity across grassland types. By quantifying both structural and dynamic aspects of resilience, this framework enhances interpretability and diagnostic precision, offering a practical tool for identifying degradation risks and ecological tipping points. Findings support region-specific adaptation strategies and provide a robust foundation for sustainable grassland governance in the face of accelerating global change.