Spatiotemporal dynamics and key climatic influences on vegetation resilience in opencast coal mine dumps after restoration

Vegetation resilience is crucial for understanding the self-repair and adaptive capacity of regional ecosystems. Opencast coal mine dumps, typical targets for ecological restoration, exhibit highly unstable vegetation ecosystems after restoration, making them prone to state shifts. However, existing studies have limitations in capturing vegetation resilience characteristics and its climatic driving mechanisms. This study addresses these deficiencies by focusing on the Pingzhuang West Opencast Coal Mine dumps, utilizing Critical Slowing Down (CSD) theory and long-term Landsat remote sensing data from 2008 to 2024. We propose ’MultiRes’, a pixel-level (30 m) method to calculate vegetation resilience. Unlike traditional fixed-window approaches, MultiRes offers adaptive window sizes with a wide range of stability. We analyzed the spatiotemporal dynamics of vegetation resilience after restoration, evaluated the effectiveness of vegetation restoration, and quantitatively assessed the impact of key climatic drivers across different phases. Results reveal that: (1) After restoration, vegetation resilience at each dump experienced three phases: initial enhancement, decline, and renewed enhancement. (2) Vegetation resilience improved significantly compared to the initial fragile ecosystem, with over 88 % of the area showing improvement, especially at Taipingdi, where the enhancement rate reached 99.85 %. (3) The influence of key climatic drivers remained consistent within each dump across all phases, particularly in the first and third phases. Areas dominated by a single climatic driver generally showed more significant changes than those influenced by combined drivers. These findings demonstrate that vegetation resilience captures stage-specific ecological patterns that NDVI alone cannot detect, thereby supporting adaptive restoration and climate-informed management in mining environments.