Identifying the vegetation destruction and restoration in surface coal mines across China over the past three decades...

It is crucial for balancing energy demands and ecological protection to understand the patterns of vegetation disturbance in surface coal mines at a national or global scale. This study developed a new method called EAuto-VDR (Enhanced Auto-VDR, Automatically identifying the vegetation destruction and restoration) to automatically identify the vegetation disturbance and revealed the disturbance patterns for more than 300 surface coal mines across various climatic contexts in China. The EAuto-VDR consists of five steps: construction of sample dataset, identification of disturbance types, automatic determination of vegetated/bare ground threshold, extraction of disturbance time, magnitude, and duration, and intelligent optimization of the results. The results show that: (1) The accuracy of EAuto-VDR reached 0.96, 0.92, and 0.90 for identifying disturbance types, destruction time, and restoration time, respectively. A dataset documenting histories of vegetation destruction and restoration has been produced and made publicly available in this paper. (2) General spatio-temporal patterns of vegetation disturbance across 329 surface coal mines of China have been revealed. Over the past three decades, surface coal mining activities in China have resulted in vegetation destruction area of 1271.34 km² totally, and 457.23 km² has been restored, and the average restoration rate is 0.36. Large-scale vegetation destruction due to mining activities began around 2003, with significant restoration activities beginning from 2010, and the "Area destroyed per ton of coal mined" (AD_t, m²/t) has been decreasing until to the latest. The "S-shaped" relationship between the cumulative vegetation destruction area and the disturbance duration, the "progressive mining" mode, and the "restoring while mining" phenomenon, were discovered. This study solved the problem of how to automatically identify the vegetation disturbance for surface coal mines in various climatic contexts, which provides an effective tool for investigating vegetation dynamics for surface coal mines at the national and even the global scale in future.