Coal mining and ecological restoration activities significantly affect land surfaces, particularly vegetation. Long-term quantitative analyses of vegetation disturbance and restoration are crucial for effective mining management and ecological environmental supervision. In this study, using the Google Earth Engine and all available Landsat images from 1987 to 2020, we employed the Landsat-based Detection of Trends in Disturbance and Recovery (LandTrendr) algorithm and Support Vector Machine (SVM) to conduct a comprehensive analysis of the year, intensity, duration, and pattern of vegetation disturbance and restoration in the Heidaigou and Haerwusu open-pit coal mines (H-HOCMs) in the Jungar Coalfield of China. Our findings indicate that the overall accuracy for extractions of disturbance and restoration events in the H-HOCMs area is 83% and 84.5%, respectively, with kappa coefficients of 0.82 for both. Mining in Heidaigou has continued since its beginning in the 1990s, advancing toward the south and then eastward directions, and mining in the Haerwusu has advanced from west to east since 2010. The disturbance magnitude of the vegetation greenness in the mining area is relatively low, with a duration of about 4–5 years, and the restoration magnitude and duration vary considerably. The trajectory types show that vegetation restoration (R, 44%) occupies the largest area, followed by disturbance (D, 31%), restoration–disturbance (RD, 16%), disturbance–restoration (DR, 8%), restoration–disturbance–restoration (RDR), and no change (NC). The LandTrendr algorithm effectively detected changes in vegetation disturbance and restoration in H-HOCMs. Vegetation disturbance and restoration occurred in the study area, with a cumulative disturbance-to-restoration ratio of 61.79% since 1988. Significant restoration occurred primarily in the external dumps and continued ecological recovery occurred in the surrounding area.

Climate change and climate extremes and their impacts on agriculture, water resources, and ecosystems have become important issues globally. Agricultural sustainability and food security are facing unprecedented challenges due to the increasing occurrence of extreme climatic events, including, notably, extreme droughts in recent years in China. In this study, a threshold determination model of extreme agro-climatic droughts (EADs) was built based on the cumulative probability distribution functions (CDF) of an agricultural drought index—the consecutive days without available precipitation (CDWAP). The CDWAP was established by combining meteorological data with the characteristics of cropping patterns and the water requirement in different growing periods of crops. The CDF of CDWAP was obtained based on the relationship of CDWAP and its occurrence frequency. Based on the model, the spatial pattern of the thresholds of EADs and the threshold exceedance time series of EADs in 500 meteorological stations were obtained, and then changes in the frequencies and intensities of EADs in China and their impacts on grain yields in rain-fed regions during the past 50 years were analyzed. The results follow: (1) The threshold value of EADs in China gradually increased from southeast to northwest. The stations of the highest value were located in the Northwest China, with the CDWAP more than 60 days, while the lowest value was in the middle reaches of the Yangzi River, with the CDWAP less than 16 days. (2) The frequencies and intensities of the EADs increased mostly in the east areas of the Hu Huanyong line, which was also the main agricultural production region in China. The North China (NC) and Southwest China (SW) regions showed the highest increasing rates of the EADs; their frequencies and intensities were 11.3% and 2.2%, respectively, for the NC region, and 9.3% and 2.7%, respectively, for the SW region. (3) Case studies in the NC, SW, and SE regions indicated that there was a negative correlation between grain yields and EAD frequency and intensity; i.e., the low grain yields often occurred in the year with relatively higher frequency or/and stronger intensity of EADs. The correlation coefficients of grain yield and EAD were generally greater than that of merely extreme climatic droughts; therefore, the study of EAD is necessary when researching the impacts of extreme drought events on grain yield.