Monitoring vegetation restoration on potentially contaminated mining land using a novel multi-source data fusion method

Accurately mapping and monitoring of vegetation restoration area on potentially contaminated mining land (PCML) is crucial for evaluating ecological recovery from mining-related pollution. However, the types and patterns of PCML’s vegetation restoration area (PCML-VRA) remain unclear at large scales, despite numerous studies conducted at single-site or local scales. In this study, we developed a PCML-VRA dataset in Inner Mongolia Autonomous Region from 2010 to 2020 using a novel multi-source data fusion method. This method integrated high-resolution satellite imagery, potentially contaminated enterprise sites data, mineral deposit database, land use/cover data, and other thematic data using the Google Earth Engine (GEE) cloud platform in combination with GIS local platform. The overall accuracy of the PCML-VRA vector dataset in Inner Mongolia was (87.80 ± 5.81)%, demonstrating high reliability and quality. Results showed that the area of PCML in 2010 was 954.34 km2, and the area of PCML-VRA in Inner Mongolia between 2010 and 2020 was 373.33 km2. Consequently, the level of vegetation restoration within PCML in Inner Mongolia was 39.12% over the decade, with gross domestic product, precipitation, wind speed, and population identified as the dominant factors. This study enhances access to high-resolution mapping of PCML-VRA, facilitating the evaluation of ecological protection and the effectiveness of mine greening initiatives.