Observation of coal mining activities in Ha Long and Cam Pha cities, Quang Ninh province, Vietnam using Sentinel-2 data

Mapping the spatial distribution of exposed coal sites over a large territory and ensuring temporal continuity is essential for environmental assessment. It helps identify relationships between coal mining activities and environmental issues such as water pollution from coal contamination or air pollution from particulate matter in the coastal area of Ha Long - Cam Pha. The carbon index has proven to be an effortless yet effective technique. It is calculated using the spectral ratio between two shortwave infrared bands (Band 11 - SWIR1 and Band 12 - SWIR2) of Sentinel-2 imagery. The selection of an appropriate carbon threshold value - used to distinguish exposed coal from other land cover types, particularly urban areas - was validated through visual interpretation of high-resolution satellite images from the Google Earth dataset. This approach, leveraging Sentinel-2 imagery with higher spatial resolution, differs from previous studies that applied complex algorithms or multiple spectral indicators to lower-resolution images for exposed coal site detection. The results, with an overall accuracy of 87.35% and a Kappa coefficient of 0.73, provide valuable support for monitoring and managing natural resource exploitation. Additionally, this method identifies illegal coal mines and seawater contamination, aiding in coal industry management, etc. These findings lay a crucial foundation for assessing the environmental impact of coal mining activities and proposing solutions for ecological restoration in coastal areas where diverse and complex socio-economic activities occur.