Spatiotemporal Distribution Characteristics and Concentration Prediction of Pollutants in Open-Pit Coal Mines

Abstract

Open-pit coal mining is characterized by multiple pollution sources, diverse types, and extensive affected areas, leading to complex air pollution with wide diffusion. Traditional fixed monitoring methods cannot address these limitations. Taking a coal mine in Xinjiang as a case study, this study developed a drone-mounted mobile atmospheric monitoring system, focusing on nitrogen dioxide (NO2) and suspended particulate matter (PM2.5 and PM10) to explore their distribution, diffusion patterns, and influencing factors. The results show distinct seasonal pollutant characteristics: NO2 and ozone (O3) dominate in summer, while particulate matter prevails in winter. The temporal distribution exhibits a bimodal pattern, with high levels in the early morning and evening hours. Spatially, higher pollutant concentrations accumulate vertically below ground level, while lower levels are observed above it. Horizontally, elevated concentrations are found along northern transport corridors; however, these levels become more uniform at greater heights. A spatiotemporal prediction model integrating convolutional neural network (CNN) and long short-term memory (LSTM) network was successfully applied to real-time pollutant prediction in open-pit coal mining areas. This study provides a reliable mobile monitoring solution for open-pit coal mine air pollution and offers valuable insights for targeted pollution control in similar mining areas.