Spatiotemporal Characteristics and Influencing Factors of Dust Pollution in Mining Areas: A Quantitative Approach Based on Correlation and Statistical Models

In response to ecological degradation caused by uncontrolled dust emissions from open-pit mines, this study selected the Hequ open-pit coal mine as the study area and established a monitoring system to collect data on TSP, PM10, PM2.5, and environmental indicators across three zones: the mining pit, the main haul road, and the coal yard. The necessity of zoning was validated using the least significant difference (LSD) method. Pollutant correlations were examined using the individual air quality index (IAQI), Pearson correlation matrix analysis, and grey relational analysis. Univariate models, multiple linear regression (MLR), and principal component analysis–multiple linear regression (PCA–MLR) were applied to quantitatively analyze dust evolution patterns and the influence of environmental factors, with model accuracy verified by the mean relative error (MRE) method. The results showed significant differences in dust concentrations among the three zones. Dust concentrations of all particle sizes in the mining pit and coal yard exceeded the secondary standard limit, whereas those on the haul road only exceeded the primary limit, with pollution intensity ranked as mining pit > coal yard > haul road and PM2.5 identified as the core pollutant in all zones. Linear relationships were significant in univariate models, and multivariate fitting outperformed univariate fitting, with MLR prediction accuracy ranked as coal yard (3.02%) > haul road (9.46%) > mining pit (10.75%). In the mining pit, TSP and PM10 exhibited a strong positive correlation with atmospheric pressure, while PM2.5 showed a strong negative correlation with relative humidity. On the haul road, all particle size fractions displayed strong negative correlations with temperature and wind speed. In the coal yard, only a strong negative correlation with temperature was observed. The PCA–MLR model improved prediction accuracy by 56.63% and 13.41% compared to the direct MLR model. Comprehensive analysis indicates that the atmospheric environment of the Hequ open-pit mine urgently requires proactive restoration measures to optimize the sustainability of the ecological environment.