Quantitative assessment of urban-rural spatiotemporal heterogeneity in air pollutants using GEE multi-source data across the Anhui province, China

Abstract

Air pollution in urban areas has garnered considerable attention in recent years. However, from an equity perspective, the air pollution situation in non-urban areas warrants more in-depth investigation. Limited research still exists on the non-urban air pollution. This study examined the urban-rural spatiotemporal heterogeneity of NO₂, CO, and O₃ in the Anhui Province from 2019 to 2023. Using Google Earth Engine (GEE), ArcGIS, and R language, the analysis integrated TROPOMI column concentration data with natural and social influencing factors. The results revealed the following: (1) The concentrations of NO₂, CO, and O₃ exhibit a declining trend during the annual plum rain season. NO₂ and CO concentrations show inter-annual fluctuating decreases, whereas O₃ demonstrates a fluctuating increase. NO₂ and CO concentrations are lowest in summer and increase synchronously during autumn and winter. The highest correlation coefficient between NO₂ and O₃ concentrations occurs in spring, at −0.946. (2) From January to March each year, the maximum concentrations of CO and O₃ are more likely to occur in non-urban built-up areas. Compared to 2019, the average area proportions of increased CO and NO₂ concentrations in urban built-up areas across all four seasons in 2023 are 8.91% and 4.93%, respectively, significantly lower than those in non-urban built-up areas (91.08% and 95.05%). Except for summer, O₃ concentrations show an increasing trend throughout the entire province. (3) The standard deviations of the multi-year average concentrations of CO, NO₂, and O₃ among the 16 prefecture-level cities are 0.002, 2.19 × 10⁻⁵, and 0.0027, respectively. This suggests that the variation in NO₂ pollution among cities is relatively small, while the spatial imbalance of O₃ pollution is pronounced, with the highest average O₃ concentrations observed in cities in northern Anhui. (4) The correlation coefficients between each air pollutant and the perimeter-area fractal dimension of water, forests, and buildings exceed 0.64, and the correlation coefficients with the aggregation index of forests and buildings exceed 0.58. These findings indicate that the complexity and dispersion of landscape patterns resulting from human disturbance may have a significant impact on air pollution levels.