Monitoring and Analysis of Relationship between Vegetation Cover and Air Pollution Using Sentinel and MODIS Satellite Data in Mazandaran Province

Urban vegetation mitigates air pollution, thereby enhancing environmental quality and human health. This study examines the relationship between vegetation cover (NDVI) and air pollution indicators (O₃, CO, AOD) in Mazandaran Province, Iran, from 2019 to 2024. Data were obtained from Sentinel-5P (O₃, CO), MODIS (AOD), Sentinel-2 (NDVI), and 60 ground-based and virtual monitoring stations across Amol, Babol, Qaemshahr, Babolsar, and Sari. Analyses included R², Pearson, Kendall Tau, Spearman correlations, mutual information (MI), distance correlation (DC), HSIC, principal component analysis (PCA), Bland–Altman, and K-Means clustering. Spatiotemporal results indicated O₃ concentrations of 0.132–0.145 mol/m², CO of 0.02–0.037 mol/m², and AOD of –0.62 to 0.97, peaking in urban areas and along with major highways. NDVI ranged from –0.26 to 0.46, declining from 2022 (–0.08 to 0.26 in 2024), with higher values in Hyrcanian forests. A consistently strong negative NDVI–AOD relationship was observed (R² = 0.552–0.644; Pearson’s r = –0.743 to –0.803; Kendall’s τ = –0.554 to –0.635; Spearman’s ρ = –0.738 to –0.820; Mutual Information = 0.832–0.989; PC1 = 87.1–90.1%), while Bland–Altman analysis indicated biases of 0.07–0.12 with strongest agreement in 2019–2021 (LoA: –0.12 to 0.26). K-Means clustering revealed three clusters per year, with NDVI thresholds of 0.15–0.20 associated with lower AOD (0.15–0.20) in vegetated areas versus higher AOD (0.19–0.25) in sparsely vegetated zones. These results highlight vegetation’s critical role in aerosol mitigation and emphasize the importance of targeted vegetation management and continuous high-resolution satellite monitoring (MODIS, Sentinel-2) to improve air quality in Mazandaran Province.