Spatiotemporal covariability between air pollution and meteorological variables over Khyber Pakhtunkhwa, Pakistan.

This study analyzed spatiotemporal covariability of O3, SO2, NO2, CO, and PM2.5 with meteorological variables (rain precipitation rate, specific humidity, pressure, temperature, wind speed, latent heat flux, and solar radiation) using satellite data in Khyber Pakhtunkhwa province, Pakistan. Inverse Distance Weighted interpolation, ordinary least square regression, Pearson correlation, Generalized Linear, and Generalized Additive models were applied. Results revealed highest annual average pollutants as; NO₂ (3.87 ± 0.73) × 1015 molecules/cm2, PM2.5 (37.91 ± 17.75) µg/m3, SO2 (6.81 ± 8.27) × 1014, CO (1.34 ± 0.52) × 1018 molecules/cm2, and O3 (7.73 ± 0.10) × 1018 molecules/cm2. Seasonally NO2 peaked in summer and spring, SO₂ in autumn, CO in spring, PM2.5 in winter while O₃ in spring with minor seasonal variations. Annual spatial distribution of SO2, PM2.5, and CO were highest in central and southern areas while O3 in the central and NO2 in the central and southeastern. Wind speed was negatively correlated with NO2 annually and in winter, summer, and autumn. Temperature positively influenced NO2 and PM2.5 annually and seasonally, while O3 positively correlated with rain and specific humidity but negatively with solar radiation and temperature in spring. In autumn, O3 exhibited a positive association with rain and negative with solar radiation. SO2 indicated positive correlations with solar radiation annually and temperature in spring, while CO showed weak associations except for a positive correlation with specific humidity in summer. GAM models slightly better captured pollutant dynamics by explaining both linear and nonlinear relationships. These findings provide crucial insights for targeted air quality management strategies and pollutant mitigation.