Vertical Structure Matters: Improving Urban PM2.5 Assessment Using Lidar at Background and Traffic-Influenced Sites

Understanding particulate pollution in Eastern Mediterranean (EM) cities is challenging due to the combined influence of local urban emissions, marine aerosols, and long-range transported desert dust. Conventional surface-based measurements often fail to detect lofted dust layers, while satellite-derived Aerosol Optical Depth (AOD) provides only column-integrated information, limiting its ability to represent near-surface PM_2.5 (fine particulate matter 2.5 micrometres or less in diameter) concentrations.  Here, we combine five years of ground-based lidar observations with high resolution satellite AOD retrievals, PM_2.5 measurements and meteorological data over the EM to investigate aerosol layering, source contributions, and column-to-surface relationships across three contrasting urban environments: regional background, urban traffic and semi-indoor sites located along highway/railroad. Lidar profiling identifies ten distinct aerosol layering types, from shallow anthropogenic layers to deep mixed structures composed of desert dust, marine aerosols, and urban pollution. We find that the AOD-PM_2.5 relationship is strongly regime-dependent, and the degree of column-surface coupling varies sharply across the three urban environments. Machine-learning models that incorporate vertical lidar features significantly improve PM_2.5 predictions across all sites, outperforming models without vertical information. Overall, our results demonstrate that reliable urban PM_2.5 assessment requires explicit consideration of vertical aerosol structure. Integrating lidar-derived features enhances the interpretation of satellite AOD and improves urban exposure estimates in complex EM atmospheres.