Relating satellite NO2 tropospheric columns to near-surface concentrations: implications from ground-based MAX-DOAS NO2 vertical profile observations

Abstract

Given the significant environmental and health risks associated with near-surface nitrogen dioxide (NO₂), machine learning is frequently employed to estimate near-surface NO₂ concentrations (S_NO2) from satellite-derived tropospheric NO₂ column densities (C_NO2). However, data-driven methods often face challenges in explaining the complex relationships between these variables. In this study, the correlation between C_NO2 and S_NO2 is examined using vertical profile observations from China’s MAX-DOAS network. Cloud cover and air convection substantially weaken (R = −0.68) and strengthen (R = 0.71) the C_NO2-S_NO2 correlation, respectively. Meteorological factors dominate the correlation (R² = 0.58), which is 31% stronger in northern regions than in the southwest. Additionally, anthropogenic emissions impact S_NO2, while topographical features shape regional climate patterns. At the Chongqing site, the negative impacts of unfavorable meteorological conditions, high emissions, and basin topography lead to significant contrasts and delays in daily C_NO2 and S_NO2 variations. This study enhances understanding of the spatial and temporal dynamics and influencing mechanisms of C_NO2 and S_NO2, supporting improved air quality assessments and pollution exposure evaluations.