Geostationary Observations Reveal Widespread Underestimation of Current Ambient NO2 from Strong Diurnal Variability

Ambient nitrogen dioxide (NO2) is a short-lived but harmful pollutant, and its accurate estimation remains challenging. Traditional polar-orbiting satellites cannot capture its diurnal variability due to their once-daily early-afternoon overpasses. Thus, we integrate GEMS geostationary satellite observations with GEOS-CF model simulations, applying a space-time machine learning model to generate the first gapless, hourly surface NO2 dataset across China with full 24-hour coverage. A distinct bimodal pattern in surface NO₂ levels is observed, exhibiting a large diurnal amplitude (71%) with a morning peak, midday decline, and stronger evening rebound. Daytime levels are driven by photochemical activity and enhanced vertical mixing associated with higher boundary layer heights (BLH), whereas nighttime concentrations are influenced by a shallow BLH and higher humidity that suppress mixing and chemical reactions. This contrast aligns with the much lower diurnal variability in tropospheric NO2 (36%). Incorporating GEMS hourly observations further improves estimates of daily (24-hour) mean surface NO2 (R2 = 0.88), particularly in rural regions, compared to results derived from polar-orbiting TROPOMI satellites. TROPOMI shows widespread underestimation of daily NO2 exposure across 85% of the domain due to its once-daily overpass in the early afternoon. Thus, it often misses high-pollution events and underreports days exceeding the World Health Organization’s daily Interim Target 1 (120 µg m-3) and the air quality guideline (25 µg m-3), respectively. Consequently, using this polar-orbiting-derived estimate results in 2,200 fewer premature deaths from acute ambient NO2 exposure, 11,500 fewer deaths from chronic exposure, and 850 fewer asthma cases compared with estimates based on geostationary GEMS observations in 2023. These findings underscore the urgent need to capture the full diurnal dynamics of ambient NO2 for accurate exposure assessment and more effective environmental health policies.