Neighborhood-scale NO2 variations and enhanced capability of human exposure assessment

This study estimated ambient long-term average NO₂ concentrations using TROPOspheric Monitoring Instrument (TROPOMI) tropospheric column NO₂ data and land use parameters in California, U.S. for the years 2018–2019. Exploiting unprecedentedly high spatial resolution of TROPOMI NO₂ (3.5×7 km prior to August 6, 2019 and 3.5×5.5 km thereafter) and the point statistics function implemented in ArcGIS (Environmental Systems Research Institute, ESRI), the NO₂ concentration estimates were downscaled to 500 m, which enabled the neighborhood-scale exposure assessment of ambient NO₂. Our satellite-land use hybrid regression model demonstrated cross-validation R²= 0.76, mean absolute error (MAE)= 1.95 ppb, and root mean squared error (RMSE)= 2.51 ppb in a comparison between site-specific average measured and estimated NO₂ concentrations. These high-resolution NO₂ concentration estimates enhanced the capability of human exposure assessment, enabling (1) the evaluation of ground NO₂ monitors to represent population exposures and (2) the attribution of micro-level NO₂ exposures. When measured NO₂ concentrations were compared to population-weighted NO₂ concentrations, calculated by using the satellite-based NO₂ estimates, in each county, the differences in NO₂ concentrations (i.e., population-weighted average NO₂ – arithmetic average NO₂ measurements) ranged from -38.6% (San Bernardino) to 82.2% (Humboldt). Though both negative and positive differences represented exposure errors without considering the spatial co-variations of NO₂ and populations, monitor-based NO₂ higher than the population-weighted NO₂ demonstrated the overestimation of population NO₂ exposures and was at least protective with current NO₂ monitoring locations in the counties. However, the opposite was detrimental, while underestimating population NO₂ exposures and likely not motivating NO₂ mitigation efforts. In addition, the high-resolution NO₂ estimates were further overlaid with parcel-level property data in Los Angeles County to attribute the spatial variation of NO₂ exposures to that of the property types. The micro-level NO₂ hotspots were identified at high-density residential complexes such as (high-rise) apartments. When the traffic impacts on NO₂ were adjusted, the NO₂ hotspots at the residential complexes still remained. This finding may suggest residential complexes as an emerging source type of NO₂ due to emissions from boilers (space heating and hot water) and other indoor-to-outdoor ventilation systems.