Spatiotemporal Patterns and Attribute Analysis of NO₂ concentrations in U.S. Cities based on satellite observations

Keywords: Air Pollution; Spatiotemporal Pattern; Sentinel-5P TROPOMI
Abstract Type: Poster Abstract

Abstract

Urban NO₂ pollution poses significant public health and environmental risks, particularly in areas experiencing rapid urbanization and industrial activity. Existing research often overlooks smaller urban areas, limits cross-city comparisons of NO₂ drivers and needs more analysis in understanding how natural and anthropogenic factors interact to influence NO₂ pollution trends across diverse urban settings. This study uses Sentinel-5P TROPOMI data to analyze NO₂ levels in multiple U.S. cities from 2019 to 2023, identifying the natural and anthropogenic factors behind intercity differences. Our findings reveal that 86% of cities experienced NO₂ increases, with annual growth rates varying widely (0.17–221.5 × 10⁻⁸ mol/m² every year). Temporal patterns reveal higher NO₂ in winter and on weekdays, lower levels in summer and on weekends. Spatial clustering highlights the northeastern U.S. and large metropolitan areas as pollution hotspots but with modest increases, while smaller cities show more substantial NO₂ growth over time. Pearson correlation analysis underscores significant relationships between NO₂ levels and key variables: elevation and temperature correlate negatively with NO₂, while wind speed, GDP, proximity to power plants, and road density correlate positively. A random forest regression model explains 85% of the spatial variability in annual average NO₂ concentrations in 2019, with natural factors accounting for 54.4% of the variation and anthropogenic factors for 45.6%. However, annual NO₂ increases from 2019 to 2023 are more strongly driven by anthropogenic influences (52.4%). These insights are crucial for policy interventions targeting NO₂ emissions across urban areas to protect public health.