Multi-city assessment of fine-scale urban building and tree impacts on air quality

This study leverages 1-metre resolution NAIP (National Agriculture Imagery Program) aerial imagery and Sentinel-5P TROPOMI (Tropospheric Monitoring Instrument) data to quantify fine-scale urban spatial patterns and nitrogen dioxide (NO2) concentrations across 22 major U.S. cities. Using Boosted Regression Trees, we examined the relationships between the spatial patterns of two key urban entities (buildings and trees) and NO2 levels across seasons, geographic regions, and climate zones. Our results show that building size and coverage explain significantly more variation in NO2 than tree patterns (p < 0.05), suggesting that pollution mitigation strategies should prioritize building design, placement, and density alongside urban greening. Despite widespread urban forest management initiatives, the effectiveness of trees in reducing NO2 appears limited unless planted in more continuous canopy patches. We also find that high-resolution imagery enhances the characterization of urban features, and that a larger spatial extent (1,500 m buffer) captures stronger associations with NO2 than a smaller one (500 m). This highlights the importance of holistic, regional-scale planning approaches that go beyond parcel-level interventions. In contrast, we observed no consistent seasonal influence on NO2–building or NO2–tree relationships.