Air pollution and urban form: Evidence from satellite data

We consider here the extent to which urban form (the physical layout of a city, e.g., sprawl versus density) can impact urban air pollution. Using available satellite-based measurements for 89 urban areas throughout the globe, we compare air pollution (nitrogen dioxide) and spatial patterns of built-up areas (urban contiguity and compactness). Linear regression modeling reveals that, as expected, air pollution is proportional to city size (urban population and land area), with larger cities having worse pollution. Consistent with Environmental Kuznets Curve (EKC) theory, the relationship between air pollution and per capita GDP is an inverse "U" shape. Rising income increases pollution levels for per capita GDP less than US$23,000 but decreases pollution for per capita GDP greater than US$23,000. Urban form - specifically, a combined measure of urban contiguity and compactness - has a modest but statistically significant impact on air pollution levels: at mean values for independent variables, 1-standard-deviation increase in contiguity/compactness yields a 14% reduction in air pollution concentrations. This finding indicates that striving to increase the contiguity/compactness of a city may reduce air pollution, an encouraging result. Additional work is necessary to confirm these initial findings.