As expected, on-road emissions are consistently large sources of both greenhouse gas and NOx emissions in both cities across all seasons. This sector also exhibits the most pronounced diurnal and day of week behavior in the inventories, so that air quality benefits may depend on what time of day these impacts are most relevant to human health. However, we find that dominant emissions sectors are not the same between cities, and have strong seasonal variation. For example, the electrical generation and residential sectors dominate CO2 emissions (~50% of total) in Boston during the winter, but only contribute 5% and 16% respectively of urban NOx emissions for the same season. In contrast, in Salt Lake City wintertime CO2 emissions are dominated by on-road gasoline and diesel vehicle emissions (~50% of total), which also contribute 60% of urban NOx emissions. We connect this inventory analysis to ambient measurements of CO2 and NOx in our study areas to confirm the importance of various emission sectors on ambient air quality and greenhouse gas concentrations over space and time. For example, the weekday-weekend difference in Boston NOx emissions (~20%), dominated by changes in diesel emissions (30%), is consistent with ambient measurements of NOx enhancements. Likewise, ambient measurements of NOx and CO2 in Boston and Salt Lake City sorted by air mass history show evidence of spatial heterogeneity in the emissions sources consistent with our prior inventories. For example, northwest of Boston the integrated NOx:CO2 emission ratio is substantially higher (1.35 g/kg) than southwest of the city (1.05 g/kg), due to the spatial distribution of electricity generating units within the city. We leverage these new observational constraints in combination with our bottom-up inventory analysis to infer how air quality benefits from specific greenhouse-gas focused interventions may depend on the intersection of neighborhood-level emissions, population spatial distribution, and prevailing meteorology.

