Scaling relationship between CO pollution and population size over major US metropolitan statistical areas

As the world's population is projected to reach 9.7 billion by 2050, large and small cities will continue to expand. There are few studies investigating how the size of cities affects air pollution. Carbon monoxide (CO), a precursor of ozone and a by-product of incomplete combustion, is a common air pollutant. The major sources of CO in the US urban areas are motor vehicles. Here we examined the scaling relation of CO concentrations over major US metropolitan statistical areas (MSAs) using Measurement Of Pollution in The Troposphere (MOPITT) surface CO retrievals and National Emissions Inventory (NEI) data. We found significant power-law scaling relationships between CO and population (r2 of 0.30 for MOPITT average CO concentration and r2 of 0.71 for NEI total CO emission). We found decreasing CO trends from 2000 to 2015 using MOPITT and EPA CO ground measurements. Sublinear scaling relationships (scaling coefficient β < 1) suggest that larger MSAs are more combustion-efficient in terms of CO emissions. We found a weaker scaling relation and smaller scaling coefficient from MOPITT CO concentrations than from NEI total CO emission data. This pattern may be attributed to the differences between the two CO datasets: annual average of monthly MOPITT CO concentration at 1° by 1° spatial resolution versus the NEI annual CO emissions compiled from emission inventories and estimated from mobile source emissions models. Future research is needed to investigate the capability of using satellite observations to study scaling relations between air pollutants and population