Comparison of spatial and temporal variations in p-PAH, BC, and p-PAH/BC ratio in six US counties

Abstract

Abstract An ambient air monitoring campaign was performed in six counties (Sacramento, CA; Maricopa, AZ; Anoka, MN; Jefferson, KY; Harris, TX; and Pinellas, FL) between January 2008 and September 2009. The purpose of this paper is to compare the spatial and temporal variability of black carbon (BC) and particle-bound polycyclic aromatic hydrocarbons (p-PAHs), across these counties using continuous monitoring instruments – an Aethalometer and a Photoelectric Aerosol Sensor reporting in units of μg m −3 and fA, respectively. We explored temporal trends in these measurements to assess the potential impact of local combustion sources on air quality. Median BC concentrations ranged from 0.13 to 0.53 μg m −3 ; and median p-PAH values ranged from 0.31 to 4.18 fA. Hourly BC and p-PAH were elevated during morning rush hour and rapidly decreased later in the morning. Nighttime increases in BC and p-PAH were also observed in most counties. Diurnal patterns of BC and p-PAH were different on weekdays compared to weekends. Profiles of hourly ratios of p-PAH/BC in combination with meteorological data can provide insight into potential sources across the sites. Hourly ratios of p-PAH/BC which peaked during early morning and late afternoon hours suggest a dominating contribution of motor vehicle sources in four of the six counties. In two counties, hourly ratios remained elevated for several hours after rush hour and did not show a distinctive peak suggesting additional sources of BC and p-PAH. Such profiles were seen in both Jefferson KY and Harris TX, and may be attributed to coal combustion, petro-chemical industry and shipping activities, respectively. These results suggest that measurements of BC and p-PAH, combined with meteorological information and emission data are potentially useful to identify combustion sources impacting air quality. More research combining BC and p-PAH measurements with detailed source apportionment data is needed to more fully evaluate the utility of these real-time measures.