Atmospheric chemistry-transport models are often used to determine the marginal impact of emissions on air quality and public health, but the uncertainties related to modeling assumptions are rarely formally characterized from the perspective of health impact calculations. Aviation emissions are of increasing interest, due to the projected growth in aviation transport coupled with the decreasing emissions from other anthropogenic sources, but the geographic scale and resolution necessary to accurately characterize public health impacts from aviation are unclear, given the unique spatio-temporal pattern of aviation emissions. In this study, we estimate the incremental contribution of commercial aviation emissions during landing and takeoff (LTO) cycles from three U.S. airports - Atlanta's Hartsfield-Jackson, Chicago's O'Hare, and Providence's T.F. Green - to fine particulate matter (PM2.5) levels using the Community Multiscale Air Quality model (CMAQ), a comprehensive chemistry-transport air quality model. To evaluate the significance of model resolution and geographic scale, we ran a one-atmosphere version of CMAQ at 36-, 12-, and 4-km resolutions, and calculated the total population health risks at various distances from each airport. in spite of significant differences in maximum concentrations attributable to aviation emissions, total population health risks over the entire model domain were largely unaffected by model resolution, with a 2% difference in estimated mortality risks between the 36-km and 12-km resolution outputs. Analyses of model scale indicated that a 108 x 108 km domain centered on the airport captured most population exposure for primary components of PM2.5, but total public health risks were dominated by populations at greater distances from the airport, given the contribution from secondary ammonium sulfate and nitrate, with 28-35% of health risks occurring more than 300 km from the airports. Our findings provide insight about the model resolution and geographic scales necessary for population risk assessment from LTO emissions, and demonstrate the robustness of risk-based prioritization across multiple grid resolutions. (C) 2011 Elsevier Ltd. All rights reserved.
