Abstract for 2017 HEI Annual Conference
Impacts of Emission Changes on Air Quality and Acute Health Effects in the Southeast 1993–2013
Armistead (Ted) Russell, Paige Tolbert, Lucas Henneman, Joseph Abrams, Cong Liu, Mitch Klein, Jim Mulholland, Stefanie Sarnat, Yongtao Hu, Howard Chang, Talat Odman, Matt Strickland, Huizhong Shen, and Abiola Lawal
Georgia Institute of Technology, Atlanta, GA, United States of America
Emory University, Atlanta, GA, United States of America
Background Researchers at Georgia Tech and Emory used air quality modeling and statistical techniques to determine the impact of regulatory policies on health outcomes in the southeastern United States. Effects of multiple national and state rules promulgated between 1993 and 2013 were investigated. Various strategies and tools were employed to separate changes in emissions, air quality, and health, notably a counterfactual approach that compared scenarios in which the only difference was the occurrence of an intervention.
Methods Long-term (1999-2013) records of daily concentrations of ambient air pollutants (ozone, NOx, SO2, CO, PM2.5, and components of PM2.5: sulfate, nitrate, ammonium, OC, and EC) were detrended to remove the fluctuations attributable to meteorological variability. Linear statistical models were applied to determine empirical sensitivities of ambient concentrations to changes in emissions from electricity generating unit (EGU) and mobile sources in Atlanta, GA. Empirical sensitivities were used in addition to similar sensitivities developed using the CMAQ-DDM to estimate uncertainty, and were combined with counterfactual–i.e. assuming no intervention–emissions estimates to create counterfactual daily ambient air pollution concentrations. Daily contrasts of observed and counterfactual ambient concentrations were utilized in conjunction with parameter estimates obtained from multi-pollutant Poisson time-series models to estimate the excess cardiorespiratory emergency department (ED) visits that would have occurred in Atlanta from 1999-2013 in the absence of regulatory actions.
Results Annual average concentrations of NOx, SO2, and CO have all fallen by at least 50% since 1999, roughly matching estimated changes in emissions. EGU regulations have had a larger impact on PM2.5 than mobile; EGU emissions changes were found to reduce mean PM2.5 in 2013 by 7.1 µg m-3 (and 3.1 µg m-3 for mobile). EGU reductions were found to decrease mean ozone by 3.2 ppb in 2013, and mobile regulations were found to increase mean ozone by 1.8 ppb. In ozone, the largest changes attributable to control programs were observed in the high and low quantiles; EGU emissions reductions were more associated with decreasing summer values, and mobile reductions with increasing winter values. Preliminary health models over the period from 1999-2013 suggest that the greatest reductions in ED visits in Atlanta attributable to changes in emissions from EGU and mobile sources occurred in 2012 and 2013, in which it was estimated that there would have been 5.9% more respiratory disease ED visits, 16.5% more asthma ED visits, 2.3% more cardiovascular disease ED visits, and 2.6% more congestive heart failure ED visits in the absence of these emission changes.
Conclusions The timing of emissions reductions from mobile and utility sources corresponds with documented implementation of specific regulatory actions. These emission changes have led to improvements in air quality, as demonstrated through a variety of different modeling techniques. While estimated reductions in ED visits attributed to the regulatory actions varied across the regulations analyzed, reductions in ED visits for all regulatory actions combined were substantial.