@misc{tong_yu_kan_2009, title={Ozone exposure and mortality}, volume={360}, number={26}, journal={New England Journal of Medicine}, author={Tong, D. Q. and Yu, S. C. and Kan, H. D.}, year={2009}, pages={2788–2788} }
 @article{tong_muller_kan_mendelsohn_2009, title={Using air quality modeling to study source-receptor relationships between nitrogen oxides emissions and ozone exposures over the United States}, volume={35}, ISSN={["1873-6750"]}, DOI={10.1016/j.envint.2009.06.008}, abstractNote={Human exposure to ambient ozone (O3) has been linked to a variety of adverse health effects. The ozone level at a location is contributed by local production, regional transport, and background ozone. This study combines detailed emission inventory, air quality modeling, and census data to investigate the source–receptor relationships between nitrogen oxides (NOx) emissions and population exposure to ambient O3 in 48 states over the continental United States. By removing NOx emissions from each state one at a time, we calculate the change in O3 exposures by examining the difference between the base and the sensitivity simulations. Based on the 49 simulations, we construct state-level and census region-level source–receptor matrices describing the relationships among these states/regions. We find that, for 43 receptor states, cumulative NOx emissions from upwind states contribute more to O3 exposures than the state's own emissions. In-state emissions are responsible for less than 15% of O3 exposures in 90% of U.S. states. A state's NOx emissions can influence 2 to 40 downwind states by at least a 0.1 ppbv change in population-averaged O3 exposure. The results suggest that the U.S. generally needs a regional strategy to effectively reduce O3 exposures. But the current regional emission control program in the U.S. is a cap-and-trade program that assumes the marginal damage of every ton of NOx is equal. In this study, the average O3 exposures caused by one ton of NOx emissions ranges from − 2.0 to 2.3 ppm-people-hours depending on the state. The actual damage caused by one ton of NOx emissions varies considerably over space.}, number={8}, journal={ENVIRONMENT INTERNATIONAL}, author={Tong, Daniel Q. and Muller, Nicholas Z. and Kan, Haidong and Mendelsohn, Robert O.}, year={2009}, month={Nov}, pages={1109–1117} }
 @article{tong_mathur_kang_yu_schere_pouliot_2009, title={Vegetation exposure to ozone over the continental United States: Assessment of exposure indices by the Eta-CMAQ air quality forecast model}, volume={43}, ISSN={["1873-2844"]}, DOI={10.1016/j.atmosenv.2008.09.084}, abstractNote={The main use of air quality forecast (AQF) models is to predict ozone (O3) exceedances of the primary O3 standard for informing the public of potential health concerns. This study presents the first evaluation of the performance of the Eta-CMAQ air quality forecast model to predict a variety of widely used seasonal mean and cumulative O3 exposure indices associated with vegetation using the U.S. AIRNow O3 observations. These exposure indices include two concentration-based O3 indices, M7 and M12 (the seasonal means of daytime 7-h and 12-h O3 concentrations, respectively), and three cumulative exposure-based indices, SUM06 (the sum of all hourly O3 concentrations ≥ 0.06 ppm), W126 (hourly concentrations weighed by a sigmoidal weighting function), and AOT40 (O3 concentrations accumulated over a threshold of 40 ppb during daylight hours). During a three-month simulation (July–September 2005), the model over predicted the M7 and M12 values by 8–9 ppb, or a NMB value of 19% and a NME value of 21%. The model predicts a central belt of high O3 extending from Southern California to Middle Atlantic where the seasonal means, M7 and M12 (the seasonal means of daytime 7-h and 12-h O3 concentrations), are higher than 50 ppbv. In contrast, the model is less capable of reproducing the observed cumulative indices. For AOT40, SUM06 and W126, the NMB and NME values are two- to three-fold of that for M7, M12 or peak 8-h O3 concentrations. The AOT40 values range from 2 to 33 ppm h by the model and from 1 to 40 ppm h by the monitors. There is a significantly higher AOT40 value experienced in the United States in comparison to Europe. The domain-wide mean SUM06 value is 14.4 ppm h, which is about 30% higher than W126, and 40% higher than AOT40 calculated from the same 3-month hourly O3 data. This suggests that SUM06 and W126 represent a more stringent standard than AOT40 if either the SUM06 or the W126 was used as a secondary O3 standard. Although CMAQ considerably over predicts SUM06 and W126 values at the low end, the model under predicts the extreme high exposure values (>50 ppm h). Most of these extreme high values are found at inland California sites. Based on our analysis, further improvement of the model is needed to better capture cumulative exposure indices.}, number={3}, journal={ATMOSPHERIC ENVIRONMENT}, author={Tong, Daniel Q. and Mathur, Rohit and Kang, Daiwen and Yu, Shaocai and Schere, Kenneth L. and Pouliot, George}, year={2009}, month={Jan}, pages={724–733} }