@article{liu_zhang_2013, title={Understanding of the formation mechanisms of ozone and particulate matter at a fine scale over the southeastern US: Process analyses and responses to future-year emissions}, volume={74}, ISSN={["1873-2844"]}, DOI={10.1016/j.atmosenv.2013.03.057}, abstractNote={Ozone (O3) and fine particle (PM2.5) formation over the southeastern U.S. are of a major concern due to high emissions of precursors and special weather conditions that are conducive to their formation. In this study, the Community Multiscale Air Quality (CMAQ) modeling system is applied to simulate the formation of major air pollutants over an area in the southeastern U.S. at a 4-km horizontal grid resolution for January, April, July, and October in 2002 and 2018. Model performance evaluation shows an overall satisfactory performance for O3 in all months and for PM2.5 in January and October at rural sites and in January, April, and October at urban sites. Large underpredictions in PM2.5 concentrations occur in April and July at rural sites and in July at urban sites, because of biases in meteorological predictions and underestimation of emissions of precursors. The model performance at 4-km in terms of O3, PM2.5 and PM2.5 components show some improvements but overall are not always better than that at 12-km. O3 chemistry is VOC-limited in urban areas and NOx-limited over the west of the mountain regions and the southern Georgia throughout the year, and VOC-limited over the rest of areas in January but NOx-limited in other months. Among all photochemical indicators examined, PH2O2/PHNO3 and O3/NOy are the most robust indicators. The domain is NH3-rich or neutral in all months, indicating a high potential for NH4NO3 formation and the sensitivity of PM2.5 formation to the emissions of SO2, NOx, and NH3. Surface O3 is accumulated primarily through vertical transport in urban, rural and coastal areas and both horizontal and vertical transport in mountain regions and produced via gas-phase chemistry at non-urban sites during daytime. The loss of O3 is attributed to gas-phase chemistry via NO titration in urban areas, and dry deposition and transport processes in rural and mountain areas. PM2.5 is produced by primary emissions and PM processes and lost through vertical and horizontal transport in urban areas. The combined effects of transport, emissions, and PM processes influence PM concentrations in other areas. The 2018 simulations project a decrease in PM2.5 concentrations and an improvement in visibility over almost the entire domain, slight decreases in O3 mixing ratios in urban areas in July and most non-urban areas in April and October but large increases in the rest of areas in other months, and a decrease in total N deposition fluxes in most areas except for central and eastern North Carolina and northern Georgia. The development of integrated emission control strategies should consider region-specific seasonality and differences in the responses of O3, PM2.5, visibility, and nitrogen deposition.}, journal={ATMOSPHERIC ENVIRONMENT}, author={Liu, Xiao-Huan and Zhang, Yang}, year={2013}, month={Aug}, pages={259–276} } @article{zhang_liu_liu_jacobson_mcmurry_yu_yu_schere_2010, title={A comparative study of nucleation parameterizations: 2. Three-dimensional model application and evaluation}, volume={115}, journal={Journal of Geophysical Research. Atmospheres (Online)}, author={Zhang, Y. and Liu, P. and Liu, X. H. and Jacobson, M. Z. and McMurry, P. H. and Yu, F. Q. and Yu, S. C. and Schere, K. L.}, year={2010} } @article{zhang_liu_liu_pun_seigneur_jacobson_wang_2010, title={Fine scale modeling of wintertime aerosol mass, number, and size distributions in central California}, volume={115}, journal={Journal of Geophysical Research. Atmospheres (Online)}, author={Zhang, Y. and Liu, P. and Liu, X. H. and Pun, B. and Seigneur, C. and Jacobson, M. Z. and Wang, W. X.}, year={2010} } @article{liu_zhang_olsen_wang_do_bridgers_2010, title={Responses of future air quality to emission controls over North Carolina, Part I: Model evaluation for current-year simulations}, volume={44}, ISSN={["1873-2844"]}, DOI={10.1016/j.atmosenv.2010.04.002}, abstractNote={The prediction of future air quality and its responses to emission control strategies at national and state levels requires a reliable model that can replicate atmospheric observations. In this work, the Mesoscale Model (MM5) and the Community Multiscale Air Quality Modeling (CMAQ) system are applied at a 4-km horizontal grid resolution for four one-month periods, i.e., January, June, July, and August in 2002 to evaluate model performance and compare with that at 12-km. The evaluation shows skills of MM5/CMAQ that are overall consistent with current model performance. The large cold bias in temperature at 1.5 m is likely due to too cold soil initial temperatures and inappropriate snow treatments. The large overprediction in precipitation in July is due likely to too frequent afternoon convective rainfall and/or an overestimation in the rainfall intensity. The normalized mean biases and errors are −1.6% to 9.1% and 15.3–18.5% in January and −18.7% to −5.7% and 13.9–20.6% in July for max 1-h and 8-h O3 mixing ratios, respectively, and those for 24-h average PM2.5 concentrations are 8.3–25.9% and 27.6–38.5% in January and −57.8% to −45.4% and 46.1–59.3% in July. The large underprediction in PM2.5 in summer is attributed mainly to overpredicted precipitation, inaccurate emissions, incomplete treatments for secondary organic aerosols, and model difficulties in resolving complex meteorology and geography. While O3 prediction shows relatively less sensitivity to horizontal grid resolutions, PM2.5 and its secondary components, visibility indices, and dry and wet deposition show a moderate to high sensitivity. These results have important implications for the regulatory applications of MM5/CMAQ for future air quality attainment.}, number={20}, journal={ATMOSPHERIC ENVIRONMENT}, author={Liu, Xiao-Huan and Zhang, Yang and Olsen, Kristen M. and Wang, Wen-Xing and Do, Bebhinn A. and Bridgers, George M.}, year={2010}, month={Jun}, pages={2443–2456} } @article{liu_zhang_cheng_xing_zhang_streets_jang_wang_hao_2010, title={Understanding of regional air pollution over China using CMAQ, part I performance evaluation and seasonal variation}, volume={44}, ISSN={["1873-2844"]}, DOI={10.1016/j.atmosenv.2010.03.035}, abstractNote={The U.S. EPA Models-3 Community Multiscale Air Quality (CMAQ) modeling system with the process analysis tool is applied to China to study the seasonal variations and formation mechanisms of major air pollutants. Simulations show distinct seasonal variations, with higher surface concentrations of sulfur dioxide (SO2), nitrogen dioxide (NO2), and particulate matter with aerodynamic diameter less than or equal to 10 μm (PM10), column mass of carbon monoxide (CO) and NO2, and aerosol optical depth (AOD) in winter and fall than other seasons, and higher 1-h O3 and troposphere ozone residual (TOR) in spring and summer than other seasons. Higher concentrations of most species occur over the eastern China, where the air pollutant emissions are the highest in China. Compared with surface observations, the simulated SO2, NO2, and PM10 concentrations are underpredicted throughout the year with NMBs of up to −51.8%, −32.0%, and −54.2%, respectively. Such large discrepancies can be attributed to the uncertainties in emissions, simulated meteorology, and deviation of observations based on air pollution index. Max. 1-h O3 concentrations in Jan. and Jul. at 36-km are overpredicted with NMBs of 12.0% and 19.3% and agree well in Apr. and Oct. Simulated column variables can capture the high concentrations over the eastern China and low values in the central and western China. Underpredictions occur over the northeastern China for column CO in Apr., TOR in Jul., and AODs in both Apr. and Jul.; and overpredictions occur over the eastern China for column CO in Oct., NO2 in Jan. and Oct., and AODs in Jan. and Oct. The simulations at 12-km show a finer structure in simulated concentrations than that at 36-km over higher polluted areas, but do not always give better performance than 36-km. Surface concentrations are more sensitive to grid resolution than column variables except for column NO2, with higher sensitivity over mountain and coastal areas than other regions.}, number={20}, journal={ATMOSPHERIC ENVIRONMENT}, author={Liu, Xiao-Huan and Zhang, Yang and Cheng, Shu-Hui and Xing, Jia and Zhang, Qiang and Streets, David G. and Jang, Carey and Wang, Wen-Xing and Hao, Ji-Ming}, year={2010}, month={Jun}, pages={2415–2426} }