Responses of future air quality to emission controls over North Carolina, Part II: Analyses of future-year predictions and their policy implications

doi:10.1016/j.atmosenv.2010.03.022

Atmospheric Environment

Volume 44, Issue 23, July 2010, Pages 2767-2779

https://doi.org/10.1016/j.atmosenv.2010.03.022 Get rights and content

Abstract

The MM5/CMAQ system evaluated in Part I paper is applied to study the impact of emission control on future air quality over North Carolina (NC). Simulations are conducted at a 4-km horizontal grid resolution for four one-month periods, i.e., January, June, July, and August 2009 and 2018. Simulated PM_2.5 in 2009 and 2018 show distribution patterns similar to those in 2002. PM_2.5 concentrations over the whole domain in January and July reduced by 5.8% and 23.3% in 2009 and 12.0% and 35.6% in 2018, respectively, indicating that the planned emission control strategy has noticeable effects on PM_2.5 reduction in this region, particularly in summer. More than 10% and 20% of 1-h and 8-h O₃ mixing ratios are reduced in July 2009 and 2018, respectively, demonstrating the effectiveness of emission control for O₃ reduction in summer. However, O₃ mixing ratios in January 2009 and 2018 increase by more than 5% because O₃ chemistry is VOC-limited in winter and the effect of NO_x reduction dominates over that of VOC reduction under such a condition. The projected emission control simulated at 4-km will reduce the number of sites in non-attainment for max 8-h O₃ from 49 to 23 in 2009 and to 1 in 2018 and for 24-h average PM_2.5 from 1 to 0 in 2009 and 2018 based on the latest 2008 O₃ and 2006 PM_2.5 standards. The variability in model predictions at different grid resolutions contributes to 1–3.8 ppb and 1–7.9 μg m⁻³ differences in the projected future-year design values for max 8-h O₃ and 24-h average PM_2.5, respectively.

Introduction

Air quality attainment for future-years posts significant challenges in emission control technologies, regulation revision and enforcement, as well as decision tool development and application. CMAQ is one of the decision tools for regulatory applications developed by the U.S. EPA. Several global and regional models including MM5/CMAQ have recently been applied to simulate future air quality and their responses to future climate changes and/or emission changes either as a result of changing climate (e.g., biogenic emissions) or as part of the emission control programs (e.g., anthropogenic emissions) (e.g., Hogrefe et al., 2004, Liao et al., 2006, Arunachalam et al., 2006, Tagaris et al., 2007, Wu et al., 2008a, Zhang et al., 2008). Most of these studies are conducted at a regional scale (≥36-km) and focus only on surface ozone (O₃). Very few focus on fine particulate matter (PM_2.5) and consider the changes in biogenic and/or anthropogenic emissions (e.g., Tagaris et al., 2007, Zhang et al., 2008). Since O₃ and PM_2.5 share common emission sources and precursors, studies excluding each other may not provide a complete description of future air quality and an emission control strategy focusing on one pollutant may not lead to an overall improvement of future air quality. For example, a reduction in emissions of volatile organic compounds (VOCs) may decrease O₃ but increase particulate nitrate formation (Meng et al., 1997, Liu et al., submitted for publication a). The effectiveness of the emission control strategies depends on chemical and meteorological conditions due to the complex, non-linear interplays of chemical and meteorological processes during the formation of O₃ and PM_2.5. For example, O₃ can be most effectively reduced by reducing NO_x under the NO_x-limited conditions or by reducing VOCs under the VOC-limited conditions. Tsimpidi et al. (2007) found that NH₃ emission control during winter time is a more effective and less costly control strategy than reducing NO_x and SO₂ emissions. Emission control strategies therefore require a careful design to reflect the characteristics of emissions, chemistry, and meteorology of regions of interest. In this Part II paper, the impact of emission control on future air quality will be evaluated by using CMAQ v4.5.1 described and evaluated in Part I (Liu et al., in press b). The effectiveness of the projected emission controls under winter and summer conditions will be assessed. As shown in Part I, the performance of MM5/CMAQ for January and July 2002 simulations is overall consistent with that reported in the literature, although some large biases occur for temperature at 1.5 m in January, precipitation in both months, and 24-h average PM_2.5 concentrations in July. Compared with the simulation at 12-km, the 4-km simulation gives slightly larger biases for maximum 1-h and 8-h average mixing ratios of O₃, 24-h average concentrations of PM_2.5 and most PM_2.5 components, and visibility parameters, but lower biases for EC and OM in January and wet deposition fluxes of NH₄⁺ and NO₃⁻ in July.

The state implementation plan (SIPs) modeling for multiple pollutants is expected to be conducted at a grid resolution of 12-km or finer (U.S. EPA, 2007). Several studies have evaluated the sensitivity of model predictions over NC to horizontal grid resolution. For example, Arunachalam et al. (2006) evaluated the impact of grid resolution on O₃ simulated by the Multiscale Air Quality Simulation Platform (MAQSIP) for 19–25 June 1996 and a few days in summers 1995–1997, respectively. Neither work assessed the impacts of grid resolution on meteorology, PM_2.5, visibility, and dry and wet deposition amounts, some of these were examined in Wu et al. (2008b) and Queen and Zhang (2008) at 4-, 12-, and 36-km for both August and December 2002 using CMAQ and an older version of the VISTAS’s emissions, and all of these have been examined at 4- and 12-km for both January and July 2002 using a variant of CMAQ and the latest VISTAS’s emissions in Liu et al. (in press b). Evaluation of such impacts has not been done for future-year simulations and for a complete set of model outputs over NC, which is another focus of this Part II paper. Such an evaluation aims to address some policy-related concerns regarding whether the model results at 4-km significantly differ from those at 12-km in terms of temporal variation, spatial distribution, and performance statistics and what species exhibit a large sensitivity; what the policy implications of such a sensitivity are to the SIP modeling; and what additional information can fine-scale simulations provide for SIP and future design values for O₃ and PM_2.5.

Section snippets

Responses of air quality to emission reductions in 2009 and 2018

Fig. 1 shows the emissions of major pollutants in current and future years and corresponding domain-average percentage reductions from the level of 2002–2009 and 2018 for these emissions and concentrations of O₃ and PM_2.5. CO, an important O₃ precursor under rural conditions, is mainly emitted from motor vehicles in NC, with a total of 4,164,158 tons in 2002 and projected 19.7% and 32.6% reductions in 2009 and 2018, respectively, from the 2002 level. NO_x, an important precursor of O₃ and PM

The sensitivity of model predictions to horizontal grid resolution

The sensitivity of model predictions to horizontal grid resolution in 2009/2018 simulations is overall similar. Fig. 7 shows absolute differences in the monthly-mean maximum 1-h and 8-h O₃ mixing ratios, 24-h average concentrations in PM_2.5, and EXT_Recon and DCV_Recon between the 4- and 12-km simulations in January and July 2009. Compared with the 12-km simulation, the 2009 4-km simulation gives higher O₃ (mostly 1–4 ppb, or < 10%, up to 6.6 ppb or 22.6%) in the mountain area but lower values

Modeled attainment test

The U.S. EPA’s modeled attainment test is performed to estimate the impact of emission controls in 2009/2018 on the attainment of 8-h average max O₃ and 24-h average PM_2.5 at a given monitor over NC. The results are summarized in Table 1, Table 2, Table 3. The site-specific current design values (DVCs) for 8-h max O₃ are obtained from NCDENR, except for four CASTNET sites where those values are obtained from Arunachalam et al. (2006). The DVCs were calculated as 5-year weighted design values at

Summary

MM5/CMAQ simulations are conducted to simulate air quality of January and July 2002, 2009, and 2018. The impact of planned emission control strategies are examined to evaluate their effectiveness in controlling PM_2.5 and O₃ levels in NC in 2009 and 2018. The impact of horizontal grid resolution on simulated O₃ and PM_2.5 and their policy implications are assessed. The planned emission control strategy will reduce O₃ levels by up to 22.5% in July 2009/2018 from its level in 2002. It will,

Acknowledgements

This work was supported by North Carolina Department of Environment and Natural Resources Division of Air Quality, the USDA Air Quality Program/National Research Initiative Competitive Grant no. 2008-35112-18758, and the National Science Foundation Career Award Atm-0348819. Thanks are due to Mike Abraczinskas, Chris Misenis, Wayne Cornelius, Karen Harris, and Hoke Kimball of NCDENR for providing VISTAS’s 12-km CMAQ inputs and results and observational data; Don Olerud, BAMS, for providing

References (20)

S. Arunachalam et al.
A quantitative assessment of the influence of grid resolution on predictions of future-year air quality in North Carolina, USA
Atmospheric Environment
(2006)
A. Queen et al.
Examining the sensitivity of MM5-CMAQ predictions to explicit microphysics schemes and horizontal grid resolutions, Part III-The impact of horizontal grid resolution
Atmospheric Environment
(2008)
S.-Y. Wu et al.
Modeling atmospheric transport and fate of ammonia in North Carolina-Part I: evaluation of meteorological and chemical predictions
Atmospheric Environment
(2008)
C. Hogrefe et al.
Simulating changes in regional air quality over the eastern United States due to changes in global and regional climate and emissions
Journal of Geophysical Research
(2004)
D.J. Jacob et al.
Seasonal transition from NO_x⁻ to hydrocarbon-limited conditions for ozone production over the eastern United States in September
Journal of Geophysical Research
(1995)
L.R. Leung et al.
Potential regional climate change and implications to US air quality
Geophysical Research Letters
(2005)
H. Liao et al.
Role of climate change in global predictions of future tropospheric ozone and aerosols
Journal of Geophysical Research
(2006)
Liu, X.-H., Y. Zhang, J. Xing, Q. Zhang, D.G. Streets, C.J. Jang, W.-X. Wang, and J.-M. Hao. Understanding of regional...
Liu, X.-H., Zhang, Y., Olsen, K., Wang, W.-X., Do, B., Bridgers, G. Responses of future air quality to emission...
Z. Meng et al.
Chemical coupling between atmospheric ozone and particulate matter
Science
(1997)

There are more references available in the full text version of this article.

Cited by (19)

Re-Naturing Cities: Evaluating the effects on future air quality in the city of Porto
2020, Atmospheric Environment
The effect of different “green” measures, such as the increase of urban green areas, the application of green roofs and the increase of surfaces albedo on urban air quality were evaluated with the WRF-CHIMERE modelling system. In order to account for the heterogeneity of urban areas, a single layer urban canopy model was coupled to the WRF model. The case study consists of a heat wave occurring in the Porto (Portugal) urban area in a future climate scenario, considering the Representative Concentration Pathway RCP8.5. The influence of the selected measures on PM10, NO₂ and O₃ concentrations was quantified and compared with a control run (without measures) simulation scenario. The results revealed that all the measures are able to mitigate the effects of heat waves by reducing the air temperature between −0.5 °C and −1 °C (maximum differences for the mean of the episode). Positive and negative effects were found in terms of air quality. The implementation of green roofs and the increase of surfaces albedo promoted an overall increase of PM10 (between +0.6% and +1.5%) and NO₂ (between +0.8% and 3.5%) concentrations, which are closely related to a decrease of vertical mixing in the urban boundary layer. The increase of green urban areas promoted an overall decrease (on average) of both PM10 and NO₂, by around −1% and −3%, respectively. The O₃ levels increased with the increase of urban green areas, mostly located over the Porto urban area. Slight differences were promoted by the implementation of green roofs. For the increase of surfaces albedo, both increases and decreases of O₃ concentrations were observed. The obtained results contribute to the knowledge of the chemical composition of the urban atmosphere and can be of great importance for stakeholders and decision-makers to deal with climate change impacts.
Climate change and pollutant emissions impacts on air quality in 2050 over Portugal
2016, Atmospheric Environment
Citation Excerpt :
the studies that only consider the effect of climate change, keeping anthropogenic emissions constant (Langner et al., 2005; Racherla and Adams, 2008; Meleux et al., 2007; Zhang et al., 2008; Fiore et al., 2011; Carvalho et al., 2010; Tai et al., 2012; Manders et al., , 2012); the studies that maintain the meteorological conditions constant (same as the historical year) in future scenarios and only change the pollutant emission scenarios (Dentener et al., 2005; Zhang et al., 2010); the studies that consider the effect of climate change along with the modification of the anthropogenic emissions in the future scenarios (Unger et al., 2006; Dentener et al., 2006; Stevenson et al., 2006; Tagaris et al., 2007; Nolte et al., 2008; Lei et al., 2012; Gidhagen et al., 2012; Coleman et al., 2013; Doherty et al., 2013; Gao et al., 2013; Jiang et al., 2013; Colette et al., 2013; Penrod et al., 2014; Trail et al., 2014; Lacressonnière et al., 2014; Markakis et al., 2014).
Changes in climate and air pollutant emissions will affect future air quality from global to urban scale. In this study, regional air quality simulations for historical and future periods are conducted, with CAMx version 6.0, to investigate the impacts of future climate and anthropogenic emission projections on air quality over Portugal and the Porto metropolitan area in 2050. The climate and the emission projections were derived from the Representative Concentrations Pathways (RCP8.5) scenario. Modelling results show that climate change will impact NO₂, PM10 and O₃ concentrations over Portugal. The NO₂ and PM10 annual means will increase in Portugal and in the Porto municipality, and the maximum 8-hr daily O₃ value will increase in the Porto suburban areas (approximately 5%) and decrease in the urban area (approximately 2%). When considering climate change and projected anthropogenic emissions, the NO₂ annual mean decreases (approximately 50%); PM10 annual mean will increase in Portugal and decrease in Porto municipality (approximately 13%); however PM10 and O₃ levels increase and extremes occur more often, surpassing the currently legislated annual limits and displaying a higher frequency of daily exceedances. This air quality degradation is likely to be related with the trends found for the 2046–2065 climate, which implies warmer and dryer conditions, and with the increase of background concentrations of ozone and particulate matter. The results demonstrate the need for Portuguese authorities and policy-makers to design and implement air quality management strategies that take climate change impacts into account.
Application of Weather Research and Forecasting Model with Chemistry (WRF/Chem) over northern China: Sensitivity study, comparative evaluation, and policy implications
2016, Atmospheric Environment
Citation Excerpt :
It should be noted that the observational data of the indicators mentioned in the Section 6 in the supplementay material are not available over China to verify simulated model sensitivity. However, the observations of another indicator, i.e., the ratios of column mass abundance of HCHO/NO2 (Tonnesen and Dennis, 2000a, b; Y. Zhang et al., 2009b, 2010b), can be calculated from the satellite observed column mass concentrations of HCHO and NO2. Fig. 5 compares the simulated and observed tropospheric column mass abundance of HCHO and NO2, and their ratios.
An extremely severe and persistent haze event occurred over the middle and eastern China in January 2013, with the record-breaking high concentrations of fine particulate matter (PM_2.5). In this study, an online-coupled meteorology-air quality model, the Weather Research and Forecasting Model with Chemistry (WRF/Chem), is applied to simulate this pollution episode over East Asia and northern China at 36- and 12-km grid resolutions. A number of simulations are conducted to examine the sensitivities of the model predictions to various physical schemes. The results show that all simulations give similar predictions for temperature, wind speed, wind direction, and humidity, but large variations exist in the prediction for precipitation. The concentrations of PM_2.5, particulate matter with aerodynamic diameter of 10 μm or less (PM₁₀), sulfur dioxide (SO₂), and nitrogen dioxide (NO₂) are overpredicted partially due to the lack of wet scavenging by the chemistry-aerosol option with the 1999 version of the Statewide Air Pollution Research Center (SAPRC-99) mechanism with the Model for Simulating Aerosol Interactions and Chemistry (MOSAIC) and the Volatility Basis Set (VBS) for secondary organic aerosol formation. The optimal set of configurations with the best performance is the simulation with the Gorddard shortwave and RRTM longwave radiation schemes, the Purdue Lin microphysics scheme, the Kain-Fritsch cumulus scheme, and a nudging coefficient of 1 × 10⁻⁵ for water vapor mixing ratio. The emission sensitivity simulations show that the PM_2.5 concentrations are most sensitive to nitrogen oxide (NO_x) and SO₂ emissions in northern China, but to NO_x and ammonia (NH₃) emissions in southern China. 30% NO_x emission reductions may result in an increase in PM_2.5 concentrations in northern China because of the NH₃-rich and volatile organic compound (VOC) limited conditions over this area. VOC emission reductions will lead to a decrease in PM_2.5 concentrations in eastern China. However, 30% reductions in the emissions of SO₂, NO_x, NH₃, and VOC, individually or collectively, are insufficient to effectively mitigate the severe pollution over northern China. More aggressive emission controls, which needs to be identified in further studies, are needed in this area to reach the objective of 25% PM_2.5 concentration reduction in 2017 proposed in the Action Plan for Air Pollution Prevention and Control by the State Council in 2013.
Influence of the choice of gas-phase mechanism on predictions of key gaseous pollutants during the AQMEII phase-2 intercomparison
2015, Atmospheric Environment
Citation Excerpt :
Such a measure is complementary to what we presented in the last section, as it represents the slope of the O3 surface in Figs. 7 and 8 at a certain point. The indicator of Martin et al. (2004) is especially useful as both quantities can be observed via satellite instruments and are hence eminently useful for 3-D model evaluation as is done in Zhang et al. (2010) or Campbell et al. (2015). CH2O/NO2 indicates VOC-limited conditions if the ratio is below 1, and NOx-limited conditions if above 1.
The formulations of tropospheric gas-phase chemistry (“mechanisms”) used in the regional-scale chemistry-transport models participating in the Air Quality Modelling Evaluation International Initiative (AQMEII) Phase 2 are intercompared by the means of box model studies. Simulations were conducted under idealized meteorological conditions, and the results are representative of mean boundary layer concentrations. Three sets of meteorological conditions – winter, spring/autumn and summer – were used to capture the annual variability, similar to the 3-D model simulations in AQMEII Phase 2. We also employed the same emissions input data used in the 3-D model intercomparison, and sample from these datasets employing different strategies to evaluate mechanism performance under a realistic range of pollution conditions.
Box model simulations using the different mechanisms are conducted with tight constraints on all relevant processes and boundary conditions (photolysis, temperature, entrainment, etc.) to ensure that differences in predicted concentrations of pollutants can be attributed to differences in the formulation of gas-phase chemistry. The results are then compared with each other (but not to measurements), leading to an understanding of mechanism-specific biases compared to the multi-model mean. Our results allow us to quantify the uncertainty in predictions of a given compound in the 3-D simulations introduced by the choice of gas-phase mechanisms, to determine mechanism-specific biases under certain pollution conditions, and to identify (or rule out) the gas-phase mechanism as the cause of an observed discrepancy in 3-D model predictions.
We find that the predictions of the median diurnal cycle of O₃ over a set of emission conditions representing a network of station observations is within 4 ppbv (5%) across the different mechanisms. This variability is found to be very similar on both continents. There are considerably larger differences in predicted concentrations of NO_x (up to ± 25%), key radicals like OH (40%), HO₂ (25%) and especially NO₃ (>100%). Secondary substances like H₂O₂ (25%) or HNO₃ (10%), as well as key volatile organic compounds like isoprene (>100%) or CH₂O (20%) differ substantially as well. Calculation of an indicator of the chemical regime leads to up to 20% of simulations being classified differently by different mechanism, which would lead to different predictions of the most efficient emission reduction strategies.
All these differences are despite identical meteorological boundary conditions, photolysis rates, as well as identical biogenic and inorganic anthropogenic emissions. Anthropogenic VOC emissions only vary in the way they are translated in mechanism-specific compounds, but are identical in the total emitted carbon mass and its spatial distribution.
Our findings highlight that the choice of gas-phase mechanism is crucial in simulations for regulatory purposes, emission scenarios, as well as process studies that investigate other components like secondary formed aerosol components. We find that biogenic VOCs create considerable variability in mechanism predictions and suggest that these, together with nighttime chemistry should be areas of further mechanism improvement.
Evaluation of concentrations and source contribution of PM10 and SO2 emitted from industrial complexes in Ulsan, Korea: Interfacing of the WRF-CALPUFF modeling tools
2014, Atmospheric Pollution Research
Citation Excerpt :
Therefore, it is essential to collect many measurements of these air pollutants, which necessitates that local governments maintain the optimized air quality network with a correct spatial distribution according to the specific emissions, orography and weather conditions to determine ambient concentrations as accurately as possible near regional sources. Specifically, air quality models have proven useful for determining the spatio–temporal distribution of air pollutants and for developing emission control policies that allocate limits to air pollutant emissions (Holmes and Morawska, 2006; Zhang et al., 2010; Ma et al., 2013). Dispersion modeling describes the transport and dispersion of air pollutants, as well as chemical and physical processes within the plume.
The Ulsan metropolitan city in Korea includes two national industrial complexes [Ulsan Petrochemical Industrial Complex (UPIC) and On–San Industrial Complex (OSIC)] that produce various industrial products. Air pollution from these industrial complexes may pose potential health risks to nearby residential areas. Therefore, WRF–CALPUFF (Weather Research and Forecasting–California PUFF) modeling systems were used to simulate concentration distributions of typical air pollutants (PM₁₀ and SO₂), and statistics are computed to determine the models' ability to simulate observations. Finally, we classified the type of business and districts in the region and evaluated their contribution to air pollutant concentrations. Five statistical metrics [Index of Agreement (IOA), Fractional Bias (FB), Normalized Mean Square Error (NMSE), and Pearson correlation coefficient (R)] indicated that the simulated values using CALMET was determined to have sufficient reliability to predict CALPUFF, and simulated concentration field using CALPUFF showed a good agreement [typical values: IOA (0.284 to 0.850 for PM₁₀, 0.412 to 0.895 for SO₂), and FB (0.043 to 0.821 for PM₁₀, –0.393 to 0.638 for SO₂)] with the observed concentrations. The maximum concentrations of PM₁₀ and SO₂ using CALPUFF were predicted to be located around OSIC and UPIC, respectively. We compared the simulated values with observed values at 14 monitoring stations, and the SO₂ tended to display better agreement to observed SO₂ values than modeled and observed PM₁₀. The source contribution analysis found that PM₁₀ and SO₂ were mostly influenced by group B (35.1%) including steel, machinery, and electronic industry nearby OSIC and group A (40.6%) including chemical industry nearby UPIC, respectively. Finally, the correlations between simulated concentrations of PM₁₀ and SO₂ and corresponding emission quantities were 0.663 and 0.528, respectively. Overall, the results of this study could be useful for designing appropriate seasonal regulations to reduce ambient concentrations of air pollutants and assisting environmental administrators to control the sources that contribute the most to degradation of air quality.
Impacts of future climate and emission changes on U.S. air quality
2014, Atmospheric Environment
Changes in climate and emissions will affect future air quality. In this work, simulations of regional air quality during current (2001–2005) and future (2026–2030) winter and summer are conducted with the newly released CMAQ version 5.0 to examine the impacts of simulated future climate and anthropogenic emission projections on air quality over the U.S. Current meteorological and chemical predictions are evaluated against observations to assess the model's capability in reproducing the seasonal differences. WRF and CMAQ capture the overall observational spatial patterns and seasonal differences. Biases in model predictions are attributed to uncertainties in emissions, boundary conditions, and limitations in model physical and chemical treatments as well as the use of a coarse grid resolution. Increased temperatures (up to 3.18 °C) and decreased ventilation (up to 157 m in planetary boundary layer height) are found in both future winter and summer, with more prominent changes in winter. Increases in future temperatures result in increased isoprene and terpene emissions in winter and summer, driving the increase in maximum 8-h average O₃ (up to 5.0 ppb) over the eastern U.S. in winter while decreases in NO_x emissions drive the decrease in O₃ over most of the U.S. in summer. Future PM_2.5 concentrations in winter and summer and many of its components decrease due to decreases in primary anthropogenic emissions and the concentrations of secondary anthropogenic pollutants as well as increased precipitation in winter. Future winter and summer dry and wet deposition fluxes are spatially variable and increase with decreasing surface resistance and precipitation, respectively. They decrease with a decrease in ambient particulate concentrations. Anthropogenic emissions play a more important role in summer than in winter for future O₃ and PM_2.5 levels, with a dominance of the effects of significant emission reductions over those of climate change on future PM_2.5 levels.

View all citing articles on Scopus

View full text

Responses of future air quality to emission controls over North Carolina, Part II: Analyses of future-year predictions and their policy implications

Abstract

Introduction

Section snippets

Responses of air quality to emission reductions in 2009 and 2018

The sensitivity of model predictions to horizontal grid resolution

Modeled attainment test

Summary

Acknowledgements

Atmospheric Environment

Atmospheric Environment

Atmospheric Environment

Simulating changes in regional air quality over the eastern United States due to changes in global and regional climate and emissions

Journal of Geophysical Research

Seasonal transition from NOx− to hydrocarbon-limited conditions for ozone production over the eastern United States in September

Journal of Geophysical Research

Potential regional climate change and implications to US air quality

Geophysical Research Letters

Role of climate change in global predictions of future tropospheric ozone and aerosols

Journal of Geophysical Research

Chemical coupling between atmospheric ozone and particulate matter

Science

Seasonal transition from NO_x⁻ to hydrocarbon-limited conditions for ozone production over the eastern United States in September