Session Presentations

Yuhong (Ruby) Tian, Winston Hao, Jeongran Yun, Eric Zalewsky, Kevin Civerolo and Margaret LaFarr

Division of Air Resources, New York State Department of Environmental Conservation

Several states in the Ozone Transport Commission (OTC) region are in non-attainment status for the 2008 and/or 2015 ozone National Ambient Air Quality Standards (NAAQS) and as such are required to submit State Implementation Plans (SIPs). The Community Multiscale Air Quality (CMAQ) model was run for 2016 and the 2023 future year at both 4 and 12 kilometer resolutions. Ozone (O₃) estimations were evaluated for 71 monitoring sites over the Northeastern Non-attainment areas.

Using the EPA 2016 V1 (fi) inventory emissions at 12km and 4km resolutions, CMAQ underestimated O₃ in May and June, and overestimated O₃ in July and August. At 12km resolution, O₃ concentrations were highly overestimated in some coastal grid cells, defined as water, due to difficulty in characterizing the land/water interface in both the air quality and meteorological models. Predicted O₃ concentrations for July and August improved significantly at 4km resolution for three coastal monitors in Groton and Greenwich CT and Hart Miller Island MD. The differences in meteorological data and total emission of Nitrogen Dioxide (NO₂) and Volatile Organic Compounds (VOC) between the two platforms were explored to examine the model performance at the six coastal CT sites. Either meteorological conditions or total emission changes explained the O₃ concentration differences for most of the sites. In addition, we also compared the 2023 future design value (DVF) calculated from the 12km and 4km platforms. For most of the land monitoring sites, the 2023 DVFs from 4km platform are very similar to those from 12km platform. At 4km resolution, sites Greenwich, Stratford, Westport and Madison will have 2023 DVFs higher than 71 ppb; sites New Haven and Groton will have 2023 DVFs lower than 71 ppb.

Our work suggests that for CMAQ model, grid resolution plays a crucial role in modeling O₃ along the land/water interface area where more accurate allocating emissions at 4km resolution improved the O₃ estimations when the meteorological conditions are not favorable for a good model performance.

Jeongran Yun, Winston Hao, Eric Zalewsky, Yuhong (Ruby) Tian, and Kevin Civerolo

Division of Air Resources, New York State Department of Environmental Conservation

Grid cell sizes ranging from ~1 km to 12 km resolution have been used for various modeling applications for ozone. Previous works suggested that smaller grid cells with more detailed emissions and meteorological inputs do not always have better model performance than 12 km resolution. However, in our other work (see Tian et al, this conference), the Community Multiscale Air Quality (CMAQ) model at 4 km resolution performed better in simulating ozone than at 12 km resolution along the land/water interface area in the northern part of Long Island Sound with more accurate emissions allocation. This area is within the nonattainment area for the 2008 and 2015 8-hour ozone National Ambient Air Quality standards (NAAQS). Both the CMAQ and the Comprehensive Air quality Model with eXtensions (CAMx) photochemical models are commonly used for State Implementation Plan (SIP) attainment demonstration modeling. The 2016 emissions modeling platform for use in SIPs and other regulatory applications has been developed over the past several years through collaborative efforts between states, EPA, multi-jurisdictional organizations (MJOs), and others. In this study, CAMx performance will be compared between 2016 v1 based modeling platforms at 12 km and 4 km resolution focusing on the land/water interface along Long Island Sound. The model will be run for both 2016 and 2023, and future ozone design values will be compared between the two grid resolutions. Emissions allocations can be different at two grid resolutions; contributions to ozone from source apportionment modeling based on the emissions allocation differences will also be investigated.

Yu Qian, Yongtao Hu, Petros Vasilakos, Zongrun Li, and Armistead G. Russell

We developed, applied, and evaluated different methods to estimate ozone-emission relationships, with a focus on the South Coast Air Basin in California (SoCAB). We conducted extensive air quality modeling of the SoCAB for a number of historical and future years to understand how well the current chemical transport models (CTMs) captured ozone dynamics and the response to controls of NOx and VOC. Those results, including the sensitivities, were analyzed independently and also used to construct ozone-emission isopleths using a modified inverse distance-weighted method. Both the 2011 and 2016 comparisons show high consistency between simulated and observed MDA8 ozone concentrations, suggesting that the CMAQ simulations capture peak ozone values in terms of their relationship to estimated emissions. We conducted a set of 15 CMAQ-HDDM simulations with different emissions levels, and 11 of the cases were used to develop ozone-NOx-VOC isopleth diagrams at locations across the basin. The CMAQ-derived isopleth for Crestline (the monitoring location with the highest modeled ozone for 2016) suggests that for much of the 1975–2018 period, peak ozone levels have been most sensitive to reducing VOCs, and, up until at least 2005, NOx controls would have led to ozone increases. Recent NOx emissions reductions, however, are pushing the SoCAB to a region where, on a percentage emissions basis, NOx emissions reductions will lead to greater ozone reductions than VOC emissions controls. We combined empirical and CTM modeling to comprehensively evaluate how well do chemical transport models capture observed ozone trends and sensitivities to emissions. The HDDM-based sensitivities capture much more spatial variation in the ozone-emissions responses. A similar trend is found for other far-inland monitoring sites (e.g., Azusa), and these are the locations that typically experience the highest observed levels in the SoCAB. This indicates that the HDDM-derived response over the emissions space is consistent with the empirically derived response, which describes the overall characteristics of the ozone-emissions relationship based on long-term estimated emissions and ODV observations.

Dale J. Allen (University of Maryland)

Shan He (New Jersey State Department of Environmental Protection)

Eric Zalewsky, Winston Hao, and Jeongran Yun (New York State Department of Environmental Conservation)

Jin Sheng-Lin (Virginia Department of the Environment)

Mike Woodman (Maryland Department of the Environment)

Jeff Underhill (New Hampshire Department of Environmental Services) 

Ozone concentrations from northern Virginia to Connecticut continue to exceed National Ambient Air Quality Standards (NAAQS). To provide guidance to air quality planners, the Ozone Transport Commission (OTC) states with the assistance of modeling centers at the University of Maryland and Rutgers University used the Comprehensive Air Quality Model with Extensions (CAMx) to examine the contribution of 322 tags (e.g., WV oil and gas, MD on-road non-diesel, PA Electrical Generating Units (EGUs)) to hourly ozone amounts over the eastern two-thirds of the U.S. under NOx-limited and NOx-saturated conditions. The simulations covered April to September and used 2016 meteorology and projected 2023 emissions based on the v1 2016 with ERTAC EGU emissions modeling platform developed by collaborative efforts between states, multi-jurisdictional organizations (MJOs), EPA and others.  The relative importance of various sectors including mobile sources, Electrical Generating Units (EGUS), area non-point sources, oil-and-gas production, and commercial marine vessels (CMVs) to hourly and 8-hour maximum ozone will be examined using averages over several different measures of high ozone days.  The role of meteorology will be highlighted by comparing day-to-day and hour-to-hour variations in the contributions of individual sectors and/or states to ozone amounts.  Temporal and spatial variations in the contribution of NOx-limited versus NOx-saturated chemistry to 8-hour maximum ozone will be illustrated.    

Yukui Li, Kristina Wagstrom

According to The American Lung Association “2021 State of the Air” report, most counties in Connecticut continue to receive failing grades for air quality, specifically due to high ozone levels. This has continued even after great effort has been made to control air pollution. As with all places, there are sensitive populations in Connecticut that are at high risk of health complications when ozone concentrations are high. This creates an urgent need to lower ozone concentrations in Connecticut. The first step to achieving this is to determine the major sources contributing to ozone in Connecticut. We use the Comprehensive Air Quality Model with Extensions with the Ozone Source Apportionment Tool (CAMx-OSAT), CAMx with the Decoupled Direct Method (CAMx-DDM), and the Community Multiscale Air Quality Modeling System with the Integrated Source Apportionment Method (CMAQ-ISAM) to estimate the impact of nine nearby point sources in New York and New Jersey and twelve nearby states on ozone concentrations in Connecticut. All three approaches estimate that boundary conditions and unapportioned sources (areas except for the selected sources inside the domain) contribute the most to ozone in Connecticut. The contribution of boundary condition remains fairly consistent, while unapportioned sources and the selected sources contribute more to ozone in Connecticut during periods with elevated levels. The nine point sources combined contribute little (<7%) to ozone in Connecticut, but emissions from the twelve states contribute substantially in Connecticut when ozone is high . Emissions from the twelve states contribute a substantial fraction (~50%) of the ozone in Connecticut, especially when levels exceed 70ppb. Among the twelve states, Ohio, Pennsylvania, and New York contribute the most to summertime average ozone. During periods of elevated ozone, New York contributes the most.

Julia Placido Moore

Leonardo Hoinaski

The air pollution is a major threat to human health, and it is related to health diseases increase. The health hazard from air pollution is different to each human population. A recurring example is a higher probability of respiratory infections. That kind of hazard is highly influenced by population sensitivity, exposure, air pollution event and adaptive capacity. Santa Catarina state in Brazil presents a significant economic and industrial development scenario in the national context, which may influence human health consequently. However, little is known about the effects of air pollution on Santa Catarina population health, there are no recent research on the subject. Thus, the research aimed to evaluate the health vulnerability due to atmospheric emissions in the Santa Catarina state.

A bivariate analysis was conducted, Pearson and Spearman correlation methods were utilized to verify the behavior between independent and dependent variables. The research utilized vehicular and burning emissions data and each municipality industry distribution to consider air pollution events. Also, for the vulnerability context were used socioeconomics index such as: Gross Domestic Product (GDP), Municipal Human Development Index (HDI-M) and Gini Coefficient. For the health data, were obtained respiratory hospitalization from Department of Computer Science of the Brazilian Unified Health System (DATASUS), segregate by ages. The analyses were performed considering all the municipalities in the Santa Catarina State and for each city’s groups in the HDI-M state range.

For the whole state, the vehicular emissions and the number of industries did not seemed to affect the number of hospitalizations. Moreover, the adults and elderly age group were liable to the burn emissions. The highest hospitalization rates were in municipalities with lower HDI-M. In addition, the segregated analysis by HDI-M showed that hospital admissions for adult and elderly populations, the vehicular emissions, and the number of industries in municipalities are linked with medium HDI-M (the lowest HDI-M in the state). The results display that municipalities that are in vulnerability (lowest HDI-M), may suffer the consequences from air pollution. However, municipalities with the highest HDI-M, despite not showing to have been affected by vehicle emissions, were affected by the wildfires. This shows that such municipalities may present a limit in relation to their adaptive capacity. Also, the elderly population demonstrated to be most affected in the analysis. The current research result on hypotheses that can guide future research at this subject and may help to understand about environmental risks to human health in Santa Catarina State.

Yiang Chen, Xingcheng Lu, Jimmy C.H. Fung

The key problems in addressing air pollution include determining the source of the pollutants and developing a means to control them. In addition to the source area and source category of pollutants, the contribution of pollutants emitted during various periods is an important factor that must be better understood for effective and efficient policymaking. The source apportionment module in the Comprehensive Air Quality Model with Extensions (CAMx) was extended and applied to analyze the temporal contributions of emissions to the concentration of atmospheric particulate matter with a diameter ≤ 2.5 μm (PM_2.5) in the Pearl River Delta (PRD) region. The results showed that in addition to cross-boundary transport, the PM_2.5 concentration in the PRD region was generally influenced by emissions on the current day and previous day. Local emissions (within the local city) from 06:00 to 17:59 on the current day accounted for approximately 30% of the PM_2.5 concentration, whereas regional sources (cross-city transport) had greater contributions during earlier emission periods. During the periods of episodic pollution, a weak wind situation hindered the transportation and diffusion of pollutants; thus, pollutants from 2 days earlier were trapped within the PRD region. Our results suggest that emission control measures should be implemented 2 days in advance when adverse meteorological conditions are predicted.

Ziqi Gao, Petros Vasilakos, Theodora Nah, Masayuki Takeuchi, Hongyu Chen, David J. Tanner, Nga Lee Ng, Jennifer Kaiser, L. Greg Huey, Rodney J. Weber, and Armistead Russell

Organic acids are among the many components produced from the secondary organic aerosol formation that can impact aerosol and cloud water acidity. The most abundant gas-phase organic acid in the atmosphere is formic acid, which has been observed in concentrations in excess of 2.5 ppbV in rural areas. However, atmospheric model simulations of formic acid are typically biased low, potentially due to biases in their direct emissions and/or chemistry. In this study, we used Community Multiscale Air Quality (CMAQ) to simulate gas-phase formic acid in Yorkville, Georgia during an intensive campaign that lasted from September to October 2016. Similar to previous studies, the diurnal trend of simulated formic acid did not match observations, being biased low with a dissimilar diel profile. Potential reasons for mismatch were investigated. Additional gas-phase reactions of isoprene and monoterpenes suggested by recent studies were added, leading to a small increase in acid concentrations, but having little impact on the diel profile. The surface reaction between OH and isoprene epoxydiols (IEPOX) was simulated, increasing the simulated concentration of formic acid by up to 150%; though the resulting simulated diurnal profile did not match the observed. Examination of the diel profile found that formic acid concentrations began to rise early in the morning before suspected precursors (e.g., isoprene, IEPOX, and methanediols) and the hydroxyl radical, and that concentrations decreased rapidly in the afternoon. The rapid, early morning rise could be reproduced with increased emissions. Given the long lifetime of formic acid, and that the condensed phase formate concentrations are small, the evening decrease is explained mainly by rapid dry deposition. This suggests that formic acid underwent bidirectional deposition/emission, e.g., depositing rapidly as dew formed at night, with subsequent re-emission during the following day as the dew evaporated. This also led to observed decreases in concentrations with height.

Melissa Buechlein, Michelle Snyder, R. Chris Owen

A model sensitivity study was conducted by comparing monitoring data to modeling at an airport using source specific emission rates and NO2 chemistry algorithms in AERMOD to explore the sensitivity of each chemistry method. Temporal and spatial emissions were estimated for one year using AEDT, background NO2, NOx, and ozone concentrations were obtained from a nearby AQS monitor, and meteorology was from the on-site airport met tower. The Ozone Limiting Method (OLM), Ambient Ratio Method 2 (ARM2), Plume Volume Molar Ration Method (PVMRM), Travel Time Reaction Method (TTRM), and Generic Reaction Set Method (GRSM) were tested to understand: source parameter sensitivity, atmospheric stability sensitivity, and overall model performance.