Session Presentations

EPAs Air QUAlity TimE Series Project (EQUATES): 2002-2017 meteorology, emissions, and air quality modeling for the Northern Hemisphere and the Conterminous United States

Kristen Foley, George Pouliot, Wyat Appel, Christian Hogrefe, Alison Eyth, Norm Possiel, Michael Aldridge, Chris Allen, Jesse Bash, Megan Beardsley, James Beidler, Sarah Benish, David Choi, Brian Eder, Caroline Farkas, Rob Gilliam, Janice Godfrey, Barron Henderson, Shannon Koplitz, Rich Mason, Rohit Mathur, Chris Misenis, Havala Pye, Lara Reynolds, Matthew Roark, Sarah Roberts, Donna Schwede, Karl Seltzer, Darrell Sonntag, Kevin Talgo, Claudia Toro, Jeff Vukovich

The US EPA has developed a set of modeled meteorology, emissions, air quality and pollutant deposition spanning the years 2002 through 2017.  Modeled datasets cover the Conterminous US (CONUS) at a 12km horizontal grid spacing and the Northern Hemisphere at a 108km using WRFv4.1.1 for meteorology and CMAQv5.3.2 for air quality modeling. New hemispheric and North American emissions inventories were developed using, to the extent possible, consistent input data and methods across all years, including emissions from mobile, fire, and oil and gas sources. This presentation will compare trends in the new emissions inventories to other emissions modeling platform data. Model estimates for ozone and PM2.5 over the CONUS will be compared to estimates from previous CMAQ version and to observational data. We will also highlight EQUATES datasets that are publicly available and can be used to support a wide variety of human health and ecological modeling applications as well as model evaluation and development.

CMAQv5.3.2 ozone simulations over the Northern Hemisphere: model performance and sensitivity to model configuration

Christian Hogrefe¹, Robert Gilliam¹, Rohit Mathur¹, Barron Henderson¹, Golam Sarwar¹, K. Wyat Appel¹, George Pouliot¹, Jeff Willison¹, Rebecca Miller², Jeff Vukovich¹, Alison Eyth¹, Kevin Talgo³, Chris Allen³, and Kristen Foley¹

¹U.S. Environmental Protection Agency

²Oak Ridge Associated Universities

³GDIT

This presentation will provide an overview of the recently completed 2002 – 2017 Community Multiscale Air Quality (CMAQ) version 5.3.2 model simulations performed over the Northern Hemisphere for EPA’s Air QUAlity TimE Series Project (EQUATES) project. After presenting a brief summary of the base model configuration which includes the Weather Research and Forecasting (WRF) version 4.1.1 model and an emissions dataset developed for EQUATES, model estimates of ozone will be compared against surface observations, satellite retrievals, and ozonesonde observations for different regions. Initial results indicate a persistent overestimation of summertime surface ozone and a tendency to underestimate springtime ozone both at the surface and aloft. Regional variations in modeled ozone trends are found to be in good agreement with observations. The presentation will also include an analysis of variability and trends in ozone and other longer-lived species simulated by hemispheric CMAQ at the boundaries of a 12km modeling domain covering the conterminous U.S. Finally, we will present results from sensitivity simulations performed for 2010 to investigate the impacts of several model configuration options (including biogenic volatile organic compound emissions, soil nitrogen oxide emissions, and WRF cumulus parameterization) on model estimates.

Evaluation of National Air Quality Forecast Capability for the Summertime: Case Studies of Year 2019-2020

Youhua Tang^1,2, Patrick C. Campbell^1,2, Barry Baker^1,2, Daniel Tong^1,2, Rick Saylor¹, Ariel Stein¹, Jianping Huang^3,4, Ho-Chun Huang^3,4, Edward Strobach^3,4, Jeff McQueen³, Fanglin Yang³, Ivanka Stajner³, Jose Tirado-Delgado^5,6, Youngsun Jung⁵ and Tanya Spero⁷

1. NOAA Air Resources Laboratory (ARL), College Park, MD. 2. Center for Spatial Information Science and Systems, George Mason University, Fairfax, VA. 3. NOAA National Centers for Environmental Prediction (NCEP), College Park, MD 4. I.M. Systems Group Inc., Rockville, MD 5. NOAA NWS/STI 6. Eastern Research Group, Inc (ERG) 7. U.S. Environmental Protection Agency, Research Triangle Park, NC

The NOAA National Air Quality Forecasting Capability (NAQFC) has been upgraded with the CMAQ 5.3.1 with CB6r3-Aero7 mechanism driven by the operational Finite Volume Cubed-Sphere (FV3)-Global Forecast System (GFS), version 16, through the NOAA-EPA Atmosphere-Chemistry Coupler (NACC). We evaluated the new system, old NAQFC (CMAQ 5.0.2 driven by North America Mesoscale Model) and a downscaled GFSv16-driven WRF/CMAQ simulation with surface and airborne observations (Joint NOAA/NASA FIREX-AQ field campaign in summer 2019). Differing from the standard offline (i.e., WRF/MCIP/CMAQ) approach, the NAQFC systems use an interpolation-based meteorological coupler that can handle non-native grid meteorological inputs, as CMAQ is designed to handle various meteorological inputs with proper mass-conservation schemes. Compared with the CMAQ 5.0.2, CMAQ 5.3.1 drastically reduced the nighttime ozone overpredictions, and improved the overall ozone scores, also with help of a new emission inventory (NEI2016v1). The different meteorological drivers also showed strong impacts. The operational FV3-GFSv16 yielded better surface wind, compared to the overpredicted WRF 10-meter wind, but the GFS tended to underpredict 2-meter temperature. Correspondingly, GFS/CMAQ yielded the lowest ozone among the three models, and underpredicted the surface ozone over the western US. The GFS also tends to have stronger diurnal variations of planetary boundary layer (PBL), or higher daytime PBL and lower nighttime PBL. The comparison with FIREX-AQ aircraft data suggests that NEI2016v1 anthropogenic emissions are reasonable for most species except toluene and ethane. Overall, the WRF/CMAQ and GFS/CMAQ exhibited similar performance though their different meteorological inputs show influence on the predictions compared to the airborne observations.

Upgrades to the Canadian Operational Regional Air Quality Deterministic Prediction Systems in 2021

Michael Moran, Sylvain Ménard, Verica Savic-Jovcic, Jack Chen, Ayodeji Akingunola, Konstantinos Menelaou, Rodrigo Munoz-Alpizar, Dragana Kornic, Hugo Landry, Qiong Zheng, Junhua Zhang, and Paul Makar

Environment and Climate Change Canada (ECCC)’s operational Regional Air Quality Deterministic Prediction System (RAQDPS) has been operational since 2001 and has undergone frequent updates since then. A special parallel wildfire version called FireWork became operational in 2016. Both systems are built on a limited-area version of the GEM-MACH chemical weather model, an in-line chemistry model embedded in ECCC’s GEM meteorological forecast model, and both use a 10-km North American grid. The 2021 upgrades to the RAQDPS and FireWork include a new set of anthropogenic emissions files based on a projected 2020 Canadian inventory and projected 2023 U.S. and Mexican inventories, improved representation of particle sedimentation, several GEM improvements, and new GEM meteorological and MACH chemical libraries. FireWork also benefits from an upgraded version of the Canadian Forest Fire Emission Prediction System (CFFEPS), which has an improved wildfire plume-rise parameterization and more detailed forest fuel parameters. This presentation will describe these changes in more detail and show their impact on air-quality forecast performance.

Uncertainty Analysis of CMAQ-derived Burden Estimation of PM2.5 Exposure with Bias Correction Technique

Cheng-pin Kuo, Joshua S. Fu

Ambient PM_2.5 causing adverse effects on human health has been recognized as a major issue for regional air quality management in developing and developed countries. To quantify the impact of PM_2.5 exposure to human health, the burden of disease (BD) is a commonly used assessment tool to provide references for air quality and health management. Technically, the BD estimation relies on reported PM_2.5 exposure risks from existing literature and modeled or monitored PM_2.5 concentration. Except for using monitoring data to access PM_2.5 exposure, CMAQ modeling is also commonly employed to provide PM_2.5 exposure estimations and has been intensively applied in previous studies. However, most CMAQ studies overlooked the bias between the CMAQ modeled results and monitoring data and failed to adjust the CMAQ results with monitoring data. Moreover, if CMAQ results were directly applied for BD calculation, the derived BD estimation could be also biased. In this study, we constructed a daily PM_2.5 exposure matrix in 2013 with 3 km*3 km resolution in Taiwan and applied the inverse distance weighting (IDW) bias correction technique to adjust the CMAQ results with monitoring data. The BD was evaluated by the emergency visits of cardiovascular disease and estimated by CMAQ original and corrected results, respectively. Our objective is to quantify the difference between applying original and corrected CMAQ results for BD estimation and emphasize the importance of bias correction of CMAQ results and its potential bias for BD estimation. Our results showed that the average daily PM_2.5 in the study regions has increased from 12.3 μg/m³ to 13.0 μg/m³, which implied underestimation (5%) of PM_2.5 exposure if directly using CMAQ original results. For BD estimations, the emergency visits by using CMAQ results (2443 visits) would be 30% lower than those by using CMAQ-corrected results (3512 visits), especially for urbanized areas. The underestimation of CMAQ results and its derived BD estimation emphasizes the importance of bias correction of CMAQ modeling data. In addition, although most modeled estimations at monitoring sites had significant performance improvement, the CMAQ corrected estimations closer to the background sites had worse performance after bias correction, which suggests the IDW bias correction technique may be limited when applying in the region with complex terrain, and more data such as land-uses or satellite images should be involved for better modeling performance.

Enhanced representation of inter-continental pollutant transport by assimilating satellite NO2 and performing NOx emissions inversions

James East, Barron H. Henderson, Sergey Napelenok, Shannon Koplitz, Brad Pierce, Allen Lenzen, Rob Gilliam, Golam Sarwar, Fernando Garcia Menendez

Long-range transport of ozone (O3) to the US has human health and regulatory impacts, and is produced, in part, from international nitrogen oxides (NOx) emissions. Accurately representing international O3 transport in hemispheric and global-scale air quality models requires accurate emissions inventories. However, developments of global bottom-up emissions inventories can be out-of-date and uncertain, particularly in developing countries, limiting the representation of international emissions and transport in air quality models. We introduce a satellite data assimilation system for inverse modeling of northern hemispheric NOx emissions which offers opportunities to improve representations of long-range O3 transport and model performance. The system can provide estimates of NOx emissions in China, India, Europe, and the US. We assimilate data from the Ozone Monitoring Instrument (OMI) and the TROPOspheric Monitoring Instrument (TROPOMI) separately in the Community Multiscale Air Quality (CMAQ) model for 2019 over the northern hemisphere and perform NOx emissions inversions for each assimilation. We analyze and compare model performance and long-range transported O3 in simulations: (1) without any satellite derived information, (2) with assimilated nitrogen dioxide (NO2), and (3) with satellite derived emissions updates. Results show large NOx emissions biases in India and China, and smaller biases in the US and Europe compared to satellite-inferred estimates. Assimilating NO2 and using satellite-inferred emissions both improve O3 and NOx model performance.

Disclaimer: The views expressed in this abstract are those of the authors and do not necessarily represent the views or policies of the U.S. Environmental Protection Agency (EPA).

Modelling the impact of wood burning, road transport and agricultural emission reduction on total and secondary inorganic PM2.5 in the West Midlands, UK using WRF-CMAQ.

Mazzeo A.¹, Zhong J.¹, Hood C.², Smith S.², Dai Y.¹, Stocker J.², Cai X.¹ and Bloss W.J.¹

1. School of Geography Earth and Environmental Sciences, University of Birmingham, Edgbaston Campus B15 2TT, Birmingham UK.

2. 2. Cambridge Environmental Research Consultants, 3 King’s Parade, Cambridge, CB2 1SJ, UK

Several progressive steps have been taken in the last decade to improve the air quality in the United Kingdom. Mitigation policies at the national level have led to significant reductions in PM_2.5 concentrations, by 23 and 26% at urban and roadside locations respectively [1]. Despite this, air pollution levels exceeding the national legal threshold still represent a matter of concern for the rapid expansion evident in some places, such as the case of the West Midlands (WM).

This area is the second-most populous region of the UK after Greater London, with more than 2,900,000 inhabitants. It includes the city of Birmingham as the UK’s second-largest city, with 1.1 million inhabitants in the city alone. Moreover, according to UK government projections, the WM represents one of the highest population growth rates (+7.5%) in the period 2015-2025 [2]. Its rapid population growth, urbanisation, and consequent reduction in the air quality of urban areas led to an integrated approach to air pollution by West Midlands Combined Authority (WMCA) to reduce citizens’ exposure to PM_2.5 on the basis of the UK Clean Air Strategy 2019 [3].

According to the Air Quality Expert Group Report of 2012, 32% of the total PM_2.5 level in the UK is generated by secondary inorganic aerosols; this percentage reaches 44% of the total concentration in the city of Birmingham [4]. Studies in the WM showed that NO₃, SO₂ and NH₄ secondary inorganic fractions are the main constituents of PM_2.5 in WM urban areas, followed by carbonaceous fractions of organic and elemental carbon (OC and EC) [5]. In order to maximise the effects of national and local environmental policies, it is important to analyse the influence that these reductions have not only on primary PM_2.5 but also on the formation of secondary inorganic aerosols.

In the context of the West Midlands Air Quality Project, the modelling system for meteorology and chemistry-transport processes, WRF-CMAQ [6-7], was implemented and validated to simulate annual concentrations of PM_2.5 for a winter a summer period of the base case year 2016. The UK National Atmospheric Emission Inventory (NAEI) [8] was merged with the regional emission inventory CAMSv3.1 to account for the most up-to-date anthropogenic emissions for the UK and northwest Europe [9]. “What-if” scenarios with reduced primary emissions were designed and tested to assess the impact on annual concentrations of PM_2.5 of both the UK Clean Air Strategy 2019 and WMCA mitigation policies related to road transport, agriculture and wood burning. The secondary inorganic fractions of PM_2.5 were also quantified for the individual and combined reduced emissions scenarios.

Results show that scenarios with combined mitigation policies lead to significant decreases in the air concentration of PM_2.5, reducing both the primary PM_2.5 and secondary inorganic aerosol fractions. In light of this, the combined transport and agriculture scenarios provide a high percentage of reduction for the whole region of the West Midlands, while the reduction of primary emissions of PM_2.5 related to wood and coal burning only generates substantial concentration change in winter periods.

Sensitivity Assessment of the Ozone and Fine Particulate Matter Pollution in the Greater Bay Area Using a Regional-to-Local Coupling Model

Xuguo Zhang, Jenny Stocker, Kate Johnson, Yik Him Fung, Teng Yao, Christina Hood, David Carruthers, and Jimmy C. H. Fung 

Ultrahigh-resolution air quality models resolving sharp concentration gradients benefit health calculations. Mitigating fine particular matter (PM_2.5) in the past decade triggered the ozone (O₃) deterioration in China. Effectively controlling the coupled two-pollutant from an ultrahigh-resolution perspective is still understudied. This study proposes a novel coupled regional-to-local scale air quality modeling system suitable for quantitative assessment of pollution mitigation pathways at various resolutions. Sensitivity scenarios relating to the control of nitrogen oxide (NO_x) and volatile organic compounds (VOC) concentrating on traffic and industry sectors are explored. The results show the concurrent controls on both sectors lead to an overall 17%, 5%, and 47% emission reduction in NO_x, PM_2.5, and VOC pollutatant concentrations, respectively. The 50% less traffic scenario leads to reduced NO₂ and PM_2.5, but increased O₃ concentrations in urban areas, revealing a VOC-limited chemical regime. The reduced industrial VOC scenario leads to reduced O₃ concentrations throughout the mitigation domain. The maximum decrease in the median hourly NO₂ metric is over 11 µg/m³, and the maximum increase in the median maximum 8-hour rolling O₃ metric is over 10 µg/m³ for the reduced traffic scenario. When both the traffic and industrial control scenarios are applied, the impact on O₃ reduces to an increase of less than 7 µg/m³. The daily-averaged PM_2.5 decreases by less than 2 µg/m³ for the reduced traffic scenario and varies little for reduced VOC scenario. An O₃ episode analysis for both controls scenario leads to O₃ decreases by up to 15 µg/m³ (8-h metric) and up to 25 µg/m³ (1-h metric) in a rural area to the northeast of the mitigation domain.

Assessment of the Impact of Lightning NOx on Air Quality over the Northern Hemisphere

J. Mike Madden¹, Daiwen Kang², James East¹, Golam Sarwar², Christian Hogrefe², Rohit Mathur², and Barron H. Henderson²

¹ORAU ORISE Research Participation Program hosted at the U.S. Environmental Protection Agency, Research Triangle Park, NC, USA.

²U.S. Environmental Protection Agency, Research Triangle Park, NC, USA

Lightning is a major natural source of nitrogen oxides (NOx), and as such, it is responsible for influencing the mixing ratios of tropospheric ozone (O3), a pollutant of interest to atmospheric scientists and regulators. Lightning-generated NOx (LNOx) is largely emitted in the mid- to upper-troposphere and is thus subject to long-distance transport that can affect O3 mixing ratios downwind. Uncertainties still exist regarding the production and distribution of LNOx, especially over hemispheric scales, and the importance of LNOx will increase as anthropogenic NOx emissions decrease over parts of the world. These concerns, therefore, merit further investigations of the release and fate of LNOx, as well as an assessment of its effects on air quality.

This study uses the Community Multiscale Air Quality (CMAQ) model under varying LNOx emission configurations to analyze the effects of these emissions on tropospheric O3 mixing ratios across the Northern Hemisphere. Furthermore, this study focuses on the representation of LNOx and O3 in the mid- to upper-troposphere. Thus, performance evaluations using satellite (e.g., Ozone Monitoring Instrument (OMI)) and ozonesonde data will be conducted to examine the spread of results across the varying CMAQ configurations. Performance evaluations using surface O3 data (e.g., Air Quality System (AQS)) will also be conducted.

Model simulations with and without lightning NOx are performed. In the simulations with LNOx, the original and adjusted data from the World Wide Lightning Location Network (WWLLN) are used to produce LNOx emissions. The adjustment factors for the WWLLN data are derived to reflect National Lightning Detection Network (NLDN) seasonal climatology over the contiguous United States, but the adjustment factors are interpolated onto the Northern Hemisphere. An additional LNOx case is conducted with Global Emissions InitiAtive (GEIA) climatological LNOx emissions. Preliminary results show differences in mid- and upper-tropospheric O3 mixing ratios that are especially pronounced during summer months across tropical and sub-tropical regions.

Disclaimer: The views expressed in this abstract are those of the authors and do not necessarily represent the views or policies of the U.S. Environmental Protection Agency (EPA).

Modeling of Nitrogen and Sulfur Deposition in the Chesapeake Watershed and Tidal Bay: Trends and Sources

Sarah E. Benish^1*, Jesse O. Bash², Kristen Foley², Sergey Napelenok², Christian Hogrefe², Wyat Appel², Lewis Linker³

¹Oak Ridge Institute for Science and Education (ORISE), US Environmental Protection Agency, Research Triangle Park, NC, USA; ²US Environmental Protection Agency, Research Triangle Park, NC, USA; ³US Environmental Protection Agency, Annapolis, MD, USA

Excess deposition of atmospheric nitrogen compounds can have harmful effects on vulnerable aquatic and terrestrial ecosystems. Chemical transport models, such as the Community Multiscale Air Quality (CMAQ) model, can simulate atmospheric deposition of compounds not routinely measured, such as organic nitrogen, as well as provide information about deposition in locations without measurements, like over complex terrain or over water bodies. Multiyear simulations from the EPA’s Air QUAlity TimE Series (EQUATES) project provide a consistent modeling framework using CMAQv5.3.2 to provide stakeholders and partners with key information necessary for ecological evaluations and nutrient assessments. In this presentation, we assess trends of total inorganic nitrogen and sulfur from 2002-2017 in nine climatically consistent regions within the contiguous United States. We focus specifically on the Chesapeake watershed and tidal Bay, an area naturally sensitive to atmospheric nutrient deposition due to its geography and home to over 18 million residents. We present precipitation and bias adjustment of wet deposition fields to demonstrate improvements in modeling wet deposition compared to the National Atmospheric Deposition Program’s (NADP) National Trends Network (NTN) wet deposition measurements and show how annual total nitrogen and sulfur budgets change over time and spatially. By applying the Integrated Source Apportionment Method (ISAM) in CMAQ to the Chesapeake Bay airshed, we identify key source regions and emission sectors to quantify their contribution to nutrient loading, one of the first ISAM applications for deposition. Preliminary results suggest mobile sources are responsible for ~24% of the total oxidized nitrogen deposition and non-poultry manure contributes ~17% to the total reduced nitrogen deposition in Winter 2016 throughout the Chesapeake Bay Watershed.

Exploring the Management Value of a Machine Learning-Based Model that Predicts Chlorophyll- Using Multi-Media Modeling Environmental Predictors

Christina Feng Chang¹, Marina Astitha¹, Valerie Garcia¹, Penny Vlahos²

¹Department of Civil and Environmental Engineering, University of Connecticut, Storrs, CT 06269, USA

²Department of Marine Sciences, University of Connecticut, Groton, CT 06340, USA

In the past, we presented a machine learning (ML) and multi-media modeling framework that was used to assess environmental variables (predictors) that affect water quality by using Lake Erie as a case study. We have since improved our water quality model by: (1) expanding our study period from an 11-year study period to a 16-year study period; (2) including additional in-lake observations with the addition of 6 monitoring stations (a total of 16 stations); and (3) utilizing outputs from updated versions of the numerical prediction models. These improvements have led to the development of a successful ML model that can predict chlorophyll-α (chl-α), a proxy for lake eutrophication and algal biomass, with a coefficient of determination of 0.81, bias of -0.12μg/l and RMSE of 4.97μg/l. In-situ chl-α measurements are provided by the Lake Erie Committee Forage Task Group and the University of Toledo for the 2002-2017 period. Meteorological weather variables from the Weather Research and Forecasting model, hydrological variables from the Variable Infiltration Capacity model, and agricultural management practice variables from the Environmental Policy Integrated Climate model for the 16-year period are used to fit a random forest ML model to predict chl-α concentrations. We discuss the importance of explanatory variables that originate from these individual modeling systems and analyze the contribution of each covariate in the model to better understand the occurrence of high chlor-α concentrations. Through various sensitivity tests, we focus on: (1) the influence of each set of predictors (meteorological, hydrological, and agricultural) by varying the usage of the ML-based model inputs; and (2) the ML model’s ability to serve as a management tool through scenario evaluations (e.g., independent or concurrent changes in agriculture management, weather, and hydrology). Lessons learned from developing and testing this ML-based approach can be used to tackle water quality problems in other lakes or coastal areas and inform policy decisions.

An evaluation of a two-decade CONUS and Northern Hemispheric timeseries of retrospective WRF simulations

Robert Gilliam, Kristen Foley, Lara Reynolds, Jesse Bash, Christian Hogrefe, Rohit Mathur, Norm Possiel, Chris Misenis, Barron Henderson, Golam Sarwar, Wyat Appel and George Pouliot

The US EPA’s Air QUAlity TimE Series (EQUATES) Project has developed a consistent set of modeled meteorology and emissions that are being used to produce multidecadal timeseries (2002-2017) of air quality model output using the Community Multiscale Air Quality (CMAQ) model. This CMAQ dataset will be leveraged for a diverse set of research and air quality management applications. To support this collective dataset that will be available for broad distribution, we present an evaluation of the nearly twenty years of meteorology that was produced using version 4.1.1 of the Weather, Research and Forecasting (WRFv4.1.1) model. WRFv4.1.1 was ran over the contiguous United States and the Northern Hemisphere for the 2002-2020 period using a horizontal grid-resolution of 12 km and 108 km, respectively, although the air quality was a few years shorter. We examine the simulated meteorology using a broad range of observations including upper-air and near-surface temperature, moisture and winds. We also use shortwave radiation observations to examine the model’s ability to represent clouds. And finally, the US-based PRISM (Parameter-elevation Regressions on Independent Slopes Model) precipitation dataset is leveraged to understand how well the model characterizes seasonal and annual precipitation. Results of the evaluation show that the model output is suitable in terms of driving air quality models as error levels are within expected bounds of recent retrospective simulations that use data assimilation. As expected, the 12 km scale simulations reduce the level of error when compared to the coarser 108 km hemispheric domain. As such, the distribution and use of these meteorological modeling datasets can help air quality modelers by providing a robust and lengthly set of inputs for air quality applications.

Modeling potential odor sources in Brunswick, GA

Xiangyu Jiang, Byeong-Uk Kim, Jim Boylan

Residents in Brunswick, GA reported foul smelling odors from December 2, 2020 to February 9, 2021. To locate potential emission sources associated with the odor complaints, we applied two modeling approaches along with a “hybrid” approach. The first modeling approach used a simple vector trajectory method with 1-minute wind speed and wind direction data from a single meteorological station to track air parcels back in time. The second approach used the HYSPLIT model to identify the hourly averaged movement of air parcels based on the 3-km HRRR meteorological data. Both approaches show that more than 80% of the back-trajectories passed through the 1-mile buffer zone surrounding a wood pulp mill and the adjacent wastewater treatment plant. When the meteorological monitoring station, potential odor sources, and odor complaints are relatively close together, the simple vector trajectory method captures more back-trajectories passing the 1-mile buffer zone compared to the HYSPLIT approach. The “hybrid approach” takes advantage of strengths from both modeling approaches. This approach showed that over 90% of the back-trajectories passed the 1-mile buffer zone. Based on this analysis, the odor source(s) are likely in and/or near the 1-mile buffer zone surrounding the wood pulp mill and the adjacent wastewater treatment plant.

Implementation of Lightning NOx production in CMAQ over the Northern Hemisphere with Lightning Flash Data from WWLLN

Daiwen Kang, Mike Madden, James East, Golam Sarwar, Christian Hogrefe, Rohit Mathur, and Barron H. Henderson

Lightning NO_x (nitrogen oxides) emissions based on lightning flash data from the National Lightning Detection Network (NLDN), a ground-based lightning detection network that covers the contiguous United States (CONUS) with high detection efficiency, have been included and analyzed in previous modeling studies using the Community Multiscale Air Quality (CMAQ) model over the CONUS domain. For modeling simulations beyond the CONUS domain, a potential alternative source of lightning flash dataset is the World Wide Lightning Location Network (WWLLN), albeit with low detection efficiency. In this study, a comparison between NLDN and WWLLN is performed over the CONUS domain, and an effort is made to adjust the WWLLN data based on the climatological ratios between WWLLN and NLDN. Lightning NO_x emissions are then estimated using lightning flashes from NLDN, WWLLN, and the adjusted WWLLN flashes in CMAQ simulations over the CONUS domain. This presentation will focus on the development of a methodology for the adaptation of the WWLLN data and the assessment of model performance over the CONUS domain to demonstrate how the flash data from WWLLN can be used and how the simulations perform with different sources of data and quality. In addition, as an important component of the impact of lightning NO_x on air pollution, the simulations of aerosol nitrate deposition are evaluated using data from the National Atmospheric Deposition Program (NADP)’s National Trends Network (NTN). Then, CMAQ simulations are performed over the Northern Hemisphere without any lightning NO_x, and with lightning NO_x produced by the original and adjusted WWLLN flashes and using the Global Emissions InitiAtive (GEIA) climatological lightning NO_x emissions. The simulation results with lightning NO_x emissions are compared to those obtained without lightning NO_x and with climatological lightning NO_x emissions. A follow-on presentation will focus on the assessment of hemispheric CMAQ simulations with varying lightning NO_x emissions.

Evaluation of the Community Multiscale Air Quality (CMAQ) version 5.3.2 over California: Preliminary analysis

Sarika Kulkarni and Chenxia Cai

The Community Multiscale Air Quality (CMAQ) version 5.3.2 (CMAQ532) that was released to the public in October 2020, has numerous scientific updates when compared to previous versions. This study will present preliminary findings from the ongoing evaluation of an air quality modeling system comprised of CMAQ532, Weather Research Forecasting v4.2.1 (WRF421) and Sparse Matrix Operator Kernel Emissions 4.8 (SMOKE48) applied over California. As a first step, the performance of CMAQ532, using State Air Pollution Research Center version 07tc with extended isoprene chemistry (SAPRC07tic) and aerosol module (AERO6) treatment of SOA, was evaluated by comparing simulated ozone and particulate matter (PM2.5) to observations as well as model results from a similarly configured modeling system which utilized the CMAQ version 5.2.1 (CMAQ521). Preliminary analysis has shown that ozone and PM2.5 predicted by CMAQ532 compare well with observed values and are able to capture the broad spatial patterns and temporal variability over the major air basins in California including San Joaquin Valley, Sacramento Metropolitan area and South Coast. However, the CMAQ532 simulated ozone and PM2.5 values are generally lower when compared to the corresponding values from the CMAQ521 simulation. Additional sensitivity analysis results will be presented to isolate the impact of specific science updates including but not limited to new aerosol module (AERO7) and the new Surface Tiled Aerosol and Gaseous Exchange (STAGE) deposition model on the simulated ozone and PM2.5 over California.

An evaluation of trends and variability in aerosol optical depth over the Northern Hemisphere simulated by CMAQv5.3.2 for the EQUATES project

R. Miller¹, C. Hogrefe², R. Mathur², J. Pleim² and K. Foley²

¹ Oak Ridge Associated Universities

² AESMD, Environmental Protection Agency

Air quality models are important tools for studying atmospheric trace gases and aerosols. Their outputs are often used to quantify adverse impacts of air quality on human health and the environment. Model simulations were performed for 2002–2017 using CMAQv5.3.2 and WRFv4.1.1 over the Northern Hemisphere. Emissions were represented by a 2002–2017 emissions dataset developed for EPA’s Air QUAlity TimE Series (EQUATES) project. In this presentation, we compare these simulations with monthly average aerosol optical depth (AOD) data from the Moderate Resolution Imaging Spectroradiometer (MODIS) on board the NASA Terra satellite to assess the model’s ability to capture observed variability and trends in the tropospheric aerosol burden. Initial findings show regional and seasonal differences between model and satellite observations. Over CONUS, the annual AOD cycle is in general agreement with satellite observations for all seasons, except summer when CMAQ AOD values are lower than observed. Linear trends through the annual average AOD reveal a decrease in AOD over the southeastern United States, Europe, and China, and an increase over India. These linear trends are similar for both the model output and satellite observations. To provide further context for the AOD analysis, we also present an analysis of ground-based AOD from AERONET and PM2.5 concentrations measured at surface monitoring networks over North America and Europe.

The Impact of Altering Emission Data Precision on Efficiency and Accuracy Regarding the Community Multiscale Air Quality Model
Michael Walters (US EPA)

The Community Multiscale Air Quality Model (CMAQ) has been a vital tool for regulatory work at the United States’ Environmental Protection Agency (EPA), State government sectors, and at numerous, academic institutions. At the EPA specifically, CMAQ itself requires a copious amount of disk space in terms of input and output as a result of temporally long simulations initialized at a fine horizontal grid scale which encompasses the contiguous United States. As a result, disk space management is strict to ensure the uninterrupted progress of ongoing research projects at the EPA. Such disk space management includes optimal data compression techniques on input and output files for all CMAQ simulations when archiving is necessary. Currently, there are several such utilities’ that compresses files using losslessness compression such as GNU zip (gzip) and Basic Leucine Zipper Domain (bzip2), however, a new approach has been created herein this study and ingested into CMAQ simulations to examine the impact on CMAQ model performance and data compression efficiency. In total, four simulations were conducted: Three using emission files which consist of 5, 4, or 3 significant digits, and the fourth case using original, unaltered emission files as the benchmark case.

Results demonstrate that CMAQ runtime was consistently reduced by roughly 1-7% for simulations using altered emission files. Additionally, the altered emission files improved disk space by 5%, 25%, and 48% compared to the unaltered emission files when using the bzip2 compression utility for files containing 5, 4 and 3 significant figures respectively. As for simulation accuracy at in-situ locations via the United States Environmental Protection Agency’s (USEPA) Air Quality System (AQS) and Ammonia Monitoring Network (AMON) stations, all 3 cases performed nearly identical to one another for ammonia (NH3), ozone (O3), and particulate matter (PM2.5). In fact, the normalized mean bias for all analyzed species differed by less than 0.01% for all bulk, regional, and temporally stratified calculations. In summary, we have shown that by reducing input CMAQ emission data to 5, 4, or 3 significant digits, simulation runtime slightly improved, disk space significantly improved, and model accuracy remained relatively unchanged with respect to the unaltered emission data CMAQ simulation.