Long-term criteria and toxic pollutants trends and community exposures over the Marcellus Shale in the U.S.
Bok Haeng Baek
Research Professor
George Mason University
Regi Oommen
Senior Scientist
Eastern Research Group
Joseph Wilkins
Assistant Professor
Howard University
Shinkuang Chang
Research Engineer
George Mason University
Chi-Tsang Wang
Postdoctoral Researcher
George Mason University

This study aims to assess trends in air quality and community exposures in the Marcellus Shale region and whether any might be explained by changes in oil and gas development-related operations or governance. The analysis will focus on local and regional exposures to criteria and hazardous air pollutants from 2002-2021, with special attention toward historically disadvantaged communities. The investigators are achieving their research aims with the following steps:

  1. Perform a long-term emissions trend analysis by integrating bottom-up oil and gas emissions and ambient measurements of criteria and select hazardous air pollutant concentrations.
  2. Conduct a long-term air quality trend analysis by applying a chemical transport model to simulate the criteria and select hazardous air pollutant concentrations over the region. 
  3. Use a more advanced chemical transport model that employs machine-learning to investigate sources of and control strategies for oil and gas emissions. 
  4. Apply a county-level community health vulnerability index approach to identify disproportionately exposed communities.

Baek Study Locations

Baek Study Locations

 

Video
Image
Baek headshot

Research Team

Bok Haeng Baek
Research Professor
George Mason University
Regi Oommen
Senior Scientist
Eastern Research Group
Joseph Wilkins
Assistant Professor
Howard University
Shinkuang Chang
Research Engineer
George Mason University
Chi-Tsang Wang
Postdoctoral Researcher
George Mason University

Project Updates

Baek Quarterly Update - December 2024

Long-term criteria and toxic pollutants trends and community exposures over the Marcellus Shale in the U.S.

Bok Haeng Baek

This study aims to assess trends in air quality and community exposures in the Marcellus Shale region and whether any might be explained by changes in oil and gas development-related operations or governance. The analysis will focus on local and regional exposures to criteria and hazardous air pollutants from 2002-2021, with special attention toward historically disadvantaged communities. The investigators are achieving their research aims with the following steps:

  1. Perform a long-term emissions trend analysis by integrating bottom-up oil and gas emissions and ambient measurements of criteria and select hazardous air pollutant concentrations.
  2. Conduct a long-term air quality trend analysis by applying a chemical transport model to simulate the criteria and select hazardous air pollutant concentrations over the region. 
  3. Use a more advanced chemical transport model that employs machine-learning to investigate sources of and control strategies for oil and gas emissions. 
  4. Apply a county-level community health vulnerability index approach to identify disproportionately exposed communities.

What's Happened

  • Developed historical abandoned wells emissions inventory.
  • Completed CAMx simulations for 2014 and 2019 years based on the latest U.S. EPA’s EQUATES criteria and hazardous air pollutant inventories. 
  • Started generating the rest of CAMx-ready meteorology and emissions inputs for additional study years.
  • Trained the DeepCTM model (a machine-learning based chemical transport model) for four oil and gas-related air toxics (benzene, toluene, ethylbenzene, xylenes) based on CAMx inputs and outputs for 2008, 2014, and 2019. 
  • Processed abandoned well emissions through SMOKE and generated the 2008, 2014 and 2019 CAMx-ready gridded emissions inputs for CAMx simulations.
  • Reviewed the CAMx outputs to estimate a county-level community health vulnerability index  for PM2.5 and ozone, two key indicators of air quality.
  • Presented “Long-term criteria and toxic pollutants trends in air quality and community exposures over the Marcellus Shale region in the U.S.” at the 2024 CMAS Conference in Chapel Hill, NC (October 2024).

  • Submitted the abstract titled “Spatial and Temporal Patterns in Air Quality and Community Exposure Disparities across the Marcellus Shale region in the U.S.” to the 2025 AAG conference in Detroit, MI (March 2025).

    Image
    baek1

Figure 1. Average PM2.5 at the county, census tract, and census block groups level. Developed by UNT and HU teams. 
 

Image
baek2

Figure 2. Total number of PM2.5 days above 35 µg/m3 (unhealthy for sensitive groups) at the county, census tract, and census block groups level. Developed by UNT and HU teams. 

What's New

  • Training the DeepCTM model for ozone based on the 2008, 2014, and 2019 CAMx inputs and outputs.
  • Conducting additional CAMx simulations to improve the accuracy of ozone and air toxics concentrations with abandoned well emissions inventory data.
  • Processing the CAMx outputs to estimate concentrations for PM2.5 and ozone at county, census tract, and census block group levels (See Figure 1 for an example for PM2.5).
  • Associating the estimated PM2.5 and ozone levels with various socioeconomic and demographic variables.
  • Estimating the number of people residing in areas with elevated PM2.5 and ozone over time (See Figure 2 for an example for PM2.5)
  • Examining the relationship between community health vulnerability index and PM2.5 and ozone at the county level.

What's Next

  • Analyze the long-term trends of emissions and measured concentrations over the Marcellus Shale region. 
  • Review oil and gas regulations over the years to assess how they may have affected emissions and air quality trends.
  • Simulate two additional CAMx annual runs for 2008 and 2014 with the new additional abandoned well emissions inventory.
  • Generate CAMx-ready gridded hourly data for abandoned well emissions over the remaining study period using the SMOKE emissions modeling system.
  • Complete the DeepCTM training for air toxics and ozone.
  • Simulate the long-term air quality concentrations using the DeepCTM based on the CAMx-ready emissions and meteorology inputs.
  • Complete the community health vulnerability index analysis for PM2.5 and ozone over the counties within the Marcellus Shale region.
Baek Quarterly Update - September 2024

Long-term criteria and toxic pollutants trends and community exposures over the Marcellus Shale in the U.S.

Bok Haeng Baek

This study aims to assess trends in air quality and community exposures in the Marcellus Shale region and whether any might be explained by changes in oil and gas development-related operations or governance. The analysis will focus on local and regional exposures to criteria and hazardous air pollutants from 2002-2021, with special attention toward historically disadvantaged communities. The investigators are achieving their research aims with the following steps:

  1. Perform a long-term emissions trend analysis by integrating bottom-up oil and gas emissions and ambient measurements of criteria and select hazardous air pollutant concentrations.
  2. Conduct a long-term air quality trend analysis by applying a chemical transport model to simulate the criteria and select hazardous air pollutant concentrations over the region. 
  3. Use a more advanced chemical transport model that employs machine-learning to investigate sources of and control strategies for oil and gas emissions. 
  4. Apply a county-level community health vulnerability index approach to identify disproportionately exposed communities.

What's Happened

  • Initiated the study in collaboration with researchers at George Mason University, Howard University, and Eastern Research Group.
  • Gathered unconventional oil and gas development (UOGD)-related emissions inventory and air quality monitoring datasets to assess air quality trends over the Marcellus Shale region.  
  • Started simulating concentrations of nitrogen dioxide, ozone, particulate matter (PM2.5), and selected air toxics including benzene, toluene, ethylbenzene, and xylene (BTEX) compounds, naphthalene, styrene, acrolein, and formaldehyde) for the period of 2002-2021. To start, they have completed simulations of hazardous air pollutants (HAPs) in the Marcellus Shale region for 2004, 2018, and 2019, which represent significant years in regulatory policy over UOGD sources.
  • Completed modeling the 2008 annual CAMx simulations for targeted air pollutants.
  • Presented the rationale for the study and research plan in a poster session at HEI's 2024 Annual Conference.

What's New

  • Reviewing and analyzing the 20-year UOGD-related emissions inventory and air quality monitoring datasets to understand long-term trends in air quality and possible associations with UOGD regulatory policy implementations.
  • Evaluating the performance of the 2008 CAMx air quality modeling simulations for ozone, PM2.5, and HAPs by comparing them with air quality monitoring datasets gathered by the ERG team. 
  • Based on preliminary outputs from the CAMx model, train an artificial intelligence modeling system to mimic the CAMx modeling performance and generate spatiotemporal air quality datasets over the Marcellus Shale region.

What's Next

  • Continue review of UOGD-related emissions inventory and monitoring datasets.
  • Conduct modeling simulations of air quality using CAMx for 2014 and 2019.
  • Using the CAMx model, analyze ozone and PM2.5 concentrations to understand potential UOGD-related sources and how they may have impacted local and regional air quality over the study period.