Bingham Research Center: 2024 Annual Report

(435) 722-1740 320 North Aggie Blvd Vernal, UT 84078 binghamresearch.usu.edu ubair.usu.edu 2024 ANNUAL REPORT Submitted to: Utah State Legislature and Uintah Special Service District 1 Submitted by: Seth Lyman, PhD Bingham Research Center Utah State University 320 N Aggie Blvd Vernal, UT 84078 Contributors: Seth Lyman, PhD Colleen Jones, PhD John R. Lawson, PhD Trevor O’Neil Pamela Gardner, PhD (editor) DOCUMENT NUMBER: REVISION: DATE: BRC241130A ORIGINAL RELEASE 30 NOV 2024 2 Acknowledgments This document reports on the activities the scientists at the Bingham Research Center have carried out with financial support from the Utah State Legislature and the Uintah Special Service District 1. We are grateful to these two entities for their ongoing support of our Uinta Basin air-quality work. In addition to the support of the Utah State Legislature and the Special Service District 1, we administer an endowment from the Anadarko Petroleum Corporation that funds student participation in air-quality research. We appreciate grant funding from the Marriner S. Eccles Foundation, the National Science Foundation, the U.S. Environmental Protection Agency, and the U.S. Department of Energy. Site access, electricity, and/or equipment at some of our monitoring stations are provided by the Utah Division of Air Quality, Scout Energy Partners, Koda Resources, and the Bureau of Land Management. Many energy companies have provided data and access to oil and gas facilities for our work. Marc and Debbie Bingham, the namesakes of the Bingham Research Center, provided initial funds to establish the Center and its facilities. We are grateful for the support of each of these groups, and the continuing support of the administrators and staff at the Uintah Basin Campus, Statewide Campuses, and President Elizabeth Cantwell and her administrative team at Utah State University. 3 Table of Contents EXECUTIVE SUMMARY 4 2024 PERFORMANCE 6 STAKEHOLDER ENGAGEMENT 21 TECHNICAL REPORTS 28 DRONE-BASED MEASUREMENT OF EMISSIONS FROM OIL AND GAS SOURCES 29 NSF - COLLABORATIVE RESEARCH: VERIFICATION OF ATMOSPHERIC MERCURY REDOX RATES 33 PUBLIC LANDS INITIATIVE – CATTAIL CONTROL MANAGEMENT STRATEGIES FOR ENDANGERED FISH HABITAT ON UTAH’S PUBLIC LANDS 35 WINTER OZONE PREDICTION MODEL 39 THE SNOW SHADOW OF THE UINTA BASIN 47 AI CHATBOTS FOR PUBLIC SCIENCE COMMUNICATION 49 EMISSIONS FROM AN ARROW L-795 PUMPJACK ENGINE: TESTS WITH AND WITHOUT MODIFICATIONS FOR ELECTRONIC FUEL INJECTION 52 UNDERSTANDING THE KEY COMPONENTS OF WINTER OZONE FORMATION 73 AIR QUALITY TRENDS 79 WINTER 2023-24 AIR QUALITY AND METEOROLOGY 91 SUMMERTIME AIR QUALITY 109 UPDATES TO LAB SPACES AND FIELD SITES 114 4 Executive Summary The Bingham Research Center was established in 2010 through a generous donation from the Marc and Debbie Bingham Family. Mission of the Bingham Research Center The mission of Utah State University’s Bingham Research Center is to conduct high-quality academic research that can be used by industry, government, and the public to develop efficient and effective solutions to environmental problems. The Center focuses on research that benefits Utah and the Uinta Basin, but scientists at the Center carry out projects around the world and strive for their work to be globally relevant. Purpose of this Report This report details the activities undertaken by the Bingham Research Center over the past twelve months. The report focuses on winter ozone research, as this is a core research area for the Center, and it serves as an annual report to the Utah Legislature and Uintah Special Service District 1, the primary funders of the Center’s winter ozone research. The report also contains information about other projects funded by other entities, the Center’s goals, and performance. This and past reports are available at: https://www.usu.edu/binghamresearch/papers-and-reports.The Center’s Management Plan is available here: https://usu.box.com/s/877z4o8nwynu3uwcze8uxj8jaant7auw. Background Information about Wintertime Ozone Ozone negatively impacts respiratory health, especially for those with lung diseases. During wintertime temperature inversion episodes, ozone in the Uinta Basin sometimes increases to levels that exceed the standard of 70 ppb set by the U.S. Environmental Protection Agency (EPA). Because of this, the portions of Uintah and Duchesne counties have been designated as nonattainment areas in the past. The Uinta Basin is one of only two places in the world known to routinely experience wintertime ozone exceeding EPA standards (Wyoming’s Upper Green River Basin is the other). Ozone forms in the atmosphere from reactions involving oxides of nitrogen (NOX) and organic compounds, and the majority of NOX and organic compound emissions in the Uinta Basin are from oil and gas development. Inversion conditions trap these pollutants near ground level, increasing their concentrations and allowing them to generate ozone. The unique mix of pollutants during inversion episodes in the Uinta Basin leads to the formation of wintertime ozone, in contrast to the fine particulate matter (PM2.5) pollution that is prevalent during winters on the Wasatch Front. 5 The number of ozone exceedance days and concentrations of ozone that occur each year are closely tied to meteorology. Years with persistent snow cover and high barometric pressure tend to have more days with strong winter inversions and high ozone. In the absence of snow cover and winter inversions, ozone concentrations in the Basin are similar to those in other rural, high-elevation locations around the western United States. Changes in emissions of organic compounds and NOX can also impact ozone levels. This past winter 2023-2024, saw no days that exceeded the EPA standards. Because of this data, the EPA is considering designating the Uinta Basin as compliant. Because wintertime ozone is relatively new to science, some aspects of the meteorology, chemistry, and emissions that allow ozone to form during winter are still poorly understood. Federal and state agencies are required by law to promulgate regulations that reduce ozone-forming emissions in the Uinta Basin. These regulations will mostly target the local oil and gas industry, which contributes to the majority of the Basin’s economy. Scientific research to better elucidate the causes and characteristics of winter ozone can help industry and regulators craft emissions reductions that maximize effectiveness and minimize costs to the local industry and economy. Since 2010, we (scientists at the Bingham Research Center) have conducted research to improve the understanding of winter ozone in the Uinta Basin. A cumulative summary of all significant research findings that relate to Uinta Basin air quality from 2010 through the present is available here: https://www.usu.edu/binghamresearch/cumulative-research-summary. Bingham Research Center Productivity The scientists, technicians, and student employees have been busy this past year. They have had thirteen papers published in scientific journals. Center scientists and staff made 23 presentations at local groups, regional, national, and international conferences. Three PhD candidates are pursuing their research under the guidance of Bingham Research scientists. Please see the section on 2024 Performance for specific details. Colleen Jones, PhD, was promoted to Lead Scientist and Research Associate Professor. John R. Lawson, PhD, was promoted to Research Assistant Professor. Bingham Research scientists are sought after to collaborate with colleagues outside the university. Currently, Center scientists have collaborated with others on grants from the Department of Energy (University of Utah, Brigham Young University, Stanford University, University of Nevada—Reno, IBM). The EPA (Utah Radon Lab, USU’s Office of Sustainability, TriCounty Health, Healthy Communities of Northeastern Utah, USU’s Transforming Communities Institute), and the National Science Foundation (BYU). 6 2024 Performance This section contains information about our performance on overall goals for Uinta Basin air quality research and performance for annual project objectives for the 2024 reporting period. Our management plan, which describes our group’s overall goals and objectives, is available at: https://usu.box.com/s/877z4o8nwynu3uwcze8uxj8jaant7auw. Research Output The basic outcomes of our research are publications and presentations that describe our work and make it available to other academics, stakeholders, and the public. The Bingham Research team had 13 reports accepted for publication in academic journals. All the peer-reviewed publications and significant technical reports are available on our website at https://www.usu.edu/binghamresearch/papers-and-reports. Peer-reviewed Publications 1. Lawson, J.R. and Lyman, S.N., 2024. A Preliminary Fuzzy Inference System for Predicting Atmospheric Ozone in an Intermountain Basin. Air, 2(3), pp.337-361. https://doi.org/10.3390/air2030020. 2. Jones, C., Tran, H., Tran, T. and Lyman, S., 2024. Assimilating Satellite-Derived Snow Cover and Albedo Data to Improve 3-D Weather and Photochemical Models. Atmosphere, 15(8), p.954. 3. Ghimire, S., Lebo, Z.J., Murphy, S., Rahimi, S. and Tran, T., 2023. Simulations of winter ozone in the Upper Green River basin, Wyoming, using WRF-Chem. Atmospheric Chemistry and Physics, 23(16), pp.9413-9438. 4. Lee, C.F., Eligar, T., David, L.M., Wilmot, T.Y., Reza, M., Hirshorn, N., McCubbin, I.B., Shah, V., Lin, J.C., Lyman, S. and Hallar, A.G., 2024. Elevated Tropospheric Iodine over the Central Continental United States: Is Iodine a Major Oxidant of Atmospheric Mercury? Geophysical Research Letters, https://doi.org/10.1029/2024GL109247. 5. Derry, E.J., Elgiar, T.R., Wilmot, T.Y., Hoch, N.W., Hirshorn, N.S., Weiss-Penzias, P., Lee, C.F., Lin, J.C., Hallar, A.G., Volkamer, R. and Lyman, S.N., 2024. Elevated oxidized mercury in the free troposphere: Analytical advances and application at a remote continental mountaintop site. Atmospheric Chemistry and Physics, 24(16), pp.96159643. 6. Gačnik, J., Lyman, S., Dunham-Cheatham, S.M. and Gustin, M.S., 2024. Limitations and insights regarding atmospheric mercury sampling using gold. Analytica Chimica Acta, 1319, p.342956. 7. Gustin, M.S., Dunham-Cheatham, S.M., Lyman, S., Horvat, M., Gay, D.A., Gačnik, J., Gratz, L., Kempkes, G., Khalizov, A., Lin, C.J. and Lindberg, S.E., 2024. Measurement of Atmospheric Mercury: Current Limitations and Suggestions for Paths Forward. Environmental Science & Technology, 58(29), pp.12853-12864. 7 8. Elgiar, T.R., Lyman, S.N., Andron, T.D., Gratz, L., Hallar, A.G., Horvat, M., Vijayakumaran Nair, S., O’Neil, T., Volkamer, R. and Živković, I., 2024. Traceable Calibration of Atmospheric Oxidized Mercury Measurements. Environmental Science & Technology, 58(24), 10706–10716. 9. Gustin, M.S., Dunham-Cheatham, S.M., Allen, N., Choma, N., Johnson, W., Lopez, S., Russell, A., Mei, E., Magand, O., Dommergue, A. and Elgiar, T., 2023. Observations of the chemistry and concentrations of reactive Hg at locations with different ambient air chemistry. Science of The Total Environment, 904, p.166184. 10. Cope, E.M., Ketcherside, D.T., Jin, L., Tan, L., Mansfield, M., Jones, C., Lyman, S., Jaffe, D. and Hu, L., 2024. Sources of atmospheric volatile organic compounds during the salt lake regional smoke, ozone and aerosol study (SAMOZA) 2022. Journal of Geophysical Research: Atmospheres, 129(17), e2024JD041640. 11. Lawson, J.R., Potvin, C.K. and Nelson, K., 2024. Decoding the Atmosphere: Optimising Probabilistic Forecasts with Information Gain. Meteorology, 3(2), pp.212-231. https://doi.org/10.3390/meteorology3020010. 12. Jaffe, D.A., Ninneman, M., Nguyen, L., Lee, H., Hu, L., Ketcherside, D., Jin, L., Cope, E., Lyman, S., Jones, C. and O’Neil, T., 2024. Key results from the salt lake regional smoke, ozone, and aerosol study (SAMOZA). Journal of the Air & Waste Management Association, 74(3), pp.163-180. 13. Stratman, D. R., N. Yussouf, C. A. Kerr, B. C. Matilla, J. R. Lawson, and Y. Wang, 2024: Testing stochastic and perturbed parameter methods in an experimental 1-km Warn-on-Forecast System using NSSL’s phased-array radar observations. Mon. Weather Rev., 152, 433–454, https://doi.org/10.1175/mwr-d-23-0095.1. Reports 1. Lyman S., Jones C., Lawson L., Mansfield M., David L., O’Neil T., Holmes B., 2023. 2023 Annual Report: Bingham Research Center. Utah State University, Vernal, Utah. https://www.usu.edu/binghamresearch/files/reports/BinghamCenter2023AnnualRepor t.pdf and https://doi.org/10.5281/zenodo.13999275. 2. Lyman S., Lin J., Tran H., 2024. Top-down Estimates of Emissions from Oil and Gas Production in the Uinta Basin: Final Project Report. Utah State University, Vernal, Utah. https://www.usu.edu/binghamresearch/files/reports/Topdowninventory_finalreport_FI NAL1.pdf. 3. Lawson, J. R., 2024: Communicating risk with possibility, not probability. arXiv [stat.AP], https://doi.org/10.48550/arXiv.2410.21664. 4. Lawson, J. R., J. E. Trujillo-Falcón, D. M. Schultz, M. L. Flora, K. H. Goebbert, S. N. Lyman, C. K. Potvin, and A. J. Stepanek, 2024: Pixels and predictions: Potential of GPT-4V in meteorological imagery analysis and forecast communication. arXiv [cs.CL], http://arxiv.org/abs/2404.15166. 8 Presentations 1. Elgiar T., Lyman S., Gratz L., David L., Volkamer R., Hallar G., December 2023, GEOSChem produces much less atmospheric oxidized mercury than calibrated mesurements. AGU Fall Meeting, San Fransicso, California. 2. Gratz L., Derry E., Lyman S., Elgiar T., Wilmot T., Volkamer R., December 2023. Determining the origins of ambient oxidized mercury at a continental mountaintop site in the western U.S. AGU Fall Meeting, San Fransicso, California. 3. Lyman S., February 2024. Our efforts to develop methods for oxidized mercury compounds. PittCon, San Diego, California. 4. Lyman S., March 2024. Uinta Basin Air Quality. Ute Indian Tribe Education Department, Fort Duchesne, Utah. 5. Lawson, J. R., M. J. Davies, and S. N. Lyman, March 2024: The Uinta Basin Snow Shadow: 2022/23 & 2023/24. 8th Science for Solutions Conference, Ogden, Utah, 28 March 2024. https://doi.org/10.5281/zenodo.13937073 6. Dhar L., Lyman S., March 2024, Understanding the Complexity of Winter Ozone Formation by Analyzing Different Chemical Mechanisms in a Box Model, Air Quality Science for Solution Conference, Ogden, Utah. 7. Lyman S., May 2024. Air Quality in the Uinta Basin. Vernal Area Chamber of Commerce, Vernal, Utah. 8. Lyman S., May 2024. Methane emissions panel. Wilkes Climate Summit, Salt Lake City Utah. 9. Lyman S., July 2024. Introduction to the Bingham Research Center. Utah Methane Coalition, Utah (online). 10. Lawson, J. R., M. J. Davies, July 2024: Fuzzy Winter-Ozone Predictions. 21st Conference on Mountain Meteorology, American Meteorological Society, Boise, ID. 11. Lyman S., July 2024. Introduction to the Uinta Basin Ozone Working Group. Utah Division of Oil, Gas and Mining Uinta Basin Collaborative Meeting, Duchesne, Utah. 12. Lyman S., Elgiar T., July 2024. Photochemical models are unable to simulate high oxidized mercury measured in the western United States. International Conference on Mercury as a Global Pollutant, Capetown, South Africa. 13. Lyman S., Elgiar T., July 2024. Advancing permeation tube-based mercury calibration. International Conference on Mercury as a Global Pollutant, Capetown, South Africa. 14. Dunham-Cheatham S., Lyman S., Lown L., Gacnik J., O’Neil T., Gustin M., July 2024. Toward a GC-MS method for quantification and characterization of ambient gaseous oxidized mercury compounds. International Conference on Mercury as a Global Pollutant, Capetown, South Africa. 9 15. Gacnik J., Dunham-Cheatham S., Lyman S., Gustin M., July 2024. Particulate-bound mercury sampling using membrane materials: biases due to adsorption of gaseous oxidized mercury. International Conference on Mercury as a Global Pollutant, Capetown, South Africa. 16. Gacnik J., Dunham-Cheatham S., Lyman S., Gustin M., July 2024. Gold sampling for atmospheric mercury analysis: insights and limitations. International Conference on Mercury as a Global Pollutant, Capetown, South Africa. 17. Gustin M., Dunham-Cheatham S., Lyman S., July 2024. Workshop Results: Measurement of Atmospheric Mercury: Assessment of new measurement and calibration methods and development of a path forward. International Conference on Mercury as a Global Pollutant, Capetown, South Africa. 18. O’Neil T., Lyman S., Elgiar T., Zager K., July 2024. Tips and Tricks of the trade: Building instrumentation for atmospheric mercury. International Conference on Mercury as a Global Pollutant, Capetown, South Africa. 19. Lyman S., September 2024. Uinta Basin air quality update. Uintah Basin Energy Summit, Vernal, Utah. 20. Dhar L., Lyman S., October 2024, Influence of Chemical Mechanism on Carbonyl and Ozone Formation During Winter, 23 Annual CMAS Conference, Chapel Hill, North Carolina. 21. Lawson, J.R., November 2024. AI is a wonderful innovation, but can we trust it? Science Unwrapped, Utah State Univ., Logan, Utah. 22. Lawson, J.R., November 2024. Fuzzy pollution predictions: communicating risk as possibility, not probability. Applied Math colloquium series, Utah State Univ., Logan, Utah. 23. Lawson, J.R., November 2024. CLYFAR (AI ozone prediction model) tutorial. Lunch & Learn Series, Uinta Basin Working Group, Utah. 24. Allred, J., Jones, C. P., November 2024. Aerial Imagery for Vegetation Monitoring PostWildfire. Utah Watershed Restoration Indicative Northeastern Region Annual Meeting, San Antonio, Texas. Funding The Bingham Research Center received $685,217 in grants and contracts, $50,000 in gifts, $400,000 in appropriations from the Utah Legislature, and $26,980 in endowment disbursements in the past 12 months for a total of $1,162,197 (Table 1). 16 •Elspeth Montague is a high school senior at Uintah High School. She started working with us in early 2024 developing web materials to display real-time and forecast air quality information. •Tristan Coxson is a high school senior at Uintah High School. He started working with us in early 2024 on laboratory analysis and instrumentation. All Students and Postdoctoral Researchers These are the students and postdoctoral researchers who have worked at the Bingham Research Center. The year the person first worked at the Center is also listed. 1. Emily Smith, undergraduate, 2012 2. Chad Mangum, undergraduate, 2013 3. Cathy Crawford, undergraduate, 2013 4. Jordan Evans, undergraduate, 2013 5. Trevor O’Neil, undergraduate, 2013 6. Trang Tran, postdoctoral researcher, 2013 7. Huy Tran, postdoctoral researcher, 2014 17 8. Colleen Jones, postdoctoral researcher, 2015 9. Cody Watkins, master’s student, 2014 10. Tate Shorthill, undergraduate, 2014 11. Tanner Allen, undergraduate, 2014 12. Lena Morgan, undergraduate, 2015 13. Felito Martinez, undergraduate, 2015 14. Sheree Meyer, graduate, 2015 15. Eric Hacking, undergraduate, 2016 16. Sandra Young, undergraduate, 2017 17. Justin Allred, undergraduate and graduate, 2017 18. Makenzie Holmes, undergraduate, 2018 19. Tyler Elgiar, undergraduate and graduate, 2018 20. Krystal White, undergraduate, 2019 21. Brant Holmes, undergraduate, 2020 22. Keirra Tolbert, undergraduate, 2021 23. Jackson Liesik, undergraduate, 2021 24. Davis Smuin, undergraduate, 2021 25. Lisa Boyd, graduate, 2022 26. Kristin Miller, undergraduate, 2023 27. Rachel Merrell, undergraduate, 2023 28. Loknath Dhar, graduate, 2023 29. KarLee Zager, undergraduate, 2023 30. Michael Davies, undergraduate, 2024 31. Elspeth Montague, high school, 2024 32. Tristan Coxson, high school, 2024 Data Management, Quality, and Dissemination Data Management As described in our management plan, all measurement data and notes we have generated during the reporting period have been stored on a cloud-based data storage server, with regular backups to local, removable hard drives. We stored all instrument maintenance, calibration, and repair information within this archival structure. We used established standard operating procedures for our work. These are publicly available here: https://www.usu.edu/binghamresearch/team_pages/standard-operating-procedures. Atmospheric Data Quality Table 2 shows a summary of data quality results for ambient air measurements we collected during 2024. The maximum uptime possible for most measurements shown in the table is approximately 95% due to maintenance and calibration periods. 18 Table 2. Data quality summary for ozone, oxides of nitrogen (NOX), carbon monoxide (CO), and organic compound data collected during 2022-23. Results are shown as averages ± 95% confidence intervals for all locations at which the indicated measurements were collected (confidence intervals are shown if the number of data points is three or more). For a list of measurements collected and sites of collection, Percent uptime indicates the percentage of the measurement period for which valid measurements were obtained. NMHC indicates non-methane hydrocarbons. N/A means not applicable. Measurement Zero calib. (ppb) Span calib. (% recov.) Percent uptime Ozone -0.4 ± 0.7 100 ± 2 93 ± 10 NO -0.0 ± 0.0 100 ± 0 89 ± 19 NO X (NO calib.) -0.2 ± 0.1 99 ± 1 89 ± 19 NO y (NO calib.) -0.4 ± 0.4 98 ± 1 91 NO X (GPT calib.) N/A 105 ± 1 89 ± 19 NOy (GPT calib.) N/A 97 ± 1 91 CO -20 ± 19 99 ± 3 97 Methane 33 ± 11 100 ± 0 68 Total NMHC 66 ± 29 103 ± 1 68 Speciated NMHC 0.1 ± 0.0 100 ± 0 94 Speciated Carbonyls 0.0 ± 0.0 98 ± 0 94 PM 2.5 (BAM) N/A N/A 97 Data Dissemination We have uploaded the winter ozone dataset for the most recent winter and an updated air chemistry and meteorology dataset for the Roosevelt, Castle Peak, and Horsepool monitoring stations to the data access page of our website, https://www.usu.edu/binghamresearch/dataaccess. We have also updated speciated organic compound data on the same web page. During the year, we gave meteorological and chemical datasets we collected to regulators, environmental consultants, and energy companies for use in their own analyses. Outcomes from Annual Air Quality Project Objectives We identified project goals for the current reporting period in Section 20 of our previous annual report, which is available at: https://www.usu.edu/binghamresearch/files/reports/BinghamCenter2023AnnualReport.pdf. In Table 3 we report on any discrepancies between planned work and actual outcomes for each of the project objectives outlined in the 2023 annual report. 19 Table 3. Outcomes of annual project objectives for the current reporting period. OBJECTIVE OUTCOMES Priority 1: Air Chemistry Operate air quality monitoring stations We completed this objective for winter 2023-24. We will continue the operation of these stations for the coming winter. Continue Investigation of Carbonyl Fluxes at the Air-snow Interface Measurements for this objective are complete, but we are still analyzing the data we obtained. We expect to release a report in early 2025. Investigate Ozone Formation in Summertime Wildfire Smoke We set up a summertime measurement system at the Roosevelt monitoring station as planned for this project. We collected several days of speciated organic compound measurements in smokey and non-smokey periods, and we are currently analyzing the data. We will continue this analysis in the coming year and may collect measurements during summer 2025 if the Roosevelt site again experiences significant wildfire smoke. Priority 2: Air Quality Modeling Improve WRF Simulations of Uinta Basin Winter Inversions We contributed to a paper published on adding noise to WRF forecasts to improve forecasts. However, Further work using WRF was paused due to concern that AI-based weather models were progressing much faster as frontier science. This caused a change of priorities, weighting the forecastsystem objective heavier. We will return to WRF simulations in 2025 as part of a U.S. Department of Energy grant on modeling the air quality impacts of carbon sequestration. The National Weather Service is currently changing national weather-forecast models and this new generation, freely available, requires further analysis of Basin weather for evidence earlier shortcomings have improved. This would motivate a return to this Priority to reassess previous findings. Develop a System for Quantitative Winter Ozone Forecasts We have produced an operational first version of our ozone-prediction model, named Clyfar. Its prototype is described in a recent paper, and we will display forecasts on a new UBAIR website for the upcoming 2024-25 Ozone Alert season. This system does not use WRF forecasts. Instead, we utilize available NOAA forecasts and use a rule-based form of artificial intelligence to produce ozone predictions. We will develop version 2 of Clyfar in 2025 ready for that year’s Ozone Alert, adding more AI elements that learn from earlier forecast accuracy. Box Model Investigation of the Impact of Chemical Mechanisms on Simulations of Winter Ozone We have completed this work and are currently preparing our findings for publication. We will continue this work with the CMAQ 3D photochemical model in the coming year. 20 Priority 3: Emissions Establish a Method for Monthly Basin-wide Pollutant Emissions Estimates This work is underway. We are using the Integrated Methane Inversion system (https://imi.seas.harvard.edu/) to ascertain Uinta Basin-wide methane emissions and determine organic compound and NOX emissions from ambient air correlations. We will have the first emission results from 2019 through 2024 in early 2025. Development of Methods to Determine Oil Storage Tank Emission Factors in the Uinta Basin This work is underway. We completed sample collection in November 2023. Lab analysis took much longer than expected but is complete as of September 2024. We have completed a statistical analysis of the laboratory data, and Birol Dindaruk at University of Houston is now performing equation of state modeling. We expect this project to be completed in mid-2025. Drone-based Measurement of Emissions from Oil and Gas Sources This is a multi-year objective. We have built the drone-based methane measurement system and tested it, and it is working properly. The next steps are to build a computer program to process the drone system data into emissions measurements, and then to test the entire system against known methane emission rates. At that point, we will be ready to conduct real emissions measurements in the field. We expect to be ready for this in mid-2025. Priority 4: Stakeholder Engagement Operate ubair.usu.edu website to display mapbased, real-time air quality information to the public We completed this goal for the reporting period. We are building a new real-time website. Operate the Ozone Alert program We completed this objective for the reporting period. Uinta Basin Ozone Working Group We completed this objective for the reporting period. More about the working group can be found at https://www.usu.edu/basinozonegroup/. 21 Stakeholder Engagement The mission of the Bingham Research Center is to generate knowledge and provide information that helps stakeholders (industry, regulators, and others) to make better decisions that will reduce emissions in the Uinta Basin. Thus, we strive to engage stakeholders in our research process to clarify and utilize the data we produce. 2024 Stakeholder Engagement Activities Ozone Alert Program At the request of oil and gas industry representatives and with input from the Utah Division of Air Quality, TriCounty Health, and several oil and gas companies, we created a program in 2017 to alert oil and gas companies when we expect high winter ozone. The program includes a web page (https://www.usu.edu/binghamresearch/ozone-alert) to describe the program and allow individuals to receive alerts. We collect their name, email address, and place of business. We send everyone on the list an email when we identify a possible local ozone formation. We forecast if an ozone episode extends longer than expected and when episodes end or are expected to end. We attempt to forecast ozone episodes up to six days in advance. The purpose of this program is to provide users with information that allows them to reduce ozoneforming pollution when it matters most. Winter 2023-24 had zero days with ozone exceeding the EPA standard of 70 ppb. There were some heavy snow events, but air temperatures were too high to support a snowpack after a heavy snowfall. One snowfall did lead to a weak inversion (where cold air begins to stagnate at the Basin floor, trapping pollution), but the snow melted before ozone levels built substantially. Periods during which we alerted subscribers that high ozone levels were likely are shown in Figure 1. 22 Figure 1. Time series of the highest daily maximum 8-hr average ozone the site observed at any monitoring site in the Uinta Basin during winter 2023-24. The EPA ozone standard is shown as a red dashed line. Periods during which USU issued ozone alerts are shown as grey shading. The program currently has 174 subscribers, with 41% from the energy industry, 23% are affiliated with government entities, 20% are members of the local public, 11% are academics, 2% are representatives of the media, and 1% are from environmental groups. Uinta Basin Ozone Working Group In 2018 we coordinated with individuals from government, industry, and environmental advocacy organizations to organize the Uinta Basin Ozone Working Group. The purpose of this group is to determine and promote actions that will reduce wintertime ozone in the Uinta Basin. The group’s website is: https://basinozonegroup.usu.edu. Marc Mansfield served as the group’s facilitator and led its steering committee through March 2024. When Marc retired from work at USU, Seth Lyman assumed this role in March 2024. Seth Lyman manages the group’s website and is responsible for group communications, with assistance from Michael Davies. Our team regularly gives presentations to the group about the science of wintertime ozone and actively participates in all working group meetings. The group’s website provides agendas for the meetings and links to presentations given. 23 Websites •ubair.usu.edu, our real-time air quality data website, has hundreds of unique users in thousands of sessions every year. •Our main website, https://www.usu.edu/binghamresearch and the ozone working group website, https://www.usu.edu/basinozonegroup, had hundreds of unique visitors over the reporting period. Information and Data Sharing •We gave dozens of presentations to many different stakeholder groups during the reporting period. Please see the Performance Report Section for a list of all presentations. •We provided our annual report, specific project reports, report summaries, and datasets to members of our stakeholder committee and others in government and industry upon request. •We have made updated datasets available at: https://www.usu.edu/binghamresearch/data-access. •We have made all our project reports and peer-reviewed papers publicly available at: https://www.usu.edu/binghamresearch/papers-and-reports. •We make Python and other code we develop for our research available on GitHub at https://github.com/Bingham-Research-Center. Community Outreach and Service •John Lawson, KarLee Zager, and Michael Davies served as judges at the Uintah School District’s middle school science fair in January 2024 and gave the new Bingham Research Center Award for the fair to Cameron DeBerard. •Seth Lyman is serving on the Advisory Board for Uinta Basin CarbonSAFE, a project to explore carbon capture and storage options in the Uinta Basin. •John Lawson participated in USU’s public-facing science presentation series Science Unwrapped in November 2024 in Logan, Utah. •Seth Lyman served as a judge for the Utah Petroleum Association’s Environmental Leadership Awards and presented the awards at the Association’s annual meeting in March 2024. •Seth Lyman was appointed to the Board of Trustees for Healthy Communities of Northeastern Utah in October 2024. •Colleen Jones, Trevor O’Neil, John Lawson, and Seth Lyman participated in USUUintah Basin’s STEAM Expo for Uinta Basin middle school students in October 2024. They staffed interactive displays with air quality and drone themes. •Seth Lyman has served on the Board of Directors for the Utah Clean Air Partnership (UCAIR) since 2021. 24 Use of Our Air Quality Research by Stakeholders These individuals and agencies use our research, excluding reports and formal presentations, which are reported in the Performance Report Section. A complete list of all stakeholder uses of our research is available at: https://usu.box.com/s/1k70hyz1dgca2kr9zuynz0locogd9uti. •The Environmental Protection Agency cited industry participation in the Bingham Center’s Ozone Alert Program as evidence of progress on atmospheric emissions that allowed them to accept a second extension of the Uinta Basin’s ozone attainment date. This action will very likely result in the Uinta Basin being officially declared in attainment of the EPA ozone standard (it has been in nonattainment since 2018). •The Utah Petroleum Association cited our research in their comments on the EPA proposed action to accept the second extension of the Uinta Basin’s ozone attainment date. •The Utah Department of Health and Human Services used data and the expertise of the Bingham Research Center in a study of asthma prevalence and causes in the Uinta Basin region. The study will be released publicly in 2025. Media Appearances The following are news articles from the reporting period that mention our work. A complete list of media mentions of our research is available at: https://usu.box.com/s/5s0busf524npd935mqfecsnvhn4cep52. •2024. USU and You with John Lawson. KVEL radio (two appearances). •2024. Bingham Research Center receives grant to study carbon capture and storage. Carbon Capture Journal. Also see basinnow.com. •2024. Utah’s air quality has been harming your health for years, but tourism is now in the crossfire. Salt Lake Tribune. •2024. The power of kindness: fostering a positive culture at Utah State University Uinta Basin. Vernal Express. •2024. PittCon Thought Leader: Seth Lyman. The Pittcon Podcast. •2024. USU Welcomes Colleen P. Jones as an Associate Research Professor. Vernal Express. •2024. Methods of Measuring Atmospheric Mercury Pollution. Azo Materials. •2024. Seth Lyman named one of 50 “influential Aggies.” Utah Statesman. •2024. Assistance that never sleeps: Embracing AI in the Uinta Basin. Vernal Express. •2024. New research tests AI’s ability to forecast severe weather. Heller Weather. •2024. Railroad Commission Approves Toxic Waste Ponds Next to Baptist Camp. Inside Climate News. 25 •2024. Bingham Research Center works to improve Utah’s climate. Utah Statesman. •2024. Meet The Secretive ‘Mastermind’ Behind Utah's Oil Boom. Huffpost. •2024. Zero exceedances of winter ozone this season. Basinnow.com. •2024. Winter ozone formation possible next week. Basinnow.com. •2023. Elizabeth Cantwell: My vision for a land-grant university in the 21st century. Salt Lake Tribune. •2023. The Uinta Basin: Part 2 —Ozone and Collaboration. EM. •2023. Research Landscapes. USU Today. •2023. Bingham Research Center Hires Senior Air Quality Scientist. USU Today. •2023. Wyoming Department of Environmental Quality Air Quality Division Receives National Award (for a model developed by Bingham Center). Wyoming Department of Environmental Quality. Annual Stakeholder Survey We conducted an online survey to learn how stakeholders feel about the Bingham Research Center and how they use our research products. We advertised the survey at our exhibitor’s booth and oral presentation at the September 2024 Uintah Basin Energy Summit. We received 22 verified responses. Survey respondents were asked, “Have you used research products or other output from the USU Bingham Research Center?” Responses included: •Air quality •Air quality tracker / alerts •Uinta Basin air quality monitoring •Enjoy reading reports and hearing presentations Survey respondents were asked, “Please provide any suggestions you may have for how the Bingham Center could improve its research products or dissemination of those products.” We didn’t receive any responses to this question. Survey respondents were asked, “Please provide any suggestions you may have for air quality research or other activities you think the Bingham Research Center should undertake.” We received two substantive responses, both from representatives of the oil and gas industry: •Collaboration with our operations / LDAR folks •More involved with industry Summary of Stakeholder Surveys from 2019 to 2023 We reviewed stakeholder survey responses from 2019 to 2023. Table 1 provides a distillation of our findings. Major themes of the surveys include: •Increased awareness of BRC oGreater promotion of branding oMore conferences & presentations 32 References 1.Gålfalk, M.; Nilsson Påledal, S.; Bastviken, D. Sensitive Drone Mapping of Methane Emissions without the Need for Supplementary Ground-Based Measurements. ACS Earth and Space Chemistry 2021, 5, 2668-2676, doi:10.1021/acsearthspacechem.1c00106. 33 NSF - Collaborative Research: Verification of Atmospheric Mercury Redox Rates Colleen Jones, PhD Seth Lyman, PhD Introduction This project is a three-year project that addresses the critical gaps in understanding the atmospheric redox chemistry of mercury (Hg), a potent neurotoxin that poses significant risks to both wildlife and human health. Mercury is emitted as elemental Hg (Hg0) but can be oxidized to its divalent form (HgII), which is more water-soluble and is rapidly deposited in ecosystems. Current uncertainties regarding the oxidants involved in this transformation and their reaction rates lead to varied predictions of HgII deposition patterns (Figure 1). Our research is a collaborative effort between Brigham Young University (BYU), the University of Utah (UU), and the University of Nevada, Reno (UNR) with National Science Foundation funding. We will verify the amount of HgII and its reaction rates by collecting compressive datasets in the field to support and change current modeling of atmospheric Hg. By generating a comprehensive dataset and refining existing measurement techniques, this project aims to reduce uncertainties in mercury redox reaction rates and improve global models of mercury deposition. The outcomes will not only enhance scientific understanding but also have broader implications for air quality predictions and public health policies. 34 Specific contributions of the proposed project to the body of atmospheric Hg research include: •A comprehensive dataset of atmospheric Hg and other relevant species (halogens, ozone, NOX, NOY, aerosols, and OH radical) in an area with some of the highest HgII ever measured in North America (Lan, 2012). •Improvements to and measurements with the only Hg measurement and calibration system that has been shown to quantitatively measure HgII compounds. •Decrease in the uncertainty of Hg-halogen reaction rates, which are key to Hg0 oxidation, and will help constrain estimates of when, where, and how atmospheric Hg impacts ecosystems. Year 1 Progress Report During Year I, our objective was to design, order supplies, and build the mobile field/lab instrumentation and measurement trailer. Students participating with our team this past year have been involved in UNR’s Mercury Journal Club with UNR students, and attended a Science for Solutions Conference in Ogden, Utah with researchers and students from all four universities. USU’s students have also learned constructing and programming skills while building the mobile field/lab instrument and measurement trailer. During Year II, we planned to deploy our mobile field/lab trailer at the shore of the Great Salt Lake. However, our original field site at the US Magnesium Plant is no longer feasible. The US Magnesium Plant is uncertain about their return to magnesium production. Currently, they are focusing on extracting other rare earth metals at their site. Because our ideal field location is no longer available, our other option is a chamber study which would allow us to test our hypothesis. We will still work collaboratively with our team on developing a chamber study at BYU utilizing their chamber and our dual-channel measurement system to verify reactions of Hg0 and HgII, and oxidants such as ozone, methane, HONO, HCHO, Nox, and Br. These compounds facilitate the oxidation of Hg0 and use advanced atmospheric modeling to interpret our findings. Acknowledgments All atmospheric mercury research at the Bingham Research Center is funded by the U.S. National Science Foundation. 35 Public Lands Initiative – Cattail Control Management Strategies for Endangered Fish Habitat on Utah’s Public Lands Colleen Jones, PhD Lisa Boyd, PhD Candidate Introduction Managing cattail encroachment in the wetlands of Stewart Lake is a costly endeavor, with annual expenses ranging from $25,000 to $50,000. Current management practices often lack scientifically validated methodologies and data analysis. Previous research indicates that the breeding success of two endangered fish species at Stewart Lake peaks when cattail cover is around 50%, with the remainder as open water. This study aims to assess the effectiveness of various cattail treatment methods at the Stewart Lake Wetland Management Area (SLWMA) and provide the Utah Division of Wildlife Resources (UDWR) with a cost-benefit analysis for each management option. We will collaborate with the UDWR, which has secured funding for cattail treatment through Utah’s Watershed Restoration Initiative (UWRI). The research will evaluate various treatment methods across study plots in the wetlands to identify the most effective approach for controlling cattail encroachment and maintaining it at 50%. We will test the hypothesis that spring treatments, applied before flooding the lake, are as effective as fall treatments conducted after the lake is drained. The treatment methods under consideration include: •Herbicide application •Grazing by goats •Controlled Burn/Fire Herbicide will be applied following grazing, using a utility task vehicle (UTV) equipped with a boom, adhering to the manufacturer’s guidelines for wetlands. The Utah Division of Wildlife Resources (UDWR) will provide funding for treatments and materials, managing the purchase, storage, and application of the herbicide. Utah State University (USU) will assess treatment effectiveness by monitoring cattail canopy cover in designated areas during spring to determine the percentage of cattail growth, and again in summer and fall to measure both cattail and open water cover. We will utilize remote sensing technology, including multispectral satellite imagery for the overall lake area and drone imagery for specific plots. Dr. Colleen Jones from the Bingham Research Center in Vernal, Utah, 36 is currently leading projects to gather multispectral drone imagery for post-wildfire vegetation and soil stability assessments. Progress Report The pretreatment drone survey was completed on July 8, 2024, with Stewart Lake filled with water. Stewart Lake began draining on September 16, 2024. UDWR. At that time, UDWR collected endangered juveniles in fish traps to collect growth indicator measurements and released them back into the adjacent Green River. During November 6-11, 2024, the initial cattail treatment survey transects and ground truthing plots were set up in partnership with UDWR. The treatment plots were designed to test single treatments as well as combination treatments (Figure 1). 37 Figure 1. Map of Stewart Lake Cattail Control Management Project’s treatment and ground truthing plots with an aerial RGB base map image collected by USU’s Bingham Research Center UAV team on July 8, 2024. A second UAV flight with a multispectral camera flew after the treatment plots were established. UDWR will perform the treatments during the fall and spring field seasons. During the winter of 2024/25 the USU graduate student will perform the remote sensing data analysis collected from the two 2024 UAV flights from the multispectral camera. Analysis includes machine learning vegetation classification to assess open water and treatment progress. As well as beginning the cost-benefit analysis of different treatments. Another cattail treatment survey will happen during the summer of 2025. USU will fly the UAV again after the summer survey to assess the treatments and finalize the cost-benefit analysis to finish the project in December 2025. 38 Acknowledgment The project is funded by USU’s Public Land Imitative Grant Program. 39 Winter Ozone Prediction Model John R. Lawson, PhD Overview Episodes of high wintertime ozone in Utah’s Uinta Basin can follow a heavy snowstorm when pressure rises and still conditions encourage formation of prolonged cold pools. The snow cover keeps the cold pool stagnant, trapping emissions from the oil and gas industry, leading to sunlight reacting with these emissions to build ozone concentrations. These conditions challenge traditional numerical weather prediction (NWP) models, which must balance computational trade-offs between achieving high resolution (necessary for capturing cold pools and boundary-layer dynamics) and ensemble forecasts (which quantify uncertainty). Highresolution NWP models can simulate small-scale atmospheric interactions essential to winter ozone dynamics, but this computational intensity limits the ability to run ensembles, making uncertainty quantification difficult. This trade-off constrains NWP model utility for capturing the episodic, uncertain nature of cold pools (and hence ozone) in the Basin, making risk-based forecasts difficult to issue. In response, we have created a lightweight solution that does not require a traditional weather and air-chemistry model. We deploy Clyfar (Welsh for “clever”) by implementing a fuzzy inference system (FIS) that synthesises domain knowledge into interpretable rules, creating ozone forecasts without requiring complex parameterisation. Clyfar addresses forecast uncertainty by reframing risk communication through possibility theory, which defines forecast outcomes in terms of plausibility. This framework, which flags high-ozone conditions as possible even under uncertain atmospheric setups, serves risk-averse stakeholders of Ozone Alert by indicating early risks of ozone exceedances. However, implementing possibility theory requires additional development to simplify the mathematics, enhancing accessibility for operational use. 40 Figure 1: A comparison of traditional yes/no logic (left, a) and fuzzy logic (right, b). With Clyfar, we improve winter-ozone forecasting by taking information from freely available federal NWP sources that are most related to cold-pool formation (and so high ozone). Fuzzy logic (Fig. 1), the core of Clyfar’s FIS, assigns degrees of truth to input conditions—such as “calm wind”—instead of strict binary states, enabling the system to avoid a synthetic sharp transition between, say, “shallow” and “deep” snow that is unknowable (e.g., the amount of snow required to create a persistent cold pool). Clyfar thus captures non-binary conditions linked to ozone formation and translates these into fuzzy rules, creating an inference model that does not require extensive computational resources when using existing NWP data as inputs. Model configuration and data sources To inform the prediction system, we obtained observation data from Synoptic Weather. Forecast data is taken from national NWP models. We begin with the NOAA Global Ensemble Forecast System (GEFS) model. 41 Description of Clyfar configuration Clyfar’s fuzzy inference system operates on inputs such as snow cover, barometric pressure, and clear-sky radiation. These are categorized into fuzzy states (adjectives such as “deep”) representing observed Basin condition states. The input variables are processed through fuzzy rules informed by known relationships between meteorological conditions and wintertime ozone formation. Clyfar generates forecasts that highlight ranges of plausible ozone outcomes. Possibility distributions over the output categories (background, moderate, elevated, extreme) signal varying degrees of ozone risk rather than a binary prediction. This approach enables Clyfar to signal potential ozone episodes with low certainty, enhancing early warning capabilities. Each rule in Clyfar’s system corresponds to expert-identified meteorological thresholds associated with elevated winter ozone. We create the categories based on adjectives of the systems that represent common states, such as wind speed being “calm” or “breezy”: only the two states matter to whether ozone remains in the Basin. We see this in the two panes of Fig. 2 showing wind speed against ozone concentration, and how we might create categories of “calm” and “breezy” relevant to ozone prediction based on the scatter plot. We see a divide around 1.5 m/s where speeds stronger than this blow pollution out of the Basin. There is not a sharp number where wind speeds preclude ozone formation. The category shape (“membership value”; curves in Fig. 3 panels) described how much each value (x-axis) is described by the category’s adjective: we walk through an example below. Rules activate based on conditions like, say, “calm wind and deep snow leads to elevated ozone”, named Rule 1 here. Figure 2: wind-speed measurements, plotted against ozone measurements (left, a) and defined categories based on data (right, b) 48 Future Work Michael is a first-year undergraduate and balances research with classes. However, we have plans to explore the following in 2025: •Use of jupyter notebooks to demonstrate analysis such as mapping snowfall •Mountain meteorology in the Basin as a whole •Python code structuring and re-use for other projects •Future application to Clyfar (Technical Report XX) •Presentation at AMS NOLA 2025 at the student conference was accepted •Future paper in the MDPI Air journal •Hydrology variation and impact on the ecosystem •Review papers on ancient climate (especially in the Uinta Mountains) Presentations Lawson, J. R., and Davies, M. J.: The Uinta Basin Snow Shadow: 2022/23 & 2023/24. 8th Science for Solutions Conference, March 2024, Ogden, Utah. https://doi.org/10.5281/zenodo.13937073 Lawson, J. R., and Davies, M. J., and Seth N. Lyman: Fuzzy winter-ozone predictions. 21st Conference on Mountain Meteorology, July 2024, American Meteorological Society, Boise, ID 49 AI Chatbots for Public Science Communication John R. Lawson, PhD Overview This project, documented by the pre-print report in the list at the section end, explores the application of OpenAI’s GPT-4 Vision (GPT-4V), experiments done late 2023, in interpreting meteorological charts and communicating weather hazards. Specifically, the work evaluates GPT-4V’s competency in delivering understandable and accurate weather forecasts, acknowledging the challenges posed by AI hallucinations—confident but incorrect outputs that could mislead in high-stakes contexts such as severe weather forecasting. The research’s core objectives were to (1) determine whether GPT-4V can interpret meteorological charts accurately and issue its own forecasts (2) evaluate GPT-4V’s capacity to communicate these forecasts in audience-specific language, enhancing clarity for diverse stakeholders, and (3) understand the limitations and potential risks in using AI for weather hazard communication. Preliminary Findings Initial findings indicate that GPT-4V shows considerable promise in interpreting and describing meteorological charts but also highlights significant challenges. For example: Accuracy of Interpretation. GPT-4V generally succeeded in interpreting large-scale weather patterns but struggled with finer details, sometimes providing incorrect reasoning in specific outlooks. This limitation suggests the need for human oversight to ensure forecast accuracy. Language and Cultural Nuances. When translating weather information into Spanish, GPT-4V performed literal translations that lacked the idiomatic precision necessary for effective communication in non-English-speaking contexts. This limitation underscores the need for improved translation mechanisms in AI, especially for nuanced topics like risk communication to specific communities. This is important is a diverse area such as the Basin. Communication Complexity: While GPT-4V’s outputs were coherent, they occasionally presented forecasts in a manner that was too simplistic for expert audiences or too complex for lay audiences, highlighting the challenge of balancing depth and clarity based on the user’s knowledge level. 50 Discussion The paper emphasizes GPT-4V’s utility in complementing human forecasters by rapidly generating preliminary interpretations and visual analyses from meteorological data. However, GPT-4V’s risk of hallucinations raises concerns, especially when providing high-stakes forecasts that influence decision-making. The need for reliable and precise language is critical in meteorology, as inaccurate or ambiguous forecasts can lead to unnecessary panic or, conversely, underestimation of severe weather threats. In the context of the Ozone Alert system integration of AI chatbots like GPT-4 could streamline the process of generating and disseminating ozone forecasts from Clyfar outputs (see Technical Report XX), which include complex, fuzzy-logic-based predictions of atmospheric ozone concentrations that require simpler language to communicate. Implementing GPT-4V for realtime feedback from users could allow the Ozone Alert system to adjust communication dynamically. For example, users could specify their level of understanding or urgency, prompting GPT-4V to adapt its language accordingly. This approach would significantly enhance user engagement and the system’s responsiveness to individual needs. Potential testing and future work Addressing GPT-4V’s translation limitations will be essential for broadening Ozone Alert’s reach across non-English-speaking communities. Fine-tuning translations or integrating AI trained specifically for meteorological and air-quality terminology could mitigate current challenges. Implementing human-in-the-loop processes for high-stakes predictions, such as severe inversion events, to cross-verify chatbot outputs before public release would safeguard against the model’s hallucination risks. Additionally, this approach would foster a model for continual improvement by providing corrective feedback to GPT-4V . In summary, GPT-4V’s integration within the Ozone Alert program holds significant potential to modernize forecast communication by enhancing the interpretability and accessibility of complex meteorological data. However, the model’s current limitations necessitate careful management to prevent miscommunication, particularly in high-risk forecasts like those associated with ozone levels in the Uinta Basin. 51 Presentations Lawson, J. R.: “A.I. is an astounding innovation, but can we trust it?”. Science Unwrapped, November 2024, USU College of Science, Logan, UT Lawson, J. R., J. E. Trujillo-Falcón, D. M. Schultz, M. L. Flora, K. H. Goebbert, S. N. Lyman, C. K. Potvin, and A. J. Stepanek, 2024: Pixels and predictions: Potential of GPT-4V in meteorological imagery analysis and forecast communication. arXiv [cs.CL]. (Conditionally accepted for publication in Artificial Intelligence for Earth Systems.) 52 Emissions From an Arrow L-795 Pumpjack Engine: Tests With and Without Modifications for Electronic Fuel Injection Seth Lyman, PhD Introduction This document reports on measurements made in February and August 2024 by scientists at Utah State University’s Bingham Research Center of emissions from an Arrow L-795 natural gasfueled engine that powered an oil pumping unit in the Uinta Basin, Utah. The engine was modified to utilize a prototype electronic fuel injection system, and we measured emissions with the fuel injection system in place and with the engine’s original carburetor. The fuel injection system was created by Integrated Power Solutions (57 N Skyline Dr, Roosevelt, Utah). Arrow L-795 and other pumpjack engines in the Uinta Basin are known to operate inefficiently (Lyman et al., 2022b), leading to high emissions of methane and non-methane organics from engine exhaust, and the prototype injection system was intended to improve the engine’s operations and decrease pollutant emissions. The Bingham Center agreed to collect the measurements detailed in this report to promote emissions reduction options in the Basin. 1. Methods Most of the methods used in this work were the same as those employed by Lyman et al. (2022a) and Lyman et al. (2022b), and much of the text in this section is taken from Lyman et al. (2022a). 1.1. Engine Measured Arrow L-795 engines have two cylinders, are two-stroke, and have a 65-horsepower capacity. The measured engine had a horizontal exhaust stack with a muffler and an internal diameter at the outlet of 102 mm. The stack was elevated about 0.6 m from the ground. 1.2. Field Measurements 1.2.1. Meteorology During each engine assessment, we measured ambient temperature and relative humidity (New Mountain NM150WX), wind speed and direction (Gill WindSonic), and barometric pressure (Campbell CS100) on a tower extended from the measurement trailer to a height of 6 m. We check these measurements against NIST-traceable standards annually. 53 1.2.2. Physical and Inorganic Chemical Measurements of Engine Exhaust We used an Ecom J2KN Pro Industrial analyzer to measure the physical and inorganic chemical properties of the exhaust. The analyzer utilized standard sensors for oxygen (O2), water vapor, carbon monoxide (CO), nitrogen monoxide (NO), and nitrogen dioxide (NO2) and an infrared sensor for carbon dioxide (CO2). The analyzer’s inlet line was connected to the exhaust stack between the engine and the muffler with a stainless tube fitting. We collected measurements every 2 sec over 10 min periods in most cases. The Ecom analyzer’s probe incorporates a thermocouple to measure exhaust temperature, and the analyzer also measured exhaust pressure. Figure 1 shows a photograph of the sample inlets. The manufacturer calibrated the Ecom analyzer after the study, and we calibrated chemical measurements one week before the study. Calibration results are available in Section 2.1.2. 54 Figure 1. Photograph of measurement inlets on the engine exhaust. The leftmost tube connecting to the exhaust is the pitot tube for flow measurement. The middle tube is for the Ecom analyzer, and the rightmost tube is the heated line for organic compound sampling. 1.2.3. Sampling Line for Organic Compounds We sampled organic compounds via a custom-built, heated sampling line. The tip was a 1.3 cm stainless steel tube with a length of approximately 0.5 m. This tip was connected to the exhaust stack between the engine and the muffler with a stainless tube fitting. 55 Downstream of the stainless tube was a stainless-steel sintered filter (7 µm pore size) and a PFA Teflon filter pack that held a 47 mm PTFE filter with 5 µm pore size. Installed after the filter pack was a 1 cm PFA Teflon tube of 15 m length. All portions of the inlet line except the stainless-steel tip were heated to 55°C, which was hotter than the dewpoint of the engine exhaust. The sample line led to a generator-powered, on-site trailer and sample gas was flushed through the line at 4 L min-1. 1.2.4. Methane We used a Los Gatos Research Fast Greenhouse Gas Analyzer with a high-concentration laser to measure methane concentrations in the exhaust gas. The analyzer pulled gas from the sample line in the trailer. 1.2.5. Non-methane Hydrocarbons and Alcohols Within the research trailer, evacuated silonite-coated 6 L stainless steel canisters with flow regulated by Alicat mass flow controllers collected gas samples from the sample line. Canisters were filled over 10-minute periods. The line to the canisters and the canisters themselves were unheated (the trailer was heated, but its temperature fluctuated with the ambient temperature and with opening and closing of doors). We analyzed the filled canisters in our laboratory (see Section 1.3.2). 1.2.6. Carbonyls An independent pump pulled exhaust gas from the sample line through Waters SepPak 2,4 dinitrophenylhydrazine (DNPH)-coated amorphous silica bead cartridges (350mg) which collected carbonyl compounds. We used two cartridges in series which allowed the second cartridge to capture any breakthrough. The line to the cartridges and the cartridges themselves were heated to 55°C to avoid water condensation. A mass-flow controller regulated flow through the cartridges. We kept cartridges refrigerated to the manufacturer-recommended temperature of <4oC before and after sampling until we completed the analysis. We were cautious to minimize any particulate contaminates on the cartridge and sample container during the handling and the collection of the samples. 1.2.7. Exhaust Flow Measurement We measured exhaust flow with the Ecom analyzer’s pitot tube, flow-measurement probe. The pitot tube was inserted into the exhaust stack between the engine and the muffler with a stainless tube fitting. The tip of the tube was placed in the center of the stack. 56 1.3. Laboratory Analyses 1.3.1. Dilution of Sample Canisters After sampling, we diluted fuel gas and exhaust gas sample canisters in the laboratory with ultrapure nitrogen gas to bring the mixing ratios of organic compounds into the range of our gas chromatography-mass spectrometry (GC-MS) system. We pressurized canisters to 3800 mbar and then diluted them with ultrapure nitrogen using an Entech 4600 dynamic diluter. The canisters required multiple stages of dilution for organic compound mixing ratios to be within the range of our GC-MS. 1.3.2. Analysis of Non-methane Hydrocarbons and Alcohols We analyzed canisters containing diluted fuel gas and exhaust gas samples for a suite of 55 hydrocarbons and three alcohols (see the list of compounds in Table 1-). We used an Entech 7200 pre-concentrator (in cold trap dehydration mode) and 7016D autosampler to concentrate samples and introduce them to a gas chromatograph (GC) system for analysis. The GC system consisted of two Shimadzu GC-2010 GCs, one with a flame ionization detector (FID) and another with a Shimadzu QP2010 Mass Spectrometer (MS). The FID detected C2 and C3 compounds, and the MS detected all other compounds. Details about the method were reported by Lyman et al. (2021) and Lyman et al. (2022b). Table 1. List of organic compounds analyzed for this project. Compound group and analytical method are also shown for each compound. Compound Compound Group Analytical method Methane Methane LGR analyzer Ethane Alkane GC-MS Ethylene Alkene GC-MS Propane Alkane GC-MS Propylene Alkene GC-MS Isobutane Alkane GC-MS n-Butane Alkane GC-MS Acetylene Alkyne GC-MS Trans-2-butene Alkene GC-MS 1-Butene Alkene GC-MS Cis-2-butene Alkene GC-MS Isopentane Alkene GC-MS N-Pentane Alkane GC-MS Trans-2-pentene Alkene GC-MS 1-Pentene Alkene GC-MS Cis-2-pentene Alkene GC-MS 2,2-Dimethylbutane Alkane GC-MS Cyclopentane Alkane GC-MS 57 Compound Compound Group Analytical method 2,3-Dimethylbutane Alkane GC-MS 2-Methylpentane Alkane GC-MS 3-Methylpentane Alkane GC-MS Isoprene Alkene GC-MS 1-Hexene Alkene GC-MS n-Hexane Alkane GC-MS Methylcyclopentane Alkane GC-MS 2,4-Dimethylpentane Alkane GC-MS Benzene Aromatic GC-MS Cyclohexane Alkane GC-MS 2-Methylhexane Alkane GC-MS 2,3-Dimethylpentane Alkane GC-MS 3-Methylhexane Alkane GC-MS 2,2,4-Trimethylpentane Alkane GC-MS n-Heptane Alkane GC-MS Methylcyclohexane Alkane GC-MS 2,3,4-Trimethylpentane Alkane GC-MS Toluene Aromatic GC-MS 2-Methylheptane Alkane GC-MS 3-Methylheptane Alkane GC-MS n-Octane Alkane GC-MS Ethylbenzene Aromatic GC-MS m/p-Xylene Aromatic GC-MS Styrene Alkene GC-MS o-Xylene Aromatic GC-MS n-Nonane Alkane GC-MS Isopropylbenzene Aromatic GC-MS n-Propylbenzene Aromatic GC-MS 1-Ethyl-3-methylbenzene Aromatic GC-MS 1-Ethyl-4-methylbenzene Aromatic GC-MS 1,3,5-Trimethylbenzene Aromatic GC-MS 1-Ethyl-2-methylbenzene Aromatic GC-MS 1,2,4-Trimethylbenzene Aromatic GC-MS n-Decane Alkane GC-MS 1,2,3-Trimethylbenzene Aromatic GC-MS 1,3-Diethylbenzene Aromatic GC-MS 1,4-Diethylbenzene Aromatic GC-MS Methanol Alcohol GC-MS Ethanol Alcohol GC-MS 64 Figure 3 also shows the percentage of carbon atoms in exhaust that were in organic molecules. This value is a carbon, atom-based measure of the amount of fuel slip. The amount of fuel slip dropped slightly in fuel injector test 3 but was generally similar in all tests except the upgraded fuel injector test. Those tests were similar to the average of previous measurements of Arrow L795 engines. The upgraded fuel injector test showed a marked reduction in fuel slip. Figure 3. Mixing ratios (i.e., concentrations) of carbon atoms in individual pollutants or pollutant groups in exhaust, in units of parts per million or parts per million of carbon. The top of the bar shows the total mixing ratio, and the coloration of the bar shows mixing ratios of individual pollutants or pollutant groups, as indicated in the legend. The leftmost bar is average results for six Arrow L795 engines measured by Lyman et al. (2022b). The whisker on the leftmost bar shows the 95% confidence interval of average total emissions. The black line with black circles shows the percentage of carbon atoms in exhaust that were part of organic compounds. Figure 4 shows emissions of non-methane organic compounds, organized by compound group. Non-methane organic compounds are equivalent to volatile organic compounds, or VOC, except that the regulatory definition of VOC does not include ethane, while non-methane organic compounds do include ethane (it is an alkane). Trends for non-methane organics and VOC are the same. Figure 4 shows that non-methane organics emissions in all the fuel injector tests, except fuel injector test 3 were higher than in tests with the engine’s original carburetor and higher than the average of exhaust measurements from other Arrow L795 engines. Fuel injector test 3, on the other hand, was lower than all other measurements. The high organics emissions in most of the fuel injector tests are due to high exhaust flow rate, since high flow rate leads to higher overall emissions. As shown in Figure 5, NOX (i.e., NO + NO2) emissions in fuel injector tests 1 and 2 were similar to emissions during the carburetor tests and the average of exhaust measurements from other 65 Arrow L795 engines. Fuel injector test 3 and the upgraded fuel injector test had the lowest emissions of NOX. Figure 4. Emissions of groups of non-methane organic compounds from the engine in each of the tests, in units of grams per hour. The top of the bar shows total emissions, and the coloration of the bar shows emissions of each group, as indicated in the legend. The leftmost bar is average results for six Arrow L795 engines measured by Lyman et al. (2022b). The whisker on the leftmost bar shows the 95% confidence interval of average total emissions. The black line with black circles shows the exhaust flow rate. 66 Figure 5. Emissions of NO and NO2 from the engine in each of the tests, in units of grams per hour. The top of the bar shows total emissions, and the coloration of the bar shows emissions of each group, as indicated in the legend. The leftmost bar is average results for six Arrow L795 engines measured by Lyman et al. (2022b). The whisker on the leftmost bar shows the 95% confidence interval of average total emissions. The black line with black circles shows the exhaust flow rate. 3.4. Time Series Data Time series of the five tests are shown in the following figures. Figure 6 shows fuel injector tests 1 and 2, Figure 7 shows fuel injector test 3, and Figure 8 shows carburetor tests 1 and 2. The cyclical nature of exhaust properties and composition in the figures coincides with strokes of the pumping unit. Figure 9 shows only two minutes of a time series so the cycle can be better visualized. It shows that the cycle of exhaust O2 levels is opposite to the cycle for NOX and CO2, and CO, but is in sync with the cycle for hydrocarbons. These cycles are on the same time scale as exhaust temperature, but with a lag. Lyman et al. (2022b) postulated that, as the engine load increases when pulling the pumping unit up, more fuel is used, which increases the exhaust temperature (and the unmeasured engine temperature) and decreases O2 and uncombusted fuel in the exhaust. Higher engine temperatures lead to more NOX production, and increased combustion leads to more CO2 and CO production. Then, as the pumping unit descends and the engine load decreases, the engine consumes less fuel, the exhaust temperature declines, and all these trends are reversed. At higher combustion temperatures, fuel slip can be expected to decrease, and NOX emissions can be expected to increase. Exhaust temperatures were similar for all five engine tests, which probably explains why fuel slip and NOX emissions were also similar for all tests. 67 Figure 6. Time series of measurements collected during fuel injection tests 1 and 2. All measurements shown were collected at 2-second intervals, except methane, which was collected at 20-second intervals. 68 Figure 7. Time series of measurements collected during fuel injection test 3. All measurements shown were collected at 2-second intervals, except methane, which was collected at 20-second intervals. 69 Figure 8. Time series of measurements collected during carburetor tests 1 and 2. All measurements shown were collected at 2-second intervals, except methane, which was collected at 20-second intervals. 70 Figure 9. Two minutes of the time series for fuel injector test 3. All measurements shown were collected at 2second intervals. Hydrocarbons are shown rather than methane. This hydrocarbon measurement was collected by an infrared sensor on the Ecom analyzer. Other measurements of methane and non-methane organics were collected as described in methods. The Ecom hydrocarbon measurement has been found to be unquantitative, but it provides an accurate measure of emission trends across time. It is shown here because it is available at 2second intervals, while other organics measurements collected for this study are only available at longer 71 intervals. Figure 10. Time series of measurements collected while the engine was being adjusted to decrease O2 intake. All measurements shown were collected at 2-second intervals except methane, which was collected at 20-second intervals. 72 Figure 10 shows a time series for a period when the fuel injection system was used to dynamically change air intake by the engine, as well as other controllable engine parameters, to increase combustion temperature and decrease fuel slip. The figure shows the percentage of carbon atoms in the exhaust that were contained within organic compounds, which is a measure of fuel slip. Measurements of non-methane organics were not available during this period, so the percentage of carbon in organic compounds was calculated using only methane data and is underestimated. Nonetheless, the figure shows that modification of engine parameters can lead to changes in fuel slip. Indeed, the figures show that fuel slip varied by more than two times. References Anneken, D., Striebich, R., DeWitt, M. J., Klingshirn, C., and Corporan, E.: Development of methodologies for identification and quantification of hazardous air pollutants from turbine engine emissions, J. Air Waste Manag. Assoc., 65, 336-346, 2015. Brettschneider, J.: Calculation of the air ratio lambda of air-fuel mixtures and the effect of errors in measurement on lambda, Bosch Tech. Ber.;(Germany, Federal Republic of), 6, 1979. EPA, U.: Compendium Method TO-11A, U.S. Environmental Protection Agency, Research Triangle Park, North Carolina, 1999. Lyman, S., Tran, H., O'Neil, T., and Mansfield, M. L.: Final Report: Air Pollutant Emissions from Natural Gas-Fueled Pumpjack Engines in the Uinta Basin, Utah State University, Vernal, Utah, 2022a. Lyman, S. N., Tran, H. N. Q., O’Neil, T. L., and Mansfield, M. L.: Low NOX and high organic compound emissions from oilfield pumpjack engines, Elem. Sci. Anth., 10, 00064, https://doi.org/10.1525/elementa.2022.00064, 2022b. Lyman, S. N., Holmes, M., Tran, H., Tran, T., and O’Neil, T.: High ethylene and propylene in an area dominated by oil production, Atmos., 12, 1, 2021. Restek: Improve Analysis of Aldehydes and Ketones in Air Samples with Faster, More Accurate Methodology, Restek, Bellefonte, Pennsylvania, 2018. Shimadzu: Nexera Application Data Sheet No.13: Ultrafast Analysis of Aldehydes and Ketones, Shimadzu, Tokyo, Japan, 2011. Uchiyama, S., Naito, S., Matsumoto, M., Inaba, Y., and Kunugita, N.: Improved measurement of ozone and carbonyls using a dual-bed sampling cartridge containing trans-1, 2-bis (2-pyridyl) ethylene and 2, 4-dinitrophenylhydrazine-impregnated silica, Anal. Chem., 81, 6552-6557, 2009. 73 Understanding the Key Components of Winter Ozone Formation & the Influence of Chemical Mechanisms Seth Lyman, PhD Introduction The Uinta Basin is located in the northeastern corner of Utah which is renowned for its natural beauty and rich energy resources. In recent years, it has faced a unique environmental challenge of elevated tropospheric ozone during the winter months (1). This phenomenon is primarily attributed to a combination of thermal inversion, snow cover, and emission of pollutants from the oil and gas industry. Uinta Basin is surrounded by high mountains, creating a bowl-shaped terrain, which is ideal for creating strong thermal inversions. During thermal inversions, colder air is trapped beneath warmer air, which traps the pollutant near ground level, leading to elevated ozone concentrations. Volatile organic compounds (VOCs) and nitrogen oxides (NOx) released from the oil and gas production pads interact with sunlight and undergo photochemical reactions to produce ozone (2). Among VOCs, Carbonyl compounds are found to be important precursors in winter ozone formation, but their chemistry during winter ozone episodes has not been thoroughly investigated. Box models can be useful in air quality studies because they simplify the atmosphere into a single box. It allows modification and adjustments of input parameters like precursor emission rates which helps in detailed analysis of atmospheric chemistry (3). Method In this study, (a) we used the F0AM Box Model along with four different chemical mechanisms to identify the specific carbonyl species that act as important precursors to winter ozone formation and determine how they form; (b) we determined the modeled emission factors for carbonyl compounds, (c) we assessed the ozone formation potential of different carbonyl compounds , and (d) analyzed the sensitivity of various primarily emitted organic species (alkane, alkene, alkyne, alcohols and aromatics) to carbonyl compounds to provide information about how different hydrocarbon species regulate the production of ozone and carbonyl compounds. As researchers, we utilized a subset of the Master Chemical Mechanism version 3.3.1 (MCMv331) (4), which consists of 3423 species and 10309 reactions as the base chemical mechanism with the F0AM Box model. Then we compared MCMv331 output with lumped chemical mechanisms such as Regional Atmospheric Chemistry Mechanism (RACM2) (5), Statewide Air Pollution Research Center Chemical Mechanism version 07 (SAPRC07) (6), and 80 Table 1. Ozone summary statistics for five sites in the Uinta Basin over 15 calendar years. All values were calculated from daily maximum 8-hr average concentrations. For 2024, only data through August are shown. Year Site Mean Median Max Min 4th High Daily Max Exceedance Days (>70 ppb) 2009 (JulyDec) Ouray 47 47 101 23 67 1 Vernal -- -- -- -- -- -- Roosevelt -- -- -- -- -- -- Whiterocks -- -- -- -- -- -- 2010 Ouray 56 54 123 20 117 45 Vernal -- -- -- -- -- -- Roosevelt -- -- -- -- -- -- Whiterocks -- -- -- -- -- -- 2011 Ouray 54 52 138 18 119 28 Vernal 55 55 95 33 84 12 Roosevelt 56 54 116 30 103 21 Whiterocks -- -- -- -- -- -- 2012 Ouray 49 50 76 18 67 1 Vernal 45 46 68 14 64 0 Roosevelt 50 51 70 14 67 0 Whiterocks -- -- -- -- -- -- 2013 Ouray 58 54 141 24 133 52 Vernal 53 52 114 20 102 32 Roosevelt 56 54 110 18 104 35 Whiterocks -- -- -- -- -- -- 2014 Ouray 48 49 91 17 79 8 Vernal 44 46 64 12 62 0 Roosevelt 51 51 63 29 62 0 Whiterocks 47 48 67 24 64 0 2015 Ouray 46 47 71 21 68 2 Vernal 43 43 67 10 64 0 Roosevelt 44 44 66 24 60 0 Whiterocks 47 47 73 25 68 2 2016 Ouray 49 48 120 20 96 11 Vernal 47 46 78 20 73 5 Roosevelt 47 47 96 20 81 5 Whiterocks 48 48 86 29 81 7 2017 Ouray 50 50 111 20 103 11 Vernal 48 49 69 25 68 0 Roosevelt 48 48 86 24 78 8 Whiterocks 46 46 76 27 66 1 2018 Ouray 48 48 72 18 67 1 Vernal 48 50 81 20 69 2 Roosevelt 49 49 79 18 71 8 81 Year Site Mean Median Max Min 4th High Daily Max Exceedance Days (>70 ppb) Whiterocks 47 46 71 22 69 1 2019 Ouray 50 51 110 21 98 16 Vernal 47 48 76 16 65 1 Roosevelt 50 51 96 19 87 10 Whiterocks 48 49 74 23 67 3 2020 Ouray 49 48 74 26 65 1 Vernal 42 41 67 14 63 0 Roosevelt 46 47 71 24 63 1 Whiterocks 47 47 73 28 65 1 2021 Ouray 50 50 73 21 72 5 Vernal 47 47 72 23 68 3 Roosevelt 48 48 77 20 72 4 Whiterocks 49 48 75 25 68 1 2022 Ouray 46 46 68 23 64 0 Vernal 46 46 66 20 63 0 Roosevelt 47 48 71 21 66 1 Whiterocks 47 47 63 27 62 0 2023 Ouray 49 49 102 14 91 27 Vernal 47 50 101 14 82 12 Roosevelt 49 50 112 11 93 23 Whiterocks 50 51 105 18 88 13 2024 (JanAug) Ouray 42 43 66 18 57 0 Vernal 46 48 74 15 68 3 Roosevelt 48 49 78 17 69 2 Whiterocks 49 50 76 24 70 3 Utah DAQ also measured ozone in Vernal during 2006 and 2007, but those data are not publicly available and are not included here. No wintertime exceedances of the ozone standard were measured in Vernal during that period. The three-year average of annual fourth-highest daily maximum 8-hr averages for a given site (using calendar years) is referred to as a design value. The design value is the value EPA uses to determine whether an airshed is in attainment of the 70 ppb ozone standard (design values of 71 and above are out of attainment). EPA used the 2014-16 period in their 2018 decision to designate the Uinta Basin as a nonattainment area for ozone. Table 2 shows the ozone design value for the past several three-year periods for the same monitoring stations shown in Figure 1. The design value for 2022-24 shown in Table 2 only includes data through August 2024. Note that these are not official design values. Regulatory agencies may exclude some data or consider other criteria in the process of determining official design values. 82 Table 2. Average of the fourth-highest daily maximum 8-hr average ozone during three consecutive calendar years for several monitoring stations in the Uinta Basin (a.k.a. ozone design values). Only 2024 data collected through August are used. Values in exceedance of the EPA standard are in bold font. Values shown may include summertime ozone events that could be excluded from regulatory consideration. Station Ouray Vernal Roosevelt Whiterocks 2013-15 93 76 75 66 2014-16 81 66 67 71 2015-17 89 68 73 71 2016-18 88 70 76 72 2017-19 89 67 78 67 2018-20 76 65 73 67 2019-21 78 65 74 66 2020-22 67 64 67 65 2021-23 75 71 77 72 2022-24 70 71 76 73 Figure 2 and Figure 3 show the number of ozone exceedances and the annual fourth-highest daily maximum 8-hr average ozone, respectively, for each entire calendar year at the same monitoring stations shown in the previous tables and figures. These figures show that air quality is extremely variable from year to year and across measurement stations in the Uinta Basin. For example, Ouray experienced more than 40 exceedance days in 2010 and 2013 but had only one in 2012, 2018, and 2020, and only two in 2015. Some exceedances shown in Figure 2 occurred during summer, not winter, and the summertime exceedances were likely due to intrusions of ozone-rich stratospheric air or wildfires. Figure 3 shows that the fourth-highest daily maximum 8-hr average ozone at the five sites is always about 60 ppb or higher. The natural summertime background ozone level in the intermountain western United States is 60 to 65 ppb (Parrish et al., 2022). During years with low wintertime ozone (2012, 2015, 2018, 2020, 2022, and 2024), the highest ozone is observed during summer and is usually in the range of 60-65 ppb. Ozone is also spatially variable. Figure 3 shows that Ouray and Roosevelt tend to have higher ozone than Vernal. 83 Figure 2. Number of annual ozone exceedances at five sites from 2010 through August 2024. The grey bars indicate years with little snow cover. Figure 3. Annual fourth-highest 8-hr average daily maximum ozone at five sites from 2010 through August 2024. The red-dashed line indicates 70 ppb, the current EPA standard for ozone. The grey bars indicate years with little snow cover. Particulate Matter Figure 4 shows a time series of all PM2.5 measurements that have ever been collected in the Uinta Basin. Exceedances of the EPA PM2.5 standard sometimes occur during winter but are more common during summer months. These summertime spikes in PM2.5 concentrations are typically caused by wildfire smoke. 84 Figure 4. Time series of daily 24-hr average PM2.5 concentrations at nine sites in the Uinta Basin, October 2009March 2024. The red-dashed line shows 35 μg m-3, the EPA standard for PM2.5. Trends in the Capacity of Winter Inversion Episodes to Produce Ozone Ozone Exceedance Days The number of exceedances of the EPA ozone standard show a decreasing trend (Mansfield and Lyman, 2021) from 2010 through 2022. Figure 2 shows all exceedance days by calendar year, including those that occur during summer. The decreasing trend is clearer in Figure 5, which shows only wintertime exceedances by winter season. Mansfield and Lyman (2021) showed that NOX emissions also decreased over the same period and Lin et al. (2021) showed that methane emissions also decreased over the same period. Mansfield and Lyman (2021) attributed the decline in ozone and its precursors to (1) a decline in energy production (which was driven mostly by a decline in natural gas production) and (2) regulatory and voluntary action by the oil and gas industry to reduce emissions. 85 Figure 5. Number of ozone exceedance days per winter season and total energy production per year in units of barrels of oil equivalent (i.e., sum of oil production and gas production scaled to equivalent amount of energy. The number of exceedance days at the monitoring station in the Basin with the maximum number in any given winter season is shown. Winters with at least 15 days with snow depth greater than 5 cm are shown in red and other years are shown in grey. Figure 5 shows that the declining trend in winter ozone reversed sharply in 2023 when deep snow cover and many strong inversions allowed for many exceedance days. The figure shows that oil and gas production have also increased, and increased emissions most likely also contributed to increased ozone in 2023. Dependence on Meteorological Conditions As Figure 5 shows, significant local production of wintertime ozone has never occurred and cannot occur without significant snow cover across the Uinta Basin (Oltmans et al., 2014). Indeed, winter ozone levels are positively correlated with snow depth (r2 = 0.50 for the Horsepool site). The number of winter exceedance days is more strongly correlated with the season-average inversion strength; however, since snow cover sometimes exists in stormy or windy conditions that don’t allow for strong winter inversion episodes, strong winter inversion episodes seldom occur without snow cover, and winter ozone needs inversions and snow to form (Mansfield, 2018). (Figure 6) Figure 6 plots the number of ozone exceedance days per season against the pseudo-lapse rate. The lapse rate is the inverse of the change in temperature with altitude. The lower the lapse rate, the stronger the inversion. Typically, lapse rates are measured by releasing temperature sensors attached to helium balloons, but these measurements are expensive and have rarely been performed in the Uinta Basin. As an alternative, Mansfield (2018) used temperature measurements from surface stations at different elevations to approximate the lapse rate (a “pseudo” lapse rate). We follow Mansfield's method of determining pseudo-lapse rates in this section. 86 Figure 6. Number of ozone exceedance days per winter season versus seasonal average pseudo-lapse rate. Pseudo-lapse rate is a measure of inversion strength (more negative value indicates stronger inversion) and is discussed at length by Mansfield (2018). The orange dashed line is a linear regression that only includes seasonal average pseudo-lapse rates less than 2 K km-1. Figure 7 shows the difference between actual ozone exceedance days per winter season and the number of exceedance days expected from the relationship shown in Figure 6-. The figure shows an apparent decreasing trend over time, meaning that fewer exceedance days than expected have occurred in recent years. This could indicate that, for similar seasonal conditions, the capacity of the Uinta Basin atmosphere to produce ozone has decreased. Figure 7. The number of excess ozone exceedance days per winter season relative to the amount expected from the relationship is shown in Figure 6-. Only years with seasonal average pseudo-lapse rates less than 2 K km-1 are shown. Trends in Ozone Precursors Mansfield and Lyman (2021) used a linear regression of daily maximum 8-hr average ozone against daily pseudo-lapse rate to “correct” ozone for inversion strength. Mansfield and Lyman 87 (2021) found that the daily pseudo-lapse rate and the daily maximum 8-hr average ozone for most winter seasons were correlated. They used the linear regression for each season to calculate the daily maximum 8-hr average ozone value that would be expected for a pseudolapse rate of -15 K km-1in that season, and they called this metric 〈[O3]〉−15. Since the 〈[O3]〉−15 normalizes for the influence of inversion strength on ozone production, they assumed that year-to-year differences in 〈[O3]〉−15 were due to differences in ozone-forming emissions, not differences in meteorology. Mansfield and Lyman showed that 〈[O3]〉−15 declined from 2010 to 2020, and they attributed this decline to declines in NOX and organic compound emissions. We applied the method of Mansfield and Lyman (2021) to methane, NOX, and non-methane hydrocarbon emissions to separate emission trends from meteorological variability. We calculated linear regressions of the pseudo-lapse rate versus daily average methane, NOY (as a proxy for NOX, since NOY is the sum of NOX and its photochemical degradation products) and total non-methane hydrocarbons measured at Horsepool and Roosevelt. We calculated separate regression equations for each site for each year for which data were available. We omitted days with relatively uncertain pseudo-lapse rates (r2 < 0.3 for the relationship between temperature and elevation). The pseudo-lapse rate was predictive of the majority of the variability in methane, NOY, and non-methane hydrocarbons (r2 or 0.61 ± 0.21, 0.63 ± 0.21, and 0.56 ± 0.22, respectively: average ± standard deviation). We determined residuals by subtracting predicted daily concentrations from measured concentrations. Daily average concentrations of methane, NOY, and non-methane organics were significantly correlated with temperature, wind speed, the number of consecutive inversion days and the number of days since the winter solstice. The residuals, however, were not significantly correlated with any of these variables, which indicates that regression against the pseudo-lapse rate was adequate to account for the influence of these other variables on daily average concentrations. Following Mansfield and Lyman (2021), we applied site-and-year-specific regression equations to ascertain the daily average concentrations that would be expected in each winter season at a pseudo-lapse rate of -15 K km-1. The results are shown in Figure 1-8 through Figure 1-10. Since this method considers the propensity of methane, NOY and non-methane hydrocarbon concentrations to increase under inversion conditions, and since no other significant meteorological correlations existed after inversion strength was taken into account, we attribute the temporal trends in the figures to changes in emissions. We acknowledge that the trends shown in the figures may be local and that more work is needed to verify whether these trends hold true for the Uinta Basin as a whole. 88 Figure 8. Daily average NOY (a proxy for NOX) at a pseudo-lapse rate of -15 K km-1 as predicted from yearand site-specific linear regressions of NOY against the pseudo-lapse rate. Whiskers show the combined uncertainty of the pseudo-lapse rate calculation and the NOY measurement. Figure 9. Daily average methane at a pseudo-lapse rate of -15 K km-1, as predicted from yearand site-specific linear regressions of methane against the pseudo-lapse rate. Whiskers show the combined uncertainty of the pseudo-lapse rate calculation and the methane measurement. 89 Figure 10. Daily average total non-methane hydrocarbons at a pseudo-lapse rate of -15 K km-1, as predicted from yearand site-specific linear regressions of non-methane hydrocarbons against the pseudo-lapse rate. Whiskers show the combined uncertainty of the pseudo-lapse rate calculation and the non-methane hydrocarbons measurement. The results for Horsepool are similar to the findings of Mansfield and Lyman (2021) and Lin et al. (2021), which all show that emissions of methane, NOX, and non-methane hydrocarbons have declined since 2013. Figures 8 through 10 do not show a meaningful trend in any of the compounds at Horsepool after 2017, even though energy production and oil and gas activity changed considerably over that time. No clear trend exists for methane and non-methane hydrocarbons at Roosevelt, perhaps because of the site’s shorter hydrocarbon measurement record. The NOY trend at Roosevelt is dominated by a strong upswing in winter 2023. Several large flares at oil wells were active near the Roosevelt site during winter 2023, and these could be the cause of high NOY in 2023. Acknowledgments This work was funded by the Utah Legislature and Uintah Special Service District 1. This work utilized weather observations obtained using the Synoptic Data PBC Weather API (synopticdata.com). References Lin, J.C., Bares, R., Fasoli, B., Garcia, M., Crosman, E., Lyman, S., 2021. Declining methane emissions and steady, high leakage rates observed over multiple years in a western US oil/gas production basin. Sci. Rep. 11, 1-12. Mansfield, M.L., 2018. Statistical analysis of winter ozone exceedances in the Uintah Basin, Utah, USA. J. Air Waste Manag. Assoc. 68, 403-414. 96 Additional information about the methods used is available in Lyman et al. (2021). Table 2 lists the organic compounds we measured. Particulate Matter Measurements We measured particulate matter with aerodynamic diameter smaller than 2.5 micrometers (PM2.5) at Horsepool with a BAM 1020 monitor. We operated the instrument according to manufacturer protocols, with leak checks, flow and mass calibrations, detector calibrations, and cleanings performed at regular intervals. We obtained particulate matter values for other sites from the EPA AQS database (https://aqs.epa.gov/api). Meteorological Measurements We deployed solar radiation sensors at Horsepool (incoming and outgoing shortwave and longwave with a Hukseflux NR01 radiometer and UV-A and UV-B with Kipp and Zonen UV radiometers), Roosevelt (incoming and outgoing shortwave with a Kipp and Zonen CNR-4), and Castle Peak (incoming and outgoing shortwave and longwave with a Hukseflux NR01 radiometer). We check these sensors against calculations of clear-sky radiation annually. We operated a suite of comprehensive, research-grade meteorological instruments at all sites operated by USU. We checked wind speed and direction, temperature, humidity, and barometric pressure against a NIST-traceable standard once annually. We checked snow depth sensors against a height standard annually. We also obtained meteorological data from the EPA AQS database. Results and Discussion Ozone Very little snow cover existed across the lower elevations of the Uinta Basin, keeping ozone well below the 70 ppb EPA standard throughout winter 2023-24 at Horsepool (Figure 1) and at sites across the Basin (Figure 2). 97 Figure 1. Horsepool daily maximum ozone, average snow depth, and daytime average total UV radiation (incoming + reflected) during winter 2023-24. Inversion periods are shown as light blue boxes. Figure 2. 8-hr average ozone from all sites listed in Table 2-1 during winter 2023-24. Table 3 provides information about ozone observed at all monitoring stations in the Uinta Basin during winter 2023-24. No monitoring station experienced exceedances of the EPA ozone standard during the winter. An exceedance occurs when the daily maximum 8-hr average ozone value at a station is greater than the EPA standard of 70 ppb. The average of the fourthhighest daily maximum 8-hr average ozone value over three consecutive calendar years is used to determine regulatory compliance with the standard. 98 Table 3. Eight-hour average ozone concentrations around the Uinta Basin, winter 2023-24. Mean Maximum Minimum 4th Highest Daily Maximum Number of Exceedances Seven Sisters 26.1 47.5 6.9 44.5 0 Castle Peak 30.3 50.7 4.3 48.9 0 Dinosaur N.M. 26.4 50.3 4.9 48.1 0 Red Wash 28.4 44.8 9.5 42.8 0 Vernal 24.0 44.0 2.8 43.8 0 Whiterocks 32.0 49.3 5.5 48.0 0 Ouray 21.3 45.0 1.3 38.9 0 Roosevelt 22.7 46.8 0.0 45.9 0 Myton 24.6 47.8 0.4 46.8 0 Horsepool 28.1 48.3 7.5 45.7 0 Rangely 29.2 50.4 5.9 48.0 0 Figure 3 shows the spatial distribution of the fourth-highest daily maximum 8-hr average ozone concentration around the Uinta Basin during winter 2023-24. All sites in the Basin exceeded the 70 ppb EPA ozone standard, but ozone tended to be lower at the Basin edges, in Whiterocks, Vernal, and Rangely. Ouray has typically had ozone in the same range as Horsepool and Seven Sisters, but this year Ouray had fewer exceedance days and lower ozone overall than those two sites. The reason for this is unclear. Figure 3. Fourth-highest daily maximum 8-hr average ozone in the Uinta Basin during winter 2023-24. Monitoring stations are shown as red circles. Active oil and gas wells are shown as orange dots. The 99 background color indicates ozone concentration and was interpolated using the inverse distance weighting method in ArcGIS Pro. There were some heavy snow events during winter 2023-24, but air temperatures were too high to support a snowpack after the heavy snowfall. One snowfall did lead to a weak inversion (where cold air begins to stagnate at the Basin floor, trapping pollution), but the snow melted before ozone levels built substantially. Snowfall and seasonal snowpack were substantial along both the Wasatch and Uinta Mountain ranges. Locals in the Basin noticed the contrast between the last two winters (see Figure 4); winter 2022-23 had snow cover across the entire Basin for months, leading to many inversion days and high ozone, in comparison to last winter with zero days of ozone. Figure 4. Satellite images of the Uinta Basin from February 2023 and February 2024, showing the difference in snow cover across the two years. Images taken from https://worldview.earthdata.nasa.gov/. Particulate Matter PM2.5 concentrations stayed below the EPA standard of 35 µg m-3 during winter 2023-24 at Vernal, Roosevelt, Rangely, and Horsepool (Figure 4-). As is typical of low-ozone winters in the Uinta Basin, PM2.5 was lower at Horsepool than at the other stations, where urban sources of particulate pollution dominate (Figure 5). 100 Figure 5. 24-hr average PM2.5 at monitoring stations around the Uinta Basin during winter 2023-24. The reddashed line indicates the EPA PM2.5 standard. Figure 6. Box-and-whisker plots of 24-hr average PM2.5 at four monitoring stations during winter 2023-24. X’s indicate average values. Lines within the boxes indicate medians. Tops and bottoms of boxes indicate the third and first quartiles. Top and bottom whiskers indicate maximum and minimum values. Circles show outliers. Comparison of Roosevelt, Horsepool, and Castle Peak Data The Horsepool and Roosevelt monitoring stations began operating in winter 2011-12 and were designed to contain nearly identical suites of instrumentation. At both stations we measure NOX with instrumentation that doesn’t bias NO2 during winter inversion episodes, while all regulatory monitoring stations in the Uinta Basin use alternative, biased instrumentation. The 101 areas surrounding the Horsepool and Roosevelt stations are different from one another. The Horsepool station is on the northern edge of an area of dense oil and gas development (mostly gas), whereas the Roosevelt station is within a small city. Oil and gas development exists within and near the city of Roosevelt (mostly oil). The two stations are at very similar elevations (Table 1). In 2017, Utah DAQ donated a NOX analyzer that we upgraded with a photolytic converter and installed at our Castle Peak monitoring station. Castle Peak is in an area of dense oil development, and its elevation is less than 100 meters higher than the Roosevelt and Horsepool stations. Figure 6 shows NOX measured at Roosevelt, Horsepool, and Castle Peak during winter 2023-24, and Figure 7 shows NOZ at Roosevelt and Horsepool. NOX is the sum of NO and NO2 which are important precursors to ozone production. NOY (not shown in the figures) is the sum of NOX and all other reactive nitrogen compounds (e.g., nitric and nitrous acids, organic nitrates, and particulate-bound nitrogen compounds). NOZ is the sum of all reactive nitrogen compounds except NOX (in other words, it is NOY minus NOX). While NOX is an ozone precursor, the compounds that comprise NOZ are mostly generated along with ozone because of photochemical reactions and are byproducts and indicators of atmospheric photochemical conditions. Figure 7. Hourly average NOX measured at Roosevelt, Horsepool, and Castle Peak during winter 2023-24. During winter 2023-24, as in previous winters, NOX was higher in Roosevelt than at Horsepool and Castle Peak (Figure 6-) and was 4.0 times higher than Horsepool on average. NOX in Roosevelt is emitted from urban sources like cars and home heating, as well as from oil and gas sources, while NOX in the vicinity of Horsepool and Castle Peak originates almost entirely from oil and gas activity. NOX at Castle Peak was 29% higher than Horsepool (p-value for a t-test of difference was <0.01). NOZ was 30% higher at Roosevelt than at Horsepool, indicating greater 102 photochemical activity at Roosevelt, especially in early winter when stagnant conditions with higher PM2.5 prevailed (Figure 7). Figure 8. Hourly average NOZ measured at Roosevelt and Horsepool during winter 2023-24. NOX at Roosevelt showed a pronounced peak in the morning and a lesser peak in the late afternoon and early evening, probably due to morning and afternoon peaks in local traffic (Figure 8-). Horsepool did not show a pronounced peak, probably because the majority of NOX emissions at the site were due to stationary, continuous sources rather than traffic-related sources. Castle Peak showed a NOX maximum in late morning, perhaps due to unique oil field traffic patterns. Figure 9. Average NOX at Roosevelt, Horsepool, and Castle Peak during each hour of the day during inversion episodes that occurred during winter 2023-24. Whiskers represent 95% confidence intervals. Methane was only 3% higher at Horsepool compared to Roosevelt. (Figure 9-), but average total non-methane hydrocarbons (TNMHC; measured as a single group of compounds with an in-situ hydrocarbon GC) were 61% higher at Roosevelt (Figure 10-). In the past, methane was much 103 higher at Horsepool and TNMHC was also higher, but new oil and gas activity near Roosevelt appears to be increasing emissions in the area. Figure 10. Hourly average methane measured at Roosevelt and Horsepool during winter 2023-24. Well maintenance activity at the Horsepool site created interference until late December, and those data are not shown. Figure 11. Hourly average total non-methane hydrocarbons (TNMHC) measured at Roosevelt and Horsepool during winter 2023-24. (ppmC is parts-per-million of carbon atoms.) Well maintenance activity at the Horsepool site created interference until late December, and those data are not shown. A spike of 55 ppmC occurred on 26 January at Roosevelt, and those data are not shown. Snow depth was low, and snow cover was intermittent during the winter (Figure 11), keeping albedo (i.e., reflectivity from the ground surface) low (Figure 12). 104 Figure 12. Snow depth at the Roosevelt, Horsepool, and Castle Peak stations during winter 2023-24. Figure 13. Shortwave albedo at the Roosevelt and Castle Peak stations during winter 2023-24. Shortwave radiation is visible light from the sun. Albedo is the percentage of radiation that is reflected by the earth’s surface. Roosevelt ozone tended to be lower than at Horsepool and Castle Peak, especially at night Figure 13). This was the case even though NOX and non-methane hydrocarbons were both higher at Roosevelt than at Horsepool, and even though snow depth and albedo were similarly low at all sites. We expect that this occurred because the atmosphere at Roosevelt has more NOX than is needed for ozone production. Too much NOX can allow NOX to react with and destroy ozone, suppressing ozone concentrations. At night, when no photochemistry occurs, ozone is not formed, but NOX can still react with and destroy ozone, leading to the larger NOX reduction at night in Roosevelt compared to the other locations. 105 Figure 14. Hourly average ozone measured at Roosevelt, Horsepool, and Castle Peak during winter 2023-24. Figure 15. Average ozone at Roosevelt, Horsepool, and Castle Peak during each hour of the day during inversion episodes that occurred during winter 2023-24. Whiskers represent 95% confidence intervals. Speciated Volatile Organic Compounds This section focuses on measurements of individual organic compounds measured from whole air canister samples and DNPH cartridge samples. As in previous years, organic compounds in the atmosphere at field sites were dominated by alkanes, especially lighter alkanes (Figure 15 and Figure 16-). Benzene, toluene, xylenes, and other aromatics were relatively low, and C8 and larger aromatics were rarely observed. The organic compound speciation at all sites was similar, indicating that the locations were all influenced by the same general source type (oil and natural gas production). 112 Figure 4. True-color satellite image from 23 July 2024. The Uinta Basin is highlighted in the blue box. Satellitedetected wildfires are shown as orange circles. 113 Figure 5. True-color satellite image from 2 August 2024. The Uinta Basin is highlighted in the blue box. Satellitedetected wildfires are shown as orange circles. Acknowledgments This work was funded by the Utah Legislature and Uintah Special Service District 1. 114 Updates to Lab Spaces and Field Sites Trevor O’Neil Seth Lyman, PhD In the past year, we have made significant improvements to both our laboratory spaces and field sites, optimizing our research capabilities and ensuring the continued high quality of our data collection. Lab Consolidation and Equipment Surplus Our lab spaces have been reorganized to enhance operational efficiency. We have combined similar analytical instruments into shared spaces to streamline maintenance, facilitate ease of use, and improve workflow. Additionally, a comprehensive review of our storage facilities identified obsolete or unused equipment, which are marked for surplus, freeing up valuable space for future acquisitions. New Instrumentation Acquisitions We purchased a new Agilent GCMS, at a significant academic discount. This advanced instrument replaces equipment that has been in service for over 14 years, offering improved accuracy and reliability in our analytical processes. At our Castle Peak site, we installed a Teledyne N500 true NO2/NOx/NO analyzer replacing an outdated instrument. This new analyzer utilizes Cavity Attenuated Phase Shift (CAPS) Spectroscopy to measure true NO2, NOx, and NO gases. This method is known for its high sensitivity and accuracy while consuming minimal power and requiring significantly less upkeep compared to traditional chemiluminescence systems. Additionally, a new dilution calibrator has been installed at the same site to ensure the continued accuracy of our ongoing air quality measurements. Field Site Management and QA/QC Enhancements We have progressed in our efforts to update our field site management plan, particularly in quality assurance and quality control (QA/QC). This includes initiatives to simplify the data analysis process, with completion expected by mid-next year. The aim is to ensure that our field operations continue to align with best practices. 115 Uninterruptible Power Supply (UPS) Installation To mitigate the risk of power outages, which have previously resulted in significant losses of both equipment and data, we installed an uninterruptible power supply (UPS) in the lab. This addition provides a critical buffer against the region’s recurrent power disruptions, safeguarding our sensitive analytical instruments. Collaboration with Utah Department of Air Quality (UDAQ) We are appreciative of our partnership with the Utah Department of Air Quality (UDAQ). Over the past year, UDAQ has loaned us equipment, allowing us to continue uninterrupted data collection at our field sites while we await repairs and replacements for aging instruments. Additionally, we have acquired several surplus instruments from UDAQ allowing us to utilize these for improving instrumentation reliability and enabling higher data capture rates going forward. Ongoing Research Collaborations Our collaboration with UDAQ extends beyond equipment loans. We have also shared our expertise in programming for Campbell dataloggers, contributing to UDAQ's ongoing transition to digital data collection at their air quality monitoring sites. This knowledge exchange underscores the value of our continued partnership in advancing regional air quality research. Special Project - Post-Storm Peak Design Improvements With student involvement, we have initiated significant improvements to the dual-channel inlet design identified in research conducted at Storm Peak Observatory by a recently graduated master’s student. We will test these updates to determine if they enhance our sampling capabilities, allowing the removal of the impactor from the flow path without compromising particulate filtration. New Mercury Research Trailer We have begun working on a new mercury (Hg) research trailer, which will support a collaborative project involving instrumentation from UNR, BYU, and USU. This trailer will play a key role in an upcoming study focused on mercury chemistry. A high school intern working for the research center this summer helped with the work, giving him exposure to some technical aspects of research which will help develop his skillset for future projects. Training and Lab Relocation We have continued to build expertise within our team, with two staff members receiving training on DNPH sample elutions needed in preparation for analysis using our Shimadzu HPLC. Additionally, our mercury analysis equipment was moved from Room 215 to Room 218, improving lab organization and efficiency in the use of gases and other resources. 116 Mobile Lab Trailer The use of our mobile lab trailer to facilitate investigative research for pump jack engine enhancements provided an opportunity to work with the industry on improving fuel utilization by large engines in the oilfield. Our interest in this was to observe the improvements in fuel slip and its potential impact on emissions reduction. Testing Heated Inlet Design for Field Measurement Stations At cold temperatures, organic compounds may be retained on Teflon filters and tubing used for sample collection. We built two identical sample inlets at the Horsepool measurement station. Both inlets consisted of a PFA Teflon filter housing with a 5 µm pore size PTFE Teflon filter, followed by PFA tubing leading into our measurement trailer. The tubing led to a flowrestricting orifice and a vacuum pump to provide continuous flushing of the line. A tee upstream of the orifice led to whole-air sampling canisters for collection of organic compound samples. We insulated one of the inlets and heated it to 25°C, and we left the other inlet unheated. We collected three pairs of heated and unheated samples and analyzed them as described in the section of this report that discusses wintertime air quality. Winter temperatures were mild during 2023-24, which may have led to smaller differences between the heated and unheated inlets. Samples were collected on the mornings of 14 February, 20 February, and 13 March, and the average ambient air temperatures during each sampling period were 4.9, 2.5, and 0.3°C, respectively. Figure 1 shows the average difference between organic compound concentration results for samples collected on the heated and unheated lines. As the figure shows, results were higher for most compounds on the heated line, which may indicate that the unheated line retained organics. Compounds that were higher, on average, in the unheated line included 2-methylpentane, methylcyclopentane, benzene, cyclohexane, and 2-methylhexane. The differences were small, however, for all compounds (average ± standard deviation of 2 ± 5%). For reference, individual compound differences in replicate analyses of the same samples were 1 ± 7% (average ± standard deviation) during winter 2023-24. Though the differences between the heated and unheated lines were similar to differences among replicate analyses of the same samples, we decided that the best practice for the future would be to heat organic compound sampling inlets at our field sites. We have installed heated inlets for all organic compound sampling (including carbonyls) for the 2024-25 winter at Horsepool, Castle Peak, and Roosevelt. 117 Figure 1. The percent difference between organic concentrations from the heated and unheated sample inlet lines, organized by order of elution during gas chromatographic analysis of the collected air samples (x axis). Only compounds with results above the method detection limit are shown.