1. Barnett Shale Coordinated Campaign Please do not cite without permission from EDF.

2. EDF Methane Studies 1. NOAA Denver- Julesburg 2. NOAA Barnett 3. Coordinated Campaign 4. UT Phase 1 5. UT Phase 2 Pneumatics Liquid Unloadings 6. HARC/EPA 7. CSU Study Methods Measurements National Scale-up 8. CSU Study Measurements National Scale-up 13. WVU Study 14. Pilot Projects 15. Gap Filling 16. Project Synthesis Results public Submitted, not yet public Almost ready for submission Not yet submitted 12 papers Barnett synthesis 11. WSU Multi-City 9. Methane Mapping 10. Boston Study 12. Indianapolis Study

3. Top-Down: regional emissions estimated by atmospheric measurements of well mixed air Bottom-Up: emission inventory based on activity factors and emission factors Top-down estimates of CH 4 emissions often are greater than bottom-up estimates. Miller et al. 2013: 1.5X higher than EPA GHG Inventory Brandt et al 2014: 1.25 – 1.75X higher than GHG Inventory Karion et al. 2016 – Oct. 16 flight photo credit: U. of Texas

4. Barnett Shale October 16 – 30, 2013

5. Harriss et al.

6. Campaign Measurements Aircraft-based top-down regional measurements –NOAA/CU/Scientific Aviation/U. Michigan/Penn. State → 8 mass balance estimates of regional CH 4 emissions Source apportionment –U. Michigan→ ethane:methane ratios to estimate regional fossil CH 4 –UC-Irvine/U. Cincinnati → stable isotope & hydrocarbon ratios of source & background air samples –Picarro/Duke → δ 13 C-CH 4 of well pad plumes & background air –Aerodyne → C2:C1 of facility plumes

7. Top-down mass balance flights

8. Campaign Measurements Bottom-up direct component measurements –West Virginia U. → 5 transmission & storage compressor stations –Washington State U. → 230 local distribution pipeline leaks & 239 M&R stations (13 in Barnett) Ground-based near-field measurements –Picarro → 186 well pads –U. Houston → 152 well pads, midstream facilities, & landfills –Aerodyne → 224 well pads, midstream facilities, & landfills –*Colorado State U. → 125 gathering stations and processing plants

9. Campaign Measurements Aircraft-based near-field measurements –Purdue → 8 midstream facilities & landfills –Princeton/UT-Dallas (remote-control model aircraft) → repeat measurements of one compressor station –Sander Geophysics/Shell Global Solutions → locations & emission rate of sources in survey areas by Markov Chain Monte Carlo analysis

10. Custom CH 4 Emission Inventory Spatially-resolved activity factors for 25-county region Two-step Monte Carlo used to estimate emissions from well pads, compressor stations, & processing plants EPA GHG Reporting Program and published emission factors used for other sources Emissions by source estimated in 4 km x 4 km cells Alternative O&G inventories developed based on EPA GHG Inventory, GHG Reporting Program, and EDGAR

11. Spatially-Resolved Activity Factors EPA Greenhouse Gas Reporting Program EPA National Emissions Inventory TCEQ Barnett Shale Special Inventory (2009) TCEQ Permits Drillinginfo DI Desktop Google Earth

12. Lyon et al.

13. O&G Facility Monte Carlo Simulations Unbiased sample –Production sites Picarro –Midstream facilities Mitchell et al. 2015 –100 compressor stations –9 small & 16 large processing plants Targeted sample (fat-tail) –Production sites U. Houston, Aerodyne functional superemitters (paper by Zavala-Araiza) –Midstream facilities U. Houston, Aerodyne, Purdue –9 stations & plants Lyon et al.

14. Other O&G Sources Well completions –location & initial production (DI Desktop) –gas wells assumed to control completion emissions Gathering & transmission pipelines –location(DI Desktop) & EFs (EPA GHG Inventory) Local distribution –2013 pipeline miles (PHMSA), M&R station counts (GHGRP), & EFs (Lamb et al. 2015) picture credit: http://www.enengineering.com/wp-content/uploads/2013/04/Natural-Gas-Storage-Production-LNG-1.jpg

15. Other Thermogenic Sources Industrial facilities –EPA GHG Reporting Program 2013 facility emissions Residential & Commercial End Use –TX 2013 gas delivered (EIA) prorated by population with assumed leak rate Gasoline & Diesel Vehicles –EPA 2011 NEI county-level emissions Natural gas vehicles –TX 2013 vehicle fuel delivered (EIA) prorated by vehicle miles traveled with assumed leak rate picture credit: https://consumeraffairs.global.ssl.fastly.net/files/news/natural_gas.jpg

16. Other Thermogenic Sources Abandoned wells –location (DI Desktop) & emission factor (Kang et al. 2014) Geologic seepage –global microseepage EF (Etiope & Klusman 2002) picture credit: http://dnr.louisiana.gov/assets/images/conservation/engineering/6-large.JPG

17. Biogenic Sources Livestock –CAFO point source (TCEQ) and county-level (NASS) 2013 cattle population & EFs (EPA GHG Inventory) Landfills – GHGRP 2013 facility emissions adjusted up 18% for ~700 non-reporting landfills Wastewater treatment –National 2013 emissions (EPA GHG Inventory) prorated by population

18. Lyon et al.

20. High-level Findings: Top-Down & Source Apportionment 76 ± 13 total & 60 ± 11 fossil metric tons CH 4 hr -1 ≈ 1.3 – 1.9% leak rate –Karion et al, Smith et al. Multiple flights are valuable for reducing uncertainty. Stable isotope and hydrocarbon ratios vary widely across region and among source types. –Townsend-Small et al., Smith et al., Yacovitch et al. Ethane is useful for source apportionment but requires sophisticated analysis to account for C2:C1 variability. –Smith et al., Townsend-Small et al.

21. High-level Findings: Facility Measurements Production, gathering, & processing sites have fat-tail distribution. –Rella et al., Mitchell et al. Direct measurements with small sample size underestimate regional average emission rates. –Johnson et al. Targeted measurements identify higher emission sites than unbiased sampling. –Lavoie et al., Lan et al., Yacovitch et al. Compressor station emissions have high temporal variability. –Nathan et al. Processing plant and compressor station emissions are often higher than GHGRP annual average, which may be due to reporting requirements and temporal variability. –Lavoie et al., Lan et al., Yacovitch et al.

22. Rella, et al. skewed distribution of emissions: Highest 5% of sites account for 57% of emissions Emissions > 4.6 kg/h CH 4. Proportional Loss Rate (f e ): Methane Emissions Methane Production Functional super-emitter approach: sites with high proportional loss rates have excess emissions resulting from avoidable operating conditions (e.g. malfunctioning equipment). High-level Findings: Functional Super-Emitters Zavala-Araiza et al.

24. High-level Findings: Emission Inventory Emission inventory = 72 (+10/-9) total & 46(+8/-6) O&G metric tons CH 4 hr -1 ≈ 1.0 – 1.4% leak rate Gathering stations emit 40% of O&G CH 4 emissions. Custom inventory is 1.5X, 2.7X, & 4.3X higher than estimates based on EPA GHG Inventory, EPA GHGRP, and EDGAR. Differences primarily are due to greater activity factors and emission factors for gathering stations. Lyon et al.

26. Environmental Science & Technology Environmental Science & Technology http://pubs.acs.org/toc/esthag/49/13 http://pubs.acs.org/toc/esthag/49/13 Paper Lead Institution Title Harriss et al.EDF Using Multi-Scale Measurements to Improve Methane Emissions Estimates from Oil and Gas Operations in the Barnett Shale, Texas: Campaign Summary Karion et al.NOAA / CIRESAircraft-based Estimate of Total Methane Emissions from the Barnett Shale Region Smith et al.U. Michigan Airborne ethane observations in the Barnett shale: Quantification of ethane flux and attribution of methane emissions Johnson et al.WVU Methane Emissions from Leak and Loss Audits of Natural Gas Compressor Stations and Storage Facilities Nathan et al. UT-Dallas / Princeton Near field characterization of methane emissions from a compressor station using a model aircraft Lavoie et al.PurdueAircraft-Based Measurements of Point Source Methane Emissions in the Barnett Shale Basin Yacovitch et al.AerodyneMobile Laboratory Observations of Methane Emissions in the Barnett Shale Region Lan et al.U. HoustonCharacterizing fugitive methane emissions in the Barnett Shale area using a mobile laboratory Rella et al.*PicarroMeasuring Emissions from Oil and Natural Gas Well Pads Using the Mobile Flux Plane Technique Townsend-Small et al. U. Cincinnati / UC-Irvine Integrating source apportionment tracers into a bottom-up inventory of methane emissions in an urban natural gas producing region Zavala-Araiza et al. EDF Towards a Functional Definition of Methane Super-Emitters: Application to Natural Gas Production Sites Lyon et al.EDFConstructing a Spatially Resolved Methane Emission Inventory for the Barnett Shale Region

27. Helicopter-based Infrared Camera Survey ~8000 well pads surveyed for hydrocarbon emissions in 7 basins (Barnett, Eagle Ford, Fayetteville, Uintah, Power River, Bakken, & Marcellus) Aircraft quantified methane emissions at a subset of sites Statistical analyses performed to predict the number and presence of detected emissions by source type

28. Funding for EDF’s methane research series is provided by Fiona and Stan Druckenmiller, Heising- Simons Foundation, Bill and Susan Oberndorf, Betsy and Sam Reeves, Robertson Foundation, Alfred P. Sloan Foundation, TomKat Charitable Trust, and the Walton Family Foundation. Acknowledgements