Energy Research and Development Division FINAL PROJECT REPORT The California Methane Survey Gavin Newsom, Governor July 2020 CEC-500-2020-047 PREPARED BY: Primary Author: Riley Duren Andrew Thorpe Ian McCubbin Jet Propulsion Laboratory California Institute of Technology 4800 Oak Grove Drive Pasadena, CA 91109-8099 Phone: 818-354-4321 http://www.jpl.nasa.gov Contract Number: 500-15-004 PREPARED FOR: California Energy Commission Yu Hou Project Manager Jonah Steinbuck, Ph.D. Office Manager ENERGY GENERATION RESEARCH OFFICE Laurie ten Hope Deputy Director ENERGY RESEARCH AND DEVELOPMENT DIVISION Drew Bohan Executive Director DISCLAIMER This report was prepared as the result of work sponsored by the California Energy Commission. It does not necessarily represent the views of the Energy Commission, its employees or the State of California. The Energy Commission, the State of California, its employees, contractors and subcontractors make no warranty, express or implied, and assume no legal liability for the information in this report; nor does any party represent that the uses of this information will not infringe upon privately owned rights. This report has not been approved or disapproved by the California Energy Commission nor has the California Energy Commission passed upon the accuracy or adequacy of the information in this report. ACKNOWLEDGEMENTS Phase 1 of this project was funded by the California Air Resources Board (ARB-NASA Agreement 15RD028 Space Act Agreement 82-19863). Phase 2 was funded by the California Energy Commission (CEC) (agreement number 500-15-004). NASA’s Earth Science Division contributed funds for both project phases. A portion of this research was carried out at the Jet Propulsion Laboratory, California Institute of Technology, under contract with the National Aeronautics and Space Administration (NNN12AA01C). The authors thank Kelsey Foster, Brian Bue, Vineet Yadav, Michael Eastwood, David Thompson, Winston Olson-Duvall, Rob Green, Charles Miller and the AVIRIS-NG team at JPL and Francesca Hopkins and Talha Rafiq at the University of California, Riverside for their contributions to this work. The authors thank Dr. Jack Kaye at NASA Headquarters for sustained support of our methane research activities, particularly exploratory airborne campaigns in California. The authors also thank Mr. Guido Franco of the CEC, Dr. Bart Croes and many members of California Air Resources Board (CARB) research staff for helpful comments on this report. We appreciate the many helpful discussions and input to flight planning and analysis from colleagues at CARB, the Bay Area Air Quality Management District, the South Coast Air Quality Management District, the California Energy Commission, Southern California Gas Company, Pacific Gas and Electric Company, Sunshine Canyon Landfill Local Enforcement Agency, and the Milk Producer’s Council. The project also benefits from methane data processing and analysis tools and a new Geographic Information System data set developed by two concurrent NASA projects: the ACCESS program’s Methane Source Finder and the Carbon Monitoring System (CMS) program’s Prototype Methane Monitoring System for California. This final report supersedes the preliminary methane point source findings previously summarized in the California Methane Survey Interim Report. The authors are responsible for the content of the paper and the findings do not represent the views of the funding agencies. i PREFACE The California Energy Commission’s (CEC) Energy Research and Development Division manages the Natural Gas Research and Development Program, which supports energy-related research, development, and demonstration not adequately provided by competitive and regulated markets. These natural gas research investments spur innovation in energy efficiency, renewable energy and advanced clean generation, energy-related environmental protection, energy transmission and distribution and transportation. The Energy Research and Development Division conducts this public interest natural gasrelated energy research by partnering with research, development, and demonstration entities, including individuals, businesses, utilities and public and private research institutions. This program promotes greater natural gas reliability, lower costs and increases safety for Californians and is focused in these areas: • Buildings End-Use Energy Efficiency. • Industrial, Agriculture and Water Efficiency • Renewable Energy and Advanced Generation • Natural Gas Infrastructure Safety and Integrity. • Energy-Related Environmental Research • Natural Gas-Related Transportation. The California Methane Survey is the combined final report for the California Methane Survey project (CEC Contract Number 500-15-004 and ARB-NASA Agreement 15RD028 Space Act Agreement 82-19863) conducted by the Jet Propulsion Laboratory. The information from this project contributes to the Energy Research and Development Division’s Natural Gas Research and Development Program. For more information about the Energy Research and Development Division, please visit the CEC’s research website (www.energy.ca.gov/research/) or contact the CEC at 916-327-1551. ii ABSTRACT Methane point source emissions play an important role in the human (anthropogenic) methane inventory and present unique opportunities for mitigation. The researchers conducted a comprehensive survey of facilities and components in California, spanning the oil and gas, manure management, and waste management sectors, using an airborne imaging spectrometer capable of rapidly mapping methane plumes. Five campaigns were conducted over several months from 2016 to 2018, resulting in the detection, geolocation, and quantification of 564 strong methane point sources. This represents a major advance in the use of remote sensing to rapidly and repeatedly assess large areas at high spatial resolution for a poorly characterized population of methane point sources. The team estimated that emissions from methane point sources in California contribute more than a third (34 to 46 percent) of the state’s methane inventory for 2016. Methane super-emitter activity occurs in every surveyed sector. Over the entire population of observed point sources, 10 percent of sources contributed nearly 60 percent of emissions. The largest methane point source emitters in California are 32 landfills and composting facilities exhibiting persistent, potentially anomalous activity. Production is responsible for nearly 80 percent of point source emissions associated with California’s oil and gas sector. Point source emissions from natural gas infrastructure are primarily associated with a relatively small number of processing plants, compressor stations, refineries, and gas fired power plants. The project identified five low pressure natural gas leaks that were subsequently repaired by operators. This work highlights the potential for efficient point source monitoring to enable mitigation of a broad class of methane super-emitters, representing a significant contribution to California’s climate stabilization targets, reduced natural gas product loss, and early warning of potentially hazardous leaks. Keywords: natural gas, methane, emissions, mapping, remote sensing Please use the following citation for this report: Duren, Riley, Andrew Thorpe, Ian McCubbin. 2020. The California Methane Survey. California Energy Commission. Publication Number: CEC-500-2020-047. iii TABLE OF CONTENTS Page ACKNOWLEDGEMENTS .........................................................................................................i PREFACE ............................................................................................................................ ii ABSTRACT ......................................................................................................................... iii TABLE OF CONTENTS ........................................................................................................ iv LIST OF FIGURES ............................................................................................................... v LIST OF TABLES ................................................................................................................ vi EXECUTIVE SUMMARY ........................................................................................................1 Introduction .....................................................................................................................1 Project Purpose ................................................................................................................1 Project Approach ..............................................................................................................1 Project Results .................................................................................................................2 Knowledge Transfer .........................................................................................................4 Benefits ...........................................................................................................................4 CHAPTER 1: Project Purpose ..............................................................................................5 Motivation........................................................................................................................5 Prior Studies ....................................................................................................................5 Project Objectives ............................................................................................................6 CHAPTER 2: Project Approach ............................................................................................7 Observing Strategy for Methane Emissions ........................................................................7 AVIRIS-NG Instrument and Methane Retrievals ..................................................................8 Airborne Survey Design .................................................................................................. 10 Data Analysis ................................................................................................................. 13 CHAPTER 3: Project Results ............................................................................................... 16 Airborne Survey Statistics ............................................................................................... 16 Survey Completeness................................................................................................... 16 Spatial, Temporal and Sectoral Distribution of Emissions................................................... 19 Sector Specific Findings .................................................................................................. 25 Oil and Gas Production and Processing ......................................................................... 25 Natural Gas Transmission, Storage and Distribution ....................................................... 30 Refineries.................................................................................................................... 35 Power Plants ............................................................................................................... 36 iv Landfills ...................................................................................................................... 36 Wastewater treatment ................................................................................................. 38 Dairies and Livestock ................................................................................................... 38 CHAPTER 4: Knowledge Transfer ....................................................................................... 43 CHAPTER 5: Recommendations ......................................................................................... 44 CHAPTER 6: Benefits to Ratepayers ................................................................................... 46 GLOSSARY AND LIST OF ACRONYMS ................................................................................. 47 REFERENCES .................................................................................................................... 48 APPENDIX A: Data Availability.......................................................................................... A-1 LIST OF FIGURES Page Figure 1: Observing Strategy ...............................................................................................7 0BFigure 2: AVIRIS-NG Flight Parameters ................................................................................8 Figure 3: Methane Absorption Signature for AVIRIS-NG.........................................................9 Figure 4: Real Time Methane Mapping .................................................................................9 Figure 5: Methane Quick-Look Products.............................................................................. 10 Figure 6: Airborne Survey Design for the Southern San Joaquin Valley ................................. 11 Figure 7: Airborne Survey Design for the Northern San Joaquin Valley ................................. 12 Figure 8: Airborne Survey Design for the Northern California ............................................... 13 Figure 9: Data Analysis Workflow ....................................................................................... 14 Figure 10: Spatial Coverage for Survey ............................................................................... 17 Figure 11: Locations of Methane Sources Detected by Survey .............................................. 20 Figure 12: Distribution of Methane Emissions from Individual Sources .................................. 21 Figure 13: Emission Histograms for Key Sectors .................................................................. 22 Figure 14: Measured Emissions vs CARB Inventory ............................................................. 25 Figure 15: Typical Methane Plumes in SJV Oil and Gas Fields ............................................... 27 Figure 16: Closeup of a Methane Plume from a Condensate Storage Tank ............................ 27 Figure 17: Variability in Source Density Between Nearby Oil and Gas Fields .......................... 28 Figure 18: Gas Processing Facility in Elk Hills ...................................................................... 30 v Figure 19: Multiple Emission Sources at Honor Rancho Storage Facility ................................ 31 Figure 20: Variety of Emission Sources at Aliso Canyon ....................................................... 32 Figure 21: Examples of Different Emission Modes at Gas Storage Facilities ........................... 33 Figure 22: Detection of Leak in Low Pressure Gas Distribution Line ...................................... 34 Figure 23: Measured vs Reported Emissions for Refineries and Power Plants ........................ 35 Figure 24: Examples of Methane Plumes from Refineries in the LA Basin .............................. 36 Figure 25: Time Series of Landfill Point Source Emissions .................................................... 37 Figure 26: Measured vs Reported Emission for Representative Landfills ................................ 37 Figure 27: Emissions from Wastewater Treatment Plants ..................................................... 38 Figure 28: Mosaic of Two Days of AVIRIS-NG Flights Over Tulare Area Dairies ..................... 39 Figure 29: Methane Point Source Variability for Dairies Near Tipton ..................................... 40 Figure 30: Close-Up of a Dairy from the Intensive Study ..................................................... 41 Figure 31: Methane Plume Observed Persistently at Dairy Methane Digester ........................ 42 LIST OF TABLES Page Table 1: AVIRIS-NG Image Parameters ................................................................................8 Table 2: Survey Completeness by Emission Sector .............................................................. 18 Table 3: Summary of Total Emissions by Sector .................................................................. 23 Table 4: Production-Normalized Emission Rates for Associated Gas Producing Fields in the SJV .............................................................................................................................. 29 vi EXECUTIVE SUMMARY Introduction Methane (CH4) is a powerful greenhouse gas and is targeted for emissions mitigation by the State of California. It is increasingly prioritized for near-term climate action given its relatively short atmospheric lifetime and the potential for rapid, focused mitigation that can complement economy-wide efforts to reduce carbon dioxide emissions. Methane is also a precursor for tropospheric ozone and is strongly linked with co-emitted reactive trace gases targeted by California air quality and public health policies. California has established a methane emission reduction goal of 40 percent below 2013 levels by 2030. Efforts to understand the state’s methane emissions are complicated by large inconsistencies between estimates of methane emissions derived from atmospheric measurements and greenhouse gas inventories. Project Purpose The team used advanced remote sensing methods to detect and characterize anthropogenic (human) methane emissions to support the state’s objectives for mitigating short-lived climate pollutants, identifying methane “hotspots” in response to AB 1496, and supporting natural gas leak detection and correction for rate payer benefit. The project was performed in two phases with funding from the California Air Resources Board (CARB) and California Energy Commission (CEC), and co-funding from the National Aeronautics and Space Administration (NASA) Earth Science Division. Phase 1 primarily used data collected in 2016 and addressed CARB priorities spanning multiple methane emission sectors relevant to point sources—defined in this study as infrastructure components or localized (typically less than 10 meters in scale) surface features that emit plumes of concentrated methane—in California. Phase 2 collected data in 2017 and 2018 and focused on CEC priorities, particularly the natural gas sector, to improve understanding of leaks and to help enable mitigation. Phase 2 also included advances in data analysis, including estimating emission rates for individual sources and assessing total statewide emissions for each sector. Project Approach California was surveyed for methane point source emissions using the Jet Propulsion Laboratory (JPL) next generation airborne visible/infrared imaging spectrometer (AVIRIS-NG) remote sensing instrument. AVIRIS-NG is capable of rapidly mapping methane plumes over large areas using absorption spectroscopy. Absorption spectroscopy can detect and quantify a targeted substance (in this case, methane) in a sample based on how the sample interacts with different wavelengths of light (in this case, sunlight). AVIRIS-NG flights for this study were conducted during five campaigns: August – November 2016, March 2017, June 2017, August-November 2017, and September-October 2018. The survey imaged approximately 59,000 square kilometers (km2) including revisits. The survey was designed to cover at least 60 percent of methane point source infrastructure in California and was guided by a newly developed geospatial data set known as Vista-CA. This technology mapped nearly 450,000 potential methane emitting infrastructure elements, spanning the oil and gas, manure management, and waste management sectors. Of that population of nearly 450,000 potential methane emitters, approximately 272,000 infrastructure elements were surveyed by the AVIRS-NG flights, including approximately 200,000 oil and gas wells and related production 1 infrastructure as well as nearly 70,000 natural gas transmission and distribution pipeline elements. The survey included multiple overflights of the same infrastructure over several years to address source persistence – a major source of uncertainty in previous studies. This represents a major advance in the use of remote sensing to rapidly and repeatedly assess large areas at high spatial resolution for a poorly characterized population of methane point sources that often appear in an intermittent or unpredictable fashion, or both. Project Results Emissions from methane point sources in California were estimated to be equivalent to 34 percent to 46 percent of the state’s methane inventory for 2016. Methane point sources were observed at a total of 564 of the surveyed facilities and infrastructure elements (0.2 percent). Super-emitter activity occurs in every surveyed sector. Over the entire population of observed point sources, 10 percent of sources contributed ~60 percent of emissions. The largest methane point source emitters in the state (43% of the total emissions in this study) are 32 landfills (including 2 composting operations). Flight imagery includes examples of strong methane plumes at these landfills associated with gaps in intermediate cover and/or leaking gas capture wells. (Intermediate cover is compacted earthen material of at least 12 inches placed on the surface of a fill where no additional solid waste will be deposited within 180 days.) These plumes represent significant mitigation opportunities. Study results suggest that the majority of waste disposal facilities emit methane as area sources or as point sources below this study’s detection limit and that landfills exhibiting point sources are a unique subpopulation. The team found that about 26 percent of methane point source emissions in California are from the oil and gas supply chain, with nearly 80 percent of that due to production. Spatially, 85 percent of point source emissions from production are concentrated in the southern San Joaquin Valley (the highest oil- and associated-gas producing region in the state), 14 percent in Los Angeles and Ventura counties, and 1 percent in the Sacramento Valley. These emissions are attributed to a variety of oil and gas production infrastructure, including well heads, gathering lines, and storage tanks. The researchers found no compelling evidence of strong methane emissions from abandoned oil or gas wells or specific to fracking operations, although more detailed analysis is recommended to confirm. Methane point source emissions from natural gas infrastructure in California appear to be due to a combination of normal process emissions and anomalous leakage at processing plants, a small number of compressor stations on transmission pipelines and underground storage facilities, gas-fired power plants, and leaks in distribution pipelines. The methane point source emissions observed at most refineries and at seven power plants in California appear to be generally higher than reported to the EPA; however, additional study is recommended to pinpoint the causes. The team estimates that California’s refineries and the outlier power plants contribute about 5 percent of the total methane point source emissions in the state. Overall, the team found that methane point source emissions from the natural gas sector in California are generally consistent with the State’s 2016 methane inventory, with the aforementioned exceptions as well as a caveat that this study was not designed to address the potential for a large number of small leaks downstream of production, processing, and transmission. In particular, this study cannot rule out large disagreements between reported 2 and actual fugitive methane emissions from the dense natural gas distribution system in some major urban areas. The team surveyed 443 confined animal feeding operations in California – an estimated 71 percent of all such facilities in the state. Manure management at large dairies in the San Joaquin Valley is recognized as one of the top methane emitting sectors in California. The survey results are consistent with this, and wet manure management – particularly settling ponds and anaerobic lagoons – is found to be responsible for about 26 percent of total methane emission from point sources in California. Methane emission sources at these facilities are diverse and complex and could benefit from additional intensive study, including on-farm measurements. Methane digesters are increasingly being deployed at California dairies in an effort to reduce the net greenhouse gas impact of each facility while offering additional revenue opportunities, such as biogas for energy production. The survey covered about 25 known dairy digesters in California, including a combination of facilities in operation and still undergoing construction. In principle, a well-functioning digester should capture methane from manure management; however, the study indicated significant and fairly persistent methane point sources at four dairy digester facilities in the San Joaquin Valley. This suggests that future monitoring for atmospheric methane around dairy digester facilities before and after digester construction could prove useful for assessing their efficacy in meeting mitigation objectives while helping operators avoid unintentional biogas product loss. A total of 58 wastewater treatment facilities across California were surveyed, of which ten exhibited methane point source plumes. Of these ten, three were persistent point sources, suggesting potentially anomalous activity. Based on AVIRIS-NG survey results, this entire sector is estimated to be responsible for about 2 percent of total methane point source emissions in California. As a general finding, with the exception of the landfill emissions, many of the methane point sources detected by this survey were highly intermittent so for every source the researchers calculated a persistence or frequency that is simply the number of observed plumes divided by the number of observations. This resulted in a median persistence of 0.20 for the entire population (mean 0.33, range 0.02 – 1.0). In some cases, the intermittent emissions can be explained by normal operations (e.g., periodic waste flushing at large dairies). In other cases, more persistent activity appears to be due to sustained venting at a small number of anaerobic digesters at dairies and wastewater treatment plants or leaking bypass valves at natural gas compressor stations. The researchers found a similar distribution of persistence (.20 to .35 on average) and emissions in the manure management, wastewater treatment, and oil and gas sectors. Persistence numbers are applied to the emission estimate for each source, effectively lowering the average emission rates for most sources. This intermittency highlights the need for more frequent sampling. The preliminary findings, including high resolution methane plume images, were shared with the operators of methane point source facilities, who provided verification with surface observations or explained the underlying mechanisms for the observed emissions, or both. Several of these collaborative efforts directly led to mitigation of the methane sources detected by the survey, including four leaking natural gas distribution lines and one leaking liquified natural gas storage tank. 3 Knowledge Transfer Much of the data analysis system used in this study was developed under parallel NASA programs. Those data analysis capabilities, including a web-based methane data portal (http://methane.jpl.nasa.gov) that displays all methane plumes detected during this study are to be transferred to the California Air Resources Board for sustained operation. Additionally, over the course of this study the project team organized multiple meetings and briefings to share and discuss interim findings with stakeholders, including staff from the Energy Commission, Air Resources Board, CalRecycle, Bay Area Air Quality Management District, South Coast Air Quality Management District, Southern California Gas company, Pacific Gas and Electric company, Milk Producer’s Council, City of Los Angeles Department of Sanitation, Sunshine Canyon Landfill Local Enforcement Agency, and several operators of individual facilities. These interactions resulted in two-way knowledge transfer, including feedback on the utility of the methane data sets as well as ground truthing and explanations about potential causes for observed emissions. Another source of knowledge transfer was the publication of a paper in the journal Nature (Duren et al., 2019). Benefits This project has provided new insights into California’s methane inventory with the first systematic assessment of the relative contributions of methane point sources, including their distribution by space, time, and emission sector. These findings may lead to improvements in California’s greenhouse gas inventory and to efforts by state and local agencies and businesses to both prioritize future investments in methane emissions mitigation and assess overall progress towards emissions reduction goals. This work also highlights the potential for efficient point source monitoring techniques to directly enable mitigation of a broad class of methane super-emitters, representing a significant contribution to California’s climate stabilization targets. Based on this research, point source emissions for the oil and gas sector in California are estimated to be 0.158 TgCH4/yr (95 percent confidence interval 0.135-0.184 TgCH4/yr). If translated to natural gas equivalent, these emissions represent about $28-$39 million in annual product loss using July 2018 United States prices. This indicates the potential value of mitigation for California ratepayers, additional to climate benefits. 4 CHAPTER 1: Project Purpose Motivation Methane is a powerful greenhouse gas and is targeted for emissions mitigation by the State of California (California 2017). Methane is also a precursor for tropospheric ozone and is strongly linked with co-emitted reactive trace gases that are the focus of air quality and public health policies, particularly in high priority regions such as the San Joaquin Valley (SJV) and the South Coast Air Basin (SoCAB). Globally, the atmospheric growth rate of methane is likely strongly influenced by anthropogenic emissions from a population of spatially condensed point sources distributed over large areas and spanning diverse socio-economic sectors. However, “bottom-up” (inventory-based) estimates of methane emissions are often in disagreement with top-down (atmospheric measurement based) estimates (Wecht et al., 2014, Turner et al., 2015, Wong et al., 2016, Jeong et al., 2017, Cui et al., 2019). Limitations in process-based understanding of methane emissions is exemplified by the ongoing scientific discussion on both the hiatus in the atmospheric growth rate of methane in the early 2000’s and the unexpected rise starting in 2007 (Kirschke et al., 2013). Emissions and process attribution remain highly uncertain but are needed to resolve key elements of the global carbon budget, generate accurate greenhouse gas inventories and inform emission mitigation decisions. A key factor is that many current methane monitoring methods (bottomup and top-down) are limited to regional or coarser scale resolution and often cannot detect individual sources or attribute fluxes to specific activity and facilities. Other methods are sufficient for studying previously known sources but are not well suited to surveying large areas for unknown sources. Hence methane emissions remain a challenging target for abatement since the locations and emission fluxes of many significant sources are still mostly unknown. These challenges are reflected in the recently enacted California AB 1496 law: “there is an urgent need to improve the monitoring and measurement of methane emissions from the major sources in California” and directs the California Air Resources Board to “undertake, in consultation with districts that monitor methane, monitoring and measurements of high-emission methane hot spots in the state using the best available and cost-effective scientific and technical methods”. Another motivation is supporting efforts by natural gas utilities to improve leak detection and repair, a general benefit to California ratepayers. Prior Studies California has benefited from a number of top-down studies focused on methane. The 2010 CalNex campaign addressed many sectors and priority regions such as the SoCAB and SJV (Ryerson et al., 2013). There has been an ongoing focus on SoCAB methane emissions and trends (Wennberg et al., 2012; Wunch et al., 2016; Wong et al., 2016), source attribution (Hopkins et al., 2016), and characterization of individual sources such as the Aliso Canyon gas leak incident (Conley et al., 2016; Thompson et al., 2016). Recent years have also seen a dramatic improvement in the ability of passive remote-sensing methods to detect and locate large methane sources. Observations from polar orbiting 5 satellites have detected strong, persistent enhancements of atmospheric methane in the Four Corners region and California’s SJV (Kort et al., 2014) and have produced spatially resolved estimates of United States methane emission trends (Turner et al., 2016). The 2017 launch of the TROPOMI instrument on the Sentinel-5 Precursor satellite should further advance spacebased methane detection for global studies (Butz et al.,2012). However, the ability of satellites to detect and quantify emissions from point sources is still limited to relatively coarse spatial scales (typically 25 km at best). Some surface measurement networks and models can resolve methane fluxes at resolutions as fine as a few kilometers but so far this is limited to a few urban testbeds (McKain et al., 2015) and in most cases is insufficient to pinpoint and attribute point sources. JPL and partners have devised a tiered observational strategy for efficiently surveying large areas for methane point sources, quantifying individual source emissions, and estimating their contributions to the net emissions of key regions and sectors. The strategy is flexible with regards to vantage points and measurement systems – enabling significant near-term progress using existing NASA remote sensing instrumentation that were developed as prototypes for next generation satellites. Over the past four years this strategy was tested with a series of exploratory airborne field campaigns over California’s Central Valley and SoCAB (Thompson et al., 2016) as well as the Four Corners region (Frankenberg et al., 2016). Project Objectives Based on the success of exploratory NASA airborne campaigns and in response to California policy needs the California Air Resources Board (CARB) and California Energy Commission (CEC) funded Jet Propulsion Laboratory (JPL) to conduct the first comprehensive airborne survey of methane point sources in the state. In this study, a point source is defined as a condensed surface feature or infrastructure component (typically less than 10 meters across) that emits a plume of highly concentrated methane. This is in contrast to an “area source” or the combined effect of many small emitters distributed over a large area (typically 1 to 100 km across) that release methane in a more diffuse fashion including anaerobic decomposition occurring with rice cultivation and enteric fermentation from livestock, both of which are better addressed with other measurement methods and not included in this study. The project technical objectives were as follows: • Use state of the art airborne instruments and methane detection algorithms to conduct a California methane source survey over key regions that are major contributors to California’s methane budget. • Measurement data will be processed into maps of large source emitters detected within the areas flown. • Provide the Energy Commission with the locations and plume characteristics of large fugitive emission sources located within the survey area for the natural gas system (CEC) and other relevant point source sectors (CARB). 6 CHAPTER 2: Project Approach Observing Strategy for Methane Emissions Data collection involved a broad airborne survey of methane point sources spanning key regions and sectors across the state by JPL’s Next Generation Airborne Visible/Infrared Imaging Spectrometer (AVIRIS-NG) in Figure 1. Figure 1: Observing Strategy Approximately 2000 individual AVIRIS-NG flight lines flown in 2016 (blue) and 2017 (green) covered over 272,000 individual facilities and infrastructure elements. Detected sources are indicated by red points with the densest clusters in the San Joaquin Valley (dairies and oil fields). The inset images show examples of representative methane plumes from different sectors: A. compressor stations at a natural gas storage facility, B. oil well, C. liquified natural gas tank, D. dairy manure management, E. wastewater treatment plant, F. landfill (Duren et al., 2019). The color scales indicate the methane concentration-length enhancement in each pixel in units of parts per million-meter (ppm-m). Surface map images: Google Earth (basemap) and AVIRIS-NG (inset images). Source: Duren et al. 2019 The airborne remote sensing method applied is not optimized for detecting and quantifying area sources and hence methane emissions from area sources such as enteric fermentation, rice cultivation and wetlands are excluded from this study. 7 AVIRIS-NG Instrument and Methane Retrievals The next generation Airborne Visible/Infrared Imaging Spectrometer (AVIRIS-NG) measures ground-reflected solar radiation from the visible to infrared spectral regions (350 to 2,500 nm). This push broom instrument has a 34° field of view and operates on high performance aircraft, allowing for efficient mapping of large regions. Increasing flight altitude affects the ground resolution, i.e., the size of each image pixel increases while the image swath increases (Figure 2, Table 1). For most of the California Methane Survey, AVIRIS-NG flew at 3 kilometer (km) above ground level, resulting in 3 meter (m) image pixels on average. Figure 2: AVIRIS-NG Flight Parameters 0B L=image swath width, V=aircraft velocity, FOV=field of view, IFOV = instantaneous FOV. Source: Murai, 1995. The methane retrieval is based on absorption spectroscopy (Figure 3) and has been used for a number of prior NASA research campaigns including Bakersfield area oil fields (Thompson et al., 2015), a campaign to the Four Corners region in Colorado and New Mexico (Frankenberg et al., 2016), Aliso Canyon (Thompson et al., 2016), and a study of California landfills (Krautwurst et al., 2017). A methane controlled-release experiment indicated consistent detection of plumes for releases as low as 14.16 m3/h (~10 kgCH4/hr) at multiple AVIRIS-NG flight altitudes and variable wind speeds (Thorpe et al. 2016). Table 1: AVIRIS-NG Image Parameters Flight altitude (meters above ground level) Image swath width (meters) Ground resolution (meters) 1,000 611 1 2,000 1,223 2 3,000 1,834 3 Source: JPL 8 Figure 3: Methane Absorption Signature for AVIRIS-NG Methane absorption signature (transmittance) plotted for the wavelength range measured by AVIRIS-NG. Strong absorptions are present between 2,200 and 2450 nm. Source: JPL The detected quantity is a mixing ratio length in units of ppm m representing the thickness and concentration within a volume of equivalent absorption. This is equivalent to an excess methane concentration in ppm if the layer is one meter thick (i.e. directly equivalent to ppb km). At a scale height of about 8 km, the total column averaged excess mixing ratio Xmethane would be about 0.000125 times the excess in ppm-m. For example, 1000 ppm-m is equivalent to an Xmethane enhancement of 125 ppb. Integrating over the physical area of the plume yields an Integrated Methane Enhancement (IME) in kg, as in Thompson et al. (2016) and Frankenberg et al. (2016), tantamount to the total observed mass of methane above the ambient background. This technique can be combined with simple steady state assumptions for a first-order estimate of a point source emission flux. Methane retrievals are performed in real time onboard the aircraft (Figure 4), which permits the instrument operator to identify and geolocate plumes in real time. Figure 4: Real Time Methane Mapping Real time methane mapping onboard the aircraft. Red methane plumes are overlaid on raw AVIRIS-NG image. Source: JPL 9 This information can be used for adaptive surveying and results communicated down to ground crews for rapid follow up. At the end of each flight day, methane quick-look data products (Figure 5) are generated and used to quickly assess results and plan future flights. After the AVIRIS-NG data is transported to JPL, scenes are reprocessed to generate methane retrievals for orthorectified scenes (planimetrically correct images with constant scale). Figure 5: Methane Quick-Look Products Methane quick-look products are generated at the end of each flight day. This example shows a plume from a leaking low-pressure gas pipeline that was confirmed and repaired by the gas company. Source: JPL Airborne Survey Design Figures 6, 7 and 8 illustrates examples of AVIRIS-NG flight planning including the diversity of emission sectors and their spatial distributions. The flight planning was governed by two primary objectives: 1) spatial coverage sufficient to map the infrastructure in the state most likely responsible for >60 percent of methane point source emissions (with >80 coverage for key sectors) and 2) sufficient number of revisits to have a reasonable probability of detecting intermittent emission sources (for example for a source that is active 25 of the time, six visits should provide a detection probability of 0.82). In addition to the broader goal to map and revisit large areas the team also conducted several intensive studies focused on gaining insight into key emission processes. One intensive focused on an area near Visalia that was mapped repeatedly over a 5-hour period to investigate the temporal variability of manure emissions from more than 100 dairies with 60minute revisit intervals. Others focused on natural gas infrastructure across southern California, gas-fired power plants during heat wave conditions and refineries in the LA basin and San Francisco Bay Area. In several cases coordinated, contemporaneous measurements were conducted with mobile on-road laboratories, fixed surface observations and other airborne systems to help validate source locations and emission estimates. 10 Figure 6: Airborne Survey Design for the Southern San Joaquin Valley Southern San Joaquin Valley Source: JPL A Geographic Information System (GIS) data set known as Vista-CA that maps potential methane emitting infrastructure across the State of California was developed by researchers at the University of California Riverside to assist with flight planning and for source attribution following detection of methane plumes (Duren et al., 2019). The Vista-CA data set applies similar methods previously used to develop a Vista-LA methane GIS data set for the greater Los Angeles area (Carranza et al., 2017). Vista-CA mapped the locations of infrastructure associated with three primary sectors (energy, agriculture, and waste) following the frameworks used by the State of California’s Greenhouse Gas Inventory and the IPCC Guidelines for GHG Reporting. Vista-CA contains 450,572 distinct pieces of potential methane emitting infrastructure and was used to guide selection of flight boxes (Duren et al., 2019). Many of the Vista-CA elements were readily derived from public data records but others were more challenging and required some new development. For example, the natural gas pipeline numbers in Vista-CA include transmission, distribution, gathering and “other” categories. The 4,599 km of gas transmission lines in California was derived from a combination of NMPS, CEC and EIA data but there is no publicly available map of distribution lines in urban areas. To address the latter gap a residential distribution line mask was constructed using parts of the California road network overlaid on raster cells classified as being 20-100 percent impervious in urban areas from the National Land Cover Dataset (NLCD; 53) and connected it to the existing NG Pipeline infrastructure using a 10km distance tolerance (Duren et al., 2019). Survey coverage was computed by using the AVIRIS-NG flight path (1800m width) rectangular 11 polygons as clip features to pull out overlapping pipelines and recalculate segment length within each AVIRIS-NG survey polygon. Oil and gas infrastructure is divided into two main categories: Production Sites (including well heads, pumpjacks, and other equipment immediately associated with extraction) and Other Production Equipment. Although the VistaCA layers include 3,356 pieces of “other Production Equipment” such as condensate tanks and waste ponds that correlate well with satellite imagery of facility infrastructure this category should also include gathering lines for which the team had very limited information. For this reason (and that more than 80 percent of production fields in the state were surveyed) the emissions results from other production equipment are not up-scaled. Dairies are another special case given the number and magnitude of methane sources and complexity in identifying which facilities are more likely to be point source emitters (see Duren et al., 2019 for details). Figure 7: Airborne Survey Design for the Northern San Joaquin Valley Northern San Joaquin Valley Source: JPL 12 Figure 8: Airborne Survey Design for Northern California Northern California Source: JPL Data Analysis The analysis for this study (Figure 9) consists of a) standard processing including calibration and orthorectification of the AVIRIS-NG image cube data, b) retrieval of CH4 column mixing ratio-lengths and generation of CH4 plume maps, c) quality control and filtering of plumes, d) geolocation and attribution of CH4 plumes to Vista-CA spatial layers, e) calculation of integrated methane enhancement (IME) and length for each plume, f) acquisition and processing of High Resolution Rapid Refresh (HRRR) reanalysis wind fields, g) emission flux estimation and uncertainty quantification for individual methane plumes, h) filtering and removal of plumes with sub-optimal shapes, are redundant/overlapping with others plumes or have excessive errors in IME and/or wind speed estimates, i) validating emission estimates with independent methods, j) averaging and scaling plume emission estimates with observed persistence to derive an annual net emission for each source, k) applying Vista-CA spatial layers to calculate net emission estimates for facilities and key sectors statewide, l) apply bootstrap analysis to determine confidence intervals for each sector and total population. Each of these steps is described in detail in Duren et al. (2019) with an overview below. 13 Figure 9: Data Analysis Workflow HRRR 10m, 80 m wind fields; 10 nearest 3km grid cells over -1 to + 1hr AVIRIS-NG pipeline Wind data ingest & validation L0 raw data NWS wind observations L1 calibrated radiance CH4 retrievals L2/L3 orthorectified data products: RGB images, grayscale CH4 images, flight line maps Calculate avg 10m wind speeds and uncertainties at plume locations, times Vista-CA spatial layers Plume analysis 1. Detect and verify valid plume 2. Assign source # 3. Geolocate plume/source origin 4. Record line #/date/time and source coordinates 5. Identify nearest Vista-CA element ID, facility name, source type and IPCC emission sector 6. Plume filtering: eliminate excessively cluttered plumes and redundant/overlapping detections 7. Calculate and apply plume aspect ratio & thresholds 8. Calculate IME, plume length, uncertainties 9. Calculate plume flux and uncertainty 10. Remove estimates with > 100% uncertainty 11. Remove estimates with > 100% disagreement w/IME proxy method Plume list Source analysis 1. Calculate average flux for each source 2. Calculate source persistence from Nobs and Mflights 3. Apply persistence scalar to calculate net source flux Source list Facility & Sector analysis 1. Calculate aggregate flux for facilities 2. Apply Vista-CA population scalars to calculate sectoral & State totals 3. Bootstrap analysis to derive confidence intervals for key sectors Colorized plume images Facility List Source: JPL The AVIRIS-NG flights conducted during this survey detected 1,181 individual methane plumes that were each attributed to a Vista-CA infrastructure element (Tables 2 and 3). Plumes were identified manually for this study. An experimental machine learning system based on a convolutional neural network was trained on a subset of plumes from this and other field studies and then used to assess potential false positives and false negatives in the manual plume list. The observed presence or absence of a plume at each source was used to calculate its persistence (frequency of occurrence); e.g., the ratio of plume occurrences to the number of overflights of a given source. Many of the sources were highly intermittent – with a median persistence of 0.20 for the entire population (mean 0.33, range 0.02 – 1.0). The survey provided a median of nine samples per source (range 1-66) for the population, translating to a median probability of 0.75 that the persistence is at least as high as reported (mean 0.82). Filtering criteria were used to eliminate plumes with noisy retrieval results and complex shapes from the overall emissions analysis. An integrated methane enhancement (IME) and plume length for each plume were computed using methods that build on those demonstrated in previous studies (Thompson et al., 2016; Frankenberg et al., 2016). Near surface wind speeds were calculated for each plume location and overflight time using NOAA’s HRRR data set (3km, hourly resolution) with validation from surface weather observations. Methane emissions and uncertainties were calculated using the IME, plume length and wind speed data for every plume. Additional filtering was then applied using the aforementioned IME proxy method to calculate emissions. Plumes emission estimates that differed by > 100 percent between the two methods were eliminated from the source emissions analysis. The net result of the filtering steps left 1050 plume emission estimates for the analysis. An average emission rate was then calculated for each source using the plume emission estimates and the observed 14 source frequency or persistence. This process resulted in emission estimates and 1sigma () uncertainties for 564 sources at 230 facilities and infrastructure elements. For most sectors, the extent of the observed methane plume was small compared to the full spatial extent of the associated facility and generally appeared in a repeatable fashion from the source to which it was attributed. For most sectors emissions for individual sources are reported, with larger facilities often including multiple sources. However, a different accounting scheme is used for landfills given the complexity of emission processes. For landfills where plumes were detected, large plumes spanning the spatial extent of the facility were observed. Additionally, in most cases the location of each landfill plume evolved significantly over time in response to daily changes in waste deposition and surface cover. The team defined each landfill with observed methane plumes as a composite source. All plume observations at a given landfill, within a single flight line, were summed to get the total facility emissions per flight line for that sample interval. This process is defined in more detail in Duren et al., 2019, SI section S2.8. 15 CHAPTER 3: Project Results The following caveats apply to these results: • The remote sensing methods applied in this project were not optimized for detecting and quantifying area sources and hence methane emissions from area sources such as enteric fermentation, rice cultivation and wetlands are excluded from this study. • With the exception of approximately 100 sources, most of the sources reported in this report have not yet been verified with surface measurements. This project was limited to remote sensing methods and was not funded to conduct follow-up surface verification. † This means that there are some residual uncertainties about source attribution that could result in misidentification of facilities and/or incorrect assignment of a source to a given emission sector. 1F1F1F 1F1F • This project was also not funded to determine which sources are normal process emissions such as periodic venting as opposed to a leak or other malfunction. A few exceptions are noted where a root-cause was confirmed (through surface follow-up measurements or through consultation with a facility operator). Airborne Survey Statistics The actual implementation of the airborne survey was influenced by the planning activity described in Section 2, response to discovery of methane plumes (e.g., follow-up observations), and impacts due to weather and aircraft availability. Survey Completeness The survey covered approximately 271,556 distinct facilities and infrastructure components (out of 449,648 candidates) spanning 21,699 km2 of land area at least once (Figure 10). A significant fraction of these flight lines were flown more than once, resulting in 54,817 km2 total area coverage (Duren et al., 2019). † See https://ww2.arb.ca.gov/our-work/programs/methane/ab1496-research for CARB Methane Hotspots Research website including follow up measurements of some sources detected by this and other studies. 16 Figure 10: Spatial Coverage for Survey As-flown AVIRIS-NG flight lines (cyan) showing area covered during the California Methane Survey. Source: JPL Compared with the Vista-CA GIS data set the survey achieved a completeness per emission sector that ranged from 32 to 100 percent (Table 2; Duren et al., 2019). Note that most of the categories shown here represent facilities or other discrete infrastructure features with the exception of transmission pipelines – as linear features the latter are reported as fraction of total length. Also, for landfills the survey focused on only the likely top emitters – the 60 facilities predicted to be responsible for 90 percent of California’s landfill methane emissions based on bottom-up estimates from CARB. 17 Table 2: Survey Completeness by Emission Sector IPCC Emission Sector Gas-fired Power Plants CARB Inventory (2014) EIA (2016) EPA FLIGHT (2016) 238 Refineries CARB Inventory (2014) EIA (2016) EPA FLIGHT (2016) 26 26 100.0% 264 107 25 461 162 46 66.0% 54.3% 538 1,131 47.6% 68,548 216,774 31.6% 23 26 88.5% 12 12 100.0% sub-totals CNG Fueling Stations LNG Fueling Stations Natural Gas Stations (non-storage compressor, dehydration, metering, odor, etc) Natural Gas Pipelines (length in km) Natural Gas Processing Plants Natural Gas Storage Fields AFDC (2017) AFDC (2017) CEC (2017) EPA FLIGHT (2016) CEC (2012) EIA (2017) NLCD (2011) NPMS (2013) U.S. Census Bureau (2017) EIA (2014) DOGGR (2016) EIA (2016) Oil and Gas: Other production DOGGR (2018) equipment Oil and Gas: Wells DOGGR (2018) sub-totals 3A2 Manure Management All dairies Dairies CAFOs with >1000 head Composting Sites 4A1 Solid Waste Disposal Sites Landfills Solid Waste Disposal Sites (landfills) 4D1 & 4D2 Wastewater Treatment Percentage of Percentage of Total Number Vista-CA IPCC of Vista-CA Infrastructure Emission Infrastructure Elements Sector Elements Surveyed Surveyed Vista-CA Infrastructure Elements 1A1 Energy Sectors 1B2 Oil and Natural Gas Number of Features Surveyed by AVIRIS (20162017) Vista-CA Infrastructure Elements Data Sources Wastewater Treatment Plants Domestic Wastewater treatment & discharge Industrial Wastewater treatment & discharge TOTALS CIWQS (2018) CARB (2015) RWSCB - Region 5 (2017) SJVAPCD (2017) CalRecycle (2015) CARB (2015) EPA FLIGHT (2016) CalRecycle (2015) CARB (2015) EPA FLIGHT (2016) CARB (2016) EPA FLIGHT (2016) other (satellite imagery) 435 54.7% 57.3% 2,872 3,356 85.6% 198,231 270,356 225,766 447,273 87.8% 890 1,544 57.6% 60.4% 64.5% 443 620 71.5% 166 430 38.6% 38.2% 270 716 37.7% 57 148 38.5% 38.5% 1 n/a 272,447 n/a 451,192 60.4% Source: Duren et al., 2019 In terms of temporal completeness the survey sampling ranged from one visit per source to multiple visits distributed over the project time span. In some cases (e.g., intensive study of dairies near Visalia and some studies of underground gas storage fields) revisit intervals as short as a few minutes were obtained over the course of a day, providing insight into diurnal variability. Most of the overflights occurred between the hours of 10 am and 3 pm local time. 18 Spatial, Temporal and Sectoral Distribution of Emissions Our analysis provided emission estimates and 1sigma () uncertainties for 564 sources at 230 facilities and infrastructure elements. The locations of the 564 confirmed point sources are shown in Figure 11, indicating that most of the strong point sources detected in this survey are concentrated in the southern half of the state- particularly the SoCAB and areas in the SJV with the largest concentrations of dairies and oil/gas fields. The point source population has a heavy-tail distribution indicating that 10 percent of the point sources are responsible for 60 percent of the point source emissions (Figure 12). This is generally consistent with previous studies of the US oil and gas supply chain (Alvarez et al., 2018, Zavala-Araiza et al., 2015, Brandt et al., 2015), but here the team also observed the super-emitter behavior in every surveyed emission sector including manure management, landfills, wastewater treatment plants and refineries. The sum of the measured source emissions is 0.511 Tg CH4/yr and a non-parametric bootstrap analysis was applied to the population of observed sources to calculate a 95 percent confidence interval of 0.433 - 0.601 Tg CH4/yr (Duren et al., 2019). The repetitive, high spatial resolution plume imagery from the California Methane Survey allowed us to characterize point source behavior and controlling processes, particularly for sectors that have not been as well studied as the oil and gas production sector. Many of these point sources are highly intermittent. In some cases, the intermittent emissions can be explained by normal operations (e.g., periodic waste flushing at large dairies) albeit with higher than expected emission rates. In other cases, more persistent activity is apparently due to sustained venting at a small number of anaerobic digesters at dairies and wastewater treatment plants or leaking bypass valves at natural gas compressor stations. A similar distribution of source persistence/frequency (20-35 percent on average) was found and emissions in the manure management, wastewater treatment and oil and gas sectors. The methods used to estimate source persistence are described in Duren et al., 2019 SI section 2.9. However, it should be noted that this survey was primarily intended to provide completely spatial sampling rather than a comprehensive assessment of source persistence. It is recommended that future studies be designed to provide frequent and uniform temporal sampling of intermittent emissions to provide more robust statistics. Solid waste management is the largest methane point source emission sector in California (Table 3) with persistent plumes only observed at 32 of 436 surveyed landfills and composting facilities. The imagery of landfills identified methane plumes associated with construction, gaps in intermediate cover and leaking gas capture wells – indicating a sub-population of anomalous emitters. The team did not detect a larger population of smaller methane point sources across the landfill sector, which suggests the majority of those facilities emit methane as area sources that are not detectable with this method. Since a significant fraction (32-100 percent) of every point source emission sector in California was surveyed, the team can upscale their measurements to estimate statewide point source emissions. Table 3 gives coverage scalars for each sector derived by combining the Vista-CA infrastructure data with the AVIRIS-NG flight coverage. For most sectors the scalar is simply the number of Vista-CA elements divided by the number of elements surveyed at least once during this study – with three exceptions where additional constraints were applied to reduce or eliminate scaling. This results in 0.618 (95 percent confidence 0.523-0.725) TgCH4 yr-1, equivalent to 34 - 46 percent of the California Air Resources Board (CARB) methane inventory 19 for 2016 (Duren et al., 2019). Solid waste management contributes 41 percent of observed point source emissions followed by 26 percent from manure management and 26 percent from oil and gas (in contrast to 32 percent, 39 percent and 25 percent of total methane emissions for those sectors according to the CARB the inventory). The rest of California’s methane budget is likely due to area sources such as enteric fermentation, rice cultivation and nonsuper-emitter landfills as well as a large number of low emission sources in the downstream natural gas supply chain that fall below the detection threshold of this survey (Ellis et al., 2010, Fitzgerald et al., 2000, Wennberg et al., 2012). Any under-estimates in the CARB inventory (Wecht et al., 2014, Turner et al., 2015, Jeong et al., 2014, Cui et al., 2019) will reduce the relative contribution of point sources to California’s total budget. Figure 11: Locations of Methane Sources Detected by Survey AVIRIS-NG flight lines (cyan) and methane sources (red points). Source: JPL The distribution of detected methane point sources by Intergovernmental Panel on Climate Change (IPCC) emission sector is summarized in Figure 13 and Table 3. This offers some insight into the potential total population of point sources in the state (e.g., fraction of sampled infrastructure where at least one methane source was detected). 20 Figure 12: Distribution of Methane Emissions from Individual Sources 10000 100% 90% 1000 80% 100 60% 50% 10 40% % of emissions Emission (kg/hr) 70% 30% 1 20% 10% 0% 5% 10 % 15 % 20 % 25 % 30 % 35 % 40 % 45 % 50 % 55 % 60 % 65 % 70 % 75 % 80 % 85 % 90 % 95 % 10 0% 0% 0 % of sources Emissions % of total Distribution of average methane emissions and 1 uncertainties for the 564 point sources detected in the survey, adjusted for source persistence (frequency) in units kgCH4/hr. The heavy tail indicates that 10 percent of detected point sources contribute 60 percent of the population total emissions. Source: Duren et al., 2019 21 Figure 13: Emission Histograms for Key Sectors Histograms indicating the density of measured methane point source emissions (adjusted for persistence) for each of the key sectors in California (kgCH4/hr). Managed waste disposal exhibits qualitatively different behavior than the other sectors, with point sources only appearing at 32 persistent, high emitting landfills – likely constituting a distinct sub-population within that sector. Source: Duren et al., 2019 22 NG Storage Fields Oil & Gas: Wells Oil & Gas: Other Production Equipment # of surveyed elements 461 264 208 132 1,131 538 216,774 68,548 26 23 12 12 1.83 7 0.007 0.013 100 1.00 37 0.015 0.015 57 1.27 44 0.022 0.028 63 1.58 6 0.002 0.003 48 2.10 5 0.005 0.010 0.009, 0.012 1.6% 32 3.16 5 0.004 0.012 0.010, 0.014 1.9% 88 1.13 5 0.004 0.004 100 1.00 11 0.009 0.009 88 1.14 107 0.048 0.054 86 1.00 120 0.066 0.066 225,766 198,231 3,356 2,872 23 0.007, 0.021 0.008, 0.023 0.015, 0.044 0.003, 0.004 0.004, 0.005 0.008, 0.010 0.046, 0.063 0.056, 0.076 % of total emissions 55 State total 95% confidence intervals (TgCH4 y-1) State Total Emissions (TgCH4 y-1) 1B2 Oil and Natural Gas CNG/LNG Fueling Stations NG Stations (nonstorage compressor, metering, etc) NG Pipeline (transmission, distribution) NG Processing Plants 26 Measured emissions (TgCH4 y-1) sub-totals 26 N sources detected Refineries 238 Sectoral Scalar 1A1 Energy Industries 435 % surveyed Gas fired power plants # of Vista-CA infrastructure elements Vista-CA infrastructure element IPCC Source Category Table 3: Summary of Total Emissions by Sector 2.1% 2.4% 4.6% 0.5% 0.7% 1.4% 8.8% 10.7% 57 n/a n/a 449,648 271,556 % of total emissions 148 State total 95% confidence intervals (TgCH4 y-1) 436 State Total Emissions (TgCH4 y-1) 1,146 Measured emissions (TgCH4 y-1) Totals 443 N sources detected 4D1, 4D2 Wastewater Treatment & Discharge 620 Sectoral Scalar 4A1 Managed Waste Disposal Dairy Confined Animal Feeding Operations Landfills & composting facilities Domestic & industrial wastewater treatment Industrial wastewater treatment: beef processing 447,273 270,356 % surveyed 3A2 Manure Management # of surveyed elements # of Vista-CA infrastructure elements Vista-CA infrastructure element IPCC Source Category sub-totals 60 1.16 259 0.137 0.158 0.135, 0.184 25.6% 71 1.40 215 0.115 0.161 0.137, 0.187 26.1% 38 1.11 32 0.229 0.255 0.175, 0.345 41.3% 39 2.60 12 0.004 0.012 0.005, 0.020 1.9% n/a 1.00 2 0.004 0.004 0.004, 0.005 0.6% 60 1.21 564 0.511 0.618 0.523, 0.725 100.0 % Summary of persistence (frequency) adjusted point source emissions by IPCC sector from this study and estimated total emissions derived with population scalars. Most of the scalars are simply the ratio of the number Vista-CA infrastructure elements to the number of surveyed elements with three exceptions highlighted in blue font (other oil and gas production equipment, landfills and industrial wastewater treatment) where scaling is further constrained or eliminated. Source: Duren et al., 2019 24 Figure 14: Measured Emissions vs CARB Inventory 0.45 0.40 0.35 0.30 0.25 0.20 0.15 0.10 0.05 0.00 1A1 Energy Industries 1B2 Oil & Natural Gas 3A2 Manure Management All sources (CARB 2015 inventory) 4A1 Managed Waste Disposal Sites 4D1 Domestic Wastewater Treatment & Discharge Point sources only (this study) Comparing statewide methane point source emission estimates from this study and the relevant sectors in the 2016 CARB inventory that are likely to include point sources (11). The whiskers indicate the 95 percent confidence intervals from this study. Source: Duren et al., 2019 The following sections provide additional findings for methane point source emission sector. Sector Specific Findings Oil and Gas Production and Processing Approximately 79 percent of the oil and gas sector emissions in the study are attributed to production in California. Spatially, 85 percent of point source emissions from production are concentrated in the southern San Joaquin Valley (the highest oil- and associated-gas producing region in the State), 14 percent in Los Angeles and Ventura counties, and 1 percent in the Sacramento Valley. There is no compelling evidence of strong methane emissions from abandoned oil or gas wells or specific to fracking operations although more detailed analysis is recommended to confirm that. The Vista-CA “Oil and Gas: Wells” category (Table 2) includes active well heads, pumpjacks, and other equipment immediately associated with extraction and also inactive wells. “Oil and Gas: Other Production Equipment” is derived the “California Statewide Oil and Gas Production or Injection Facility Boundary” data set from DOGGR https://maps.conservation.ca.gov/doggr/metadata/FacilityBoundaries.html. To the team’s knowledge that is the best publicly available database on locations of oil and gas production infrastructure in California that may emit methane including permanent tanks, flowlines, headers, gathering lines, wellheads, heater treaters, pumps, valves, compressors, injection equipment, production safety systems, separators, manifolds, and pipelines. However, that 25 database does not likely cover all such equipment statewide, and also excludes production equipment known to emit methane, such as separators, water tanks, acid gas removal units, and dehydrators. The team had limited information about the spatial distribution of some components such as gathering lines. For these reasons, and the fact that more than 80 percent of production fields in the state were surveyed, no attempt was made to upscale the emissions results from other production equipment. Production-normalized emissions for the largest associated-gas producing ‡ oil and gas fields in the SJV covered by this survey were estimated by dividing the net methane emission estimates for each facility by reported gas production numbers (DOGGR 2016), using a methane content for associated gas derived from composition measurements (USGS 2007). For 15 of the top associated gas producing fields in the SJV the team found a mean production-normalized emission rate of 4.2 percent (range 1.5 – 82.9 percent, Table 4). Recall that the plume imagery only detects point sources greater than the detection limit, so these estimates for those fields are likely conservative. Fields with lower production tend to have higher production-normalized emission rates. For the fields in Table 4 the seven with lowest gas production are responsible for 2 percent of the associated gas but 41 percent of the methane point source emissions. Both of these findings are in good agreement with another recent study that modeled oil and gas production emissions across the US using sparse measurements from major oil and gas basins, predicting higher production-normalized emissions from lower producing fields and a production-normalized emissions rate of 4.8 percent for gas production in the SJV that is significantly higher than the 1.5 percent mean rate for the entire US (Omara et al., 2018). 2F 2F2F 2F2F The prevalence of methane plumes varies significantly by oil and gas field. The Poso Creek and Kern Front fields exhibited some of the highest density of methane plumes in the state – both dramatically higher than other nearby oil fields such as Kern River and Round Mountain (Figure 17). There is no apparently correlation with oil and gas produc-tion in this case. All of these fields are relatively low gas producers whereas Kern River produced nearly six times as much oil in October 2016 as each of the other three fields. A recently released whitepaper based on an independent airborne study of the same three oil fields reported a similar spatial distribution of methane sources (Jones et al., 2020). ‡ This refers to associated gas production as reported by DOGGR, not oil production. 26 Figure 15: Typical Methane Plumes in SJV Oil and Gas Fields Commonly observed methane point sources in oil and gas fields include storage tanks, well heads and (potentially) gathering lines. Example shown is in Kern Front. Source: JPL Figure 16: Closeup of a Methane Plume from a Condensate Storage Tank Example of a typical storage tank (from Kern Front oil field). This plume was consistently through at least September 2017 after initial detection from NASA airborne campaigns since 2014, suggesting a mechanism other than normal intermittent pressure relief valve actuation. Source: JPL 27 Figure 17: Variability in Source Density Between Nearby Oil and Gas Fields Poso Creek Kern Front Kern River Round Mountain Significantly higher densities of methane sources (red markers) were observed in Poso Creek and Kern Front oil fields than others in eastern Kern County. Source: JPL 28 Table 4: Production-Normalized Emission Rates for Associated Gas Producing Fields in the SJV facility Field 1 Field 2 Field 3 Field 4 Field 5 Field 6 Field 7 Field 8 Field 10 Field 9 Field 11 Field 12 Field 13 Field 14 Field 15 Total methane Methane Methane production CH4 mole 2015 Net Gas production production emissions normalized fraction of Production (BCF) TgCH4/yr kgCH4/hr kgCH4/hr emission rate associated gas 56.2 0.9142 104360 1569 1.5% 0.86 13.3 0.1975 22543 88 0.4% 0.79 10.3 0.1714 19571 294 1.5% 0.88 8.1 0.1046 11937 485 3.9% 0.68 4.5 0.0698 7968 341 4.1% 0.82 4.5 0.0749 8551 1446 14.5% 0.88 4.0 0.0605 6910 200 2.8% 0.80 2.8 0.0466 5320 570 9.7% 0.88 0.8 0.0126 1442 1302 47.4% 0.88 0.7 0.0113 1294 202 13.5% 0.80 0.36 0.0054 621 578 48.2% 0.79 0.27 0.0040 455 283 38.3% 0.79 0.26 0.0040 461 312 40.4% 0.83 0.15 0.0024 273 450 62.2% 0.86 0.12 0.0019 221 202 47.7% 0.84 totals 191926 8321 4.2% 0.83 totals for fields producing < 1 BCF 4767 3328 41.1% Estimated leakage rates for the largest associated gas producing fields in the southern San Joaquin Valley. The equivalent average methane production is derived from 2016 reported annual production in billion cubic feet or BCF (DOGGR 2016) and CH4 mole fraction measurements from another study (USGS 2007). The 7 lowest producing fields in this table are responsible for 2 percent of associated gas production and 41 percent of methane emissions. Source: JPL Some facilities such as the sour gas production plant and gas injection facility in Elk Hills oil field presented unique combinations of methane plumes associated with compressor operation and flaring stacks (Figure 18). 29 Figure 18: Gas Processing Facility in Elk Hills Gas processing facility in Elk Hills showing methane plumes from two of three large compressors (top) and one or more flaring stacks (bottom). These are all highly intermittent sources. Source: JPL Natural Gas Transmission, Storage and Distribution • The team estimated that 4.9 percent of the point source emissions observed in this study can be attributed to these sectors (1.9 percent for transmission and distribution pipelines, 1.4 percent for storage and 1.6 percent for non-storage gas stations). The Vista-CA data set used for source attribution for this study derived spatial maps of gas transmission lines in California from a combination of NMPS, CEC and EIA data. However, there is no publicly available map of distribution lines in urban areas. To address this issue the team constructed a residential distribution line mask using parts of the California road network overlaid on raster cells classified as being 20-100 percent impervious in urban areas from the National Land Cover Dataset and connected it to the existing NG Pipeline infrastructure using a 10km distance tolerance (Duren et al., 2019). “Natural Gas Stations” in Table 2 include 158 non-storage compressor stations as well as dehydration stations, metering stations, odor stations, pressure limiting stations, regulation stations, storage stations, taps, and valves; 48 percent of these natural gas stations were surveyed. • Approximately 32 percent of transmission and distribution pipelines in the state were surveyed at least once. For the transmission sector a small number of plumes were observed at compressor stations including at least one persistent source at a shutdown 30 stack (suggesting a possible leaking bypass valve). No leaks in transmission pipelines were observed. All of California’s 12 active underground gas storage facilities were surveyed multiple times (Thorpe et al 2019). Some were surveyed extensively, particularly Honor Rancho (Figure 19), Aliso Canyon (Figure 20) and MacDonald Island (Figure 21c, d). Figure 19: Multiple Emission Sources at Honor Rancho Storage Facility Attribution of multiple methane emission sources detected at Honor Rancho gas storage facility. (a) AVIRIS-NG result shows a persistent methane plume for source 1 (yellow box in (a) and (b)). (b) Source 2 and/or source 3 appeared intermittently (red box). (c) Close-up of source 1 with high-resolution satellite imagery (Google Earth). The facility operator confirmed that source 1 is the facility’s emergency shut down stack – likely due to a leaking isolation valve. (d) Close-up of sources 2 and 3 - likely from the rod pack vents for reciprocating compressor units 2 and 4. Source: Thorpe et al, 2019 In 2016, Aliso Canyon was in a standby state and no obvious plumes were present. In August 2017 injection operations resumed at Aliso Canyon and plumes were regularly observed at the compressor station’s shutdown stack in addition to intermittent plumes observed at other infrastructure in the area that are most likely associated with oil production rather than gas storage operations (Figure 20). 31 Honor Rancho presented a persistent methane plume at an emergency shutdown stack, due to a leaking isolation valve (Figure 19a, c). The operator confirmed this mechanism and subsequently mitigated it. An intermittent source was also observed from the compressor units – likely from the rod pack vents for one or two reciprocating compressors when those units were operated (Figure 19b, d). MacDonald Island presented relatively large plumes on multiple occasions associated with venting from shutdown stacks and one compressor (Figure 20 c, d). Plumes were also observed at Gill Ranch (Figure 20a, venting from shutdown stack), Lodi (Figure 20b, likely dehydrator venting, Kirby Hills (Figure 20, venting from shutdown stack) and Wild Goose (Figure 20f, compressor loss). Figure 20: Variety of Emission Sources at Aliso Canyon Characterizing methane emissions from diverse activities in Aliso Canyon. Aliso Canyon field spans roughly 14 km2 and DOGGR records indicate 251 wells (red dots) - 115 of which are linked to the deeper gas storage reservoir (blue circles) and the rest connected to shallower oil-producing formation (combination of active, idle and plugged wells). Subpanels (a) through (d) indicate methane plumes observed with AVIRIS-NG overlaid on true color imagery that correspond to the locations shown in the map of the Aliso Canyon field from a pumpjack (a), tank (b), drill rig (c), and blowdown stack at the storage facility’s compressor station (d). Source: Thorpe et al., 2019 32 Figure 21: Examples of Different Emission Modes at Gas Storage Facilities (a) Gill Ranch venting from shutdown stack, (b) Lodi –likely dehydrator venting, (c) and (d) McDonald Island south and north platforms - venting from shutdown stacks and one compressor, (e) Kirby Hills – venting from shutdown stack, (f) Wild Goose compressor loss. Source: Thorpe et al., 2019 California’s natural gas distribution infrastructure spans several large urban areas. Four leaks in distribution lines were detected (two each in the LA basin and Bakersfield area) and shared the data with the gas company to guide repairs. In each case follow-up AVIRIS-NG flights verified the repairs were successful. 33 Figure 22: Detection of Leak in Low Pressure Gas Distribution Line Example of gas leak detection and repair in Chino Hills. Left: AVIRIS-NG flight pattern, middle: real-time detection software on airplane, right: processed methane plume image and geolocation of source to within 10 meters. Source: JPL, see https://photojournal.jpl.nasa.gov/catalog/PIA22467 34 Figure 22 illustrates the AVIRIS-NG search pattern covering a 60 km2 area in 30 minutes, realtime detection with the onboard software, and determination of the source location to within 10 meters. The gas company was notified and dispatched technicians to the site who promptly confirmed and repaired the leak. Refineries Methane emissions from the Energy Industries sector in California are significantly higher than reported by the CARB inventory (Fig. 14) and appear to be strongly influenced by refineries although this sector is only responsible for 2.4 percent of the estimated total for statewide point source emissions. This was attributed to intermittent strong sources that translate to average emissions that with one exception are significantly higher than reported to the EPA (Fig.23). Strong methane plumes were observed at nearly every refinery sampled in this study. There appears to be a diverse set of sources at refineries, ranging from storage tanks (either venting from relief valves or leaks) to unknown sources (Figure 24). Figure 23: Measured vs Reported Emissions for Refineries and Power Plants 700 Average emissions (kgCH4/hr) 600 500 449 400 325 300 268 190 200 188.9 120 100 102 81 6 20 0 2.8 0.22 25 21.7 14.2 24.2 18.6 7.16 0.04 12.2 75.8 34 0.79 er Po w Po w er pl an t 1 pl Po an t2 w er pl an Re t 3 fin er Re y 1 fin er Re y 2 fin er Re y 3 fin er Re y 4 fin er Re y 5 fin er Re y 6 fin er Re y 7 fin er Re y 8 fin e Re ry 9 fin er y1 0 0.74 46 41 This study EPA 2017 Comparing hourly average emissions derived from annual total emissions in EPA’s Greenhouse Gas Reporting Program (GHGRP) for facilities participating in that program in 2017 (EPA 2018) with persistence adjusted average emission estimates from this study for facilities with at least 6 overflights. With two exceptions the GHGRP emissions are significantly lower than observed by AVIRIS-NG. Source: Duren et al., 2019 35 Figure 24: Examples of Methane Plumes from Refineries in the LA Basin (Left) venting or leaking storage tanks, (Right) methane from combustion sources. Source: JPL Power Plants The team surveyed 238 natural gas-fired power plants in the state – an estimated 55 percent of all such facilities – including an intensive campaign during a heat wave in Los Angeles to assess the potential for additional fugitives during peak demand periods. Plumes at only seven such facilities were observed. While some of the observed emissions are larger than those reported to the EPA (Figure 23), the team concluded that this is not a major methane emitting sector for California – collectively responsible for about 2.1 percent of total emissions from the population of point sources. The research team acknowledges the possibility that a few cogeneration facilities located within the oil and gas fields currently attributed to the oil and gas sector may be appropriately classified under the power generation sector. However, study results would not significantly change were this reclassification made. Landfills To prioritize flight hours for this sector, CARB’s database of landfill methane emissions and the Vista-CA data set were used to identify 436 likely highest emitters, collectively predicted to contribute 90 percent of managed waste disposal emissions in California. The observed point sources at 30 landfills and two composting facilities include some of the largest outliers in the overall source population and collectively are the highest emitting point source sector in California – representing about 43 percent of the total (Figure 14). The high-resolution images suggest that some of the strong methane plumes at these landfills may be associated with gaps in intermediate cover, delays in construction projects and/or leaking gas capture wells – all indicating a significant mitigation opportunity; however follow up study is recommended to confirm this. Figure 25 provides an example of a reduction in the number and size of methane point source plumes over time at a large landfill in Southern California due to mitigation efforts from the operator that were in part informed by the data from this study (Cusworth et al., 2020). The varying degrees of agreement and disagreement between our measurements and bottomup accounting for the landfills illustrated in Figure 26 is representative of the total population 36 of landfills that exhibit point sources. The fact that the team did not detect a larger population of smaller methane point sources across the landfill sector suggests the majority of those facilities emit methane as area sources and/or point sources below the detection limit. They concluded that the landfills exhibiting point sources in California are a unique sub-population. Figure 25: Time Series of Landfill Point Source Emissions Example of tracking trends in methane point source emissions over time at a large landfill in southern California. Left: September 2016, Center: October 2017, Right: October 2018. Source: JPL Figure 26: Measured vs Reported Emission for Representative Landfills 6000 5000 Emissions (kgCH4/hr) 4000 3000 2000 1660 1000 778 769 618 394 211 0 Landfill 1 Landfill 2 Landfill 3 Landfill 4 Scientific Aviation AVIRIS-NG Landfill 5 Landfill 6 EPA Comparing landfill emissions reported to the EPA for 2017 (EPA 2018) with persistence adjusted average emission estimates from this study and mean values from a series of coordinated Scientific Aviation flights (CARB 2018b) – the last 4 of which were not contemporaneous with AVIRIS-NG flights. Since Scientific Aviation measures the net facility emissions (area + point sources) and AVIRIS-NG only measures point sources, the latter will be lower than the former in many cases Source: Duren et al., 2019 37 Wastewater Treatment We estimate that this sector is responsible for about 2.6 percent of total methane point source emissions in California. A total of 57 domestic wastewater treatment facilities were surveyed. Only 12 exhibited methane point source plumes - three of which were persistent, suggesting potentially anomalous activity (see Figure 26 for examples). While the team did not explicitly include industrial wastewater treatment and discharge in the Vista-CA data set or flight planning, a large beef processing facility with methane plumes emanating from on-site pits was detected. After confirming the latter were not associated with dairies or other nearby infrastructure (using satellite imagery) this single facility was allocated to emission sector 4D2 “Industrial Wastewater > Production processed - Red meat”. Figure 27: Emissions from Wastewater Treatment Plants Examples of persistent methane point sources at a small number of wastewater treatment facilities. Left: Hyperion treatment plant, Right: Santa Clara/San Jose plant. Source: JPL Dairies and Livestock The team surveyed 443 Confined Animal Feeding Operations (CAFOs) in California – an estimated 71 percent of all such facilities in the State. CAFO manure management – particularly at large dairies in the San Joaquin Valley (SJV) – is recognized as one of the top methane emission sectors in California. The survey results are consistent with this given that wet manure management – particularly settling ponds and anaerobic lagoons – is responsible for about 26 percent of methane total methane emission from point sources in California. A robust assessment of the individual and net emissions from dairies and other livestock facilities in California is complicated by several factors. Figure 28 indicates one such factor: the complex spatial gradients of near-surface atmospheric methane that manifests in portions of the SJV in response to the dense concentration of emission sources (large dairies) and/or the effects of “pooling” from wind and other meteorological variables. This figure also raises the question: why weren’t methane point sources detected at more dairies? Detecting and attributing methane plumes to individual point sources can be challenging in the presence of strong methane enhancements over large areas – essentially a “contrast” problem. In such areas there is a risk both of over-estimating the emissions of individual dairies (by convolving the 38 flux with nearby facilities) and also under-estimate the net emissions of the area. This represents an active area of measurement science and is a priority for future attention. Figure 28: Mosaic of Two Days of AVIRIS-NG Flights Over Tulare Area Dairies 10 km The raw grayscale image overlays represent areas with lower (black) and higher (white) levels of atmospheric methane. The striking gradient seen here suggests accumulation of enhanced levels of methane in these areas due to the combination of many strong emitters and low wind conditions. Atmospheric transport modeling will likely be required to disentangle those two effects. Blue squares indicate the known locations of dairies from the Vista-CA data set. Red markers indicate methane point sources detected during these overflights. Source: JPL Another complexity involves the inherent variability of dairy methane emission processes. The primary driver for methane point source emissions from manure management involves the use of water and anaerobic conditions that promote methanogenesis. Dairies are dynamic facilities in that water and wastes are moved around each facility over the course of the day on a given duty-cycle, translating to methane point sources that can vary significantly on time-scales of hours – as anaerobic layers in lagoons are disturbed and as methane laden water is transported around the facility including irrigation for adjacent fields. 39 Figure 29: Methane Point Source Variability for Dairies Near Tipton AVIRIS-NG repeatedly flew the same flight lines over an area with about 100 large dairies near Tipton with a roughly 45-minute revisit interval per facility over a 5-hour period. The colors indicate the number of times a source was observed during that period. Some of the sources (red circles) were persistent – others were highly variable. This variability is common in many emission sectors and illustrates the need for frequent sampling. Source: JPL This diurnal, management-driven variability is likely somewhat independent of seasonal variability in emission fluxes driven by changing temperatures. This short-term variability can have an impact on detectability as illustrated in Figure 29 and 30 (e.g., surveys with insufficient revisit frequency can fail to detect sources through aliasing). 40 Figure 30: Close-Up of a Dairy from the Intensive Study 103.1 kg/hr 69.3 kg/hr 51.6 kg/hr 137.6 kg/hr 59.8 kg/hr Each image shows methane plumes for snapshot in time, each separated by about 45 minutes - indicating significant temporal variability in emissions. In this case the variability is attributed to water flushing manure from feedlots on the right side of each image through settling ponds in the middle. Source: JPL Methane digesters are increasingly being deployed at California dairies in an effort to reduce the net greenhouse gas impact of each facility while offering additional revenue opportunities such as biogas for energy production. The survey covered about 25 known dairy digesters in the state including a combination of facilities in operation and still undergoing construction. In principle a well-functioning digester should capture methane from manure management however the study indicated the presence of significant methane point sources at four facilities in the SJV. Figure 31 shows an example of a persistent methane plumes at a dairy digester. The biogas operator for this facility indicated that the cause was likely manual venting during maintenance activity. This suggests that future monitoring for atmospheric methane around these facilities before and after digester construction could prove useful for assessing their efficacy in meeting mitigation objectives while helping operators avoid unintentional biogas product loss. 41 Figure 31: Methane Plume Observed Persistently at Dairy Methane Digester The difference in plume appearance between the two dates is attributed primarily to different wind speeds. Source: JPL 42 CHAPTER 4: Knowledge Transfer Much of the new data analysis system used in this study was developed under parallel NASA programs. All data on methane plumes detected in this study (following quality control filtering), as well as the new GIS methane data set (Vista-CA), are now available through a prototype Methane Source Finder web-based methane data portal (https://methane.jpl.nasa.gov). That portal and those data sets are designed to improve their accessibility and relevance to a diverse set of stakeholders. The team is exploring options to transfer sustained operation of the methane data portal to the California Air Resources Board in the future. Additionally, during this study, the project team organized multiple meetings and briefings to share and discuss interim findings with stakeholders, including staff from the CEC, CARB, CalRecycle, Bay Area Air Quality Management District, South Coast Air Quality Management District, Southern California Gas company, Pacific Gas and Electric company, Milk Producer’s Council, City of Los Angeles Department of Sanitation, Sunshine Canyon Landfill Local Enforcement Agency, and several operators of individual facilities. These interactions resulted in two-way knowledge transfer, including feedback on the utility of the methane data sets as well as ground truthing and explanations about potential causes for the observed emissions. Another source of knowledge transfer was the publication of a paper in the journal Nature (Duren et al., 2019). 43 CHAPTER 5: Recommendations The team found that methane point sources across all sectors are significant contributors to California’s methane budget. The prevalence of methane super-emitter activity observed in key sectors also suggests significant mitigation potential for California – particularly for landfills, dairies, and oil and gas production. With these lessons in mind, the team makes the following recommendations for future attention by the State of California and other stakeholders. • Detecting, quantifying, and attributing point source emissions to specific infrastructure elements on an ongoing basis can improve the scientific understanding of regional methane budgets and inform policy and planning activities that reduce methane emissions. • The highly intermittent and stochastic nature of many point sources underscores the need for persistent, wide area monitoring systems (Duren et al., 2019). The most effective approach will likely involve a tiered observing system with components for detecting strong point source emissions such as those described here, as well as other components optimized for monitoring area emissions below the detection limit of AVIRIS-NG (Duren et al., 2012; Cusworth et al., 2020b). The combination of methods that collectively provide high spatial resolution and high frequency sampling over large areas could also help disentangle the relative contributions of point and area sources to regional methane budgets. • Future developments in high performance imaging spectroscopy have the potential to address area sources (e.g., from enteric fermentation, rice cultivation, and wetlands) and reduce the detection limits for point sources by a factor of 10 or more. In particular, improving the spectral resolution of an AVIRIS-NG class imaging spectrometer from the current 5 nm to 1 nm in the 2100-2400 nm range, without sacrificing other aspects of instrument performance, would enable dramatic advances (Thorpe et al., 2016b). • One of the largest sources of uncertainty in emission estimates for individual methane point sources is the general coarse space-time resolution of near surface wind speed data at most locations. Improvements in wind speed data, either through simultaneous remote sensing or denser surface observations and/or higher resolution weather reanalysis products, could reduce emission uncertainties by 20-50 percent or more. • This project has demonstrated the ability of regional scale monitoring systems to detect the footprint of large anomalous methane emissions and of airborne imaging spectrometers to find and pinpoint leaks in natural gas infrastructure. However, the data analysis is complex and time-consuming. Future improvements in measurement and analysis frameworks could support operational, rapid-response versions of such systems, which would be particularly valuable for hazardous leak detection. • The mitigation examples in this study were primarily limited to natural gas transmission, storage, and distribution infrastructure and a single landfill. Opportunities abound to further evaluate and apply these methods in other key sectors, particularly anaerobic 44 digesters at dairies and wastewater treatment plants, as well as key infrastructure within oil and gas fields. Establishing pilot projects to further facilitate data sharing and collaborative mitigation could provide scientists and technologists with the feedback necessary to make these systems more relevant and effective (Hopkins et al., 2016). • A concerted effort to diagnose and mitigate some of the point source emissions observed at 32 landfills and composting facilities could provide a significant advance towards meeting the State’s target of reducing methane emissions. It could also help address related priorities, including air-quality, environmental justice, and the potential economic potential of biogas as an alternative energy source. 45 CHAPTER 6: Benefits to Ratepayers This project has provided new insights into California’s methane budget. Specifically, the team conducted the first systematic assessment of the relative contributions of methane point sources, including their distribution by space, time, and emission sector. These findings may lead to improvements in California’s greenhouse gas inventory and to efforts by state and local agencies and businesses to both prioritize future investments in methane emissions mitigation and to assess overall progress towards emission reduction goals. This work also highlights the potential for efficient point source monitoring techniques to directly enable mitigation of a broad class of methane super-emitters, representing a significant contribution to California’s climate stabilization targets. The researchers estimated 0.158 (95 percent confidence 0.135-0.184) TgCH4/yr in point source emissions for the oil and gas sector in California (Duren et al., 2019). If translated to natural gas equivalent, this represents approximately $28-$39 million in annual product loss at July 2018 US city-gate gas prices. This indicates the potential value of mitigation for California ratepayers, additional to climate benefits. 46 GLOSSARY AND LIST OF ACRONYMS Definition Term ACCESS NASA Advancing Collaborative Connections for Earth System Science program AVIRIS-NG Next Generation Airborne Visible/Infrared Imaging Spectrometer CARB California Air Resources Board CEC California Energy Commission CH4 Methane CO Carbon monoxide CO2 Carbon dioxide EPA United States Environmental Protection Agency Enhance Enhancement GIS Geographic Information System H2O Water vapor HITRAN High-resolution transmission molecular absorption database HRRR High resolution rapid refresh reanalysis IME Integrated methane enhancement IPCC Intergovernmental Panel on Climate Change JPL Jet Propulsion Laboratory N2O Nitrous oxide NCEP National Centers for Environmental Prediction O2 Oxygen Ppm-m Parts per million meter. Representing the thickness and concentration within a volume of equivalent absorption that is equivalent to an excess methane concentration in ppm if the layer is one meter thick. SF San Francisco SJV San Joaquin Valley SoCAB South Coast Air Basin TBD To Be Determined UTC Universal Time Coordinated XCH4 Total column averaged methane excess mixing ratio 47 REFERENCES Alvarez, R., D. Zavala-Araiza, D.R. Lyon, D.T. Allen, Z.R. Barkley, A.R.Brandt, K.J. Davis, S.C. Herndon, D.J. Jacob, A. Karion, Anna , E. Kort, B.K. Lamb, T. Lauvaux, J.D. Maasakkers, A.J. Marchese, M. Omara, S.W. Pacala, J. Peischl, A.L. Robinson, P.B. Shepson C. Sweeney, A. Townsend-Small, S.C. Wofsy, S.P Hamburg, Science 21 (2018). Brandt, A., G. A. Heath, E. A. Kort, F. O'Sullivan, G. Pétron, S. M. Jordaan, P. Tans, J. Wilcox, A. M. Gopstein , D. Arent , S. Wofsy, N. J. Brown, R. Bradley, G. D. Stucky, D. Eardley, R. Harriss, Methane Leaks from North American Natural Gas Systems, Science 343, 6172, 733-735 (2015). Butz, A. (2012), TROPOMI aboard Sentinel-5 Precursor: Prospective performance of methane retrievals for aerosol and cirrus loaded atmospheres, Rem. Sens. Environ. 120 (2012) 267–276 , doi: 10.1016/j.rse.2011.05.030 California (2014) Senate Bill SB 605 Short-Lived Climate Pollutant Strategy. California (2015) Assembly Bill AB 1496 Methane Mitigation Act. California (2017) Short Lived Climate Pollutant Strategy (2017), https://www.arb.ca.gov/cc/shortlived/meetings/03142017/final_slcp_report.pdf CARB 2018a, California Greenhouse Gas Emission Inventory - 2018 Edition, California Air Resources Board, https://www.arb.ca.gov/cc/inventory/data/data.htm, 2018. CARB 2018b, CARB AB1496 research program, Airborne Facility-Level Methane Emissions Study, retrieved October 1, 2018, https://ww2.arb.ca.gov/ourwork/programs/methane/ab1496-research Carranza, V., Rafiq, T., Frausto-Vicencio, I., Hopkins, F., Verhulst, K. R., Rao, P., Duren, R., and Miller, C. (2017): Vista-LA: Mapping methane emitting infrastructure in the Los Angeles megacity, Earth Syst. Sci. Data Discuss., https://doi.org/10.5194/essd-2017-65 , in review. Conley, S., et al. (2016), Methane emissions from the 2015 Aliso Canyon blowout in Los Angeles, CA, Science. Cui, Y., et al. (2019), A Multiplatform Inversion Estimation of Statewide and Regional Methane Emissions in California during 2014–2016, Environ. Sci. Technol. Cusworth, D.H., Duren, R.M., Thorpe, A.K., Tseng, E., Thompson, D.R., Guha, A., Newman, S., Foster, K., Miller, C.E. (2020). Using remote sensing to detect, validate, and quantify methane emissions from California solid waste operations. Env. Res. Let. 15 Cusworth, D. H., Duren, R. M., Yadav, V., Thorpe, A. K., Verhulst, K., Sander, S., et al. (2020b). Synthesis of methane observations across scales: Strategies for deploying a multitiered observing network. Geophysical Research Letters, 47, e2020GL087869. https://doi.org/10.1029/2020GL087869 48 DOGGR (2016). Annual Report of California Oil and Gas Production Statistics, Department of Oil Gas and Geothermal Resources, http://www.conservation.ca.gov/dog/pubs_stats/annual_reports/Pages/annual_reports. asp Duren, R., A. Thorpe, K.T. Foster, T. Rafiq, F. M. Hopkins, V. Yadav, B.Bue, D.R. Thompson, S. Conley, N. Colombi, C. Frankenberg, I.McCubbin, M.Eastwood, M.Falk, J. Herner, B. E. Croes, R. Green, C. Miller, California’s methane super-emitters, Nature, 575 (180-184) (2019). Doi 10.1038/s41586-019-1720-3 EPA Greenhouse Gas Reporting Program, https://www.epa.gov/ghgreporting/ghg-reportingprogram-data-sets, retrieved October 20, 2018. Ellis, J.L.,A. Bannink, J. France, E. Kebreab, J. Dijkstra, Evaluation of enteric methane prediction equations for dairy cows used in whole farm models, Global Change Biology 16, 3246–3256, (2010). Fitzgerald, G.J., K.M.Scow, J.E.Hill, Fallow season straw and water management effects on methane emissions in California Rice, Glob. Biogeo. Cyc. 14, 767-776, 2000. Frankenberg, C., Thorpe, A.K., Thompson, D.R., Hulley, G., Kort, E.A., Vance, N., Borchardt, J., Krings, T., Gerilowski, K., Sweeney, C. and Conley, S., 2016. Airborne methane remote measurements reveal heavy-tail flux distribution in Four Corners region. Proceedings of the National Academy of Sciences, p.201605617. Hopkins, F.M., J.R. Ehleringer, S.E. Bush, R.M. Duren, C.E. Miller, C.T. Lai, Y.-K. Hsu, V. Carranza, J.T. Randerson (2016). Mitigation of methane emissions in cities: how new measurements and partnerships can contribute to emissions reduction strategies, Earth’s Future, doi: 10.1002/2016EF000381 Jeong, S., Y. Hsu, A. E. Andrews, L. Bianco, P. Vaca, J. M. Wilczak, M. L. Fischer, A multitower measurement network estimate of California’s methane emissions, J. Geophys. Res. Atmos 118, 1–13, (2013) Jeong, S., D. Millstein, M. L. Fischer, Spatially Explicit Methane Emissions from Petroleum Production and the Natural Gas System in California, Environ. Sci. Tech., 48 (10), 59825990 (2014). Jeong, S., et al. (2017), Estimating methane emissions from biological and fossil-fuel sources in the San Francisco Bay Area, Geophys. Res. Lett., 44, 486–495. Doi 10.1002/2016GL071794 Jones,K.L., Tratt.D.M. (2020), Mapping Methane Super-Emitters in Oil and Gas Fields: A Tiered Remote Sensing Strategy, Center for Space Policy and Strategy, Aerospace Corporation. Kirschke, S., et al. (2013) Three decades of global methane sources and sinks. Nat. Geosci. 6, 813-823. 49 Kort EA, et al. (2014) Four corners: The largest US methane anomaly viewed from space. Geophys. Res. Lett .41(19):6898–6903. Krautwurst, S., Gerilowski, K., Jonsson, H.H., Thompson, D.R., Kolyer, R.W., Thorpe, A.K., Horstjann, M., Eastwood, M., Leifer, I., Vigil, S. and Krings, T. (2017). Methane emissions from a Californian landfill, determined from airborne remote sensing and insitu measurements. Atmospheric Measurement Techniques Discussions . Omara, M., N. Zimmerman, M.R. Sullivan, X. Li, A.Ellis, R. Cesa, R. Subramanian, A. A. Presto, and A. L. Robinson, Methane Emissions from Natural Gas Production Sites in the United States: Data Synthesis and National Estimate, Environ. Sci. Tech. 52 (21), 12915– 12925 (2018). Ryerson, T. B., et al. (2013), The 2010 California Research at the Nexus of Air Quality and Climate Change (CalNex) field study, J. Geophys. Res. Atmos., 118, 5830–5866, doi:10.1002/jgrd.50331 Thompson, D.R., Leifer, I., Bovensmann, H., Eastwood, M., Fladeland, M., Frankenberg, C., Gerilowski, K., Green, R.O., Kratwurst, S., Krings, T. and Luna, B., 2015. Real-time remote detection and measurement for airborne imaging spectroscopy: a case study with methane. Atmospheric Measurement Techniques, 8(10), pp.4383-4397. Thompson, D.R., Thorpe, A.K., Frankenberg, C., Green, R.O., Duren, R., Guanter, L., Hollstein, A., Middleton, E., Ong, L. and Ungar, S., 2016. Space‐based remote imaging spectroscopy of the Aliso Canyon methane superemitter. Geophys. Res. Lett., 43(12), pp.6571-6578. Thorpe, A. K., Frankenberg, C., Aubrey, A. D., Roberts, D. A., Nottrott, A. A., Rahn, T. A., Sauer, J. A., Dubey, M. K., Costigan, K. R., Arata, C., Steffke, A. M., Hills, S., Haselwimmer, C., Charlesworth, D., Funk, C. C., Green, R. O., Lundeen, S. R., Boardman, J. W., Eastwood, M. L., Sarture, C. M., Nolte, S. H., Mccubbin, I. B., Thompson, D. R., McFadden, J. P. (2016). Mapping methane concentrations from a controlled release experiment using the next generation Airborne Visible/Infrared Imaging Spectrometer (AVIRIS-NG). Remote Sensing of Environment. Thorpe, A. K., Frankenberg, C., Green, R. O., Thompson, D. R., Aubrey, A. D., Mouroulis, P., Eastwood, M. L., and Matheou, G. (2016b), The Airborne Methane Plume Spectrometer (AMPS): Quantitative imaging of methane plumes in real time, 2016 IEEE Aerospace Conference, doi:10.1109/AERO.2016.7500756. Thorpe, A.K., Duren, R., Conley, S., Prasad, K., Bue, B., Yadav, V., Foster, K., Rafiq, T., Hopkins, F., Smith, M. and Fischer, M.L., Thompson, D.R., Frankenberg, C., McCubbin, I.B., Eastwood, M., Green, R.O., Miller, C.E., (2020)., Methane emissions from underground gas storage in California, Environ. Res. Lett. 15 045005 Turner, A. J., et al (2015), Estimating global and North American methane emissions with high spatial resolution using GOSAT satellite data, Atmos. Chem. Phys., 15, 7049-7069, doi:10.5194/acp-15-7049-2015. 50 Turner, A. J., D. J. Jacob, J. Benmergui, S. C. Wofsy, J. D. Maasakkers, A. Butz, O. Hasekamp, and S. C. Biraud (2016), A large increase in U.S. methane emissions over the past decade inferred from satellite data and surface observations , Geophys. Res. Lett., 43, doi:10.1002/2016GL067987. USGS 2012, U.S. Geological Survey Professional Paper 1713, Petroleum Systems and Geologic Assessment of Oil and Gas in the San Joaquin Basin Province, California; 2007. Wecht, K.J., Jacob, D.J., Sulprizio, M.P., Santoni, G.W., Wofsy, S.C., Parker, R., Bosch, H., and Worden, J. (2014) Spatially resolving methane emissions in California: constraints from the CalNex aircraft campaign and from present (GOSAT, TES) and future (TROPOMI, geostationary) satellite observations. Atmos. Chem. Phys. Discuss. 14, 4119-4148. Wennberg et al (2012), On the Sources of Methane to the Los Angeles Atmosphere, Env. Sci. Tech. Wong, C. K., Pongetti, T. J., Oda, T., Rao, P., Gurney, K. R., Newman, S., Duren, R. M., Miller, C. E., Yung, Y. L., and Sander, S. P.: Monthly trends of methane emissions in Los Angeles from 2011 to 2015 inferred by CLARS-FTS observations, Atmos. Chem. Phys. 16, 13121-13130 (2016). Wunch, D., Toon, G. C., Hedelius, J. K., Vizenor, N., Roehl, C. M., Saad, K. M., Blavier, J.-F. L., Blake, D. R., and Wennberg, P. O.(2016): Quantifying the loss of processed natural gas within California's South Coast Air Basin using long-term measurements of ethane and methane, Atmos. Chem. Phys., 16, 14091-14105, https://doi.org/10.5194/acp-1614091-2016, 2016. Zavala-Araiza, D., D.R. Lyon, R. A. Alvarez, K.J. Davis, R. Harriss, S.C. Herndon, A. Karion, E.A. Kort, B.K. Lamb, X. Lan, A J. Marchese, S.W. Pacala,, A. L. Robinson, P.B. Shepson, C.Sweeney, R.Talbot, A.Townsend-Small, T. I. Yacovitch, D. J. Zimmerle, S.P. Hamburg, Reconciling divergent estimates of oil and gas methane emissions, Proc. Nat’l. Acad. Sci. 112 (51) 15597-15602 (2015). 51 APPENDIX A: Data Availability Methane plumes images, Vista-CA layers, and regional scale methane emission products for California can be viewed at the Methane Source Finder prototype data portal at https://methane.jpl.nasa.gov/ AVIRIS-NG calibrated radiance and reflectance products can be ordered from the AVIRIS-NG data portal https://avirisng.jpl.nasa.gov/alt_locator/ Retrieved methane images from flight lines in this study are available for download at https://doi.org/10.3334/ORNLDAAC/1727 Vista-CA infrastructure spatial layers are available for download at https://doi.org/10.3334/ORNLDAAC/1726 An electronic archive of the following products has also been delivered to CARB and CEC. 1. Georeferenced image files (GeoTIFF format) for the 1349 methane plumes detected during this study that passed quality control checks 2. Plume list: source ID, latitude, longitude, detection date, detection time (UTC), source type, IPCC sector, IME/r (kg m-1), IME/r (kg m-1), U10 (m s-1), U10(m s-1), Qplume(kg h-1), Q (kg h-1) [last two fields intentionally blank for those plumes lacking emission estimates due to quality control and filtering] 3. Sources list: source ID, latitude, longitude, source type, IPCC sector, number of overflights, persistence, confidence in persistence estimate, persistence adjusted average source emissions Qsource(kg h-1), Q (kg h-1) where IME = integrated methane enhancement (total mass of methane in plume) r = plume length IME/r = uncertainty (standard deviation) in IME/r estimate U10 = total (vector) wind speed at 10 meters above ground level U10 = uncertainty in wind speed at 10meters above ground level Qplume = instantaneous plume emission rate (single observation) Qsource= average, persistence adjusted source emission rate Q = total uncertainty in emission rate A-1