Preliminary measurements from the natural gas system in

Preliminary measurements from the natural gas system in California: from well to downstream of the meters June 1 st, 2015 Marc L. Fischer (mlfischer@lbl. gov) • Motivation and Overview – NG methane contribution as a CA climate pollutant – Bottom-up estimates of natural gas (NG) methane emissions • CALGEM-NG – – Regional Top-down Study of SF Bay Area Methane Airborne Facility Scale Measurements Localized Mobile Plume Integration (MPI) Emissions from Residential Buildings 1

Team Acknowledgements LBNL: Seongeun Jeong, Toby Walpert This work was supported by the California Energy Commission, Natural Gas Environmental Research Program UC Davis: Stephen Conley, Ian Faloona UCI: Tianyang Zhu, Don Blake SJSU: Matt Llyod, Neil Larau, Craig Clements Picarrro: Eric Crosson, Chris Rella BAAQMD: David Fairley, Phil Martien, Saffet Tanrikulu CEC: Guido Franco, Simone Brant PG&E: François Rongere, Gerry Bong UC Berkeley: Allen Goldstein, Abhinav Guha NOAA-CCG: Arlyn Andrews , Laura Bianco, Ed Dlugokencky, Jim Wilczak, Steve Montzka, Ben Miller, Pieter Tans CIT: Sally Newman, Debra Wunch, Paul Wennberg UCR: Jingsong Zhang, Mixtli Campos NASA Ames: Laura Iraci, Emma Yates, Matt Johnson JPL: Riley Duren, Christian Frankenberg, Charles Miller UCSD: Ray F. Weiss, Ralph Keeling, Peter Salameh Earth. Networks: Christopher D. Sloop, Bob Marshall CARB: Ying-Kuang Hsu, Abhilash Vijayan, Jorn Herner, Bart Croes, 2 Fischer, Jim Vernon Hughes, Marc Vayssières, Richard Bode, Joseph Narady, and Webster Tassat, Mac Mc. Dougall, Ken Stroud

Problem Overview • Natural gas provides 43% California’s fossil fuel energy • Methane is a potent short lived climate pollutant – A 3% well-to-burner methane leak approximately equals climate forcing from remaining 97% gas combusted to CO 2 on 20 year timescale – Pre-meter distribution and post-meter consumption leakage paths deserve attention • CA and US now moving to control CH 4 emissions – Current inventories typically lower than measurements in facility or regional studies – Likely ~ 10% California’s methane emissions from NG • CEC survey project provides new look across CA natural gas infrastructure 3

Overview of Natural Gas System • Many sub-sectors may contribute to total NG emissions • Post-meter emissions previously neglected • More complete representation captures emissions per energy delivered Courtesy G. Franco and S. Ziaja adapted from US-EPA 4

Bottom-up Natural Gas Inventories Vary • Combine US-EPA emission factors and GIS activity data § Production: Conventional and enhanced recovery wells § Transmission, compression, and storage § Distribution 0. 3% of regional consumption (1 st guess) § Prior CALGEM-NG emissions (Jeong et al. , 2014) still lower in So. CAB than recent top-down (Peischl et al. , 2013) CALGEM-NG Methane Flux (nmol m-2 s-1) § Resulting CALGEM-NG emissions likely 200 -400 Gg CH 4 yr-1 § Livestock and landfills likely much larger in CA so NG 10 -20% of total CA CH 4 5

CALGEM-NG CH 4 Collaboration • Regional Emissions – Tower Collaborations • Facilities Leakage – Aircraft Observations • Localized Plumes – Mobile Plume Integration – Building Studies 0. 1 °× 0. 1 ° LBL & Picarro Plume UCD Airborne UCI VOC SJSU Lidar LBL Residential 6

Estimating Regional NG Emissions • NG: Total CH 4 emission ratio only large in urban/production areas 0. 1 °× 0. 1 ° • Evaluate emissions in SF Bay Area • Light alkane composition (ethane: methane) key for attribution Total NG Methane NG CH 4 Total CH 4 nmol/m 2/s NG: Total Ratio nmol/m 2/s San Joaquin Valley So. CAB 0. 1 °× 0. 1 ° 7 Regions

Napa 450 Fremont 400 San Jose 350 San Rafael Pittsburg 300 Vallejo 250 Richmond 200 Concord 150 Redwood City 100 Santa Rosa 50 Livermore 0 2 09 -1 7 20 05 -0 20 04 20 20 02 - 20 01 8 99 - 19 96 -9 5 19 93 -9 19 1 San Francisco 89 -9 19 • Collaboration with BAAQMD (Fairley et al. , 2014, ES&T) • Estimate CH 4 emissions CH 4: CO correlations + CO emission inventory CH 4 emissions (Gg. CH 4/year) Bay Area Methane 500 – Total CH 4 1. 5 – 2. 0 x BAAQMD Inventory – AQ focused sites likely biased toward CO emissions – Likely NG significant urban contribution • Future: include VOC tracers to quantify NG emissions from distribution 8 Inventory ch 4*

Airborne Studies of Point Sources Validated with Controlled Release Tests Methane Plume Controlled PG&E methane and ethane release in natural gas § Release rate ~ 15 kg CH 4 hr-1 § Flight data estimates recover both Wind methane and ethane to within 20% § Recover ethane: methane ratio to better than 10% Wind * note 0. 1 Gg CH 4 yr-1 = 11. 4 kg CH 4 hr-1 Courtesy Steven Conley

Bay Area NG Storage and Petroleum Refining § Bottom-up estimates for average storage emissions 80 kg CH 4 hr-1 (US - EPA) § Airborne measurements at four storage facilities (June, 2014 – May, 2015) § Two sites non-detection, one site small ~ 11 Kg CH 4 hr-1 § Five flight days at 4 th site larger and variable 80 - 300 kg CH 4 hr-1 § C 2 H 6: CH 4 matches PG&E ( ~ 5% vol) § Initial measurements at three refineries (Feb-May, 2015) § Large variation in emissions 30 - 250 kg CH 4 hr-1 § and ethane: methane ratios 6 -20% 10

San Joaquin Valley Production § Flights April -June, 2014 § Clear downwind enhancements of CH 4 and ethane on both days § Emissions from example production field § Flight data 14 +/- 5 Gg CH 4 yr-1 § CALGEM bottom-up 10 Gg CH 4 yr 1 § Another larger field shows episodic emissions, apparently correlated with well completion 11

Local Measurements: LBNL Mobile Plume Integration (MPI*) System Gas System CH Plume at 1, 2, 4 m agl 4 • Cross-wind integral of CH 4 enhancement flux quantifies plume emissions 4 m 2 m 1 m – Sample inlets can be set to 4 -8 m above ground – Anemometry of wind velocity • Recent system developments – Tests at LBNL and PG&E facilities – Better than 30% accuracy with 3 passes in most conditions – Multi-analyzer system w/ 13 CH 4 allows NG attribution for strong plumes Wind direction * Patent Pending 12

Capped Gas Wells in Sacramento Delta • Plan drive using CA Dept. Cons. well map data • 1 day drive-by of 12 capped wells MPI System – Quantify one plume 130 +/- 40 sccm (5 passes) Methane Plume – Detected 3 plumes 40 -350 sccm (one pass each) – Non-detect for 2 sites – 7 sites not downwind of public road CH 4 Enhancement in vertical plane Capped well 13

Bakersfield Distribution System § Survey 80 km of Bakersfield public streets § Detect 20 large leaks above elevated varying background § 40% emissions found within 0. 5 km of large distribution pipes § Plume integrations yield total emissions of 6. 4 kg CH 4 hr-1 § Scaling by area suggests total emissions ~ 90 kg CH 4 hr-1 § Comparing with consumption suggests ~ 0. 3% distribution leakage § Consistent with CALGEM-NG distribution estimate CH 4 enhancements (green), distribution (orange) and transmission (blue) pipelines 14

Residential Leakage § Measurements § Depressurize house producing controlled inflow of outdoor air § Measure CH 4 enhancement relative to outdoor air § 13 CH /12 CH used to identify gas vs. 4 4 biological methane § Results from 10 SF Bay homes § Average leak rate 7 +/- 2 sccm equal ~ 0. 16% of consumption § Indoor CH 4 enhancements show NG 13 CH signature 4 § New CEC project underway to measure 50 -75 homes across CA housing stock Outdoor CH 4 C 0 Air flow in Q Q Ci Ci E leak C 0 Q + E = C i Q E = Q (Ci - C 0) Measured indoor (white) and outdo (grey) methane during calibrated indoor leak (red) 15

Combustion Appliance Leakage § Methods § Mass balance gathering total exhaust stream as with building level test § Ratio of CH 4: CO 2 enhancements + fuel use yields emissions § Examples: tank-less water heaters, clothes driers, gas range § Three tank-less water heaters emitted 80 - 300 sccm CH 4 § 1 hr operation ~ equal 1 day of quiescent house leakage § Two clothes driers emitted ~ 10 sccm emissions in continuous operation § One gas range emitted ~ 5 sccm in continuous operation 16

Summary and Next Steps § Work sponsored by CEC identifying key components of natural gas CH 4 emissions from CA § Emissions from production and distribution sectors uncertain and likely underestimated in state current inventories § Production emissions episodic -> continuous observation § Distribution emissions diffuse -> NG tracers § Atmospheric measurements can quantify emissions reductions at multiple scales § Need to identify critical gaps in mitigation activities § Energy systems, agriculture, waste management 17