introduction to the phase i report of the california ... · proposal. an additional action recently...

of 6

Introduction to the Phase I Report of the California Methane Survey from the Staff of the California Air Resources Board (CARB)

October 2, 2017

The State of California has passed two important pieces of legislation regarding the understanding and control of sources of methane. Assembly Bill 1496 (Thurmond, Chapter 604, Statutes of 2015) requires the State to monitor methane hotspots, while Senate Bill 1383 (Lara, Chapter 395, Statutes of 2016) sets a goal of reducing methane emissions by 40% from 2013 levels by 2030. To inform policies and regulations to achieve these goals, and to improve the understanding of methane emissions, the State of California has put in place a comprehensive measurement and research program.1 This Phase 1 report provides initial results from one of those efforts, called the California Methane Survey, in which CARB, the California Energy Commission, and NASA’s Jet Propulsion Laboratory (JPL) collaborated to use remote sensing to survey key infrastructure in California and identify large methane point sources. In many cases, the State has already begun working to reduce the emissions this Phase I report describes, and the research helps inform the efficacy of those efforts as well as helps identify additional reduction strategies.

The work has been accomplished by using a NASA JPL airborne infrared imaging spectrometer that can ‘see’ and quantify methane plumes as it is flown over large areas with infrastructure types known to be responsible for methane emissions, such as dairies, landfills, oil and gas wells, natural gas reservoirs and pipelines, and refineries. This method is able to identify large point sources, i.e., those that release concentrated emissions and provide a significant methane contrast to background levels. The method is not able to identify area sources, i.e., those emissions that occur over a wider area (such as enteric fermentation from dairy cows and beef cattle) or those sources that emit less than about 10 kilograms of methane per hour. While some sources were visited multiple times, the results in this report generally represent a snapshot of emissions. In addition, since Phase I of the survey did not include resources to conduct root-cause attribution for every source, the identification of a source in this report does not necessarily mean that there was an unintended leak or malfunction. The identified source could also be part of normal operational emissions. Additional analysis on sources will be conducted during Phase II of the study and as part of CARBs ongoing efforts.

In Phase II, additional analysis on the results from this Phase I of the project will include calculating the emissions rate for each of the identified plumes and additional airborne surveys of some of the priority sources and areas. These analysis will provide insights

1 More information on the CARB Methane Research Program: https://www.arb.ca.gov/research/methane.htm

https://www.arb.ca.gov/research/methane.htm

of 6

on whether the identified plumes are intermittent, whether they represent normal emissions or leaks, whether they are well represented in emission inventories, and whether they present an opportunity for mitigation of methane emissions in California. The survey partners intend to discuss the results with stakeholders, make more systematic ground-level emission measurements at selected sources, and work with various partners on follow-up activities to reduce emissions. The field campaign for Phase II of the study will be completed in the fall of 2017, and a final report for the entire effort will be available in late 2018.

The study focuses on three important methane emission sources in the State: dairies, landfills, and the hundreds of thousands of potential sources in the oil and gas supply chain (wells, processing facilities, refineries, storage tanks, and transmission and distribution pipelines). The following sections describe the legislation and regulations recently put in place in California to address emissions from these sources and provide important context for the results presented in the Phase 1 report.

I. Methane Control for Dairies

In addition to setting an overall methane reduction target, Senate Bill 1383 set a specific target of a 40% reduction from the dairy and livestock sector and recognizes the potential for anaerobic digestion (AD) to direct renewable methane to beneficial uses such as offsetting fossil fuel energy sources. Accordingly, Senate Bill 1383 added provisions designed to encourage the development of AD at dairy operations. These include:

• Developing at least five pipeline-interconnected biomethane pilot projects;

• Increasing the certainty of revenue streams from environmental credits by developing a pilot financial mechanism;

• Providing guidance on the impact of regulation on environmental credit values;

• Developing infrastructure and procurement policies to encourage biomethane development;

• Considering the adoption of methane reduction protocols; and

• Convening a stakeholder group to address barriers to dairy methane reduction projects.

Further solidifying the State’s commitment to reducing dairy methane emissions, the Legislature passed Assembly Bill 1613 (Chapter 310, Statues of 2016) that included a $50 million appropriation to the California Department of Food and Agriculture (CDFA) to distribute between two dairy-specific programs. The first, the Alternative Manure Management Program (AMMP)2, provides as much as $16 million for non-AD based 2 More information on the AMMP: https://www.cdfa.ca.gov/oefi/AMMP/

https://www.cdfa.ca.gov/oefi/AMMP/

of 6

methane reduction projects. The second program, the Dairy Digester Research and Development Program (DDRDP)3, initiates a second round of funding that builds upon the success of the first round of disbursements by providing as much as $36 million for AD system development. These resources demonstrate the Legislature’s clear intent to reduce dairy methane emissions while supporting the beneficial uses of biogas produced using AD.

These programs are now underway. The first round of Dairy Digester Research and Development Program funds have already been disbursed and the projects they supported are at or near completion. Funding awards under the second round will be announced by the end of 2017. Extensive stakeholder outreach to the environmental justice community, government agencies, industry, and university researchers concerning methane reduction and AD system deployment has begun in earnest. Dairy pilot project solicitations are being developed, as is a pilot financial mechanism proposal. An additional action recently taken by the State was to designate a dairy AD project as a California Sustainable Freight Action Plan4 pilot project. This project will demonstrate the potential of dairy biomethane to reduce freight transport emissions by displacing diesel-powered vehicles with clean, renewable-biomethane-powered vehicles.

Anaerobic digestion is an important option for controlling dairy methane emissions and is an established technology globally. It can capture and control as much as 98%5 of the methane emissions that would otherwise be released to the atmosphere. Although AD has demonstrated significant potential to reduce methane emissions, the technology is in the early stages of maturation. AD systems have been consistently improving and now exhibit solid reliability, longevity, and potential to generate revenue from the sale of energy products. As expected with any new technology, during the initial operational deployment of AD situations may arise including leaks or higher-than-predicted venting that can impact the desired methane control efficiency. Characterizing potential AD fugitive emissions through the use of tools like the NASA JPL flyovers can help achieve and maintain the expected methane reductions by helping digester operators optimize their systems and reduce revenue loss.

The NASA JPL flyovers are effective in identifying instantaneous emission point sources that represent a snapshot but require additional information regarding facility operations to distinguish between transient, maintenance-related leaks, and persistent, unexpected emissions sources that may merit further focused on-site attention. The NASA JPL

3 More information on the DDRDP: https://www.cdfa.ca.gov/oefi/ddrdp/ 4 More information on the California Sustainable Freight Action Plan: http://www.casustainablefreight.org/ 5 Biogas collection efficiency for dairy AD systems can be as high as 98%, as noted on pg. 46 (Appendix A Emissions Factor Tables – Quantification Methodology, Table A.3) of the Livestock Projects Compliance Offset Protocol: https://www.arb.ca.gov/cc/capandtrade/protocols/livestock/livestock.htm

https://www.cdfa.ca.gov/oefi/ddrdp/

http://www.casustainablefreight.org/

https://www.arb.ca.gov/cc/capandtrade/protocols/livestock/livestock.htm

of 6

flyovers yield important information about digester leaks and maintenance events, but can also assess emissions from digesters during normal operations and at times when they are operating at their expected control efficiencies. For this reason, NASA JPL has done multiple flyovers of key dairy regions and will continue to do so during Phase II of the project.

CARB is looking forward to working with NASA JPL and the dairy industry to develop a better understanding of AD system optimization. This important learning opportunity can help inform efforts to address leaks—possibly including the development of new leak detection and repair protocols. Such protocols can help ensure maximum methane capture, minimum revenue loss, and improved long-term reliability.

II. Methane Control for Landfills Accepting Organic Wastes

Landfills have long been identified as a significant source of methane emissions in California. Landfilled organic materials break down anaerobically into methane, which escapes through cover materials, becoming a fugitive emission. Because organic wastes constitute a significant portion of California’s waste stream, the State is moving to divert and manage them. This requires not only keeping organics out of landfills, but also improving the infrastructure for diverting and/or recycling organics, including minimizing and recovering edible food wastes; and incentivizing conversion process such as composting and anaerobic digestion that yield valuable energy and soil amendment products.

CARB, in partnership with local, State, and federal entities is working to address methane and related emissions through implementation of various programs such as the Landfill Methane Regulation, and the organic waste diversion provisions of Senate Bill 1383 (Lara, Chapter 395, Statutes of 2016).

Under CARB’s Landfill Methane Regulation6, which became effective in 2010, owners and operators of certain uncontrolled municipal solid waste landfills are required to install gas collection and control systems. The regulation also requires existing and newly installed gas collection and control systems to operate in an optimal manner.

As required by AB 1383, CalRecycle, in consultation with CARB, is developing regulations7 to reduce disposal of organic waste by 50% by 2020 and 75% by 2025 (reductions are measured against a 2014 baseline). CalRecycle is planning to adopt the necessary regulations by the end of 2018. As specified in AB 1383, these regulations can take effect on or after January 1, 2022.

6 https://www.arb.ca.gov/regact/2009/landfills09/landfills09.htm 7 http://www.calrecycle.ca.gov/organics/

of 6

Reducing the disposal of organics in landfills would align California with a growing range of organic waste diversion efforts underway in other jurisdictions. For example, San Francisco and Alameda Counties require that food waste be separated and kept out of the landfill, and various cities, including Los Angeles and San Francisco, have plans in place to achieve zero-waste.

Although the organics diversion efforts described above will reduce landfill methane emissions over time, the organic wastes in place in California’s landfills will continue to emit methane for decades to come. Monitoring efforts like this program will help quantify the rate of emissions reductions from California’s landfills. They will also help identify malfunctions in landfill gas extraction and control systems, enabling more timely repairs. This approach is being actively tested at the Sunshine Canyon Landfill in Los Angeles County where NASA JPL is working closely with the landfill operator and local agencies to share data and offer feedback as they take steps to reduce emissions. As such, the NASA JPL flyovers are an important component in California’s suite of programs aimed at reducing landfill methane emissions.

III. Methane Control for the Oil and Gas Sector

CARB and CEC have recognized that the oil and gas sector is the source of significant methane emissions. Accordingly, CARB recently adopted a regulation8 to reduce methane emissions from oil and gas production, processing, and storage in California. Beginning on January 1, 2018, oil and gas operators will need to: 1) enhance or begin leak detection and repair (LDAR) procedures to control fugitive methane emissions from the sector; 2) control methane emissions from uncontrolled oil-water separators and storage tanks; 3) replace venting pneumatic devices with devices that do not vent methane; 4) control methane leaks from compressors; and 5) undertake enhanced monitoring at underground natural gas storage facilities.

In addition to this CARB regulation, the California Public Utilities Commission (CPUC) adopted a set of Best Practices for Methane Leakage Abatement and Emissions Reductions9 this year, which will lead to methane emission reductions from the natural gas transmission and distribution sectors. These Best Practices complement the CARB regulation by addressing the segments of the oil and gas sector that the CARB regulation does not address.

Enforcement of the new CARB and CPUC regulations, as well as continued enforcement of the air districts’ current volatile organic compound (VOC) rules for oil and gas sources, should start reducing the number and severity of the detectable methane plumes identified in this sector by the NASA JPL flyovers. For example, pump

8 https://www.arb.ca.gov/cc/oil-gas/oil-gas.htm 9 http://www.cpuc.ca.gov/General.aspx?id=8829

of 6

jacks and tanks—two of the most frequent sources of high methane emissions included in the oil and gas portion of the study—will be subject to the methane LDAR in CARB’s regulation starting on January 1, 2018. In addition, continued surveillance of the kind being done by the NASA JPL flyovers will help evaluate the success of (and compliance with) the CARB oil and gas methane regulation. CARB recently published the Refinery Emergency Air Monitoring Assessment, Objective 2 report,10 which includes recommendations to improve emergency air monitoring, as well as monitoring of ongoing routine emissions, at California’s major refineries and the communities that surround them. Along with CARB’s planned community air monitoring near oil and gas facilities, the recommended actions will provide valuable information that can be used to further enhance the State’s ability to evaluate reductions in methane and other pollutants from these sectors.

10 https://www.arb.ca.gov/fuels/carefinery/crseam/o2reamarmainfinal.pdf

CALIFORNIABASELINEMETHANESURVEY–INTERIMPHASE1REPORT

i

CALIFORNIABASELINEMETHANESURVEY

InterimPhase1report

PREPAREDBYTHEJETPROPULSIONLABORATORYFORTHECALIFORNIAAIRRESOURCESBOARD

ANDCALIFORNIAENERGYCOMMISSION

OCTOBER1,2017

RileyDuren,AndrewThorpeandStanleySander

NASAJetPropulsionLaboratoryCaliforniaInstituteofTechnology

Copyright 2017 California Institute of Technology. U.S. Government sponsorship acknowledged.


ii

Acknowledgements. Phase1of this projectwas fundedby theCaliforniaAirResourcesBoard ARB-NASA Agreement 15RD028 Space Act Agreement 82-19863. Phase 2 will befundedbytheCaliforniaEnergyCommission(agreementnumber500-15-004).Partofthisresearch was carried out at the Jet Propulsion Laboratory, California Institute ofTechnology,undercontractwiththeNationalAeronauticsandSpaceAdministration.TheauthorsthanksTalhaRafiq,BrianBue,IanMccubbin,MichaelEastwood,DavidThompson,Winston Olson-Duvall, Valerie Carranza, Andrew Aubrey, Tom Pongetti, Charles Miller,RobertGreenand theAVIRIS-ng team(JPL),FrancescaHopkins (UniversityofCalifornia,Riverside),andLiyinHe(Caltech)fortheircontributionstothiswork.TheauthorsthankDr. Jack Kaye at NASA for support of our methane research activities, particularlyexploratory airborne campaigns in California and the Four Corner’s region. We alsoacknowledgeNASAandNISTsupportfortheCLARSinstrumentonMtWilsonaspartoftheMegacities Carbon Project. The project also benefits frommethane data processing andanalysis tools and a new Geographic Information System data set developed by twoconcurrentNASAprojects: theACCESSprogram’sMethaneSourceFinder and theCarbonMonitoringSystem(CMS)program’sPrototypeMethaneMonitoringSystemforCalifornia.


i

ACRONYMS,ABBREVIATIONS,ANDDEFINITIONS

ACCESS:NASAAdvancingCollaborativeConnectionsforEarthSystemScienceprogram

AVIRIS-NG:NextGenerationAdvancedVisible/InfraredImagingSpectrometer

CARB:CaliforniaAirResourcesBoard

CEC:CaliforniaEnergyCommission

CH4:Methane

CLARS:CaliforniaLaboratoryforAtmosphericRemoteSensing

CO:Carbonmonoxide

CO2:Carbondioxide

COMEX:CO2andMEthaneeXperiment

Enhance.:Enhancement

FTS:FourierTransformSpectrometer

FTIR:Fouriertransforminfraredspectroscopy

GFIT: Gas Fitting algorithm used to fit themeasured CLARS-FTS spectrum to a spectralmodelderivedfromlaboratoryandtheoreticallinelists.

H2O:Watervapor

HITRAN:High-resolutiontransmissionmolecularabsorptiondatabase

IME:Integratedmethaneenhancement

IPCC:IntergovernmentalPanelonClimateChange

LABS:LosAngelesBasinSurveys

MarkIVFTIR:Fouriertransforminfraredspectroscopyinterferometerdeployedbyballoon

Megacities Carbon Project: Project to develop and test methods for monitoring thegreenhousegasemissionsofthelargesthumancontributorstoclimatechange:citiesandtheirpowerplants(https://megacities.jpl.nasa.gov/portal/)

N2O:Nitrousoxide

NCEP:NationalCentersforEnvironmentalPrediction

O2:Oxygen


ii

ppmm: Representing the thickness and concentration within a volume of equivalentabsorptionthatisequivalenttoanexcessmethaneconcentrationinppmifthelayerisonemeterthick.

SCD:Dry-airslantcolumndensity

SF:SanFrancisco

SJV:SanJoaquinValley

SoCAB:SouthCoastAirBasin

Spectralon:PanelcomposedoffluoropolymerthathasthehighestdiffusereflectanceofanyknownmaterialorcoatingandexhibitshighlyLambertianbehavior.

SVO:SpectralonViewingObservations

TBD:ToBeDetermined

TCCON:TotalCarbonColumnObservingNetwork

UTC:UniversalTimeCoordinated

XCH4:Totalcolumnaveragedmethaneexcessmixingratio


i

1.

1. EXECUTIVESUMMARY..................................................................................................................................12. PROJECTOVERVIEW......................................................................................................................................42.1.MOTIVATION....................................................................................................................................................4

2.2. Priorstudies..........................................................................................................................................42.3. ProjectObjectives...............................................................................................................................5

3. METHODS............................................................................................................................................................63.1. Tieredobservingstrategyformethaneemissions...............................................................63.2. AVIRIS-NGinstrumentandmethaneretrievals.....................................................................73.3. Airbornesurveydesign....................................................................................................................93.4. Pointsourceattribution,quantificationandvalidation...................................................143.5. CLARSmethanemonitoringforLosAngelesBasin............................................................15

4. INTERIMRESULTS........................................................................................................................................194.1. Airbornesurveystatistics.............................................................................................................20

4.1.1. Surveycompleteness...................................................................................................204.1.2. Distributionofmethanesources...........................................................................23

4.2. Sectorspecificfindings...................................................................................................................264.2.1. Oilandgasproduction................................................................................................274.2.2. Naturalgastransmission,storageanddistribution.......................................304.2.3. Refineries 324.2.4. PowerPlants....................................................................................................................334.2.5. Landfills 334.2.6. Wastewatertreatment...............................................................................................354.2.7. Dairiesandlivestock....................................................................................................35

4.3. SouthCoastAirBasinmethanetrendsandvariability.....................................................395. CONCLUSIONS.................................................................................................................................................456. REFERENCES...................................................................................................................................................46APPENDIXA–POINTSOURCEDATABASE....................................................................................................1APPENDIXB–CATALOGOFPOINTSOURCEIMAGERY.........................................................................1


1

1. ExecutivesummaryJPLisapplyingadvancedremotesensingmethodstodetectandcharacterizeanthropogenicmethaneemissionsinCaliforniatosupporttheState’sobjectivesformitigationofshort-livedclimatepollutants(California2013,2017),identificationofmethane“hotspots”inresponsetoAB1496(California2015),andsupportingnaturalgasleakdetectionandcorrectionforrate-payerbenefit.TheprojectisbeingimplementedintwophasesonbehalfoftheCaliforniaAirResourcesBoardandCaliforniaEnergyCommission(CEC).Phase1primarilyuseddatacollectedin2016andaddressesCARBprioritiesspanningIntergovernmentalPanelonClimateChange(IPCC)methaneemissionsectorsrelevanttopointsourcesinCalifornia.Phase2willcollectdatain2017andfocusonCECpriorities,particularlythenaturalgassectortoimproveunderstandingofleaksandtohelpenablemitigation.ThisinterimreportsummarizesthePhase1activityandfindingsincludinglessonsthatwillinformdatacollectionandanalysisstrategiesforPhase2aswellasfutureresearchprojects.AfinalreportwillbeproducedattheconclusionofPhase2. Phase 1 data collection included an airborne remote sensing survey of nearly 180,000individual facilities and infrastructure components across California between Septemberand November 2016. Over 15,000 km2 of land area was imaged at spatial resolutionsrangingfrom1to3meters.Thisrepresentsaphase1surveycompleteness(comparedtoall known relevant infrastructure in the State) ranging from 20 to 100% per emissionsector;forexample,35%ofknownpowerplants,38%oflandfills,and50%ofdairiesweresampledatleastonce.The phase 1 airborne survey identified 329 unique methane point sources* with highconfidence. Some of the sources were observed repeatedly, providing some insight intotheir variability. Based on these observations a Source Database (Appendix A) wasdevelopedthatincludesforeverysourcethelocation,emissiontype,andsize(expressedasan average atmospheric enhancement and plume length). Additionally, phase 1 of thisproject extended an established time-series (2011-2015) of methane emissions in theSouthCoastAirBasin(SoCAB)throughspring2017.Allphase1projectobjectiveshavebeenmetorexceeded.Thefollowinginterimfindingsare based on a preliminary analysis of data collected in phase 1 with conservativeassumptions.AquantitativeestimateofindividualandnetpointsourceemissionfluxesanduncertaintieswillbeprovidedinthePhase2 finalreport. Thefollowing interimfindingsaresubjecttorevisioninphase2followingadditionalanalysisandvalidation.

* Methanepointsourcesareemissionsourcesthatarespatiallycondensed–typicallyoriginatingfromasurfacelessthan10metersacross.Pointsourcesoftengeneratewell-definedplumesofmethanegassimilartosmokereleasedfrompowerplantsandfires.Thisisincontrasttoareasourcesthatreleasemethaneinamorediffusefashionoversurfacesspanninghundredofmeterstomanykilometers(e.g.,entericfermentation,ricecultivation,andwetlands).Pointsourcesrefertothespatialcharacteristicsofthesourcenottheroot-cause(e.g.,apointsourcecanresultfromexpectedprocessemissionsorunplannedfugitiveemissionssuchasleaks).Thisreportdoesnotattempttoassignroot-causetopointsources.


2

Interimfindings:

1. Strongmethaneplumes†areobservedatarelativelysmallfraction(<0.2%)ofCalifornia’sinfrastructurewiththepotentialformethanefugitiveemissions.Methaneplumesappearmorefrequentlyinspecificsectors,particularlyrefineries(94%),landfills(86%),gasstoragefacilities(25%),anddairies(22%).[Seecaveatinfinding2]

2. Detectingandquantifyingmethaneemissionsfromsomesectors–particularlydairies,naturalgastransmission,storageanddistributionsystems,andpetroleumrefineries–iscomplicatedbysignificantvariability.Failuretoproperlysampleepisodicactivitycouldresultinsignificantunder-estimationoftheprevalenceandnetemissionsofmethanesources.Hencethesourcedistributionpresentedhereislikelyconservative.Futureeffortstoassessandmitigateemissionswouldbenefitfrommorefrequentandsustainedmonitoring.

3. Theobservedspatialdistribution,sizeandatmosphericenhancementsofmethaneplumesintheSanJoaquinValley(SJV)indicatethatdairiesarelikelytheleadingcauseofmethanepointsourceemissionsinthatregion.However,resolvingtherelativecontributionsofmanuremanagementpointsourceemissionsandentericfermentationareasourceemissionswillrequireadditionalobservationsandmodelingbeyondthescopeofthisprojectandperhapsapriorityforfutureresearch.

4. LargemethaneplumesthatpersistformonthsareobservedatsmallnumberofdairydigestersintheSJV;thesearelikelyassociatedwithmanualventingfromfacilitiesundergoingmaintenanceorconstruction.

5. MethaneplumesareobservedatmultiplelocationswithinandinthevicinityofdairiesintheSJVandarelikelyassociatedwithdifferentstagesofmanuremanagementaswellascropirrigationwithwastewater.Thespatialdistributionofdairymethaneemissionsiscomplexandrequiresadditionalstudytofullyunderstandthecontrollingprocesses.

6. ThedenseconcentrationsoflargedairiesinsomepartsoftheSJVresultsinstrongmethaneenhancementsoverlargeareas,complicatingeffortstoattributeemissionstoindividualfacilities.However,thisprojectalsodemonstratedtheabilitytodirectlyimagethoselargeregionalenhancementsandthepotentialforfuturequantificationofareamethanefluxesusingremotesensing.

7. Theprevalenceandmagnitudeofmethaneplumesatlandfillsvariessignificantlybetweenfacilitiesandisnotwell-predictedbybottom-upestimates.Additionalstudyisrequiredtofullyunderstandthecontrollingprocessesincludingtherelativeinfluenceofwastevolumesandmanagementpractices.

8. Themajorityofmethaneplumesobservedinoil-andgas-producingfieldsinCaliforniaoccurinKernCounty.Additionalstudyisrequiredtodeterminewhetherthisisprimarilyduetothehigherproductionratesinthatcountyorotherfactors.

†Forthisreportstrongplumesaredefinedasthosewithemissionratesexceedingthenominaldetectionthresholdoftheairborneimagingspectrometerusedinthissurvey(about10kgCH4/hrassuminga3meter/secondwind).


3

9. Thisprojecthasdemonstratedtheabilityofairborneimagingspectroscopytodetectandpin-pointaleakingundergroundnaturalgasdistributionlineinaneighborhoodandprovidenearreal-timenotificationtooperatorstoguiderepairefforts.

10. Asimpleanalysisofplumeenhancementsforthesourcesdetectedinthisstudysuggeststhatasmallfractionofsourcesinnearlyeverysectorlikelydominatesthenetemissionsofthepopulation.Furtheranalysisandemissionfluxestimationinphase2willberequiredtoconfirmtheextentofpotential“super-emitter”behavioranditscontributiontoCalifornia’snetmethanebudget.

11. PersistentmonitoringofSoCABregionalmethaneemissionsovera5yearperiodindicatessignificantseasonalvariabilitythatislikelyassociatedwithnaturalgasuse.Additionalresearchisrequiredtoattributewhichpart(s)ofthegassupplychainintheSoCABareresponsibleforthisvariabilityandtodeterminewhetherthisextendstootherurbanareasintheState.

12. Persistent,spatiallyresolvedremotesensingofatmosphericmethaneenhancementsacrosstheLosAngelesbasindetectedtheimpactoftheAlisoCanyongasleakandreturntonormalconditions,suggestingthepotentialforfuturemethanehotspotmonitoringforanomalousemissionevents.


4

2. ProjectOverview

2.1. MotivationMethaneisapowerfulgreenhousegasandistargetedforemissionsmitigationbythe

StateofCalifornia [California2017]. Methane isalsoaprecursor for troposphericozoneandisstronglylinkedwithco-emittedreactivetracegasesthatarethefocusofairqualityand public health policies, particularly in high priority regions such as the San JoaquinValley(SJV)andtheSouthCoastAirBasin(SoCAB).Globally,theatmosphericgrowthrateofmethaneis likelystrongly influencedbyanthropogenicemissionsfromapopulationofspatiallycondensedpointsourcesdistributedoverlargeareasandspanningdiversesocio-economic sectors. However, “bottom-up” (inventory-based) estimates of methaneemissions are often in disagreement with top-down (atmospheric measurement based)estimates[Wechtetal2014,Turneretal2015,Wongetal2016,Jeongetal2017].

Limitations inprocess-basedunderstandingofmethaneemissions isexemplifiedbythe ongoing scientific discussion on both the hiatus in the atmospheric growth rate ofmethaneintheearly2000’sandtheunexpectedrisestartingin2007[Kirschkeetal2013].Emissionsandprocessattributionremainhighlyuncertainbutareneededtoresolvekeyelements of the global carbonbudget, generate accurate greenhouse gas inventories andinform emission mitigation decisions. A key factor is that many current methanemonitoringmethods (bottom-up and top-down) are limited to regional or coarser scaleresolutionandoftencannotdetectindividualsourcesorattributefluxestospecificactivityandfacilities.Othermethodsaresufficientforstudyingpreviouslyknownsourcesbutarenotwell suited to surveying largeareas forunknownsources.Hencemethaneemissionsremainachallengingtargetforabatementsincethelocationsandemissionfluxesofmanysignificantsourcesarestillmostlyunknown.Thesechallengesarereflectedintherecentlyenacted California AB1496 law: “there is an urgent need to improve the monitoring andmeasurement of methane emissions from themajor sources in California” and directs theCalifornia Air Resources Board to “undertake, in consultation with districts that monitormethane, monitoring and measurements of high-emission methane hot spots in the Stateusing the best available and cost-effective scientific and technical methods”. Anothermotivation is supporting efforts by natural gas utilities to improve leak detection andrepair:ageneralbenefittoCaliforniarate-payers.

2.2. PriorstudiesCalifornia has benefited from a number of top-down studies focused onmethane.

The2010CalnexcampaignaddressedmanysectorsandpriorityregionssuchastheSoCABand SJV (Ryerson et al 2013). There has been an ongoing focus on SoCAB methaneemissionsandtrends(Wennbergetal2012;Wunchetal2016;Wongetal2016),sourceattribution (Hopkins et al 2016), and characterization of individual sources such as theAlisoCanyongasleakincident(Conleyetal2016;Thompsonetal2016).

Recent years have also seen a dramatic improvement in the ability of passiveremote-sensingmethods to detect and locate largemethane sources. Observations frompolar orbiting satellites have detected strong, persistent enhancements of atmosphericmethaneintheFourCornersregionandCalifornia’sSanJoaquinValley[Kortetal.,2014]andhaveproducedspatiallyresolvedestimatesofUSmethaneemissiontrends[Turneret


5

al, 2016]. The planned 2017 launch of the TROPOMI instrument on the Sentinel-5Precursor satellite should further advance space-based methane detection for globalstudies[Butzetal2012].However,theabilityofsatellitestodetectandquantifyemissionsfrom point sources is still limited to relatively coarse spatial scales (typically 25 km atbest). Some surface measurement networks andmodels can resolve methane fluxes atresolutions as fine as a fewkilometers but so far this is limited to a fewurban testbeds[McKain et al 2015] and in most cases is insufficient to pinpoint and attribute pointsources.

JPL and partners have devised a tiered observational strategy for efficientlysurveyinglargeareasformethanepointsources,quantifyingindividualsourceemissions,and estimating their contributions to the net emissions of key regions and sectors. Thestrategy is flexiblewith regards to vantage points andmeasurement systems – enablingsignificant near-term progress using existingNASA remote sensing instrumentation thatweredevelopedasprototypes fornextgenerationsatellites.Over thepast twoyears thisstrategywastestedwithaseriesofexploratoryairbornefieldcampaignsoverCalifornia’sCentral Valley and SoCAB [Thompson et al 2016] as well as the Four Corners region[Frankenberg et al 2016]. Additionally, persistent monitoring of SoCAB total methaneemissionsbyJPL’sCaliforniaLaboratoryforAtmosphericRemoteSensing(CLARS)overtheperiod2011-2015[Wongetal2016]demonstratedtheability toassessvariability in theemissionsofakeyregion.

2.3. ProjectObjectives

Based on the success of exploratory NASA airborne campaigns and in response toCalifornia policyneeds theCaliforniaAirResourcesBoard (CARB) andCaliforniaEnergyCommission(CEC)arefundingJPLtoconductthefirstcomprehensiveairbornesurveyofmethanepoint sources in the State. Additionally, CARB contributed to extendingCLARSdatacollectionandanalysisformethaneemissionsintheSoCABthroughSpring2017.

Theprojecttechnicalobjectivesforphase1areasfollows:1. Preparea SurveyAreaFlightPlan that covers all keyemissions sectors for

methane point sources in California including dairies, feedlots, landfills,wastewater treatment facilities, natural gas supply chain, oil production,processing and refineries, natural seeps, etc including specific areas ofinterestidentifiedbyCARBandCECstaff.

2. Conduct the California Baseline Methane survey using airborne imagingspectrometer(s) to detect and characterize fugitive methane emissionsourcesintheSurveyAreaFlightPlanandprovideaflightcoveragereport.Asresourcesallow,providefollow-upsurveysofstrongemitters.

3. Deliver maps of methane enhancements (both individual point sourceplumes and regional enhancements for the SoCAB) and a Statewide pointsourcedatabaseincludingsourcelocations,plumesizes,andemissiontypes.Pointsourcemapstobedeliveredinelectronic,georeferencedformats.

4. Deliver a report summarizing the project purpose, approach, findings andconclusions.


6

3. Methods

3.1. TieredobservingstrategyformethaneemissionsPhase 1 data collection involved a two-tiered approach: 1) a broad airborne survey ofmethane point sources spanning key regions and sectors across the State by JPL’s NextGeneration Advanced Visible/Infrared Imaging Spectrometer (AVIRIS-NG) and 2)persistent regional-scale monitoring of methane emissions in the South Coast Air Basin(SoCAB)by JPL’sCalifornia Laboratory forAtmosphericRemote Sensing (CLARS) facilityonMtWilson.Thefirsttierisprovidinganunprecedentedbaselineassessmentofmethanepointsources in theState.Thesecondtier isextendingapreviouslyestablishedmethaneemissions time-series for the SoCAB. Figure 3-1 illustrates representative CLARS andAVIRIS-NGdataproducts(thelatterisactuallyanAVIRIS-classicimage).

Figure3-1ExampleofobservationsofmethaneassociatedwiththeAlisoCanyongasleak.Left:CLARSmapofcolumnmethane/CO2correlationratiosshowingalargeareaofenhancedmethane(red)extendingacrosstheSoCABonDecember25,2015–theleaksourceisjustoutsidetheCLARSfieldofregardinthisfigure.Right:AVIRIS-CimageofAlisoCanyonandthegasleakplumeextendingover>2kmonJanuary12,2016[Thompsonetal2016].

The focus for theairbornesurvey forphase1(andtheexclusive focus forphase2) isonmethane point sources. The airborne remote sensing method applied here are notoptimized fordetectingandquantifyingareasourcesandhencemethaneemissions fromareasourcessuchasentericfermentation,ricecultivationandwetlandsareexcludedfromthisstudy.


7

3.2. AVIRIS-NGinstrumentandmethaneretrievalsThe next generation Airborne Visible/Infrared

Imaging Spectrometer (AVIRIS-NG)measures ground-reflected solar radiation from the visible to infraredspectral regions (350 to 2,500 nm). This push broominstrument has a 34° field of view and operates onhigh performance aircraft, allowing for efficientmapping of large regions. Increasing flight altitudeaffectsthegroundresolution,i.e.,thesizeofeachimagepixel increases while the image swath increases(Figure 3-2, Table 3-1). For most of the Fall 2016campaign,AVIRIS-NGflewat3kmabovegroundlevel,resultingin3mimagepixels.

The methane retrieval is based on absorptionspectroscopy (Figure 3-3) and has been used for anumber of prior NASA research campaigns includingBakersfield area oil fields (Thompson et al., 2015), acampaign to theFourCorners region inColoradoandNewMexico (Frankenberg et al., 2016), Aliso Canyon(Thompson et al., 2016), and a study of Californialandfills (Krautwurst et al., 2017). A methanecontrolled release experimentindicated consistent detection ofplumes for releases as low as 14.16m3/h (~10 kgCH4/hr) at multipleAVIRIS-NG flight altitudes andvariable wind speeds (Thorpe et al.2016).

The methane retrieval is alinearizedmatched filter thatmodelsthebackgroundofradiancespectraasa multivariate Gaussian having meanμ and covariance Σ. We estimate background parameters using the image data in theappropriate pushbroom cross track location. This compensates for subtle uncorrectedvariations in radiometric response or dark current level by different focal plane arrayelements, permitting accuracy beyond what is possible from a purely first-principleslaboratory calibration (Thompson et al., 2015). The matched filter tests the nullbackgroundcaseH0againstthealternativeH1inwhichthebackgroundisperturbedbyasignalt:

𝑯𝟎: 𝒙 ~ 𝑵 𝛍,𝚺 (1)𝑯𝟏: 𝒙 ~ 𝑵 𝛍 + 𝛂𝐭,𝚺 Inthisequation,αisascalingoftheperturbingsignal.Thematchedfilterα*(x)istheoptimaldiscriminantbetweenthesehypotheses.Itestimatesαbyprojectingthemean-

Figure3-2.AVIRIS-NGflightparameters:L=imageswathwidth,V=aircraftvelocity,FOV=fieldofview,IFOV=instantaneousFOV(Murai,1995).

Table3-1.AVIRIS-NGimageparameters.

Flightaltitude(metersabovegroundlevel)

Imageswathwidth(meters)

Groundresolution(meters)

1,000 611 12,000 1,223 23,000 1,834 3


8

removed spectrum onto the target, after whitening to account for covariance. Theestimatortakestheform:𝜶∗(𝒙) = (𝒙!𝝁)

𝑻𝜮!𝟏𝒕𝒕𝑻𝜮!𝟏𝒕

(2)

We write the estimatedcoefficient as α*. Carefullydefining the target signature tpermits a quantitativeinterpretation of this value interms of physically-meaningfulscene properties. Specifically,wedefinethetargetsignaturetas thechange inradianceunitsof the background caused byadding a unit mixing ratiolength of methane absorption.The additional absorption actsas a thin Beer-Lambert attenuation ofthebackgroundμ.Our target signatureis the partial derivative of measuredradiancewithrespecttoaenhancementin absorption path length l by anoptically-thin absorbing layer ofmethane.Atthebackgroundlevelofl=0wehave:𝒕 = 𝝏𝒙

𝝏𝒍 = −𝝁 𝒆!𝜿𝒍𝜿 = −𝝁𝜿 (3)

Here, κ represents the unitabsorptioncoefficientandμisthemeanradiance as before. The detectedquantity α* is a mixing ratio length inunits of ppm m representing thethickness and concentration within avolumeofequivalentabsorption.Thisisequivalent to an excess methaneconcentrationinppmifthelayerisonemeterthick(i.e.directlyequivalenttoppbkm).Atascaleheightofabout8km,thetotalcolumnaveragedexcessmixingratioXmethanewouldbeabout0.000125timestheexcessinppm-m.Forexample,1000ppm-misequivalenttoanXmethaneenhancementof125ppb. IntegratingoverthephysicalareaoftheplumeyieldsanIntegratedMethaneEnhancement(IME)inkg,asinThompsonetal.(2016)andFrankenbergetal.(2016),tantamounttothetotalobservedmassofmethaneabovetheambientbackground.This

Figure3-3.methaneabsorptionsignature(transmittance)plottedforthewavelengthrangemeasuredbyAVIRIS-NG.Strongabsorptionsarepresentbetween2,200and2450nm.

Figure3-4.Realtimemethanemappingonboardtheaircraft.RedmethaneplumesareoverlaidonrawAVIRIS-NGimageswithestimatedpeakenhancement(ppm-m)andplumesourcecoordinates.


9

techniquecanbecombinedwithsimplesteadystateassumptionsforafirst-orderestimateofapointsourceemissionflux.Critically,awindspeed-correctedIMEisproportionaltothesourcestrength;fieldmeasurementsoffluxfromasubsetofsourcescanserveasareferenceforinferringfluxofmanydifferentsites. methaneretrievalsareperformedinrealtimeonboardtheaircraft(Figure3-4),whichpermitstheinstrumentoperatortoidentifyandgeolocateplumesinrealtime.Thisinformationcanbeusedforadaptivesurveyingandresultscommunicateddowntogroundcrewsforrapidfollowup.Attheendofeachflightday,methanequick-lookdataproducts(Figure3-5)aregeneratedandusedtoquicklyassessresultsandplanfutureflights.AftertheAVIRIS-NGdataistransportedtoJPL,scenesarereprocessedtogeneratemethaneretrievalsfororthorectifiedscenes(planimetricallycorrectimageswithconstantscale).

3.3. Airbornesurveydesign

Thephase1airbornesurveywithAVIRIS-NGwasdesignedtomapkeyregionsinCaliforniaandinfrastructurewiththepotentialformethanepointsourceemissions.AGeographicInformationSystem(GIS)datasetknownas“Vista-CA”thatmapspotentialmethaneemittinginfrastructureacrosstheStateofCaliforniaisbeingdevelopedbyJPLandUCRiversideaspartofNASA’sMethaneSourceFinderproject.TheVista-CAdatasetappliessimilarmethodsusedtodevelopa“Vista-LA”methaneGISdatasetforthegreaterLosAngelesareaaspartoftheMegacitiesCarbonProject[Carranzaetal2017].Vista-LAmappedthelocationsofinfrastructureassociatedwiththreeprimarysectors

Figure3-5.methanequick-lookproductsaregeneratedattheendofeachflightday.Thisexampleshowsaplumefromaleakinglow-pressuregaspipelinethatwasconfirmedbylocalgascompany.

Figure3-6.AVIRIS-NGflightboxesforFall2016campaignsurveyingtheenergysector(red),non-energysector(yellow),andmixtureofthesecategories(orange).


10

(energy,agriculture,andwaste)followingtheframeworksusedbytheStateofCalifornia’sGreenhouseGasInventoryandtheIPCCGuidelinesforGHGReporting.Geospatialmodellingwasappliedtopubliclyavailabledatasetstopreciselygeolocatefacilities.ForPhase1ofthisproject,apreliminaryversionofVista-CAcontaining299,644distinctpiecesofpotentialmethaneemittinginfrastructurewasusedtoguideselectionofflightboxes(Figure3-6)thatwereorganizedbroadlyintothreecategories,theenergysector(red),non-energysector(yellow),andregionsthatcontainamixtureofthesecategories(orange).ThisgroupingofsectorswaspartlydrivenbytheneedtodividethecampaignintotwophasestoaddressCARBandCECprioritiesgiventhatCECfundingfortheprojectisdedicatedtoenergysectoremissions.Figures3-7,3-8,3-9and3-10providefinerscalemapsoftheplannedflightlines(showninmagenta)forphase1alongwithkeyVista-CAelementsforfourprimaryareas.Theflightplanshowninfigure3-10wasadjustedtoaccountforreducedflighthoursintheSanFranciscoBayAreaduetoweatherconditionsinFall2016andotherconstraints.Note:theversionofVista-CAusedinphase1isstill

undergoingdevelopmentandreview;furtherrevisionsarelikelyinphase2.

Figure3-7Phase1airbornesurveydesignfortheLosAngelesBasin

LosAngelesBasin


11

Figure3-8Phase1airbornesurveydesignforthesouthernSanJoaquinValley

SouthernSanJoaquinValley


12

Figure3-9Phase1airbornesurveydesignforthenorthernSanJoaquinValley

NorthernSanJoaquinValley


13

Figure3-10Phase1airbornesurveydesignfortheNorthernCalifornia

Inadditiontotheseflightboxes,thereweretwofocusedmini-intensiveflight

campaigns.Thefirstinvolvedamulti-dayeffortfocusedonthenaturalgastransmissionandstoragesectorsintheSoCAB.Thesecondmini-intensivefocusedonanareanearVisaliathatwasmappedrepeatedlyovera5hourperiodtoinvestigatethespatialandtemporalvariabilityofdairymanureemissions.Additionally,formanyofthesourcesdetectedinphase1,revisitoverflightswereconducted.Inotherwords,flighthoursweremanagedtobalancebetweenbroadspatialcoverageandtemporalsampling.

NorthernCalifornia


14

3.4. Pointsourceattribution,quantificationandvalidationDuringtheflightcampaign,alistofcandidatemethaneplumeswascompiledusing

end of day quick-look products. Using this list as an initial guide,methane retrievals fororthorectified sceneswere then analyzedwith automated detection software to confirmand geolocate methane plumes. Analysts reviewed each plume candidate to confirm itsvalidityand toattribute themost likely source. To confirma sourceoneormoreof thefollowingcriteriawererequired:a)sourceplumeissostrongastobeunambiguouswithasingledetection,b)plumeclearlyoriginatesfromsurface,c)plumesaredetectedmultipletimesat thesame locationunderdifferentwindconditions,d)plumeorigin isco-locatedwithplumeofanotherspecies(e.g.,ammoniaorCO2),e)sourceisconfirmedbyfollowup,f)sourcewaspreviouslyknown.Inmanycasesfollow-upflightswereconductedoverhighprioritysourcecandidatesforconfirmationandtoassesstheirpersistence.

Sources thatpassedoneormoreof theabovecriteriawereassignedamediumtohighconfidencelevelandincludedintheSourceDatabase(AppendixA).Lowerconfidencesourcecandidateswereexcluded fromthe tableas likely falsealarms. Information in thetablealsoincludesthebestestimateofsourcelocation,typeandmagnitudeandlikelyIPCCemissionsectorbasedonanalysisofplumes,truecolorimagery,andinfrastructuremaps.

Eachplumeobservationisassignedauniqueidentificationnumbersincesomewereobservedmultipletimesforagivensource.Forexample,20plumeswereobservedforsourceS00006.Insomecasestheapparentoriginofplumesforagivensourceshiftedsomewhatbetweenrepeatobservations.Imageanalysisoftheseplumesincombinationwithadditionalinformation(i.e.infrastructuremaps,analysisofhigherspatialresolutionsatelliteimagery)wasusedtoassignamost-likelylocationforeachsource(Fig.3-11).

For each plume, an Integrated Methane Enhancement (IME) was calculated byintegratingoverthephysicalareaoftheplume.Thiswasdonebyfirstcalculatingthemassofmethanepresentineachimagepixelasfollows.kg = 𝐩𝐩𝐦 ∗𝐦

𝟏∗ 𝟏𝟏𝐄𝟔 𝐩𝐩𝐦

∗ 𝐩𝐢𝐱𝐞𝐥 𝐫𝐞𝐬. (𝐦)∗𝐩𝐢𝐱𝐞𝐥 𝐫𝐞𝐬. (𝐦)𝟏

∗ 𝟏𝟎𝟎𝟎 𝐋 𝟏 𝐦𝟑

* 𝟏 𝐦𝐨𝐥𝐞𝟐𝟐.𝟒 𝐋

*𝟎.𝟎𝟏𝟔𝟎𝟒 𝐤𝐠𝟏 𝐦𝐨𝐥𝐞

(4)


15

TheIMEisthencalculatedbyintegratingoverall pixels exceeding a specified methanethreshold in a given plume. The IME andplume lengths listed for a given source inAppendixAarebasedontheaverageofthosequantitiesforeachsource.The IME and plume length can ultimately becombined with wind speed information toestimate point source emission rates asfollows

Emissionrate(kg/hr) = 𝐈𝐌𝐄 (𝐤𝐠)

𝟏 ∗𝐖𝐢𝐧𝐝 𝐬𝐩𝐞𝐞𝐝 (𝐦)𝐬 ∗

𝟏𝐩𝐥𝐮𝐦𝐞 𝐥𝐞𝐧𝐠𝐭𝐡 (𝐦)

∗ 𝟑𝟔𝟎𝟎 𝐬 𝟏 𝐡𝐫

(5)Windspeeddataisavailableformanyareasbutincludesvaryingdegreesofuncertaintythatdirectlytranslatetouncertaintyinderivedemissionrates.Inphase2plumemodelingandvalidatedwindinformationwillbeappliedtoquantifyemissionrateestimatesforindividualsources.Winddataandemissionrateestimatesarenotincludedforphase1noraretheyreflectedintheanalysisinthisreport.

3.5. CLARSmethanemonitoringforLosAngelesBasinTheCaliforniaLaboratoryofAtmosphericRemoteSensing(CLARS)islocatedatanaltitudeof1670mabovesealevelwithapanoramicviewoftheLosAngelesbasin(Fig.3-12).TheCLARS-FTS instrument quantifies atmospheric column abundances of methane, CO2, CO,N2O,H2OandO2using reflected sunlight in the0.7-2.3µmregionusinga JPL-builthigh-resolutionFourierTransformSpectrometer(FTS).Itoperatesintwomeasurementmodes:SpectralonViewingObservations(SVO)andLosAngelesBasinSurveys(LABS).IntheSVOmode,the instrumentquantifiesthebackgroundfreetroposphericcolumnabundancesofmethane,CO2andotherspeciesabovetheLosAngelesbasinbymeasuringreflectancefromaSpectralon®platelocatedattheCLARSsite.IntheLABSmode,theinstrumentsamplestheslantcolumnabundancesofmethane,CO2andotherspeciesbymeasuringthespecularscatteredradiancefrom33discretelocations(orreflectionpoints)inthebasin(Fig.3-12).We selected these reflection points to achieve uniform optimal spatial and temporalcoverage of the Los Angeles basin. The number, locations and repeat frequencies of thereflection points can bemodified easily to meet specific measurement requirements. Ineach measurement cycle, we collect one set of LABS measurements and five SVOmeasurements.MultipleSVOmeasurementsareperformedpermeasurementcyclesothat

Figure3-11.ExampleofmethanesourceattributedtoaventstackattheHonorRanchogasstoragefacility.Theredmarkersindicatethedetectionofmultipleplumesoverseveralweeksthatwereusedtoinfertheexistenceofthesource(whitemarkercenteredonventstack).


16

anyvariabilityinthebackgroundduringeachmeasurementcycle,whichtypicallylastsfor90minutes,canbecaptured.Thereare5to8measurementcyclesperday,dependingonthe timeof the year. TheCLARS facility, CLARS-FTS instrument,measurement approach,retrievalalgorithms,dataproductsanduncertaintiesaredescribedbyFuetal.andWongetal.[Fuetal.,2014;CKWongetal.,2016;KWWongetal.,2015].

BasedontheBeer-LambertLaw,thedry-airslantcolumndensity(SCD)–thetotalnumberof absorbing molecules per unit area along the sun-Earth-instrument optical path – isretrieved formethane at 1.67µm, CO2 at 1.60µm, and O2 at 1.27µm using amodifiedversionoftheCLARSGasFitting(GFIT)algorithmdevelopedatJPL[Fuetal.,2014;Wunchet al., 2011]. The GFIT algorithm fits the measured CLARS-FTS spectrum to a spectralmodelderived from laboratory and theoretical line lists.These line lists are the sameasthose used in retrievals obtained by instruments that are part of NASA’s Total CarbonColumnObservingNetwork(TCCON)network.Theprimarysource istheHigh-resolutiontransmission (HITRAN) 2012 line list with modifications in some cases from empirical“pseudolinelists”derivedfromatmosphericspectra.Thelinelistforsolarabsorptionlinesisanexamplewhereempiricalobservationsareused.

TheGFITalgorithmwasdesignedforusewithspectrometersthatobservethedirectsolarbeam,e.g.theFTIRinstrumentsintheTCCONnetwork,andtheMarkIVballoonFTIR.Forthisviewinggeometry,theobservedradianceisdominatedbythedirectsunandthereforetheopticalpath iswell-defined. ForCLARSobservationshowever, aerosol scattering canmakeasignificantcontribution.Underhighlypollutedconditionswhereaerosolloadingissignificant,orwhentherearehighorlowcloudsintheopticalpath,theCLARSretrievalsbecome significantly more uncertain and some degree of filtering and/or flagging isrequired to identify these measurements. Using a numerical radiative transfer model,Zhangetal.estimatedthebiasinCLARSCO2andmethaneretrievalswithmoderateaerosolloading[Zhangetal.,2015].Whilethisworkproposedanewmethodforbiascorrection,thisapproachwasnotemployedinthepresentworkbecausethecorrectionalgorithmhasnotyetbeenimplementedintheCLARSoperationalretrievalcode.Rather,theretrievalsofmolecular oxygen areused to filter data that are affectedby aerosol scattering.Using6-hourNationalCentersforEnvironmentalPrediction(NCEP)surfacepressures,theO2SCDiscalculatedusingtheknowndry-airatmosphericmixingratioof0.2095.ThesevaluesarecomparedwithCLARSO2retrievals;soundingsthatdepartfromtheNCEPvaluesbymorethanacertainfractionarefilteredfromthedatastream.

The retrieved SCDs ofmethane and CO2 are converted to slant column-averaged dry airmixingratios,XmethaneandXCO2,bynormalizingtotheretrievedSCDofO2(SCDO2)(Eq.6).

XGHG = !"# !"!!"# !!

× 0.2095(6)

Individual retrievals are analyzed with multiple post-processing filters to ensure dataquality.SpectraarefilteredoutwhentheresidualrootmeansquareerrorsofthefitstotheGFITradiativetransfermodelexceedapre-definedthreshold.Theseareusuallyassociatedwith aerosols, high and low clouds, electrical or mechanical noise, and other transient


17

behavior.Details about theCLARS-FTSdesign,data retrieval algorithmanddata filteringprocessaredescribedinFuetal.(2014),Wongetal.(2015,2016).FollowingWongetal.(2015,2016),wecalculatedtheexcessXmethaneandXCO2,duetotheemissionsfromthebasin, by subtracting the corresponding SVOmeasurements from the LABS observations(Eq.7).

XGHG!" = XGHG!"#$ − XGHG!"#(7)SeveralstudieshavereportedstrongcorrelationsbetweenmethaneandCO2measuredinthePBLinsourceregions(Peischletal.,2013;Wennbergetal.,2012;Wunchetal.,2009;S.Newman,personalcommunication,2014).Slopesofmethane–CO2correlationplotshavebeenidentifiedwithlocalemissionratiosforthetwogases.SincetheuncertaintyinmethaneemissionsisconsiderablylargerthanthatinCO2emissions,wemayusethecorrelationslopetoreducethemethaneemissionuncertainties.Inthisstudywedeterminedthespatialvariationofmethane:CO2ratiosoriginatingfromCLARS-FTSmeasurementsbetweenSeptember,2011andDecember,2016.Wehavedeliveredtothedataportalthebackground-correctedvaluesofthemethane:CO2ratio(R)foreachsounding.Thedatafilealsoincludesthenameofeachreflectionpoint,measurementdayandtime(centroidoftheinterferometerscan),targetlat/lon,Rvalue,1-sigmauncertaintyoftheRvalueandtwodataflagswhichindicatethepresenceoflowandhighclouds.Thisinformation,combinedwithahigh-qualityCO2emissioninventorydownscaledtotheLAbasincanbeusedtoestimatemethanefluxes(asdescribedinWongetal2015andWongetal2016).ToexploretheoverallmonthlyvariabilityofRduringthe2011throughearly2017period,wecalculatedtheweighted-averageregressionslopeamongallthereflectionpointsintheLAbasinusingequation8.InEq(8),riistheregressionslopeforreflectionpointi,wiistheweightingfactorwhichisdefinedasthereciprocalofthesquareofthe1σuncertaintyofthemethane:CO2regressionslope,σi.

i iCLARS imonthly

ii

rwR

w=∑∑

(8)

where 2

1i

i

wσ

=

FollowingWongetal.(2015,2016)weestimatemonthlymethaneemissionsfromtheSouthCoastAirBasinusingthemethane:CO2regressionslope,R,determinedfromtheCLARSobservationsandaninventory-basedestimateofmonthlyCO2emissions(eqn.9)

4

4 2

2

| | CHtop down CLARS inventoryCH monthly monthly CO monthly

CO

MWE R E

MW− = × × (9)

where, iMW isthemolecularweightofmethaneorCO2.


18

AsdiscussedbyWongetal.(2016),thereareseveralchoicesavailablefor2|inventory

CO monthlyE .Theseuseanumberofdifferentgrids,underlyingdatasourcesandapproachestodownscalingtotheSCAB.WehaveelectedtousetheHestiafossilfuelCO2emissionsdataproductwhichprovidessectoralbottom-upemissionsatthebuildingandstreetlevelonhourlytimescales(http://hestia.project.asu.edu).

Figure 3-12. Relative location of the CLARS facility on Mt Wilson and basin reflectionpoints.


19

4. InterimResultsConservativeassumptionswereusedforapreliminaryanalysisofdatacollectedinphase1.Theinterimfindingsaresummarizedinsection1andhere.Thefollowingimportantcaveatsapplytotheseresults:

1. Theremotesensingmethodsappliedinthisprojectwerenotoptimizedfor

detectingandquantifyingareasourcesandhencemethaneemissionsfromareasourcessuchasentericfermentation,ricecultivationandwetlandsareexcludedfromthisstudy.

2. Estimatesofemissionrateswillnotbeavailableuntilwindinformationandothervariablesareproperlyaccountedwithplumemodelingandotheranalysesinphase2.Theobservationsandfindingsassociatedwithmethanesourcesectorsdescribedinthisphase1interimreportarethereforelimitedtofrequencyofoccurrence,plumesize,plumeenhancement(concentrationofmethaneinplumerelativetobackgroundcleanerair)andattributionofplumestolikelysourceinfrastructure.

3. TheVista-CAGISdatasetisstillundergoingreviewandrevision.Whilethisinformationwasusedtosupportthefollowinganalysis,relatedfindingsaboutsourceattributionmaychangeinphase2.

4. Airbornedatacollectionwon’tbecompleteuntilphase2andhencefindingsaboutthedistributionofsourcesinthisphase1interimreportareincomplete.

5. Giventhedegreeofvariabilityandepisodicactivityobservedformanysourcetypesanditispossiblethatthisinterimreportunder-estimatestheactualsourcepopulation.Thisisparticularlytruefordairiesandothersectorswithhighlyvariableprocesses.Follow-upobservationsofkeysectorsandareasinphase2ofthisprojectandrelatedNASAprojectsmayhelpfurtherconstrainthedegreeofvariabilityandprovideamorecompleteassessment.

6. Withafewexceptions,mostofthesourcesreportedinthisphase1interimreporthavenotyetbeenverifiedwithsurfacemeasurements.Thisprojectislimitedtoremotesensingmethodsandwasnotfundedtoconductfollow-upsurfaceverification.Thismeansthattherearesomeresidualuncertaintiesaboutsourceattributionthatcouldresultinmisidentificationoffacilitiesand/orincorrectassignmentofasourcetoagivenemissionsector.Additionaleffortwillbemadeinphase2toimproveattribution.

7. Giventheabovecaveatsitisprematuretomakedefinitivestatementsabouttherootcauseformostofthemethanesourcesdetectedinphase1.Inotherwords,atthistimeitcannotbedeterminedwhichsourcesarenormalprocessemissionssuchasperiodicventingasopposedtoaleakorothermalfunction.Afewexceptionsinphase1arenotedwherearoot-causewasconfirmed(throughsurfacefollow-upmeasurementsorthroughconsultationwithafacilityoperator).


20

4.1. AirbornesurveystatisticsTheactualimplementationofthephase1airbornesurveywasinfluencedbytheplanningactivitydescribedinSection3.3,responsetodiscoveryofmethaneplumes(e.g.,follow-upobservations),andimpactsduetoweatherandaircraftavailability.

4.1.1. SurveycompletenessThePhase1surveycoveredapproximately177,784distinctfacilitiesandinfrastructurecomponents(outof229,644candidates)spanning15,216km2oflandareaatleastonce(Figure3-1).Asignificantfractionoftheselineswereflownmorethanonce,resultingin23,176km2totalareacoverage.Approximately132hoursofflightswereconductedbetweenSeptember10andNovember4,2016.Table4-1summarizesflighthoursrelativetothethreebroadregionsshowninFigure4-1,northofFresno(white),betweenFresnoandLompoc(black),andsouthofLompoc(gray).Ofthe132totalflighthours,38hours(28.8%)wereusedtomaptheregionnorthofFresno,62.5(47.3%)betweenFresnoandLompoc,and31.5(23.9%)southofLompoc.FlighthoursarealsoorganizedinTable4-2byAVIRIS-NGflightboxesforthe3majorcategoriesofemissionsasshowninFigure3-6.

Figure4-1As-flownflightlines(green)forAVIRIS-NGFall2016survey.ThecoverageofthecompletedlinesisnotidenticaltotheboxesshowninFigure3-6duetomodificationstotheinitialflightplanandtheadditionoftwomini-intensiveflightcampaignstomaphighpressurenaturalgastransmissionlinesintheLABasinandanareanearVisalia.Regional


21

boxesindicatethreeregionsofstudy,northofFresno(white),betweenFresnoandLompoc(black),andsouthofLompoc(gray).

Table4-1Phase1flighthoursbyregions(seeFigure4-1)

Table4-2Phase1flighthoursbymajoremissioncategoriesreflectingtherelativefocusofCECandCARB:energy(e.g.,oilandgas),non-energysectors(e.g.,landfills,manuremanagement,wastewatertreatment,etc),andmixtureofthesecategories(seeFigure3-6)

ComparedwiththeVista-CAGISdatasetthephase1surveyachievedacompletenessperemissionsectorthatrangedfrom20to100%(seeFigure4-2andTable4-3).Notethatmostofthecategoriesshownhererepresentfacilitiesorotherdiscreteinfrastructurefeatureswiththeexceptionoftransmissionpipelines–aslinearfeaturesthelatterarereportedasfractionoftotallength.Also,forlandfillsthephase1surveyfocusedononlythelikelytopemitters–the60facilitiespredictedtoberesponsiblefor90%ofCalifornia’slandfillmethaneemissionsbasedonbottom-upestimatesfromCARB.

Region Flighthours PercentageNorthofFresno,CA 38.0 28.8%Center(FresnotoLompoc,CA) 62.5 47.3%SouthofLompoc,CA 31.5 23.9%Total 132.0 100.0%

Flightstargeting Flighthours PercentageEnergysector 56.5 42.8%Non-energysector 35.5 26.9%Mixtureofenergyandnon-energy 40.0 30.3%Total 132.0 100.0%


22

Figure4-2Phase1surveycompletenessbyinfrastructuretype

Table4-3Phase1surveycompletenessbyIPCCSector

Intermsoftemporalcompletenessthephase1surveysamplingrangedfromonevisitpersourcetomultiplevisitsdistributedoverSeptembertoNovember2016.Insomecases(e.g.,intensivestudyofdairiesnearVisaliaandsomestudiesofundergroundgasstoragefields)revisitintervalsasshortasafewminuteswereobtainedoverthecourseofaday,providinginsightintodiurnalvariability.Mostoftheoverflightsoccurredbetweenthehoursof10amand3pmlocaltime.

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

0.90

1.00

IPCC Sectors Methane Emitting Sector Total Features in Vista-CA

Features surveyed in phase 1

Fractional completeness

Petroleum Refineries 17 16 0.94Powerplants 468 166 0.35Transmission Pipelines (miles) 11,939 2,393 0.20Compressor Stations (excluding gas storage) 32 21 0.66Gas Storage Fields 12 12 1.00Distibution pipelines TBD TBDCNG Stations 334 93 0.28LNG Stations 46 17 0.37Oil & Gas Production 227,276 176,712 0.78

3A2 Manure Management Dairies, digesters, other livestock 1,709 851 0.504A1 Managed Waste Disposal Landfills (top emitters) 60 22 0.374D Wastewater Treatment & Discharge Wastewater Treatment Plants 152 43 0.28TBD unknown TBD TBD

229,621 177,771 0.77Totals

1B2 Oil and Natural Gas

1A1 Energy Industries


23

4.1.2. DistributionofmethanesourcesThelocationsofthe329confirmedpointsourcesareshowninFigure4-3,indicatingthatmostofthestrongpointsourcesdetectedinthissurveyareconcentratedinthesouthernhalfoftheState-particularlytheSoCABandareasintheSJVwiththelargestconcentrationsofdairiesandoil/gasfields.

Figure4-3Spatialdistributionofsources(redmarkers)andphase1flightlines(green).

ThedistributionofdetectedmethanepointsourcesbyIPCCemissionsectorissummarizedinFigure4-4andTable4-4andbyinfrastructuretypeinTable4-5.Notethatthereissignificantuncertaintyregardingtheassignmentofsourcestosector1A1(combustionrelatedactivity)andsector1B2(oilandgasfugitives)forrefineries.Thiswillbestudiedfurtherinphase2.Table4-4offerssomepreliminaryinsightintothepotentialtotalpopulationofpointsourcesintheState(e.g.,fractionofsampledinfrastructurewhereatleastonemethanesourcewasdetected).Additionalspatialand


24

temporalsamplinginphase2shouldhelpwithscalingsurveyresultstothetotalpopulation.

Figure4-4DistributionofsourcesbyrelevantIPCCemissionsectors

Table4-4DistributionofsourcesbyrelevantIPCCemissionssectorsandfractionoftotal

infrastructurepopulation(fromVista-CA)

SourceoccurencebyIPCCSector

1A1EnergyIndustries

1B2Oil&NaturalGas

3A2ManureManagement

4A1ManagedWasteDisposalSites

4D1WastewaterTreatment

unknown

IPCC Sectors Methane Emitting Sector

Petroleum RefineriesPowerplantsTransmission Pipelines (miles)Compressor Stations (excluding gas storage)Gas Storage FieldsDistibution pipelinesCNG StationsLNG StationsOil & Gas Production

3A2 Manure Management Dairies, digesters, other livestock4A1 Managed Waste Disposal Landfills (top emitters)4D Wastewater Treatment & Discharge Wastewater Treatment PlantsTBD unknownTotals

1B2 Oil and Natural Gas

1A1 Energy Industries

Number of CH4 sources

Fraction of Features with >=1 CH4

source15 0.940 0.000 0.001 0.053 0.2510 01 0.06

99 5.6E-04186 0.2219 0.860 04

329 1.9E-03


25

Table4-5providesmoresourcecategoriesthanTables4-3and4-4.Thisisparticularlytrueforoilandgasproductionfields.Alsonotethattable4-5alsoincludesan“unknown”category,indicatingthatsomesourceshavenotyetbeenconclusivelyattributedtoaknowninfrastructuretype.Oil/gasstackreferstoflaringstacksandothercombustionorventtowersassociatedwithoilandgasproduction.

Table4-5Distributionofsourcesbyinfrastructuretype.

Aspreviouslydiscussed,estimationofemissionrateswillberequiredtoprovidequantitativeassessmentsofthecontributionsofeachmethanesourcerelativetothepopulationofpointsourcesandtheStatemethaneinventory.Thatworkisplannedforphase2.MeanwhileFigure4-5plotsthemeanIntegratedMethaneEnhancement(IME)

Sourcetypes Sourcesdetected Percentagecropirrigation 3 0.9%dairy/manure 176 53.5%dairy/manure-digester 4 1.2%gascompressor 3 0.9%gasdistributionline 1 0.3%gasLNGstation 1 0.3%gasCNGstation 0 0.0%gasstoragefacility 3 0.9%landfill 19 5.8%oil/gascompressor 2 0.6%oil/gasdrillrig 2 0.6%oil/gasgatheringlines(TBD) 10 3.0%oil/gaspumpjack 33 10.0%oil/gasstack 3 0.9%oil/gastank 19 5.8%oil/gasunknowninfrastucture 28 8.5%oil/gaswastelagoon 1 0.3%otherlivestock 3 0.9%refinery 14 4.3%unknown 4 1.2%watertreatmentplant 0 0.0%Totalsources 329 100.0%


26

valuesestimatedfor304sources(robustIMEestimateshavenotyetbeenderivedfortheremaining25sourcesduetocomplicationssuchascontrastissues).IMEestimatesarenotdirectlyequivalenttoemissionratessincethelatterdependonwindspeedsthatvaryspatiallyandtemporally.However,thatvariabilityisminimizedheresomewhatbyaveraginggiventhatmanyofthesesourcesweresampledonmultipledatesandtimesduringthephase1campaign.HencethemeanIMEvaluespresentedhereofferasimplemethodforassessingtherelativestrengthofthepointsourcesdetectedinphase1.Thelogarithmiccurveinfigure4-5indicatesthat10%ofthesourcescontributenearly60%ofthetotalIMEinthesourcepopulation.AstudyofthesourcedatabaseinAppendixAindicatesthatthetop20sourcesincludeemittersfromeverysector.Thiswillbefurtherexploredinphase2.

Figure 4-5 Distribution of mean Integrated Methane Enhancements (IMEs) for most of the point sources detected in

phase 1 and their cumulative contributions to the total.

4.2. Sectorspecificfindings

Eachofthefollowingsectionsprovideexamplesofplumeimagesandfindingsforeachkeymethanepointsourceemissionsector.Theappendixofthisreportcontainsmanymoreexamplesbutthesearerepresentative.Asdiscussedpreviously,inphase2theseanalyseswillbeupdatedtoincorporatequantitativeemissionfluxestimates–towardsassessingtherelativeandnetcontributionofpointsourcestomethaneemissionsfromkeysectorsandregions.

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

0.01

0.10

1.00

10.00

100.00

1 11 21 31 41 51 61 71 81 91 101 111 121 131 141 151 161 171 181 191 201 211 221 231 241 251 261 271 281 291 301

%oftotal

Integrated

Methane

Enh

ancemen

t(kg)

MeansourceIME cumulaCve%oftotal


27

4.2.1. OilandgasproductionOver75%ofmajoroilandgasproductionfieldsinCaliforniaweresurveyedatleastonceinphase1.NotableexceptionswerefieldsintheSutterButtesandSantaMariaareafieldsthatwillbesurveyedinphase2.Generally,thevastmajorityofmethanesourcesassociatedwithoilandgasproductionappearinKerncounty.Inparticular,KernFront,ElkHillsandMidwaySunset-wheremostoftheState’soilandassociatedgasproductioniscurrentlyunderway–arenoteworthy.Onlyafewstrongmethanepointsourceshavebeendetectedsofarinotheroilandgasfieldssurveyedinphase1.InKerncounty,Storagetanksandwellheads/pumpjackswereresponsibleforthelargestfractionofobservedoilandgassources(Figures4-6and4-7).Manyofthesesourcesseemtopersistformonthsbasedonrepeatedflightsduringthephase1studyandearlierNASAairbornesurveysin2014and2015.Someoilandgasproductionsourcesmayincludeleaksingatheringlines.TenpotentialgatheringlinesourceswereidentifiedhoweverthesearechallengingtopinpointandvalidateandsoremainTBD.

Figure4-6.TypicalmethanesourcesinKernFrontoilfield.Commonsourcesincludestoragetanks,wellheadsand(potentially)gatheringlines.


28

Figure4-7CloseupoftypicalmethaneplumefromagasstoragetankinKernFrontoilfield.ThisplumehaspersistedovermultipleyearsincludingprecursorNASAcampaignssince2014.


29

Theprevalenceofmethaneplumesvariessignificantlybyoilandgasfield.ThePosoCreekandKernFrontfieldsexhibitedthehighestdensityofmethaneplumesintheState–bothdramaticallyhigherthanothernearbyoilfieldssuchasKernRiverandRoundMountain(Figure4-8).BasedonDOGGRdataforFall2016thereisnoobviouscorrelationwithproductionrates(e.g.,PosoCreekandKernFrontoilproductionrateswerelessthan20%theKernRiverproductionratesforthatperiod).Additionaldatacollectionandanalysiswillbeconductedinphase2tofurtherstudypotentialroot-causes.

Figure4-8Significantlyhigherdensitiesofmethanesources(redmarkers)wereobservedinPosoCreekandKernFrontoilfieldsthanothersineasternKernCounty.

PosoCreek

KernFront

KernRiver

RoundMountain


30

SomefacilitiessuchasthegasproductionplantandgasinjectionfacilityinElkHillsoilfieldpresenteduniquecombinationsofmethaneplumesassociatedwithcompressoroperationandflaringstacks(Figure4-9).Asnotedpreviously,attributionofsourcestocombustionvsfugitivescarriessomeresidualuncertaintyandwillbestudiedfurtherinphase2.

Figure4-9GasprocessingfacilityinElkHillsshowingmethaneplumesfromtwoofthreelargecompressors(top)andoneormoreflaringstacks(bottom).

4.2.2. Naturalgastransmission,storageanddistribution

Longlinearfeaturessuchasgastransmissionpipelineswerenotapriorityinphase1.Onlyabout20%oftransmissionlinesintheStatewerecovered–primarilylimitedtoathreedayintensiveassessmentofthenaturalgassupplychainintheSouthCoastAirBasin.Nomethanesourcesdirectlyassociatedwithtransmissionpipelineleaksweredetected.Fortransmissionlinecompressorstationsassociatedonlyoneexhibitedanobservablemethaneplumeduringoverflights.Additionalattentionmaybegiventothetransmissionsectorinphase2.


31

EachoftheState’s12activeundergroundgasstoragefacilitiesweresurveyedatleastonceinphase1.Someweresurveyedmultipletimes,particularlyHonorRancho,MacDonaldIsland(Figure4-10),andAlisoCanyon.AlisoCanyonwasinastandbystateduringthephase1campaignandwhilethereweresignsofmethaneintheareanoobviousplumeswerepresent.HonorRanchoandMacDonaldIslandpresentedrelativelylargeplumes.BothapersistentandepisodicsourcewereobservedatHonorRancho.Thepersistentsourcewaspotentiallyassociatedwithaleakingbypassvalveandemergencyshutdownstack.Theepisodicsourceappearedtoberelatedtocompressoractivityduringgasinjection.Plumeswereobservedatafewoftheotherstoragefacilitiesbuttheywererelativelysmallandbarelydetectable.Giventheobservedepisodicnatureofmethaneemissionsatstoragefacilitiesadditionalattentionwillprobablybedevotedtothissectorinphase2.

Figure4-10ExampleofmethaneplumeatMcDonaldIslandgasstoragefacility,south

complex.

California’snaturalgasdistributioninfrastructurespansseverallargeurbanareas.Thephase1surveyprimarilyfocusedonpriorityareasintheSouthCoastAirBasin.Asingleleakinalowpressuregasdistributionlinewaspin-pointedinthevicinityofChinoHillsfollowingdetectionfrompersistentregionalmonitoringbytheMegacitiesCarbonProject‡.Figure4-11illustratestheAVIRIS-NGsearchpatterncoveringa60km2areain30minutes,real-timedetectionwiththeonboardsoftware,anddeterminationofthesourcelocationtowithin10meters.Thegascompanywasnotifiedanddispatchedtechnicianstothesitewhoconfirmedandrepairedtheleakwithin24hours.Phase2willlikelydevoteadditionalattentiontodistributioninfrastructureinotherurbanareas.‡ https://megacities.jpl.nasa.gov/portal/


32

Figure4-11DetectionofleakingasdistributionlineinChinoHills.Left:AVIRIS-NGflightpattern,middle:real-timedetectionsoftwareonairplane,right:processedmethaneplumeimageandgeolocationofsourcetowithin10meters.

4.2.3. Refineries

Strongmethaneplumeswereobservedatnearlyeveryrefinerysampledinphase1.Thereappearstobeadiversesetofsourcesatrefineries–rangingfromstoragetanks(eitherventingfromreliefvalvesorleaks)touncombustedmethanefromvariousprocesses–seeFigure4-12.Notethatthereissignificantuncertaintyregardingtheassignmentofsourcestosector1A1(combustionrelatedactivity)andsector1B2(oilandgasfugitives)forrefineries.Thiswillbestudiedfurtherinphase2.

Figure4-12ExamplesofmethaneplumesfromrefineriesintheLAbasin.(Left)ventingorleakingstoragetanks,(Center)unknownsource(TBDinfrastructure),(Right)incompletecombustionfrom3flaringstacks.Notethesmallerconcentrationscalefortheimageontheright.


33

4.2.4. PowerPlants

Despitesurveying166powerplants(over30%oftheState’sfossilfueledgeneratingstations)therewasnoevidenceoflargemethaneplumesfromthissector.Thismaybeexpandedinphase2tocovermorepowerplantsincludingcogenerationfacilitiesState-wide.

4.2.5. LandfillsIn order to prioritize flight hours for this sector, CARB’s database of landfill methaneemissions(over370facilities)wasusedtoidentify60likelyhighestemitters,collectivelypredicted tocontribute90%of landfillemissions inCalifornia.22of these facilitiesweresurveyedatleastonce(severalweresurveyedmultipletimes)inphase1.Methaneplumeswereobservedat all but twoof the22 surveyed facilities. A smaller fractionof landfillsexhibitedverylargemethaneplumes.Figure 4-13 illustratesmultiple strongmethane plumes at the Sunshine Canyon landfill.Phase1 flights identifiedstrongmethaneplumesthatpersisted insectionsof this facilityover several months. While these features are observed at a subset of other landfills,SunshineCanyonconsistentlypresentedthelargestplumesunderallconditions.

Figure 4-13 Four persistent methane plumes at Sunshine Canyon landfill. Left: rawmethane images(whitepixels indicatehighermethaneenhancements) forSeptember11,2016,center:sameforOctober3,2016,right:satelliteimageofthefacility.Thevariabilityin plume shape is due to changing wind conditions. Red markers indicate the averagelocationofthe4sourcesobservedatthefacility.


34

Inphase1therewasagenerallackofagreementbetweenCARBbottom-upestimatesofindividuallandfillemissionratesinSouthernCaliforniacomparedtothemeansourceenhancementsestimatedbyAVIRIS-NG.Asimpleanalysisfor10landfillsintheSoCABpredictedtobehighemittersispresentedinFigure4-14.TheabsenceofabluebaratagivenfacilityindicatesthatAVIRIS-NGdidnotdetectanystrongmethaneplume.OneexplanationfortheobserveddisagreementatPuenteHillslandfillisthatfacilitywasshut-downbetweentheCARBbottom-up2014estimateandtheFall2016AVIRIS-NGoverflights.MethaneplumeswereobservedatPuenteHillslandfillduringaNASAairbornestudyinJuly2014.Figure4-14doesnotprovideadirectcomparisonofemissionsandhencethebluebarsshouldnotbeinterpretedtomeanthatmeasuredemissionsin2016werelowerthaninthe2014CARBestimate.

Figure4-14Relativeassessmentof10landfillsinSoCalpredictedtobehighmethaneemitters:comparingCARB2014bottom-upemissionsestimates(unitskgCH4/hr)foreachfacilityandFall2016AVIRIS-NGmeanintegratedmethaneenhancementestimates(unitskg)forobservedplumes.Notethecommonlogarithmicscale.Thesequantitiesarenotdirectlyequivalent.Inparticular,untilemissionestimatesarederivedfromtheAVIRIS-NGenhancementmeasurementsitwouldbeprematuretointerpretthisplotasmeaningactualemissionsarelowerthanthe2014CARBestimates(infact,theoppositeismorelikely).

FewerflightswereconductedovernorthernCaliforniainphase1thanplanneddue

tocloudyconditionsthereinFall2016andotherfactors.ThistranslatedtoreducedcoverageforlandfillsandwastewatertreatmentfacilitiesintheSanFranciscoBayArea.SubsequentflightsovertheseareaswereconductedinJune2017aspartofaNASAairbornecampaign.Aneffortwillbemadeinphase2toincorporatethosenewdatasetsintothisanalysisincludingquantitativeestimatesofemissionrates.

1

10

100

1000

PuenteHillsLF

SunshineCanyon

OlindaAlphaSLF

FrankR.Bowerman

LopezCanyonLF

SchollCanyonLF

ChiquitaCanyon

BKKWestCovina(ClassIandIIILFs)

BradleyAveEast&West

TolandRoad

AVIRIS-NGenhancement(kg) CARBemissions(kgCH4/hr)


35

4.2.6. WastewatertreatmentWastewatertreatmentfacilitiesweresurveyedinphase1butprimarilylimitedto

theSoCABwherenomajormethaneplumeswereobserved.Thismayberevisedinphase2pendinganalysisofdatacollectedduringtheaforementionedNASAJune2017campaign.

4.2.7. DairiesandlivestockLivestockmanuremanagement–particularlyintheSanJoaquinValley(SJV)–isrecognizedasoneofthetopmethaneemissionsectorsinCalifornia.OurPhase1surveyresultsareconsistentwiththisgiventhatwetmanuremanagement–particularlysettlingpondsandanaerobiclagoons–isresponsibleforover50%ofmethanepointsourceplumesobservedinthisstudy.Thephase1studycoveredabout850knowndairiesintheState–abouthalfofthetotalpopulation.ArobustassessmentoftheindividualandnetemissionsfromdairiesandotherlivestockfacilitiesinCaliforniaiscomplicatedbyseveralfactors.Figure4-15indicatesonesuchfactor:thecomplexspatialgradientsofnear-surfaceatmosphericmethanethatmanifestsinportionsoftheSJVinresponsetothedenseconcentrationofemissionsources(largedairies)and/ortheeffectsof“pooling”fromwindandothermeteorologicalvariables.Thisfigureraisesthequestion:whyweren’tmethanesourcesdetectedatmoredairies?Detectingandattributingmethaneplumestoindividualpointsourcescanbechallenginginthepresenceofstrongmethaneenhancementsoverlargeareas–essentiallya“contrast”problem.Insuchareathereisariskbothofover-estimatingtheemissionsofindividualdairies(byconvolvingthefluxwithnearbyfacilities)andalsounder-estimatethenetemissionsofthearea.Thisrepresentsanactiveareaofmeasurementscienceandisapriorityforfutureattention.


36

Figure4-15 MosaicoftwodaysofAVIRIS-NGflight linesoverTularecountydairies.Therawgrayscaleimageoverlaysrepresentareaswithlower(black)andhigher(white)levelsof atmospheric methane. The striking gradient seen here suggests either a denseconcentrationofemissionsourcesaroundWaukenaandGoshenand/oranaccumulationofenhanced levels of methane in these areas due to meteorological effects. Atmospherictransportmodelingwilllikelyberequiredtodisentanglethoseeffects.Bluesquareindicatethe known locations of dairies. Red markers indicate methane point sources detectedduringtheseoverflights.

Another complexity involves the inherent variability of dairy methane emissionprocesses. The primary driver for methane point source emissions from manuremanagement involves the use of water and anaerobic conditions that promotemethanogenesis.Dairiesaredynamicfacilitiesinthatwaterandwastesaremovedaroundeachfacilityoverthecourseofthedayonagivenduty-cycle,translatingtomethanepointsourcesthatcanvarysignificantlyontime-scalesofhours–asanaerobiclayersinlagoonsare disturbed and as methane laden water is transported around the facility includingirrigationforadjacentfields.

10km


37

Figure4-16 Plotofmethanepoint sourcevariability foran intensivestudyof50dairiesnear Tipton. AVIRIS-NG repeated the same flight lineswith a roughly 40minute revisitintervalperlineovera5hourperiod.Thecolorsindicatethenumberoftimesasourcewasobserved during that period. Some of the sources were persistent – others were morevariable.

This diurnal, management-driven variability is likely somewhat independent of

seasonalvariability inemissionfluxesdrivenbychangingtemperatures. Thisshort-termvariabilitycanhaveanimpactondetectabilityasillustratedinFigure4-16and4-17(e.g.,surveyswithinsufficientrevisitfrequencycanfailtodetectsourcesthroughaliasing).


38

Figure4-17Close-upofadairyinfigure4-15.Eachimageshowsmethaneplumesforsnap-shotintime,eachseparatedbyabout40minutes-indicatingsignificantdiurnalvariabilityinemissions.

Methane digesters are increasingly being deployed at California dairies in an effort toreduce the net greenhouse gas impact of each facility while offering additional revenueopportunitiessuchasbiogasforenergyproduction.Thephase1surveycoveredabout22knowndairy digesters in the State including a combination of facilities in operation andstill undergoing construction. In principle a well-functioning digester should capturemethanefrommanuremanagementhoweverourphase1studyindicatedthepresenceofsignificantmethanepointsourcesatfourfacilitiesintheSJV.Figure4-18showsanexampleof a persistentmethane plumes at a dairy digester. The biogas operator for this facilityindicated that the cause was likely manual venting during maintenance activity. Thissuggeststhatfuturemonitoringforatmosphericmethanearoundthesefacilitiesbeforeandafter digester construction could prove useful for assessing their efficacy in meetingmitigationobjectiveswhilehelpingoperatorsavoidunintentionalbiogasproductloss.

103.1kg/hr 51.6kg/hr 137.6kg/hr

69.3kg/hr 59.8kg/hr


39

Figure4-18Exampleofmethaneemissionsplumeobservedrepeatedlyatmethanedairydigester in the SJV. The difference in plume appearance between the two dates isattributedtodifferentwindspeeds.

4.3. SouthCoastAirBasinmethanetrendsandvariability

As described in section 3.5, JPL’s CLARS facility on Mt Wilson provides persistentmonitoring ofmethane fluxes in the SouthCoastAirBasin. TheCLARS tracer-tracer fluxestimationmethod provides basin-scale,monthly averagedmethane emission estimates.This project contributed to extending the established CLARS time series (2011-2015)throughspring2017.

TheCLARSdataproducts for thisworkconsistof text filespostedon the JPLMegacitiesCarbon Project data portal§. There is a read_me file which discusses the format andinterpretationofeachdatafield.

§ https://megacities.jpl.nasa.gov/public/Los_Angeles/Remote_Sensing/CLARS_RRNES/


40

Figure4-19showstheextendedtimeseriesofmonthlymethaneemissionscalculatedusingeq.(9)[section3.5].Shadedareasrepresentthe1σuncertaintiesinthederivedmethaneemissions.UncertaintiesarepropagatedfromtheuncertaintiesinCLARSXmethane(XS):XCO2(XS)regressionslopesandCO2emissions.ForCO2,weassumeda10%uncertaintyintheHestiamonthlyCO2emissions.Fig.4-20plotstheseasonablevariabilityofCLARS-FTSinferredmethaneemissionsforeachyearfrom2011-2016.AsnotedinWongetal(2016),whilethechoiceofCO2inventorycanaffecttheestimatedmethaneflux,themethane/CO2regressionslopeRisapurelymeasuredquantityanditshowsconsistentseasonalvariability.AlsoasnotedinWongetal.(2016),theinferredmethaneemissionsestimatesshowabimodaldistributionwithpeaksduringthesummerandlatesummer/earlyfall.AstrikingfeatureinFigure4-19istheapparentabsenceoftheAlisoCanyongasleak’simpactontheSoCABmethaneflux.Infact,thebasinfluxdidincreasesignificantlyandconsistentlywiththeknownAlisoCanyonleakratehoweverthenormalseasonalvariabilityintheSoCABmakestheimpactdifficulttoseeatthisscale.TheAlisoCanyongasleakeffectivelyincreasedthedurationofthenormalwinterspikeinSoCABmethaneemissionsbyabouttwomonths.

Figure4-19CLARStime-seriesofSoCABmonthlyaveragemethaneemissionsfrom9/2011-3/2017.Thegrayshadingindicatesuncertaintybounds.TheredboxcoverstheperiodoftheAlisoCanyongasleakinfall2015-winter2016.TheblueboxprovidesareferenceforcomparingtheAlisoCanyonincidentwiththefall/winterspikein2013-2014.


41

Figure4-19ComparisonofSoCABmethaneseasonalvariabilitybyyear.

Fig.4-20showstheannualSoCABmethaneemissionsfrom2011-2016.Thecentralvaluesforeachyearshowasmalldecreasingslopebeginningin2013butthetrendisnotstatisticallysignificantwithintheuncertaintiesoftheemissionsestimationmethod.


42

Figure4-21CLARStime-seriesofSoCABannualmethaneemissions.

Additionally,CLARS’sdaily,spatiallyresolvedscansofXmethane/XCO2correlationratiosacrossmuchoftheSoCABdemonstratedtheabilitytodirectlydetecttheatmosphericimpactoflargeanomalousmethaneemissionsourcessuchastheAlisoCanyongasleak.Figures4-22and4-23provideexamplesofCLARSmethanemapsduringandfollowingthatincident.FutureimprovementsinCLARSdataprocessingandanalysisframeworkscouldenabletheabilitytorapidlydetectandlocatelargemethanehotspotsintheSoCAB.Suchacapability,ifcombinedwithrapid-responseairborneimagingspectroscopysuchasAVIRIS-ngtopreciselygeolocatesources,couldbeimportantincaseswheresuddenleakonsetoccursinremoteorunmonitoredlocations(and/orincasesinvolvingmethanethathasnotbeentreatedwithanodorizingagent).


43

Figure4-22CLARSmapofXmethane/XCO2correlationratiosformeasurementcycle3(earlyafternoon)onDecember25,2015.Thehigh(red)valuesindicateaplumeofmethaneextendingfromtheSanFernandoValleyacrosstheLAbasin.TheyellowstarindicatesthelocationoftheAlisoCanyongasleak(wellSS25)source.


44

Figure4-23Sameasfigure4-18butforearlyafternoononFebruary8,2016(top)andFebuary11,2016(bottom).TheAlisoCanyonleakwaspluggedonthemorningofthe11th.


45

5. ConclusionsJPLisapplyingadvancedremotesensingmethodstodetectandcharacterizeanthropogenicmethane(methane)emissionsinCaliforniatosupporttheState’sobjectivesformitigationofshort-livedclimatepollutants(California2013,2017),identificationofmethane“hotspots”inresponsetoAB1496(California2015),andsupportingnaturalgasleakdetectionandcorrectionforrate-payerbenefit.Phase1ofthefirstcomprehensivesurveyofmethanepointsourcesinCaliforniahasbeencompleted.Allobjectivesforphase1havebeenmet.Phase1primarilyuseddatacollectedin2016andaddressesCARBprioritiesspanningIPCCmethaneemissionsectorsrelevanttopointsourcesinCalifornia.Phase2willcollectdatain2017andfocusonCECpriorities,particularlythenaturalgassector,toimproveunderstandingofleaksandtohelpenablemitigation.ThisinterimreportsummarizesthePhase1activityandfindingsincludinglessonsthatwillinformdatacollectionandanalysisstrategiesforPhase2aswellasfutureresearchprojects.AfinalreportwillbeproducedattheconclusionofPhase2.Significant,preliminaryinsightshavebeengainedinphase1regardingthedistributionofmethanepointsourcesandsomeofthecontrollingprocesses.InadditiontothefindingssummarizedintheExecutiveSummarysomemajorlessons-learnedinclude:

1. Moreattentionisrequiredtoaddressthechallengeofepisodicactivityinkeymethaneemissionsectionsincludingrepeatedairborneremote-sensingsurveysandpersistentregional-scalemonitoring.Thisisparticularlytruefordairyemissionsandwilllikelyrequireadditionalresearchbeyondthescopeofthisprojecttoanswerkeyquestionsaboutrelativeandtotalemissionfluxesandcontrollingprocesses.

2. Futureadvancesinimagingspectroscopyhavethepotentialaddressareasources(e.g.,entericfermentation,ricecultivation,andwetlands)aswellasthepointsourcesdescribedhere.Inparticular,improvingthespectralresolutionofanAVIRIS-NGclassimagingspectrometerfromthecurrent5nmto1nmcouldenablesuchadvances[Thorpeetal2016b].

3. Thisprojecthasdemonstratedtheabilityofregionalscalemonitoringsystemstodetectthefootprintoflargeanomalousmethaneemissionsandofairborneimagingspectrometerstofindandpinpointrelativelysmallleaksinnaturalgasinfrastructure.Futureimprovementsinmeasurementandanalysisframeworkscouldsupportoperational,rapid-responseversionsofsuchsystems.

Asplanned,additionaldatacollectionandanalysisinphase2oftheprojectwillbeperformedtovalidatetheinterimfindingsandaddresstheopenquestionsreportedhere.Thatphase2activityisplannedtoincludeanother6-8weekairbornecampaigninFall2017followedbyanalysisandinterpretation.ItmayalsoincorporatedataandfindingsfromrecentNASAfundedresearchcampaignsinCalifornia.Akeynewelementforthephase2reportwillbeinclusionofemission/fluxestimates.


46

6. References

Butz,A.(2012),TROPOMIaboardSentinel-5Precursor:Prospectiveperformanceofmethaneretrievalsforaerosolandcirrusloadedatmospheres,Rem.Sens.Environ.120(2012)267–276,doi:10.1016/j.rse.2011.05.030California(2014)SenateBillSB605Short-LivedClimatePollutantStrategy.

California(2015)AssemblyBillAB1496MethaneMitigationAct.California(2017)ShortLivedClimatePollutantStrategy(2017),https://www.arb.ca.gov/cc/shortlived/meetings/03142017/final_slcp_report.pdf

Carranza,V.,Rafiq,T.,Frausto-Vicencio,I.,Hopkins,F.,Verhulst,K.R.,Rao,P.,Duren,R.,andMiller,C.(2017):Vista-LA:MappingmethaneemittinginfrastructureintheLosAngelesmegacity,EarthSyst.Sci.DataDiscuss.,https://doi.org/10.5194/essd-2017-65,inreview.Conleyetal(2016),Methaneemissionsfromthe2015AlisoCanyonblowoutinLosAngeles,CA,Science.

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 remotemeasurements reveal heavy-tail flux distribution in Four Corners region. Proceedings oftheNationalAcademyofSciences,p.201605617.

Fu,D.,T.J.Pongetti,J.F.L.Blavier,T.J.Crawford,K.S.Manatt,G.C.Toon,K.W.Wong,andS.P.Sander(2014),Near-infraredremotesensingofLosAngelestracegasdistributionsfromamountaintopsite,AtmosMeasTech,7(3),713-729,doi:10.5194/amt-7-713-2014.

Jeong,S.,etal.(2017),Estimatingmethaneemissionsfrombiologicalandfossil-fuelsourcesintheSanFranciscoBayArea,Geophys.Res.Lett.,44,486–495.Doi10.1002/2016GL071794Kirschke,S.,etal.(2013)Threedecadesofglobalmethanesourcesandsinks.Nat.Geosci.6,813-823.

KortEA,etal.(2014)Fourcorners:ThelargestUSmethaneanomalyviewedfromspace.GeophysResLett41(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 emissionsfrom a Californian landfill, determined from airborne remote sensing and in-situmeasurements.AtmosphericMeasurementTechniquesDiscussions.

Ryerson,T.B.,etal.(2013),The2010CaliforniaResearchattheNexusofAirQualityandClimateChange(CalNex)fieldstudy,J.Geophys.Res.Atmos.,118,5830–5866,


47

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. andLuna,B., 2015.Real-time remotedetectionandmeasurementforairborneimagingspectroscopy:acasestudywithmethane.AtmosphericMeasurementTechniques,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.andUngar,S.,2016.Space-basedremoteimagingspectroscopyofthe Aliso Canyonmethane superemitter. Geophysical Research Letters, 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).MappingmethaneconcentrationsfromacontrolledreleaseexperimentusingthenextgenerationAirborneVisible/InfraredImagingSpectrometer(AVIRIS-NG).RemoteSensingofEnvironment.Thorpe,A.K.,Frankenberg,C.,Green,R.O.,Thompson,D.R.,Aubrey,A.D.,Mouroulis,P.,Eastwood,M.L.,andMatheou,G.(2016b),TheAirborneMethanePlumeSpectrometer(AMPS):Quantitativeimagingofmethaneplumesinrealtime,2016IEEEAerospaceConference,doi:10.1109/AERO.2016.7500756.Spectrometer(AVIRIS-NG).RemoteSensingofEnvironment,179,pp.104-115.

Turner,A.J.,etal(2015),EstimatingglobalandNorthAmericanmethaneemissionswithhighspatialresolutionusingGOSATsatellitedata,Atmos.Chem.Phys.,15,7049-7069,doi:10.5194/acp-15-7049-2015.Turner,A.J.,D.J.Jacob,J.Benmergui,S.C.Wofsy,J.D.Maasakkers,A.Butz,O.Hasekamp,andS.C.Biraud(2016),AlargeincreaseinU.S.methaneemissionsoverthepastdecadeinferredfromsatellitedataandsurfaceobservations,Geophys.Res.Lett.,43,doi:10.1002/2016GL067987.Wecht,K.J., Jacob,D.J.,Sulprizio,M.P.,Santoni,G.W.,Wofsy,S.C.,Parker,R.,Bosch,H.,andWorden,J.(2014)SpatiallyresolvingmethaneemissionsinCalifornia:constraintsfromtheCalNex aircraft campaign and from present (GOSAT, TES) and future (TROPOMI,geostationary)satelliteobservations.Atmos.Chem.Phys.Discuss.14,4119-4148.

Wennbergetal(2012),OntheSourcesofMethanetotheLosAngelesAtmosphere,Env.Sci.Tech.

Wong,K.W.,Pongetti,T.J.,Oda,T.,Rao,P.,Gurney,Kevin.R.,Newman,S.,Duren,R.M.,Miller,C.E.,Yung,Y.L.,andSander,S.P.:MonthlytrendsofmethaneemissionsinLos


48

Angelesfrom2011to2015inferredbyCLARS-FTSobservations,Atmos.Chem.Phys.(2016),doi:10.5194/acp-2016-232Wunch,D.,Toon,G.C.,Hedelius,J.K.,Vizenor,N.,Roehl,C.M.,Saad,K.M.,Blavier,J.-F.L.,Blake,D.R.,andWennberg,P.O.(2016):QuantifyingthelossofprocessednaturalgaswithinCalifornia'sSouthCoastAirBasinusinglong-termmeasurementsofethaneandmethane,Atmos.Chem.Phys.,16,14091-14105,https://doi.org/10.5194/acp-16-14091-2016,2016.

introduction to the phase i report of the california ... · proposal. an additional action recently...

Documents