1

February17,2017

ClerkoftheBoardCaliforniaAirResourcesBoard1001IStreetSacramento,California95814

KairosAerospaceCommentsontheModifiedTextfortheProposedRegulationforGreenhouseGasEmissionStandardsforCrudeOilandNaturalGasFacilities

WeatKairosAerospacecommendtheARBonitscommitmenttocurbmethaneemissionsfromthe oil and gas industry. The proposed regulation is one of the strongest,with the broadestapplicability,inthenation.Wewould,however,respectfullyrecommendamodificationtotherulethatwouldachievesimilarmethanereductionsatalowercosttotheregulatedindustry.WefocusourcommentsonLDARforfugitivemethaneemissions.

The fugitive methane emissions detection technology landscape is highly dynamic, withinnovationhappeninginrealtime.Innovationintechnologyinvariablyleadstobetteroutcomesatlowercosts.However,thecurrentdraftisprescriptiveintermsofLDARtechnology,sacrificingthe opportunity for commercialization of new technologies that could both improve overallenvironmental outcomes and lower costs. The inflexible nature of the LDAR regulation alsoimpactsdifferentoperatorsdifferently,assignificantvariationscanexistinoperators’facilities,emissions,andcostsofcontrol.

ARBshouldallowoperatorstoseekapprovaltoallowalternativecompliancepathways,aslongasthealternativeisatleastaseffectiveinreducingmethaneemissionsvolumeasquarterlyOGI-based LDAR. ARB should also establish clear criteria bywhich the equivalence of alternativeprogramswillbejudged,andensurethattheapprovalprocessistransparentandopentopublicparticipation.OnlybyallowingspacefornewandinnovativetechnologiescanARBensurethatitisachievingitsgoalofmaximumenvironmentalimpactatminimalcost.

ManyOptionsExist,andMoreAreinDevelopment

AsARBknows,thereisalreadyawiderangeofproventechnologiesandimplementationsonthemarket,withdifferentdetectionlimits,frequencies,andunderlyingscience.ARBhasundertakenprojectswithavarietyof instruments– identifyingmethanehot-spots, implementingatieredmethaneobservationsystem,usingaerialsurveysforAlisoCanyon-relatedmonitoring.1,2ARBrecentlypostedaresearchreport,“EnhancedInspection&MaintenanceforGHGandVOCsatUpstreamFacilities,”3thatevaluatestheeffectivenessofsixdifferentinstrumentsandfindsthethreeremotesensinginstruments(RMLD,IRcamera,andPicarroSurveyor)tobesimilarintheir

1“CaliforniaMethaneMonitoringforClimateActionandPublicSafety,”presentedbyRileyDuren.June2016.https://www.arb.ca.gov/cc/oil-gas/Miller_2016-06-14%20-%20Duren_methane_hot_spots.pdf2“AirborneEstimationofSurfaceEmissions,”presentedbyStephenConley.June2016.https://www.arb.ca.gov/cc/oil-gas/Conley_Presentation_ARB%20%281%29.pdf3“AirResourcesBoardRFPNo.13-414:EnhancedInspection&MaintenanceforGHG&VOCsatUpstreamFacilities–Final(Revised),”preparedbySageATCEnvironmentalConsultingLLCfortheCaliforniaAirResourcesBoard.December2016.https://www.arb.ca.gov/cc/oil-gas/sage_i&m_ghg_voc_dec2016.pdf

2

ability todetectemissions.ARB isalsoapartof the ITRC’sEvaluationof InnovativeMethaneDetectionTechnologiesteam,whichisworkingtoevaluateandcomparedifferenttechnologies,andhasparticipatedinseveralworkshopsandconferencesdiscussingwaystoaddressfugitivemethaneemissions.TheseeffortsindicatetousthatARBisawareoftheplethoraofoptions,andthewaystheycanbecombinedtoyieldbetterandmorecost-effectiveresults,yettheregulationitselfdoesnotreflectthisunderstanding.

Inadditiontotheseexistingoptions,thereareseveralinnovativetechnologiesatdifferentpointsalongthespectrumfromconcepttocommercialization.Theseoptionsallhaveadifferentsetofmonitoring frequencies and technical capabilities. On one end of the spectrum there aretechnologies still in development, for example in ARPA-E’s MONITOR program and EDF’sMethaneDetectorsChallenge.Ontheotherend,forexample,thereisKairosAerospace,whichhas patented and newly commercialized an aerial-based methane imaging system calledLeakSurveyorthatleadstogreatermethanereductionatalowercostthantheOGIrequiredbythecurrentdraftregulation.Oursystemallowsoperatorstoconductlow-cost,high-frequencysurveysoftheirfieldstofindandfixlargeemittersfaster.Weimagemethaneinfalsecolorfromfixed-wing aircraft, similar toNASA JPL’s AVIRIS system, combinedwith simultaneous opticalimagery, toproducegeoreferencedopticalmapswithmethaneplumespinpointedandsized.Oncewescreenalargeareaandidentifytheemissionsources,targetedgroundcrewsfollowupto identifyandrepair thespecificcomponentthat is leaking.Thisapproachresults ina loweroverallcosttooperators,ascomparedtosendinggroundcrewstoeverysite.Andmorefrequentsurveysforlargeleaksresultinagreaterenvironmentalimpactthaninfrequentgroundsurveys.

It is, of course, critical that any alternative LDAR program achieves at least the sameenvironmentalimpactastheLDARprogramlaidoutintheregulation.ComparingdifferentLDARprogramsispossiblethroughcomputermodels,combinedwithreal-worlddemonstrations.

EquivalentEnvironmentalImpactistheImportantThing

Atabasiclevel,theemissioncontroleffectivenessofanyLDARprogramisafunctionofboththeabilityofthetechnologyusedtodetect leaksandthefrequencyofmonitoring.Anequivalentprogrammayrequiremorefrequentmonitoring,ifitsmassratethresholdfordetectingleaksishigher,becausehighermassemissionsreductions fromlarge leaks foundearlierareoffset tosomedegreebysmallerleakswhichgoundetected.Indeed,thisisreasoningputforthbytheEPAin2006inaproposedamendmenttoallowOGIasanalternativeworkpracticetoMethod21,4andextendseasilytoarangeofmonitoringinstruments.Thisreasoningisalsousedina2004AmericanPetroleumInstitutereport5,whichstates:

“Lowerleakdefinitionsforrepairdonotnecessarilyleadtobetteremissionscontrolsince,astheleaklevelisdecreased,fewadditionalleakingcomponentsareaddedtotherepairgroupand these contribute very little to theoverallmass emissions. The Smart LDAR

4EPAAlternativeWorkPracticetoDetectLeaksfromEquipment,71FR17401(April6,2006)(tobecodifiedat40CFR60).5“SmartLeakDetectionandRepairforControlofFugitiveEmissions,”preparedbyICFInternationalforAmericanPetroleumInstitute.June2004.

3

approach…focuseson identifyingandrepairing thehighest leakerssince theseare thesourceofalmostallthemassemissions.Equivalenceisobtainedbymorequicklyfindingand repairing these large leaks, which more than offsets the emission rates fromcomponents with low ppmv readings that leak for longer periods. The use of opticalimaging could provide a more cost effective approach to more quickly find the highleakers.”

Usingacomputermodel(alongwithlaboratorytestingandfieldvalidationofinstruments)solvesmanyofthereal-worlddifficultiesofcomparingtheequivalencyofdifferenttechnologiesinthefieldovertime.ThisapproachwasalsosupportedbytheEPAinits2006amendmentmentionedabove,andtheAPIinits2004report,indicatingbroadstakeholderagreementonitsusefulnessandvalidity.ItisthereforealsotheapproachweatKairosusetoarriveatouremissionsreductionestimates.Weuseanopen-sourcemodelcalledtheFugitiveEmissionsAbatementSimulationTestbed(orFEAST).6ThismodelsimulatesnaturalgasleakageovertimeunderdifferentLDARprograms.Weusedapower-lawdistribution7fittedtothe2011FortWorthAirQualitySurvey8datatogeneratetheleaksfromthegasfield.Wethenmodeledtheaveragereduction(relativetoanullscenario)inmethaneemissionswithquarterlyOpticalGasImagingsurveysandsemi-monthlyAerialMethaneImagingsurveys(whichinthiscaseisKairosAerospace’sLeakSurveyortechnologywithaminimumdetectionlimitof120g/hror500ppm-m).

Wefindthatsemi-monthlyLeakSurveyorsurveysresult in87%reductionofmethaneemittedandquarterlyOGIsurveysresultin76%reductionoveranullscenario9overfiveyears.

6FEASTDocumentation:https://pangea.stanford.edu/researchgroups/eao/sites/default/files/FEASTDocumentation_0.pdf7SeeAppendixB.8CityofFortWorthNaturalGasAirQualityStudy.EasternResearchGroupetal.fortheCityofFortWorth,2011.URL:http://fortworthtexas.gov/uploadedFiles/Gas\_Wells/AirQualityStudy\_final.pdf9Thenullscenarioismodeledasoperatorsrandomlynoticingandfixingleaksinthenormalcourseofoperations,withnoexplicitLDARprogram.

0%

20%

40%

60%

80%

100%

1Year 5Years 10Years

%CH4

Red

uced

MethaneReducedOverTimebyDifferentLDARPrograms

AerialMethaneImaging(semi-monthly) OpticalGasImaging(quarterly)

4

Therefore, LeakSurveyor would be a more effective LDAR program than OGI, which is anapproved technology in the current regulatory draft. (See Appendix A for more detail onLeakSurveyor’s capabilities, and Appendix B for more detail on FEAST.) Incidentally, theLeakSurveyorprogramisalsocheaperthantheOGIprogram,despitetheincreasedfrequencyofsurveys, since each LeakSurveyor survey is an order of magnitude less expensive than aconventionalOGIsurvey.LeakSurveyoralsooffers leakratequantification,direct imagingandgeoreferencedresults,andanauditablerecordofsurveysandresults.

The same FEASTmodel can be used to compare other technologies as well, by adding newmodulestotheprogram.Thusanoperatorcanchoosethemostcost-effectiveLDARprogrambased on their own needs and constraints, and ARB can estimate whether the proposedalternative compliance programwill reduce emissions from an operator at least asmuch asquarterlyOGI,ultimatelyresultinginmoreefficientregulation.

Conclusion

In LDAR, the current ARB draft actually falls behind EPA’s NSPS OOOOa regulations, whereinsteaditshouldbeanopportunityforARBtoleadtheEPA.CaliforniaishometoSiliconValleyinnovationandalong-timeleaderinairqualityregulation,yettheserulesdonotleavespaceforinnovation.NSPSOOOOaallowsanalternativecompliancepathwayforLDARofnewormodifiedwellsitesandcompressorstations,butthelackofcriteriaforjudgingequivalenceisunclearandhasledtoconfusionfrommanyoperators,andthedemonstrationperiodisprohibitivelylongandcostly.ARB’soilandgasregulationappliestobothnewandexistingsources,makingLDARevenmorecostprohibitivetooperators,yetdoesnotincludeanalternativecompliancepathwayoption.

We strongly urge the agency to adopt an alternative compliance pathway that is robust,minimally prescriptive, and specifically creates an entry point for demonstrated methanedetectingsolutions.Suchanapproachwillhelpcatalyzearacetothetopintechnology,reducecostsfortheregulatedcommunity,andboostenvironmentaloutcomes.

WegreatlyappreciatetheopportunitytobepartofthedialogueonGHGstandardsinCalifornia,andlookforwardtoworkingwithARBandindustrytowardourcommongoals.Wearehappytoanswerquestionsordiscussanythinginthesecommentsingreaterdetail.

StevenDeiker,ChiefExecutiveOfficerKairosAerospace777CuestaDrive,Suite202MountainView,CA94040

5

AppendixA:LeakSurveyor(AerialMethaneImaging)

LeakSurveyor isanewmethanedetectionservicethatcombinesanaerialmethanedetectioninstrumentwithaproprietarydataanalysispipeline.Ourgoalistoprovidecustomerswithclear,actionableinformationaboutwhere,andhowbig,theirmethaneemissionsare.LeakSurveyorisdesignedtosurveylargeareasfrequentlyatanaffordablecost–asingleinstrumentcancover30,000acresinoneday.Onceanemissionisdetectedaerially,thecustomersendsadirectedgroundcrewtoidentifyandrepairtheparticularcomponentthatisleaking.Thishybridapproach–frequentaerialsurveyscombinedwithtargetedfollow-upinspections–savesacompanytimeandmoney,andsavesmoremethanethansendinggroundcrewstoeverysite.Thisisbecause:

1) Sourcesareoftenremoteandmostarenotemittingsignificantamountsofmethane,sosendingcrewstoeachonewastesthetimeofengineersatameaningfulcost.

2) Finding a single large emission a day earlier can be more impactful than findinghundredsofsmallemissions,sotimeisbestspentlookingforthelargeemissions.

3) We can frequently survey unsafe- or difficult-to-monitor components that wouldotherwiserarelybesurveyed,andrecovergasthatwouldotherwisegoundetectedbyground-basedmonitoringmethodsforlongperiodsoftime.

Ultimately, LeakSurveyor stops the large emitters that causemost of the environmental andeconomicdamageearlier,foragreatercumulativeimpact.

HighlySpecificMethaneImaging

TheLeakSurveyorinstrumentconsistsofthreeparts:1)anopticalcameraforvisualverificationof sites, 2) aGPS and inertialmonitoring unit to record precise positions, and 3) a patentedspectrometer that detects methane. The spectrometer is sensitive to infrared sunlight thatreflectsofftheground.Whenthislightpassesthroughaplumeofmethane,themethaneabsorbscertainfrequenciesandletsotherspassthrough.Thespectrometeridentifiesthoseabsorptionfeatures,andassociatesthemwithaparticularpositionontheground.Thismakesoursystemhighlyspecificformethane,asitistheonlymoleculethatleavesthisparticularsignatureonthespectrum,andavoidssignalconfusionfromothergaseslikepropane.Itisalsospecifictolocation;the resolutionon theground (~20 feet) iswell-matched tomost gasplumes.Asa result,weproducedirect imagesof plumes, overlaidon simultaneously capturedoptical imagery.AfterreviewingthespecificationsandcapabilitiesofothercategoriesofmethanesensingtechnologieswebelieveLeakSurveyorbelongstoanewclassofinstrument,whichwehavebeencalling“AerialMethaneImagery.”

LeakSurveyor’sresolutionallowsustodistinguishbetweenseparatepointsourcesofmethaneanddifferentiatesusfromair-samplingtechniques.Intheimagebelow,forexample,weseparatethemethaneplumefromnaturalgasproductionfromthemethaneplumefromthefloodedricefield nearby, and would only report the former. What’s more, we distinguish between thelocationandconcentrationofseparateemissionswithinasinglesite,suchthatthegroundcrewfollowingupontheemissionscangodirectlytoaspecificareatoidentifyandrepairaspecificcomponent.

6

Clear,Actionable,PrioritizedResults

Theimagebelowisanexampleoftheresultsweprovide.Methaneplumes,highlightedinblue,areoverlaidonopticalimageryofasurveyareaaboutamilewidewithtenseparatewellpads.

Figure 1: LeakSurveyor’s direct point source methane imaging separates the gas production-related emission (on the left) from the nearby flooded field-related emission (on the right). This allows distinct attribution of methane between different sources.

MethanefromGasProduction

MethanefromFloodedField

Figure 2: Sample LeakSurveyor false color image; blue areas represent methane plumes.

7

TheminimumdetectionthresholdforLeakSurveyorincontrolledsettingsis500ppm-m.Witha5mphwind,thiscorrespondstoanemissionrateof10Mscf/day(thatis,8kg/hr,or2g/s.)Ourconservativereal-worlddetectionlimitis2500ppm-m,whichcorrespondstoanemissionrateof50Mscf/day.Intherealworld,LeakSurveyordetectsastatisticalfractionofemissionsbelow50Mscf/day.Aswecontinuetocollectvastquantitiesofdata–wehaverecentlysetuppartnershipstodosowithanumberofoperators,non-profits,andgovernmentagencies–wewillcontinueto refine the “probability of detection” curve we have built that shows the probability ofdetectingemissionsasafunctionoftheemissionrateandofexternalconditionslikeweatherandterrain.

LeakSurveyordistinguishesbetweendifferentsizesofmethaneemissionswithaprecisionwithin±25%; inFigure4above,areasofdarkblue represent lowerconcentrations, ranging towardslighterareaswhichrepresenthigherconcentrations.Ourabilitytoidentify,overawidearea,thesizeofemissionsallowsprioritizationof repairwork to stop thebiggestemissions fastest forgreater impacton theenvironment, and improveson thequalityofdataonwhichacademicresearchandpolicydecisionsarecurrentlybased.Figure3belowshowsacontrolledmethanereleasewherewesimultaneouslyoperatedLeakSurveyorfromanairplaneflyingat3,000ft.,aFLIRGasFinder320IRcamerapointedatthereleasevalvefrom50ft.away,andaMethod21analyzerheld20ft.fromthevalve.TheLeakSurveyorresultsshowaclearrelationshipbetweenthemethane release size and our signal size. Aswith ourminimumdetection threshold,wecontinuetorefineourquantificationabilityaswecollectmoreLeakSurveyordatacombinedwithground truthmeasurements.Wealsocontinue toconduct calibration studieswithcontrolledreleases,oftenflyingacontrolledreleaseatthesametimeasweareflyingoperators’fields.

Figure 3: Side-by-side comparison of emissions monitoring technology results during a controlled methane release. LeakSurveyor is able to distinguish between different leak sizes.

8

Faster,Cheaper,Safer

TheLeakSurveyorinstrumentiseasilymountedonlightaircraftandflownatstandardgeneralaviationaltitudesof3,000ft.,makingitordersofmagnitudefasterthanagroundcrewandabletoaccessterrainthatwouldbedifficultordangeroustoreachbycar.LeakSurveyorisalsofasterandsaferthanhelicopters.Itcanflylongerandfartherthancommerciallyavailabledrones,whichrarelyhavebatterylivesofmorethan30minutes,limitingtheirflightrangeandincreasingtheircost.Dronesarealsosubjecttocomplexandshiftingstate-by-stateregulatoryissues.

PlugandPlayService

Weoperatethepodasaservice,sothereisnotraining,calibration,orinstrumentmaintenanceorrepairsneededonthepartofourcustomers.Weeliminatethepossibilityofoperatorerrororvariation,asallprotocols,frompre-surveycalibrationtopost-surveydataqualityassurance,areperformedby highly trained Kairos engineers or automatically through our cloud-based datapipeline.Resultsarethusdirectlycomparablefromflighttoflight.Eachpodattachestoaplanewithnotools,wiring,ormodificationsrequired(seeimagebelow),whichmeanswedonotneedFAAapproval.

Figure 4: LeakSurveyor covers orders of magnitude more area than a ground crew, allowing frequent revisits.

Figure 5: Pod attaches to the wing strut of light aircraft with no tools, modifications, or wiring connections.

9

UniqueSoftwareCapability,ContinuousImprovement

Ourproprietarysoftwareanddataanalysiscapabilitiesareuniqueinthemarket.Ourmodern“bigdata”analyticspipelineimmediately,automatically,anduniformlyprocessesthedatawecollectintousableform–whetherthatformisanindividualreportthatintegrateswithexistinginfrastructuremanagementsoftwareusedbyoperators,oramainstreamtoollikeGoogleEarth.Ourprocessreducesthepossibilityofusererrorininterpretingresults,lowerscompliancecosts,andincreasesreportingaccuracy.Anadditionalbenefitisthattheaccuracyandcorrectnessofeachnewsurveywillbeimprovedbytheoverallanalysisofallthedatawehaveevercollected,asresultstakenovertimeareusedasafeedbacklooptoimprovetheLeakSurveyorservice.

NoCapitalExpendituresonEquipmentorTraining

BecauseweoperateLeakSurveyorasaservice,therearenoupfrontexpendituresoncapital(e.g.instruments or trucks) or labor (e.g. training or hiring). Thismeans that starting up an LDARprogram can yield savings for a company immediately. It also gives companies flexibility incraftinganLDARprogramthatworksfortheirparticularneeds.Forexample,acompanythatusesanIRcameraforoneareamaywanttouseLeakSurveyortomonitoranotherhard-to-reacharea.AcompanythathasonlyenoughlaborcapacitytoconductsemiannualsurveyscanincreaseitsmonitoringfrequencytoquarterlyorevenmonthlywithLeakSurveyor.

LowerCostPerSurveyThanAlternatives

LeakSurveyoralsoreducescostspersurveybecausewehavegreatlyreducedtheneedforon-the-groundsurveyors,themostsignificantcost inatraditionalprogram,whetheranoperatorconductshisownsurveysorhiresa thirdparty.Asaquickback-of-the-envelopecomparison,Carbon Limits10estimated that it costs $600 to hire a contractor to survey a singlewell site,whereas LeakSurveyor costs $100 perwell site. Realistically,most of the operatorswe havespokentoreportthatanIRcameracontractorcostsmorethan$600perwellsite,particularlyasthe companies still need to send in-house engineers to accompany the contractors on-site;anecdotally,thecosthasalsoincreasedduetoincreaseddemandandashortageofbothOGIcamerasandqualifiedcontractors.Wehavealsoruncostanalysesusingassumptionsfromothersources,forvarioussizesofproducerandforothermonitoringoptions(i.e.owninganIRcameraversushiringacontractor,usingaMethod21analyzerinsteadofanOGI).

SimplificationofReportingandRecordkeeping

Inadditiontonoupfrontcostsandlowongoingcosts,ourpost-surveycostsareloweraswell–oursoftwareandsecuredatapipelinestreamlinesnecessaryactivities likerecordkeepingandreportingforcompliancepurposes,aswellasmorein-depthanalysis.Wegenerateautomaticreportsregardingdataqualityandcompletenessduringeachsurveyandfortheentiretyofthesurvey.Allofourrawdataandmetadataisstoredinthecloudindefinitelyforrecordkeepingandtime-seriesanalysesforoperators.AndourautomatedsurveygreatlyreducesthetimeittakestorecordcomponentlocationsandIDsforanyfugitiveemissionsdetected.

10 “Quantifying Cost-effectiveness of Systematic Leak Detection and Repair Programs Using Infrared Cameras,” Carbon Limits SA. March 2014. URL: http://www.catf.us/resources/publications/files/Carbon_Limits_LDAR.pdf

10

AppendixB:FEAST,theFugitiveEmissionsAbatementSimulationTestbed

For an in-depth description of the FEAST model, please see the documentation online athttps://pangea.stanford.edu/researchgroups/eao/sites/default/files/FEASTDocumentation_0.pdf.Thisappendixismeanttoserveasabasicsummaryofthemodel’sstructureandhowweusedittomodeltheefficacyofdifferentLDARprograms.

Asdescribedinthedocumentation:“FEASTsimulatestheleakagefromanaturalgasfieldasafunctionof timeunderdifferentLDARprograms. ItdefinesanLDARprogramasa technologyused for leakdetection, the implementationof thedetection technology,and the leak repairprocess.Basedonaplumesimulation,FEASTappliesdetectioncriteriaforseveralLDARmethods,identifiestheleaksthatwillbedetectedundereachLDARprogram,andremovesthemfromthesetofleaksattheappropriatetime.ThetotalgassavedbytheLDARprogramiscalculatedasthetime-integrateddifferencebetweentheleakageinanullscenarioandascenariowiththeappliedLDAR program. The null scenario represents the status quo: it allows a steady leakage ratethroughtimeasnewleaksareproducedandoldleaksarerepairedrandomlywithoutanexplicitLDARprogram.”

TheinputdataforFEASTincludefieldparameters(numberofwells,numberofcomponentsperwell, average distance betweenwells), a dataset of leak rates, atmosphere data in order tosimulate the gas plumes (measured wind speeds and directions), and LDAR parameters(detectioncriteria,timetofindleaks).

Leak distribution: The FWAQS study sampled 375 well sites one time, which represents arelatively small sample size unlikely to capture the full range of extreme results. In order toaccountforsuper-emitters,weextendedtheFWAQSdatawithapowerlawdistribution,whichisusuallyusedtomodeldatawhosefrequencyofaneventvariesasapowerofsomeattributeof that event – in this case the event is a leak and the attribute is leak size. Thepower lawdistribution we used has an upper bound of 500 Mscf/day (meaning the simulation nevergeneratesaleaklargerthanthat.)Wechose-1.75forthepower,orexponent,inthepowerlawformulay = ax% , to match the national top-down estimates of petroleum-sector methaneemissions. If theexponent ismorenegative, excessivenumbersof tiny leaks are required toproducethemethanefromthetop-downestimates.Thisresultsinmanymoreleaksthanwells,suchthatthetypicalwellwouldhavethreetofiveleaks.Thisisoutofsyncwithreality,asmanywellsandfacilitiesarenot leakingatall. Iftheexponent is lessnegative(closertozero),thenmoreof themethane comes from truly enormousemissions.Wedon'thaveenoughdata tosupportthat,althoughitcouldstillbetrue.

LeakSurveyor:WeaddedamoduleforLeakSurveyor,whichisanimplementationofamethodwehavetermedAerialMethaneImaging.Weformedourexpectationsforperformancebasedonextensivecalibrationtests,bothinacontrolledsetting–whereacontrolledmethanereleasewasgradually turnedupwhilea LeakSurveyorplane flewbackand forthoverheadandan IRcamera operator and aMethod 21 instrument operator stood on the ground and recordedimagesandreadingsforcomparison–andintherealworld,wherewecollectedresultsforanoperatorandthenperformedgroundtruthmeasurementstovalidateourfieldresults.Thisreal-

11

worldperformancewasthenaddedtotheFEASTsimulationwiththeconsultationofoneoftheoriginalauthorsofthecodetoensurecorrectintegration.

ThechartbelowshowsonesimulationoftheFEASTmodeloverfiveyears(labeledasJanuary2018-December 2022.) The blue line represents total emissions from the 1,100 wells in thenatural gas fieldwith a semi-monthlyAerialMethane Imagingprogram,which in this case isLeakSurveyor.TheredlinerepresentstotalemissionsfromthenaturalgasfieldwithaquarterlyOGIprogram.Thegreylinerepresentstotalemissionsfromthefieldinthenullcase,whereleaksare randomly fixed in the normal course of operations. Over the time represented, theLeakSurveyor program reduces methane from the null scenario by 87%, and OGI reducesmethaneby76%.

Thechartbelowshowsthecumulativevolumeofmethanereducedovertimerelativetothenullscenario.

0

200

400

600

800

1000

1200

1400

Jan-18

Apr-18

Jul-1

8

Oct-18

Jan-19

Apr-19

Jul-1

9

Oct-19

Jan-20

Apr-20

Jul-2

0

Oct-20

Jan-21

Apr-21

Jul-2

1

Oct-21

Jan-22

Apr-22

Jul-2

2

Oct-22

Metha

neEm

itted

PerDay(g/s)

GasFieldEmissionsWithDifferentControlTechniques

AerialMethaneImaging(semi-monthly) OpticalGasImaging(quarterly) Null

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

1Year 5Years 10Years

MillionsofT

onso

fCH4

MethaneReducedOverTimebyDifferentLDARPrograms

AerialMethaneImaging(semi-monthly) OpticalGasImaging(quarterly)