GHG Emissions Associated with Proposed Natural Gas ...
17
1 GHG Emissions Associated with Two Proposed Natural Gas Transmission Lines in Virginia i Summary of GHG Emission Estimates The primary purpose of this white paper is to estimate possible greenhouse gas (GHG) emissions associated with several proposed new interstate natural gas transmission lines that would run through parts of Virginia. By “associated” emissions we mean the major GHG emissions that are estimated to occur (a) from operation of the transmission pipelines, (b) from the upstream stages of production and processing of the natural gas that is intended to go into to those transmission pipelines, and (c) from combustion of the transported natural gas. (The analysis excludes leaks from local distribution lines, which we assume would be avoided if the gas will be combusted in large plants connected closely to the transmission lines; however, local distribution lines are a major source of methane emissions and would need to be accounted for—in addition to combustion emissions—if deliveries are first made to local gas distributors.) The four major interstate natural gas transmission lines and their daily throughputs of gas proposed in Virginia are the Atlantic Coast (ACP, 1.5 bcf/day), the Mountain Valley (MVP, 2.0 bcf/day), the WB Xpress Project to expand the capacity of the Columbia Gas Transmission pipeline by 1.3 bcf/day), and the Appalachian Connector (up to 2 bcf/day), for a total of 6.8 bcf/day. Our emission estimates for the Atlantic Coast (ACP) and Mountain Valley (MVP) pipelines are summarized in Figures 1 and 2, respectively. The base case (in the first column of the Figures) is from a published analysis: that of Laurenzi and Jersey (2013), referred to here as the “ExxonMobil” analysis since the authors are employees of that Corporation and used data from drilling sites owned by it. In addition we developed three alternative cases based on different assumptions than used in the ExxonMobil results, although one of those cases is derived directly from the ExxonMobil results. In general, the four cases fall into two pairs (labeled ExxonMobil and”EX”) that amount to a low and a higher estimate of upstream emissions of methane (CH4), while estimated carbon dioxide (CO2) remain the same for all cases. Within each pair the difference in carbon dioxide-equivalent (CO2eq) total emissions is due to two different assumptions about how methane is weighted—known as the Global Warming Potential (GWP) of methane. (More detail on the quantitative contributions of CO2 and CH4 in the four cases is given in Tables 1 and 2 in the next section.) For comparison to those pipeline-associated GHG emissions, a seventh column in the Figures shows the total reported emissions of GHGs in Virginia in 2014 from EPA’s Greenhouse Gas Reporting Program.
Transcript of GHG Emissions Associated with Proposed Natural Gas ...
1
GHGEmissionsAssociatedwithTwoProposedNaturalGasTransmissionLinesinVirginiai
SummaryofGHGEmissionEstimates
Theprimarypurposeofthiswhitepaperistoestimatepossiblegreenhousegas(GHG)emissionsassociatedwithseveralproposednewinterstatenaturalgastransmissionlinesthatwouldrunthroughpartsofVirginia.By“associated”emissionswemeanthemajorGHGemissionsthatareestimatedtooccur(a)fromoperationofthetransmissionpipelines,(b)fromtheupstreamstagesofproductionandprocessingofthenaturalgasthatisintendedtogointotothosetransmissionpipelines,and(c)fromcombustionofthetransportednaturalgas.(Theanalysisexcludesleaksfromlocaldistributionlines,whichweassumewouldbeavoidedifthegaswillbecombustedinlargeplantsconnectedcloselytothetransmissionlines;however,localdistributionlinesareamajorsourceofmethaneemissionsandwouldneedtobeaccountedfor—inadditiontocombustionemissions—ifdeliveriesarefirstmadetolocalgasdistributors.)ThefourmajorinterstatenaturalgastransmissionlinesandtheirdailythroughputsofgasproposedinVirginiaaretheAtlanticCoast(ACP,1.5bcf/day),theMountainValley(MVP,2.0bcf/day),theWBXpressProjecttoexpandthecapacityoftheColumbiaGasTransmissionpipelineby1.3bcf/day),andtheAppalachianConnector(upto2bcf/day),foratotalof6.8bcf/day.OuremissionestimatesfortheAtlanticCoast(ACP)andMountainValley(MVP)pipelinesaresummarizedinFigures1and2,respectively.Thebasecase(inthefirstcolumnoftheFigures)isfromapublishedanalysis:thatofLaurenziandJersey(2013),referredtohereasthe“ExxonMobil”analysissincetheauthorsareemployeesofthatCorporationanduseddatafromdrillingsitesownedbyit.InadditionwedevelopedthreealternativecasesbasedondifferentassumptionsthanusedintheExxonMobilresults,althoughoneofthosecasesisderiveddirectlyfromtheExxonMobilresults.Ingeneral,thefourcasesfallintotwopairs(labeledExxonMobiland”EX”)thatamounttoalowandahigherestimateofupstreamemissionsofmethane(CH4),whileestimatedcarbondioxide(CO2)remainthesameforallcases.Withineachpairthedifferenceincarbondioxide-equivalent(CO2eq)totalemissionsisduetotwodifferentassumptionsabouthowmethaneisweighted—knownastheGlobalWarmingPotential(GWP)ofmethane.(MoredetailonthequantitativecontributionsofCO2andCH4inthefourcasesisgiveninTables1and2inthenextsection.)Forcomparisontothosepipeline-associatedGHGemissions,aseventhcolumnintheFiguresshowsthetotalreportedemissionsofGHGsinVirginiain2014fromEPA’sGreenhouseGasReportingProgram.
2
0
10
20
30
40
50
60
70
80
90
100
ExxonMobilGWP25 ExxonMobilGWP84 3XGWP25 3XGWP84 VAGHGRP2014
Figure2.GHGEmissionsfromMVPforFourCasesComparedwithEPAGHGRP2014VAStaBonarySources(MMTCO2eq/year)
Sta?onarySources
Combus?onofDeliveredGas
Transmission&Storage
Produc?on&Processing
TheissueofwhichGWPtochoosecanbebypassedbycomputingthetime-dependentradiativeforcingdueseparatelytoCO2andCH4.Figure3showstheresultsofcalculationsofradiativeforcingcomputedbyasimplemodel.However,insteadofshowingradiativeforcinginconventionalunitsofwatts/meter2weshowthetotalthermalheatingeffectontheplanetofGHGemissionsfromallfourpipelineprojects,consistingoftheradiativeforcingmultipliedbythetotalsurfaceareaoftheEarthplus,forcomparison,themuchsmallergenerationofheatgeneratedbycombustionofthenaturalgasdeliveredbythepipelines.NotethatthethermaleffectofCO2persistslongafteroperationscease(weshowitfor300years),andwilllastforcenturiesafterthat.ThebasisforthisgraphisexplainedinmoredetailintheDiscussionsectionbelow.
Amoredetailedexplanationoftheresultsisgiveninthenextsection.Asubsequentsection,DiscussionofAssumptionsandResults,describestheunderlyingbasisandcomparesourresultstootherstudiesfromtherecentliterature.Followingthatsectionwepresentsomerecommendationsbasedontheresultsandlessonslearnedinanalyzingtheliteratureonemissionsfromthenaturalgasfuelcycle.
DetailedDescriptionofResultsTheExxonMobilanalysisproducedresultsbasedonemissionvaluesperunitoutputofahypotheticalnaturalgaselectricpowerplant(KgCO2eq/MWh),andwescaledtheirGHGemissionsvaluestocorrespondtothepotentialmaximumnaturalgasthroughputoftherespectivepipelines(1.5Bcf/dayforACPand2.0Bcf/dayfortheMVP).Wechosethis
4
studybecauseitwasapartialLCAanalysis(oftheproductionatthewellheadstage),provideddetailedresultsforprocessstepsseparatelyforcarbondioxide(CO2),methane(CH4)andnitrousoxide(N2O)emissions,andpertainedtoconditionsspecifictonaturalgasfromhydraulicfracturingproductionintheMarcellusshaleregion,whichisidentifiedasthesourceforthetwopipelinesinquestion,includingsomemeasurementsmadeontheCorporation’sownwelloperations.However,whiletheseExxonMobilestimatesareusefulasastartingpoint,theymaynotberepresentativeofallfrackingoperationsintheMarcellusorothershaleregions.Infact,otherestimatesofoverallemissionsfromthatregionsuggestmuchhigherfugitiveemissionsofmethane,anditisclearthatsomeoperatorsareresponsibleformuchmoreemissionsperunitofproductionthanothers.Forthatreasonwealsopresentanalternativesetofestimatesformethaneemissionsfromtheoverallproductionandprocessingstage,asdiscussedbelow.Notethatneitheroftheseestimatesappearstoconsidertheproblemofpost-productionleaks,which,asdocumentedbySchlumberger,mayemergemanyyearsafterawellhasbeencappedandtakenoutofoperations.Figure1andTable1showresultsapplicabletotheACPpipeline,whileFigure2andTable2showsimilarresultsfortheMVPpipeline.Forsimplicity,weaggregatedtheoriginalauthors’moredetailedprocesslevelresultsintothreemajorfuelcyclestages:1ProductionandProcessing(i.e.,operationsupstreamofthetransmissionline),TransmissionandStorage,andCombustionofthedeliveredpipelinegas(assumingnolocaldistribution).(CO2eqemissionsofN2OareneglectedinTables1&2asrelativelysmallcomparedtotheGHGimpactsofmethaneandCO2emissions.)WebelievethatassessingGHGemissionsfromallthreemajorfuelcyclestages,notjustthetransmissionpipelinestage,isimportantbecausethesenewpipelinesareintendedtocollecttheproducedgasesandtransportthemtoneworexpandedmarketsinVirginiaandNorthCarolina,andpossiblyeventoforeignexportterminals.Hence,thepipelineswilltendtogenerateoratleastsupportadditionalusesofnaturalgasthatarguablywillresultingreatergasproductionandcombustionandtheirassociatedemissions.Someoftheusesmayincludenewindustrialplantsownedbyforeigncompaniesthatareattractedtotheregionbytheavailabilityofcheapernaturalgassuppliesthanavailableabroad.Pipelineproponentshavebeentoutingsucheconomicdevelopmentasabenefitoftheirpipelines.Thetwoprincipleissuesinmakingmethaneleakageestimatesare:1)whatistheactualleakagerateofmethanefromvariousstagesofthenaturalgasfuelcycle?and2)whatistheappropriatechoiceofglobalwarmingpotential(GWP)(orothermethod)toapplywhencomparingemissionsofCO2tootherGHGs,especiallytomethane?Thereason1Thefuelcycleapproachmeansanalysisofoperationalimpactsofallrelevantstagesfromextractionthroughuseanddispositionofwastes;alifecycleanalysis(LCA)approachextendstheanalysistoconsiderationoftheindirectimpactsofmanufacturingandtransportingtheequipmentandtherawmaterialsthatgointothestagesandisevaluatedovertheestimatedlifetimeofthecapitalfacilities.
5
therearefourcolumnsinthetwotablesandfirsttwofiguresisbecausewemadealternativechoicesforbothofthoseissues.InTables1and2thefirstcolumnisfromthegenericestimatesgivenbyLaurenziandJersey(exceptforthescalinguptoeachpipeline).Notethatthescale-upassumesthepipelinesoperateatfullcapacity24/7/365becausewehavenoestimatesfromtheproponentsabouttheirplannedoperatingschedule.ThesecondcolumnadjuststhemethaneCO2eqemissionvalues(thefirstcolumnwasbasedonEPA’s100-yearGWPassumptionof25)tothe20-yearGWPof84fromIPCCAR5whensummingtoobtaintotalCO2eqemissionsfromeachstage.Thethirdandfourthcolumns(3X)increasethemethaneemissionsfromProductionandProcessing(butnotthetransmissionorcombustionstageemissions)bymultiplyingColumns1and2,respectively,byafactorofthreetoreflectresultstypicaloftop-downhighermethaneemissionmeasurementsintheMarcellusandothershalebasins.ThereasonforthischoiceisexplainedbelowintheDiscussionsection.ThosetwoadjustmentsincreasetheupstreamproductionandprocessingemissionsinColumn4byafactorof4.9andthetotalsystememissionbyafactorof1.7relativetocolumn1.(NotethattheCH4emissionvaluesshownintheTablesareinmillionmetrictonnes(MMT)ofmethane,notCO2eq.)TheCO2eqvaluesfromthefourcolumnsinthetalesarealsoshowngraphicallyinthebarchartsofFigures1and2.Forcomparison,Virginia’stwolargestsourcesofCO2eqGHGemissionsin2014weretheChesterfield(7.22MMT)andClover(5.67MMT)coal-firedpowerplants.TheColumn1totalinTable1fromtheACPpipeline(40.7)iscomparabletothetotalcontributionfromthe177GHGsourcesinVirginia(49.4MMTCO2eq)fromEPA’sGreenhouseGasReportingProgram(GHGRP)in2014,whilethetotalinTable2fromtheMVPpipelineconsiderablyexceedsit.2(However,onlypartoftheemissionsinTables1and2wouldoccurinVirginia.)TheseVirginiaGHGRPvaluesalsoarecomparedagainstthepipelinevaluesinFigures1and2.ObviouslythecomparabletotalsforthehighermethaneemissionsassumedinColumns3and4ofthetwotableswouldbeevenhigher,butonlyColumns1and3shouldbecomparedwithEPA’sGHGvaluessincethelatteralsoassumeaGWPof25.EmissionsfromtheothertwoproposedpipelineswouldnearlydoublethetotalemissionsfromtheACPandMVPforatotalof185MMTCO2eqataGWPof25inthebasecase,theExxonMobilrates,or3.7timestheEPAGHGRPtotalforVirginia.
2ThisisbasedonEPA’s“Flightdatabase”fromtheirGreenhouseGasReportingsystem,butthatdatabaseexcludesGHGemissionsfromonshoreoilandgasproductionatthestatelevel,henceitdoesnotincludetheemissionsfromcoalbedmethaneextractionoperationsinVirginia,forexample.Also,thelistof177largesourcesincludessomethatreportedzeroemissionsin2014comparedwithsubstantialemissionsinprioryearsandEPAgenerallyassumestheGHGRPreportedemissionsunderestimateactualtotalssomewhat.Onlylargesourcesarerequiredtoreport,andthedatabasedoesnotincludetransportationandmanysmallsources
6
TABLE 1. Generic GHG Emission Estimates for the ACP Pipeline GHG Emissions by gas and fuel cycle stage
ExxonMobil* (w/CH4 GWP=25 over 100 years)
Adjusted ExxonMobil* (w/CH4 GWP=84 over 20 years)
Top-Down Higher CH4 Leakage Estimate** (w/CH4 GWP=25)
Top-Down Higher CH4 Leakage Estimate** (w/CH4 GWP=84)
Production & Processing CO2 (MMT CO2/year) 3.60 3.60 3.60 3.60 CH4 Losses (MMT CH4/year) 0.107 0.107 0.321 0.321 Total CO2eq Emissions (MMT/year) 6.3 12.6 11.6 30.6 Transmission & Storage CO2 (MMT CO2/year) 1.27 1.27 1.27 1.27 CH4 Losses (MMT CH4/year) 0.058 0.058 0.058 0.058 Total CO2eq Emissions (MMT/year) 2.7 6.1 2.7 6.1 Combustion of Delivered Gas CO2 (MMT/year) 31.7 31.7 31.7 31.7 Grand Total GHG Emissions (MMT CO2eq /year)
40.7 50.4 46.0 68.4
* ExxonMobil means the ANALYSIS analysis of Laurenzi & Jersey (2013); note that this was a generic analysis, not specific to the ACP pipeline. The values here represent a conversion from their numbers in terms of emissions/MWh into emissions/SCF, which are multiplied times the ACP capacity of 1.5 Bcf/day to get the MMT/year values shown here. These values assume full-time operation 24/7/365. ** Assumes 3 X ExxonMobil CH4 Production & Processing emissions (see discussion)
7
TABLE 2. Generic GHG Emission Estimates for the MVP Pipeline
GHG Emissions by gas and fuel cycle stage
Exxon-Mobil* (w/CH4 GWP=25 over 100 years)
Adjusted Exxon-Mobil* (w/CH4 GWP=84 over years)
Top-Down Higher CH4 Leakage Estimate** (w/CH4 GWP=25)
Top-Down Higher CH4 Leakage Estimate** (w/CH4 GWP=84)
Production & Processing CO2 (MMT CO2/year) 4.8 4.8 4.8 4.8 CH4 Losses (MMT CH4/year) 0.143 0.143 0.428 0.428 Total CO2eq Emissions (MMT/year) 8.4 16.8 15.5 40.8 Transmission & Storage CO2 (MMT CO2/year) 1.7 1.7 1.7 1.7 CH4 Losses (MMT CH4/year) 0.077 0.077 0.077 0.077 Total CO2eq Emissions (MMT/year) 3.6 8.1 3.6 8.1 Combustion of Delivered Gas CO2 (MMT/year) 42.3 42.3 42.3 42.3
Grand Total GHG Emissions (MMT CO2eq /year)
54.3 67.2 61.3 91.2
* ExxonMobil means the ANALYSIS analysis of Laurenzi & Jersey (2013); note that this was a generic analysis, not specific to the MVP pipeline. The values here represent a conversion from their numbers in terms of emissions/MWh into emissions/SCF, which are multiplied times the MVP capacity of 2.0 Bcf/day to get the MMT/year values shown here. These values assume full-time operation 24/7/365. ** Assumes 3 X ExxonMobil CH4 Production & Processing emissions (see discussion) _____________________________________________________________________________________________
DiscussionofAssumptionsandResultsThetwoprincipleissuesinmakingtheseestimatesare:1)whatistheactualleakagerateofmethanefromvariousstagesofthenaturalgasfuelcycle,and2)whatistheappropriatechoiceofglobalwarmingpotential(GWP)(orothermethod)toapplywhencomparingemissionsofCO2tootherGHGs,especiallytomethane?Bothofthosequestionshavebeenissuesforseveraldecades.Neitheriscompletelysettledtoday.Wehaveapproacheditinourestimatesbychoosingalowerandhighervalueforeachfactor,andalsoproducedaseparateanalysisthatobviatestheGWPissue.
8
Theissueofleakageratesremainsunresolvedandaverycontroversialtopic.Thewaychosenheretorepresentarangeofopiniononleakageratesfromtheupstreamproductionandprocessingstagesistoshowalowerestimate(theso-calledExxonMobilvalues,whicharesimilartoEPA’semissionfactors)vs.ahigherestimate(the“3X”or”Top-DownHigher”valuesinColumns3&4)asexplainedfurtherbelow.ChoiceofGWP.TheGWPissueisnowquitewellunderstoodscientificallybutremainscontroversialinthepolicyandpoliticalarenas.TheissuewithaGWPselectionisthattheUNadopteda100-yearGWPaspartoftheKyotoProtocol.EPAalsoadopteditbecauseoftheneedtohaveaspecificwaytoweightvariousGHGsandvalueemissiontradeoffsandtobeconsistentwithInternationalreportingrequirements.However,forotherpurposessuchasevaluatingmitigationstrategiesandlonger-termtradeoffs,manyclimatescientistsandpolicyanalysts,includingthelatestIPCCreports,nowunderstanditslimitations.Forstrategicpurposestherearealternativesolutionsforcharacterizingtherelativeimpactsavailableintheliterature(e.g.,Alvarezetal.2012)thatrenderthatchoiceirrelevant.However,forsimplicityherewesimplycomputemethaneeffectsfortwowidelydifferentvaluesoftheGWPtoillustratetherange:EPA’svalueof25(thatwasbasedontheIPCCAR42007reportfora100-yeartimeframe)andwasusedbyLaurenziandJersey,andtheIPPCAR52013valueof84fora20-yeartimeframe.Webelievethatthelatestscientificestimatesshouldbeappliedandthatthereisnoscientificjustificationforpreferringa100-yearovera20-yearvalues,especiallysincemanyoftheGHGmitigationgoalsoftheU.S.(forexample,theU.S.pledgetotheUNFCCCprocessfor2025)willoccurovermuchshorterperiodsoftime,closertoa20-yearperiod.WealsoshowinFigure3theresultsofasimplemodelthatshowsthetemporalevolutionofplanetaryheatingduetotheemissionsofCO2andCH4(separately)plusheatingfromcombustionofthedeliveredgasesfromallfourpipelineprojects.Forthischartweusedthehighermethaneemissionrates(columns3and4inthetables).PlanetaryheatingfromtheGHGemissionsmeanstheincrementalradiativeforcingatthetopoftheatmosphereduetotheemittedgases.OursimplemodelissimilartothatdescribedbyAlvarezetal.2012,althoughweuseupdatedparametersbasedonthelatestestimatesoftotalgreenhousegasconcentrationsintheatmosphereanddisplayourresultsinabsolutetermsasplanetaryheating.Ourmodelwillbedescribedinmoredetailinasubsequentpaper.ThisapproacheliminatestheneedforusingGWPsandprovidesmoreinformation.ProductionandProcessingStages.EstimatesofGHGemissionsfromnaturalgasproduction,processingandgatheringpipelinetransportoperationsdifferwidelyandcurrentlyareverycontroversial.Briefly,thereisanunresolveddisconnectbetweentwogeneralapproachestoestimatingemissions:so-calledbottom-upmethodsthatsumupmeasurementsand/orgenericemissionfactor-basedestimatesforindividualoperationsandequipmentintheoverallprocess,versustop-downmethodsbasedonmeasuringconcentrationsofmethaneintheatmosphereforsomeregioninwhichtherearenaturalgasand/oroilproducingoperations,thentranslatingthosemeasurementsintoestimatesofemissionsassociatedwithnaturalgasandoilproduction,processingand
9
otherstages(dependingonwhatoperationsareoccurringinthestudyregion).Thosetwoapproachesleadtoestimatedemissionsthatcandifferbyasmuchasanorderofmagnitude.Figure4belowshowssomeexamplesoftop-downcomparedwithEPAbottom-upestimates.Notethatseveraltop-downestimatesshowninFigure4haveamedianvalueofabout10%leakage,comparedwiththeEPAestimatebetween1and2%.Tables1and2beginwithoneestimate(Columns1&2)ofabottom-upapproach,theExxon-Mobilstudy,whichisnearthelowerendoftherangeofsuchestimates,(althoughthereareevenlowerones).Itamountstoabout1.12%leakageofmethanefromtheupstreamproductionandprocessingstagesofMarcellusshalefracking,inparticularintheSouthwesternPennsylvaniapartofthatregion.Wealsogivehypothetical(3X)estimates(Columns3&4)(basedonmultiplyingtheExxonMobilresultsbyafactorof3)thatwebelievearerepresentativeofthemiddleofthetop-downestimatesandalsoarecomparabletothehigherendofbottomupestimates),whichisequivalenttoupstreammethaneemissionsofabout3.4%.TheExxonMobilresultsformethaneemissionappeartoberoughlyinthesamerangeassomeotherbottom-upestimatesnearthelowend,includingvaluesbasedonEPA’sGreenhouseGasEmissionInventory.Thereareanumberofgeneralissueswithmostbottom-upstudies,includingthedifficultyofassuringthatindividualmeasurementsmadetodetermineemissionfactorsarerepresentativeofthebroaderindustryoperations,andthatmostmeasurementshavebeenmadebyorincloseassociationwiththeproducingindustrythathasavestedindustryinshowinglowemissions.(Itisdifficulttomakedetailedmeasurementsatasitewithouttheoperator’scooperation,andtherealwaysisaquestionaboutwhethertheoperatormaydothingsdifferentlywhenheknowsresearchersorgovernmentinspectorsarepresent.)Theparticularhigh-endestimateformethaneleakageweusehere(3.4%)iscomparabletothetop-downresultsreportedinthestudybyPetronetal.(2014),viz.4.1±1.5%.However,thatpertainedtonaturalgasproductionfromacombinationofoilandgaswellsandsupportinginfrastructure.Thatstudyinvolvedatmosphericstudiesusingvariouscombinationsofground-basedairmonitors,aircraftmeasurements,andothermeasurementsofmethaneandVOCconcentrations.Therehavebeenrelativelylargeuncertaintyboundsontop-downmethods.(SeeboundsshowninFigure4below,butalsothenewerZavala-Araizaetal.,2015studydiscussedbelow.)Theadvantageoftop-downestimatesisthattheytendtocaptureallthemethaneemissionsinaregion,includingnaturalgasindustrysourcesthatmayhavemuchhigheremissionsthanrepresentedbyemissionfactors(andthereismuchevidencethatafewlargeleakagesourcesaccountforadisproportionatecontributiontototals).Theirresultwasnowhereneartheworst-caseleakageexampleamongtopdownstudies,someofwhichfoundvaluesofmethaneleakageontheorderof10%ormore,asshowninFigure4.Aleakagerateof3.4%isalsoconsistentwithhigherestimatesusingbottom-upmethodsfromtheliterature[forexample,seeBrandtetal.(2014)].AtmosphericmeasurementsdonotmeasureCO2emissions,soweusethesameCO2estimatesfromLaurenziandJerseyinthiscolumninTable1.AlsonotethatatmosphericmeasurementsdonotnecessarilycapturealltheindirectLCAvaluessincesomeofthosemayapplyto
10
operationsoutsidetheproducingareas,butthosetendtobethesmallerpartofthetotalemissions.AveryrecentreportbyZavala-Araizaetal.(2015)reconcilesbottom-upandtop-downestimatesintheBarnettshaleoilandgas-productionregionofTexas.Itaugmentsconventionalbottom-upinventories,accountsforhighemitters,andcomparesthemtotop-downaircraftstudiesinwhichethanemeasurementsareusedtocorrectforbiogenicsources.Theirbottom-upinventoryis1.9timesestimatedemissionsbasedontheEPAGHGIprogram,andrepresentsamethaneleakagerateof1.5%(1.2—1.9%).TheAircrafttop-downmeasurementsoffossilmethaneaveragedabout10%higherthanthebottom-upestimates,butstillwithinthetop-downuncertaintybounds.ThoseresultsfortheBarnettregionindicateasignificantlysmallerleakageratethanthePetronetal.(2014)resultsobtainedintheDenver-Julesburggasandoilproductionregion.TheZavala-Araizaresults(amethaneleakagerateof1.5%forupstreamproductionandprocessingstages)suggestamediumleakagecaseinbetweenourbaseExxonMobilvalueandthe“Higher3X”leakageestimateof3.4%incolumnsthreeandfour.Ofcourse,neitherofthoseestimatesfromotherbasinsnecessarilypertainstotheMarcellusshalegasproductionregion,sowecannotsaywhetherourassumedmediumandhighvaluesintheTablesandFiguresareconsistent.Wedonotclaimthatthevalueof3.4%usedhereisavalidupperestimatefortheMarcellusregion,butonlythatitillustratesthepotentialimpactofahigherestimatethatisslightlysmallerthanatop-downresultfromanotherregionthatinvolvedparticularlycomprehensivemeasurements.AreportbyMarcheseetal.(2015)givesestimatesofemissionsfromthegasprocessingandgatheringpipelinestages(whichstagesareincludedinourestimatesofProductionandProcessing).Generallytheyfoundthattheirmeasurementsof16gasprocessingplantswereevenlowerthanEPA’semissionfactors,butmeasurementsof114gatheringpipelinefacilitieswereoftenmuchhigherthanEPAemissionfactors.Afewofthesmallergatheringfacilitiesappeartohaveleakageratesexceeding10%ofgasthroughput,butmostweremuchlessthanthat.Marcheseetal.didconcludethat:
“WhilethereisuncertaintyindetermininggatheringfacilityemissionsfromtheEPAGHGI,theresultsofthisstudysuggestthattheGHGIsubstantiallyunderestimatesemissionsfromgatheringfacilities.
TheMarchesestudyindicatesthatemissionsfromgatheringlinesmaybeconsiderablylargerthanestimatedintheExxonMobilanalysis.However,suchincreasedmethaneemissionspresumablywouldalreadybeaccountedforinbroadregiontop-downstudiesthatarethebasisforourmediumandhighermethaneestimates,sotheredoesnotappeartobeaneedtofactorthatintoourresultsincolumnsthreethroughsix.Arecentreport,ConcernedHealthProfessionalsofNewYorkReport(2015),foundthat(p.52-57):
11
“Leakagefromfaultywellsisanissuethattheindustryhasidentifiedandforwhichithasnosolution.AccordingtoSchlumberger,oneoftheworld’slargestcompaniesspecializinginfracking,aboutfivepercentofwellsleakimmediately,50percentleakafter15years,and60percentleakafter30years.DatafromPennsylvania’sDepartmentofEnvironmentalProtection(DEP)for2000-2012showoverninepercentofshalegaswellsdrilledinthestate’snortheasterncountiesleakingwithinthefirstfiveyears.Leaksposeseriousrisksincludingpotentiallossoflifeorpropertyfromexplosionsandthemigrationofgasorotherchemicalsintodrinkingwatersupplies.
“Leaksalsoallowmethanetoescapeintotheatmosphere,whereitactsasamorepowerfulgreenhousegasthancarbondioxide.Indeed,overa20-yeartimeframe,methaneis86timesmorepotentaheataccumulatorthancarbondioxide.Thereisnoevidencetosuggestthattheproblemofcementandwellcasingimpairmentisabating.Indeed,a2014analysisofmorethan75,000compliancereportsformorethan41,000wellsinPennsylvaniafoundthatnewerwellshavehigherleakageratesandthatunconventionalshalegaswellsleakmorethanconventionalwellsdrilledwithinthesametimeperiod.Industryhasnosolutionforrectifyingthechronicproblemofwellcasing/cementleakage.”
CombustionStage.CO2emissionsfromthenaturalgas-firedcombustion(e.g.,powerplant)stagedependmainlyontheamountofgasconsumed,whichinthiscaseissimplythethroughputofthepipeline,andslightlyonthecompositionofnaturalgas(whichchangestheCO2percubicfoot).EffectivelyweusedthelatterfactorfromLaurenziandJerseysincetheybaseditontypicalpipelinenaturalgasproducedintheMarcellusshaleregion(ratherthanEPA’snominalemissionfactor).Anycombustionuseofthetransmissionlinenaturalgasthroughputwouldgivethesameresult.However,naturalgasdeliveredfurtherforusethroughlocaldistributionlineswouldhavehigheroverallCO2eqemissionsbecauseofthesubstantialextraleakageofmethaneinmanydistributionsystems.GHGemissionspublishedbyLaurenziandJerseyfromthisstagearejustfromcombustion,arenotbasedonalifecycleanalysis,anddonotaccountforanyleakageofmethaneorunburnedmethaneinthepowerplantexhaustorpre-combustionhandling.WhilewecouldnotfindadefinitiveemissionfactorfromEPAformethanespecifictoNGCCpowerplants,NETL(2010)givesthefactor8.56E-06kg/MWhforNGCCplants3.ThatwouldbenegligiblecomparedwiththeCO2emissions.TransmissionandStorage(T&S)Stage.Ourbaseestimateforthisstageisbasedonadifferenttreatment.TheExxonMobilanalysisdidnotbasetheirestimateonalife-cycleanalysisoradetailedcalculationofemissionsfrompipelinefacilities.Rather,ittakes
3Methaneemissionfactorsvarywiththetypeofcombustionprocess;methaneandN2Oemissionsfromsimplegasturbinesandotherenginesusedtopowerpipelinecompressorsarenotassmall;e.g.,EPA’sAP-42GHGemissionfactorsfornaturalgas-firedturbinesare0.003lb/MMBtuforN2Oand0.0086forCH4,whichtogetheramounttoabout1.4%oftheCO2emissionswhentheAR520-yearGWPsforthosegasesareapplied(268forN2O).
12
2009EPAestimatesoftotalT&SfugitivemethaneemissionsandtotalCO2fromcompressorstocalculatetheratiotototalnaturalgaswithdrawalsthatyear.Thatresultsinanaverageleakagerateof0.45%ofmethaneandanaverageamountofCO2emissionsof82Kg/MMScfoftransportedgas.Weonlyhavelimitedinformationaboutthetwoproposedpipelines,suchaslengths,sizes,compressorhorsepower,andmaximumgasthroughputperday.Theredonotappeartobeanyemissionfactorsavailabletoestimatepipelineemissionsbasedonlyonthoseparameters.GiventhoselimitationsandthegenericnatureofinformationfromtheLaurenziandJersey(2013)paperabouttheassumptionsanddatafortheiremissionestimatesoftheTransmissionandStoragestage,itdidnotappearfeasibletoestimatehowtheirgenericestimatesofmethaneshouldscalewithvariouspipelineparameters,otherthanadirectscalingwithpipelinethroughputcapacity.WealsonotethatGHGemissionestimatesfromthepipelineproponentsdonotyetappeartobeavailable.Thatmayespeciallybeimportantforthedirectemissionvaluesforpipelineoperations.TheanalysisofLaurenziandJersey(2013)assumesa0.45%CH4leakrateintransmissionbuttheydonotstatespecificassumptionsabouttransmissionmiles,compressorHPandotherfactors.Rather,theyassumeafractionoftotalEPAestimatesforpipelineCH4leakageandcompressorCO2emissionsin2009basedonthefractionofgrossgaswithdrawals.TheACPandMVPtransmissionpipelines,totaling554.6milesand294miles,respectively,maynotbetypicalofthelengthandleakageratesimplicitintheLaurenziandJerseyanalysis.Itwouldbedesirabletoupdatethoseestimateswhenmorespecificinformationbecomesavailable.Subramanianetal.(2015)recentlypublishedanonsitestudyofcompressorstationemissions.Itincludesmeasurementsofmethaneemissionsfrom47transmissionlinecompressorandstoragesites.Thisisclaimedtobethemostcomprehensivesetofmeasurementssincethe1996jointEPA/GasResearchInstitutestudy.However,themeasuredfugitivemethaneemissionestimatesvarybyseveralordersofmagnitudeamongstationsandthestudyfoundnocorrelationbetweenemissionsandcompressorhorsepower.Thoseresults,togetherwithresultsofotherstudies,indicatethattherearelargevariationsinemissionsamongdifferenttechnologiesusedinequipment,probablyintheamountofeffortcompaniesspendonmaintenanceofthingslikesealsoncompressors,valves,andleaks,andperhapsalsointheeffortsspentonmonitoringtodetectleaks.4Becauseofthewidevarianceintheseresultsandthelackofclearcorrelationtopipelineparameterssuchastotalhorsepowerandsizeofpipeline,wewereunabletousetheresultstoreplaceorcomparedirectlywiththoseoftheExxonMobilstudy.4AnEPAbackgroundstudy,EPA(2014),preparedforanalysisofaproposedNSPSstandard,estimatedthefollowingmethaneemissionsachievablepercompressorforeachofthethreetypesoftransmissioncompressor:27.1metrictonne/yearforreciprocating,126forcentrifugalwithwetseals,and15.9forcentrifugalwithdryseals,butthoseestimatesapparentlydonotincludealltheothercomponentsatacompressorstation,whichinpracticecancontributesubstantialemissionsduetoleakage,ventingandexhaustemissions.
13
Zimmerleetal.(2015)publishedarecentstudyoftheU.S.naturalgastransmissionlineandstoragesystem(T&S)methaneemissions.Thisstudy’sestimatedoverallUStransmissionandstoragesectoremissionsfor2012as1503Gg/yr,whichwerewithintheirstatisticaluncertaintyofEPA’sGHGIestimatedvalueof2071Gg/yr.Theyalsofoundsuperemitterstationsthatappeartobeduetoequipmentorcontrolmalfunctions.OnecancomparethoseleakageestimateswiththeU.S.totalvaluethattheExxonMobilstudyusedasthebasisfortheirgenericestimateofpipelineemissions,whichwas2,115Gg/yrfor2009,or0.45%oftotalgasproduction.Sincetotalgrosswithdrawalsin2012wereabout16.5%largerthanin2009,theZimmerlestudyvalueof1503Gg/yrcorrespondstoamethaneleakagerateofabout40%lessthantheExxonMobilstudy,orabout0.27%ofgrosswithdrawals(apparentrangeof0.23to0.39%).However,bothofthoseestimatesrefertoaveragesoveranationalmixofdifferentpipelinesofdifferentsizes,agesandcapacities,soitisquestionablewhethertheycanbeapplieddirectlytospecificnewtransmissionpipelineprojects.TheZimmerleetal.studyincludestheresultsfromSubramanianetal.(2015)atindividualcompressorstationandstoragesites,butapparentlyextendstheanalysis.Theyfitalltheirresultstoseveraldifferentmodelsinordertodrawconclusionsabouttheoverallpopulationofsites,includingtheU.S.totalT&Semissionscitedabove.However,itagainitisdifficultforustointerpretthoseresultsintermsofspecificestimatesfortheACPandMVPpipelines.
14
Fig.4.ChartfromSchneisingetal.(2014).(Figureandcaptioncopied
directlyfromFigure7oftheirreport)
15
ConclusionsandRecommendations
ThepotentialtotalGHGemissionsassociatedwiththesetwoproposednewpipelinescouldgreatlyincreaseemissionsfromthisregionfordecadesintothefuture.Hence,inanerawhereclimatechangemitigationwillrequirereducingGHGemissionssharply,decisionmakersneedtoconsiderwhetherapprovaloftheseprojectsisconsistentwithnationalandinternationalgoalsforclimatemitigation.GiventheobservedwidevariationinmethaneemissionsandtheveryhightotalpotentialGHGemissions,itisimportantthatthetransmissionpipelinecompaniesandFERCprovidecompletelife-cycleestimatesofmethaneandCO2emissionsfromtheirprojectsfortheEISfortheirproposedpipelineprojects,togetherwithdetaileddocumentationoftheirassumptionssothatthepotentialGHGemissionsandotherenvironmentalimpactsofthepipelinestagecanproperlybejudged.ItisclearthatexpandinggasusageandsupportingitwithnewpipelinesandproductionimpliessubstantiallygreatertotalGHGemissionsthanappearwhenagenciesoradvocatesfocusononlyonestageatatimeandignoretheindirectimpactsoftheimmediateproject.FERCmustrecognizethattheemergingworldcommitmenttocutGHGemissions,asevidencedbytherecentUNFCCCCOP21agreementinParis,willmeanthattheoperatinglivesofnewnaturalgasinvestmentsarelikelytobesubstantiallyshorterthanthetraditionalassumptionthatapipelinewilloperateforthirtyormoreyears.Expandinginvestmentsbasedonsuchrosyassumptionswillleadtosubstantialstrandedinvestments,inadditiontoincreasedglobalwarmingfromexcessiveGHGemissions.Theseareamplegroundsforrejectingcertificateapplicationsforexpandednaturalgaspipelinecapacity.Ataminimum,pipelineinvestorsshouldbeplacedatriskforunder-recoveryofinvestmentsasaresultofovercapacityfortransportationofnaturalgasthatcannotcontinuetobeburnedathistoric,letaloneexpanded,levelsforseveraldecadesintothefuture.Furthermore,ifFERCdecidestoalloweitheroftheproposedpipelinestoproceed,itshouldrequiredetailedmaintenanceandemissionmonitoringplansfornewandassociatedexistingpipelinesandcompressorstationsadequatetopreventleaksanddetectallreleasesofmethanetotheatmosphereinatimelyfashionsothatsubstantialleakscanquicklyberemedied,bothforpublicsafetyandtominimizetheclimateimpactsofGHGemissions.
References
Alvarezetal.(2012):Alvarez,R.A.,etal.,“Greaterfocusneededonmethaneleakagefromnaturalgasinfrastructure”,PNAS,109,6435–6440,April24,2012.
Brandtetal.(2014):Brandt,A.etal.“MethaneLeaksfromNorthAmericanNaturalGasSystems”,Science343,733(2014)DOI:10.1126/science.1247045
16
ConcernedHealthProfessionalsofNewYork,“CompendiumOfScientific,Medical,AndMediaFindingsDemonstratingRisksAndHarmsOfFracking(UnconventionalGasAndOilExtraction)”ThirdEdition(October14,2015),p.52-57.“ExxonMobilReport”LaurenziandJersey(2013):Laurenzi,J.,andG.Jersey,“LifeCycleGreenhouseGasEmissionsandFreshwaterConsumptionofMarcellusShaleGas”,Env.Sci.&Technol,dx.doi.org/10.1021/es305162w.Marcheseetal.(2015):Marchese,A.J.,etal.,“MethaneEmissionsfromUnitedStatesNaturalGasGatheringandProcessing”,DOI:10.1021/acs.est.5b02275Environ.Sci.Technol. NETL (2010): “LifeCycleAnalysis:NaturalGasCombinedCycle(NGCC)PowerPlant“,DOE/NETL-403/110509. Petronetal.(2014):Pétron,G.,etal.(2014),“AnewlookatmethaneandnonmethanehydrocarbonemissionsfromoilandnaturalgasoperationsintheColoradoDenver-JulesburgBasin”,J.Geophys.Res.Atmos.,119,6836–6852,doi:10.1002/2013JD021272.Schneisingetal.(2014):Schneising,O.,J.P.Burrows,R.R.Dickerson,M.Buchwitz,M.Reuter,andH.Bovensmann(2014),RemotesensingoffugitivemethaneemissionsfromoilandgasproductioninNorthAmericantightgeologicformations,Earth’sFuture,2,doi:10.1002/2014EF000265Subramanianetal.(2015):“MethaneEmissionsfromNaturalGasCompressorStationsintheTransmissionandStorageSector:MeasurementsandComparisonswiththeEPAGreenhouseGasReportingProgramProtocol”,DOI:10.1021/es5060258,Environ.Sci.Technol.2015,49,3252−3261 EPA (2014): EPA OAQPS, “ OilandNaturalGasSectorCompressorsReportforOilandNaturalGasSectorCompressorsReviewPanel”,April2014.
G.Vaidyanathan,“Whichoilandgascompaniesareleakingthemostmethane?”http://www.eenews.net/climatewire/2015/06/26/stories/1060020954(June26,2015).
Zavala-Araizaetal.(2015):Zavala-AraizaD.etal.,“Reconcilingdivergentestimatesofoilandgasmethaneemissions”,www.pnas.org/cgi/doi/10.1073/pnas.1522126112Zimmerleetal.(2015):“MethaneEmissionsfromtheNaturalGasTransmissionandStorageSystemintheUnitedStates”,Environ.Sci.Technol.2015,49,9374−9383.
17
iPreparedfortheVirginiaChapterSierraClubwithcontributionsbyRichardH.Ball,Ph.D.,volunteerSustainableEnergyChair,WilliamPenniman,Esq.,volunteerConservationChair,andKirkBowers,PE,PipelinesProgramManager,VirginiaChapter.
