LBNL513E
Carbon Sequestration Risks and Risk Management
[Link],[Link],[Link]
EnvironmentalEnergyTechnologiesDivision
LawrenceBerkeleyNationalLaboratory
Berkeley,California94720
Acknowledgment:ThisworkwassupportedinpartbyWESTCARBthroughtheAssistant
SecretaryforFossilEnergy,OfficeofCoalandPowerSystems,throughtheNationalEnergy
TechnologyLaboratory(NETL),andbyLawrenceBerkeleyNationalLaboratoryunder
DepartmentofEnergyContractNo.DEAC0205CH11231.
Disclaimer:ThisdocumentwaspreparedasanaccountofworksponsoredbytheUnitedStates
[Link],neithertheUnited
StatesGovernmentnoranyagencythereof,norTheRegentsoftheUniversityofCalifornia,nor
anyoftheiremployees,makesanywarranty,expressorimplied,orassumesanylegal
responsibilityfortheaccuracy,completeness,orusefulnessofanyinformation,apparatus,
product,orprocessdisclosed,orrepresentsthatitsusewouldnotinfringeprivatelyowned
[Link],process,orservicebyitstrade
name,trademark,manufacturer,orotherwise,doesnotnecessarilyconstituteorimplyits
endorsement,recommendation,orfavoringbytheUnitedStatesGovernmentoranyagency
thereof,[Link]
expressedhereindonotnecessarilystateorreflectthoseoftheUnitedStatesGovernmentor
anyagencythereoforTheRegentsoftheUniversityofCalifornia.
Abstract
Thissectionsummarizestheriskassessmentapproachesandriskmanagementmethodsthat
canbeappliedtostudygeologiccarbonsequestrationinCalifornia,includingCO2capture,
[Link]
[Link],
[Link],pathwaysforhumanand
environmentalexposureduringoperationsandpostoperations,andhowriskschangeoverthe
timescalesassociatedwithstorageastheCO2migratesandreactsinthesubsurface
environment,arealsodiscussed.
Table of Contents
1.0 Introduction.......................................................................................................................... 1
2.0 BackgroundonRiskAssessmentandRiskManagement.............................................. 1
2.1. RiskAssessment ............................................................................................................. 1
2.1.1. Exposureassessmentaspartofriskassessment .................................................. 3
2.2. RiskManagement........................................................................................................... 4
3.0 TheProblemofUncertainty............................................................................................... 4
4.0 CarbonSequestrationHazards.......................................................................................... 6
4.1. WhatCanGoWrong?.................................................................................................... 7
4.1.1. CO2DoseResponseRelationship ........................................................................... 8
4.1.2. Hazard1:ExposureDueToPipelineLeaks.......................................................... 9
4.1.3. Hazard2:ExposureduetoLeakagefromGeologicalStorage ........................... 11
4.1.4. Hazard3:LeakagefromGeologicalStoragetoGroundwater ........................... 13
4.1.5. Hazard4:LeakagefromGeologicalStoragetoFossilFuelAssets .................... 15
4.1.6. Hazard5:LeakageEliminatesBenefitsofGeologicalStorage ........................... 15
4.1.7. Hazard6:Inducedfracturingorseismicity........................................................... 15
5.0 Discussion............................................................................................................................. 16
6.0 References............................................................................................................................. 18
1.0 Introduction
Theintentofgeologiccarbondioxide(CO2)sequestration,alsoknownasgeologiccarbon
sequestration,istopreventanthropogenicCO2fromenteringtheatmosphere,thereby
[Link],theCO2willbesafelytransportedto
sequestrationsitesandinjecteddeepunderground,whereitwillbepermanentlytrapped,with
nonegativeimpactontheterrestrialenvironmentoronundergroundresourcessuchas
[Link],carbonsequestrationwillbeneitherperfectly
[Link],and
evaluationoftheseriskswillbeimportantformakingdecisionsconcerningwhereandhow
sequestrationshouldbeperformed.
Thegoalofariskassessmentistoquantifythelikelihoodofharm(orloss)andtopresentthe
resultsinaformatthatassistsdecisionmakerswhomustacttotolerate,mitigate,oreliminate
[Link]
assessedrisks,comparethecostsofreducingtheserisktobenefitsgained,comparetherisksto
thesocietalbenefitsderivedfromincurringtherisk,andtoestablishpoliticalandinstitutional
[Link],riskmanagement,andtheir
relationshipstopolicydecisionscanbefoundinKahnemanetal.(1982),Ruckelshaus(1983),
NRC(1984and1996),andKammenandHassenzahl(1999).
Thissectionsummarizestheriskassessmentapproachesandriskmanagementmethodsthat
canbeappliedtostudygeologiccarbonsequestrationinCalifornia,includingCO2capture,
[Link],
pathwaysforhumanandenvironmentalexposureduringoperationsandpostoperations,and
howriskschangeoverthetimescalesassociatedwithstorageastheCO2migratesandreactsin
thesubsurfaceenvironment.
2.0 Background on Risk Assessment and Risk Management
Tounderstandandmanagethethreatsposedbyatechnology,decisionmakersoftenrelyon
riskassessmentasaprocesstoidentifythelikelihoodthatsomeaction(orinaction)willresultin
harm,andrelyonriskmanagementtocommunicateand/[Link]
introductiontobothconceptsastheyrelatetocarbonsequestration,inthefollowing
[Link],1996providesmoredetailifdesired.
2.1. Risk Assessment
Riskassessmentrequiresnotsimplyanevaluationofwhatdeleteriouseffectsarepossible,but
alsoanassessmentofthelikelihoodoftheseeffects.
Inanearlyandwidelycitedpaper,KaplanandGarrick(1981)notedthatariskassessment
shouldaddressthreequestions:
1. Whatcangowrong?
2. Howlikelyisittohappen?
3. Whataretheconsequences?
1
Thefirstquestionisansweredbyahazardassessment:definingaccidents,failures,orexposure
sequencesbeginningwiththeirinitiatingevent,followedbyanychainofeventsthateither
[Link]
[Link],
inwhichachemicalreleaseleadstoexposure,whichinturnleadstoaninternaldose,whichin
[Link]
[Link]
addressedbytheendstateifthesequenceoccurs,whatistheultimateimpact?which
expressesconsequencesassomenumberoffatalities,injuries,ordiseasesinthehumanhealth
riskassessment;astheexpectedeffectonspeciesorecosystemsinecologicalriskassessment;or
asdollarslostinafinancialriskassessment.
TheSocietyforRiskAnalysis(SRA2007)hasbroadlydefinedriskasthepotentialfor
realizationofunwanted,adverseconsequencestohumanlife,health,property,orthe
environment;estimationofriskisusuallybasedontheexpectedvalueoftheconditional
probabilityoftheeventoccurringtimestheconsequenceoftheeventgiventhatithas
[Link]
summationoverarangeofpotentialeventsequencesoftheconsequencesfromthatsequence
[Link]
instrumentaldefinitionofriskastheexpectedvalueofaprobabilitydistributionofimpacts.
FromthisapproachKaplanandGarrick(1981)defineriskasatripletsetcomposedof
sequences,probabilities,[Link]
quantitatively:riskistheexpectedvalueofanundesirableconsequence:
n
Risk = ( probi )(consequencei ) (51)
i=1
Here,iistheithhazard,andprobiandconsequenceiaretheprobabilityandconsequence(over
sometimeinterval)[Link],the
[Link]
anaerospacevehicle,achemicalprocessplant,acivilstructureoranuclearpowerplant,risk
maythenbeexpressedasanaverageannualprobabilityofconsequencesuchasmission
failurerate,orchronicexposurerate,orcoredamagefrequency,oracutefatalityrate.
Theexpressionofriskintermsoftheexpectedvalueofundesirableconsequences,asin
Equation51,isnottheonlymethodtodefineorquantifyrisk,andisoftennotadequateifused
[Link]
thisway,[Link],PolicyAmightbe
guaranteedtokill100peopleeachyear,whilePolicyBhas,eachyear,a1/1000chanceofkilling
100,000people.BothofthesehavethesameriskasdefinedbyEquation51(anexpected
valueof100deathsperyear)[Link]
drawbackofquantifyingriskasasinglenumberisthatthenumbersometimesdependson
quantitiesthatareveryhardtoestimate,suchastheoddsofanextremelyunlikelybut
[Link],itmaybealmostimpossibleto
quantifyeitherthe(tiny)probabilityofamajorearthquakeinducedbyCO2injection,orthe
2
(enormous)costifitoccurs,buttheriskasdefinedbyEquation51mightdependstronglyon
theproductofthesetwonumbers.
Ratherthanattempttoquantifyriskwithasinglenumber,asinEquation51,itcanbebetterto
generatemoresophisticatedsummariesofrisks,suchassummariesofthebestplausiblecase
andworstplausiblecase,orstatisticaldistributionsofcostandhazardestimates.
Riskassessmentstartswithhazardidentification,whichreferstoidentifyingallpossible
[Link]
identification,[Link]
[Link]
characterizationincludesthreeprincipalelements:(1)identifyallofthescenariosinwhichthe
negativeeffectsofthehazardwouldberealized,(2)quantifythenegativeconsequences
associatedwitheachscenario,and(3)assessthemagnitudeandsourcesofuncertaintythatlimit
theaccuracyoftheestimatesinparts(1)and(2).
2.1.1. Exposure Assessment as Part of Risk Assessment
AswillbediscussedinSection5.4,oneofthehazardsassociatedwithCO2sequestrationis
exposureofpeople,animals,[Link]
exposurefromatoxicagent,hazardidentificationcomprisesidentifyingsourcesandpathways,
andestimatingenvironmentalconcentrationsinexposuremediasuchasambientairandindoor
[Link]
willberealized,andquantifyingtheresultingexposures;thisiscalledexposureassessment.
Riskassessmentfurtherincludesquantifyingadoseresponsefunctionandusingitto
determinethenegativeeffectsofthepredictedexposures.
Exposureassessmentisanimportantcomponentofariskassessmentthataddressesresidential
[Link]
[Link]
informationontheintensity,frequency,anddurationofhumanexposurestoaharmful
[Link]
informationforrealworldscenariostoestimatethelevelsatwhichhumanpopulations
children,workers,thetotalpopulation,themostexposedindividual,orothergroupsmaybe
exposedforcomparisonwiththedoselevelsdevelopedinthedoseresponseanalysis.IfCO2
sequestrationweretotakeplaceataspecificlocationandinagivenmanner,whatexposures
wouldbeexpectedforanimals,plants,andpeopleinthearea,nowandintothefuture?
Doseresponseassessmentforaparticularsubstanceevaluatesdataontherelationshipbetween
thedoseofthesubstanceadministeredorreceivedandthelevelofaspecificadverseeffectthat
[Link](asa
resultofexposure)totheagentcorrespondwithadversehealthresponsesisthecritical
[Link],riskassessorsreviewstudiesoftestanimals
orpeopleexposedthroughtimeandtrytoascertainhowtheaccumulateddoselinkswith
[Link]
inferenceoflowdoseeffectsfromhighdoseexperimentsiscalledthelowdoseextrapolation
3
[Link]
[Link]
[Link]
anassessmentofthevariationsinresponsesuchasdifferencesinsusceptibilitybetweenyoung
[Link]
termsofbodyweight,metabolism,breathingrates,bloodflowandotherbiologicalparameters
[Link]
sequestration,thedoseresponseassessmentstepattemptstodeterminetherelationship
betweenthemagnitudeofexposuretoCO2andtheprobabilityandseverityofadverseeffects.
Ifaplant,animal,orpersonisexposedtoagivenconcentrationoveragiventimeperiod,what
(ifany)negativeimpactswillresult?
AgreatdealofinformationisavailableconcerningCO2toxicity;someoftherelevantmaterialis
summarizedinalatersectionofthisreport.
2.2. Risk Management
Oncetherisksassociatedwithanactionhavebeenidentifiedandquantified,thedecisionmaker
orregulatordevelopsabasisforevaluatingtheserisksandthen,asnecessary,developingand
[Link]
processistoestablishthesignificanceoftheestimatedrisk,comparethecostsofreducingthis
risktobenefitsgained,comparetheestimatedriskstothesocietalbenefitsderivedfrom
incurringtherisk,andcarryingoutthepoliticalandinstitutionalprocessofreducingrisk.
Therearefourtypesofanalysesthatarecommonlyusedintheriskmanagementprocess:risk
benefit,costbenefit,riskrisk,[Link]
comparisonoftherisksaddedbyanactivitytotheconcurrentbenefits(usuallyeconomic)
[Link](indollars)ofreducingrisk
tothebenefits(inequivalentdollarsoranappropriatesurrogate)gainedbyreducingrisk.A
riskriskanalysisestablishesthesignificanceofanestimatedriskbycomparingittosomeother
acceptableorfamiliarrisksuchasthoseassociatedwithbackgroundexposures,household
accidents,occupationalhazards,naturaloccurrences,[Link]
tocompareriskreductionperunitcostamongseveraloptionsfordealingwiththesamerisk.
3.0 The Problem of Uncertainty
[Link]
[Link]
uncertainty,itisimportanttodistinguishbetweenrandomvariations(orvariability),andlack
[Link],ascurrentlyconceivedandpracticed,accountsforthesetwo
typesofuncertainty(NRC1996),referredtoasaleatory(randomvariationsandchance
outcomesinthephysicalworld)andepistemic(lackofknowledgeaboutthephysicalworld).
Moreresearch(bothobservationalandtheoretical)canreduceepistemicuncertainty,but
aleatoryuncertaintycanonlybebettercharacterized,butnotreduced,withmoreresearch.
Aneffectiveandusefulriskassessmentposesachallengetotheassessorinmanyways.
Considerasanexample,[Link]
4
uncertainaboutwhetherwehavecapturedallofthemostimportanteventsinourportfolio
[Link],sourcesarenotwell
characterized,andexposurepathwaysarecomplexandvarysubstantiallyamongindividuals
andamongdifferentpopulationsintermsofage,location,andsocioeconomicstatus.
Quantifyingtherelationshipbetweenchemicalexposureandtheprobabilityofanadverse
healtheffectisalsocomplicatedbytheneedtoextrapolateeffectsfromhighdosestolower
[Link],littleistypicallyknownaboutinterindividual
[Link]
[Link]
[Link],
developingquantitativeriskestimatesinthefaceofsubstantialuncertainty,andappropriately
characterizingthedegreeofconfidenceintheresultsarerecurringthemesinhealthrisk
assessmentaswellasotherfieldsofriskassessmentecologicalrisk,accidentrisks,global
climatechange,economicrisk,etc.
Decisionmakersoftenoperateunderthepremisethat,withsufficientfunding,scienceand
technologywillprovideanobviousandcosteffectivesolutiontotheproblemsofprotecting
[Link],inrealitytherearemanysourcesofuncertainty
[Link]
[Link]
ofuncertainty,[Link]
developedtheoryofdecisionmakingunderuncertainty,whichhasbeendescribedinChernoff
andMoses(1959),Lindley(1985),andBerger(1985)amongothers.
Oneoftenusedmethodforaddressinguncertaintyinriskassessmentsisthecompoundingof
upperboundestimatesinordertomakedecisionsbasedonahighlyconservativeestimateof
[Link](as
describedinthetextscitedabove).Thislatterapproachleavesthedecisionmakerwithno
flexibilitytoaddressmarginsoferror;toconsiderreducibleversusirreducibleuncertainty;to
separateindividualvariabilityfromtruescientificuncertainty;ortoconsiderbenefits,costs,and
comparablerisksinthedecisionmakingprocess.
[Link]
[Link]
uncertainty,itisimportanttoconsideravarietyofplausiblehypothesesabouttheworld;
consideravarietyofpossiblestrategiesformeetingourgoals;favoractionsthatarerobustto
uncertainties;favoractionsthatareinformative;probeandexperiment;monitorresults;update
assessmentsandmodifypolicyaccordinglyandfavoractionsthatarereversible(Ludwigetal.
1993).Inordertomakeriskassessmentconsistentwithsuchanapproach,itshouldincorporate
aformalquantitativetreatmentofuncertaintiesintheriskcharacterizationstep.
Initsdraftreportontreatmentofuncertaintyinexposureassessment,theInternational
ProgramonChemicalSafety(IPCS2007)proposedfourtiersforaddressinguncertaintyin
exposureassessmentrangingfromtheuseofdefaultassumptionstosophisticatedprobabilistic
[Link]:
5
Tier0:Defaultassumptions:singlevalueofresult,nouncertaintycharacterization.
Tier1:Includesqualitativebutsystematicidentificationandcharacterizationof
uncertainties
Tier2:Includesquantitativeevaluationofuncertaintymakinguseofboundingvalues,
intervalanalysis,andsensitivityanalysis.
Tier3:Includesprobabilisticassessmentswithsingleormultipleoutcomedistributions
reflectinguncertaintyandvariability.
Althoughthesetiersweredevelopedforchemicalsafety,theyprovideagoodbasisfor
discussinguncertaintiesinothercontextsaswell,includingcarbonsequestration.
EvenaTier1analysismaybesufficientinsomecases:ifitisclearthatagivenriskisextremely
lowthenitmaybedeemedacceptablewithouttheneedtoquantifyit,whileanextremely
highriskmaybeunacceptable.
Tier3riskassessmentsareusuallyexpectedtobebettermoreinformativeandmoreaccurate
[Link],however,theaddedcomplexityofTier3
[Link]
statisticalmodels(orparametervalues)usedinaTier3analysismaybesohighthatthereis
littlepointinproceedingpastaTier2analysis.EitheraTier2orTier3analysisrequires
attemptingtoidentifyallsignificanthazards,toquantifytheprobabilityofoccurrence,andto
determinewhichhazardscontributethemosttotherisk;assuch,eithercanbeextremely
valuabletodecisionmakers.
Riskmanagersmustconsidertheirroleinreducingorincorporatinguncertaintyaswellasin
[Link](1)investingtocollect,storeand
manageinformation;(2)investingtoimprovetheknowledgebase(i.e.,togeneratenew
knowledge);(3)formalizingprocessesusedtocollect,useandprocessinformation;(4)
formalizingprocessestocalculateandcommunicateuncertainty;and(5)adjustingtherisk
assessmentprocesstomitigatethepracticalimpactofuncertainty.
4.0 Carbon Sequestration Hazards
Nowthatwehavediscussedtheriskassessmentandriskmanagementingeneralterms,we
[Link]
hazardsfromcarbonsequestrationareeconomic,ecological,andhumanhealthrelated,and
havebeenstudiedextensivelybyBensonetal.2007andtheIPCC2005(especiallyChapters5
and6).Ofthehazardsreported,thosepertainingtosequestrationinCaliforniainclude:
1. Injury,death,orecosystemdamagecausedbyexposuretoharmfullevelsofCO2;
2. DamagetogroundwaterresourcesthroughdirectcontaminationwithCO2,or
contaminationbyleachingoftoxicmaterialfromsurroundingrock,duetoacidification
ofwaterbydissolvedCO2;
6
3. Damagetomineralresourcesthrough(a)contaminationofnaturalgasoroilbyCO2,
(b)pressureinducedmigrationofmobilefossilfuelsinawaythatcomplicates
extractionorrendersitinfeasible,(c)sorptionofCO2intocoal,or(d)alterationof
economicallyimportantores,minerals,gravels,etc.,byCO2orbyacidifiedwater;
4. Inducedorenhancedseismicityduetoincreasedfluidpressuredeepunderground;
5. Injuryordeathtoworkersthroughheavymachineryaccidentsandotherindustrial
accidentsnotcoveredinitem1above.
Therisksassociatedwiththehazardslistedabovearenotfixedquantities:theydependonwhat
sitesareusedforsequestration,onthepipelinetechnologiesandoperatingpracticesusedto
detectandfixleaks,[Link]
sequestrationsites,pipelineroutes,andmanagementstrategieswillbeanimportantchallenge
forriskassessorsifwidespreadcarbonsequestrationistobecomeareality.
[Link]
example,CO2pipelinesalreadyexist(mostlytoprovideCO2forenhancedoilrecovery),and
[Link]
[Link]
hasalsolongbeenusedforpollutantsotherthanCO2,[Link]
ofCO2forenhancedoilrecoveryoperationscloselyparallelsinjectionforcarbonsequestration.
Inshort,thereisalreadyafairamountofrealworldexperiencethatcanhelptoquantifythe
riskassociatedwiththehazardsidentifiedabove.
4.1. What Can Go Wrong?
CO2sequestrationwillinvolvelargeconstructionprojects.CO2extractionequipmentwillbe
installedatpowerplants;pipelinesandpumpingstationswillbebuilt;andinjectionwellsand
[Link]
[Link]
wellunderstoodandarenotuniquetocarbonsequestration.
Theprevioussectiondiscussessomeofthehazardsassociatedwithgeologiccarbon
[Link]
[Link],theliteratureonthistopichas
identifiedmechanismsandscenariosforhowCO2transportandstoragecouldcauseharmto
humans,toecosystems,andtoresources;[Link]
considersomeofthesehazardsinmoredetail.:
1. LeakageofCO2frompipelinesorpumpingstationstothesurfaceorshallowsubsurface
couldleadtodamagingorlethaleffectsonpeople,animals,andplantsthrough
asphyxiationortoxicity.
2. LeakageofCO2fromgeologicsequestrationsitestothesurfaceorshallowsubsurfacecould
leadtodamagingorlethaleffectsonpeople,animals,andplantsthroughasphyxiation
ortoxicity:
7
a. GradualleakagecouldallowCO2tobuildupintopographicdepressionsorin
enclosuressuchasbasements.
b. GradualleakageofCO2fromgeologicstoragesitesintolakescouldcreatea
deepwaterlayerofwaterthatissaturatedCO2;overturningofthelakecould
rapidlyreleasethiswatertotheatmosphere.
LeakageofCO2fromgeologicrepositoriesintogroundwatercoulddegradewater
quality,eitherdirectlyorbyacidifyingthewatersothatitleachestoxiccompoundsout
ofrockorsoil.
LeakageofCO2fromgeologicrepositoriesintofossilfuelassetssuchasnaturalgasoroil
depositscouldincreasethecostsofrecoveringorrefiningthoseassetsorcouldrender
themunusable.
LeakageofCO2fromgeologicrepositoriescouldbesolargethatthereislittleorno
reductioninatmosphericCO2concentrationscomparedtothenosequestrationcase.
InjectionofCO2intoporousrockathighpressurecouldinducefracturing,whichcould
(a)increasepermeability,thusincreasingtheleakageriskslistedabove,or(b)activate
otherwisedormantfaults,inducingpotentiallydamagingearthquakes.
Items1and2bothinvolvetoxiceffectsfromCO2,sowebeginbysummarizingtherisksfrom
CO2exposure.
4.1.1. CO2 Dose-Response Relationship
Carbondioxideoccursnaturally,andallanimals(includinghumans)havealongevolutionary
historyofexposuretoseveralhundredpartspermillion(ppm)[Link]
[Link]
toincreaseoverthenexthundredyearsormoreduetocontinuedfossilfuelburning,andis
likelytoexceed600ppmbytheendofthiscentury.
Humans,likeotheranimals,aretolerantofCO2concentrationsmuchhigherthannormal
withoutknownilleffects.CO2concentrationsaboveabout800ppmcanleadtoaperceptionof
staleair,[Link](suchassubmarinesailors)
havebeenexposedto1000ppm(i.e.0.1%CO2)forseveralweeks,againwithnoknowneffects;
however,mostofthesehighlyexposedpeoplehavebeenhealthyyoungmales,andonemight
speculatethatsomesubgroups,suchaspeoplewithdecreasedlungfunction,mightbemore
susceptibletoharmundertheseconditions.
Fromabout1000ppm,to1500ppm,[Link]
breathingratebuthasnootherknownphysiologiceffects.From1500ppmto3000ppm,the
CO2concentrationinbloodincreasesabovenormallevels,makingthebloodacidic(acondition
calledacidosis).Asignificantincreaseinrespiratoryrate,andsomediscomfort,setinatabout
30,000ppm(3%)ofairborneCO2,andabove5%theeffectsbecomesevereandlossof
consciousnesscanoccur.
Bensonetal.(2007)notethatExperimentsconductedonagroupexposedtoupto3%CO2for
manyweeksandshorttermexposurestoevenhigherconcentrationshaveshownthatalleffects
arereversibleexceptforprolongedcoma,theconsequencesofprolongedhypoxia(lackof
8
oxygen),[Link]
atconcentrationsabove25to30%.DeathsfromcatastrophicreleasesofCO2areknownfrom
industrialaccidentsandnaturaldisasters.
Federaloccupationalsafetyandhealthregulationslimitworkplaceexposuretoanaverageof
lessthan5000ppm(0.5%)fora40hourworkweek.
Asforecosystemeffects,Bensonetal.(2007)notethatEcosystemimpactsfromexposureto
[Link]
thaninvertebratestoelevatedCO2,soanysmallscale,shorttermgasleakswouldhaveminimal
[Link],incontrast,couldsuppressrespirationintherootzoneorresultinsoil
acidification,[Link]
thecontrolledexperimentshavefocusedonthemoderateincreasesinCO2concentrationsthat
areexpectedtooccurduetoatmosphericbuildupofCO2fromthecontinueduseoffossilfuels
[Link]
increasesinCO2concentrationsstimulateplantgrowth,whiledecreasingthelossofwater
[Link],treekillsassociatedwithsoilgas
concentrationsintherangeof20to30%CO2havebeenobservedatMammothMountain,
California,[Link]
informationisavailableintheintermediaterangeof2to30%.Inaddition,informationonthe
toleranceofaquaticecosystemstoshortterm,catastrophicreleaseswasnotfoundandmay
needtoberesearched.
4.1.2. Hazard 1: Exposure Due to Pipeline Leaks
CO2frompowerplantsorotherindustrialoperationsislikelytobetransportedtostoragesites
viapipelines.AbovegroundornearsurfaceCO2pipelinesprovidethemostdirectrouteto
harmfulhumanexposureortodeleteriousimpactsonanimalsandplantsbyproducingalocal
highconcentrationofCO2andgeneratingexposuressufficienttoharmorkillpeople,plantsor
[Link](2003)formoreinformationanddiscussion.
Whileanimportantrisk,precautionsthatarecommonforexistingpipelinesfordistributing
materialsthatarefarmoredangerouscanbetakentominimizethelikelihoodofamajor
[Link],pipelinedistributionsystems,includingCO2pipelines,haveafairly
goodsafetyrecord.LongdistanceCO2pipelinesintheU.S.totalabout2600kmandproduced
noinjuriesorfatalitiesoverathirteenyearperiodstudiedbyGaleandDavidson(2002).
Geologiccarbonsequestrationwouldrequirebuildingmanymorepipelines,someofwhich
wouldlikelypassthroughorneardenselypopulatedareassincetheymustoriginateatpower
stations,sopasthistorywithCO2pipelinesdoesnotguaranteethatfuturepipelineswillbe
equallysafe.
NaturalgaspipelinesareofcoursedifferentfromCO2pipelinesinseveralsignificantways,the
[Link]
295,000milesofnaturalgastransmissionpipelinesintheU.S.,andabout1.4millionmilesof
[Link]
Administration,duringthe10yearperiodfrom19942003(inclusive),795naturalgaspipeline
9
accidentsresultedin26fatalities,97injuries,and$[Link]
fatalitiesandinjurieswereduetofiresorexplosionsassociatedwithpipelineleaks,almosta
[Link],20%ofincidentswereduetocorrosion.
SinceCO2,inthepresenceofelevatedhumidity,ismorecorrosivethannaturalgas,corrosion
[Link]
usecarbonmanganesesteel,whichisnotcorrodedbyCO2ifhumidityinthepipelineislow,
[Link]
expensive,soitwilllikelybeeconomicallyimportanttodrytheCO2beforetransport,toavoid
theneedforstainlesssteel.
PipelinedamagemayoccurmorefrequentlyinCaliforniathanelsewherebecauseofseismicity.
Outrightbreakageofpipelinesduringearthquakes(orotherlandmovementssuchas
landslides)[Link],gradual
shiftsinthegroundmaycausestressesthatleadtosmallleaks.
[Link]
concentrationofCO2willbedeterminedbythecompetitionbetweentheratethatCO2leaks
fromthepipelineandentersthedepression,andtheratethatCO2escapesthedepressedarea
[Link]
regularly,[Link]
applytoothertypesofdensegasesinpipelines,soanalysisofthisriskshouldinclude
examinationofexistingdataonotherpipelines.
AlargepipelineleakresultingfromamajorrupturewouldreleasealargequantityofCO2very
[Link],endangeringpeople
[Link]
pipelinemonitoring(andperhapsimmediatelyapparenttopeopleinthevicinityaswell).As
withtheriskofasmallleak,theriskofalargeleakisnotuniquetoCO2,sodataonothertypes
ofpipelinescanprovideusefulquantitativeinformationonthestatisticaldistributionofleak
sizesandfrequenciesthatcanbeexpected.
StepscanbetakentominimizerisksfromCO2pipelines(asfromanyhazardousmaterial
pipelines).Stepsinclude:
1. Sitepipelinesawayfrompopulousareaswhenpossible;
2. Avoidrunningpipelinesnearhomesinsheltered,populatedvalleyswhereleakingCO2
couldaccumulatetodangerouslevels;
3. Monitorpipelinesregularlyforcorrosion;
4. Monitorconstantlyforleaks;
5. Installsafetyvalvestoshutoffthepipelineintheeventofalargeleak;
6. ConsideraddingodoranttoCO2,asisdonewithnaturalgas,toallowpeopletonotice
smallleaks.WeareunawareofstudiesonthecostsoreffectivenessofodorantsforCO2;
givenCO2slowtoxicityitisquestionablewhetherthisstepwouldbeworthwhilein
10
[Link],suchaspipelinesthatnecessarilypass
throughpopulousareas.
4.1.3. Hazard 2: Exposure Due to Leakage from Geological Storage
Risk from slow, steady discharge
ReleasesofCO2thataremuchtoosmalltocausewidespreaddeathorinjurycanstillcause
acutelocalproblems,[Link],animal,andplantfatalitiesfromCO2
[Link],
Californiaprovidesagoodillustration.
MammothMountainisayoungvolcano.In1990,followingaseriesofsmallearthquakesthe
previousyear,CO2beganleakingupwardsalongfaultsfromanaturalsubterraneanreservoirof
CO2.TheamountofCO2leakagehasbeenestimatedatabout300tonsperdayfortheentire
faultsystem,whichcoversseveralsquaremiles,buttheleaksareconcentratedincertainareas
thatthereforehavelocallyhighCO2concentrations.ThehighconcentrationofCO2insoilsin
someareashaskilledover100acresoftrees,andtheairborneconcentrationsarehighenoughto
[Link]
werekilledin2006whentheyfellintoadepressioninthesnowthatcontainedaveryhigh
concentrationofCO2.
AnotherinstructiveexampleofCO2leakagetothesurfaceisprovidedbytheCrystalGeyser
[Link],withoutbeing
properlyplugged,[Link]
likeageysereverydayortwo,andcanejectupto40tonnesofCO2mixedwithwaterduringa
[Link],anddoesnothavedepressionsthat
[Link],
evenwhenstandingunderthetenfoothighjetofcarbonatedwaterejectedfromthegeyser.
However,calculations(Bogenetal.2006)suggestthatinsomeunusualweatherconditions,CO2
concentrationsveryclosetothegroundcouldreachlethallevels,therebyinjuringorkilling
smallanimalsorpotentiallyproneanimalsorpeople(e.g.,campers).
Theexamplesdiscussedaboveconcernleakagetotheopenair,butleakagedirectlyinto
[Link],
forexample,couldleadtoahighconcentrationofCO2;infact,sincewindwouldnotbeableto
causemixingthatwoulddilutetheCO2,theconcentrationinabasementwouldbehigherthan
wouldresultfromthesameleakagerateintoanoutdoordepression.
AstheMammothMountainandCrystalGeyserexamplesindicate,leakagecanoccurvia
naturallyoccurringfaultsandfractures,[Link]
thoughttopresentthemostimportantriskofleakagefromgeologiccarbonsequestrationsites,
since(ifnotproperlyplugged)theycanprovideaverylowresistancepathwayfromdeep
[Link],abandonedwellsarelikelytobepresentinsomeofthe
geologicformationsthatareattractiveforcarbonsequestration,sincesomeofthesamegeologic
11
featuresthatmakeasitealikelylocationtofindnaturalgasoroilalsomakeitattractivefor
[Link]
becausetheyhavealreadyproventobeabletoretainfluidsunderpressureformillionsof
years.
Acrudecalculationshowsthatriskofacuteinjuryduetoslowleakagefromageologic
[Link]
a1000MWcoalfiredpowerplant,for10,000days(about27years).Atarateof30,000tonnes
perday,thetotalamountofCO2pumpedundergroundis300,000,[Link]
leaksoutuniformlyoveraperiodof300,000days(820years),thiscorrespondstoaleakagerate
of1000tonnesperday,oraboutthreetimestheestimatedemissionsatMammothMountain.
Thiscalculationassumes,unrealistically,thatallofthegasleaksout,withnoneofittrappedfor
alongtimeingeologicformationsorgroundwater,anditignoresthefactthattheleakagerate
[Link],itsuggeststhatifthe
geologicsequestrationfailssothatasubstantialfractionofthesequesteredCO2leakstothe
surfacewithinafewhundredyears,therateofleakagetothesurfacecouldbehighenoughto
causeseverelocalimpacts,includingpotentiallossoflife.
Possibleactionstoreducerisksfromleakageinclude:
1. SelectsequestrationreservoirsitesthatarelikelytoretainmostoftheirCO2foratleast
thousandsofyears,notjusthundreds.
2. Locateandplugabandonedboreholes,andmonitorthemforleakage.
3. Selectsitesthatarenotnearlargepopulationconcentrations.
4. Monitorsitesforleakageandhaltoperations,orreducethetotalamountofCO2
sequestered,ifanunacceptablyhighleakagerateisdetected.
Risk from a fast, large discharge
AlthoughnaturalundergroundCO2reservoirsexistthroughouttheworldinvolcanicallyactive
areas,[Link]
scalefatalitiesfromsuchreleasesareextremelyrare,[Link]
famousexample,whichisthemostdamagingonrecord.
In1986,over1700villagersand3500cattlelivingnearLakeNyosinCameroonwerekilled
whenthelakesuddenlyreleasedabout1.4milliontonnesofCO2(1tonne=1metricton=1000
Kg,approximatelyequalto1ton).Forcomparison,a1000MWcoalfiredpowerplantemits
about30,000tonnesofCO2perday;theLakeNyosreleaserepresentedabout50daysof
emissionsfromalargepowerplant.
TheLakeNyoseventoccurredasfollows.CO2fromvolcanicactivityhadaccumulatedfor
[Link],theCO2
remainedinsolution([Link]),butwaterthatmovedupwardsbyevenasmall
amountexperiencedlowerpressure,sothatCO2cameoutofsolutionandformedbubbles,just
[Link],
drawingmoreofthedeep,saturatedlakewaterupwiththemandthuscontinuingtherelease.
12
CO2isheavierthanair.TheCO2floweddownthevalleybelowthelake,smotheringseveral
villages.
Theeventthattriggeredtheinitialupwardmovementofsaturatedwaterisnotknown,butmay
[Link]
saturation,evenasmalldisturbancecantriggerarelease,andfuturereleasesatLakeNyosare
inevitableunlessmitigationisperformed:thelakestillcontainsanestimated300,000tonnesof
CO2,andnatural(volcanic)CO2entersthelakeatanestimated5000tonnesperyear.
TheCO2releaserateatLakeNyos,between1.2and1.6milliontonnesofCO2releasedtotheair
withinhours,wasvastlyhigherthancouldoccurbydirectleakagefromageologic
sequestrationsitetotheearthssurface:thelakehadaccumulatedCO2thatleakedfrom
[Link]
leakagefromageologicsequestrationsiteifageologicCO2sourceleaksintoalakeunder
[Link]:(1)thesequestrationreservoirwouldhavetofail,
sothatCO2leaksatafairlyrapidrateintoadeepportionofthelake,throughafractureorwell
thatconnectstotheundergroundsource,(2)thelakeisverticallystratifiedsothatdeepwater
doesnotcirculatetothesurfaceandtherebydissipateitsCO2asquicklyasitaccumulates,and
(3)thelakeisdeepenoughthataverylargeamountofCO2canaccumulatebeforea
[Link](4)
thereisasubstantialhumanpopulationnearthelake,nearorbelowthelakessurfaceelevation.
GiventhespecialconditionsthatarerequiredforaLakeNyostypeevent,theriskofsuchan
[Link]
example,thereareonlythreeknownsitesintheworldwheredeeplakewaterissaturatedwith
CO2fromnaturalsources.
SeveralstepscanbetakentominimizethechanceofaLakeNyostypeevent:
1. Selectsitesthat,eveniftheyleak,[Link]
thelaketothegeologicreservoir,thepresenceoffaultsunderthelake,andthedepthof
thelakeshouldallbetakenintoaccount.
2. Selectsitesthatarenotnearlargepopulationconcentrations.
3. Ifasequestrationsiteislocatednearalarge,deeplakeintowhichleakageispossible,
occasionalmonitoringoftheCO2concentrationdeepinthelakeshouldbeperformed.
Mitigationispossible:aninexpensive,selfperpetuatingsystemhasbeenoperatingat
LakeNyossince2001,forexample.
4.1.4. Hazard 3: Leakage from Geological Storage to Groundwater
AfractionoftheCO2injectedintothegroundwillreadilybedissolvedintowater,including
[Link],dissolutionofCO2
intowaterisdesirable,andsomesequestrationprojects(includingtheSleipnerprojectinthe
NorthSea)deliberatelyinjectCO2intodeepsaline(saltwater)reservoirs.TheCO2saturated
brineisexpectedtobecomedenserthanthesurroundingunsaturatedbrineandwilltherefore
sink,[Link]
13
formationsprobablyexceeds1000GigatonnesofCO2andmayexceed10000Gt,andisprobably
muchlargerthanthestoragecapacityofdepletedoilandgasfields.
Althoughoftendesirable,[Link]
beacidified,whichmayallowittodegradegeologicformationswhichmaytoxiccompounds
[Link],water
contaminatedwithtoxinsorwithCO2couldmigrateandcontaminategroundwateraquifersor
otherresources.
Bensonetal.(2007)notethatIndustrialliquidwastedisposalbydeepwellinjectionwas
[Link],ninebilliongallonsper
yearofhazardous,industrialandmunicipalwastesareinjectedinto485ClassIwells.In1987,
thecostofliquidhazardouswastedisposalrangedfrom$49to$207/[Link]
wasmixed,withmanyexamplesofwellfailuresandcontaminationofdrinkingwateraquifers.
Failureswereattributedto(1)poorcharacterizationoftheconfiningunits;(2)improperwell
completiontechniques;(3)useofwellconstructionmaterialsthatwereincompatiblewiththe
wastestreamsand,consequently,corroded;(4)inconsistentorinadequatemonitoring;and(5)
[Link]
monitoring,progressivelymorestringentregulationswereputinplacetomakethepracticeof
industrialwastedisposalbyliquidinjectionsafer.By1988,thecurrentsetofregulationswas
[Link]
reported.
ContaminationofdrinkingwaterwithCO2,orwithmetalsorchemicalsleachingoutofrock
duetoCO2acidificationofthewater,couldbeverycostlyiftheaffectedaquiferisusedbyalot
[Link]:onceanaquiferis
contaminated,itwillremainso.
[Link],experiencewithundergrounddisposalofliquid
chemicalsprovidesusefulinformationabouttheabilitytoassessgeologicformationsfor
sequestrationpotential,[Link],someoftheinformationisnot
perfectlyapplicabletoCO2because(1)thequantityofCO2thatwouldbeinjectedinto
sequestrationsitesdwarfstheamountoftoxicmaterialinjectedintowastestoragesites;and(2)
CO2isbuoyant,whereasmosttoxicmaterialinjectedintowastewellsisapproximatelyneutral.
Ontheotherhand,(3)CO2isfarlesstoxicthanmostmaterialsthatareinjectedforhazardous
wastedisposal.
Severalstepscanbetakentominimizethechanceofleakagetogroundwater:
1. Selectsitesthatarefarfromaquifersthatareormayinthefuturebeusedforpotable
water.
2. Selectsitesthat,eveniftheyleak,[Link]
conductivityofthesurroundinggeologicformations,theproximityofthegroundwater,
andthepresenceoffaultsshouldallbetakenintoaccount.
14
3. Ifasequestrationsiteislocatednearanimportantgroundwateraquifer,occasional
monitoringoftheCO2concentrationdeepintheaquifershouldbeperformed.
4.1.5. Hazard 4: Leakage from Geological Storage to Fossil Fuel Assets
Muchlikehazardsofleakagefromgeologicalstoragetogroundwater,undergroundinjectionof
CO2atpressuressubstantiallyhigherthanformationpressurescanleadtoseepagetofossilfuel
depositsalongfaults,fissures,[Link],
CO2contaminationmaydecreasethevalueofthefossilfuelsowingtoexcessivecostsfor
removing,orseparating,CO2beforethefossilfuelcanbetransported.
Humanhazardsandrisksarelargelysecondary,incurredmostlyfromthestepstakento
[Link],capture,storage,andre
sequestrationeachcreatehazardstohumanexposureandaredescribedintheabovesections.
Theprobabilityofleakagetoafossilfuelassetislikelytobesimilartotheprobabilityofleakage
[Link]
faults,orunknownandunsealedwellsinthevicinityofthestoragesiteandthefuelasset.
Possibleactionstoreducerisksfromleakagetofossilfuelsinclude:
1. SelectsequestrationreservoirsitesthatarelikelytoretainmostoftheirCO2foratleast
thousandsofyears,notjusthundreds.
2. Selectsitesthatarenotnearfossilfuelsassets.
3. Monitorsitesforrapidleakage([Link])andhalt
operations,orreducethetotalamountofCO2sequestered,ifrapidleakageisdetected.
4.1.6. Hazard 5: Leakage Eliminates Benefits of Geological Storage
Leakage,inadditiontoenablingthehazardsdiscussedintheprevioussubsections,alsoreturns
carbonintotheatmosphere.SincethewholepointofcarbonsequestrationistopreventCO2
fromenteringtheatmosphereforhundredsorthousandsofyears,leakageatamuchfasterrate
[Link],ifthefractionofsequesteredCO2that
leakswithinafewhundredyearsisveryhigh(morethan70%orso)thensequestrationcould
actuallybeworsethansimplyemittingtheCO2totheatmosphereinthefirstplace:additional
CO2isproducedingeneratingtheenergyneededforsequestration(forremovalofCO2fromthe
wastegasstream,forpipelinetransport,andforpumpingunderground),sothetotalamountof
CO2generatedishigherifsequestrationisperformedthanifitisnot.
Inadequatelysealedboreholes(fromfossilfuelextractionorexploration)arethoughtbethe
[Link]
[Link]
adequatelysealed,andadditionalworkcanbeperformedifnecessarytosealthembetter.
4.1.7. Hazard 6: Induced Fracturing or Seismicity
GeologiccarbonsequestrationrequiresinjectinglargequantitiesoffluidsupercriticalCO2
[Link]
rockislesslikelytofracturethanislowporosityrock,sinceporousrockwillallowmorefluid
15
[Link]
aretheinjectionrateandthestrengthoftherock.
[Link]
rockformationsthatareintendedtoconfinetheCO2typicallytheformationsabovethe
injectionsitethencontainmentmaybelostandCO2mayescapetocontaminategroundwater,
[Link],instead,fracturingoccursintheformationsthatareintendedto
storetheCO2,thenfracturingmaybeirrelevant,aslongastheconfiningformationsremain
intact.
Anadditionalhazardisinducedseismicity:[Link]
Nicholson(1987)notedthatinjectionusuallytriggersactivityinaseismicallyunstablearea,
[Link]
pressurizedfluidsthemselvesarenotlargeonageologicscale,andcannotprovideenough
[Link],however,triggerthereleaseof
energyalreadystoredinafault.
SminchakandGupta(2002)saythatWhilemostseismiceventstriggeredbydeepwell
injectionaretoosmalltobenoticed,moderateearthquakeshaveoccurredduetoinjection
[Link],twoearthquakeswithRichtermagnitudeof5.1and5.2weretriggeredin
Denver,Colorado,in1966.In2007ageothermalprojectinBasel,Switzerland,wasputonhold
(andcouldpossiblybecancelled)aftertwosmallearthquakes(Richtermagnitudes3.1and3.4)
werecausedbyinjectionofwaterintohotrocksatadepthof5km.
Possibleactionstoreducerisksfrominducedfracturingorseismicityare:
1. Choosesequestrationsitestoavoidformationsorlocationsthatarepronetofractureor
earthquakes.
2. Setinjectionratessothatpressuresremainbelowthefracturepressureofsurrounding
rocks.
3. Monitorforseismicactivityandreduceinjectionratesiftheactivityistooextreme.
Somelevelofinducedseismicitymaybedeemedtolerablebydecisionmakers,butgaining
[Link],theoccasionalearthquakeofRichter
magnitude2.5wouldcausenodamageandindeedwouldbeunnoticedbymanypeople,but
mightstillprovepsychologicallyunsettlingtomanypeople.
5.0 Discussion
Geologiccarbonsequestrationpresentstheopportunitytogreatlyreducetherateofinjectionof
CO2intotheatmosphere,[Link]
carbonsequestrationalsoentailsrisk,andtheriskisunknownandmaypossiblybelarge.
[Link]
shutoffvalves;sequestrationsitescanbechosenthatareawayfromhighlypopulatedareas,
and/orareknowntohavecontainednaturallyoccurringnaturalgasforalongtime;andsoon.
16
[Link]
tomographymethodstocharacterizelikelyleakagepathways;computermodelscanbe
improvedtocapturepreviouslyignoredphenomenaortoincorporatehigherspatialresolution;
datafromsequestrationsitesalreadyinusecanbeusedtobetterquantifygeologicparameters
andotherquantities.
[Link],quantifyingrisk
[Link]
[Link]
informationcoulddiscoverpreviouslyunknownproblems,suchasfaultsorfractures,thus
[Link]
coulddemonstratethattherearefewfaultsorfractures,andconfirmthatthesiteisacceptable.
Theriskhasntchanged,buttheknowledgeoftheriskhaschanged,andthiscanaffect
decisions.
Predictingthefuturecourseofacarbonsequestrationsiteisparticularlychallengingbecause
thefuturestretchesforsuchalongway:tobeeffectiveatlimitingclimatechange,studies
suggestthatsequestrationsiteshavetoretainmostoftheirCO2foratleastseveralhundredsof
years,[Link]
timescales,butverylongcomparedtothetimescalesoverwhichdeepsubterraneanphenomena
havebeenmonitoredorpredicted.
[Link]
technologiesandmethodsimprove,computermodelsimproveinbothtemporalandspatial
resolutionandintheirabilitytohandlenewphenomena,geophysicalunderstandingincreases,
andeveryyearthatpassesprovidesanotheryearofdatafromexistingcarbonsequestration
[Link],theinformationneededtomakedecisionsaboutsiteacceptabilitygetsbetter
withtimeandwithexperience.
Thesituationcallsforadaptivedecisionmaking:
Startwithsites,technologies,andactionsforwhichtheriskisbelievedtobeacceptable
now,[Link],[Link]
thisprinciplemayimplylimitedinitialdeploymentofgeologiccarbonsequestration.
Insomecases,theinformationthatcanbegainedfromanaction,orfromusingacertain
site,maybeanimportantcriteriontoconsider,inadditiontotheprobabilityofsuccess.
Iftwositesortechnologiesarejudgedtobeacceptableintermsofriskandtotalcost,but
onewouldprovideinformationthattheotherwouldnot,theonethatprovidesnew
[Link],eachsequestrationsiteis
notjustaplacetodiscardCO2,itisalsoanexperimentthatwillprovidevaluabledata
foruseinfuturesequestrationefforts.
17
6.0 References
Benson,S.M.,[Link],[Link],[Link],[Link]
[Link]
BerkeleyNationalLaboratoryreportLBNL51170.
Berger,[Link],NewYork.
BogenK.T.,[Link],[Link],[Link]/GISModelfor
[Link]
[Link]
SymposiumonSiteCharacterizationforCO2GeologicalStorage,BerkeleyCalifornia,
March2006.
Chernoff,H.,[Link],[Link]
byDover,NewYork,1986.
Gale,J.,[Link].TransmissionofCO2:[Link]:Sixth
InternationalConferenceonGreenhouseGasControlTechnologies,GaleJandKayaY,
[Link],p.517522,2003.
IntergovernmentalPanelonClimateChange(IPCC).2005.IPCCSpecialReportonCO2Capture
andStorage.
Kammen,D.M.,[Link]?:ExploringEnvironmental,
Health,[Link],
Princeton,[Link],P.,[Link],[Link]
Uncertainty:[Link],NewYork.
Kaplan,S.,[Link].1(1):1127.
Lindley,[Link],NewYork.
Ludwig,D,[Link],[Link],resourceexploitation,and
conservation:Lessonslearnedfromhistory.Science260:[Link].
McKone,[Link]:TheMerging
ofScience,Judgment,andValues.FoodControl7(2):6976.
NationalResearchCouncil(NRC).[Link]
AcademyPress,WashingtonD.C.
NationalResearchCouncil(NRC),[Link]:InformingDecisionsinaDemocratic
[Link],WashingtonD.C.
Ruckelshaus,[Link],Risk,andPublicPolicy.Science221:10261028.
SocietyforRiskAnalysis(SRA)[Link]/resources_glossary.php.
18
SminchakJ.,[Link],[Link],[Link]
[Link].
Wesson,R.L.,[Link].
[Link].OpenFileReport87331.
19