FEMSMicrobiologyEcology,92,2016,fiw166 doi:10.1093/femsec/fiw166 AdvanceAccessPublicationDate:8August2016 Minireview MINIREVIEW Hydraulic fracturing offers view of microbial life in the D o w n lo deep terrestrial subsurface a d e d Paula J. Mouser1,∗, Mikayla Borton2, Thomas H. Darrah3, Angela Hartsock4 from h and Kelly C. Wrighton2 ttp s ://a c 1DepartmentofCivil,EnvironmentalandGeodeticEngineering,TheOhioStateUniversity,2070NeilAvenue, a d e 470HitchcockHall,Columbus,OH43210,USA,2DepartmentofMicrobiology,TheOhioStateUniversity,484 m ic West12thAvenue,Columbus,OH43210,USA,3DivisionsofSolidEarthDynamicsandWater,Climateandthe .o u p Environment,SchoolofEarthSciences,TheOhioStateUniversity,125SouthOvalMall,275MendenhallLab, .c o Columbus,OH43210,USAand4DepartmentofBiology,UniversityofAkron-WayneCollege,1901Smucker m Road,Orrville,OH44667,USA /fem s ∗Correspondingauthor:DepartmentofCivil,EnvironmentalandGeodeticEngineering,TheOhioStateUniversity,2070NeilAve,470HitchcockHall, ec /a COonleumsebnutesn,OceHs4u3m21m0,aUryS:AM.iTcerlo:b6i1a4l2c4o7m44m2u9;nEit-imesaiinl:[email protected],redoxandchemicaladditives. rticle Editor:GerardMuyzer -a b s tra c ABSTRACT t/92 /1 1 Horizontaldrillingandhydraulicfracturingareincreasinglyusedforrecoveringenergyresourcesinblackshalesacrossthe /fiw globe.Althoughnewlydrilledwellsareprovidingaccesstorocksandfluidsfromkilometerdepthstostudythedeep 1 6 biosphere,wehavemuchtolearnaboutmicrobialecologyofshalesbeforeandafter‘fracking’.Recentstudiesprovidea 6/2 frameworkforconsideringhowengineeringactivitiesalterthisrock-hostedecosystem.Wefirstprovidedataonthe 40 2 geochemicalenvironmentandmicrobialhabitabilityinpristineshales.Next,wesummarizedatashowingthesamepattern 9 1 acrossfracturedshales:diverseassemblagesofmicrobesareintroducedintothesubsurface,eventuallyconvergingtoalow 8 b diversity,halotolerant,bacterialandarchaealcommunity.Datawesynthesizedshowthattheshalemicrobialcommunity y g predictablyshiftsinresponsetotemporalchangesingeochemistry,favoringconservationofkeymicroorganisms ue s regardlessofinputs,shalelocationoroperators.Weidentifiedfactorsthatconstraindiversityintheshaleandinhibit t o biodegradationatthesurface,includingsalinity,biocides,substratesandredox.Continuedresearchinthisengineered n 0 ecosystemisrequiredtoassessadditivebiodegradability,quantifyinfrastructurebiocorrosion,treatwastewatersthat 9 A returntothesurfaceandpotentiallyenhanceenergyproductionthroughinsitumethanogenesis. p ril 2 Keywords:fracking;shale;Marcellus;halotolerantbacteria;naturalgas;methanogens 01 9 ANOVERVIEWOFBLACKSHALEGEOLOGY record (Tourtelot 1979; Wignall 1994). Black shale basins are ANDTHEHYDRAULICFRACTURINGPROCESS found across the globe, with the largest known reserves in China, North America, Russia, Argentina, Australia, Brazil, as Organic-richmudstones,commonlycalledblackshales,consti- well as several African and European countries (Fig. 1A). Al- tute∼35%ofallsedimentarydepositsmakingthemthemost though exploited for various purposes (e.g. reservoirs for car- abundant rock type preserved in the global sedimentary rock bonsequestration,targetsfornuclearwastestorageandsources Received:8April2016;Accepted:12July2016 (cid:4)C FEMS2016.Allrightsreserved.Forpermissions,pleasee-mail:[email protected] 1 2 FEMSMicrobiologyEcology,2016,Vol.92,No.11 D o w n lo a d e d fro m h ttp s ://a c (D) a d e m ic .o u p .c o m /fe m s e c /a rtic le -a b Figure1.(A)Globallocationsofblackshaleformationswithassessednaturalgasresourceestimates(gray)andformationsthathavenotyetbeenassessedfortheir s resourceestimates(brown);basemapmodifiedfromtheUSEnergyInformationServicesAssessedResourcesBasinMap,2013.Readersarereferredtoreferenced tra c microbialstudiesfromUSshalegasbasinssummarizedinTable1.(B)Thenaturalfracturenetworkand(C)acarbonate-mineralizedveinfromMarcellusshalecore; t/9 (D)backscatteredelectronimageofamechanicallypolishedsectionofUtica/PointPleasantshale,horizontalfieldwidth∼300μm.Organicmatterandporesaredark 2 /1 graytoblack,silicatesanddolomitearemediumgray,calciteislightgrayandpyriteiswhite;(E)secondaryelectronimageofadualbeamFocusedIonBeam/Scanning 1 ElectronMicroscopy(SEM)sliceofMarcellusShale,horizontalfieldwidth∼20μm.Organicmatterandporesaredarkgray,pyritemediumgrayandsilicates/carbonates /fiw lightgray.ImageacquiredattheMolecularFoundry,LawrenceBerkeleyNationalLaboratory(LBNL);(F)SEMimageofmicrobialcellsandotherparticulatescollected 1 6 ona0.2μmfilterfromMarcellusshaleflowbackfluids.SEMimagesacquiredattheSubsurfaceEnergyMaterialsCharacterizationandAnalysisLaboratory,Schoolof 6 EarthSciences,TheOhioStateUniversity. /24 0 2 9 for precious metals), interest in black shales for energy de- Phanerozoic,whichincludesthelateOrdovician,themiddleDe- 1 8 velopment has intensified over the last decade following the vonian,thelateCarboniferoustoearlyPermian,thelateJurassic b y widespreadapplicationoftwonewlycoupledtechnologies:hori- andthemiddleCretaceous(ArthurandSageman1994;Wignall g u zontaldrillingandhydraulicfracturing.These‘unconventional’ 1994;Ryderetal.1998;Sagemanetal.2003;Pollastro2007;Arthur es formation development methods have dramatically expanded et al. 2008; Martini et al. 2008; Straeten et al. 2011; Trabucho- t o n recoveryoftightlyboundhydrocarbonsfromlowpermeability Alexandre et al. 2012; Kerschke and Schulz 2013). During the 0 9 rockmatrices(Kerr2010;Touretal.2010).Asaresult,natural formationofblackshales,depositionoforganic-richmaterials A p gUaSsafnrodm6%blaocfkgslohbaalelsncauturrreanltglyasacpcrooudnutcstifoonr,mwoitrhepthearcnen30ta%geosf oincgcufrrroemduhnigdherpcroimndairtyiopnrsoodfulcotwiviotyxyogrenra(pWidigsneadllim19e9n4t)a,trieosnulitn- ril 2 0 expectedtogrowinthecomingdecades(EIA2013).Despitethe stagnantmarinesystems(Tourtelot1979;WignallandNewton 19 importanceofblackshalestohumankind,westillhaveasur- 2001).Stagnationandanoxiapromotedefficientorganiccarbon prisinglypoorunderstandingofthephysical,chemicalandbi- preservation(Tourtelot1979;Wignall1991a,b,1994),influencing ologicalprocesseswithinfracturedblackshalesandfluidsre- itsdarkgraytoblackappearance(Fig.1BandC)andgeochem- coveredfromtheseformations,especiallyascomparedtoother ical signature (Wignall 1994), with black shales having higher hydrocarbon-bearingformationssuchassandstonesandlime- carboncontentandtypicallymorereducedoxidationstatethan stones.Theseaspectswilldirectlyinfluencetheshalewell’smi- grayshales.Blackshaleshaveporositiescommonly<5%(Soeder crobialecosystemsaswellasthewastefluidstheyproduce. 1988;Onstottetal.1998;Curtis2002;RyderandZagorski2003; Blackshalesareorganic-richsedimentaryrocksdominated Lash2006;LashandBlood2007;Buchwalteretal.2015)andsmall by fine- to extremely fine-grained silt, clay and quartz grains pore-throatsizes(Fredricksonetal.1997),resultinginexceed- containingvariableamountsofcarbonates,organicmatterand inglylowpermeabilitiesinthenano-Darcyrange(Onstottetal. pyritization (Fig. 1B–E) (Tourtelot 1979; Wignall 1994). These 1998;Lashetal.2004;LashandBlood2007;Engelderetal.2009, shaleswerelargelyformedduringfivegeologicperiodsinthe 2012). Mouseretal. 3 D o w n lo a Figure2.(A)Typicalcrosssectionofahorizontal-drilledandhydraulicallyfracturedwellintheUSAppalachianBasinand(B)idealizedcartoonoffractureinfluencing d e biogeochemicalchangesafterfracturing,includingnaturalfracturesandveinsintheshale,zonesoforganicmatterandinjectedcompoundsincludingbiomass. d fro m Theageofblackshaleimpactsburialdepth,paleopasteur- widthofahumanhairandpropagatingthesefracturesseveral h izationtemperaturesandpressures,aswellasthetimeframe hundredmetersintotheformation(Fig.2B)(Arthuretal.2008; ttp s ffoacrtohrysdraofcfoarrdboanwdiedveelorapnmgeento.fTpoogteetnhteiar,l vhaarbiaittaiotsnsfoirnbtuhreiesde Kteirnagn2d01te0n).sFroafctthuoruinsganudsessoofnlitaevresroafgceh1e0m–1i5camlsil(l∼io1n%liotefrtsheoftowtaa-l ://ac a (indigenous)microorganisms,ormicrobesintroducedoverge- volume)duringthedevelopmentofonehorizontalproduction d e ologictime.Moreover,thedepthsoftheshaleformationscur- well(ClarkandVeil2009;Rush2010;Gregoryetal.2011;Nicot m ic rentlybeingconsideredforpetroleumdevelopmentrangefrom and Scanlon 2012; Olsson et al. 2013; Jackson et al. 2014; Ven- .o u less than 300 m to more than 4000 m below the surface with goshetal.2014).Fracturingfluidsarecomposedofwater,sandor p .c varying thicknesses (20–200 m) (Evans 1995; Ryder et al. 1998; proppants,acidandmanyorganicandinorganicchemicaladdi- o m Pollastro2007;Arthuretal.2008;Martinietal.2008;Langeetal. tivesthatcleanthewellbore,movetheproppantintonewlycre- /fe 2013;Olssonetal.2013),impactingcurrentinsitupressureand atedfissures,alterwettabilitybetweenthefluidandtheshale, m s temperatures, which are known constraints for microbial life protectthewellfromcloggingandcorrosionandlimitgrowthof e c (RothschildandMancinelli2001;PicardandDaniel2013).Black microorganismsdetrimentaltoproduction(Rush2010;Soeder /a shaleshavetotalorganiccarbon(TOC)contenttypicallyranging 2010;Gordallaetal.2013;Stringfellowetal.2014;Kekacsetal. rtic le from2%to9%,withvaluesashighas20%(Jenden,Drazanand 2015;ElsnerandHoelzer2016).Microbialactivityinoilandgas -a b Kaplan1993;ArthurandSageman1994;Wignall1994;Kerschke wellscancausenumerousproductionissuesrangingfromgas s andSchulz2013).Inadditiontothecarbonamount,thedistri- souring and biocorrosion to reduced efficacy of chemical ad- tra c butionandlabilityoforganicmaterialscanvarywidely(Fig.1D ditivesandfractureclogging.Understandingthediversityand t/9 2 andE)basedondepositionage,grainsizeanddiagenetichistory, functionofmicrobesinhabitingblackshalescanthereforepre- /1 1 withramificationsformicrobialmetabolismandgrowthrates servecapitalinvestments(e.g.wellinfrastructure)andimprove /fiw (Schlegeletal.2013;Wuchteretal.2013;Buchwalteretal.2015). theoveralleconomicsofenergyrecoveredfromthesesystems. 1 6 Thesetypesoflowpermeability,highcarboncontentrockfor- Thisreviewtakesabroadlookatmicrobiallifeinblackshales 6 mations are considered unconventional oil and gas reservoirs prior to and after hydraulic fracturing, building on a previous /24 0 becauserelativetosandstoneorlimestoneformationsthatcon- summaryofmicrobialdynamicsandcontrolinshalesystems 2 9 stituteconventionalreservoirs,hydrocarbonsareeitherphysi- byGasparandcolleagues(2014)withacomparativeanalysisof 18 callytrappedbythesmallporethroatsorfailtodiffuseoutof publicallyavailable16SrRNAsequencedatacollectedacrossge- b y theshalematrixateconomicallyrecoverablerates.Asaresult, ographicallydifferentshales,andincorporatingmorerecently g u hydrocarbonrecoveryfromblackshaleshasonlybecomefeasi- publishedinformationonthefunctionalpotentialofkeytaxa es blebycouplinghorizontaldrillingandhydraulicfracturingstim- generated from metagenomic data and shale isolate studies. t o n ulationtechnologies. Thisbodyofliteraturesuggeststhathydraulicfracturingintro- 0 9 Engineered phases in the development of a shale well in- ducesbiomassandcreatesnewgeochemicalconditionsinthe A p cshluadleetfhoremfoaltloiowninwgitphriimnsatrayllsatteipons:o(if)vmeurtlitcipalledrpirllointegcttoivaebcoavseinthges dbeetewpeteenrredsiftfreiarelnsutbsshuarlfeacfoertmhaattisounpsp.oGritvemnictrhoebinaalsccoelnontiizmatpiloen- ril 2 0 to safeguard sensitive aquifers and formations during drilling mentationofhydraulicfracturingandtheinaccessibilityofre- 19 activities, (ii) horizontal/lateral drilling within black shale to searcherstothesesites,moststudiestodateare16SrRNAgene lengthsuptoseveralthousandmeters,(iii)stagingthelateral surveys.Thus,muchoftheinformationtodayaboutlifeinfrac- boreholewithperforatedcasingandpackerstoseparatesmall tured shale lacks detailed insight on the metabolic functions (∼200m)regionsforfocusedstimulationand(iv)hydraulicfrac- withtwoexceptions(Dalyetal.2016;Mohanetal.2014).Wesum- turingoftheblackshaleinisolatedstagesbypositivelysurging marize the literature assessing microbial life in pristine shale alargefluidvolumeunderhighpressure(upto12000psior80 usingshalecorescollectedbycleandrillingortracermethods. MPa) into the formation, resulting in an extended network of Wealsocompiledinformationdescribingmicrobialcommuni- fracturesintotheformation(Fig.2A)(Arthuretal.2008;Kargbo ties in drill muds and waters from holding ponds (containing et al. 2010; Soeder 2010; Gordalla et al. 2013; Vidic et al. 2013). freshwaterforinjectionduringhydraulicfracturing),produced Stimulationwithhydraulicfracturingimprovesrockconnectiv- fluids(fluidsthatreturnfromtheshaletothesurfaceafterhy- ityuptoseveralordersofmagnitudebyexpandingshalefrac- draulicfracturing)andwastetanks(containersthatstorepro- turesfrommicron-sizedopeningstofracturesgreaterthanthe duced fluids for future treatment or recycling in other wells). 4 FEMSMicrobiologyEcology,2016,Vol.92,No.11 Fromthisknowledge,wediscussfutureresearchneedstoad- (Onstottetal.1998),adeepgoldmineinSouthAfrica(Onstott dressemergingindustrialandenvironmentalproblemsassoci- etal.2003;Chivianetal.2008),ahardrockrepositoryinSweden atedwithshaleenergydevelopment. (KotelnikovaandPedersen1998;HavemanandPedersen2002; Pedersen2012),adeephighertemperatureColoradosandstone shale (Colwell et al. 1997) and a shallower sandstone shale in MICROBIALLIFEINPRISTINESHALE: NewMexico(Fredricksonetal.1997;Krumholzetal.1997). EVIDENCEFORINDIGENOUSHARDROCKLIFE Withintheexistingbodyofwork,FredricksonandBalkwill’s review(2006)highlightingtheimportanceofshaleorganicmat- Ourunderstandingofmicrobialdiversityfrompristineshalehas ter in supporting microbial metabolic activities along shale- remainedpoorlyconstrained(Edwardsetal.2012;Colwelland boundingformationshasgreatlycontributedtoourknowledge D’Hondt2013)owinginparttotheexpenseandchallengesasso- of deep microbial life. These views are supported by Schlegel ciatedwithrockcoringatthousandsofmetersbelowtheground andcolleagues’(2013)characterizationoflabileandrecalcitrant surface(Lehmanetal.1995;McKinleyandColwell1996;Peder- organicmatterfromNewAlbanyshale.Thesestudiessuggest D o senetal.1997).Liketheirsubseafloorcounterparts,microbesin- thatnativeshalebacteriawithfermentativeandsyntrophicca- w n habitingdeepshalemusthavethecapacitytowithstandhigh pabilitiesareinvolvedinthebreakdownofkerogenunderelec- lo temperatures(>50◦C)andpressures(>30MPa),acidicpH,nu- tron acceptor-limited conditions, producing fermented prod- ad e trient limitations, high salinities, fluid movement constraints ucts that support methanogenesis or respiration where other d andtheabsenceoflight(Onstottetal.1998;Schrenketal.2010; electronacceptorsareavailable.Significantlyhighersulfatere- fro m Fichter et al. 2012; Olsson et al. 2013; Picard and Daniel 2013; ductionactivityandacetateproductioninshale-sandstonein- h Santillan et al. 2015). Black shale formations differ from other terface cores, as compared with core taken from within the ttp s deep seafloor and terrestrial environments in being abundant shale,suggestedenhancedmicrobialactivityoccursalongshale ://a inorganiccarbon(FredricksonandBalkwill2006;Schrenketal. formationinterfaceswhereporesizesarelargerandfermenta- c a 2010)yetconstrainedbysubmicronporespaces(Colwelletal. tionproductsdiffusetoareaswithalternativeelectronacceptors de m 1997;Fredricksonetal.1997;Langeetal.2013),featuresthatlimit (Fredricksonetal.1997;Krumholzetal.1997;Krumholz2000). ic microbialmobilityandtheadvectionordiffusionoffluidsand Alongwithsulfate-reducingandacetogenicbacteria(Fredrick- .o u gasesthroughporesandfractures. sonetal.1997;Krumholzetal.1997),iron-reducers(Colwelletal. p .c Ofalltheconstraintsonmicrobiallifeinnativeshales,space 1997)andmethanogens(Onstottetal.1998)haveallbeenculti- o m is probably the most significant. When shale pore throats are vatedfrompristineshalecores.Thesestudiessuggestthatdeep /fe small (i.e. <0.2 μm), microbial activity appears to be limited shalesofmanyages,mineralogiesandgrainsizessupportmi- m s tolarger,interconnectedporesandfractures(Fredricksonetal. crobiallifepriortofracturing. ec 1fi9ll9e7d;Bwuitchhworagltaenricetmala.2tt0e1r5()B,ourchmwaatrltixerveotidasl.,w20h1i5c)hoarrefotrympiactaiollny /artic MICROBIALANDBIOGEOCHEMICAL le fluids(i.e.gasesorbrines)(Onstottetal.1998).Ontheotherhand, -a SIGNATURESINHYDRAULIC-FRACTURED b duetothehighTOCcontent,blackshalesarefissileanddomi- s natedbyopenandinterconnectedfracturesorbedding-parallel SYSTEMS trac planesofweakness(Lashetal.2004;Engelderetal.2009).Asa t/9 result,naturalfracturesinblackshale(Fig.1B)canpermitfluid Inadditiontoindigenousorganisms,inputfluids/mediaintro- 2/1 ducedduringthedrillingorstimulationprocessvianon-sterile 1 andgasmovementincludingnutrientrechargefromfreshme- muds, equipment or well completion source waters may also /fiw teoricwateroverlonggeologicperiods(Martinietal.2003,2008; 1 provideinoculaoftheshaleenvironment(Table1).Studiesof 6 McIntoshetal.2010;OsbornandMcIntosh2010;Schlegeletal. 6 the various fluids associated with all aspects of the hydraulic /2 2011b; Darrah et al. 2015). Therefore, microbes may have mi- 4 gratedalongwithsubstrates,electronacceptors,hydrocarbons fracturingprocesshavebeguntoidentifythemicrobialtaxaand 02 9 and brines through these formations via large-scale meteoro- associatedbiogeochemicalconditionsimpactingkeyshaleen- 18 logicalconnectionsafterpaleopasturization(Colwelletal.1997; ergyandinfrastructureissuesandarediscussedbelowingreater by Martinietal.1998,2003;Onstottetal.2006;SherwoodLollarand detail. gu e Ballentine2009)enablingslowmicrobialproliferation. s Like other deep subsurface rocks, which report viable cell Drillingwatersandmudsareamicrobialinoculumto t on numbersontheorderof102–105cellsg−1or103–105cellsmL−1 0 shale 9 pore fluids (Stevens and McKinley 1995; Pedersen et al. 1997; A p K20o0te3l)n,bikioomvaaassnddePnesdietyrsiennp1r9is9t8in;Leeshhmalaeniseltoawl.,2w0i0t1h;cOenllsntoutmtebtearls. Dsurrilflaincegamnuddmsaluinbtraicianteprtehsesudrreislltbhirt,oumgohvoeurtotchkecbuotrteinhgoslet.oStphee- ril 20 estimatedat101–105cellsg−1(Onstottetal.1998)andphospho- cialblendingtanksmixwater,powderedclaysandthickenersto 19 lipid fatty acids concentrations <32 pmoles g−1 rock (Lehman producedrillingmudswithaviscosityanddensitythatbalances et al. 1995; Colwell et al. 1997; Fredrickson et al. 1997; Onstott pressureswithintheformation(Struchtemeyeretal.2014).Com- etal.1998).AlthoughtherearenoDNA-basedstudiesdescribing paredtoothermarineandterrestrialdrillingexpeditions(On- indigenousmicroorganismsinPaleozoic-agedblackshales(Or- stottetal.1998;Colwelletal.2011),muchlessisknownabout dovician,Devonian,CarboniferousandPermian,255–485million microbialcommunitiesintroducedtoshaleviathedrillingpro- yearsago)priortohydrocarbondevelopment,studiesonlater cesses. depositedMesozoic-agedshale(JurassicandCretaceous,70–200 To date, one study has described the biomass and micro- millionyearsago)alongwithothercloselyrelatedfine-grained bialassemblagesofdrillingmudsusedforblackshaledrilling. sedimentary rocks provide evidence of viable microorganisms Struchtemeyerandcolleagues(2011)compareddrillingmateri- livingintheporesand/orfracturesofsubsurfacerocks.Exam- alsfromsevenwellsinTexas’sBarnettshale(Table1).Theau- pleswheremicroorganismshavebeendescribedfrompristine thors reported that drilling waters were low in biomass (103– rockincludeayounger,butcomparabledepthshaleinVirginia 10516SrRNAgenecopies/g)untilmixedwithmudcomponents, Mouseretal. 5 Sampletypesareclassifiedbasedonmedia(i.e.mudseralwellchemistryincludeswaterqualityparameters, Reportedmetadata Ammo-SpecificDissolvedGeneralnium,organicoxygenwelltotalni-Meta-com-orThio-chemistrytrogenbolitespoundsredoxsulfateMethane XXX XXXX XXX X Downloaded from mplescollectedfromblackshalewellsites.yflowbackversuslateproducedfluids).Gen Impound-LatefluidsAgedwellsmentsor49days–2yearsSeparatorwaste2yearsoroldertankstanks EnrichmentsEnrichmentsfromthreefromonewellwellsagedaged265–17monthsmonths Samplesfromthreetimepointsfortwowellsupto328days Day82and328afterfracturinginMarcellus;enrichmentfrom98daysafterfracturinginUtica TwoTwogas-waterproducedsepara-watertorstankssampledsampledover3–6over6monthsmonths https://academic.oup.com/femsec/article-a portingmicrobialcommunitiesandassociatedmetadatainsang(i.e.storagepondsversusproducedfluids)andage(i.e.earlelements. DrillingStorageEarlySamplemuds/ponds/Injectedfluids<Formationdescriptionwaterstanksfluids49days BurkettandFoursampleMarcellusenrichments;Shale,PAthreeforTerminalrestrictionfragmentlengthpoly-morphism(TRFLP)analysis;oneforDNAsequencingMarcellusNineteenSamplesSamplesShale,PAtotalsamples;fromfrom3–4threewellsthreetimefromsamewellspointsforpad,fluidseachwellcollectedwithinfrominjectedfirst14fluids,earlydaysandlateproducedfluidsMarcellusFivesamplesSampleDay7andShale,PA;frominputfromone13afterUticaShale,andproducedwellinfracturingOHfluidsfromMarcellusinoneMarcellusMarcelluswell;onesampleenrichmentfromaUticawellBarnett17totalShale,TXsamples;twogas-waterseparatorsandtwoproducedwatertanksatseparatesites;collectedon2–6repeattimes bstract/92/11/fiw166/2402918 by guest on 09 April 2019 ofstudiesregeoffracturiorinorganic Sequencinganalysis T-RFLPandIllumina,16SrRNA 454-Roche,16SrRNAbacteriaandarchaea IlluminaHiSeqmetage-nomicsbacteriaandarchaea 454-Roche,16SrRNA Table1.Summaryversuswaters),staelectronacceptors NumberonFig.1AAuthors etal1Akob.(2015) etal2Cluff.(2014) etal3Daly.(2016) etal4Davis.(2012) 6 FEMSMicrobiologyEcology,2016,Vol.92,No.11 e n a h et M X e d Thio-sulfat X e vn data Dissoloxygeorredox Reportedmeta SpecificorganicMeta-com-bolitespounds X X D Ammo-nium,totalni-trogen ownloa Generalwellchemistry X X X ded from Impound-mentsorwastetanks https://a or ca Agedwells2yearsSeparatoroldertanks Allthreewellsmorethanadecadeold Samplesfortwotimepoints,sixwellsareco-mingledintankpriortoseparator demic.oup.com/fem Latefluids49days–2years Biofilmsamplesfromcorrosioncouponssubmergedat2–3depths sec/article-a Earlyfluids<49days Samplesfromallthreewells,timingundis-closed Day1and9afterfracturing bstract/9 2 Injectedfluids Samplesfromthreewells /11/fiw1 6 Storageponds/tanks Pre-fracpondwatersfromtwosources Sourcewaterforonewell 6/24029 1 8 Drillingmuds/waters by gue SampleFormationdescription HaynesvilleThirteentotalShale,samples;TX/AR/LAfluidsfromthreewells,differentpadsinHaynesville,comparedtoonewellinPiceanceBasin(CO);deployedsteelcartridgesatmultipledepthsinallthreewellsAntrimFourtotalShale,MIsamples;threewellssampledin2009,onealsosampledin2002BarnettFourtotalShale,TXsamples;twotimeperiodsin2012forco-mingledtankcollectingproducedfluidsforsixwells;threewerereplicatesfromsampleperiodMarcellusThreetotalShale,PAsamples;onewell,samplesfromsourcewater,earlyandlateproducedfluids.SamesamplesasetalMohan.(2013)ES&T st on 09 April 2019 Sequencinganalysis FLXamplicon,16SrRNA Sangersequencing,16SrRNAarchaeaandbacteria Illumina,16SrRNAofarchaeaandbacteria Illumina,MiSeqmetage-nomics (continued) Authors etalFichter.(2012) etalKirk.(2012) etalLiang.(2016) etalMohan.(2014) Table1. NumberonFig.1A 5 6 7 8 Mouseretal. 7 e n a h et M X e d Thio-sulfat e vn data Dissoloxygeorredox X X Reportedmeta SpecificorganicMeta-com-bolitespounds X D Ammo-nium,totalni-trogen X ownloa Generalwellchemistry X X X X X ded from Impound-mentsorwastetanks Threeimpound-ments,sampledatsurface,middleandbottomdepths https://a Separatortanks Bakkensamplefrombottomofseparatortank cademic.o u Agedwells2yearsorolder 11wells,nowellagereported p.com/fem Latefluids49days–2years Dupli-catesfrom187day Onewell,8monthsafterfracturing Marcellussamplecollected18monthsafterfracturing sec/article-a Earlyfluids<49days Dupli-catesfromthreetimepointswithinfirst9days Bakkensamplecollectedaftergelwasbroken bstract/9 Injectedfluids Dupli-catesfromonewell SamplefromMarcelluswellcon-tainingchemicalsbutnosand 2/11/fiw1 6 Storageponds/tanks Sourcewaterforonewell Twofreshwa-terholdingpondsusedforonewell 6/24029 1 8 Drillingmuds/waters by gue SampleFormationdescription MarcellusSixtotalShale,PAsamples;onewell,fluidscollectedfromsourcewaters,injectedfluids,earlyandlateproducedfluidsMarcellusNinetotalShale,PAsamples;threebiocide-treatedimpound-ment,threeuntreatedim-poundment,threepretreatedandaeratedimpound-mentEagleFord,ThreetotalTXsamples;twoholdingpondsandoneproducedwatersample NewAlbany11totalShale,ILsamplescollectedfromwellheadMarcellusFourtotalShale,PAsamples;andBakkeninjectedandShale,NDproducedfluidsfromonewellinMarcellus;producedfluidsafterbreakerandfromseparatortankinBakken st on 09 April 2019 Sequencinganalysis 454-Roche,16SrRNA Sangersequencing,16SrRNAarchaeaandbacteria 454-Roche,16SrRNA TRFLP,archaea Illumina,16SrRNA (continued) Authors etalMohan.(2013b) etalMohan.(2013a) Santillanetal.(2015) Schlegeletal.(2011a) etalStrong.(2014) Table1. NumberonFig.1A 9 10 11 12 13 8 FEMSMicrobiologyEcology,2016,Vol.92,No.11 e n a h et M X X e d Thio-sulfat e vn data Dissoloxygeorredox Reportedmeta SpecificorganicMeta-com-bolitespounds X D Ammo-nium,totalni-trogen ownloa Generalwellchemistry X X X ded from Impound-mentsorwastetanks Barnettsample60dayspost-frac;Marcellus139to5yearspost-fracfromfivetanks https://a Separatortanks Twogas-watersepara-torssampledover661–787dayspost-frac cademic.o u Agedwells2yearsorolder p.com/fem Latefluids49days–2years Samplesfromallthreewellsafter5months sec/article-a Earlyfluids<49days Samplesat24hand2months bstract/9 Injectedfluids Samplesfromblenderafterchemicalandsandaddition SamplesfromoneBarnettandoneMarcellussitepre-injection 2/11/fiw1 Storageponds/tanks Holdingpondandholdingtankafterbiocideaddition 66/24029 Drillingmuds/waters Watersandmudsfromsevenwellsites 18 by gue SampleFormationdescription BarnettFourteentotalShale,TXsamples;sevenwellsfromdifferentpads,drillingwatersanddrillingmudscollectedateachwell.BarnettEighttotalShale,TXsamples;twowellsfromdifferentpads,holdingpond,pre-fracandearlyflowbacksamplescollectedBarnettFifteentotalShale,TX;fluidsamples;MarcellustwoShale,PApre-injection,nineproducedwatertanks,fourseparators AntrimThreetotalShale,MIsamples;threecloselyspacedwells,fluidscollected5monthsafterfracturingwithN2foam.Onewellwasrefracturedthreetimes st on 09 April 2019 Sequencinganalysis 454-Roche,16SrRNA 454-Roche,16SrRNA Illumina,16SrRNAandmetage-nomics;archaea IlluminaMiSeqfor16SrRNA (continued) Authors Struch-temeyeretal.(2011) Struch-temeyeretal.(2012a) etalTucker.(2015) Wuchteretal.(2013) Table1. NumberonFig.1A 14 15 16 17 Mouseretal. 9 resulting in an order of magnitude or more increase to 106– ing the fracturing process. Microbial communities in holding 107 gene copies g−1. Drilling muds, which had lower diversity ponds were most closely related to typical freshwater bacte- than drilling fluids, were dominated by Firmicutes (29%–90%) ria,includingBacteriodetes(Flavobacterium),Betaproteobacteria (primarilyEubacteriumspp.;Thermoanaerobacter,Acetivibriogen- (Comamonadaceae)andGammaproteobacteriaincludingfam- erawithinBacilli;andRuminococcuswithinClostridia),along iliesAeromonadaceae,AlteromonadaceaeandChromatiaceae. withGammaproteobacteriasuchasPseudomonadalesandAl- Microbial communities sampled in the Barnett holding ponds teromonadales (7%–67%). In drilling waters, mixing with the were3–4timeslowerbiomassandhaddifferentmembersfrom higherbiomasscontainingmudssignificantlydecreasedtherel- those in Haynesville shale, demonstrating that source water ative abundance of dominant phyla (e.g. Actinobacteria, Aci- andsitehandlingcanplayamajorroleinthetypesandnum- dobacteria,Gemmatimonadetes).Onepossibleexplanationfor bers of bacteria sent down the well. To better identify organ- thechangeincommunitystructureandbiomassisthecompo- ismscapableofcolonizingwellinfrastructure,steelcartridges sitionofthedrillingmuds;mudscontainlabilethickeners(e.g. weredeployeddownwellandbiofilmsanalyzed.Interestingly, cellulose,nuthulls,cedarfibers,xanthamgum),formsofsul- communities observed on steel coupons, including members D o fur(e.g.barite,lignosulfonates)andcolonizationsurfaces(ben- oftheProteobacteria(e.g.Ralstonia,Sphingomonas,Pseudomonas, w n toniteclays)thatmayselectforcertaintaxaandincreaseoverall Stenotrophomonas), Spirochaeta, Acinetobacter and Acidobacterium lo a biomass(Struchtemeyeretal.2014). differed greatly from those observed in produced fluids, sug- d e gestingthatplanktonicsamplesmaynotfullyrepresentshale- d Colonizationofengineeredfluidsandinfrastructureis attachedcommunities(Fichteretal.2012).Takentogether,these from anothermicrobialsourceforshale studieshighlightthathydraulicfracturing,inadditiontoprovid- h ingfluids,nutrientsandelectronacceptorstosupportlife,also ttp s Several published studies have described the presence and providesthebiomasssourcetocatalyzethebiogeochemicalre- ://a membership of microorganisms in holding ponds, tanks and actionsoccurringinfracturedshales.However,itisunclearfrom c a blenders that serve an important on-site water management existingstudieswhetherthepresenceorabundanceofcertain de m roleduringthewellcompletionprocess(Table1).Thesestud- taxaininjectedfluidswillconstituteaninevitableeconomicloss ic ies provide additional insight into the organisms being in- forindustrythroughdetrimentalactivities. .o u jected into the deep subsurface during hydraulic fracturing. p .c Engineered fluids can host diverse phyla derived from many o Biocidesareusedtocontroldeleteriouseffects m dinifcfleurdeinntgfreAschtiwnaotbearcsteoruirac,esB(aec.tge.rloaikdeest,esgroaunnddwPartoetreoobrarcivteerrisa) ofmicrobialgrowthintheshalesystem /fem s (Alpha-,Beta-andGamma-)(Fichteretal.2012;Struchtemeyer Among the chemical additives used at these sites, the pres- ec amnedmEblsehrashhiepdi2n01th2ae;sMeoflhuaidnsetsaele.m20s13tob;bCelusftfreotnagl.ly20i1m4p).aNcotetadbblyy, etoncmeicarnodbieoflfoegcitsitvse.nTehses oofilbainodcidgeassminadyusbteryoifngvreesattsessitginnitfiecraenstt /article themixtureoffreshwatersourceswithrecycledproducedwa- resourcesinbiocidestoprotectthewellandassociatedsurface -a b toefrtafrxoamaspsroecviaioteudslywidthrilmleadrisnhealeenvwireollnsm,reensutslt(ien.gg.iCnobseigtinaa,tPusreeus- icnefrrnass,trinuccltuudrein(ge.wg.epllipsionugr,inpgo,nrdess,ertavnoikrsp)lfurogmginpgr,ogdeuncetriaotniocnono-f strac doalteromonas)(Cluffetal.2014)andbrines(e.g.Roseovarius)in toxicandmalodorousgasses,equipmentcorrosionandproduct t/92 injectedfluids(Mohanetal.2013b). degradation(seereviewbyGasparetal.(2014)onmicrobialcon- /1 1 Differences in the chemical additives and the machinery trolinthesesystems).Biocidesareappliedtosourcewatersused /fiw usedtoblendfracturefluidsmayalsorepresentasourceofmi- forinjectedfluidsaswellasimpoundmentpondsandtanksthat 1 6 crobesorselectiveagentsthatcaninfluencethemicrobialcom- store produced fluids before treatment or reuse (Mohan et al. 6/2 munityinjectedintothesubsurface.Tothisend,Struchtemeyer 2013a),withselectiondependentongeology,chemicalcompat- 40 andElshahed(2012a)characterized microbialcommunitiesby ibility,costandefficacyintargetingacid-andsulfide-producing 29 16S rRNA gene sequencing and cultivation in holding ponds, bacteria(Johnsonetal.2008).Biocidalcompoundsusedbyin- 18 tanksformixingbiocidesandsourcewaters,andblendersused dustryvarygreatlyintheircomposition,structureandmodeof by formixinghydraulicfracturingchemicalsandproppantpriorto action(seeKahrilasetal.(2014)forareviewofthesecompounds), gu e wceelllleisntjiemctaitoensiinndtwicoatBeadrnthetattsahcaidle-pwreoldlsu.cMinogstbapcrotebraiabl(eAnPuBm)abnedr abniodcisdeevseroanlsthtuedrieedsuhcatvioenevoafluviaatbeldetcheellsefafencdtivcoenmemssuonfitsypecocimfic- st on sulfate-reducingbacteria(SRB),metabolismsimplicatedinbio- position,withmixedresults(Johnsonetal.2008;Rimassaetal. 09 corrosion,werereducedbyanorderofmagnitudeafterbiocide 2011;Struchtemeyeretal.2012b;Vikrametal.2014;Erkenbrecher A p ttrheraotumghenbtl.eInndtienrge,stininjegclyt,ioanddanitdioflnoawlcboancktapcetrtiiomdesrwesituhltbeidocinidneos ecteanlt.2st0u1d5i;eSsahnativlleansheotwaln.2t0h1e5;eLffiiacnagcyetoafls.o20m1e6)o.fFothreexmamosptlceo,mre-- ril 20 1 furtherreductionincultivatablecells,suggestingapossibleinhi- monlyusedbiocides(e.g.glutaraldehyde)decreasesinthepres- 9 bitionorcross-linkingeffectfromchemicalmixingormicrobial enceofsoils(McLaughlinetal.2016),organics(Struchtemeyer tolerancetobiocidesthroughthewellcompletionunitprocesses etal.2012b)orhighsalinity(Vikrametal.2014;Erkenbrecheretal. (Kahrilasetal.2014). 2015).Additionally,certainbiocidesmayalsobemoreeffective In an effort to assess the capacity for microbial induced againstshale-specifictaxa(Liangetal.2016)whomaybecapable corrosion in unconventional gas wells, Fichter and colleagues ofdevelopingresistancetobiocidesaftercontinuedexposureto (2012)analyzedsourcewatersandproducedfluidsfromthree producedfluids(Vikrametal.2014;Liangetal.2016).Coupled wells in the Haynesville shale (TX-LA). Like the Barnett shale withlittleavailableknowledgeastohowbiocidesmayinteract studyabove,fluidswereassessedbytargetedcultivationofAPB withotherchemicaladditivesusedduringfracturing(Kahrilas and SRB taxa. Fluid from holding ponds contained ∼4 × 106 etal.2014),moreempiricalresearchisneededtoaccuratelydose cellsmL−1 withalargepercentageofcultivatableAPB(1×106 anddesigntargetedbiocidalcompoundsfordominanttaxain cellsmL−1)aswellasSRBinlessernumbers(1×103–1×104 theengineeredshalesystemaswellastounderstandtheirfate cellsmL−1),suggestingviable,deleteriousbacteriaabounddur- intheenvironment. 10 FEMSMicrobiologyEcology,2016,Vol.92,No.11 2013b (A) (B) D o w n lo a d e d fro m h ttp s ://a c Figure3.(A)Geographiclocationsofpublishedstudiesthatincludedmicrobial16SrRNAgenedatacollectedfromUSblackshaleformations,includingAntrim(MI), a d Barnett(TX),Burkett(PA),Marcellus(PA)andHaynesville(TX-LA).(B)Presenceorabsenceof16SrRNAgenesfrommicroorganismsmostcommonlydetectedacross e formations.Presenceisconsideredtobearelativeabundanceof>0.05%.Thecoloredcirclescorrespondtowelllocationsin(A).Differencesinthetypeofmicrobial mic analysesaredenotedinitalicsundertheauthor’sname. .o u p Shale-derivedfluidshavedistinctmicrobialmembers seriessamplingincludingearlyandlateproductionwatersam- .co m Though shales are not created equal (differing in well depth, plesfromMarcellusshalewells(Mohanetal.2013b;Cluffetal. /fe 2014),andtwootherstudieswith‘snapshot’lateproductionwa- m shaleformation,periodofproduction,geology,biogeochemistry, s tersamplesfromBurkett(Akobetal.2015)andBarnett(Davis e etc.),strongphysical(e.g.temperature)andchemical(e.g.salin- etal.2012)shales. c/a ity)featuresaresharedacrossformationsthatmayresultinen- Forthetimeseriesstudies,sourcefluidsforhydraulicfractur- rtic vironmentalfiltering.Todate,mostpublicationshaveusedava- le ingshowmarkedlydistinctinitialmicrobialcommunitiesacross -a riety of 16S rRNA gene-based approaches (clone libraries, 454 the shale wells despite studies with the same operator (Cluff bs pyrosequencingandIlluminaMiSeqtargetingtheV4region)to etal.2014).Hydraulicfracturingflowbackwatersatearlytime tra characterizemicrobialcommunitystructureandmembershipin points(1–14dayspost-fracturing)exhibitdramaticshiftsinthe ct/9 producedfluidsfromblackshalewells(Table1).Timeseriesdata 2 microbialcommunity(Fig.4).Afterinjectionintothedeepsub- /1 tahnadtipnrcolduudceemdiflcuroidbsiaolvaenraltyimseesoarfeinlpimutitferadcdtuureetfloutihdse,dfliofwficbualctky sriucrhfmaceen,tfloufiddsifrfeetruernntinmgictroobthiaelstauxrfaacceomimpamreeddiatoteilnypsuhtoflwuiedns-, 1/fiw 1 in obtaining samples. An additional barrier is the capacity to with dynamic taxa changes over the next 2 weeks. Halotoler- 66 extract and amplify DNA from fluids because of low biomass antandhalophilicGammaproteobacteria(e.g.Marinobacter,Vib- /24 andinhibitors(Akobetal.2015;Liangetal.2016).Unfortunately, 0 rio,PseudomonasandAcinetobacter)arehighlyenrichedinthese 2 manyofthepublishedstudieshavenotdepositedsequencing 9 earlyfluids,whereasingleOTUcanaccountforupto48%ofthe 1 datainpublicdatabasesordonotprovideaccompanyingrela- relativeabundance.Inparticular,PseudomonasandAcinetobacter 8 b tiveabundanceormetadatatolinksequenceabundancesand generaarekeymicrobialindicatorsofearlytimepointsacross y g samplenames.Figure3summarizes16SrRNAgenedatafrom Marcelluswells(Fig.4).Otherkeytaxathatservetodistinguish ues studies that had publically accessible data. To date, only two earlyfromlatesamplesincludeArcobacterspp.intheEpsilon- t o studies,bothfromhydraulicallyfracturedMarcellusshale,in- n proteobacteria(upto24%inday9)andMarinilabiliaspp.(16%in 0 clude input and time series data; one focused exclusively on 9 day7)intheBacteroidetes(Fig.4). A bsaamctperliinag(Maothlaatnerettiaml.e20p1o3ibn)twshanildetihnecloutdheedrianrcclhuadeead(aCdludfiftieotnaall. welOlenncteerflsopwrboadcukctvioonluamnedsosvtearrttimtoewthaneesaalfinteitry2i–n4flwueideskss,tatbhie- pril 2 2014)(Fig.3). lizearound90000mgL−1(intheMarcellusshale)(Mohanetal. 019 Toexaminemicrobialmemberssharedacrossshalesofvary- 2013b;Cluffetal.2014)asthefluidsreachequilibriumwithfor- ingagesanddepths,wecompiledthepublicallyavailabledata mationbrines.Surprisingly,inspiteoftheseinitialdifferences sets from the studies in Table 1 to generate a shale-wide op- ininputandearlyflowbacksamples,themicrobialcommuni- erational taxonomic unit (OTU, 97% identity in 16S rRNA se- tiesconvergetoacharacteristicandstatisticallydifferentstruc- quence)table.Methodsaredescribedinourpipelinerepository tureinproductionwateratlatertimepoints(49–328dayspost- (github.com/lmsolden/Qiime-pipeline)andalldatafilesarepro- fracturing). Consistent with a changing microbial community vided(SupportingInformation).Givenalackofavailablemeta- structure,Shannon’sdiversityindicesareindistinguishablebe- dinatwah,wicehctlhaesysiwfieedrethdeet1e6cSterdRN(eAarsleyq<u4e9ndceasysb,alsaeted>up49ondathyseatifmteer tweeninputandearlyflowbacksamples(Hinput =3.5;Hearly = 3.2),butdiversitydecreasessignificantlyintimepointscollected fracturing).Weconductedcomparative,non-parametricmulti- after49days(Hlate=0.4).Aligningwiththisdecreaseinmicrobial dimensionalscaling(NMDS)analysesincludingsequencingdata diversityisthechangeinabundancein16SrRNAgenesbelong- from four studies; the two aforementioned studies with time ingtoasinglegenus.HalanaerobiumintheFirmicutesbecome