Ann.N.Y.Acad.Sci.ISSN0077-8923 - PDF document

Ann.N.Y.Acad.Sci.ISSN0077-8923 ANNALSOFTHENEWYORKACADEMYOFSCIENCES Issue: BlavatnikAwardsforYoungScientists Platetectonicsandplanetaryhabitability:currentstatus andfuturechallenges JunKorenaga DepartmentofGeologyandGeophysics,YaleUniversity,NewHaven,Connecticut Addressforcorrespondence:JunKorenaga,DepartmentofGeologyandGeophysics,YaleUniversity,P.O.Box208109, NewHaven,CT06520-8109.jun.korenaga@yale.edu Platetectonicsisoneofthemajorfactorsaffectingthepotentialhabitabilityofaterrestrialplanet.Thephysicsofplate tectonicsis,however,stillfarfrombeingcomplete,leadingtoconsiderableuncertaintywhendiscussingplanetary habitability.Here,IsummarizerecentdevelopmentsontheevolutionofplatetectonicsonEarth,whichsuggesta radicallynewviewonEarthdynamics:convectioninthemantlehasbeenspeedingupdespiteitssecularcooling, andtheoperationofplatetectonicshasbeenfacilitatedthroughoutEarth’shistorybythegradualsubductionof waterintoaninitiallydrymantle.Theroleofplatetectonicsinplanetaryhabitabilitythroughitsinuenceon atmosphericevolutionisstilldifculttoquantify,and,tothisend,itwillbevitaltobetterunderstandacoupled core–mantle–atmospheresysteminthecontextofsolarsystemevolution. Keywords: terrestrialplanets;mantledynamics;planetarymagnetism;atmosphericevolution Introduction UnderwhatconditionscanaplanetlikeEarth—that is,aplanetthatcanhostlife—beformed?Thisques- tionofplanetaryhabitabilityhasbeenaddressed countlesstimesinthepast, 1–3 asitisdeeplycon- nectedtotheoriginoflife,perhapsthemostfasci- natingprobleminscience.Inthelastdecadeorso, researchactivitiesinthiseldhavebeeninvigorated, fueledbyarapidlyexpandingcatalogofextraso- larplanets. 4–6 Thehabitabilityofaplanetdepends onanumberoffactorsincluding,forexample,the massofthecentralstarandthedistancefromit, theatmosphericcomposition,orbitalstability,the operationofplatetectonics,andtheacquisitionof waterduringplanetaryformation.Themassofthe stardeterminestheevolutionofitsluminosity,and theheliocentricdistanceofaplanetaswellasthe volumeoftheatmosphereanditscompositionthen controlthesurfacetemperatureoftheplanet.The surfacetemperaturehastobeinacertainrange sothatwecanexpectthepresenceofliquidwa- terprovidedthatwaterexists,andorbitaldynamics affectthestabilityoftheplanetaryclimate.Plate tectonicscontrolstheevolutionoftheatmosphere throughvolcanicdegassingandsubduction,andit isalsoessentialfortheexistenceofaplanetarymag- neticeld,whichprotectstheatmospherefromthe interactionwiththesolarwind.Thesefactorsaf- fectingplanetaryhabitabilityarethusinterrelated tovariousdegrees.Whetherornotplatetectonics isoperatingonaplanet,forexample,wouldgive risetovastlydifferentscenariosforitsatmospheric evolution,affectingthedenitionofthehabitable heliocentricdistance,thatis,thehabitablezone. Thefocusofthiscontributionisonplatetecton- ics.Platetectonicsreferstoaparticularmodeof convectioninaplanetarymantle,whichismadeof silicaterocks,andsofaritisobservedonlyonEarth. Earth’ssurfaceisdividedintoadozenplatesorso, andtheseplatesaremovingatdifferentvelocities. Mostgeologicalactivities,suchasearthquakes,vol- caniceruption,andmountainbuilding,occurwhen differentplatesinteractatplateboundaries.There- alizationthatEarth’ssurfaceisactivelydeforming viaplatetectonicswasachievedthroughthe1960s and1970s,revolutionizingalmostallbranchesof earthsciences.Platetectonicsisafundamentalpro- cess,yetwestilldonotunderstanditinasatisfac- torymanner.Forexample,whereasthepresent-day doi:10.1111/j.1749-6632.2011.06276.x Ann.N.Y.Acad.Sci.1260(2012)87–94 c  2012NewYorkAcademyofSciences. 87 Platetectonicsandplanetaryhabitability Korenaga platemotionisknowninconsiderabledetail, 7 re- constructingpastplatemotionbecomesquitedif- cultonceweenterthePrecambrian(before540 millionyearsago),andeventhegrosscharacteristics ofancientplatetectonicsisuncertain. 8 Naturally, whenplatetectonicsstartedtooperateonEarthis stillcontroversial. 9 Partofthedifcultyoriginates inthepaucityofobservations;wehavefewerge- ologicalsamplesfromgreaterages.Thesituation isevenmorecompoundedbythelackoftheoret- icalunderstanding.Geophysicistshaveyettoform aconsensusonwhyplatetectonicstakesplaceon Earthandnotonotherterrestrial(e.g.,Earth-like) planetssuchasVenusandMars. 10 Thephysicsof platetectonicsisstillincomplete,andthiscreatesa seriousimpedimenttothediscussionofplanetary habitability.Underwhatconditionscouldplatetec- tonicsemergeonaplanet,andhowwoulditevolve throughtime?Withoutbeingabletoanswerthese questions,itwouldbenearlyimpossibletopredict theatmosphericevolutionofagivenplanetandthus itshabitability. Fundamentalissuesregardingthephysicsofplate tectonicsmaybeparaphrasedbythefollowingques- tions:howdidplatetectonicsevolveinthepast?,why doesplatetectonicstakeplaceonEarth?,andwhen didplatetectonicsrstappearonEarth?Consider- ableprogresshasbeenmadeontherstquestionin thelastdecade,andthisprogressturnsouttohelp betteraddressthesecondandthirdquestionsaswell. Inthefollowingsections,Iwillrevieweachquestion onebyoneandconcludewithasynthesisofcurrent statusaswellasmajortheoreticalchallengestobe tackledinthecomingyears. Howdidplatetectonicsevolve? Asplatetectonicsisjustonetypeofthermalcon- vection,itisreasonabletospeculateontheevo- lutionofplatetectonicsonthebasisofuidme- chanics.Earth’smantleinthepastwasgenerally hotterandthusprobablyhadlowerviscositythan present.Elementaryuidmechanicstellsusthat thisreductioninviscosityshouldhaveresultedin morevigorousconvection,thatis,higherheatux andfasterplatetectonics. 11 Geologicalsupportfor suchfasterplatetectonicshaslongbeenlacking, 12 butthislackofobservationalsupportisusuallynot takenseriouslybecausegeologicaldatabecomevery scarceandmoredifculttointerpretinthePrecam- brian.Thenotionoffasterplatetectonicsinthepast, however,hasbeenknowntopredictanunrealistic thermalhistorycalled“thermalcatastrophe,” 13 un- lessoneassumesthatEarthcontainsconsiderably moreheat-producingelementsthanthecomposi- tionmodelsofEarthindicate(Fig.1).Thiscanbe understoodbyconsideringthefollowingglobalheat balance: C dT dt = H ( t )– Q ( t ) , (1) where C istheheatcapacityoftheentireEarth, T is averageinternaltemperature, t istime, H isinter- nalheatproductionowingtothedecayofradioac- tiveisotopes,and Q isheatlossfromthesurface bymantleconvection.Bythenatureofradioactive decay,theinternalheatproductionmonotonically Figure1. Thermalhistorypredictionforfourcombinations ofheatowscalingandinternalheatproduction(seeRef.17for modelingdetail).Thenewscalingofplatetectonicspredictsrel- ativelyconstantheatuxindependentofmantletemperature, whereasclassicalscalingpredictshigherheatuxforhotterman- tle.TheUreyratioisameasureoftheamountofheat-producing elementsinthemantle,andthechemicalcompositionmodels ofEarthsuggestthatitspresent-dayvalue( Ur 0 = H(0)/Q(0) )is relativelylow,  0.3. 14 Constantheatuxwithalowpresent-day Ureyratio( solid )istheonlyonethatcanreproducetheobserved concave-downwardthermalhistorywithanaveragecoolingrate of  100KGa –1 ( circles ). 27 Inthisprediction,Earthwaswarming upduringtherstonebillionyears;suchasituationispossible withtheefcientcoolingofthemagmaocean. 50 Classicalscal- ingwithalowUreyratioresultsinthermalcatastrophe( gray line ).ClassicalscalingwithahighUreyratio( graydashedline ) canreproduceareasonablecoolingrate,butathermalhistory isconcaveupward.ConstantheatuxwithahighUreyratio ( dashedline )resultsintoocoldathermalhistory. 88 Ann.N.Y.Acad.Sci.1260(2012)87–94 c  2012NewYorkAcademyofSciences. Korenaga Platetectonicsandplanetaryhabitability decreaseswithtime,withaneffectivehalf-lifeof aboutthreebillionyears.Theconvectiveheatloss canbeparameterizedasafunctionofaveragetem- peratureas Q  T m , (2) wheretheexponent m ispredictedtobe  10bythe classicaltheoryofthermalconvection; 11 heatloss isextremelysensitivetoachangeininternaltem- perature.Thepresent-dayinternalheatproduction H(0) isonlyabout30%oftheconvectiveheatloss Q(0), 14 soabout70%ofheatlossmustbebalanced bytherapidcoolingofEarth;thatis,Earthmust havebeenmuchhotterthanatpresenttoexplain thepresent-daythermalbudget.Becauseconvective heatlossrisessharplywithincreasingtemperature (Eq.2),however,heatlossmusthavebeenextremely highinthepast,resultinginamoresevereimbal- ancebetweenheatproductionandheatloss,aswe considerfurtherbackintime.Thispositivefeedback iswhatcausesthermalcatastropheinthemiddleof theEarthhistory.Theonlywaytopreventit,while keepingtheclassicalscaling( m  10),istoassume thatinternalheatproductionisclosetoconvective heatlossatpresent,thatis, H(0)  Q(0) ,butthisvi- olatesourunderstandingofthechemicalbudgetof Earth.Thisconictbetweenthegeophysicaltheory ofmantleconvectionandthegeochemicalmodel ofEarthhasinspiredavarietyofproposals(see Ref.14forreview),manyofwhichhideanexcessive amountofheat-producingelementsinthedeep, inaccessiblemantle—apossiblebutrather adhoc solution. Anovelsolutionwassuggestedin2003based ontheeffectofmantlemeltingonmantleconvec- tion. 15 Fasterplatetectonicsinthepastisbasedon simpleuidmechanicsthatdonotcapturerealis- ticcomplicationsassociatedwithsilicaterocks.An importantdifferencefromclassicalthermalconvec- tionischemicaldifferentiation;whenthemantleis risingtowardthesurface,itusuallymelts,andthis meltingcanaffectmantledynamics.Uponmelt- ing,impuritiesinthemantle,mostnotablywater, arelargelypartitionedintothemeltphase,leav- ingtheresidualmantleverystiff. 16 Ahottermantle inthepastmeansmoreextensivemelting,making thickerstiffplatesandslowingdownplatetecton- ics.Consideringboththephysicsandchemistryof Earth’smantlethuspointstoanentirelyopposite prediction:slowerplatetectonicsinthepast,which isequivalenttousing m  0inEq.(2).Withthis nonclassicalscalingofplatetectonics,ithasbecome possibletoreconstructareasonablethermalhis- torywithoutviolatingthegeochemicalconstraints (Fig.1).Thissolution,whichwasfurtherelaborated in2006, 17 metconsiderableskepticismbecauseofits counterintuitivenature.Somedoubtedtherobust- nessofthegeochemicalconstraintsontheamount ofheat-producingelements,buttheuncertaintyof themantlecompositionhasbeenshowntobetight enoughtodiscountsuchleeway. 18 Otherssuspected thattherelativecontributionofheat-producingele- mentsmaybeincreasedbyloweringtheestimateon present-dayheatuxinstead, 19 , 20 butthispossibility hasbeenshowntobeinconsistentwithavailablege- ologicalrecords. 21 , 22 Additionally,thecounterintu- itivepredictionwasbasedonanapproximatetheory (knownastheboundarylayertheory)withseveral simplifyingassumptions,andsomequestionedthe validityofthisapproach. 23 Recently,however,the originalpredictionhasbeengivenfulltheoretical supportfromextensivenumericalsimulationand scalinganalysis. 24 , 25 Equallyimportantistheappearanceofnewde- cisiveobservations.In2008,thecompilationof thegeologicalrecordsofancientpassivemargins waspublished,whichindicatesthatthetempoof platetectonicsinthepastwasindeedslowerthan present. 26 In2010,thethermalhistoryofEarth’s uppermantlewasreconstructedbyapplyingthe latestpetrologicaltechniquetoanextensivecom- pilationofPrecambrianvolcanicrocks(Fig.1). 27 Theconcave-downwardnatureofthisthermalhis- toryisparticularlyimportant,asitprovidesstrong supportforthenotionofslowerplatetectonicsand therelativelylowabundanceofheat-producingel- ementsatthesametime;itisimpossibletorepro- ducethiscurvaturebyassumingfasterplatetec- tonicsforahottermantle.Mostrecently,anew constraintontheabundanceofheat-producingele- mentsinthemantlewasreportedbasedongeoneu- trinoobservations,whichfavorstherelatively lowabundanceasindicatedbythegeochemical estimate. 28 Theradicallynewviewontheevolutionofplate tectonics,therefore,hasbeencorroboratedboth theoreticallyandobservationallyinrecentyears,and ithasbecomedifculttorefutethenotionofslower platetectonicsinthepast,howevercounterintu- itiveitmightbe.Actually,whatiscounterintuitive Ann.N.Y.Acad.Sci.1260(2012)87–94 c  2012NewYorkAcademyofSciences. 89 Platetectonicsandplanetaryhabitability Korenaga isasubjectivematter,andinthiscase,itislargely educational.FasterplatetectonicsforahotterEarth ispredictedbytheuidmechanicsofanearlyiso- viscousuid.Notheoreticaljusticationexistsfor itsapplicabilitytoEarth’smantle.Theclassicalthe- oryisstillwidelyusedinplanetarysciences,but itsimplyfailstoreproducethethermalhistoryof thebest-understoodplanet(Fig.1).Therewouldbe littlemeritinextrapolatingatheorythatcannotex- plainEarthtootherterrestrialplanets,forwhichwe haveconsiderablyfewerobservationalconstraints. Thisisespeciallytruewhendiscussingthedynamics ofEarth-like,potentiallyhabitableplanets. Whydoesplatetectonicshappen? Therearetwofundamentallydifferentmodesof mantleconvection:(1)platetectonicsand(2)stag- nantlid. 29 Instagnant-lidconvection,theentiresur- faceofaplanetformsarigidsphericalshell,and convectioncantakeplaceonlyundertheshell.In platetectonics,thesurfaceisbrokenintopieces, mostofwhichcanreturntothedeepmantle,en- ablinggeochemicalcyclesbetweenthesurfaceand theinterior.Amongthefourterrestrialplanetsin oursolarsystem,Earthistheonlyplanetthatex- hibitsplatetectonics,andtheotherthree(Mercury, Venus,andMars)arebelievedtobeinthemode ofstagnantlid. 30 Itiseasytoexplainwhyplatetec- tonicsdoesnottakeplaceonotherplanets,because stagnant-lidconvectionisthemostnaturalmodeof convectioninamediumwithstronglytemperature- dependentviscosity,suchassilicaterocksthatcon- stituteaplanetarymantle. 29 Mantleviscosityis extremelyhighatatypicalsurfacecondition,sovir- tuallynodeformationisexpectedthere.Themode ofplatetectonicsispossible,therefore,onlywhen someadditionalmechanismexiststocompensate theeffectoftemperature-dependentviscosity.On- goingdebatesaremostlyregardingthisadditional weakeningmechanism. Inadditiontoductiledeformationcharacterized byviscosity,silicaterockscanalsodeformbybrittle deformationsuchascrackingandfaulting.Weak- eningbybrittlemechanismsislimitedbyfrictional strength, 31 however,andwithatypicalfrictional coefcientoforder1,brittleweakeningisinsuf- cienttocauseplatetectonics. 32 Inordertosimulate platetectonicsinnumericalmodels,therefore,ithas beenacommonpracticetoassumeamuchlower frictioncoefcient,butaphysicalmechanismthat couldleadtosuchalowcoefcienthasbeenpoorly understood. 33 Oneplausiblemechanismisareduc- tioninaneffectivefrictioncoefcientbyhighpore uidpressure,withwaterbeingtheuidmedium. Forplatetectonicstooccurwiththismechanism (i.e.,tobreakathickstagnantlid),however,wa- terhastobetransportedtosubstantialdepthsand thenisolatedfromthesurfacetoachievehighpore uidpressure.Themereexistenceofsurfacewater doesnotguaranteeeitheroftheserequirements.If deepwaterisconnectedtothesurface,forexam- ple,itwouldbeathydrostaticpressure,meaning thatporeuidpressureistoolowtoachieveasuf- cientlylowfrictioncoefcient.Inthisregard,the thermalcrackinghypothesis, 33 inwhicharigidlid isextensivelyfracturedbystrongthermalstressand thenlatersealedbyhydrationreactions,hassofar beentheonlytangiblemechanismthatcouldgen- erateplatetectonicsinthepresenceofsurfacewater (Fig.2). Whendiscussingthemodeofmantleconvec- tion,itisimportanttoavoidbeingtrappedina chicken-and-eggsituation.Thebendingofasub- ductingplate,forexample,mayfractureandweaken theplatebyhydration, 34 butonecannotinvokethis mechanismfortheonsetofplatetectonics;aweak- eningmechanismmustbeoperationalevenwith- outplatetectonics.Thesamecautionappliesto variousdynamicweakeningmechanismsassociated withearthquakedynamics. 35 Findingaphysicalmechanismforweakeningis justonesideofthecoin.Theothersideistounder- standthecriticalstrengthofasurfacelidthatcanbe overcomebyconvectivestressexertedbytheman- tlebelow.Bothsidesarenecessarytounderstand underwhatconditionsplatetectonicscanhappen. Thisissuehasbeenstudiedbyvariousauthorsus- ingnumericalsimulation(e.g.,Refs.36and37),but inmostpreviousattempts,thetemperaturedepen- dencyofmantleviscositywasnotstrongenoughto beEarth-like.Aquantitativecriterionfortheonset ofplatetectonicswasfoundin2010,fortherst timewithrealisticmantleviscosity,whileconduct- inganumberofnumericalsimulationstoestablish thescalingofplatetectonicsforthermalevolution modeling. 24 Revisitingthenotionofslowerancient platetectonicswiththisnewcriterionturnsout toyieldanintriguinginsightfortheinitiationof platetectonicsonEarth,asdiscussedinthenext section. 90 Ann.N.Y.Acad.Sci.1260(2012)87–94 c  2012NewYorkAcademyofSciences. Korenaga Platetectonicsandplanetaryhabitability primary thermal cracks secondary cracking caused by high residual stress and shallow serpentinization viscous lower half Figure2. Schematicillustrationforrheologicalevolutionwithinaplateunderoceans. 33 Optimalreleaseofthermalstressdeveloped inacoolingplateisachievedbyacascadecracksystem(primarycracks).Anyresidualstresswilleventuallybereleasedbysecondary crackpropagationifpartialcrackhealingbyshallowserpentinizationraisesthepressureoftrappedseawatertolithostaticpressure. Thestiffestpartofplateduetostrongtemperature-dependentviscositycanthusbepervasivelyweakenedbythermalcrackingand subsequenthydration. Whendidplatetectonicsstart? Basedoneldevidence,manygeologistswouldcon- curwiththeoperationofplatetectonicsbackto about3billionyearsago, 38 , 39 butanythingbeyond thatiscontroversial.Earth’shistoryisdividedinto foureons:theHadean(4.6–4.0billionyearsago),the Archean(4.0–2.5billionyearsago),theProterozoic (2.5–0.54billionyearsago),andthePhanerozoic (0.54billionyearsagotopresent).TheHadean– Archeanboundaryisdenedbytheageoftheold- estrockfoundonEarth.TheArchean–Proterozoic boundaryisdenedbytherelativeabundanceof rocks—thatis,rocksofArcheanagesaremuchrarer thanthoseofyoungerages.Thesedenitionsofthe geologicaltimescaleindicatethatndingunam- biguousgeologicaldatafortherstappearanceof platetectonicsonEarth’shistory,whichmightbe intheHadeanera, 40 wouldbequitechallenging. Buildingatheoreticalfoundationforthisproblem isthusofcriticalimportance. Asmentionedearlier,slowerplatetectonicsinthe pastresultsfromtheformationofthickerstiffplates bymoreextensivemelting.Thickerplatesslowdown platetectonics,butiftoothick,theycouldpoten- tiallyjeopardizetheoperationofplatetectonicsit- self.Thelikelihoodofshuttingdownplatetectonics inthepastcanbehighbecausemuchlowerconvec- tivestressisexpectedforahotter,lessviscousmantle beneathplates;thickerplatesandweakerconvective stressbothacttoimpedeplatetectonics.Indeed,a quantitativeassessmentofthispossibilityusingthe newcriterionindicatesthatplatetectonicsisviable onlyforthelastonebillionyears, 41 whichgrossly contradictswithgeologicalevidence. Onepossibleresolutiontothisconundrumcame fromanapparentlyunrelatedthreadofthought, thoughinhindsightitisanaturalextensionofthe existingtheory.Byincorporatinggeologicalcon- straintsonthepastsealevelintothermalevolution modeling,onecanreconstructthehistoryofocean volume,andslowerplatetectonicscorrespondsto greateroceanvolumeinthepast. 42 TheArchean oceansareestimatedtohavebeenabouttwiceasvo- luminousasthepresentoceans,andthisdifference inoceanvolumeisroughlyequivalenttotheamount ofwaterstoredinthepresentmantle.Earth’sman- tlecouldhavebeendrierandthusmoreviscousin thepast,andifthisexchangeofwaterbetweenthe mantleandtheoceansistakenintoaccount,theop- erationofplatetectonicsbecomesviablethroughout Earth’shistory. 24 , 41 Ithaslongbeensuggestedthat Ann.N.Y.Acad.Sci.1260(2012)87–94 c  2012NewYorkAcademyofSciences. 91 Platetectonicsandplanetaryhabitability Korenaga platetectonicscouldresultinnetwaterinuxtothe mantle, 43 , 44 andmodelingthemantleasanopen systemnowappearstobeanecessity,ratherthanan option. Platetectonics,therefore,couldhavestartedon Earthshortlyafterthesolidicationofaglobal magmaocean,whichprobablyexistedonlyforthe rstfewtensofmillionsofyearsofEarth’shistory. 45 OneinterestingndingfromEarth’sevolutionwith ahydratingmantleisthatthesubductionofwater isessentialtomaintainadrylandmass,whichin turnplaysanindispensableroleinstabilizingthe climate.Withoutadrylandmass,nosilicateweath- eringcouldtakeplacesothattheatmosphericcom- positioncouldnotberegulatedefcientlybycarbon cycle. 1 Summaryandoutlook Anemergingviewontheevolutionofplatetecton- icsonEarthcanbesummarizedbythefollowing. Platetectonicsstartedprobablyintheveryearly Earth,shortlyafterthesolidicationoftheputative magmaocean.Theonsetofplatetectonicswasfacili- tatedbyaninitiallydrymantle,whichhassincebeen slowlyhydratedbyplatetectonics.WhileEarthhas beencoolingdown,platetectonicshasbeenspeed- ingup,insteadofslowingdown.Thisisbecause acoldermantleleadstothinner,easilydeformable platesandbecausetheeffectofhydrationonviscos- itytendstocanceltheeffectoftemperature.This scenariohasbeenshowntobeinternallyconsis- tentanddynamicallyplausiblebythescalinglaws ofplatetectonics.Thoughbeingunconventionalin nearlyallaspects,itistheonlyhypothesisthatis consistentwithallofmajorobservationsrelevantto Earth’sevolution,includingpetrologicalconstraints onthethermalhistory,geochemicalconstraintson thethermalbudget,andgeologicalconstraintson thetempoofplatetectonics,themodeofmantle convection,andtheglobalsealevelchange. Wateristhusexpectedtoplayfundamentalroles intheinitiationofplatetectonicsanditsevolution overEarth’shistory.Thephysicsofelementarypro- cessesinvolvingwaterintheabovescenario,how- ever,stillrequiresconsiderablefuturedevelopment. Theplausibilityofthethermalcrackinghypothesis, forexample,needstobetestedfurtherbymod- elingthephysicochemicalevolutionofamultiple cracksystem.Thehypothesishasindirectobserva- tionalsupportthroughitsimpactoneffectivether- malexpansivity, 46 , 47 butmoredirectevidencemay beobtainedbylarge-scaleeldexperiments.Ad- ditionally,therateofwatertransporttothedeep mantlebysubductionshouldbequantiedfrom rstprinciples.Thoughitisahighlycomplexprob- leminvolvingpetrology,mineralphysics,anduid mechanics,itssolutionisessentialforatheorywith predictingpower. Ifaterrestrialplanetstartsoutwithadryman- tleandsurfacewater,whichappearstobealikely initialconditionforsubsolidusmantleconvection, 3 theonsetofplatetectonicsisprobablyjustiable. Predictingitssubsequentevolutionis,however,still aformidabletask.Oneofthemajoruncertainties istheamountofsurfacewater.Onthebasisofthe planetaryformationtheory,theoriginofEarth’swa- terisoftenconsideredtobeintheoutersolarsys- tem, 48 andthedeliveryofwaterisahighlystochastic process.Thequantityofwatertobedelivereddoes nothavetobelarge;oneoceanworthofwatercor- respondstoonly0.02%ofEarth’smass.Adifcult partishowtomaintainitoverthegeologicaltime. Theexistenceofaplanetarymagneticeld,which couldprovideashieldagainstsolarwinderosion, 49 dependsontherateofcorecooling.Earth’sther- malhistoryindicatesthatthemantlewaswarm- ingupintheearlyEarth(Fig.1).Thecorecould stillhavebeencoolingduringthattimeifthecore wasinitiallysuperheatedbyitsformationprocess, buttoanswerwhetherthecoolingratewassuf- cienttodriveaplanetarydynamo,modelingthe thermalevolutionofacoupledcore–mantlesystem wouldbecritical.Understandingtheatmospheric evolutionofagiventerrestrialplanet,oritshab- itabilityatlarge,therefore,requiresustocreatea uniedtheoreticalframeworkthatspansfromthe solarsystemevolutiontothedynamicsofplanetary interior. Acknowledgments ThisworkwassponsoredbyaMicrosoftA.Richard NewtonBreakthroughResearchAward.Theauthor thanksNormSleepforaconstructivereview. Conßictsofinterest Theauthordeclaresnoconictsofinterest. References 1.Kasting,J.F.&D.Catling.2003.Evolutionofahabitable planet. Annu.Rev.Astron.Astrophys. 41: 429–463. 92 Ann.N.Y.Acad.Sci.1260(2012)87–94 c  2012NewYorkAcademyofSciences. Korenaga Platetectonicsandplanetaryhabitability 2.Gaidos,E.,B.Deschenes,L.Dundon, etal .2005.Beyond theprincipleofplentitude:areviewofterrestrialplanet habitability. Astron.J. 5: 100–126. 3.Zahnle,K.,N.Arndt,C.Cockell, etal .2007.Emergenceofa habitableplanet. SpaceSci.Rev. 129: 35–78. 4.Marcy,G.W.&R.P.Butler.1998.Detectionofextrasolar giantplanets. Annu.Rev.Astron.Astrophys. 36: 57–97. 5.Rivera,E.J.,J.J.Lissauer,R.P.Butler, etal .2005.A  7.5M- circleplusplanetorbitingthenearbystar,GJ876. ApJ 634: 625–640. 6.Borucki,W.J. etal .2011.Characteristicsofplanetarycandi- datesobservedbyKepler,II:analysisoftherstfourmonths ofdata. ApJ . 736: 19,doi:10.1088/0004-637X/736/1/19. 7.DeMets,C.,R.G.Gordon&D.F.Argus.2010.Geologically currentplatemotions. Geophys.J.Int. 181: 1–80. 8.Bleeker,W.2003.ThelateArcheanrecord:apuzzleinca.35 pieces. Lithos 71: 99–134. 9.Condie,K.C.&V.Pease,Eds.2008. WhenDidPlateTectonics BeginonPlanetEarth? GeologicalSocietyofAmerica. 10.Bercovici,D.2003.Thegenerationofplatetectonicsfrom mantleconvection. EarthPlanetSci.Lett. 205: 107–121. 11.Schubert,G.,D.Stevenson&P.Cassen.1980.Wholeplanet coolingandtheradiogenicheatsourcecontentsoftheearth andmoon. J.Geophys.Res. 85: 2531–2538. 12.Kr ¨ oner,A.&P.W.Layer.1992.Crustformationandplate motionintheearlyArchean. Science 256: 1405–1411. 13.Christensen,U.R.1985.Thermalevolutionmodelsforthe Earth. J.Geophys.Res. 90: 2995–3007. 14.Korenaga,J.2008.Ureyratioandthestructureand evolutionofEarth’smantle. Rev.Geophys. 46:RG2007, doi:10.1029/2007RG000241. 15.Korenaga,J.2003.Energeticsofmantleconvectionand thefateoffossilheat. Geophys.Res.Lett. 30: 1437, doi:10.1029/2003GL016982. 16.Hirth,G.&D.L.Kohlstedt.1996.Waterintheoceanicman- tle:implicationsforrheology,meltextraction,andtheevo- lutionofthelithosphere. EarthandPlanetaryScienceLetters 144: 93–108. 17.Korenaga,J.2006.Archeangeodynamicsandthethermal evolutionofEarth.In ArcheanGeodynamicsandEnviron- ments .K.Benn,J.-C.Mareschal&K.Condie,Eds.:7–32. AmericanGeophysicalUnion,Washington,D.C. 18.Lyubetskaya,T.&J.Korenaga.2007.Chemicalcom- positionofEarth’sprimitivemantleanditsvariance, 1,methodsandresults. J.Geophys.Res. , 112: B03211, doi:10.1029/2005JB004223. 19.Grigne,C.,S.Labrosse&P.J.Tackley.2005.Convective heattransferasafunctionofwavelength:implicationsfor thecoolingoftheEarth. J.Geophys.Res. 110: B03409, doi:10.1029/2004JB003376. 20.Stevenson,D.J.2007.Aplanetaryscientistforeseesashiftin thedebateaboutEarth’sheatow. Nature 448: 843. 21.J.Korenaga.2007.Eustasy,supercontinentalinsulation,and thetemporalvariabilityofterrestrialheatux. EarthPlanet Sci.Lett. 257: 350–358. 22.Loyd,S.J.,T.W.Becker,C.P.Conrad, etal .2007.Timevari- abilityinCenozoicreconstructionsofmantleheatow:plate tectoniccyclesandimplicationsforEarth’sthermalevolu- tion. Proc.Nat.Acad.Sci.USA 104: 14266–14271. 23.Davies,G.F.2009.EffectofplatebeningontheUreyratio andthethermalevolutionofthemantle. EarthPlanetSci. Lett. 287: 513–518. 24.Korenaga.J.2010.Scalingofplate-tectonicconvection withpseudoplasticrheology. J.Geophys.Res. , 115: B11405, doi:10.1029/2010JB007670,2010. 25.RoseI.R.&J.Korenaga.2011.Mantlerheologyandthe scalingofbendingdissipationinplatetectonics. J.Geophys. Res. 116: B06404,doi:10.1029/2010JB008004. 26.Bradley,D.C.2008.Passivemarginsthroughearthhistory. EarthSci.Rev. 91: 1–26. 27.Herzberg,C.,K.Condie&J.Korenaga.2010.Thermalevo- lutionoftheEarthanditspetrologicalexpression. Earth PlanetSci.Lett. 292: 79–88. 28.Gando,A. etal .2011.PartialradiogenicheatmodelforEarth revealedbygeoneutrinomeasurements. NatureGeosci. 4: 647–651,doi:10.1038/ngeo1205. 29.Solomatov,V.S.1995.Scalingoftemperature-andstress- dependentviscosityconvection. Phys.Fluids 7: 266– 274. 30.Schubert,G.,D.L.Turcotte&P.Olson.2001. MantleCon- vectionintheEarthandPlanets .CambridgeUniversityPress, Cambridge,UnitedKingdom. 31.Scholz,C.H.2002. TheMechanicsofEarthquakesandFault- ing .CambridgeUniversityPress,Cambridge,UnitedKing- dom. 32.Moresi,L.&V.Solomatov.1998.Mantleconvectionwitha brittlelithosphere:thoughtsontheglobaltectonicstylesof theEarthandVenus. Geophys.J.Int. 133: 669–682. 33.Korenaga.J.2007.Thermalcrackingandthedeephy- drationofoceaniclithosphere:akeytothegenera- tionofplatetectonics? J.Geophys.Res. , 112: B05408, doi:10.1029/2006JB004502. 34.Ranero,C.R.,J.PhippsMorgan,K.McIntosh&C.Reichert. 2003.Bending-relatedfaultingandmantleserpentinization atthemiddleAmericatrench. Nature 425: 367–373. 35.Korenaga,J.2010.Onthelikelihoodofplatetectonicson super-Earths:doessizematter? ApJ 725: L43–L46. 36.Richards,M.A.,W.-S.Yang,J.R.Baumgardner&H.-P. Bunge.2001.Roleofalow-viscosityzoneinstabilizingplate tectonics:implicationsforcomparativeterrestrialplanetol- ogy. Geochem.Geophys.Geosys. 2: 2000GC000115. 37.Stein,C.,J.Schmalzl&U.Hansen.2004.Theeffectofrhe- ologicalparametersonplatebehaviorinaself-consistent modelofmantleconvection. Phys.EarthPlanetInter. 142: 225–255. 38.Hoffman,P.F.1997.TectonicgenealogyofNorthAmerica. In EarthStructure:AnIntroductiontoStructuralGeologyand Tectonics .B.A.vanderPluijm&S.Marshak,Eds.:459–464. McGraw-Hill,NewYork. 39.VanKranendonk,M.J.,R.H.Smithies,A.H.Hickman& D.C.Champion.2007.Review:seculartectonicevolution ofArcheancontinentalcrust:interplaybetweenhorizontal andverticalprocessesintheformationofthePilbaraCraton, Australia. TerraNova 19: 1–38. 40.Hopkins,M.D.,T.M.Harrison&C.E.Manning.2010. ConstraintsonHadeangeodynamicsfrommineralinclu- sionsin � 4Gazircons. EarthPlanetSci.Lett. 298: 367– 376. Ann.N.Y.Acad.Sci.1260(2012)87–94 c  2012NewYorkAcademyofSciences. 93 Platetectonicsandplanetaryhabitability Korenaga 41.Korenaga,J.Thermalevolutionwithahydratingmantleand theinitiationofplatetectonicsintheearlyEarth. J.Geophys. Res. doi:10.1029/2011JB008410,inpress. 42.Korenaga,J.2008.Platetectonics,oodbasalts,andthe evolutionofEarth’soceans. TerraNova 20: 419–439. 43.Ito,E.,D.M.Harris&A.T.Anderson.1983.Alterationof oceaniccrustandgeologiccyclingofchlorineandwater. Geochim.Cosmochim.Acta 47: 1613–1624. 44.Smyth,J.R.&S.D.Jacobsen.2006.Nominallyanhydrous mineralsandEarth’sdeepwatercycle.In Earth’sDeepWater Cycle .S.D.Jacobsen&S.vanderLee,Eds.:1–11.American GeophysicalUnion. 45.Solomatov,V.2007.Magmaoceansandprimordialmantle differentiation.In TreatiseonGeophysics .Vol. 9 ,G.Schubert, Ed.:91–119.Elsevier,Amsterdam. 46.Korenaga,J.2007.Effectivethermalexpansivityof Maxwellianoceaniclithosphere. EarthPlanetSci.Lett. 257: 343–349. 47.Korenaga,T.&J.Korenaga.2008.Subsidenceofnormal oceaniclithosphere,apparentthermalexpansivity,andsea- oorattening. EarthPlanetSci.Lett. 268: 41–51. 48.Morbidelli,A.,J.Chambers,J.I.Lunine, etal .2000.Source regionsandtimescalesforthedeliveryofwatertotheEarth. Meteorit.PlanetSci. 35: 1309–1320. 49.Lundin,R.,H.Lammer&I.Ribas.2007.Planetarymag- neticeldsandsolarforcing:implicationsforatmospheric evolution. SpaceSci.Rev. 129: 245–278. 50.Sleep,N.H.2000.Evolutionofthemodeofconvection withinterrestrialplanets. J.Geophys.Res. 105: 17563– 17578. 94 Ann.N.Y.Acad.Sci.1260(2012)87–94 c  2012NewYorkAcademyofSciences.