Tsunamigenic Splay Faults Imply a Long‐Term Asperity in Southern Prince William Sound, Alaska

Boise State University

ScholarWorks

CGISS Publications and Presentations

Center for Geophysical Investigation of the Shallow

Subsurface (CGISS)

4-16-2019

Tsunamigenic Splay Faults Imply a Long‐Term

Asperity in Southern Prince William Sound, Alaska

L. M. Liberty

Boise State University

D. S. Brothers

U.S. Geological Survey

P. J. Haeussler

U.S. Geological Survey

This document was originally published inGeophysical Research Lettersby Wiley on behalf of the American Geophysical Union. Copyright restrictions may apply. doi:10.1029/2018GL081528

L.M.Liberty1 ,D.S.Brothers2 ,andP.J.Haeussler3

1_{DepartmentofGeosciences,BoiseStateUniversity,Boise,ID,USA,}2_{PacificCoastalandMarineScienceCenter,U.S.}

GeologicalSurvey,SantaCruz,CA,USA,3_{U.S.GeologicalSurvey,Anchorage,AK,USA}

Abstract

Coseismic slip partitioning and uplift over multiple earthquake cycles is critical to understandingupper‐platefaultdevelopment.Bathymetricandseismicreflectiondatafromthe1964 Mw9.2GreatAlaskaearthquakerupturearearevealseafloorscarpsalongthetsunamigenicPatton Bay/CapeCleare/MiddletonIslandfaultsystem.Thefaultssplayfromamegathrustwhereduplexingand underplatingproducedrapidexhumation.Trenchwardoftheduplexregion,thefaultsproduceacomplex deformationpatternfromoblique,south‐directedshorteningattheYakutat‐Pacificplateboundary.Spatial andtemporalfaultpatternssuggestthatHolocenemegathrustearthquakeshadsimilarrelativemotionsand thussimilartsunamisourcesasin1964.Tsunamisduringfutureearthquakeswilllikelyproducesimilar run‐uppatternsandtraveltimes.Splayfaultsurfaceexpressionsthusrelatetoplateboundaryconditions, indicatingmillennial‐scalepersistenceofthisasperity.Wesuggeststructureofthesubductedslabdirectly influencessplayfaultandtsunamigenerationlandwardofthefrontalsubductionzoneprism.

Plain Language Summary

Weidentifyprominentseafloorscarpsthatshowasimilarpatternof tectonicupliftoverthepast20to30subductionzoneearthquakesinthewesternPrinceWilliamSoundarea ofAlaska.Ourresultssuggestthatplateboundaryconditionshavebeenfixedthroughmanyearthquake cyclesandthatsubductedplateboundaryconditionsinfluenceseafloorupliftpatterns.Weconcludethat tsunamipatternsobservedduringthe1964earthquakewilllikelyrepeattoreproducerun‐upandtraveltime observations.Mappingstructuresalongplateboundariesiscriticaltounderstandingtsunamisourcesin subductionzones.

KeyPoints:

• Megathrustsplayfaultsthatsurface

nearPrinceWilliamSound,Alaska,

showclearevidenceforrepeatedsea flooruplift

• Thespatialandtemporalsplayfault characterindicatesapersistenceofa plateboundaryasperityrelatedto

thesubductedYakutatterrane

• Tsunamisgeneratedfromcoseismic faultmotionduringfuture

megathrustearthquakeswillshow

repeatingpatternsasduringthe 1964event SupportingInformation: •SupportingInformationS1 Correspondenceto: L.M.Liberty, lliberty@boisestate.edu Citation:

Liberty,L.M.,Brothers,D.S.,& Haeussler,P.J.(2019).Tsunamigenic splayfaultsimplyalong‐termasperity

insouthernPrinceWilliamSound,

Alaska.GeophysicalResearchLetters,

46,3764–3772.https://doi.org/10.1029/ 2018GL081528

Received30NOV2018

Accepted19MAR2019

Acceptedarticleonline25MAR2019

Publishedonline15APR2019

©2019.AmericanGeophysicalUnion. AllRightsReserved.

1.

Introduction

Thrustfaultsthatsplayfromamegathrustwithinsubductionzoneaccretionarywedgescanposemajorseis­ micandtsunamihazards,yetlittleisknownaboutthespatialandtemporalcontrolsonthisfamilyoffaults. Surfaceruptures duringsubduction zoneearthquakes can highlightpatterns ofcoseismic motion(e.g., Fujiwaraetal.,2011;Henstocketal.,2006),paleoseismicandgeodeticobservationscanprovideestimates ofrecurrenceintervalsandpatternsofuplift/subsidence(e.g.,Atwater&Hemphill‐Haley,1997;Cisternas etal.,2005;Saillardetal.,2017;Shennanetal.,2014;Siehetal.,2008),andthermochronologymeasurements canprovideregionalupliftratesoverthousandsofearthquakecycles(e.g.,Enkelmannetal.,2015;Ferguson etal.,2015;Haeussleret al.,2015).However,detailedslippartitioningandupliftpatternsovermultiple earthquakecyclesremainsunknown.Constraintsontheseparametersarecriticaltounderstandingfault evolution,therelationshipoffaultstoknownplateboundaryasperitiesorlockedzones,thepaleoseismic record,andtsunamigenesis.

Usingatightgridof40sparkerseismicprofiles,coupledwithnewhigh‐resolutionseafloorimageryand legacygeophysicaldata,wecharacterizeacomplexfaultsystemthatdevelopedbyobliqueanddip‐slipshort­ eningaboveamegathrust.Thesefaultsliewithintheprimaryruptureareaofthe1964Mw9.2GreatAlaska earthquake,immediatelyoutboardofthesubductedYakutatterraneboundaryandoffshoretheMontague IslandareaofPrinceWilliamSound(PWS;Figure1).Giventhattherecurrenceintervalforlargeearth­ quakesisestimatedat500to600years(Carver&Plafker,2008;Shennanetal.,2014)andthat~50mm/year ofN30°Wplateconvergenceisdocumented(Elliottetal.,2010),weexamineHoloceneupliftpatternsor motionoverthe past20 to 30post‐glacialearthquake cycles.From long‐termupliftpatterns and from 1964earthquakeobservations,wesuggestthatmuchofthelast500to750mofplateshorteningwasaccom­ modatedalongaseriesofsubparallelsplayfaults.Assplayfaultsarearelatively common,albeitpoorly known feature of accretionary complexes, our results provide a rare glimpse into surface‐rupturing

Geophysical

Research

Letters

10.1029/2018GL081528

processes.Despitethecomplexityofthisfaultsystem,thedatashowapatternofpersistentseafloorruptures andgrowthfaultingbetweenregionswithpresumablylowershorteningrates.Weusebathymetricimagesto identifythetectonictsunamisourcesfromthe1964earthquakeandseismicdatatorefinethelateHolocene deformationhistoryofthesesplayfaults.

Figure1.BathymetricmapofsouthernAlaskashowingtheepicentralareaforthe1964Mw9.2earthquake(yellowstar),thehingelinebetween1964upliftfrom

subsidence(red),significant1964surfacerupturingfaults(lightred),magneticcontours(50‐nTinterval)thatdefinethetrailingedgeofthesubductedYakutat

terraneorSMA(yellow),selectTrans‐AlaskaCrustalTransect(TACT)andU.S.GeologicalSurveyseismicprofilelocations,1964tsunamirun‐uplocationsand

directions(purplearrows),Plafker(1969)tsunamitraveltimes(parentheses),andcalculatedsourceregion(lightpurplelines).Theinsetmapshowsthestudyarea,

trench,andthenonsubductedportionoftheYakutatterrane(Y)andKodiakIsland(K).TwocrosssectionsalongTACTseismicprofilesshowkeyupper‐plate

interpretations,withshadedareasrepresentingthecrustalduplexingzonebeneathadécollement.Selectfaults:PBF=PattonBayfault,CCF=CapeClearefault;

MIF=MiddletonIslandfault;MSF=MontagueStraitfault;WRF=WesselsReeffault;Otherabbreviations:PB=PugetBay;WB=WhidbeyBay;MI=Middleton

Island.TheFigure2boxrepresentsthezoneofmaximumsurface/seafloordisplacementsdocumentedduringthe1964earthquake(Libertyetal.,2013;

Plafker,1969).PWS=PrinceWilliamSound.

2.

The

1964

Great

Alaska

Earthquake

The1964Mw9.2earthquakerupturedan800kmby250kmarea,causingtsunamisalongtheGulfofAlaska coastline(Plafker,1969).Theearthquakeinitiatedat~25‐kmdepthbeneathnorthernPWS(Figure1),and highmomentreleaseareaswereidentifiednearthesouthwestextentofMontagueIsland(Figure2andsup­ portinginformationFigureS1)andimmediatelysouthofKodiakIsland(Christensen&Beck,1994;Johnson etal.,1996).ReleasefromthePWSasperityproduced21mofhorizontalsurfacedisplacementneartheedge ofthesubductedYakutat terrane(Plafker,1969; Figures1and2), wherenear‐flatslab subductionand underthrusting was interpreted to intersect the steeper‐dipping Pacific plate interface(Brocher et al., 1994;Kimetal.,2014).Theregionofinferredduplexing(Haeussleretal.,2015;Libertyetal.,2013)andmax­ imumslipduringthe1964earthquake(Plafker,1969)wascoincidentwiththeSlopeMagneticAnomaly (SMA)lineament thatmarks the southwestern edgeof thesubductingYakutat terrane(Brocheret al.,

1994;Bruns,1983;Kimetal.,2014;Figure1).TheruptureliftedwesternMontagueIslandandtheadjacent seafloorasmuchas12malong listricthrustfaults thatsplayfroma décollement(Libertyet al.,2013; Haeussleret al.,2015;Figure1).BecauselittletrenchwardmotionwasrecordedonMiddletonIslandin 1964 (Figure 1), Plafker (1969) concluded that horizontal shortening from this earthquake was accommodatedalmostentirelyalongfaultsthatlieonthecontinentalshelf.Fewlarge(M>5),post‐1964 earthquakes have been recorded inthe PWS area, which now appears to be completely locked (e.g., Freymuelleretal.,2008;Zwecketal.,2002).

Immediatelyafterthe1964 earthquake,tsunami run‐upwasdocumented atnumeroussitesonKodiak Island, Kenai Peninsula, PWS, and Middleton Island (Plafker, 1969; Figure 1). Although submarine landslide‐inducedtsunamiswereinitiatedalongthedeepfjordcoastlineswithinminutesofgroundshaking (e.g.,Brothersetal.,2016;Haeussleretal.,2014;Parsonsetal.,2014),thetectonictsunamitraveltimesran­ gedfromabout20mintomanyhours,indicatingsourcesontheAlaskancontinentalshelf.WhilePlafker (1969)usedtraveltimestoinferthatthetsunamisoriginatedalongtheoffshoreextensionofthePatton BayfaultsystemthatsurfacesonMontagueIsland(Figure1),tsunamisourceswerenotlinkedtospecific seafloorscarps.

Figure2.BathymetricmapwithnewmultibeamsurveythathighlightslineamentsrelatedtothePattonBayandCape

Clearefaults(blackarrows).Selectseismicprofiles(red),1964shorelineupliftmeasurementsinmeters(greendots),

and1964horizontaldirectionandmotioninmeters(redarrows)from1964earthquake(Plafker,1969).Distances(inkm)

arelabeledalongeachseismicprofile.ThewafflepatternontheCapeCleareBankbathymetryresultsfromcombining

sparse(1929)andtrackline(1965)bathymetricsurvey.InsetmapshowsaportionoftheCapeClearethrustand

back-thrustbestdescribedbyprofile2014_22(FigureS5).YellowlinesarefaultsmappedbyPlafker(1969).MMS=Mineral

ManagementServices;USGS=U.S.GeologicalSurvey.

3.

Methods

and

Data

Sincethevelocityoftsunami wavesisdirectlyproportionaltowaterdepth(Murty,1977),werefinethe Plafker(1969)tsunamitraveldistancesusingbathymetricsurveydata.Herewecombinetraveltimesfrom arun‐updatabase(Plafker,1969)withdigitalbathymetricdatatoidentifyseafloorscarpsresponsiblefor tsunami generation(Figure 1).Eyewitness accounts fromPuget, Whidbey, and ResurrectionBays near

Geophysical

Research

Letters

10.1029/2018GL081528

Sewardsuggesttwotectonicsourcescausedwaverun‐upsthatarrivedapproximately20and30minafterthe earthquake(Figure1).Plafker(1969)noteda north‐directedwaveandupwardfirstmotion(assuminga directtravelpath)forbothPugetandWhidbeyBays,whereasthetsunamionMiddletonIslandoriginated fromwestoftheislandwithadownwardfirstmotion.

TomapfaultsandrecordsliphistoryrelatedtocoseismicreleaseofthePWSasperity,weconductedahigh‐ resolutionbathymetricandsubbottomprofilingsurveyaroundtheJunkenTrough(FigureS1).Wechose thisfocusareaduetoitsproximityto thesplayfaultswithsurfaceruptureduringthe1964earthquake (Figure2). Thisglaciallyscouredcross‐shelftroughlikely hasa near‐continuousHolocenesedimentary record(e.g.,Jakobssonetal.,2014)andislocatedadjacenttoMontagueIsland.

Approximately27km2ofmultibeambathymetrydatawereacquiredaboardtheAlaskaDepartmentofFishand GameVesselR/VSolsticewithaResonSeaBat7111(100kHz;301beams).Duringdatacollection,tidalvalues werereferencedtoMeanLowerLowWaterandsoundingswereeditedandprocessedusingResonPDS2000 software. A final raster grid was created at 10‐m cell spacing and loaded into ESRI ArcGIS and QPS Fledermausforinterpretationand comparison toearlierbathymetric data(Figure2). Additionally,high‐ resolutionsingle‐channelseismicreflectionprofilesweresimultaneouslyacquiredusinga500‐J SIG2‐Miile minisparkersource(Balster‐Geeetal.,2019).Acousticfrequenciesbetween150and750Hzprovidedmeter‐ scaleresolutionandpenetrationofupto500ms(~350‐mdepth).Wecomplementthesenewdatawithlegacy bathymetricandseismicdata.HereweincludetheBoiseStatesparkerprofile2011_1(Libertyetal.,2013),the U.S.GeologicalSurvey(USGS)airgunprofile81_12,andtheMineralManagementServices(MMS)airgunpro­

file404.TheUSGSseismicprofile,acquiredin1981,consistedofa21.7‐Lairgunarray,50‐mshotspacing,anda 2.4‐km,24‐groupstreamer(Fruehnetal.,1999).TheMMSprofile,acquiredin1975,consistedofa29.5‐L air gunarrayanda3.6‐km96‐groupstreamer(Liberty,2013).Wedepthconvertedsparkerseismicprofilesusing avelocityof1,500m/sandairgunseismicprofilesusingstackingvelocitiesobtainedduringprocessing. KeyhorizonsidentifiedwithournewseismicsurveyrepresentmajorchangesinHolocenesedimentdeposi­ tionthroughoutthePWSandGulfofAlaskaregions(Finnetal.,2015;Haeussleretal.,2015;Libertyetal., 2013).Withourdataset,weidentifythreeseismicstratigraphicpackagesaboveacousticbasement(Figures3 andS2toS5).Theuppermostsequence,unitI,haswell‐definedcontinuoushorizontalreflectorsthatunder­ lietheseafloor.UnitIisusuallynotpresentbeneaththeshallowshelfregion,anditiscommonly10to30m thickintheJunkenTroughandothercross‐troughregionsalongtheGulfofAlaska(e.g.,Carlson,1989).We interpretthisunitasrelatedtothemillennial‐scaledepositionoffine‐grained,suspendedsedimentderived primarilyfromtheCopperRiverdelta(Jaegeretal.,1998;Kuehletal.,2017).

UnitIIhasaclearangularunconformityatitsbase,oftenmarkedbyathinacousticallytransparentbasal layer.Abovethebase,theunitfeaturesmoderate‐amplitudeparallelreflectorsandrangesfrom25to50m thick. We interpret this unit as being deposited soon after glaciers retreated from their Last Glacial Maximum(LGM)position,likelybetween17 and14ka(Kopczynskietal.,2017;Mann&Peteet,1994; Misartietal.,2012).Depositionofthisunitwasmostlyfocusedwithinthedeeptroughsthatliebelowthe earlyHolocenesealevels(Shugaretal.,2014).Weinferthesesedimentsweredepositedonasurfacethat wasabradedandleveledduringLGMiceadvances,andwereconstructtheHolocenedeformationhistory oftheactivefaultswiththisassumption.

UnitIIIconsistsofsedimentsthatliebelowtheLGMunconformityandaboveCenozoicacousticbasement. TheacousticcharacterofunitIIIstrataisvariable,withsomesedimentpackageshavingstrongparallel reflectorsandotherswithlateralvariabilityandweakornocoherentreflections.TheunitIIIstrataarelikely contemporaneouswiththelatePliocene/earlyPleistoceneYakatagaFormationofMiddletonIsland(Lagoe etal.,1993;Taliaferro,1932).

4.

Accretionary

Wedge

Thrusts

4.1. PattonBayFault

Themostdramaticexpressionofsurfacefaultruptureinthe1964earthquakewasalongthePattonBayfault system,identifiedonsouthernMontagueIsland(Plafker,1969).Thesystemoffaultsincludesenechelon, 45–70°northwest‐dippingnortheast‐trendingreversefaultswithupto8mofupliftinresponseto12mof south‐directeddifferentialshorteningacrossthePattonBayfault(Figure2).AsinferredbyPlafker(1969),

wedeterminethatthe19.5‐min(WhidbeyBay)and20‐min(PugetBay)tsunamiarrivalsalongtheKenai Peninsulain1964areconsistentwithmotionalongthePattonBayfault(Figure1).Thisfaulthasaclear seafloor expression and lies subparallelto the Kenai Peninsulamargin (Figure 2). From bathymetric differencing, Liberty et al. (2013) found the largest scarp along the Patton Bay fault immediately southwestofMontagueIslandontheseaflooroftheCapeCleareBank.There,a40‐m‐deepsubmerged wave‐cutplatformexperienced8to12 mofupliftin1964.Libertyetal. (2013)suggestedthatboththe faultscarpheightandamountof1964upliftdecreasefarthertothesouthwest.

Figure3.Northwest‐southeastseismicprofiles(seeFigure2forprofilelocations).Greenandredlinesrepresentkey

boundaries,wherethenumbersaboveeachfault(blacklines)representoffsetsfortheseafloor(SF),unitIbase(red

line),unitIIbase(green),andunitIIIbase(triangles).Thenumbersthatlieadjacenttomappedfaultsrepresentmeasured

faultdip.USGS81_12airgunprofileoverlapsthe2014_47andextendsacrosstheMiddletonIslandfault.MMS404extends

acrosstheCapeClearefaulttobeyondtheshelf.Tb=tertiaryacousticbasement;M=seafloormultiple.Seismic

profilelabelsaredistancesinkilometers.MMS=MineralManagementServices;USGS=U.S.GeologicalSurvey.Fault

abbreviationsarethesameasinFigure1.

AlongtheeasternmarginoftheJunkenTrough,ournewmultibeamresultsshowa3.7‐kmnorthweststepin thePattonBayfaultscarp,aswellasastepintheCapeClearfaultscarp(seebelow;Figure2),suggesting

Geophysical

Research

Letters

10.1029/2018GL081528

nonplanestrainand likelytranspressionaldeformationtoaccountforan obliquefaultorientationwith respecttorelativeplatemotion.Theseismicdatashowevidenceforrepeatedupliftalongthetwo40°to 70°north‐dippingreversefaultsegmentsofthePattonBayfault,withnorth‐dippingbedrocksurfaceson bothsidesofthefault(Figure3).Seismicprofileswithinthestep‐overzoneshowboththrustfaultstrands havesimilarsliphistories(profiles2014_47and2014_44;FiguresS2andS3),andthesefaultstransition westwardtoanarrowkinkband(2014_42;FigureS4)andsinglefaultnearthewesternJunkenTroughmar­ gin(2014_22;FigureS5).Seismicallytransparentbedrocklieswithin100moftheseaflooralongthelength ofthefault.Thislistricfaultmergeswiththemegathrustatabout18‐kmdepth(Libertyetal.,2013). Post‐LGMdepositionalpatternssuggestthegreatestthrowonthePattonBayfaulthasremainedonsouth­ ernMontagueIslandandtheadjacentCapeCleareBank,asinthe1964earthquake.WhereasthePattonBay faultonMontagueIslandandCapeCleareBankexperiencedupwardof10mofupliftduringthe1964earth­ quake,weinterpretaverageupliftbeneaththeJunkenTrough,andfartherwestontheJunkenBank,ofno morethanafewmetersperHoloceneevent.Specifically,wemeasure32mofpost‐LGMdisplacementonthe centralJunkenTroughprofile2014_42(FigureS4),12.9mofpost‐LGMdisplacementonprofile2014_22 (FigureS5),and9.4mofpost‐LGMdisplacementalongtheJunkenBankprofilefartherwest(Figure3). Assumingarecurrenceintervalof535years(e.g.,Shennanetal.,2014),weestimateanaverageslipperevent ofabout1mpereventbeneaththeJunkenTrough(~2mm/year)andabout0.3mper500‐yeareventbelow theJunkenTrough(~0.6mm/year).ThisestimatedHoloceneslipdistributionisconsistentwith1964slip models(Ichinoseetal.,2007;Johnsonetal.,1996),whichlocatealargeslippatchsurroundingsouthwestern MontagueIsland.Thelong‐termdeformationpatternindicatesthePattonBayfaulthasremainedstrongly coupledtothePWSasperityareaformost,orall,oftheHolocenerecord.Theregionofthegreatestslipis coincidentwiththelocationoftheSMAandliesatthesouthernlimitsofthecrustalduplexinginterpreted fromcrustalseismicprofiles(Figure1;Haeussleretal.,2015;Libertyetal.,2013).Ourbathymetricandseis­ micinterpretations, coupledwithtsunami traveltimes, stronglysuggest thatthePatton Bayfault near MontagueIslandwasresponsiblefortheearlytsunamiarrivalontheKenaiPeninsula.

4.2. CapeCleareFault

Ournewbathymetric datashowtheCapeClearefaultforminga 69‐m‐highseafloorscarpontheCape CleareBank(Figure2).Bathymetricdifferencingindicatesupliftofmorethan10monthisfaultduring the1964earthquake (Libertyetal.,2013), butthere wasnoidentifiedsurface ruptureoftheMontague Islandportionofthefault(Plafker,1969).SeafloorupliftalongtheCapeClearefaultincreasestothesouth­ westacrosstheCapeCleareBank(Libertyetal.,2013).Althoughpre‐1964bathymetricmeasurementsare sparsewithintheJunkenTrough,ournewbathymetrydataareconsistentwithsignificantupliftin1964. BeneaththeeasternJunkenTroughmargin,bathymetricdatashowa1.8‐kmnorthweststepintheCape Clearefaultandaseriesofbedrockknobsinthefault'shangingwall(Figure2).Thesparkerseismicprofiles definea~70°north‐dippingfaultwithnorth‐dippingreflectorsthatlieabovebedrockinthehangingwall andmostlyflat‐lyingreflectorsinthefootwall(Figure3).Inadditiontoobliqueconnectorfaultsthatstep betweenthrustsbeneaththeJunkenTrough,foldedstratabetweenstrands(profile2014_47;FigureS2)sug­ gestlocalizedshortening.Abackthrustandrelatednormalfaultonprofile2014_42(FigureS4)definesa smallgraben.Weestimatethatthesesmallthrustsandthebackthrustlikelymergebetween1‐and2‐km depths.BedrockexposuresinthehangingwallmakeitdifficulttoestimateHolocenesliprates;however, theshallowbedrocksurfacesuggestsgreaterHolocenesliponthisfaultthanonthePattonBayfaultbeneath theJunkenTrough.Flat‐lyingfootwallreflectorssuggestlittleHolocenedeformationimmediatelysouth ofthefault.AswiththePattonBayfault,thezoneofthegreatestcoseismicupliftappearsimmediatelysouth oftheSMA(Figure1).AlthoughtheCapeCleareandPattonBayfaultscanbemappedseparately,their proximityandthenumberoffaultstepoversindicatethatonalargescale,thesefaultsshouldbeconsidered togetherasasingularfaultsystemwhereshallowslipspanstwofaultstrands.

The10‐to12‐mindelaybetweenthefirstandsecondtsunamiwavesobservedinWhidbeyandPugetBaysis consistentwithverticalseafloordisplacementabout20to30kmfarthersouthfromthefirstsource.While theactiveCapeClearefaultislocatedonly10kmseawardofthePattonBayfault,imprecisetraveltime observationscouldpointtothisfaultasatsunamisource.Alternatively,motionona~100‐kmlong,west‐ to‐east, north‐side upanastomosingsea floorscarpbetween theJunken Troughand MiddletonIsland, whichwetermtheMiddletonIslandfault(Figures1and2),mayalsobeatsunamisource.

4.3. MiddletonIslandFault

Plafkeretal.(1978)mappedaseriesoffaultsandfoldsfromseismicreflectiondataintheeasternGulfof Alaskaregion.Theyidentifiedfaults,bothnorthandsouthofMiddletonIsland,whichlikelyupliftedthe island during earthquakes, and tilted the island between earthquakes (Savage et al., 2014). West of MiddletonIsland,ourlegacybathymetricmapshowstheMiddletonIslandfaultasawest‐trendingscarp upto 23‐m‐high, locatedimmediatelysouthof, and parallel to, theSMA (Figure 1). Slipon thisfault diminishestothewest,wherewemeasurea4.5‐m‐highscarptotheeastoftheJunkenTrough(Figures2 and3).Plafker(1969)measured3.5mofupliftneartheeasternlimitsofthisfaultduringthe1964earth­ quake,Savageetal. (2014) documentedfivepriorearthquakes thatcaused40 mofupliftofMiddleton Island,andHaeussleretal.(2015)documentedamajorthrustfaultimmediatelyeastoftheislandthatsplays from the 12–15‐km‐deep megathrust (Figure 1). Airgun profiles USGS 81_12 and MMS 404 showthe Middleton Island fault as a 6‐km‐wide series of parallel thrust faults with deformed footwall strata (Figure3).Deformationshallowsalongthenorthernfaultstrands,indicatinganorthwardageprogression offaultstrandsandthatthelatestfaultmotionsarefoundneartheSMA.

Althoughwedidnotacquirenewdataacrossthisfault,scarpsandfaultsidentifiedfromlegacyseismicand bathymetricdatasuggestupliftwestofMiddletonIslandlikelycausedthesecondofthe1964tsunamiarri­ valsinWhidbeyandPugetBaysandcausedtheinitialoutflowofwaterforthetsunamiatMiddletonIsland (Figure1).WeinfertheMiddletonIslandfaulthasremainedactivethroughoutlateHolocenetimebecause ofthelargeseafloorscarpwithrespecttosingleearthquakefaultmotionandbecauseoftherepeateduplifts documentedonMiddletonIsland.ThefaultscarpsterminatetothewestneartheCapeClearefault,where weobserve a grabenbetween the twofault systemsand abroadanticline inthe footwall blockofthe MiddletonIslandfault(Figure3).GiventheheightoftheseafloorscarpandslipacrosstheLGMunconfor­ mityonprofileUSGS81_12,thisfaultlikelyproducedtsunamisduringHoloceneearthquakes.Again,this faultislocatedupdipof, andparallelto, plateinterfaceduplexingand theSMA(Haeussleret al.,2015; Libertyetal.,2013;Figure1),andthereislittleevidenceofactivedeformationsouthofthisfaultonthe MMS404profile(Figure3).

TheMiddletonIslandfaultmatchesthetsunamitraveltimeanddirectionrecordedonMiddletonIslandin 1964,isconsistentwithafaultthatcaused5mofupliftonMiddletonIslandin1964(e.g.,Savageetal.,2014), andisassociatedwitha23‐m‐highseafloorscarpneartheHinchinbrookChannel,andHaeussleretal. (2014)imagedtheMiddletonIslandfaultasasplayfromthemegathrust(Figure1).Tsunamigenesisfrom thisfaultwaslikelyin1964andwilllikelyrepeatduringfutureearthquakes.

5.

Evidence

for

a

Long

‐

Lived

Asperity

Thepost‐1964platelockingabovethePWSasperity(Zwecketal.,2002)suggeststhesameregionwiththe largestcoseismicmotionin1964isaccumulatingstrainforthenextmegathrustearthquake.Basedonthe deformationpatternsofthesplayfaults,thecorrelationbetweenthefaultingandtheregionofhighcoupling insouthernPWS and the correlationwith theSMA, we suggestthe PWS asperityhas remained fixed throughout the Holocene, with the same faults coseismically accommodating upper‐plate shortening. Thus,theimpressivePattonBay,CapeCleare,andMiddletonIslandfaultseafloorscarpsresultedfrom repeatingruptures.Thesefaultswilllikelyproducesimilarupliftduringfutureearthquakes.Regardlessof thedetailedslipdistributionduringaparticulargreatearthquake,tsunamigenesisfromanyofthethree faultswillresultintraveltimedifferencesofonlyafewminutes.

Thecomplexsurfaceexpressionsoftheidentifiedfaultsliewithinanarrowzoneimmediatelyseawardofthe SMA,wherelistricthrustfaultssplayfromtheplateboundary(Figure1;Brocheretal.,1994;Haeussleretal., 2015;Libertyetal.,2013).Conversely,thesetoflargesubparallelsurfaceruptures,expressedasbathymetric scarps,areconstrainedtothe~150‐kmextentofthenorthwest‐trendingSMA.ThissuggeststhattheSMA definesthePWSasperity,andearthquakesfromthisasperityproducerepeatedtsunamigenicsurfacerup­ tures.Thegrowth faultingindicatethese faultshad similardisplacements duringmostHoloceneearth­ quakesandthatthePWSasperitythereforehaspersistedthroughmostHoloceneearthquakes.

Megathrustsplayfaultingisoftenexhibitedneartheouterridgesofaccretionaryprisms(e.g.,Beceletal., 2017;Collotetal.,2008;Mooreetal.,2007;Parketal.,2002).Incontrast,weshowobliqueshorteningalong

Geophysical

Research

Letters

10.1029/2018GL081528

thePWSsubductionzonesegment,locatedonthecontinentalshelf,thatisrespondingtoalateraltransition from shallowanglesubductionof theYakutat terraneto steeper‐dipping Pacificplate subduction(e.g., Brocheretal.,1994;Eberhart‐Phillipsetal.,2006;Kimetal.,2014).Theseresultssupporttheobservations ofBeceletal.(2017)thatthemorphologyandstructureofthesubductedslabcansubstantiallyinfluence thepatternofshortening,splayfaultgeneration,andtsunamigenerationinsubductionzones.Ourobserva­ tionsextendthisobservationbeyondthefrontalsubductionzoneprism.

6.

Conclusions

WeshowevidenceforrepeatedseafloorrupturesrelatedtoaregionofhighmomentreleaseofthePWS asperity.Faultsthatsplayfromthemegathrusthaverespondedwithsimilarcoseismicsurfacerupturepat­ ternsduringHolocenetime(perhapsthepast20to30earthquakes).Wesuggestthesesamefaultswillrup­ tureduringthenextmegathrustearthquake,producingsimilartsunamirun‐upandtraveltimesasduring the1964earthquake.Detailedseafloorandsubbottommappingfromanewbathymetricandseismicsurvey, coupledwithlegacygeophysicaldata,providesspatialandtemporalviewsofmegathrustbehaviors.Wecon­ cludethatthesurfaceexpressionofthesplayfaultsistiedtoplateboundaryconditions,indicatingapersis­ tenceofasperitiesduringmultipleearthquakes.Theseobservationsmayapplytoothersubductionzone systems with high tsunami hazards, especially where splay faults may surface far from the trench. Mappingplateboundaryandupper‐platestructuresisacriticalsteptowardunderstandingtsunamisources insubductionzones.

Acknowledgments

TheU.S.GeologicalSurveyEarthquake

HazardsProgramawardG11AP20143

andG13AP00021,inpart,fundedthis

work.WethanktheAlaskaDepartment

ofFishandGameandR/VSolstice

Captainandcrewforasuccessful

cruise.WethankDavidFinlayson,

AliciaBalster‐Gee,GerryHatcher,and PeteDartnellforhelpwithgeophysical dataacquisitionandprocessing.

ProMAXseismicprocessingsoftware

wasprovidedbyLandmarkGraphics

CorporationStrategicUniversity Alliancegrantagreement2013‐UGP‐

009000.MatthiasDelescluseandRob

Witterprovidedimprovementstothis manuscript.Anyuseoftrade,product, orfirmnamesisfordescriptive purposesonlyanddoesnotimply

endorsementbytheU.S.Government.

Legacybathymetricdataareavailable fordownloadonline(https://maps. ngdc.noaa.gov/).Legacyseismicdata areavailablefordownloadonline (https://walrus.wr.usgs.gov/namss/).

Seismicandbathymetricdataacquired

aspartofthisstudyareavailableonline (http://www.sciencebase.gov/).

References

Atwater,B.F.,&Hemphill‐Haley,E.(1997).Recurrenceintervalsforgreatearthquakesofthepast3,500yearsatnortheasternWillapaBay, Washington.U.S.GeologicalSurveyProfessionalPaper,1576,108.https://doi.org/10.3133/pp1576

Balster‐Gee,A.F.,Brothers,D.S.,Finlayson,D.P.,Liberty,L.M.,Dartnell,P.,Hatcher,G.A.,etal.(2019).Bathymetry,acousticback­ scatter,andminisparkerseismic‐reflectiondatasetscollectedsouthwestofMontagueIslandandsouthwestofChenegaIsland,Alaska duringfieldactivity2014‐622‐FA.U.S.GeologicalSurveyDataRelease.https://doi.org/10.5066/P9K1YQ35

Becel,A.,Shillington,D.J.,Delescluse,M.,Nedimović,M.R.,Abers,G.A.,Saffer,D.M.,etal.(2017).Tsunamigenicstructuresina creepingsectionoftheAlaskasubductionzone.NatureGeoscience,10(8),609–613.https://doi.org/10.1038/NGEO2990

Brocher,T.M.,Fuis,G.S.,Fisher,M.A.,Plafker,G.,Moses,M.J.,Taber,J.J.,&Christensen,N.I.(1994).Mappingthemegathrustbeneath thenorthernGulfofAlaskausingwide‐angleseismicdata.JournalofGeophysicalResearch,99(B6),11,663–11,685.https://doi.org/ 10.1029/94JB00111

Brothers,D.S.,Haeussler,P.J.,Liberty,L.,Finlayson,D.,Geist,E.,Labay,K.,&Byerly,M.(2016).Asubmarinelandslidesourceforthe devastating1964Chenegatsunami,southernAlaska.EarthandPlanetaryScienceLetters,438,112–121.https://doi.org/10.1016/j. epsl.2016.01.008

Bruns,T.R.(1983).ModelfortheoriginoftheYakutatblock,anaccretingterraneinthenorthernGulfofAlaska.Geology,11(12),718–721. https://doi.org/10.1130/0091‐7613(1983)11<718:MFTOOT>2.0.CO;2

Carlson,P.R.(1989).SeismicreflectioncharacteristicsofglacialandglacimarinesedimentintheGulfofAlaskaandadjacentfjords.

MarineGeology,85(2–4),391–416.https://doi.org/10.1016/0025‐3227(89)90161‐8

Carver,G.,&Plafker,G.(2008).PaleoseismicityandneotectonicsoftheAleutianSubductionZone—Anoverview.InJ.T.Freymueller,P.J. Haeussler,R.L.Wesson,&G.Ekström(Eds.),ActivetectonicsandseismicpotentialofAlaskaGeophysicalMonographSeries,(pp.43–63).

Washington,DC:AmericanGeophysicalUnion.https://doi.org/10.1029/179GM03

Christensen,D.H.,&Beck,S.L.(1994).Theruptureprocessandtectonicimplicationsofthegreat1964PrinceWilliamSoundearthquake.

PureandAppliedGeophysics,142(1),29–53.https://doi.org/10.1007/BF00875967

Cisternas,M.,Atwater,B.F.,Torrejón,F.,Sawai,Y.,Machuca,G.,Lagos,M.,etal.(2005).Predecessorsofthegiant1960Chileearthquake.

Nature,437(7057),404–407.https://doi.org/10.1038/nature03943

Collot,J.Y.,Agudelo,W.,Ribodetti,A.,&Marcaillou,B.(2008).Originofacrustalsplayfaultanditsrelationtotheseismogeniczoneand underplatingattheerosionalnorthEcuador–southColombiaoceanicmargin.JournalofGeophysicalResearch,113,B12102.https://doi. org/10.1029/2008JB005691

Eberhart‐Phillips,D.,Christensen,D.H.,Brocher,T.M.,Hansen,R.,Ruppert,N.A.,Haeussler,P.J.,&Abers,G.A.(2006).Imagingthe transitionfromAleutiansubductiontoYakutatcollisionincentralAlaska,withlocalearthquakesandactivesourcedata.Journalof GeophysicalResearch:SolidEarth,111,B11303.https://doi.org/10.1029/2005JB004240

Elliott,J.L.,Larsen,C.F.,Freymueller,J.T.,&Motyka,R.J.(2010).Tectonicblockmotionandglacialisostaticadjustmentinsoutheast AlaskaandadjacentCanadaconstrainedbyGPSmeasurements.JournalofGeophysicalResearch,115,B09407.https://doi.org/10.1029/ 2009JB007139

Enkelmann,E.,Valla,P.G.,&Champagnac,J.D.(2015).Low‐temperaturethermochronologyoftheYakutatplatecorner,St.EliasRange (Alaska):bridgingshort‐termandlong‐termdeformation.QuaternaryScienceReviews,113, 23–38.

Ferguson,K.M.,Armstrong,P.A.,Arkle,J.C.,&Haeussler,P.J.(2015).Focusedrockupliftabovethesubductiondécollementat MontagueandHinchinbrookIslands,PrinceWilliamSound,Alaska.Geosphere,11(1),144–159.https://doi.org/10.1130/ GES01036.1

Finn,S.P.S.P.,Liberty,L.M.L.M.,Haeussler,P.J.P.J.,&Pratt,T.L.T.L.(2015).Landslidesandmegathrustsplayfaultscapturedbythe lateHolocenesedimentrecordofeasternPrinceWilliamSound,Alaska.BulletinoftheSeismologicalSocietyofAmerica,105(5), 2343–2353.https://doi.org/10.1785/0120140273

Freymueller,J.T.,Woodard,H.,Cohen,S.C.,Cross,R.,Elliott,J.,Larsen,C.F.,etal.(2008).ActivedeformationprocessesinAlaska,based on15yearsofGPSmeasurements.InJ.T.Freymueller,P.J.Haeussler,R.L.Wesson,&G.Ekström(Eds.),Activetectonicsandseismic potentialofAlaska(Vol.179,pp.1–42).https://doi.org/10.1029/179GM02

Fruehn,J.,vonHuene,R.,&Fisher,M.A.(1999).Accretioninthewakeofterranecollision:TheNeogeneaccretionarywedgeoffKenai Peninsula,Alaska.Tectonics,18(2),263–277.https://doi.org/10.1029/1998TC900021

Fujiwara,T.,Kodaira,S.,No,T.,Kaiho,Y.,Takahashi,N.,&Kaneda,Y.(2011).The2011Tohoku‐Okiearthquake:Displacementreaching thetrenchaxis.Science,334(6060),1240–1240.https://doi.org/10.1126/science.1211554

Haeussler,P.J.,Parsons,T.,Finlayson,D.P.,Hart,P.,Chaytor,J.D.,Ryan,H.,&Liberty,L.(2014).Newimagingofsubmarinelandslides fromthe1964earthquakenearWhittier,Alaska,andacomparisontofailuresinotherAlaskanfjords.InSubmarinemassmovements andtheirconsequences(pp.361–370).Cham:Springer.https://doi.org/10.1007/978‐3‐319‐00972‐8_32

Haeussler,P.J.P.J.,Armstrong,P.A.P.A.,Liberty,L.M.L.M.,Ferguson,K.M.K.M.,Finn,S.P.S.P.,Arkle,J.C.J.C.,&Pratt,T.L.T.L. (2015).FocusedexhumationalongmegathrustsplayfaultsinPrinceWilliamSound,Alaska.QuaternaryScienceReviews,113, 8–22. https://doi.org/10.1016/j.quascirev.2014.10.013

Henstock,T.J.,McNeill,L.C.,&Tappin,D.R.(2006).SeafloormorphologyoftheSumatransubductionzone:Surfaceruptureduring megathrustearthquakes?Geology,34(6),485–488.https://doi.org/10.1130/22426.1

Ichinose,G.,Somerville,P.,Thio,H.K.,Graves,R.,&O'Connell,D.(2007).Ruptureprocessofthe1964PrinceWilliamSound,Alaska, earthquakefromthecombinedinversionofseismic,tsunami,andgeodeticdata.JournalofGeophysicalResearch,112,B07306.https:// doi.org/10.1029/2006JB004728

Jaeger,J.M.,Nittrouer,C.A.,Scott,N.D.,&Milliman,J.D.(1998).Sedimentaccumulationalongaglaciallyimpactedmountainous coastline:NortheastGulfofAlaska.BasinResearch,10(1),155–173.https://doi.org/10.1046/j.1365‐2117.1998.00059.x

Jakobsson,M.,Andreassen,K.,Bjarnadottir,L.R.,Dove,D.,Dowdeswell,J.A.,England,J.H.,etal.(2014).ArcticOceanglacialhistory.

QuaternaryScienceReviews,92, 40–67.https://doi.org/10.1016/j.quascirev.2013.07.033

Johnson,J.M.,Satake,K.,Holdahl,S.R.,&Sauber,J.(1996).The1964PrinceWilliamSoundearthquake:Jointinversionoftsunamiand geodeticdata.JournalofGeophysicalResearch,101(B1),523–532.https://doi.org/10.1029/95JB02806

Kim,Y.,Abers,G.A.,Li,J.,Christensen,D.,Calkins,J.,&Rondenay,S.(2014).Alaskamegathrust2:Imagingthemegathrustzoneand Yakutat/PacificplateinterfaceintheAlaskasubductionzone.JournalofGeophysicalResearch:SolidEarth,119,1924–1941.https://doi. org/10.1002/2013JB010581

Kopczynski,S.E.,Kelley,S.E.,Lowell,T.V.,Evenson,E.B.,&Applegate,P.J.(2017).LatestPleistoceneadvanceandcollapseofthe Matanuska–Knikglaciersystem,Anchoragelowland,southernAlaska.QuaternaryScienceReviews,156,121–134.https://doi.org/ 10.1016/j.quascirev.2016.11.026

Kuehl,S.A.,Miller,E.J.,Marshall,N.R.,&Dellapenna,T.M.(2017).RecentpaleoseismicityrecordinPrinceWilliamSound,Alaska,USA.

Geo‐MarineLetters,37(6),527–536.https://doi.org/10.1007/s00367‐017‐0505‐7

Lagoe,M.B.,Eyles,C.H.,Eyles,N.,&Hale,C.(1993).TimingoflateCenozoictidewaterglaciationinthefarNorthPacific.Geological SocietyofAmericaBulletin,105(12),1542–1560.https://doi.org/10.1130/0016‐7606(1993)105<1542:TOLCTG>2.3.CO;2

Liberty,L.M.(2013).Retrieval,processing,interpretationandcatalogingoflegacyseismicreflectiondata,GulfofAlaska,USGeological SurveyFinalTechnicalReport(17P).Retrievedfromhttps://earthquake.usgs.gov/cfusion/external_grants/reports/G12AP20078.pdf Liberty,L.M.,Finn,S.P.,Haeussler,P.J.,Pratt,T.L.,&Peterson,A.(2013).MegathrustsplayfaultsatthefocusofthePrinceWilliam

Soundasperity,Alaska.JournalofGeophysicalResearch:SolidEarth,118,5428–5441.https://doi.org/10.1002/jgrb.50372

Mann,D.H.,&Peteet,D.M.(1994).ExtentandtimingofthelastglacialmaximuminsouthwesternAlaska.QuaternaryResearch,42(02), 136–148.https://doi.org/10.1006/qres.1994.1063

Misarti,N.,Finney,B.P.,Jordan,J.W.,Maschner,H.D.,Addison,J.A.,Shapley,M.D.,etal.(2012).EarlyretreatoftheAlaskaPeninsula GlacierComplexandtheimplicationsforcoastalmigrationsofFirstAmericans.QuaternaryScienceReviews,48, 1–6.https://doi.org/ 10.1016/j.quascirev.2012.05.014

Moore,G.F.,Bangs,N.L.,Taira,A.,Kuramoto,S.,Pangborn,E.,&Tobin,H.J.(2007).Three‐dimensionalsplayfaultgeometryand implicationsfortsunamigeneration.Science,318(5853),1128–1131.https://doi.org/10.1126/science.1147195

Murty,T.S.(1977).Seismicseawaves:Tsunamis(No.198),BulletinoftheFisheriesResearchBoardofCanada(Vol.198).Departmentof FisheriesandtheEnvironment,FisheriesandMarineService.Ottawa(Canada)IX,337S.,331

Park,J.O.,Tsuru,T.,Kodaira,S.,Cummins,P.R.,&Kaneda,Y.(2002).SplayfaultbranchingalongtheNankaisubductionzone.Science,

297(5584),1157–1160.https://doi.org/10.1126/science.1074111

Parsons,T.,Geist,E.L.,Ryan,H.F.,Lee,H.J.,Haeussler,P.J.,Lynett,P.,etal.(2014).Sourceandprogressionofasubmarinelandslideand tsunami:The1964GreatAlaskaearthquakeatValdez.JournalofGeophysicalResearch:SolidEarth,119,8502–8516.https://doi.org/ 10.1002/2014JB011514

Plafker,G.(1969).TectonicsoftheMarch27,1964Alaskaearthquake:U.S.GeologicalSurveyProfessionalPaper543–I,74p.,2sheets, scales1:2,000,000and1:500,000.Retrievedfromhttps://pubs.usgs.gov/pp/0543i/

Plafker,G.,Bruns,T.R.,Carlson,P.R.,Molnia,B.F.,Scott,E.W.,Kahler,R.,&Wilson,C.(1978).Petroleumpotential,geologichazards, andtechnologyforexplorationintheoutercontinentalshelfoftheGulfofAlaskaTertiaryprovince.USGeologicalSurveyOpenFile Report78–490,50p.

Saillard,M.,Audin,L.,Rousset,B.,Avouac,J.‐P.,Chlieh,M.,Hall,S.R.,etal.(2017).Fromtheseismiccycletolong‐termdeformation: LinkingseismiccouplingandQuaternarycoastalgeomorphologyalongtheAndeanmegathrust.Tectonics,36,241–256.https://doi.org/ 10.1002/2016TC004156

Savage,J.C.,Plafker,G.,Svarc,J.L.,&Lisowski,M.(2014).ContinuousupliftneartheseawardedgeofthePrinceWilliamSoundmegathrust: MiddletonIsland,Alaska.JournalofGeophysicalResearch:SolidEarth,119, 6067–6079.https://doi.org/10.1002/2014JB011127 Shennan,I.,Bruhn,R.,Barlow,N.,Good,K.,&Hocking,E.(2014).LateHolocenegreatearthquakesintheeasternpartoftheAleutian

megathrust.QuaternaryScienceReviews,84, 86–97.https://doi.org/10.1016/j.quascirev.2013.11.010

Shugar,D.H.,Walker,I.J.,Lian,O.B.,Eamer,J.B.,Neudorf,C.,McLaren,D.,&Fedje,D.(2014).Post‐glacialsea‐levelchangealongthe PacificcoastofNorthAmerica.QuaternaryScienceReviews,97,170–192.https://doi.org/10.1016/j.quascirev.2014.05.022

Sieh,K.,Natawidjaja,D.H.,Meltzner,A.J.,Shen,C.C.,Cheng,H.,Li,K.S.,etal.(2008).Earthquakesupercyclesinferredfromsea‐level changesrecordedinthecoralsofwestSumatra.Science,322(5908),1674–1678.https://doi.org/10.1126/science.1163589

Taliaferro,N.L.(1932).GeologyoftheYakataga,Katalla,andNichawakdistricts,Alaska.BulletinoftheGeologicalSocietyofAmerica,

43(3),749–782.https://doi.org/10.1130/GSAB‐43‐749

Zweck,C.,Freymueller,J.T.,&Cohen,S.C.(2002).Three‐dimensionalelasticdislocationmodelingofthepostseismicresponsetothe1964 Alaskaearthquake.JournalofGeophysicalResearch,107(B4),2064.https://doi.org/10.1029/2001JB000409