Oblique strike-slip faulting of the central Cascadia submarine forearc

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 102,NO. B4, PAGES 8217-8243,APRIL 10, 1997 Oblique strike-slip faulting of the central Cascadia submarine forearc Chris Goldfinger and LaVeme D. Kulm Collegeof OceanicandAtmospheric Sciences, OregonStateUniversity,Corvallis Robert S. Yeats, Lisa McNeill, and Cheryl Hummon Departmentof Geosciences, OregonStateUniversity,Corvallis Abstract. At leastnineWNW trendingleft-lateralstrike-slipfaultshavebeenmappedon the Oregon-Washington continental marginusingsidescan sonar,seismicreflection,andbathymetric data,augmentedby submersible observations.The faultsrangein lengthfrom 33 to 115 km and crossmuchof the continental slope.Five faultsoffsetboththeJuande FucaplateandNorth Americanplatesand crossthe plateboundarywith little or no offsetby the frontal thrust. Leftlateralseparation of channels, folds,andHolocenesediments indicateactiveslipduringthe HoloceneandlatePleistocene.Offsetof surficialfeaturesrangesfrom 120to 900 m, anddisplaced subsurface piercingpointsat the seawardendsof the faultsindicatea minimumof 2.2 to 5.5 km of total slip. Near their westerntips,fault agesrangefrom 300 ka to 650 ka, yieldinglate Pleistocene-Holocene slip ratesof 5.5 + 2 to 8.5 + 2 mrn/yr. The geometryand slip directionof thesefaultsimpliesclockwiserotationof fault-boundedblocksaboutverticalaxeswithin the Cascadiaforearc. Structuralrelationships indicatethatsomeof thefaultsprobablyoriginatein the Juande Fucaplateandpropagate intotheoverlyingforearc.The basement-involved faultsmay originateas shearsantitheticto a dextralshearcouplewithin the slab,asplate-couplingforcesare probablyinsufficientto rupturethe oceaniclithosphere.The setof sinistralfaultsis consistent with a modelof regionaldeformation of the submarine forearc(definedto includethedeforming slab)by right simplesheardrivenby obliquesubduction of theJuande Fucaplate. junctions and includes the smaller Explorer plate to the north, Introduction whichmaynst be presently subducting [Rohrand Furlong, The anelasticresponse of forearcsto obliquesubduction can be highly variable. The mostcommonlyrecognizedform is arc-parallel strike-slip faulting. This type of strain partitioning has been recognizedin Sumatra [Fitch, 1972; Jarrard, 1986;McCaffrey, 1991], the Kurils [Kimura, 1986], 1995]. JDF-NOAM convergence is estimated as 40 mm/yr, directed 062 ø at 45øN along the deformation front (rotation poles of DeMets et al. 1990]). No active arc-parallel faults equivalent to the MTL or Great Sumatran fault have been identified onshorein Cascadia. Snavely [1987] inferred that the Philippines[Karig et al., 1986],SouthAmerica[Dewey the Fulmar fault, a north striking dextral strike-slip fault, and Lamb, 1992], and other forearcs[Beck, 1983]. The offsets the continental slope and outer shelf in Oregon by Sumatranforearc translatesalong the arc-parallel Great about 200 km and attributed an abrupt truncation of the Sumatran fault, with the slip rate controlledby convergencebasaltic Siletzia terrane to this fault (Figure 2). The Fulmar obliquity[Fitch, 1972;Jarrard, 1986;McCaffrey, 1991]. In fault exhibits small offsets of Quaternary strata in southern Japan,the Median TectonicLine (MTL) servesa similar role Oregon but was mainly active in the Eocene [Snavely, 1987]. in the subductionof the PhilippineSea plate [Fitch, 1972; Some discontinuousarc-parallel faults have been identified in Sangawa,1986]. The Aleutianarc exhibitsbotharc-parallel the Cascadia forearc, both onshore and offshore which may translationand rotation of obliquely orientedblocks about accommodate some northward translation of the Cascadia vertical axes [Geist et al., 1988; Ryan and Scholl, 1993]. forearc (Figure 2) [Weaver and Smith, 1983; Blake et al., Both of thesemechanisms are favoredin obliquesubduction 1985; Niem et al., 1992; Kelsey and Bockheim, 1994; because the high-angle faults concentratehorizontal shear Goldfinger, 1994; Goldfinger et al., 1992a]. more effectivelythan the dippingsubductioninterface[Fitch, Paleomagnetically determined clockwise rotations of 1972]. The Cascadiasubduction zoneconsists of two smallplates, theGordaandJuande Fuca(JDF),subducting to thenortheast beneaththe North Americanplate(NOAM) (Figure1). The subduction systemis bounded to thenorthandsouthby triple Copyright1997 by the AmericanGeophysical Union. Paper number 96JB02655. 0148-0227/9 7/96JB-02655509.00 coastal basalts in Oregon and Washington suggest that a processof dextral shearof the forearchas operatedthroughout the Tertiary. Miocene (12-15 Ma) Columbia River Basaltsin western Oregon are rotated 10-30ø clockwise, and Eocene Siletz River Volcanics are rotated up to 90ø clockwise [Wells and Heller, 1988; England and Wells, 1991]. Mechanisms proposed to explain these rotations include microplate rotation during terrane accretion,basin and range extension, distributed small block rotation, or a combination of these (see Wells and Heller [1988] for summary). 8217 8218 GOLDFINGER ET AL.: CASCADIA STRIKE-SLIP FAULTING -130 ø -125 o -120 o deep-tow sidescan processing system. Processing included I I towfish positioning, geometric 'and speed corrections, correctionfor towfish attitude, georeferencingof image pixels to a latitude-longitudegrid, histogramequalization,and image enhancement. The final imagery was then integrated with other data layers in a raster/vector Geographical Information System(GIS) for analysis[Goldfinger and McNeill, 1996]. BC 50 ø Regional.sidescandata [EEZ-SCAN 84 ScientificStaff, 1986] using the (GLORIA) shallow-towed system were also used in interpretation of margin structure. The GLORIA imagery was georeferenced to Hydrosweep bathymetry collected by us in Washington (1993), or existing National Oceanic and Atmospheric Administration (NOAA) SeaBeam bathymetryin Oregon in order to producea consistentspatial Seattle WA Juan de Fuca Plate 45 ø -- ...L :. Pacific • Plate • o '.! D data set. In our structural interpretationswe have used about 30,000 km of seismic reflection profiles collected by academic institutions,the U.S. GeologicalSurvey,NOAA, and the petroleumindustry[Goldfingeret al., 1992a]. The seismic AMS-150 reflection profiles vary widely in quality, depth of penetration, and navigation accuracy and range from singlechannel sparker records navigated with LORAN A to 144- -,•[i .... "•, .--SeaMARCIA channel digitalprofiles navigated withGPS. An important Gorda •' Plate \ 400 I -----' Portland datasetis a multichannel seismic surveyconducted off central Oregonin 1989[MacKayet al., 1992]. Approximately 2000 CA Me, i I km of NAVSTAR-navigated 144-channel reflection profiles were collected and processed through time migration at the University of Hawaii. Three of the Oregon strike-slip faults Figure 1. Tectonicmap of the Cascadiasubductionzonefront withinthis survey(Figure3), Juande Fuca plate system. ApproximateSeaMARC 1A intersectthe deformation allowing detailed structural analysis, as well asdetermination sidescan coverage from ThomasThompson cruiseTT 020 shownshaded,boxesindicateapproximateAlpha Marine Systems(AMS) 150-kHz surveyareas. of fault slip rates. Fault Retrodeformation The abundance of seismicandsidescan dataallowedus to determinethe slip rates on five transversestrike-slip faults on Wehave identified nineWNWtrending transverse structures theOregon andWfishington margin.Themethod forsliprate deformingthe Cascadiaaccretionary prismOff Oregonand Washington. Eight 0f thesestructuresare left-lateral strike- determinationis describedhere, as it is similar for all five faults. We utilized the simple geometry of trenchward slip faults based on measuredoffsetsof surfaceand subsurface thickening abyssal plainsediment wedges, asdetermined by features. Five of the transverse faults can be traced across the plateboundaryinto the subducting Juande Fucaplate,up to 21 at leastonetrench-parallel andtwo trench-normal reflection profiles adjacent to each strike-slip fault (Figure 4). If the km seaward of the deformation front. Documentation of these geometry of the pre-faulting trenchwardthickening wedges is faults comesfrom sidescansonardata, swath bathymetry, simple: three profiles can define the wedge geometry seismic reflection data, and field observations from sufficiently to use them as three-dimensional(3-D) piercing submersibles. In thispaperwe summarize detailedsurveysof points,the offsetsof which representsthe net slip of the these faults on the central Cascadia margin of Oregon and Washington and discusstheir origin, structural significance, and implications for deformationof the submarineforearc. fault. The three Oregon faults used many more profiles than did the Washingtonfaults, which had the minimum numberof profiles needed. Reflection data indicate that the geometry of the wedges is simple; and the layers that bound them are approximatelyplanar over the distancesinvolved. Goldfinger Methods et al. [1996a] also calculatedthe net slip of one fault (the Data Acquisition and Processing High-resolution sidescansonar data were collected with a deep-towed SeaMARC1A 30-kHzsystem capable of imaginga 2-kmor a 5-km swathwidthwithspatialresolutions of 1 and 2.5 hi, respectively.An AlphaMarineSystems (AMS) 150kHz system was i•sed to collect sidescan data on the continentalshelvesof Oregonand Washington,with a 1-km Wecoma fault) using isopach plots of two subsurfaceunits. Using the wedge geometry capturestotal offset by the fault, which often is underestimated in measurements of offset isopachsdue to drag folding and local velocity effectsnear the fault. The five faultsfor whichWeusedthistechnique havea verticalcomponent of offset and showpronounced growth swathwidth and0.5-m resolution.Approximate coverageof strata on the downthrowh block; thus we were able to easily distinguishsynfaultingand prefaultingunits (Figure 3). We identified the point in the section at which interval thickness thesesurveysis shownin Figure 1. All sidescandata were located using Global Positioning System (GPS), then processedusing Oregon State University's (OSU's) Sonar block to thinner on the downthrown block. Synfaulting sedimentary units could not be used for this determination across the fault reversed from thicker on the downthrown -127 ø -126 ø -125 ø -124 ø Vancouver :123 ø -122 ø 8219 Idand t Ql.tcnpic Peninsula 48 ø shelf edge 5.5 _+ /i 6.7+_3 Gra• Harbor 47 ø ;) juande ' z Fuca Willapa Bay plate 46 ø -'-"::::•"-<::•::: :8.5._+2. ..... ....... ::•;!:':5,7 +_ NorthAmerican plate S?':w':•:."'i ;"' 6.6 :.ñ........ 45 ø ,Yaquina Bay Alsea Bay ":::::::::::::::::::::::::: . ":.:•: 0 : '.::<. .i'. :';'i"2':":::.-:.-:-.:ii .-- .. •:i?•:;'-.. .. :;:!:.i•11:•: ' ";::.i;:L.".. '•i western edge of Siletzia terrane 44 ø ' ":'.-:.:'-,w ::::"":.::F.;:•:-:;;11:'-:::'.-".9<:-'... z .......... ;i:::•'• .•"•.• ;S::•;:;:•/•:: -'•:.';::i =======================':.'::!':.5' :Z;:..:-':"-:i--"::** Coos Bay 43 ø .... ........ Cape Blanoo 0 - M>3,5 = M<3,5 shelf edge ":'..... ..:.: . .... 42 ø Figure 2. Active tectonicmap of the Oregon-Washington continentalmargin Faultsand anticlines shown;synclinesdeletedfor clarity. JDF-NOAM vectorof DeMetset al. [1990] is shown. Siletzia terraneboundarybasedon magneticsandreflection-refraction data. Strike-slipfaultsshownwith slip rates.inmrn/yr. NNF, North Nitinat fault; SNF, SouthNitinat fault; WCF, Willapa Canyonfault; WF, Wecomafault; DBF, Daisy Bank fault; ACF, Alvin Canyonfault; HSF, HecetaSouthfault; CBF, Coos Basinfault;TRF, Thompson Ridgefault.NF,NitinatFan;AF,Astoria Fan.Majordepocenters are shownby stipple:WB, Willapa Basin;AB, AstoriaBasin;NB, NewportBasin;CBB, CoosBay Basin. Submarine Banks are NB, Nehalem Bank; HB, Heceta Bank; CB, Coquille Bank. All offshore seismicityis shown,with BlancoFractureZoneandGordaplateeventsremoved.Focalmechanisms 1, 2, and 3, are M 3.3, M 7.1, andM 5.8 slabearthquakes, respectively,andare discussed in the text. 8220 GOLDFINGER ET AL.: CASCADIA STRIKE-SLIP FAULTING o o / / / • / / / / :i.:.-"'--i!½i:.::i::!i½:•i:.:-?i•i•i;. ::::•::i'.-'?.-'-"•)½'.-":: .... •" I ::½!½ii½!i;ii::i'..%;D;f; •,.-"i!!iii• ,' -:..:.: • ::::::::::::::::::::::::::::::: .-.--:•5•55½•5½ ..-• .- :::.•:•-:::....:: .::½:•..-.. ,.:;.•$ ...... O •: • :::::::::::::::::::::::::::: ':..:::::::.::..; :.'..-::... :.v•.:.•-•::. ................................ . ::..:....•:: :..,.:,.:•,,.::.• o'5 / / ,-0 .;!i;•;:i::'!-:.ii;;;:•-;½:•i:½...::!' :'";'""'•!!•!!i •'• 4,'.-.:': / E / (;:)•S) •IAII,L "I•AV/:EI, 0 0 0 0 0 (0:::!$):::!I/•I.L AVM-OMJ. GOLDFINGER ET AL.: CASCADIA STRIKE-SLIP FAULTING a. Restored to pre-faulting geometry (600 ka) ... .•....... •' -.-•:.•::•....•:•-•);•:•::;:.. :.::2• ....... 8221 determinedthe fault offset required to producethe observed thicknesschange for two pre-faulting units (units 2 and 3 in Figure 3). If Z 1 and Z 2 are the thicknessesof the offset unit acrossthe fault, at is the angle betweenthe bottom and top contactsof the unit (defining the eastwardthickeningwedge), and 13is the plan view angle betweenthe seismicprofile and the fault, net fault slip (S) is then given by S = [Z1-Z2/tan at] cos 13.For the three Oregon faults, we also used a trial and error restorationusing the reflection profiles directly. To find the best fit offset required to restore the abyssal plain geometry, we iteratively tested the offset to find the best match of between 10 and 18 individual reflectors within the prefaulting section using the margin-parallel and marginnormal reflection profiles and correcting for the difference betweenfault strike and profile strike (as with 13above). The geometryused in both methodsis illustratedschematicallyin b. Unrestored,post-faultinggeometry(present-day) Figure 4. The error estimate includes the minimum and seismic profile [5 Wecoma fault maximum fault separationthat could be accommodatedwithin the interpretationof the seismicdata. We estimatedthe age of faulting by convertingthe two-way travel time to the base of the growth stratato depthin meters. This conversion used an average velocity of 1680 m/s, calculatedfor the upper 400 m of the Nitinat fan from Ocean Drilling Program(ODP) drilling and reflectiondata [Hyndman and Davis, 1992; Shipboard Scientific Party, 1994]. We derived fault age by using a net sedimentation rate of 100 crn/1000 years for the Nitinat fan calculated from the 1993 ODP drill sites [Shipboard Scientific Party, 1994] and 110 cm/1000 years for the Astoria fan [Goldfinger, 1994; Figure 4. Cartoon diagram illustrating strike-slip fault Goldfinger et al., 1996a] calculatedusing the thicknessand restoration. (a) Abyssal plain section restored to age of the fan. To establish the age of the Astoria fan, we prefaultingconfiguration(600 ka). (b) Unrestoredsection correlated a prominent seismic reflector at the base of the fan as imaged on MCS line 37 (present-day). The three- from the Wecoma fault to Deep Sea Drilling Project (DSDP) dimensional (3-D) wedgeboundingunits 2 and 3, as well Site 174A (70 km southeast)usinga seismicreflectionprofile as individual reflector-bound intervals within the units, connecting the drill site to the fault [Kulm et al., 1973a, b; form the 3-D piercing points which were matchedacross Goldfinger et al., 1996a]. A profound lithologic change the fault to determinefault slip. Unit 2 resolvedthe fault offset somewhatbetter becauseit thickenedmore rapidly to the eastthanunit 3. Relationusedto calculatenet slip is given in the text. The variablesusedare shownhere:c• is the angle betweenthe boundingreflectorsof the principal seismicunits and B is the angle betweenthe strike of the seismicprofile used in restoration,and the strike of the fault. Z1 and Z2 are thicknesses of the restoredunit in the upthrownand downthrownblocks,respectively.Example between is for the Wecoma section is subperpendicular to the sediment transport direction. Thus it is unlikely that significant time transgressionhas occurredbetween Site 174A and the Wecoma fault. For the other two Oregon faults (Daisy Bank and Alvin Canyonfaults) we also usethe baseof the fan to estimatefault age, although it is likely to be slightly older due to the more southerly (distal) position of these faults. Although additional uncertainties exist in sedimentation rates, age of fault. because the presence of growth strata invalidates the assumptionthat measuredgeometric changeson the profiles result from horizontal fault offset only. In practice,we used two methods to determine fault slip. For all five faults, we determined the wedge geometry from the seismic grid, then sand turbidites of the fan and silt turbidites of the abyssal plain sequence was observed at the depth of this reflector in the drill cores. Biostratigraphic analysis of the coresfrom Site 174 yields an age of 760 + 50 ka (J. C. Ingle, Stanford University, written communication, 1995). We infer that the age of the base of the fan is approximatelythe same at the Wecoma fault becausewe observeno significant onlap or offlap relationshipswithin several hundred vertical meters of the base of the fan and the trend of the seismic the fan, and seismic velocities, we use the same estimate of Figure 3. The (a) Wecoma,(b) DaisyBank,and(c) Alvin Canyonfaultsimagedon a N-S 144-channel migrated reflectionprofile (MCS line 37) approximately3 km seawardof the deformationfront. Seismicunitbounding reflectors usedfor faultrestoration areshownin white. Baseof growthstratais shownwith whitearrows.Growthstratathickenon thedownthrown block;prefaultingunits2 and3 thin on the downthrownblock as a resultof strike-slipmotion. (d) Regionalseismicreflectionline showingthe basementpopupstructures (P) thatsimilarlyoffsetoceaniccrustat the threefaults. Fourth faultunderthekilometer scaleisa tearfaultin thesedimentary section only. 8222 GOLDFINGERET AL.:CASCADIASTRIKE-SLIP FAULTING •900 •000 •00 •00 •000 oo o •Oø •,• •900 ?•000 oo•$ GOLDFINGER ET AL.: CASCADIA STRIKE-SLIP FAULTING error (+ 50 kyr) here becausewe are unable to quantify these errors independently. The age of the fault is then simply the depth in meters of the oldest growth strata, divided by the sedimentationrate in meters per 1000 years. The age of first vertical motion for each fault is calculatedin this way, and the net slip can then be divided by the age to obtainthe slip rate of the fault. Any pure strike-slipmotion (no vertical component)that occurredprior to vertical displacementwould result in an underestimateof the age of the fault, and thusour derivedslip rate valueswill be too high. In two cases(Wecoma and North Nitinat faults) we were able to calculate slip rates independentlyusing measuredoffsetsof late Pleistocenechannels(120 m and 150 m, respectively)and estimated channel ages. Seismic reflection records showing the truncation of deep-sea channel walls and numerous sediment hiatuses in cores from the axial part of these channelsdocument the erosive characterof the coarse-grained late Pleistoceneturbidity currents in this region [Griggs and Kulm, 1973]. We estimatethe age of the offset channelwalls to be 12-24 ka, consistent with incision during the last episode of high turbidity-current activity during the Wisconsin low stand [Goldfinger et al., 1992a]. We assume fault motion was constant during the Holocene and late Pleistocene. Using this age range, we calculated a latest Pleistocene-Holocene slip rate of 8.3 + 4 and 8.5 + 4 mm/yr for the North Nitinat and Wecoma faults, respectively. These rates were similar to, but somewhathigher than, the estimates from retrodeforming the fault. The higher late Pleistocene slip rate is expected, since the point at which the growth stratawere measured,originally locatedat the fault tip, is now 20 km landwardof the tip. Since fault slip ratesdecreasefrom the centerof a fault toward its tip [Bilham and Bodin, 1992; $cholz et al., 1993], the retrodeformation method using the entire movement history includes low slip rates from the fault's early history and shouldyield a lower averagerate. The slip rate may also have varied in time over the life of the fault. Cascadia Transverse Strike-Slip Faults North Nitinat Fault Two prominent Washington margin faults, the North 8223 fan (Figure 2). We surveyedthe NNF in 1993 usingSeaMARC 1A with both 2-km and 5-km swaths and coincident Hydrosweepswathbathymetry. The 5-km swathsurveyshows that the fault intersects the deformation front at 47ø24'N, strikes 282 ø, and extends 20 km seaward of the deformation front into the Juan de Fuca plate. The NNF intersectsand crossesthe deformationfront at a large slumpon the landward vergent frontal-thrust anticline. The NNF cuts and offsets slump debris within the arcuateslump, as well as the slump scarpand the crestof the marginal ridge, 150 m left laterally (Figure 5). At the deformationfront the fault offsetslandward vergent protothrustsby 300-400 m. Figure 5 shows that these structures and the frontal thrust anticline bend 10-15 ø to the left where they intersectthe NNF. A mud volcanostraddles the fault near a right (restraining) bend 9 km seawardof the deformation front (Figure 6). The 2 km swath revealed a submarine channel parallel to the base of the continental slope on the abyssalplain. This channel is cut and offset 150 m left-laterally by the NNF (Figures 5 and 6). The offset channel wall has subsequentlyslumped (Figure 6), but the left separationis still apparentusing the wider swath in Figure 5 to map the channel wall. We estimatethe age of the offset channel wall to be 16-20 ka, spanningthe last Pleistocene lowstand at 18 ka, when turbidite activity and channel downcutting were at a maximum [Nelson, 1968; Griggs and Kulm, 1973]. This timing is consistentwith activity in other submarine channels on the Cascadia margin, although minor channel cutting continuedinto the Holocene [Nelson, 1968]. The Holocene-latePleistoceneslip rate of this fault, basedon the observed offset and inferred channel age, is 8.3 + 1.0 mm/yr, the error range reflecting uncertainty in the timing of the channel cutting episode. Further evidence of the leftlateral motion on this fault is the presenceof a mud volcanoat a restraining bend in the fault (Figure 6). We infer that the mud volcano results from elevated pore fluid pressuresdue to the transpressional fault geometry,resultingin the extrusion of deeply buried abyssalplain sediments. Restoration of fault motion on the abyssal plain using three seismicprofiles (University of Washington(UW) cruise TT-063, 1971) yielded a net left-lateral slip of 2.2 + 0.3 km for the NNF. The inferred age of the oldest growth strata adjacentto the fault is 400 + 50 ka, which we take as the time of initial fault motion on the abyssal plain near the front. We assume that horizontal and vertical Nitinat fault (NNF) and South Nitinat fault (SNF), were deformation initially recognizedby one of us (L.D. Kulm) in a 1971 air gun motion began concurrentlyand were continuousover the life seismicsurveyon the southernflank of the Nitinat submarine of the fault. From the calculatednet slip and age for the fault Figure 5. (top) SeaMARC 1A sidescansonarimageof the intersectionof the North Nitinat fault and the plateboundaryoff centralWashington.Light areasare highbackscatter.Fault offsetsthe landwardvergentthrustridge, slumpdebris,slumpheadwallscarps,andan abyssalplain channel. The fault is not offsetlaterallyat the plateboundarythrust.Swathwidth is 5 km. (bottom)Interpretationof the sidescan image. A local 20ø bend in the strike of accretionarywedge anticlinesoccursacrossthe fault, which widensand branchesinto multiple tracesat the deformationfront. Offset of abyssalplain channelis shownin detail in Figure 6. Extensiveslumpingof the backlimbof the landwardvergentfrontal anticlineof the accretionarywedgemay be triggeredby strike-slipfaulting. Two episodesof slumping are indicatedby the glide pathsof the youngerslump blockswhich have overriddenearlier slump debris. Someinterpretationdetailsarebasedon 3-D stereovisualizationof the imageryandcoregistered swathbathymetry. 8224 GOLDFINGER ET AL.: CASCADIA STRIKE-SLIP FAULTING GOLDFINGER El' AL.: CASCADIA STRIKE-SLIP FAULTING -125.98 -125.96 ß':.:.•.•: :-:-5 •-....:. •:...::::::::::::::::::::::::::: .• ..•:•::•::•. -125.94 8225 -125.92 I I Kilometers 1 0 SLUMP SCAR 47.22 47.2O -125.28 -125.26 -125.24 -125.22 47.O8 47.O6 b Figure 7. Two SeaMARC 1A 5-km swathsalong the North Nitinat fault showing the variable deformation styles along this structure. (a) The third and fourth thrust ridges landward of the deformationfront. The westernridge bendssharplyleft and is offset by individual E-W left-lateral faults. The easternridge is also offset along a larger scarpsubparallelto the main trend of the NNF, shownby white dashes. (b) The easternfault tip, showingen echelonfracturesat lower left, and a northerntracethatwe interpretascontrollinga headwarderodingchannel.Overalltrendof fault zoneis shownby white dashes. we calculate a sliprateof 5.5 + 2 mrn/yrfor theNorthNitinat continental slope was variable. We observed accretionary fault. Air gun records from the University of Washington surveydid not penetrateto oceanicbasement;thus we do not know if the NNF or SNF offset the basaltic oceanic crust. We traced the NNF southeastward across the continental slopewith sidescan sonarusinga 5-kin swath. Evidenceof faultingwas observedover a total lengthof 115 km from the westerntip of the NNF to the upper continentalslope. The expression of the NNF in the fold-thrust belt of the wedge folds acrossthe lower and middle continentalslope that were offset left-laterally, as well as sharplybent, and in several casesoffset left-laterally by possibleR shears(black arrows in Figure 7a). Subduedlinear traceswere observedin the intervening synclinal basins. We also observedregions of multiple parallel scarps,en echelon fault strands(Figure 7b), and sigmoidal bends in accretionary wedge thrust anticlines. Determinationof net slip on the continentalslope 8226 GOLDFINGER ET AL.: CASCADIA STRIKE-SLIP FAULTING portion of the fault is problematic,as we know of no datable piercing points. Based on the observation that net slip We have mappedtwo major active reversefaults on the SW Washington shelf that appear to control the location of the reaches a maximum Willapa Basin depocenter(Figure 2) [Cranswickand Piper, 1992] and Willapa Bay onshore (L. McNeill et al., Listric normal faulting of the northern Oregon and Washington continental margin, Cascadia subductionzone, submittedto Journalof GeophysicalResearch,1996). Older structures bend sharplyseawardjust westof the mappednorthernfault (Figure 2). We investigated the northern of these E-W to WNW trending structureswith the AMS 150 sidescansystem in 1994, suspectingthat this structuremight be related to the SNF. We observedlinear WNW trendingbackscatter patterns in the sedimentbut found no surface faulting in the thick at a fault's center, we infer that the maximum horizontal slip is larger than the 2.2-km slip observednear the seawardfault tip. The NNF is bestexpressed in the accretionary wedge thrust ridges and is subdued or absent in the intervening basins. This morphology is also characteristicof the South Nitinat, Willapa Canyon, Wecoma, and Alvin Canyon faults and, to a lesser extent, the Daisy Bank fault. However, the strike-slipfaults in southernOregon are well imaged even in the synclinal basins. We infer that the principal reason for this difference in surficial expression is probably related to a different origin for the southern Oregon faults, discussedin a subsequentsection. Second,the low vertical separation across the faults results in little expression of faulting in the broad flat-bottomed basins in northern Oregon and Washington as comparedwith the steep continental slope of southernOregon. We speculatethat the latitudinal difference in surfaceexpressionmay also be due to the relatively high sedimentation rates in northern Oregon and Washington as compared with southern Oregon [e.g., Sternberg, 1986; Barnard, 1978], which would tend to bury or subdue fault traces in the synclinal basins. The greater sediment supply is suggestedby the greater thickness of abyssalplain sediment cover due to the Astoria, Nitinat and Willapa submarine fans: 2.5-2.8 s two-way travel time off northern Oregon and central Washington, 2.0 s off southern Oregon. South Nitinat Fault The South Nitinat fault (SNF, Figure 2) intersects the deformation front at 47ø05.3'N, 19 km south of the North Nitinat fault. The SNF strikes 283ø, intersectingthe base of slope at a 7.7-km left step in the deformation front, and extends19 km seawardof the deformationfront on the abyssal plain. The base of slope channel offset by the North Nitinat fault also crosses the trace of the SNF adjacent to the accretionarywedge, where the seawardflank of the first thrust ridge forms its easternbank. The channel bends sharply left as it crossesthe fault, suggestiveof left-lateral offset, but the channel bathymetry is so subdued that offsetting relationships could not be clearly distinguished in the sidescandata. In a N-S seismicprofile (UW TT-063, line 32) the SNF appearsvirtually identical to the NNF, with a downto-the-south vertical separationand thickened growth strata on the downthrown block. Eastward thickening prefaulting abyssal plain units thin abruptly from north to south across the fault, indicating left-lateral slip on the SNF. By restoring fault motion, we obtain a net slip for the SNF of 2.0 _+0.8 km. The restoration reflects greater uncertainty than the NNF restorationdue to the somewhatpoorer seismic record for the SNF. The depth in the Nitinat fan section at which growth strata appear on the downthrownside of the fault is 302 m; thus, using the Nitinat Fan sedimentationrate of 100 cm/1000 years, the age of the fault is - 300 _+40 ka. From the age and net left-lateral separationwe calculate a slip rate of 6.7 _+3 mm/yr for the South Nitinat fault. We obtainedsidescanand Hydrosweep bathymetric data for the seaward40 km of the fault and attempted, unsuccessfully,to survey the landward part of the fault based on evidencefrom GLORIA regional sidescandata. GLORIA-sidescanimagery suggeststhat this fault may extend70 km fartherlandward. Holocene sediments. Willapa Canyon Fault The Willapa Canyonfault (WCF) intersectsthe deformation front at 46ø18'N, 8 km south of the outlet of Willapa submarine canyon onto the abyssal plain (Figure 2). This fault cuts the accretionarywedge but doesnot extendinto the abyssalplain as a detectable surface rupture. The seaward projection of the WCF is crossed 4.7 km west of the deformation front by University of Washington reflection line TT 79-1, which showsno disruptionof the abyssalplain section,confirming the lack of lower plate involvementwest of the plate boundary. On the accretionarywedge,the Willapa Canyonfault is poorly expressedover much of its lengthbut is well expressednear its center,on the mid continentalslope, where it offsetsleft-laterally a northwesttrendingchannelby approximately900 m. The overall strike of the fault is 280ø. No age information is available for the channel; thus we are unableto determinethe age or slip rate for the WCF. Wecoma Fault We surveyedthe entire length of the Wecoma fault using SeaMARC 1A with a 5-km swathwidth. We investigatedthe continental slope and outer shelf sectionsof the fault, as well as the abyssal plain portion from a different illumination angle than the 1989 survey. The new data supportearlier interpretationsof the continuity of the Wecoma fault as an active structureacrossthe slope [Goldfinger et al., 1992b; 1996a], and we were able to trace this fault to its apparent eastern tip over a total length of 95 km (Figure 2). The Wecoma, Daisy Bank, and Alvin Canyon faults each offset the abyssalplain sedimentarysectionand the underlyingbasaltic oceanic crust (Figure 3) [Goldfinger et al., 1992b, 1996a; Appelgateet al., 1992; MacKay, 1995]. Near the deformation front, all three fault zones have two main strandsdipping steeplytowardeach otherthat definepopupstructures (Figure 3d). The Wecomafault stepsto the right on the abyssalplain, forming a restraining step. A doubly plunging anticline between the two fault strands appearsto be causedby this restrainingstep(left end of Figure 8). This upwarpis coredby oceaniccrust, basedon the reflectiondata and modelingof a magneticprofile over the upwarp [Appelgateet al., 1992; Goldfingeret al., 1992b, 1996a]. Goldfingeret al. [1992b, 1996a] estimatethe ageof the Wecomafaultto be 650 + 50 ka andthenetleft-lateralslipto be 5.5 + 0.8 km. Fromthe ageandnetleft-lateralseparation, they calculatea slip rate of 8.5 + 2 mrn/yrfor the Wecoma fault. Offset of a dated channel 5 km west of the deformation ............................. ......:.: ............................. ...•.:•;:'.• :. ,..:... ,..:... &, ... ,..:... ,..:... •:. ,..:..... ::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::;; L,', ::::::::::::::::::::::::::::::::::::::::::::::::::: ½iii*;•: .. • :::::::::::::::::::::::::::::::::: ;:('i:i!',.; ;;;"-.;. ,..:....: \ ,iii.:: .................. ,"•i:;!ii! .............. t.' i;' --',................... .................. • ....... $ .'. :;:?;! ............................ ('- .-',• . :: -..• ,..t.. ,..t.. ,..t...,.'., =========================================== to - -'-,........... L• &' &'" ' • ============================================ "--"'================================================ ................... •, •, •:::::: ........................... . ::::::::::::::::::::::::::::::::::::::::::::::::::::::: .;':.': ::::::::::::: ..:::.."::: :::::::::::::::::::::::::::: :::::::::::::::::::::::::::::::::::::::: ::::::::::::::::::::::::::::::::::::::: ::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::: !iiiiiiii;i!:'ii!ii!i!ii ::::::::-::::: :::; ....... ,• ii•!i•..*..,..t...•,:. .... .: .-..-..,.,..t.. ,..t.. ,;,.I:"' :::::::: ::::."." ...s. •....-.-. ....... ....*-..:..-..:..4.-. '.......... • .•:..;: :.:.:.: ..... -- • •1•o ' ;---'-,-'-,-'-,•. ','• .......... ::•i•i! .......................... (,•ß• :.:..... &,•:.•, •...$... ,2 ..... ..t,..:.....½-....-. ... ,..:... ,..:....-. •. •" ----.;,..:..., ....... o ....................... *'........ '•' '-'-• -*. • ';: • ": ';: •' *-: '--'i::::•4i';: ';: ß •o *' ' - 'q:!•'"' r.,,* .........................:..... E• --'';:'.%•.;.,,*-' • ..c: ' ":'" '' ...:......... : ............ ;; ';.:;;; ?..-½•?;;; .,,,:....•,•,,½¾. . '-:.i•ii:::,:'•-* ' ..... ... &, (•1 •,•, *** :;iii!? :'" '• %7 ........ ' :'*" • '' .... '""'.: ..•::•i' ,t,"_;,t."'""'....-. '.' ..... , ... '-;:'.'.' '.,.::'•" a•:. ;" ;;; , &, , ...• . . . ...:::..::. ,;... ..' -' '" . . '..$ ..... .. ß . .... "' :, ..' "' "' •,.,,.: •:'' ,.., ." ' ..... ,',..' ,•, . ..., ß .:..':½• ,. '"'•.,_,,,,• ". "/ ß ,.•f•,7½( -"':.'-;..4:i .'•'-"',:::-,,.,;":•. •, *................. ;;..::;i ........ ""•'"•••••••••••.:.'..•,..:.,:.,•%':$/•.,.:"•':'""' -"'"'"""'"' ".-".•' : :. . :... ß -.. . '-.."..'-"-"•""".•".-."!•:-i.-'.":-'.:-•.-'.:.:.. ß,..-"" • '. ' -"--':a•. ß - ':-'•".-'•.-::.*..'.-f*:*'•4....,"•;.."•--...-"•'.-.:-. -.'s:.'-":'""--'-'-"-'•.,,•..-'•.•4-.'•. ' ' •:i:ii• .'>'"'""'"'"' '"":':"' '•'*-;•'"•"•:•:•:•.......'-..-;:'""'"'""'"""":""'•" '"'";•:••••!;'.-. '.--i.--'...ii.., ................ •••4•: '"'""'"'•' .::, .... -. ....:,...-. ............. •;•:..•;. .•/'.;.!.-'*..---"-". ....... :; ...: ,;;?,?,i(,,..'" L..'"..... ':*;½;'- "' • '/ LD ":ii ß ' ii' ß ' .."':'-" :.i: ........ .... .;.'',,. ,- ,: ... ... -"'. '%)::i,.:. '" ....... ..... .... ' .'z Z• ' .. ß, ::-::•....:s.. .. :.;..:;.: .•,, '..\ '!i•:.. .''':"!ill;''. '$i. :•!•.:.:.. ... ?::. : . " . ... -•: ' ' '. ":;•; ,-, ß.•........ ::--:.::-!.' , .,.•..:*,.",.,'t,' "",-.•?-,.w .•'-•.--,".•". :•,.,• ß ß , 'i!:•ii;ß ',,'/'.'.•',"•' ß . '%':.;;,.,'%,' '.:½., ;,.--. o "7 . >'. : 8228 GOLDFINGER ET AL.: CASCADIA STRIKE-SLIP FAULTING front allowedindependentcalculationof a slip rate of 8.5 _+4 mm/yr for the Wecoma fault [Appelgate et al., 1992; Goldfingeret al., 1992b, 1996a]. Figure8 showsa SeaMARC 1A 5-km sidescanswath over the Wecoma fault on the abyssal plain and lower continental slope. The surficial fault trace splitsinto two faults defining a popupstructurejust seawardof the fault and associated positive flower structure, which abruptly truncate an adjacent syncline (Figure 10). OSU sparker line SP-54 shows a similar truncationby the highangle fault, and a possiblesecondvertical strand2.2 km to the north (Figure 11; 8 km easton the E-W profile). the deformation records terminated at 44ø54.8'N, 124ø35.93'W. front. Both strands extend across the frontal two thrust anticlines, beyond which the northern strand may die out. Between the two strands,the frontal thrust vergence reverses to seaward from the regional landward vergence direction. Multichannel reflection data do not clearly image thesefaults in the complexaccretionarywedge. Nevertheless, the sonarimagery showsshearzonesexposedat the crestsof thesetwo ridges,as well as bendsand offsetsof the anticlinal axes(Figure8). The secondthrustridgeis offsetby a seriesof NE-SW trendingleft-lateralen echelonfaultsbetweenthe two strandsof the Wecoma fault, and the secondanticlinal ridge bendssharplyto the left as it crossesthe fault (Figure 8). The third ridge terminates at the fault. Extensive methane-rich fluid venting was observedalong the southernstrandon the frontal thrust anticline [Tobin et al., 1993]. Bedding attitudes taken from the submersibleAlvin show that dips are rotated into parallelismwith the Wecoma fault within this samearea, which is on the seaward limb of the seawardvergent frontal thrust(Figure 8) [Tobinet al., 1993]. The Wecoma fault, like the Daisy Bank, Alvin Canyon, and the three Washingtonfaults, is subduedin surfaceexpression in the lower slope basins but prominent on the anticlinal ridges and on the middle and upper continental slope. SeaMARC 1A imagery showsthat the Wecoma fault consists of multiple surficial fault tracesin a zone of deformationthat increasesin width landward. Young folds on the lower slope are commonly deflected or offset to the left at the fault and show marked increased elevation near the fault. On the middle to upper slope, several prominent folds bend sharply to an orientation parallel to the Wecoma fault from the regional NNW fold trend. These and several other folds parallel the Wecoma fault for a distanceof 15 km (Figure 9). The area shown in Figure 9 appears to be either a large right (compressional)step or complex deformation between two strandsof the Wecoma fault. The right side of Figure 9 is near the eastern tip of the Wecoma fault, and the sonar image shows the fault bending toward the northeastin a probable horse-tail splay. Fold axes in the southeastand northwest parts of Figure 9 are offset left-laterally. Additionally, leftsteppingen echelon folds along the fault indicate the overall shear senseof the Wecoma fault is sinistral [e.g., Wilcox et al., 1973; Harding and Lowell, 1979; Sylvester, 1988], in agreementwith the observedseparations.The entire Wecoma fault zone lies in a broad parallel swale visible in SeaBeam bathymetry. A proprietary seismic reflection profile shows The surface trace of the Wecoma fault on the sidescan This area of the outermostshelf was covered with ripples and sandwaves in unconsolidatedsand, confirmed by two Delta submersible dives in 1992 and 1993. The unconsolidated and mobile sedimentmade locatingthe easternfault tip problematic. The Wecoma fault appearsto terminate near the inferred western limit of the Siletzia terrane shown in Figure 2. The terrane boundary is inferred on the basis of the strong magnetic signature of the underlying Siletz block and by velocity analysisof a wide-angle reflection profile just to the southof the Wecoma fault [Trdhu et al., 1995]. SeaBeambathymetry data and numerousreflection profiles show this boundaryto mark a profoundchangein structuralstyle acrossthe forearc. The rapidly deforming fold-thrust belt of the active accretionarywedgegivesway to gentleopenfolds landwardof the Siletzia boundary. SeaBeamdata reveal a seriesof short NW trending en echelon folds just seawardof the boundary, suggestiveof right-lateral shear. Snavely [1987] infers that the dextral Fulmar fault controls the location of the seaward edgeof the Siletziablock, and Trdhuet al. [1995] suggestthat a fault observedoverlying the terrane boundary may be the active Fulmar fault. Our data also show a fault near the western edge of Siletzia (eastern fault in Figure 11), although the SeaMARC 1A datashowthis fault to be NW trending.Thuswe are uncertain whether this structure is related to the $iletzia boundaryor is the easterntip of the Wecoma fault. Daisy Bank Fault We surveyedthe entire lengthof the Daisy Bank fault (DBF; fault B of Goldfingeret al. [1992b] using the SeaMARC 1A system on the abyssal plain and slope and the AMS 150 system on the upper slope and shelf. Figure 3 shows the Daisy Bank fault immediately west of the deformationfront. The DBF extends 21 km seaward of the deformation front onto the abyssalplain where surfaceand subsurfaceexpressiondie out. The main fault trace intersectsa 150-m-highridge along the boundarybetweenthe landwardvergentthrustrampandthe fault. MCS lines 37 (Figure 3b) and 19 (not shown)show this ridge to be a southwestvergentthrust ridge boundedby the DBF on its southern flank. The main strand of the DBF stepsto the right at the westernend of this anticlinalridgeand continuesto the northwest.We interpretthe fold as a pressure ridge developed between the two overlapping fault strands similar to the fold alongthe Wecomafault [Goldfingeret al., 1992a; 1996b]. Basement reflectors show two offsets, one Figure 8. (top) SeaMARC 1A sidescanimage of the Wecoma fault zone on the abyssalplain and crossingthe frontal and secondaccretionarywedge anticlines. (bottom)Interpretation. The singlestrandedfault bifurcates,defininga popupstructurejust seawardof the frontalthrust. Both fault strands crossboth ridgesand the interveningbasin,althoughthey are faint in the basin. The secondridge is offsetby a seriesof possiblesyntheticleft-lateralfaults. The secondridgebendssharplyat the southern fault strand,forming a steep,highly reflectivescarpat right. The third ridge (not shown)terminatesat the southernfault strand. Strike and dip symbolsin and north of western shear zone from Alvin observations [Tobin et al., 1993]. GOLDFINGER ET AL.: CASCADIA STRIKE-SLIPFAULTING 8229 8230 GOLDFINGER ET AL.: CASCADIA STRIKE-SLIP FAULTING Figure 10. Line drawingof Shell Oil Companysparker line 7380 on the middle continental slope off central Oregon. Locationis shownon Figure 9. The Wecoma fault truncatesa NNW trendingsyncline.We interpretthis structure as a transpressionaloblique-slippingfault segment thatis accommodating a regionalrightstepof the Using SeaMARC 1A sidescanimagery,we tracedthe Daisy Bank fault zone acrossthe lower continentalslope (Figure 12). As with the other northern Oregon and Washington faults, the fault morphologyis subduedon the lower slope relative to the upper slope or abyssal plain. The DBF is characterizedby discontinuous fault tracesthat disruptthrust anticlinesand, to a lesserdegree,the interveningbasins. One 3 to 4-km long strandterminatesat the foot of a thrustridge, producing gullies and a prominent slump scarp. Farther seaward,severalsplaysof the DBF truncatethe frontal thrust anticline of the accretionary wedge. Reversals of vertical separationoccur along this ridge, with several tens of meters of relief evident along the main splay. The DBF crossesthe boundarybetweenthe Juande Fuca and North Americanplates in a 1 km-wide fault zone that appears to have localized slumpingof the seawardlimb of the frontal thrust. Daisy Bank, on the upper continental slope, is one of several uplifted Neogene structuralhighs off Oregon [Kulm Wecoma and Fowler, 1974]. The DBF bounds the southernflank of Vertical exaggeration = 3.3:1 WECOMAFAULT fault zone. Daisy Bank; a second less prominent strand of the fault boundsthe northernflank. SeaMARC 1A sidescanimagery up-to-the-north andoneup-to-the-south, thatdefinea "popup" and multichannel (MCS) and single channel (SCS) seismic reflection data show that the Daisy Bank fault is a wide of thebasement acrosstheDBF on MCS line 37 (Figure3b). The intersection of the DBF and the accretionary wedge structuralzone, within which Daisy Bank is upliftedas a horst marks an abrupt transition in structural domains between between two strands of the main fault. The main fault zone is seawardvergentthrusts(seawarddirectedthrusting)southward 5-6 km wide northwestof Daisy Bank, wideningaroundthe to 42ø10'N and landwardvergentthrusts(landwarddirected oblong bank, then narrowing to a single strand to the thrusting) northward to BarkleyCanyonat 48ø12'N[MacKay southeast. The traces of the fault strands are straight, et al., 1992; MacKay, 1995; Goldfinger et al., 1996b]. implying a near verticalfault. Probabledrag foldsof exposed Goldfinger et al. [1996b] interpret a progressivevergence strata, with a left-lateral sense of motion, are visible in reversalfrom southto north along a 15-km distancealong the sidescanimagerysoutheast of the bank (Figure 12). Mapping margin south of the DBF. The 1989 MCS data suggest from seismicreflectionprofilesindicatesleft-lateraloffsetsof progressiveundercuttingof an originally landwardvergent NNW trending accretionarywedge fold axes at Daisy Bank ridge by a younger seawardvergent thrust. The vergence (Figure 12). Scarp heights measured from the Delta reversal that appears to be localized by the DBF has submersiblerange from tens of centimetersto 47 m. The net apparently progressedfrom south to north, suggesting uplift of the southernflank of Daisy Bank by both foldingand faulting is about 130 m. From Delta, we traced one of the passageof the DBF beneaththe margin. We determined the geometry of pre-faulting eastward Daisy Bank scarps into an area of low relief and mud thickeningabyssalplain sedimentwedgesusing MCS lines deposition. We observeda fresh scarpstriking290ø across 05-08 for the Daisy Bank fault. Retrodeformationof the the unconsolidated Holocene mud, which is visible in the abyssalplain sectionyielded a net left-lateral separationof AMS 150-kHz sidescanimages(Figure 12). Stratigraphic 2.2 +_0.5 km. The depthto the oldestgrowthstratais 360 m, relationships indicate post12-kamotionon thissegment of determined from MCS line 37. From the sedimentation rate of 94 cm/1000 years,we obtainan age of 380 + 50 ka for the DBF. The age andnet slip valuesyield a slip rateof 5.7 + 2 mm/yr for the Daisy Bank fault. WEST 0.4- A_ wecøma fa theDaisyBankfault[Goldfinger et al., 1996b]. The mapped eastern tip of theDaisyBankfaultoverlies the Siletziaterraneboundary as doestheWecomafault,thuswe inferthattheDaisyBankfaultprobably terminates against EAST ult - 0.4 0.5- 0.6- 0.6 _• 0.7- 0.7•, 0.8- 0.8 Figure 11. OSU single-channel sparkerprofileSP-54. A prominent strandof the Wecomafault juxtaposes ananticline andsyncline at left. Thesecond subvertical faultat rightis probably alsoa strike-slipstructure andmayeitherbe thedyingnorthernstrandof theWecomafaultor theFulmarfault. SeeFigure 16 for locationandtext for discussion. GOLDFINGER ETAL.:CASCADIA STRIKE-SLIP FAULTING o o 8231 8232 GOLDFINGERET AL.: CASCADIA STRIKE-SLIPFAULTING the Siletziaterrane,givinga mappedlengthof 94 km. Further detailsof the DaisyBankfault are givenby Goldfingeret al. [1996b]. Alvin Canyon Fault The Alvin Canyonfault (ACF), namedfor its proximityto a series of Alvin dive sites, is similar in structuralstyle to the DaisyBankfault. Its structural detailsarenot aswell known, since it has not been surveyedwith sidescansonaron the continental slope. The fault was mapped from 1989 SeaMARC 1A sidescandata on the abyssalplain and from seismicreflectionprofiles and SeaBeamswathbathymetric dataon the continentalslope.The Alvin Canyonfault extends approximately 7 km seawardof the deformation front,based on sidescandata and two reflectionprofile crossingson the abyssal plain. The intersectionof the fault with the deformationfront is markedby a 3.7-km left stepin the basal thrust and initial ridge of the accretionarywedge. Linear patternsof authigeniccarbonates mappedon the lowerslope extend landward from the intersection of the ACF and the base of slope,parallelto the fault trace [Carsonet al., 1994]. Carson et al. [1994] used reprocessedGLORIA sidescan imagerywith the topographic effectsof reflectivitysubtracted from the image to identify zonesof carbonateprecipitation. The highly reflectivecarbonate-bearing sedimentsare the productof fluid ventingof methane-rich porefluidsthathave been sampleddirectlyby submersible in the immediatearea [Kulrn and Suess,1990] and are commonlyassociated with fault zoneselsewhereon the slope [Kulrn and Suess,1990; Tobin et al., 1993; Goldfingeret al., 1996b]. We retrodeformedprefaultingstratalwedgesusingMCS lines 01-05 to define the geometryof prefaultingunits. We calculatea net 2.2 + 0.5 km of left-lateral separationfor the Alvin Canyonfault. We estimatea sedimentation rate of 79 cm/1000yearsfor the Astoriafan sectionat the Alvin Canyon fault. A depth of 310 m for the oldestgrowth stratawas determinedfrom MCS line 37 usingthe depthin two-waytime and the sedimentation rate. Using the sedimentation rate and depth values, we obtain an age of 380 + 50 ka for the initiation of motion on the fault. These values yield a slip rate of 6.2 + 2 mm/yr for the Alvin Canyonfault. Like the Daisy Bank fault, the Alvin Canyonfault widens from a singlestrandon the abyssalplain to a 6-km-widefault zoneon the upperslope. The Alvin Canyonfault hasstrong morphological similarities to thebettersurveyed Wecomaand DaisyBankfaults. An unnamed submarine bankis upliftedas a horst between two strands of the fault, similar to the structuralsettingof Daisy Bank. Like Daisy Bank, this horst hasbeentruncatedby erosionduringPleistocene low sealevel stands.On the abyssalplain, Holoceneseafloorsediments are offsetby the Alvin Canyonfault, suggesting that it is likely to be an active structureon the slope, but this cannot be confirmed without further investigation. Heceta Bank Structure and Heceta South Fault The Heceta Bank structure is centered at approximately uplifted bank. Swathbathymetricdata from a NOAA shallow water system (BSSS) revealed that Heceta Bank has at least one submergedPleistocene lowstand shoreline flinging its western and northwesternflanks (C. Goldfinger, unpublished data, 1993). A smooth wave-cut platform seaward of the former shorelineis clearly visible for a length of 45 km in the bathymetric data, as are former beach berms and probable subaerialdrainagefeatures. The submergedshorelineis tilted to the south, deepeningfrom 115 m at the northernend of HecetaBank to 225 m adjacentto the HecetaBank structure.A possiblecontinuationof the submergedshorelineand wavecut platform southof the fault suggests> 200 m of vertical separationacrossthis structure. However, a Klein 50-kHz sonar survey in 1992 failed to identify any offset surficial featuresfrom which to characterizethe horizontal separation (if any) acrossthis structure. Sidescandata from this survey and subsequentdives with the submersibleDelta identified areas of tabular carbonatedepositionindicative of methanebearingfluid venting at the top of the scarp,but this evidence was subduedin comparisonto other active faults. The scarp itself had a relatively low slope, and no evidence of surface rupturewas found in severalobliquetraverseswith Delta, nor on the sidescan imagery. An unmigrated multichannel reflection profile (USGS line WO 77-05) shows only weak evidence for faulting and suggestsa principally monoclinal structure. The present surface morphology may in part be erosional accentuation of the monocline by wave action during Pleistocenelow stands. We concludethat either the Heceta Bank structureis a monocline(presumablyoverlyinga more deeply buried fault) or a strike-slipfault in which lateral motion is not resolved on the reflection profile. The Heceta South fault lies 15 km west of the Heceta Bank fault and was originally mappedas being the same structure [Goldfinger et al., 1992a]. Subsequentinterpretationof SeaMARC 1A sidescanand swathbathymetrydata showsthat they are two separatestructures.The Heceta Southfault is 35 km in length and is composedof multiple segmentsstriking 2930-325ø (Figure 13). The scarpof the seawardfault segment forms the northern rim of a slump scar on the lower continentalslope (Figure 13). This slump,approximately7.5 km in diameter, involvedapproximately 14 km3 of accreted material from the accretionmywedge. Of three known slumps of this magnitude on the Oregon and Washington margins, two occur at the intersectionof WNW trending strike-slip faults with the deformation front (North Nitinat fault, Figure 5; Heceta South fault, Figure 13). The Heceta South fault appearsto extend several kilometers into the abyssal plain, suggestedby the linear truncationof debris from the slump along the projection of the fault, although no seismic reflection data crossingthe possibleabyssalplain extension of the fault are available to confirm this. Alternatively, the linear trendsin the slump debrismay have originatedalong the fault prior to slumpingand were then translatedonto the plain duringthe slumpevent. Sidescanandbathymetricdata revealed no piercing points from which to determine horizontalseparationfor this fault. 43ø56'Non the continentalslope(Figure2), strikes308ø, and Coos Basin Fault hasa minimumlengthof 16-20 km (Figure13). This structure The Coos Basin fault intersectsthe base of slopeat a 1.2bounds the southeasternflank of Heceta Bank, a complex frontat 44ø04'N(Figure2). anticlinorium that has undergoneup to 1000 m of post- km left stepin the deformation Mioceneuplift [Kulrnand Fowler, 1974]. The HecetaBank We observedno evidencefor abyssalplainruptureseawardof structureis a strikinglinearfeaturethat abruptlytruncates the the intersection of the fault and the deformation front in GOLDFINGER ET AL.:CASCADIASTRIKE-SLIP FAULTING o 8233 8234 GOLDFINGER ET AL.: CASCADIA STRIKE-SLIP FAULTING sidescanimages. Near-surfacefaulting approximatelyalong the seaward extension of both the Coos Basin fault and Thompson Ridge faults is visible on two-channel seismic reflection records on the abyssal plain [EEZ-SCAN 84 Scientific Staff, 1986]. These faults could not be tied to the observedupperplate faults,nor are their strikesknown from the single seismic profile. The surficial expressionof the Coos Basin fault is an approximately 5-km-wide zone of deformation, the major elements of which are visible in GLORIA sidescanimagery. Two main scarpsoffsetanticlinal ridgeson the lower slopeleft-laterally,and the broadzone is expressedbathymetricallyas linear trendsstriking2880-293ø. Two measured left-lateral offsets of surficial and 158 m are visible in a 10-km-wide collected in 1993. features of 125 SeaMARC The two well-defined 1A mosaic strands of the fault diverge southeastwardinto a broader structuralzone which loses definition about 35 km landward of the base of the continental slope. Thompson Ridge Fault The Thompson Ridge fault intersects the base of the continentalslopeat a 6 km left stepin the deformationfront at latitude 43ø16.5'N (Figure 2), and, like the Coos Basin fault, does not extend into the abyssalplain as a surficial featurebasedon SeaMARC 1A imageryandSeaBeamdata. The ThompsonRidge fault is the best expressedbathymetrically of the nine Cascadiastrike-slipfaults mappedto date. The main scarpat the deformationfront along the left step is 650 m in height, north block up. The fault zone is clearly observedin SeaBeambathymetryas a patternof disruptionof accretionarywedge thrust ridges,and like many of the other faults, its component strands diverge slightly to the southeast.Figure 14 showsa perspectiveshaded-reliefimage of NOAA SeaBeamswath bathymetrydata in the Thompson Ridge fault area, gridded to a 100-m point spacing. In map view, a patternof left stepsand sigmoidalbendingand offset of crossingfolds indicatesleft-lateral shear. The SeaBeam bathymetry suggeststhree left stepping anticlinoria (Figure section and the basement are offset by these structures [Goldfinger et al., 1992b; 1996a; Appelgate et al., 1992; MacKay, 1995]. These three faults dip steeply to the northeast and bound the southern flanks of asymmetrical basementpop ups (Figure 3d). Magnetic modeling of the Wecoma fault zone indicates that the basement is both offset and significantlyupwarpedbeneatha pressureridge anticline formed at a right step in the fault trace [Goldfinger et al., 1992b; 1996a; Appelgate et al., 1992]. The basementpopups clearly show involvement of the basementin the transverse deformation of the Cascadia forearc, and indicate that these three faults are slightly transpressional.The reflection data for the Washington faults did not reach basement;however, their striking similarity to the Oregonfaults suggestssimilar origins and thus probable basementinvolvement. Although the 1989 reflection data clearly resolvebasement offseton the threeOregonfaults,we are unableto discriminate between throughgoingrupture of oceanic crust or a more superficialdetachmentof upper crustalblocks such as may have emplaced basaltic blocks in the Franciscansubduction complex in northern California [e.g., Kirnura et al., 1996]. However,high pore fluid pressureand very low wedgetaperin northern Oregon and Washington, discussedfurther below, suggestthat the lower continentalslopeis poorly coupledto the subductingplate. The basementoffsetsobservedseaward of the deformation front, therefore, are unlikely to be the result of detachment of basaltic blocks as a direct result of interactionbetweenthe accretionarywedgeand the slab. We note a consistentlongitudinal pattern of deformation along the faults that extend seawardof the base of slope: strong expressionon the abyssal plain, poor expressionon the lower slope,and strongexpressionon the upperslopeand outermost shelf. For several reasons, we infer that this pattern is most consistentwith faults that originate in the slab and propagateup throughthe upper plate. Pore fluid pressurein accreted sedimentsnear the plate boundary is known to be high. Negative polarity reflections in d6collementzones at the toe of the Oregon slope and a subhorizontalmaximum compressivestresssuggesthighly 14), indicatingthe overallshearsenseof theThompson Ridge overpressured dilatant zones at the plate boundary[Moore et fault is sinistral. Other related folds parallel the strike of the al., 1995a]. Fluid pressures in the range of 0.9-0.95 fault zone. Crossing thrust ridges step to the left and are lithostatic are estimatedfor the Barbadosaccretionarywedge elevated at the fault zone, a distinctive morphology also based on recent logging from drilling resultsfrom ODP Leg observed at the Wecoma and Daisy Bank faults. This 156 [Moore et al., 1995b]. During the 1993 SeaMARC 1A elevation difference, in the case of the Wecoma fault, is the survey we observedmud volcanoeson the lower slope and result of a compressional flower structurecomposedof fault splayswith a small reversecomponentof motion basedon plain off Oregon and Washington, implying lithostatic fluid pressureswithin the accreted section. The lower slope off seismic reflection records. We infer that much of the bathymetricexpressionof the strike-slipfaults observedon northernOregon and Washingtonhas a very low wedge taper the slopeis due to the superimposition of flower structures on and widely spacedlandwardvergentfolds,furtherindicationof the coevalor olderthrustridges.Equipmentdamageprevented a very weak d6collementdue to high porefluid pressurein this the acquisitionof all but a shortsegmentof sidescandataover region [Seely, 1977; Davis et al., 1983; MacKay, 1995]. We speculate that the longitudinal pattern of morphologic the ThompsonRidge fault. expressionalong the transversefaults can be explained by reduced transmission of lower plate fault slip across the Discussion overpressuredand poorly coupled d6collementbeneath the lower slope. Rejuvenationof the faults on the upper slope is Basement Involvement consistent with progressive dewatering of the wedge, Five of the strike-slip faults (North and South Nitinat, resultingin better interplatecouplingas would be expectedfor Wecoma, Daisy Bank, and Alvin Canyon faults) crossthe the rearwardpart of the wedge. The landwardwidening fault plateboundary andareobserved in bothupperandlowerplates zones(Figure 2) are also consistentwith fault slip transmitted (Figure2). Reflectiondatafor thethreeOregonfaultssuggest upward through the eastwardthickeningaccretionarywedge. that seawardof the deformation front,the entiresedimentary The slip distribution on the Wecoma fault, determined by GOLDFINGER ET AL.: CASCADIA STRIKE-SLIP FAULTING 8235 ':.;!•.?....., .. ß ... ß..•?::•. .... •.:::...: :::::::: ====================== 8236 GOLDFINGER E-'I'AL.: CASCADIA STRIKE-SLIP FAULTING GOLDFINGER ET AL.:CASCADIASTRIKE-SLIP FAULTING isopachplots of two abyssalplain stratigraphicunits, shows that slip increaseslandwardalong the fault from its western tip to the frontal thrust (Figure 15), landward of which isopachingwas not possible[Goldfinger et al., 1996a]. The observeddecreasein morphologicalexpressionacrossthe plate boundary,where slip is still increasingin lower plate units, is most consistentwith the fault propagatingupward acrossthe poorly coupled d6collement.Alternatively, older episodesof slip may be preservedat the landwardends of the faults but may be overprintedby compressionaldeformation 8237 seaward geometry of the Cascadia subduction zone off Washington. This geometryis inferred to resultin archingof the slab beneath Washington, with a more steeply dipping slab both to the north and to the south [Michaelson and Weaver, 1986; Crosson and Owens, 1987]. This slab observedin SeaBeambathymetry,may be sucholder traces. What mechanisms mightbe responsible for transverse rupture of the subductingslab? Three modelsthat mightexplainslab rupture are interplate coupling stresses,membranestrain due to slab geometry, and slab stressesimposedby slab-mantle configuration results in strike-parallel membrane strain as modeled by Creager et al. [1995, Figure 3] for the Bolivian syntaxis. Subductionof the JDF slab is expectedto result in similar N-S compression,and this model is supportedby sparsefocal mechanisms[Spence, 1989]. A strike-slip focal mechanism indicating right-lateral slip on a NW striking plane or left slip on a NE striking plane is located 30 km seawardof the westerntip of the Daisy Bank fault (Figure 2) [Spence, 1989]. Either nodal plane for this event is consistentwith N-S compression;however, N-S compression in the slab is inconsistentwith the nearbyWNW striking leftlateral faults. These faults shouldbe dextralif they are driven by N-S compression. The evidencefor N-S compressionjust seaward of the left-lateral faults, though sketchy, suggests that the principal horizontal stressdirection is significantly different seawardof the plate boundarythan in the subducting slab and suggeststo us that membranestraindue to subduction geometry or other mechanismsmay be operative but that it must be overprintedby someother mechanismthat drives the interaction. sinistral on the lower slope. We infer that the significant deformation of the accretionarywedge is the expressionof faulting originating in the subductingplate, transmittedto a passivelydeforming upper plate. The passageof the basementfaults beneaththe accretionarywedgeshouldleaveolderfault tracesthat may be progressively overprinted by compressional deformation (Figure15). The four faultswithoutlowerplateexpression, as well as other transverse folds and linear features These models are discussed below. we have transverse faults. Recently, Scholz and Campos[1995] proposeda dynamic model of interplate coupling and decoupling at subduction Origin of the Transverse Faults zones that incorporates the hydrodynamic resistanceof the Giventhe evidencefor deformation of bothplates,in which motion of the slab through the viscous mantle. In their plate did the deformationoriginate? Wang et al. [1995] model, the mantle is considered stationary in a hot spot calculate that interplate stresses<10 MPa result from JDF referenceframe. This "sea anchor"force, in a mantle system platesubduction, consistent with low observed stressdropsin that is fixed to the hot spotframe, shouldhavetrench-parallel subductionearthquakes,assumingthat stressreleaseis close and trench-normal components in the case of oblique to complete [e.g., Kanarnori, 1980; Magee and Zoback, subduction[Scholzand Campos,Figure 1] (Figure 15). The 1993]. High heatflow in Cascadiaplacesthe downdipbrittle trench-parallelcomponentof hydrodynamicmantle resistance ductiletransitionfor crustalrocksin the outer-shelf-upperto subductionis an unbalancedforce that setsup a shearcouple slope region [Hyndrnan and Wang, 1995]. Between the in the plane of the slab. For Cascadia,the shearcouplewould poorly coupled accretionarywedge and the shallow brittle be dextral, and antithetic shears should be left-lateral and eastductiletransitionmay be a very narrowzoneof significant west to northwesttrending. We proposethat this shearcouple interplatefrictional stress(C. Goldfinger,manuscriptin may be responsiblefor the observedtransverseruptureof the slab we have observed in the central Cascadia forearc. Limited preparation, 1996). We estimate that this narrow zone of couplingis unlikely to be responsiblefor ruptureof the slab seismicity is consistent with a model of transverse oceaniclithosphere,which has a shearstrength,for any shearingin the subductingslab. The largest instrumentally conditionsof temperature and porepressurewithin the brittle recordedearthquakein Cascadia(1949, mb= 7.1; 47.13øN, regime, greater [e.g., Brace and Kohlstedt, 1980] than the 122.95øW [Baker and Langston,1987]) occurredin the Puget Soundregion in Washington(Figure 2) at a depthof 54 km in interplatecouplingstresses inferredby Wanget al. [1995]. Another possible scenariofor intraslab deformation could the upper part of the subductingplate (Figure 2) [Baker and be the introduction of membrane strain due to the concave Langston, 1987; Ludwin et al., 1991]. The focal mechanism Figure 15. Blockrotationmodelfor thecentralCascadia forearc.SeaBeam bathymetry shaded from the north. The WecomaandDaisyBank faultsare shown,with the DaisyBankfault exposedin foreground.Well-mapped faulttracesarein solid;discontinuous tracesaredashed.The arc-parallel component of obliquesubduction createsa dextralshearcouple,whichis accommodated by WNW trendingleft-lateralstrike-slip faults.We propose thatshearing of theslabdueto obliquesubduction is responsible for thefaultsinvolvingoceaniccrust. "Seaanchor"forceandcomponents shownat lower right[Scholz andCampos, 1995].WF, Wecomafault;DBF,DaisyBankfault;FF,Fulmarfault; "pr" pressureridge;"DB", Daisy Bank. "OT?",possibleold left-lateralfault strand. Arrow headsandtails showstrike-slip motion.Whitearrowsat western endof Wecoma faultshoweastward increasing slip calculatedfrom isopachoffsets[Goldfingeret al., 1996a]. Dots at easternend of Wecomafault indicate the locationof the faultsshownon Figure 11. Schematicblockrotationmodelshownat lower left, showingdeformation of parallelogram ABCD in dextralsimpleshear[afterWellsandCoe, 1985]. 8238 GOLDFINGER ET AL.: CASCADIA A •k I C R and Thrust Extension R' Shears Faults Fractures Folds All structures Superimposed STRIKE-SLIP FAULTING is strike-slip with a preferredfault plane striking east-west_+ 15ø (azimuth 255-285 ø) and has nearly pure left-lateral slip [Baker and Langston, 1987]. While this strike-slip mechanismindicatessheafing within the Juan de Fuca plate, the fault planes of the 1949 event are not parallel to the convergencedirection, negating a possible tear fault origin [Weaver and Baker, 1988]. A similar left-lateral strike-slip solution was determined for a small 1981 earthquake (magnitude 3.3) in the upper Juan de Fuca plate on the southernWashingtoncoast (Figure 2) [Taber and Smith, 1985; Weaver and Baker, 1988]. The seismicity plotted on Figure 2 reveals several other interesting patterns. The Washingtoncoast strike-slip event lies at the easternend of an east-westlinear trend of earthquakesthat lie roughlyon the projectionof the SouthNitinat fault, and are near a major eastwest reversefault mappedon the shelf. Two eventsplot near the Wecomafault, and severalplot near the landwardend of the Alvin Canyon and Willapa Canyon faults. The depthsand locations of all of these events, however, are suspectdue to poor azimuthal station coverage (R. Ludwin, personal communication, 1995). Could the basement involved faults be reactivated structures in the oceanic basalt? Comparison of the strikes of the oblique faults to the seafloor magnetic anomaly map of Wilson [1993] shows that the three Washington faults, with D strikes of about 283 ø, could be reactivated small fracture zones C A perpendicularto the magnetic anomalies.The Oregon faults strike 292 ø to 325ø, deviating from an orthogonal to the magneticlineationsby 12ø to 45ø and thus are unlikely to be related to preexisting basement structure. We have also consideredconjugate shorteningof the forearc of the type describedby Lewiset al. [1988]; however,extensivemapping of margin structures[Wagner et al., 1986; Clarke, 1990; Goldfinger et al., 1992a; C. Goldfinger and L. McNeill, manuscriptin preparation,1996] has not identified structures conjugateto the WNW faults. We have mappedtwo majorNE trendingstructures,one on the slopeoff northernOregonand one off centralWashington,but both of theseare clearly leftlateral tear faults of the accretionaryfold-thrustbelt. Kinematic T1 T1 Model: Upper Plate Deformation Northeasterlydirectedsubductionresultsin a dextral shear couple in the North American plate and, as suggestedabove, possiblyin the Juande Fuca plate as well (Figures15 and 16). Figure 16. Strike-slipmodels,convergence shownat 062ø for centralOregonusingpoleof DeMetset al. [1990]. (a) Simple shearmodel showingR, R', and P shears.(b) Structuretypesandorientations expectedin overallright simpleshear[after Sylvester, 1988]. (c) Block rotation modelin a dextralshearcouple[afterWellsand Coe,1985] (d) Map viewinteraction of basement strike-slip faultwith a growingaccretionary wedge. Shadingindicateslower plate; no shadingindicatesupperplate. (top) Time 1, (bottom)time 2. Remanenttraces,showndashed, may remain as active structuresor be overprintedby compressional deformation.As the basementfault moves, theaccretionary wedgeadvances, maintaining theyouthof the intersection point. GOLDFINGER ET AL.: CASCADIA The existenceof active sinistralfaults in the upper plate, both related and unrelated to basementfaulting, suggeststhat the North American plate may be respondingto both passive strain transmittedfrom the slab and to dextral shear imparted STRIKE-SLIP FAULTING 8239 0 ZZ ZZZ by interplate coupling. In the following discussionwe presenta model for deformationof the upperplate, which is observable,but a similar model may also be operativein the subductingslab. Pure shear deformation of the Cascadia margin is well expressed in the fold andthrustbelt of the accretionary wedge. For the most part, these structuresare subparallel to the margin and representthe responseof the upper plate to the normal component of plate convergence. In model experiments of simple shear and in earthquake ground ruptures,five setsof fractureshave been observed:R and P or synthetic shears, with the same motion sense as the simple shear couple; R' or antithetic shears, with a motion sense opposite to the main shear; tension fractures or normal faults oriented at about 45 ø to the main shear zone; and Y shears or faults parallel to the shear couple. Strike-slip, reverse, and normal faults of the orientations shown in Figure 16b are consistentwith a simple shear model of overall right-oblique shear [e.g., Wilcox et al., 1973; Sylvester, 1988, and referencestherein], althoughnone of the experimentalmodels involved subduction-drivensimple shear. Comparisonof the right simple shearmodel (Figure 16) with the structuralmap of the Oregon-Washingtonmargin in Figure 2 suggeststhat the WNW left-lateral faults are consistent with R' shears antithetic to the right shearcoupledriven by oblique subduction.WNW to NW trending folds and thrust faults of the middle continental slope to outer continentalshelf are also expected in the right simple shearmodel. Y shears,or faults parallel to the main shear couple, would be difficult to detect within the similarly oriented structural grain of the accretionarywedge. We note that the three best mappedfault zonesthat crossthe deformation front, the North Nitinat, Wecoma, and Daisy Bank faults, bend 3-6 ø southward on the continental slope from their abyssal plain strike (Table 1). The south bending of these two fault zones could be due either to passageof the basement faults beneath the wedge (Figure 16d) or a componentof dextral arc-parallelsheardistributedacrossthe accretionarywedge, similar to that suggestedfor the onshore forearcby England and Wells (1991). In the case of the forearc rotation we propose, the shear couple is set up by oblique insertionof the subductedplate into the mantle, and thus two margin-parallel faults may not be required. Linkage betweenthe upperand lower plate faults for anything but a very short time requires slip along the inboardedge of the rotatedterranein the upperplate. Trdhu et al. (1995) suggestthat a small fault observedto overlie the edge of the Siletzia terranein central Oregon may representan active dextral fault that decouplesthe active wedge from the inboard Siletzia terrane. Our data are somewhatsupportiveof this hypothesis. We observe a vertical, probably strike-slip fault (easternfault in Figure 11) that also overlies the edge of the Siletzia terrane (see Figure 2 for location). SeaMARC 1A data show this fault to be NW trending, and we initially correlated -H -H c•. -H -H -H c•. c•. it as the northern strand of the Wecoma fault. Alternatively,it could be the samefault shownby Trdhuet al. [1995], 7 km to the south. SeaBeambathymetryrevealsthat the edge of the Siletzia terraneis overlainby a north trending zone of short, doubly plungingen echelonfolds suggestiveof •g• •o..• << 8240 GOLDFINGER ET AL.: CASCADIA STRIKE-SLIPFAULTING dextral shear at depth (shown on SeaBeambathymetryin Figure 15) A geometricrequirementof a set of parallelstrike-slip faults with the same motion sense is that the intervening blocks and the faults themselvesmust rotate (Figure 16) [e.g., Freund, 1974]. We infer that the Cascadialeft-slip faults segmentmuch of the continental slope in Oregon and Washingtoninto elongateblocks rotatingclockwiseabout verticalaxes. This styleof deformationhasbeenproposed for the Aleutianforearcon a muchlargerscale[Geistet al., 1988; Ryan and Scholl, 1993] and has been documented in other tectonicenvironments[e.g., Cowan et al., 1986;Beck, 1989; Garfunkel, 1989; Jackson and Molnar, 1990]. Although Cascadialacks seismicity-definedblocks, paleomagnetically determined clockwise rotation of basalts in western Oregon and Washingtonindicatesthat clockwiserotationhasoccurred throughoutmostof the Tertiary. A smoothcoastwardincrease in the rotation of the Miocene Gingko and Pomonamembers of the CRBG is consistent with a model of distributed deformation of the forearc and inconsistentwith microplate docking models [England and Wells, 1991]. The viscous sheetdeformationmodelof Englandand Wells [ 1991] implies a northwardtransportof the forearcat 13-15 mm/yr sincethe Miocene, or about 75% of the presentmargin-parallelplate convergence. Such northward transport of the forearc is consistentwith our simple shearmodel. A block rotation model implies geometricspaceproblems at the marginsof the rotatedblocks. If the pivotsare fixed to the North Americanplate, compression betweenthe blocksis required. Such compression is observedalongthe Wecoma, Daisy Bank, Alvin Canyon, and ThompsonRidge faults off Oregon but not along the Washington faults. Without compression alongthe block edges,the blocksmusttranslate northwardrelative to the North American plate. Small gaps and overlapsoccurat the endsof the blocks(Figure 16). On the abyssalplain, compression is taken up by anticlinesand horsetailsplaysnear the westerntips of the Wecoma,Daisy Bank, and Alvin Canyonfault and by splaysat the North and South Nitinat faults. At the landward block ends, the complexity of the accretionarywedge is such that minor deformation such as this would be difficult to discern. small offsets of the frontal thrust and the relatively straight trends of the fault zones across the slope suggestthat the accretionarywedge and subductingplate are not converging along the expectedJDF-NOAM plate vector. The forearcmay deforming internally and/or translating northward due to oblique subduction[Pezzopaneand Weldon,1991; Wells and Weaver, 1992; McCaffrey and Goldfinger, 1995]. Alternatively, subductionhas ceased or dramatically slowed within the last 0.6 Ma. The latter hypothesisconflicts with abundantevidence for modem subductionincluding Holocene shorteningof the accretionarywedge apparentin numerous seismic reflection records throughoutthe Cascadia margin; onshoregeodetic data indicating shorteningapproximatelyin the convergence direction [Savage and Lisowski, 1991; Dragerr et al., 1994]; and paleoseismologicevidenceof large, probably subduction-related earthquakesin the coastalbays of Oregon, Washington, and California (e.g., Atwater, 1987, 1992; Darienzo and Peterson, 1990; Nelson et al., 1995). The continuity of the oblique strike-slip faults acrossthe plate boundaryand the evidencethat subductionhas occurred through the Holocene to the present suggest to us that the submarine forearc is translating northward, either as a microplate or in small block translations and or rotations. The extensiveinternal deformationof the forearc supportsthe latter hypothesis. We find supportfor northwardtranslation in the active structuresof offshore Washington. Structural trends of late Quaternary folds and thrust faults off western Washington are nearly north-southon the continentalshelf (Figure 2) [Wagneret al., 1986; C. Goldfingerand L. McNeill, manuscript in preparation, 1996], and are parallel to the margin (~ 020ø) on the continentalslope. Contemporaryfold and thrust trends in the Juan de Fuca Strait, between Washingtonand VancouverIsland, strike NW, parallelto the strait. Active shorteningis apparently occurring acrossthe strait, in contrast to the east-west shorteningoccurring west of Washington, suggestingthat the Washington forearc is colliding with a more rigid Vancouver Island buttress, as suggestedby Wells and Weaver [1992]. Goldfinger [1994] and McCaffrey and Goldfinger [1995] calculated the total rate of forearc deformation in the arc- parallel direction as a result of slip on the nine strike-slip The fact that five of the nine mappedfaults crossthe plate faults (Table 1). A componentof extensionbetween points boundarymust be reconciledwith northeasterlysubductionat on any two blocks occursas a result of left-lateral motion on 40 mm/yr. The JDF plate shouldhavetraveled10 to 24 km to the intervening strike-slip fault. Slip on all nine faults the northeastrelative to the North American plate during the translatesthe northernend of the rotateddomainnorthwardby 0.3 to 0.6 Ma elapsed since the initiation of left-lateral the sum of these arc-parallel components relative to the faulting (Table 1). Thus we might expect to observe southernend, assumingno shorteningacrossthe blocks. The horizontaloffsetsof the strike-slipfaults where they crossthe net arc-parallel extensionrate is 10.5 mm/yr from the five deformation front, but we have observed few such offsets. We faults with known slip rates. If we assignthe lowest known suggestthat the interactionbetweenthe obliquefaultsand the slip rate value, 5.5 mm/yr, to the faults with unknownslip growing accretionarywedge tends to reducethe amountof rates, the extension rate is 17.4 mm/yr, or 87% of the 20 lateraloffsetexpectedat the deformationfront (Figure16d). If mm/yr tangential componentof convergence[Goldfinger, the deformation front was fixed in east-westposition during 1994; McCaffrey and Goldfinger, 1995]. If correct, this the known life span of the faults, the expectednorthward implies that the oblique faults alone can accountfor most of offset across the plate boundary due to northeasterlyplate the oblique componentof subduction. Forearc deformation motionwouldbe 4-11 kin. Goldfingeret al. [1996b] estimate rates have recently been linked to subduction earthquake that the deformationfront advanced14-21 km westwardduring magnitude[McCaffrey, 1993]. Using the observeddifference thisperiodbasedon the ageof uplift of thrustridges[Kulmet between subductionearthquakeslip vectorsand plate motions al., 1973a]. The rapid advance of the deformationfront to estimate total rates of forearc deformation, McCaffrey implies that the point of intersectionbetweenthe strike-slip [ 1993] finds a negativecorrelationbetweenrapidly deforming faults and the frontal thrustis always very young,reducingthe forearcsand the largestsubductionearthquakes.The physical expectedoffset of the frontal thrust. Some possibleminor explanationfor this is that earthquakemagnitudeis ultimately offsets may be interpretedin Figure 5. Nevertheless,the linked to the ability of the forearc to store elastic strain GOLDFINGER ET AL.: CASCADIA STRIKE-SLIP FAULTING 8241 energy; thus weak forearcs generate smaller earthquakes shorteningand arc-parallelstrike-slipand translation. The [McCaffrey, 1993]. McCaffrey and Goldfinger [1995] high slip ratesof the strike-slipfaults,coupledwith the lack calculate that the extensivestrike-slipfaulting in the central of offset of these faults as they cross the plate boundary, Cascadia forearc may limit subductionearthquakesto about imply that the seawardaccretionarywedgeis not movingat MW 8.0. the expectedconvergence raterelativeto the subducting plate. Last, the process of block rotation of the Cascadia submarineforearc may have occurredfor a longer period of time than the agesof the datedfaults. We speculatethat new oblique faults periodically rupture the lower plate, and sometimesboth plates, then slip for a relatively short period of time, perhapstaking advantageof basementweaknesses. Motion on individual faults may ceaseafter a short period of time, after which the abyssal plain fault traces would be subducted or accreted, depending on the local vergence direction and dtcollement position. Accreted upper plate faults may remain as active deformation zones or may be altered by mass wasting, deposition,erosion, or subsequent accretionary wedge faulting. In SeaBeam bathymetry and GLORIA sidescandata we have observedmany poorly defined WNW trending lineations that may be the remnants of previous episodes of strike-slip faulting, one of which is shown on Figure 15. Alternatively, if the faults are no older than the datedgrowthstrataon the abyssalplain, the observed deformationcould be very young, reducingthe effect of JDF motion on the time history of upper plate deformation. Conclusions Using sidescan sonar, seismic reflection profiles, and swath bathymetric data, we have mapped a set of WNW trending left-lateral strike-slip faults that deform the Oregon and Washingtonsubmarineforearc. Evidence for left-lateral separation includes offset of accretionary wedge folds, channels,and other surficial features; sigmoidal left bending of accretionary wedge folds; and offset of abyssal plain sedimentary units. Five of these faults cross the plate boundary, extending 5-21 km into the Juan de Fuca plate. Using offset of subsurface piercing points and offset of approximately dated submarinechannels, we calculate slip rates for these five faults of 5.5 to 8.5 mm/yr. Little or no offset of these faults by the basal thrust of the accretionary wedge is observed. Holocene offset of submarinechannels and unconsolidated sediments is observed in sidescan records and directly by submersible. The strike-slip faults are most likely driven by dextral shearingof the subductingslab and propagateupwardthrough the overlying accretionarywedge. Tangential hydrodynamic drag causedby obliqueinsertionof the slab into the mantleis a possibledriving mechanism. Four sinistralfaults observed in only the upper plate may be reinanenttracesof previous basement-driven deformation. Alternatively, a similar, though unrelated dextral shear couple driven by interplate couplingmay drive thesefaults and may augmentdeformation of the upper plate for all the sinistralfaults. A model of overall right-lateral simple shear of the submarine forearc is consistent with the observed surface We conclude that the accretionary wedge is rotating and translatingnorthward,driven by the tangentialcomponentof Juande Fuca-North American plate convergence. Acknowledgments. We thank the crews of the research vesselsThomas Thompson(University of Washington), and supportvesselsCavalier andJolly Roger; Delta pilots Rich and Dave Slater,Chris Ijames, and Don Tondrow;membersof the Scientific Party on cruises from 1992 to 1993 during whichmos•of thedatawerecollected; KevinRedman,David Wilson, Tim McGiness,Wolf Krieger, Chris Center, and Kirk O'Donnell from Williamson and Associates of Seattle Washington, our sidescancontractors;Ed Llewellyn for cowriting Sonar, OSU's sidescan-sonar processingsoftware;and Chris Fox, and Steve Mutula of NOAA for their assistancewith the multibeam data. Multibeam bathymetry data were collectedby NOAA andprocessed by the NOAA PacificMarine and EnvironmentalLaboratory,Newport Oregon. Thanks to Chris Fox and Bob Dziak, also of NOAA Newport, Oregonfor the T wave earthquakelocationsshownin Figure2. We thank Eric Geist, Roger Bilham, and Greg Moore for thoroughand helpful reviews. This researchwas supportedby National ScienceFoundationgrantsOCE-8812731 and OCE-9216880; U.S. Geological Survey National Earthquake Hazards Reduction Program awards 14-08-0001-G1800, 1434-93-G2319, and 1434-93-G-2489; and the NOAA Undersea Research Programat the West CoastNational UnderseaResearchat the Universityof Alaska grantsUAF-92-0061 and UAF-93-0035. References Appelgate,T. B., C. Goldfinger,L. D. Kulm, M. MacKay, C. G. Fox, R. W. Embley,andP. J. Meis, A left-lateralstrike-slipfault seawardof the central Oregon convergentmargin, Tectonics,11, 465-477, 1992. Atwater, B. F., Evidencefor great Holoceneearthquakesalong the outercoastof WashingtonState, Science,236, 942-944, 1987. Atwater,B. F., Geologicevidencefor earthquakes duringthe past2000 years along the Copalis River, southerncoastalWashington, J. Geophys.Res.,97, 1901-1919,1992. Baker, G. E., and C. A. Langston,Sourceparametersof the 1949 magnitude7.1 South Puget Sound,Washington,earthquakeas determined from long-period body waves and strong ground motions,Bull. Seisrnol.Soc.Am., 77, 1530-1557, 1987. Barnard, W. D., The Washington continental slope: Quaternary tectonicsandsedimentation, Mar. Geol.,27, 79-114, 1978. Beck, M. E., Jr., On the mechanismof tectonictransportin zonesof obliquesubduction,Tectonophysics, 93, 1-11, 1983. Beck,M. E. Jr., Block rotationsin continental crust:examplesfrom North America, in Paleomagnetic Rotations and Continental Deformation, edited by C. Kissel, and C. Laj, pp. 1-16, Kluwer Acad. Norwell, Mass., 1989. Bilham,R., andP. Bodin,Faultzoneconnectivity: sliprateson faultsin the San FranciscoBay area, California, Science, 258, 281-284, 1992. faults, which may be R' or antithetic shears to the overall right-shear couple. The major strike-slip faults define elongate blocks that, becauseof their orientation and sinistral slip direction, must rotate clockwise. We infer that the Blake, M. C., D.C. Engebretson,A. S. Jayko, and D. L. Jones, Tectonostratigraphic terranes in southwest Oregon, in Tectonostratigraphic Terranesof the Circum-PacificRegion,edited deformation of the submarine forearc (defined to include the Brace,W. F., andD. L. Kohlstedt, Limitsonlithospheric stress imposed lower plate) is highly strain-partitioned into arc-normal by D. G. Howell,pp.147-157, Circum-Pac. Counc. forEnergyand Miner. Resour.,Houston,Tex., 1985. by laboratory experiments, J. Geophys. Res.,85, 6248-6252,1980. 8242 GOLDFINGER ET AL.: CASCADIA STRIKE-SLIP FAULTING Carson,B., E. Seke,V. Paskevich,and M. L. Holmes,Fluid expulsion sites on the Cascadia accretionaryprism: Mapping diagenetic depositswith processedGLORIA imagery, J. Geophys.Res., 99, 11,959-11,969, 1994. Clarke, S. H., Jr., Map showing geologic structuresof the northern California continentalmargin, scale 1:250,000, U.S. Geol. Surv. Map, MF-2130, 1990. Cowan, D. S., M. Botros, and H. P. Johnson, Bookshelf tectonics: Rotatedcrustalblockswithin the SovancoFractureZone, Geophys. Res. Lett., 13, 995-998, 1986. Cranswick, D. J., and K. A. Piper, Geologic framework of the Washington-Oregon continental shelf-Preliminary findings, in Proceedingsof the 1991 ExclusiveEconomicZone Symposium on Mapping and Research:WorkingTogetherin the Pacitic EEZ, U.S. Geol. Surv. Circ. 1092, p. 146-151, 1992. Creager, K. C., L. Y. Chiao, J.P. Winchester, and E. R. Engdahl, Membrane strain rates in the subducting plate beneath South America, Geophys.Res.Lett., 22, 2321-2324, 1995. Crosson, R. S., and T. J. Owens, Slab geometry of the Cascadia subductionzone beneathWashingtonfrom earthquakehypocenters andteleseismic converted waves,Geqphys. Res.Lett.,14, 824-827, 1987. Darienzo, M. E., and C. D. Peterson,Episodictectonicsubsidenceof late Holocene salt marshes, northern Oregon central Cascadia margin, Tectonics,9, 1-22, 1990. Davis, D., J. Suppe,and F. A. Dahlen, Mechanicsof fold-and-thrust belts and accretionarywedges, J. Geophys.Res., 88, 1153-1172, Oregon,in Subduction:Top to Bottom,Geophys.Monogr.Ser., vol. 96, editedby G.E. Beboutet al., pp.65-74,AGU, Washington, D.C., 1996b. Griggs, G. B., and L. D. Kulm, Origin and developmentof Cascadia deep-seachannel,J. Geophys.Res.,78, 6325-6339, 1973. Harding,T. P., and J. D. Lowell, Structuralstyles,their plate-tectonic habitats,and hydrocarbon trapsin petroleumprovinces,AAPG Bull., 63, 1016-1058, 1979. • Hyndman, R. D., and E. E. Davis, A mechanismfor the formationof methane hydrate and seafloor bottom simulating reflectors by verticalfluid expulsion,J. Geophys.Res.,97, 7025-7041, 1992. Hyndman, R. D., and K. Wang, The rupture zone of Cascadiagreat earthquakesfrom currentdeformationand the thermal regime,J. Geophys,Res.,100, 22,133-22,154,1995. Jackson,J., and P. Molnar, Active faulting and block rotationsin the western Transverse Ranges, California, J. Geophys.Res., 95, 22,073-22,087, 1990. Jarrard,R. D., Terranemotionby strike-slipfaultingof forearcslivers, Geology,14, 780-783, 1986. Kanamori,H., Stateof stressin the Earth'slithosphere,in Proceedings of the 1979 EnricoFermi SummerSchool,Varenna,Italy, Physicsof the Earth's Interior, pp. 551-553, editedby A.M. DziewonskiandE. Boschi, North-Holland, New York, 1980. Karig, D. E., D. R. Sarewitz, and G. D. Haeck, Role of strike-slip faultingin the evolutionof allochthonous terranesin the Philippines, Geology,14, 852-855, 1986. Kelsey, H. M., and J. G. Bockheim,Coastallandscapeevolutionas a 1983. function of eustasy and surface uplift rate, Cascadia margin, southernOregon,Geol. Soc.Am. Bull., 106, 840-854, 1994. DeMets, C., R. G. Gordon, D. F. Argus, and S. Stein, Current plate Kimura, G., Oblique subductionand collision:Forearctectonicsof the motions,Geophys.J. lnt., 101,425-478, 1990. Kuril arc, Geology,14, 404-407, 1986. Dewey, J. F., and S. H. Lamb, Active tectonics of the Andes, Kimura, G., S. Maruyama, Y. Isozaki, and M. Terabayashi,WellTectonophysics, 205, 79-95, 1992. preserved underplating structure of the jadeitized Franciscan Dragert, H., R. D. Hyndman, G. C. Rogers, and K. Wang, Current complex,PachecoPass,California,Geology,24, 75-78, 1996. deformationand the width of the seismogenic zoneof the northern Kulm, L. D., and G. A. Fowler, Oregoncontinentalmarginstructure Cascadiasubductionthrust,J. Geophys.Res.,99, 653-658, 1994. EEZ-SCAN 84 Scientific Staff, Atlas of the exclusive economic zone, and stratigraphy: A test of the imbricate thrust model, in The western conterminous United States, scale 1:500,000, U.S. Geol. Geologyof ContinentalMargins, editedby C. A. Burke and C. L. Surv., Misc. Invest. Ser., I- 1972, 1986. Drake, pp. 261-284, Springer-Verlag,New York, 1974. England, P., and R. E. Wells, Neogenerotationsand quasicontinuous Kulm, L. D., and E. Suess,Relationof carbonatedepositsand fluid deformationof the Pacific northwestcontinentalmargin, Geology, venting, Oregon accretionaryprism, J. Geophys.Res., 95, 889919, 978-981, 1991. Fitch, T. J., Plate convergence,transcurrent faults, and internal deformationadjacentto southeastAsia and the westernPacific, J. Geophys.Res., 77, 4432-4460, 1972. Freund, R., Kinematics of transform and transcurrent faults, Tectonophysics, 21, 93-134, 1974. Garfunkel,Z., Regionaldeformationby block translationand rotation, in PaleomagneticRotationsand ContinentalDeformation,editedby C. Kisseland C. Laj, NATO ASI Ser. C, 254, 181-208, 1989. Geist, E. L., J. R. Childs, and D. W. Scholl; The origin of summit basinsof the Aleutian Ridge: Implicationsfor block rotationof the arc massif, Tectonics, 7, 327-341, 1988. Goldfinger, C., Active deformation of the Cascadia forearc: Implications for great earthquake potential in Oregon and Washington,Ph.D. thesis, Oreg. StateUniv., Corvallis,1994. Goldfinger,C., andL. C. McNeill, Casestudyof GIS dataintegration 8915, 1990. Kulm, L. D. et al., Site 174,Initial Rep.DeepSeaDrill. Proj., 18, 97167, 1973a. Kulm, L. D., R. A. Prince,and P. D. SnavelyJr., Site Surveyof the northernOregoncontinentalmarginand AstoriaFan, Initial Rep. Deep Sea Drill. Proj., 18, 979-987, 1973b. Lewis,S. D., J. W. Ladd,andT R. Bruns,Structuraldevelopment of an accretionaryprism by thrust and strike-slip faulting: Shumagin region, Aleutian Trench,Geol. Soc.Am. Bull., 100, 767-782, 1988. Ludwin, R. S., C. S. Weaver, and R. S. Crosson,Seismicity of Washingtonand Qregon,in Neotectonics of North America,DNAG CSMV-1, editedby D. B. Slemmons, E. R. Engdahl,D. Blackwell', and D. Schwartz, pp. 77-98, Geol. Soc. of Am., Boulder, Colo., 1991. MacKay, M. E., Structuralvariationand landwardvergenceat the toe of the Oregonaccretionary prism,Tectonics,14, 1309-1320,1995. andvisualization in submarine tectonicinvestigations: Cascadia MacKay, M. E., G. F. Moore, G. R. Cochrane,J. C. Moore, and L. D. Kulm, Landward vergenceand oblique structuraltrends in the subduction zone,Mar. Geod.in press,1996. Oregonmarginaccretionaryprism:Implicationsand effect on fluid Goldfinger, C., L. D. Kulm, and R. S. Yeats, Neotectonicmap of the flow, Earth Planet. Sci. Lett., 109, 477-491, 1992. Oregon continental margin and adjacent abyssal plain, scale Magee, M. E., and M.D. Zoback,Evidencefor a weak interplate 1:500,000, Open File Rep. 0-92-4, Oreg. Dep. of Geol. and Miner. Ind., Portland, 1992a. thrust fault along the northern Japan subductionzone and implications for themechanics of thrustfaultingandfluidexpulsion, Goldfinger, C., L. D. Kulm, R. S. Yeats, B. Appelgate,M. MacKay, Geology,21,809-812, 1993. and G. F. Moore, Transversestructuraltrends along the Oregon convergent margin:Implications for Cascadia earthquake potential, McCaffrey, R., Slip vectorsand stretchingof the Sumatranforearc, Geology,20, 141-144, 1992b. Geology,19, 881-884, 1991. Goldfinger,C., L. D. Kulm, R. S. Yeats,B. Appelgate,M. MacKay, McCaffrey,R., On the role of the upperplatein greatsubduction zone and G. R. Cochrane,Active strike-slipfaulting and folding of the earthquakes, J. Geophys,Res.,98, 11,953-11,966,1993. Cascadia plate boundary and forearc in central and northern McCaffrey, R., and C. Goldfinger, Forearc deformationand great Oregon, in Assessingand Reducing Earthquake Hazards in the earthquakes:Implications for Cascadia earthquakepotential, Pacific Northwest, edited by A.M. Rogers, T. J. Walsh, W. J. Science, 267, 856-859, 1995. Kockelman,and G. Priest,, U.S. Geol. Surv. Prof. Pap., 1560, p. Michaelson,C. A., and C. S. Weaver, Upper mantle structurefrom 223-256 1996a. teleseismicP wave arrivalsin Washingtonand northernOregon,J. Goldfinger, C., L. D. Kulm, R. S. Yeats, C. Hummon, G. J. Huftile, A. Geophys.Res.,91, 2077-2094, 1986. R. Niem, and L. C. McNeill, Oblique strike-slip faulting of the Moore,J. C., G. F. Moore,G. R. Cochrane,andH. J. Tobin,NegativeCascadiasubmarineforearc:The Daisy Bank fault zone off central polarityseismicreflectionsalongfaultsof the Oregonaccretionary GOLDFINGER ET AL.: CASCADIA STRIKE-SLIP FAULTING prism: Indicatorsof overpressuring, J. Geophys.Res.,100, 12,89512,906, 1995a. Moore, J. C. et al., Abnormalfluid pressures andfault-zonedilationin the Barbadosaccretionaryprism: Evidencefrom loggingwhile drilling,Geology,23, 605-608, 1995b. Nelson, A. R. et al., Radiocarbonevidencefor extensiveplateboundaryruptureabout 300 years ago at the Cascadiasubduction zone, Nature, 378, 371-374, 1995. 8243 associated with the subduction of the Juande Fucaplatebeneaththe Olympic Peninsula,Washington,Bull. Seismol.Soc.Ant., 75, 237249, 1985. Tobin, H. J., J. C. Moore, M. E. MacKay, D. L. Orange,and L. D. Kulin, Fluid flow alonga strike-slipfault at the toe of the Oregon accretionaryprism: Implicationsfor the geometryof frontal accretion, Geol. Soc.Am. Bull., 105, 569-582, 1993. Tr6hu,A., G. Lin, E. Maxwell,andC. Goldfinger,A seismicreflection Nelson,C. H., Marine geologyof the Astoria deep-seafan, Ph.D. profile across the Cascadia subduction zone offshore central dissertation, 287pp.,Oreg.StateUniv., Corvallis,1968. Oregon: New constraints on methane distribution and crustal Niem, A. R., N. S. MacLeod,J. Snavley,P. D., D. Huggins,J. D. structure,J. Geophys. Res.,100, 15,101-15,116,1995. Fortier,H. J. Meyer, A. Seeling,andW. A. Niem, Onshore-offshore Wagner, H. C., L. D. Batatian,T. M. Lambert,and J. H. Tomson, geologiccrosssection,northernOregonCoastRangeto continental Preliminarygeologicframeworkstudiesshowingbathymetry, slope, Oreg.Dep. Geol.Miner.Ind.,Spec.Pap.26, 10pp,1992. locationsof geophysical tracklines andexploratory wells,seafloor Pezzopane,S. K., and R. J. Weldon II, Tectonicrole of Holocenefault geologyand deepergeologicstructures,. magneticcontours,and activityin Oregon,Tectonics,12, 1140-1169, 1991. inferredthicknessof Tertiary rocks on the continentalshelf and Rohr,K. M., andK. P. Furlong,Ephemeral platetectonics at theQueen uppercontinental slopeof southwest Washington betweenlatitudes Charlottetriplejunction,Geology,23, 1035-1038,1995. 46øN and48ø 30'N and fromthe Washington coastto 125ø 30'W, Ryan,H. F., andD. W. Scholl,Geologicimplications of gmatinterplate OpenFile Rep.86-1, 10 pp.,Wash.Div. of Geol.andEarthResour., earthquakes alongthe Aleutianarc, J. Geophys.Res.,98, 22,135Olympia,1986. 22,146, 1993. Sangawa,A., A history of fault movementsince late Pliocenein the centralpartof southwest Japan,Bull. R. Soc.N. Z., 24, 75-85, 1986. Savage,J. C., andM. Lisowski,Strainmeasurements andpotentialfor a greatsubduction earthquakeoff the coastof Washington, Science, 252, 101-103, 1991. Wang,K., T. Mulder,G. C. Rogers, andR. D. Hyndman, Caseforvery low coupling stress ontheCascadia subduction fault,J. Geophys. Res., 100, 12,907-12,918, 1995. Weaver,C. S., andG. E. Baker,Geometry of theJuandeFucaplate beneathWashingtonand northernOregonfrom seismicity,Bull. Seismol.Soc.Am., 78, 264-275, 1988. Scholz,C., andJ. Campos,On the mechanism of seismicdecoupling Weaver,C. S., andS. M. Smith,Regionaltectonicandearthquake andbackarc spreadingat subduction zones,J. Geophys.Res,100, hazard implications of a crustal fault zone in southwestern 22,103-22,115, 1995. Scholz, C. H., N.H. Dawers, J.-Z. Yu, and M. H. Anders, Fault growthandfaultscalinglaws:Preliminary results,J. Geophys. Res., 98, 21,951-21,961, 1993. Seely, D. R., The significanceof landwardvergenceand oblique structuraltrendson trenchinnerslopes,in IslandArcs,Deep Sea Trenches andBack-ArcBasins,MauriceEwing$er.vol. 1, editedby M. TalwaniandW.C. Pitman,pp. 187-198,AGU, Washington, D.C., 1977. Shipboard Scientific Party,Site892,in G.K.westbrook, B. Carson, R. J. Musgrove,et al., Proc.OceanDrill. Program,Initial Rep.,part 1, 146,:CollegeStation,TX (OceanDrillingProgram), 234-300,1994. Snavely,P, D., Jr., Tertiary geologicframework,neotectonics, and petroleumpotentialof the Oregon-Washington continental margin, in Geologyand ResourcePotentialof the ContinentalMargin of WesternNorthAmericaandAdjacentOceanBasins-Beaufort Seato Washington,J. Geophys.Res.,88, 10,371-10,383,1983. Wells,R. E., andR. S. Coe,Paleomagnetism andgeologyof Eocene volcanic rocks of southwestWashington,implicationsfor mechanisms of tecti•nicrotation,J. Geophys.Res.,90, 1925-1947, 1985. Wells, R. E., and P. L. Heller, The relativecontribution of accretion, shear, and extension to Cenozoic tectonic rotation in the Pacific Northwest,Geol. Soc.Am. Bull., 100, 325-338, 1988. Wells, R. E., and C. S. Weaver,Rotatingcrustalblocksandbig earthquakes in westernOregonand Washington,Geol. Soc.Am, Abstr. Programs, •4, 89,1992. Wilcox,R. E., T. P. Harding,andD. R. Seely,Basicwrenchtectonics, Am. Assoc.Pet. Geol. Bull., 57, 74-60, 1973. Wilson, D. S., Confidence intervals for motion and feformation of the Juande FucaPlate, J. Geophys. Res.,98, 16,053-16,071,1993. Baja California, edited by D. W. Scholl, A. Grantz, and J. G. Vedder,pp. 305-335, Circum-Pac.Counc.for Energyand Miner. Resour., Houston, Tex., 1987, Spence,W., Stressoriginsand earthquakepotentialin Cascadia,J. Geophys.Res., 94, 3076-3088, 1989. Sternberg, R. W., Transport and accumulationof river-derived sedimenton the Washingtoncontinentalshell J. Geol.Soc.London, 143, 945-956, 1986. C. Goldfinger andL. D. Kulm,Collegeof Oceanic andAtmospheric Sciences,Oregon State University, Corvallis, OR 97331. (email: gold@oce.orst.edu; lkulm@oce.orst.edu) C. Hummon, L. McNeill, and R. S. Yeats, Departmentof Geosciences, OregonStateUniversity,Corvallis,OR 97331. (email: hummonc@ucs.orst.edu; Imcneill•'oce.orst.edu; yeatsr@bcc.orst.edu) Sylvester,A. G., Strike-slipfaults, Geol. Soc.Am. Bull., 100, 16661703, 1988. (ReceivedFebruary13, 1996;revisedAugust23, 1996; Taber, J. J., Jr., and S. W. Smith, Seismicityand focal mechanisms acceptedAugust28, 1996.)