Geoscience Canada. Volume 3. Nunber 4 269

in the area of Washington and Oregon seemed necessary geometrically. This suggestion was reiterated by Tobin and Sykes (1968) in their study of the seismicity ofthe regionanddevelopedin

established the presence and basic bathymetric trench at the foot of the Canada's Active configuration of an active spreading continental slope, the absence of a ocean ridge system ofl the west coast of clearly defined eastward dipping Beniofl Western Margin - Vancwver Island. The theory of the earthquake zone and the apparently geometrical mwement of rigid quiescent state of present volcanism The Case for lithospheric plates on a spherical earth have all contributed toward a (introducedas the 'paving stone' theory) widespreaa uncertalnfy as to whether was tirst ~ro~osedand tasted bv subddctlon 1s cbrrently taking place ~c~enziiand Parker (1967) using data (e.8.. Crosson. 1972: ~rivast&a,1973; R. P. Riddihough and R. D. Hyndnan from the N. Pacific and in the global Stacey. 1973). Past subduction, fran Victoria Geophysical Observatory, extension of the theory by Morgan several tens of m.y. ago until at leastone Earth Physics Branch, (1 968) this reglon again played a major or two m y. ago. IS more generally Department Energy, Mines and role Morgan alsoconcluded that a small accepted ana it seems to us that the Resources, block eitof theJuandeFuca ridge (the case for contemporary subduction, in a 5071 W. Saanich Rd. . Fig. 2) was moving large part, must beargued upona Victoria B. C. independently of the main . continuation into the present of the The possibility that theJuan de Fuca processes shown to be active in the Summary plate is now underthrustingNorth recent geological past. We review the evidence for subduction America was first suggested by In this article we anempt to treat at the continental margin of south- McKenzie and Parker (1967). They systematically each of themainsources western British Columbia and conclude noted that consumptionof of information on whether or not that, although the pattern is complex and changing, south of 50°N convergence has continued to the present day.

lntmduetlon Some of the questions that most commonly arise in studies of the recent tectonics of south-western British Columbia are 'when and where has subduction taken place along the continental margin?' and 'is subduction still going on today?'. Being actively involved in both onshore and onshore studies of this region, we felt that it would be usefulto review all the important data which bear on these questions. While admitting that our special interests have led us to emphasize particular aspects of the many fields of and geophysics relevant tothe problem, the exercise has nevertheless considerably clarified our own ideas. It will, we hope, be similarly useful to others whether or not they agree with our conclusions that subduction has occurred and is still occurring. The ocean floor off western Canada occupies a fundamental position in the ~~~~~~kanmalymap (black, .ogamma) development of the theory of plate from Ranand Mason (1961). Arrows show tectonics. The magnetic anomaly maps most recently anomaly. subduction is taking place. Firstly, the earthquake data extend only about 50 Vancwver Island. This is relative motions between the America. years into the past and representa very of the -ridge-trench type Pacific and Juan de Fuca plates overthe short, recent, time interval. and is not necessarily stable in the terms past I0my. based largely on sea-floor defined by McKenzie and Morgan magnetic anomalies; secondly, the Present Plate Conflgurallon (1969) and may move with time. The indications of subduction at the margin The olate boundaries of the NE. Pacific transform fault is the right lateral Queen such as sedimentary structures and dip reg& have been progressively Charlotte fault which runs northwest of the oceanic basement; thirdly, the established by many studies over the along the continental slope west of the effects of subduction to be expected on past 10 years, principally by marine Queen Charlotte Islands, meeting the the continent such as arc vulcanism, geophysical surveys (bathymetry, Aleutian trench in southeastern Alaska. heat flow and gravity; and finally the magnetics, reflection seismic and heat The ridge is composed of a complex evidence from seismicity. We feel flow) and earthquake locations in the sequence of en-echelon segments however that is is important to area off thewashington, Oregon and connected by fracture zones running emphasize that each source of British Columbia coast. The plate south from the Dellwood Knolls infomation has a different timescale. boundaries based on the most recent spreading centre, through the Explorer The offshore magnetic anomalies, for dataare shown in Figure2. Ridge to the . The example, can give information back Of critical importance forthe history of trench (although not a bathymetric several tens of m.y. but their time subduction is the location of the point of 'trench') or line of convergence, is near resolution is little better than one my.: intersection of three main plates (atriple the base of the continental slope off the junction) near the northern end of coast of and Washington. Not all of these plate boundaries are equally established and several are in doubt. One complication is the very recent 'ridge jumping' that has been suggested for the northern ends of both the Juan de Fuca and Explorer Ridges. For the former, reflection seismic studies COLUMBIA reported by Barr and Chase (1974) and by Davis and Lister (1976) have established that the present axis of spreading is located in the westerly of two spreading centres about 20 km Dellwood Knol apart. Both spreading centres are expressed by pronounced valleys and a positive magnetic anomaly. The eastern centre was probably active within the past 0.7 my. At the northern end of the a similar situation has been proposed by Srivastava ef a/. (1 971), on the basis of bathymetric, magnetic and heat flow data. A more radical possibility was suggested by Barr (1974), principally from the evidence of earthquake locations, that the Juan de Fuca - America boundary now extends northeast from the northern endof the Juan de Fuca Ridge and that the extends south- eastwards to the same position. This configuration would make the Dellwood Knolls and Explorer Ridge spreading segments inactive. While the absence of any detectable magnetic or bathymetric Figure 2 Dossible EDIJDboundarv Dotted lineis discontinuity corresponding with an Tectonic features. AP - America plale, n,pornetca norrnern eogeol s~w~cttw Bp - Brooks Peninsula, EP - Explorerplate, D,are Dasned..ne!nlano!swestern im rot active extension of the Queen Charlotte Gvb - Garibaldivolcanic belt, JP - Juan de volcanism. Triangles - Quaternary Fault makes Barr's suggestion unlikely, Fuca plate, PP - Pacific plale, PRh - Paul volcanoes, squares -Miocene volcanic the existenceof a diffuse line of Revere , SFz - Sovanco fracture centres. Arrows are probable piate morions epicentres trending north-east from the zone. Dashedlineeasfwards from SF2 is relative lo America plate. northern end of theJuan de Fuca Ridge WienceCanada Volume 3. Number 4

is confirmed in a recent earthquake geomagnetic reversal was presented by of between 0' and 20") From a gradual compilation (Fig. 3) and may be signifi- Vine (1968). Using a time scale derived change in anomaly direction and thus cant This is commented upon later. lrom a profile across the East Pacific ridge orientation, principally from four to Clearly, the area is of some Rise. Vine established a history going six m.y, ago, a progressive change in complexity. It was described by Vine back 15 m.y. The oldest dateable spreading direction has been implied. (1 966) as one in which there has been magnetic anomalies east of the Juan de Atwater (1 970), taking a right lateral 'faulting and a gradual stilling and re- Fuca rloge (now at tne margln and Pac~l~c-Amercanplate mollon deduce0 orientation ofthe ridge crest'. Me~rd oarallel to the Wasn~naton-Oreaon" - from theGulfot Cal~lorn~aand Juan oe and Alwater (1968) described the coast) were estimated to be about nine Fuca Ridge spreading parallel to the deformation in this offshore region as m.y. old. Asa complete record older than Blanco Fracture Zone at 5.8 cm yr-'. 'exceptionally complex both because of this liestothe west ofthe ridge System. deduced a present convergence at the a continent nearby and because of a subduction must have taken place continental margin of25cmyr-1 inaNE. relatively recent change in direction within the last nine m.y. direction. Silver (1 971 )subsequently of spreading'. A number of authors have considered the influence of a major NE. subsequently used Vine's analysis to transcurrent fault from the ridge to the Oflshon Magnetic Anomalicn and look at the predicted motions in more margin and showed that while it was Relatire PIale Motions detail. There is general agreement on active (210 0.5 m.y. ago) subduction may Through the dating of individual spreading rates but the absence of have been considerably less at the anomalies, magnetic suweys are the clecdrly defined transform faults has British Columbia margin (approximately most powerfultool available for madefhe accurate determination of one cm1yr)than tothe south (3 to 4 deducing past motions of thesea floor spreading directions more difficult. In cmlyr). Recent more detailed studies of away from spreading ridges. For the different studies spreading has been the magnetic anomaly geometry for the northeast Pacific, the first detailed assumed either to be perpendicular to swthern part of the Juan de Fuca plate combining of the magnetic anomaly the magnetic anomaly trend or parallel off Vancwver Island could have pattern with the chronology of to the Blanco fracture zone(a difference become virtually coupled to the America plate in the recent past.

Flgum 3 Cornpilafion of eanhquake epicenlres from Sel~molcgicalSeries. Earth Physrcs branch, EMR. Canada (see Milne el al.. 1976 MS). All of the above studies have worked to the Blanco fracture zone. It seems to although now close to 49"N. has moved from the southern edge of the Juan de US that these two approaches may north at approximately 45 km per million Fuca plateand considered that the nevenhdess eventually prove to be years. South of it the convergence rate Blanw fracturezone defines the reconcilable through the action of the at the margin has decreased frwn 5.5 direction of relative motion between the transcurrent faults mentioned above. cmlyrat ninemy. to3.6 cmlyr Juan de Fuca and Pacific plates. This One of us (Riddihough.1976) has (perpendicular subduction of 3.2 cmlyr) approach seems to result in particularly recently re-examined the fine structure in the past one million years, a result large uncertainties in relative plate of spreading rate variations on the Juan which is Similar to that of Chase etal. motions off British Columbia. Working de Fuca and Explorer ridges. Again (1975). However, to the north of it (the from the northern end of the Juan de taking spreading orthogonal to the ridge ) convergence at the Fuca ridge system and making the but using the more recent reversal margin has decreased to a greater critical assumption that spreading chronology of Talwani etal. (1971). this extent, from 5.5 cmlyrto about 2.0 directions were perpendicular to the examination suggests: 1)the Explorer cmlyr (perpendicular subduction rate of magneticanwnaly lineation.Chaseeta1. and Juan de Fuca plates (Fig. 2) have 1.4 cmlyr) in the same period (Fig. 4b). (1975) recently computed a moved independently at least since The movement Of the triple junction or convergence rate for theJuan deFuca - eight m.y, ago and 2) convergence northern limit of subduction off northen America plates of 5.2 cm yr-' at N45"E vectors at the continental margin have Vancouver can be also assessed from from 10 to f~em.y. decreasing to4.2 cm been variable both in space and time estimates of relative plate motions. yr-'at N43"E from five m.y, to the (Fig. 4a) The boundaly between the Atwater (1970) suggested that it has present. This is a much higher Explorer and Juan de Fuca plates and remained close to its present location convergence rate than that computed between the two appropriate overthe last tO m.y Chase etal. (1975) by Atwater (1970) and comes from convergence regimes has in general calculated that it should have moved taking spreading perpendicular to the terms been an eastward extension ofthe 15 km NW. from 10 lo five m.y. ago and magnetic anomalies rather than parallel Sovanco Fracture Zone (Fig. 2) which further65 krn from 5 to0 my. Theslow

(a)

10 9 I 1 8 3 4 3 2 1 0 I Juan deI Fuca America I / //A/>//1 r r 1 / America rr r t

10 1,9 8 1 I 5 4 3 2 1 # m years be. -3 o-3 cmr per year (b)

Juan de Fuco - America Explorer - America

1 ."-.-.. I 1 1

*I 5 a m 5 I m. pan b.p. .

Flgum 4 interactions Subduction rates perpendicular Convergence veclors forJuan de to margin are shown in lower graphs (from Fuca/America and Explorer/Amer~ca Rddihough, 1976 MS). Geoscience Canada Volume 3. Nunber 4 273

northward movement is supported by trench deposit because its gecmary is from deformation of the slope seismic reflection interpretatiwls of controlled by the structural trench, and sediments, that there is strong evidence sedimentary structures such as tnOSeOf considers Me older sequences to have for convergence continuing up to the Tln~nelar (1972) whlchsuggestthatthe been deposited on flat lying seafloor gwlogical 'present' south of 50°N. triple junction was stable near Brooks further offshore that was later However, as the time resolution of Peninsula (Fig. 2) at least until 1hreem.y. transported eastward and down-warped sedimentary studies, particularly in the ago. A detailed study of themagnetic toward the trench axis. light of extremely high post-glacial anomalies (Riddihough. 1976 MS) using The continental slope itself changes sedimentation rates, may be only about the best estimate of the Pacific-America in character southwards from steep and 10.000 years. a very recent cessation motion, does confirm that the junction narrow off the Queen Charlotte Islands may not be detectable. remained stable close to Brooks and northern Vancouver island to a Peninsula from 10 to 4 m.y. before much wider, irregular slope off southern Inland Vulcantsm, Heat Flow moving to the nwthwest at a rate of the Vancwver Island. Washington and and Gravlty order of 1.8 cm yr-1. Using convergence Oregon. In broad terms this seems to be The subduction of oceanic lithosphere vectors from the detailed spreading a symptom of the change in the under a continent results in a nwnberof history and the abovemotionofthetriple interaction regime along the margin from characteristic processes. These extend junction, one may obtain a theoretical transform faulting to compression several hundred kilometres inland of the position for the northern edge of the plate and subduction. subduction zone itself and are subducted under Vancouver lsland Details of this interaction, between the principaliy manifested by characteristic during the last 10 m.y. (Fig. 2). The edge sediments at the foot of theslope and the panems of 'arc' or calc-alkaline strikes northeastward, the plate edge continent, have been studied bothoff vulcanism, heat-flow and gravity. aged 10 m.y. (assuming a 45" dip) now Vancwver lsland and off the In Washington. Oregon and northern being some 200 km inland from the Washington coast. OffVancouver lsland California,the vulcanism resulting from continental margin. as far north as Brooks Peninsula,Tiffin el Juan de Fuca plate subduction is well a/. (1972)and Murray andTiffin (1974) defined by the Cascade chain of Bathymetry, Baement Dlpand conclude that up to at least the lower volcanoes which occurs paralld to and Sediment Deformation at the Margin Pliocene (4 m.y. ago), many of the betwen 150 and 200 km inland from the Bathymetrically,theJuan de Fuca and structures of the continental slope are continental slope (e.g.. Dickinson.1970). Explorer Ridges are evident as broad characteristic of compression. They Although there have been few major ridges with central troughs interrupted also observe considerable recent recorded eruptions in historical times. by transverse ridgelscarp features faulting and note that deformation this chain is clearly active. In northern corresponding with the Sovanco and increases in intensity along the margin Washington, its northern trend tums to Revere-Dellwoodfracture zones. To the to the swtheast Barr (1 974) concludes the northwest and it continues into east of the Juan de Fuca Ridge there is that deformation continued at least until southern British Columbia up to

an almost leatureless~ ~ olain of low relief. the late Pleistocene. Off the Washinaton a~~roximatelv52"N. In British the Cascadia basin, which continues to coast Silver (1972) came to a simila; dd~umbia,it i$represented by a lineof the edge of the continental slope and conclusion and suggested a recent Miocene volcanic centres (the contains up to four km of flat-lying convergence rate of two cm yr-I from Pemberton volcanic belt. Souther. 1970 sealmenl Selsm c ref ect80ndata (e g . sedlmenlaryevlaence tntnesame area SW NE T~ff~nelal1972 Murrav ana Ttff~n. Carson eta1 (1 974) lnlerorel antrcl~nal 1974). the attenuation of magnetic ridges in the most rkcenisediments at anomalies (Barr. 1974) and gravity the outer edge of the slope as indicating measurements (Srivastava. 1973) have deformation which is continuous up to all been interpreted to indicate a the present. Both von Heune and Kulm significant eastward dip ol the oceanic (1973) and Scho11(1974) also deduce basement beneath thesesediments and that there is good evidence of very below the continental slope which is recent deformation and uplift in attributable to the effect of subduction. response to compression at the SchoIl(1974) regards the region near Washington-Oregon slope. Von Heune the foot of the slope as a 'structural and Kulm compute a convergence rate trench'which has been filled withland of from 1.6 to 2.7 cmlyr. derivedsediments. OR the Washington A typical seismic profile of the base of coast this trench fill is represented by a the continental slope off southern 1 I 200 to 500 m lhlck ILrblOltesequence Vanco~verlslana s shown In Flgure5. 0 10 20 krn aooarentlv accum~lateaw thln the oasl the strJctJre revealedcomoares closelv , . ,~~~~~ one m.y. Below this is a sequenceof with the deformation of the most recent FlgumS landward dipping Miocene to sediments at an outer anticlinal ridge Line drawing of fyprcal seismic profile across Pleistocene, finer grained, beds which examined by Carson et dl. (1 974). We the base of lhe conlinenlal slope off southern do not thicken appreciably toward the Conclude therefore, both from the dip of Vancouver Island (alter Barr. 1974). Nole margin. Scholl considers the upper the oceanic basement into a sediment deformaliw ollhe most recenlsedimenls in turbidite sequenceas a 'true' axial filled trench along from the margin and the firslanliclinalridge.

1976). Quaternary volcanism being volcanic arc also corresponds with pronounced low lust inland ofthe trench. limited to the NNW. Garibaldi volcanic changes from high to low Bouguer a plateau of normal heat flowto about belt which extends only to 51 ON (seeFig. gravityandchanges in anumberof other 200 km inland and high heat flowbeyond 2). The youngest recorded activrty in this geophysical parameters, (Fig. 6). Heat flow values for most ofthe belt is 2500 years ago. Souther (1976) There is evidence for this pattern in models of Hasebe etal. aresomewhat states that Garibaldi 'are north-western U.S.A. (Blackwell, 1969, higher than those observed but chemically similar to those of the 1975) and it has been shown by one of reasonable model parameters can be Cascade province of western U.S.A.. US to beclearly defined in south-westem readily found that give good agreement. and like the latter, their analyses plot British Colurnbia(Hyndman. 1976a). A The time required for the surface heat along a typical calc-alkaline trend that is number of authors have produced flow and other thermally related universally associated with Benioff zone thermal models that exhibit swneof the parameters to respond to a change or characteristics of this pattern but the magmatism'. Souther relates the J+termination of subduction is probably changes in trend ofthe British Columbia model that most successfully several million years for areas near the volcanic belts to changes in offshore reproduces the observed heat flow marginal 'low' and several tens of million plate geometry and it is interesting to pattern is by Hasebe et a1 (1970). This years for the inland'high'Thus whilethe note that the change from the NW. model fits the data for south-western heat flow pattern clearly indicates Pemberton belt to the NNW. Garibaldi British Columbiavery well, showing a subduction within the last few million belt occurred close in time to the sharp 25* change in Explorer-America convergence at four my. suggested by \ - \ Riddihough (1976 MS) (see Fig. 4). \ Coast Low Inland High Following the same reasoning, it is '\ 2 - \ possible that the present lull is a reaction -3 '\ to the initiation of nearly N-S Explorer- -z - . America convergence within the last X 2 I- onemy. (Fig.4.0.5 m.y.) LL \ Estimation of the time scale required B - \ for a change in subduction to be I O looI oI looI ZooI 300I 400I reflected in a change In volcanism is difficult. However, if changes are in response to a stress regime. as the Distance (km) above speculations suggest. it may be geologically rapid. Thus theexistenceof Juan de Fuco Van. lalana Volcanic Arc recent volcanism can probably betaken to imply that subduction has occurred 0 OOC and has continueduptoat least 0nem.y. 500 ago. The reduced level of volcanism in IOOU the Cascade - British Columbiavolcanic 1500 belt as compared to volcanic arcs such E I00 as Japan and the Aleutians seems to us -Z to be broadly explicable by the much C slower subduction rates involved (1 to 3 H cmlyr asopposedto5to8 cmlyr). n Another observed characteristic of 200 subduction zones is a band of low heat First melting and upwelling flow on the continental side of the active trench and much higher than normal \\. - heat flow further inland (eg., McKenzie 300L and Sclater. 1968; Hasebe el a/, 1970). The 'low zone is produced by the cold oceanic lithosphere which, acting as a heatsink beneaththecontinent. absorbs heat that would otherwise reach the surface (dehydration of hydrated minerals may complement this process). The inland 'high', which m commences at the volcanic arc. probably results from upwelling Figum 6 Stnkrng slab 1s below !he cr!t!cal remperalure derived from the melting of the sinking Schemallc cross-seclronof the conlrnenlal lorearlhquakes Heal flow values are shown slab at a depth of several hundred margfn olsouthweslern Bnl~shColurnb~a wrrh above anda smoothedBouguer gravlly kilometres The transition at the ~lluslralrvesolherms The dotredarea m the proflie shown below Geoscienee Cenada Volume 3. Nunber 4 275

years, a recent termination of narrow gravity 'high' coincides with the the relative plate motions is thus subduction would not be evident. heat flow minimum. inhibited. A group of earthquakes A characteristic pattern has been The Bouguer gravity 'low' (<-I 00 apparently in the regions of the Paul observed in thegravity field across mgal) which extends eastwards from the Revere fracture zone (Azimuth 128') almost all margins where subduction is volcanic arc is explicable as a result of and the Sovanco fracture zone(Azimuth occurring (e.g.. Grow and Bowin. 1975). higher temperatures and partial melt in t 10" to 120") studied by Chandra Previously published interpretations of the mantle. Assuming a mantle density (1 974) indicated right-lateral strike slip the gravity field over the continental below 60 km which is anomalously low motion at azimuths from 152O to 1No. If margin of south-western British by the order of 0.05 glcmj, both the such directions represented the mlion Columbia have not in general. observed gravity ellect and the higher between the Explorer and Pacificplates. considered the existence ofa dow- average elevation can be explained. the Explorer plate would be moving

going slab and other manifestations of Such a result~ is~~ similar to that of Stacev away from, rather than converging on subduction. They are, nevertheless. (1973) and is in conformity with a the American plale. However, as very broadly consistent with it. For example. number of studies of the inland region complex interactions are clearly taking Couch (1969) postulated a steeply (e.g., Berry eta1 1971; Berry and place in this area, we do not believe that dipping boundary separating high Forsyth, 1975) which have shown it to such a conclusion is valid. density mantle beneath the continent have both a thin crust and a thin hot Within the America plate. Rogers from lower density mantle beneath the lithosphere underlain by a pronounced (1976 a, bMS) has analysed the first ocean. Similarly. Stacey (1 973) showed asthenosphere. motions of a number of larger that the 'plateau' of near zero Bouguer The response of the density earthquakes and concludes that the gravity anomaly that lies over Vancouver distribution producing the gravity field to region can be divided into twodistinct lsland(Earth Physics Branch. 1974) change in or tenination of subduction areas. North of 4g0, in central between the oceanic 'high' (>+I00 is, like heat flow, quite slow. The low Vancwver Island the earthquakes are mgal) and the inland 'low' (-1 00 mgal) gravity inland of the volcanic arc will shallow and consistent with strike-slip required anomalously high density probably persist lor several tens of faulting in response to north-swth beneath the island. millions of years although the gravity compression. Such faulting could either One of us (Hyndman. 1976b) has high nearer the coast may disappear trend NW. (right-lateral) or NE. (lefl- recently shown that thegravity field of mud? more quickly, within several lateral). The vector interaction at the the region is compatible with the million years, as thesubducted slab margin from offshoremagnetic data (Fig. features expected in an actively heats and the eclogite reverts to lower 4) does in fact show N-S convergence subducting situation; the magnitude of density phases. between the American and Explorer the gravity anomalies is however, less plates for the most recent (0.5 m.y.) than lor other subduction zones Selsmlclly epoch. Analysis ofmotionalonga because the 'trench' is sediment filled Major subduction zones around the proposed eastward extension of the and the convergence rate is lower. world are characterized by the presence Sovanco Fracture Zone (Riddihough, Figure6 shows a Bouguer gravity profile of a narrow plane of earthquakes, the 1976 MS) also suggests that a NE. across southern Vancouver lsland. Benioltzone. steeply dipping beneath of oriented left lateral fault in the Profiles across southern Washington thevolcanic arc from the trench to a downgoing lithoshpere beneath and Oregon aresimilar. The maingravity depth of up to700 km. Thelackof such a Vancwver Island is quite probable: the features are explained qualitatively by Beniofl zone under western North northern edge of the downgoing plate the cross-section in the centre of the America has been held as primary (Fig. 2) is also oriented in the same figure. The zero Bouguer gravity evidence against subduction. We feel, direction. The crustal earthquakes 'plateau' probably arises frmthe thick however, that there are strong grounds observed may thus be related to the cold lithosphere occurring between the for concluding that earthquakes deeper coupling of one or all of thesestresses to trench and the volcanic arc and may than70 km are not to beexpectedinthis the America plate. rellt?Ct a crust and mantle structure reglon and that the observed seismicity South of 49"N. in theGulf Islands - rathersimtlar to stable continental areas. is, in fact, consistent with continuing Puget Sound area a high concentration Superimposed on this'plateau'there isa subduction. of earthquakes has been recorded. narrow +30 mgal h~ghover western There is considerable seismicity in the Although again they are predominantly Vancouver Island (also observed near region (Fig. 3) (e.g.,Milne. 1967; Milne er shallow, deeper earthquakes (down to the Coasts of southern Washington and a1 1976 MS; Crosson, 1972) and, in 60 km)occurwhich areconsistent with a Oregon) that is probably associated w~th consequence, little doubt that there is downgoing slab interpretation. Here. high denstty in the downgoing slab. intense tectonic activity of some form. McKenzie and Julian (1971 ), in a study Grow and Bowin (1975) haveshownthat Theoffshore seismicity generally of the travel times lrom the 1965 Seattle such a density contrast must exist and is follows the Oueen Charlotte fault. and earthquake, proposed the existence of largely caused by the basalt-ecologite the Explorer and Juan de Fuca Ridge an eastward dipping high velocity slab. phase transformation in the oceanic system. Epicentres, however, are rather lsaacks and Molnar (197l)andChandra crust (an increase in density from 2.9 to imprecisely located on specific ridges or (1 974) obtained adepth of about 60 km 3.5 g/cm3) and the low lemperature 01 transforms and may be systematically with extension stress parallel to the dip the s~nkingoceanic lithosphere with displaced (Wetmiller. 1971; K. Lee. pers. for the same event. Rogers (I976b MS) respect to thesurrounding material.This commun.). Accurate determination of finds similar depths and stress regimes for two other events in the area, further We thus conclude that seismicity data plate and the continent. This interaction supporting continuing subduction to the supports continuing subduction south of appears to have changed in direction present although confirming that the about 4g0N, i.e., theJuan de Fuca - very recently and to be rapidly slowing. geometry of the downgoing slab is American convergence. It supports but Earthquake data and recent sediment probably complex (Fig. 6). is not conclusive,for convergence to the deformation show evidence for The reason that earthquakes in the north of 4g0N, i.e., between the Explorer continuing compression and a stress region are limited to amaxirnum depth of and American plates. regime compatible with the interaction about 60 km is. in our opinion. readily but no clear evidence for an actively seen. The foci of deep BenioH Summlng-up underthrusting oceanic lithosphere. earthquakes are in general limited to the From reviewing the aspects of geology Nevertheless it is of critical importance sinking oceanicplate. Such aplate has a and geophysics relevant to subduction to realise that because of thevery slow long thermal time constant, so that under south-western British Columbia, it sinking rateinvolved, and the ywngand startingcold at the surface, its core will seems clear to us that there is definitive 'very thinoceanic plate being subducted. remalncoo~and se~sm~c~nstab~llty or ev,dence tor convergence ano many of the pnenomena normal y brlltle fract~reoccur, to a considerable subd~ctlonso~th of a DOlnl on the assoc~atedwltn s~bouctlonwtll bee lher depth. The surrounding asthenosphere margin near50 "N, at ieast until a few absent or not readily detectable. is hot and deforms plastically. The million years ago. This evidence is maximum depth of earthquakes is set by strongest from offshore magnetic the depth below which all of the sinking anomalies, from sediment deformation Atwater. T., 1970. Implications of plate plate exceeds the critical temperature along the continental margin and from tectonics for the Cenozoic tectonic for seismic instability or fracture. The onshore heat-flow values, gravity evolution of westernNorth America: Bull. two most important factors that anomalies and volcanic activity. As we Geol. Soc.Amer. v. 81, p. 3513-3536. determine this depth are: 1 )the rate of have stressed, the accuracy of the convergence or sinking speed of the evidence from these sources for the Barr, S. M., 1974. Structure and tectonics of the continental slope west of plate - the faster the plate is sinking the timing of subduction activity is very further it will descend before the critical variable and could not, in any case, be Vancouver Island: Can. Jour. Earth. Sci.. temperature is exceeded: and 2) the expected to prove or disprove v. l1.p 1187-1199. thickness of the subducted plate when it conclusively the existence of some form Barr. S. M. and R. L. Chase, 1974, commences to sink thethicker the of present subduction. Earthquake data Geology of the northern end of Juan de plate, the longer its thermal time however, seem to us to indicate very Fuca Ridge and sea-floor spreading: constant and the further it will penetrate strongly that at least south of4g0N, Can. Jour. Earth. Sci.. v. 11, before completely heating above the subduction of oceanic lithosphere (the p. 1384-1406. critical temperature. Juan de Fuca plate) beneath the 0erry.M. J. and D. A. Forsyth. 1975. lsaacks eta1 (1966) and McKenzie continent, IS continuing. Structure of the Canadian Cordillera (1 969) showed that the maximum Between 4g0N and Brooks Peninsula. from seismic refraction and other data: earthquake depth is approximately near 50°N, marine magnetic evidence Can. Jour. Earth. Sci. v. 12, p. 162-208. proportional to the sinking rate: Morgan suggests that convergence at the (in Deffeyes. 1972) showed that the margin may be distinct from that to the Berry. M.J.. W. R. Jacoby, E. R. Niblett depth depends directly on theageofthe south and becontrolled by the relative and R. A. Stacey. 1971. A review of sea floor beingsubducted, thethickness motion between the Explorer oceanic geophysical studies in theCanadian Cordillera: Can. Jour. Earth. Sci., v 3. of the oceanic lithosphere being Latitude CN) approximately proportional to the square p 788-801 80 48 46 44 root of its age (e.g., Parker and Blackwell, D. D., 1969. Heat-flow Oldenburg. 1973). Using relations determinations in the north-westem obtained from compilations of world :"'lo United States: Jour. Geophys. Res.. subduction zones, oneof us (Hvndman. 20 \ lo' v. 74. 0. 992-1007. 1976b) has estimatied maximum depths Blackwell, D. D., 1975, Terrestrial heat- for Benioff earthquakes in the region flow and its implications for the location (Fig. 7. and cross-section in Fig. 6). The of geothermal reservoirs in Washington: variation with latitudc ariscs both from ~nEnergy Resources of Washington: variation in sinking rateand in the age(or Information Circular 50, Dept. Nat. thickness) of the plate being subducted. Resources, Div. Geol. and Earth The maximum depth of earthquakes of Resources.Olympia. Wash., U.S.A. 70 km for a subduction rate of three Carlson, R. L., 1976. Cenozoic plate cmlvr at 45O dioand 10 m.v. old sea "" floor, is in good agreement with the convergence in the vicinity ofthe Pacific North-west: Ph.D. Thesis, Univ. observed maximum of about 60 km for Flgure7 Washington and Oregon. The predicted ~h,~~~d,~t~d~~~i~~~depth of Washington. maximum depth under Vancouver earthquakes under British Columbia, Island (Explorer plate) is less that30 km. WashingtonandOregon. Geoscience Canada. Volume 3. Number 4 277

Carson. 0.. J. Yuan. P. B. Myers, and W. Hyndman, R. D.. 1976b.Geophysical Raff. A. D. and R. G. Mason. 1961. D. Barnard. 1974, Initial deep-sea structure of the subduction zone of Magnetic survey off the west coast of sediment deformation at the base of the southwestern Canada: Trans. Amer. North America. 40"N to 52"N Latitude: Washington continental slope: a Geophys Union. v. 57, p. 334. Geol. Soc. Amer Bull.. v. 72, p. 1267- response to subduction: Geology, v. 3. Isaacks. 8, and P. Molnar, 1971. 1270. p. 561 -564. Distribution ol stresses in the Riddihough. R. P.. 1976MS. Lithospheric Chase.R.L.,D.L.Tiffin, andJ.W. Murray, descending lithosphere from a global plate interactions off Canada's west 1975. The western Canadian survey of focal mechanism solutions of coast during the last 10 million years: continental margin: Can. Soc. Petrol. mantle earthquakes: Rev. Geophys.. Earth Physics Branch Contribution Geol.. Mem. 4, p. 701 -721. v. 9, p. 103-174. (in prep.). Chandra. U.. 1974. Seismicity. Isaacks. 0.. J. Oliver, and L. R. Sykes. Rogers. G. C. 1976a. The Vancouver earthquake machanisms and tectonics 1968. Siesmology and the new global Island earthquake of 5 July 1972: Can. along the western coast of North tectonics: Jour. Geophys. Res., v. 73. Jwr. Earth Sci.. v. 13, p. 92-101. America from42"N to61 ON: Bull. Seism. p. 5855-5899. Rogers. G. C. 1976b MS, Earthquake Soc. Amer.. v. 64, p. 1529-1549. McKenzie. D. P.. 1969. Speculations on mechanism solutions near Vancouver Couch. R. W., 1969. Gravity and the consequences and causes of plate Island: Earth Physics Branch Structures ofthe crust and sub-crust in motions: Geophys. Jour. v. 18. p. 1-32. Contribution (in prep.). the north-east Pacific Ocean west of McKenzie. D. P, and 8. Julian, 1971,The Scholl. D. A. 1974. Sedimentary Washington and British Columbia: Ph.D. Puget Sound, Washington earthquake sequences in North Pacific trenches: in Thesis, Oregon State Univ. and the mantle structure beneath the C. A. Burk, and C. L. Drake, eds.. The Crosson. R. S.. 1972. Small earthquakes. north-western United States: Geol. Soc. Geology of Continental Margins: New structure and tectonics of the Puget Amer. bull..^. 82,p. 3519-3524. York. Springer-Verlag. p. 493-504. Sound region: Bull. Seism. Soc. Amer.. McKenzie. D. P. and W. J. Morgan. 1969. Silver, E. A,. 1971, Small v.62.p. 1133-1171. Evolution of triple junctions: Nature. in the north-eastern Pacific: Geol. Soc. Davis, E. E. andC. R. B. Lister, 1976, v. 224, p. 125-133. Amer. Bull., v. 82, p. 3491 -3496. Tectonic structure of theJuan de Fuca McKenzie. D. P. and R. L. Parker. 1967, Silver. E. A,. 1972. Pleistocene tectonic Ridge Jour. Geophys. Res.. (in press). The North Pacific: an example of of the continental slope off Deffeyes, K. S.. 1972. Plume convection tectonics on a sphere: Nature. v. 21 6. Washington: Marine Geology. v. 13. with an upper mantle temperature p. 1276.1 280. p. 239-249. inversion: Nature, v. 240, p. 539-544. McKenzie, D. P. and J. G. Sclater, 1968, Souther, J. G.. 1970. Volcanism and its Dickinson. W. R.. 1970. Relation of Heat now inside the island arcs ofthe relationship to recent crustal andesites, granites and derivative north-western Pacific: Jour. Geophys. movements in the Canadian Cordillera: sandstones to arc-trench tectonics: Rev Res., v. 73, p. 31 37-31 79. Can. Jour. Earth Sci.. v. 7, p. 553-568. Geophys.,v. 8, p. 81 3-860. Menard. W, and T. Atwater. 1968. Souther, J. G. 1976, volcanism and Earth Physics Branch, 1974. Bouguer Changes in direction of sea-floor tectonic environments in the Canadian anomaly mapof Canada: Gravity Map spreading: Nature. v. 219. p. 463-467. Cordillera - a second look: in W. R. A. Baragar. L. C. Coleman, andJ. M. Hall, Ser. 74-1. Earth Phys. and Dept. Energy. Milne. W. G., 1967. Earthquake eds., Volcanic Regimes in Canada: Geol. Mines and Resources. Ottawa. epicentres and strain release in Canada: Assoc. Can. Spec. Paper No. 16, Grow. J. A. and C. 0. Bowin. 1975, Can. Jour. Earth Sci.. v. 4, p. 797-81 4. ,:- -.---\ ,"I pfcas,. Evidence for high density crust and Milne. W.G., R.D.Hyndman, K. Lee, R. P. Srivastava. S. P., 1973. Interpretation of mantle beneath the ChileTrench Riddihough, and G. C. Rogers, 1976 MS. gravity and magnetic descending lithosphere:Jour Geophys. seismicityof western canada: ~~~h across the continental margin of British Res.. v. 80, p. 1449.1 458. Physics Branch Contribution (in prep.). Columbia, Canada:Can. Jour. Earth Sci., Griggs, D. T. 1972. The sinking Morgan. W. J., 1968, Rises, trenches, v. to, p. 1664-1677. lithosphere and the focal mechanisms of great faults and crustal blocks: Jour. Srivastava, S. P., D. L. Barren. C. E. Keen. deep earthquake: ;nE. C. Robertson,ed., Geophys, R~~,,v. 73, p, 1959.1982. The Nature of the Solid Earth: New York. K. S.Manchester, K.G.Shih. D. L. Tiffin. McGraw-Hill, p. 361 -384. Murray, J. W. and D. L. Tiffin, 1974. R. L. Chase, A.G.Thomlinson. E. E. Patterns of deformation, sedimentation Davis, and C. R. B. Lister. 1971. Hasebe, K.. Fuiii,and S. "yeda. 970, N. and tectonism, southwestern Canadian preliminary analysisof geophysical Thermal processes under island arcs: contin,tal ac,~601, measurements north of Juan de Fuca Tectonophys.. v. 10, p. 335-355. Belgique, v. 97, p. 169-183. Ridoe: Can. Jour. Earth Sci.. v. 8. Hyndman. R. D.. 1976a. Heat flow Parker. R. L. and D. W. Oldenburg. 1973. P. 615-628. measurements in the inlets of Thermal model of ocean ridges: Nature, Stacey, R. A,1973. Gravity anomalies, southwestern British Columbia:Jour. v. 242, p. 137-139. crustal structure and plate tectonics in Geophys. Res..v. 81. p. 337-349. the Canadian Cordillera: Can. Jour. Earlh Sci.. v. to. p. 61 5-628. Talwani M.. C. C. Windisch, and M. G. Langseth. 1971, Rekjanes Ridge crest: a detailed geophysical study; Jour Geophys. Res.. v. 76. p. 473-51 7. Tiffin. D. L.. 8.E. 8. Cameron, andJ. W. Murray. 1972. Tectonic and depositional history of the continental margin off Vancouver Island. B. C.: Can. Jour. Eallh Sci., v. 9, p. 280-296. Tobin. D. G. and L. R. Sykes, 1968, Seismicity and the tectonics of the northern Pacificocean: Jour. Geophys. Res.. v. 73, p. 3821 -3845. Vine F. J.. 1966. Spreading of the ocean floor; new evidence: Science, v. 154. p. 1405-141 5. Vine. F. J., 1968. Magnetic anomalies associated with mid-ocean ridges: in R. A. Phinney, ed., The History of the Earth's Crust: Princeton Univ. Press, p. 73-89. Vine, F. J. and D. H. Matthews. 1963. Magnetic anomalies over ocean ridges: Nature. v. 199. p. 947-949. Vine. F. J. and J. T. Wilson. 1965, Magnetic anomalies over a young oceanic ridge off Vancouver Island: Science, v. 150. p. 485-489. von Heune, R. and L. D. Kulm, 1973. Tectonic summary of leg 18: in Initial Reports of the Deep Sea Drilling Project. v. 28. Washington, U. S. Govt. Printing Off., p. 961 -976. Wetmiller. R. J.. 1971. An earthquake swarm on the Queen Charlotte Islands Fracture Zone: Bull. Seism. Soc. Amer.. v. 61, p. 1489-1505. Wilson, J. T., 1965, Transform faults, oceanic ridges and magneticanomalies southwest of Vancouver Island: Science. v. 150. p. 482-485. Contribution of the Earth Physics Branch, No 631 MS received August 23.1976