ML20151H106
| ML20151H106 | |
| Person / Time | |
|---|---|
| Site: | Trojan File:Portland General Electric icon.png |
| Issue date: | 04/12/1988 |
| From: | Spence W INTERIOR, DEPT. OF, GEOLOGICAL SURVEY |
| To: | |
| Shared Package | |
| ML20151H012 | List:
|
| References | |
| NUDOCS 8808010145 | |
| Download: ML20151H106 (56) | |
Text
!
f
\\
i ANOMALOUS SUBDUCTION AND THE ORIGINS OF STRESSES AT CASCADIA P
by William Spence National Earthquake Ir. formation Center U.S. Geologicat Survey P. O. Box 250G, MS 967 i
Denver, Colorado 80225 l
December 16,1987 k
i' 1
$DR808010145 880615 p
ADOCK 050co344 PDC 1
,_y.,-_-
.---_..-~~_.,-_--.,_-__--_,_,,,,.-,__,,.--_~m, y-...-_,_-,
ABSTRACT This framework paper on the seismicity and tectonics at the Cascadian plate system indicates that a primary regional stress is northerly compression, e en though the Juan de Fuca plate genatally is thought to be subducting N50*E. New and existing earthquake l
focal mechanism data show that this compression is pervasive throughout the Gorda. Juan c
de Fuca Explorer plate system and much of the adjoining section of North American piste.
i s
Modeling, using a discrete element code, indicates that this north trendmg compression is due to the Pacinc plate being driven into the Gorda block and Juan de Fuca plate (at the
\\
Mendocino and Blanco fractute zones), causing compression of the plate system northwards
)
to the 45* W shift of the coutline at Vancouver Island. At the contacts between the offshore 4
b plate system and the overriding plate, this compression is coupled into the overriding plate. Detailed studies of the magnetic anomaly patterns of the Juan de Fuca plate show I
that this plate's absolute subduction rate recently hu slowed 60?ii, from 45 km/m.y. at
\\
(
ti.5 m.y. ago to 17 km/m.v. at 0.5 m.y.
ago (Riddihough,1984). Sfagnetic anomaly 1
lineations cannot resolve plate motions for the fut 500,000 years; thus present.Qy details ofsubduction of the Juan de Fuca plate must be inferred from recent geologic, tectonic, and
(
i seismic data. Severalindependent lines of evide ejndicate that the Cascadian subduction interface is locked and that subduction is not occurring aseismically. The great resistance to subduction provided by the Cucadian interface thrust zone recently has caused the i
Explorer subplate and the south Gorda block to move independently from the main Juan i
de Fuca plate, and subduction of the Gorda block probably has stopped. A comparison of the cro<s.section of Wuhington's subduction so.te with cross sections for six other well.
t 4
resolved subduction sones shows that the longest downdip length ofinterface thrust zone j
is beneath Washington's Olympic blountains. This relates to the youth of plate subducted i
l beneath Wuhington and may imply relatively great resistance to subduction. Earthquake l
focal mecisanisms indicate extensional stresses that trend downdip within the subducted
)
plate beneath the Puget Sound region. This extension is consistent with stresses, due to the
- k 2
J
o slab pull force, observed in other subducted plates. The slab pull force acting at a locked interface thrust zone is shown to be a likely cause of the geodetically observed warping and northeast trending compression along the CascaSa coast. The observed fragmentation of the offshore Juss. de Fuca plate, the slowing of subduction of the Juan de Fuca plate, the strong inBuence c( the Pacinc plate's motion in causing stresses at the Cascadia plate j
system, and the lack c.! thrust faulting earthquakes at the Cnscadia interface thrust zone suggest that in plate driving forces may no longer dominate the Juan de Fuca plate's motion and : hat a long term cessation of subduction at Cascadia is in progress. Subduction and associated large earthquakes are most likely beneath Washington but beneath Oregon subduction is much slower or already may have ceased. The northward compression in the Coscadia plate system may be be capable of cznsing large, crustal earthquakes in and offshore of Vancouver. Washington, Oregon, and northern California.
INTRODUCTION 1
The Washington Oregon trench is filled with 12 km of terrigenous deposits (Scholl, 1974; von Huene and Kulm,1973), and the presence of this shallow trench reflects recent subduction of young oceanic lithosphere (lieaton and Hartzell,1986). Although active volcanism at the Cascade volcanic chain indicates subducted plate at least to depths of about 100 km (Isacks and Barazangi,1977), most of the subducted lithosphere lacks earth-quakes (Weaver and Michaelson,1987; Weaver and Baker,1987). This subducted plate l
and the Gorda Juan de Fuca Explorer plate system that is seaward of the trench are rem-nants of the Farallon plate, which formerly underthrust much of western North America (Atwater,1970). This offshore plate system is sandwiched between the San Andreas and Queen Charlotte transform fault systems, with the northwestward motion of the Pacific plate causing right lateral shear on each transform fault, i
The plate motion history of the Juan de Fuca remnant is preserved in its pattern of s
(
magnetic anomaly reversals (Raff and Mason,1961; Vine,1968; Peter and Lattimore,1969:
j 3
J
Elvers et al.,1973). The changes in shape of the anomalies" (fanning and pseudofaults) in the Juan de Fuca plate allow detailed reconstructions of this plate's recent motions (Hey, 1977; Riddihough,1984; Wilson et al.,1984; Nishimura et al.,1984). Riddihough's recon-structions of the absolute motions of the Juan de Fuca plate system show how it has been slowing down and fragmenting over the last 6.5 m.y., while maintaining a northeasterly subduction direction (Figure 1). This absolute velocity of the Juan de Fuca plate is the same as the velocity of this plate into the asthenosphere. The convergence rate between the Juan de Fuca plate and the overriding North American plate is the vector sum of their individual absolute velocities; the convergence rate determines both the seismic coupling between plates and the characteristic maximum earthquake at the plate interface. How-ever, the absolute velocity of the Juan de Fuca plate reflects the in-plate driving forces for that plate's motion, and determines that plate's contribution to coupling at the plate interface. At 6.5 m.y.B P. the absolute rate of the Juan de Fuca plate was about 45 km/my (4.5 cm/yr). By 2.5 m.y.B.P. and 0.5 m.y.B.P. the absolute rates had slowed to about 25
(
km/my and 17 km/my, respectively. Similarly, Nishimura et al. (1984) find the most recent absolute motion of the Juan de Fuca plate to be 10-20 km/my. Thus the absolute motion of the Juan de Fuca plate has slowed by about 60% over the interval 6.5-0.5 m y.B.P., to become one of the slowest moving plates on Earth.
At 500,000 years ago the absolute pole of the Juan de Fuca plate was in northern California, implying much slower plate convergence at Oregon than at Washington State (Riddihough,1984). At this time the Corda block contained its own rotation pole,impiying that this block was neither subducting nor being overridden but was moving with the North American plate. The southeast corner of the Gorda block is the Mendocino triple junction. This triple junction continues to be moved northward, as the San Andreas fault is lengthened and as the Juan de Fuca plate system is reorganized. The E.tplorer subplate began to act independently of the Juan de Fuca plate at about 4 m.y. ago (Riddihough, 1984). These recent plate motions at Cascadla indicate that the subduction of the Juan de 4
Fuca plate is anomalously complicated. Magnet c anomaly lineations cannot resolve plate i
motions for the last 500,000 years; thus present-day details of subduction of the Gorda-Juan de Fuca. Explorer plate system must be inferred from recent geolcgic, tectonic, and seismic data.
POTENTIAL FOR A GREAT EARTHQUAKE AT THE CASCADIA SUBDUCTION ZONE Dased on analogies with subduction zones that appear to be similar to the Cascadia subduction zone or on geodetic and seismicity data at the Cascadia subduction zone, nu.
mesm's workers argue that the Juan de Fuca plate continues to subduct the North American plate (Ando and Balazs,1979; Rogers,1979: Savage et al.,1981; lleaton and Kanamori, 1984; Taber and Smith,1985; Weaver and Michaelson,1985; Baker arid Langston,1987).
The age of the Juan de Fuca plate at the trench is a very young 8-10 m.y.B.P. Subduction of relatively young oceanic lithosphere often is associated with great earthquakes (Ruff and Kanamori,1980). This implies that plate convergence and subduction at Cascadia may be accompanied by great earthquakes.
A global summary of the characteristic maximum earthquake for various subduc' ion zones vs. convergence rate and age of the subducting oceanic lithosphereis shown by Figure 2 (adapted from Ruff and Kanamori,1980). This figure shows a strong inverse relationship i
i between the age of subducting oceanic lithosphere and the characteristic maximum earth-quake. This implies that the properties of the subducting oceanic lithosphere dominate over those of the overriding plate in determining the size of the characteristic maximum earthquake. The characteristic maximum earthquakes for various subduction. zones have been related to the sizes of the main driving forces of oceanic plates (Spence,1987), which are functions of the ages of oceanic lithospheres. For a given subduction zone, the ridge push force, due to lendward increasing density gradients of oceanic lithosphere, generally is considerably smaller than the slab pull force, due to the negative buoyancy of subducted 5
plate (Forsyth and' Uyeda,1975; Carlson,1983). This effect increases for increasing age of oceanic lithosphere. At an interface thrust zone the decrease in coupling due to the slab pull is more significant than the increase in coupling due to the the ridge push force and the seaward advance of the overriding plate, leading to the results of Figure 2. However, the 4
slab pull force generally is so much larger than the ridge push force that it dominates over the ridge push force in loading stresses that cause subduction zone earthquakes (Spence, 1987).
In Figure 2 the slo.ving of the convergence rate between the Juan de Fuca and North American plates is shown with convergence rate data for other subduction zones. Because i
2 the absolute velocity of the North American plate has remained at about 2.2 cm/yr this J
slowing convergence at Cascadia entirely is due to the slowing of the absolute velocity of the Juan de Fuca plate (Riddihough,1984). This implies a corresponding gradual-
~
decrease in the in plate driving forces of the Juan de Fuca plate, particularly a decrease of 4
(
the slab pull force of the subducted Juan de Fuca plate. This slowing may also be aided by increasing resistance to subduction at the shallow interface, due to increasing plate buoyancy of possible increasingly younger plate entering the subduction zone.
3 Great earthquakes at southern Chile, Colombia, C. Mexico, and SW Japan, asrociated with subducting oceanic lithosphere of about the same age as the Juan de Fuca plate (see Figure 2), have fostered suggestions that the Cascadia subduction zone has the potential i
for producint great earthquakes (Heaton and Kanamori,1984; Heaton and Hartzell,198G and 1987). However, the subduction setting at Cascadia has significant differences from the subduction settings for those great earthquakes. The Mw 9.5,1960 Chile earthquake began at interface contact with 25 m.y.-old oceanic plate, and dynamically ruptured into l
Interface contact with oceanic plate <4 m.y. old (Spence,1987). The absolute motion rates for subducting oceanic plates at the Colombia (Kanamori and McNally,1982) and l
Chile source zones are several times greater than for the Juan de Fuca plate (Minster et al.,1974), implying much stronger slab pull forces at Colombia and Chile. The SW Japan 1
6 l
source zone is near an arc arc intersection and interaction between subducted slabs causes C
lateral stretching of the Philippine Sea plate (Ukawa,1982), much unlike deformation in the subducted Juan de Fuca plate. The 1932 Jalisco, Mexico earthquake occurred at the Rivera plate, which has only about 1/3 the area of the Juan de Fuca plate (Singh et al,1985). liere the oceanic lithosphere is about 9 m.y. old and the convergence rate is only about 2 cm/yr. The Jalist.o source perhaps is the closest analog to possible seismic subduction at Cascadia. in general each source zone for great earthquakes that has been termed analogous with Cascadia has its own very unique properties that make the nature of stress accumulation there different from stress accumulation at Cascadia.
The rate of slowing of the absolute motion of the Juan de Fuca plate indicates that it may be ir. valid to extrapolate to the present this plate's most recently known motions based on 500,000 yr-old geomagnetic anomalies. The lack ofinstrumentally recorded earthquakes at the interface between the subducting Juan de Fuca plate and the overriding North American plate (Crosson,1983; Taber and Smith,1985) has led to uncertainty as to
\\
whether subduction of the Juan de Fuca plate is active and if active, whether it is seismic.
Stresses near plate interfaces at subduction zones typically show compression in the direction of plate convergence (Nakamura and Uyeda,1980). Focal mechanisms of earth-quakes in the St. *< elens seismic zone have P axes parallel to the theoretical direction of plate convergence, and these earthquakes in the overriding plate have been interpreted as evidence for a locked interface thrust zone (Weaver and Smith,1983). The focal mecha-nistns of earthquakes within the subducted Juan de Fuca plate, beneath or downdip of the interface thrust zone of northwest Washington, generally have T axes that are downdip l Table 1; Figure 3], also consistent with a locked interface thrust zone (Spence,1967).
Ilowever, understanding the tectonic framework at Cascadis is complicated by focal mech-anistns of many earthquakes in the Gorda block and near Vancouver having N.S trending axes of maximum compressive stress (Bolt et al.,1968; Rogers,1979; liyndman and We-ichert,1903) l Figure 3l. The focal mechanisms of many shallow, crustal earthquakes in 7
the Puget Sound depression also have N-S-trending axes of maximum compressive stress (Crosson,1972 and 1983; Weaver and Smith,1983).
Recent work by Atwater (1987) suggests that coseismic subsidence of the Washington.
coast is the explanation for several sudden burials of coastal estuarine vegetation over the last 7000 years. Geodetic studies indicate that the coast from Vancouver Island to south central Washington is subject to.NE-SW compression, consistent with deformation at tu active subduction ::one. However, the Washington Oregon coast is being uplifted and tilted Inndward (Ando and Balazs,1979; Savage et al.,1981; Reilinger and Adams, l
1982; Riddihough,1982; Adams,1984). This tilt is opposite to that often observed when stress is accumulating at a locked interface thrust zone, and the interpretation of these tilt data seem ambiguous (Weaver and Smith,1983).
j The objective of this paper is to evaluate the ridge push and slab pull forces at the-1 Cascadia subduction zone, and to understand the interaction of the Gorda Juan de Fuca-Explorer plate system with the motion of the Pacific plate. In this study, the spatial I
distribution of regional earthquakes and their focal mechanisms are integrated with known plate driving forces to explain the origin of stresses within the lithospheric plate elements j
in and around the Cascadia subduction zone, thus more sharply focussing the present debate on the potential for large or great earthquakes to occur at Cascadia.
Figure 3 contains the seismicity and focal mechanism data on which this study primar-ily is based. Earthquakes are concentrated within the Gorda block, the Explorer subplate, and at the Blanco fracture zone but are sparse within the Juan de Fuca plate and along l
l the coasts of Oregon and Washington The plotted earthquakes, m6 > 5.1, are the result of merging catalogs from the U.S. Coast and Geodetic Survey, the International Seismological Centre, the U.S. Geological Survey, Oregon State University, The University of Washing-ton, and the Earth Physics Branch, Ottawa. Except as noted, only earthquakes since 1964 are included because earlier earthquakes may be poorly located. Focal mechanism data i
for the larger and significant earthquakes are included for events back to 1934, and these
(
i 8
1 J
nineteen focal mechanisms are shown in Figure 3 and in Ta'ble 1. Care was exercised only to include reliable focal mechanisms, whose data are not clustered near nodal planes. New focal mechanisms are shown in detail in Figure 4. Assuming that the regional axes of greatest and least compressive stress, ai and o, generally are aligned with the P and s
T-axes of earthquake focal mechanism solutions (Zoback et al.,1987), these nineteen focal l
mechanisms then reflect much of our present knowledge of stresses at the Cascadia region.
P STRESSES in the r
GORDA, JUAN de FUCA, and EXPLORER LITHOSPHERES F
Stresses in and near the Gorda block Numerous recent geologic studies indicate that the Atendocino triplejunction is being i
driven northward by the Pacific plate (e.g. Fox et al.,1953; Sarna-Wojcicki et al.,1980), at a rate comparable to that of the Pacific plate's motion of 58 km/my. This is an ' unstable' triple junction, leading to the creation of new plate $undary at the northwards extending San Andreas fault (Dickinson and Snyder,1979). Subducted plate exists south of this triple junction, as described by Jachens and Grisom (1983)
Focal mechanisms 1 and 2 (Figures 3 and 4; Table 1) are for the two largest earthquakes since 1964 on the hiendocino fracture zone and they each are consistent with right lateral, strike-slip faulting on the blendocino fracture zone. The axes of compression for these earthquakes nearly are parallel to the absolute motion of the Pacific plate. Focal mechanism i
studies of small earthquakes at the eastern hiendocino fracture zone include some consistent with the Pacific plate's causing northward compression at that fracture zone (Seeber et al.,1970; Simila et al.,1975).
i 1
Within the Corda block, focal mechanisms for sizeable earthquakes suggest right-lateral, NW strike slip faulting (Bolt et al.,1968). Numerous mapped faults in the Gorda l
j block trend north to northwest (Carver et al.,1986), overprinting the north to northeast I
r trend of magnetic anomalies (Silver,1971; Riddihough,1980). The largest known earth-4 9
I d
a
(-
quake in the Gorda block is the 1980, Ms 7.3 strike-slip event (focal mechanism 3 of Figures 3 and 4, and Table 1). Aftershocks of this earthquake trend 150 km southwest from the main shock, nearly to the Mendocino fracture zone (Eaton,1981). Thus the preferred fault plane for this main shock indicates left-lateral, strike slip faulting. Earthquakes 3,4, and 5 are strike-slip events cnd their axes of maximum compressional stress trend approximately northward. At the Gorda block, the focal mechanisms for earthquakes 1-5 indicate that the westward component of the Pacific plate's motion is accomodated by right-lateral, strike slip along the hiendocino fracture zone, and that the northward component of the Pacific plate's motion leads to north directed compression, with erithqua'<es occurring or' a set of conjugate faults in the Gorda block.
The high level of seismicity in the o!Tshore Gorda block is consistent with this block being coupled to the overriding North American plate and resisting the northward push t
from the Pacific plate. This supports Riddihough's (1984) conclusion that the Gorda block j
(-
may be very strongly coupled to the overriding plate. Earthquakes 3 and 5 initiated their r
ruptures in or beneath the continental shelf, consistent with north-trending compression in the Gorda block being transferred to the overriding plate. The highest uplift rate for the Cascadia subduction zone is at Cape hiendocino (Lajole et al.,1933), probably reflecting the compression near the Mendocino triple junction. Focal mechanisms for several small t
earthquakes, at depths 30 to 87 km in the subducted Gorda block, show normal, strike-slip, and thrust faulting (Cockerham,1984; Walter,1986), unlike deformat!on in typical suducted plate. This stress complication might be expected as the plate subducted be-
{
neath the Gorda block and attached to the North American plate would be subjected to
' rudder' forces.
The northern bouadary of the Gorda block generally is taken to be near the landward extension of the Blanco fracture zone, because of the high intraplate seismicity south of that extenslor Owever there is no clear tectonic feature that corresponds to a boundary
(
between the Gmda block and the Juan de Fuca plate. The distortion of magnetic anomalies e
10
in the Gorda block' has,been shown to be due to internal deformation there (Wilson,1986; i
Stoddard,1987). The spreading history of the South Gorda ridge has been independent from that of the North Gorda ridge for the last 2-3 m.y. (Riddihough,1980). At the Gorda ridge,in the context of a non-subducting Gorda block, the mapped normal faulting (Atwater and Mudie,1968) and a normal-faulting earthquake (focal mechanism 6) would be due to the Gorda ridge's response to the motion of the Pacific plate.
Stresses in the ofshore Juan de Fuca plate l
l Evaluation of stresses within the offshore Juan de Fuca plate is made difficult by the scarcity of earthquakes there, particularly where they would be expu:ted '. the trench I
and laterface thrust zone. The Juan de Fuca ridge essentially is aseismic. The Blanco fracture zone, however, has frequent earthquakes. The Pacific plate has a component of motion into the Blanco fracture zone and this fracture zone is rotated about 15' clockwise
[
from its expected orthogonality to the Juan de Fuca ridge. The Blanco fracture zone j
consists of a series of strike-slip faults that are offset by extensional basins (Embley et al., in press). Bolt et al. (1965) found some normal faulting earthquakes at the Blanco 4
fracture zone, possibly related to these extensional basins. The two largest earthquakes on the Blanco fracture zone in Figure 3 are shown by the nearly identical strike slip focal 4
mechanisms 7 and 8 (also see Figure 4 and Table 1). The preferred fault planes for these
~
earthquakes strike parallel to the Blanco fracture zone. The P axes for these mechanisms 1
)
trend about 35' north of the Pacific plate's motion vector but are at right angles to the j
possible motion vector of the Juan de Fuca plate. This implies that the strike slip faulting on the Blanco fracture zone is caused by the motion of the Pacific plate rather than motion l
of a subducting Juan de Fuca plate. These large strike slip earthquakes are concentrated at q
the Blanco ridge, at the eastern third of the Blanco fracture :one. Iback (1981) Interpreted a sediment wedge at the south side of the Blanco ridge as evidence for compression acting 4
4 across the transform. Thus, the mapped structure and the earthquake focal mechanisms 11
]
^
at the Blanco fracture zone indicate that the westward component of the Pacific plate's
(
motion is accomodated by right lateral, strike-slip faulting and extensional basins, and that the northward component of the Pacific plate's motion leads to compression acting across the Blanco fracture zone.
The only two earthquakes known to be interior to the offshore part of the Juan de Fdca plate are indicated by focal mechanisms 10 and 11 (Figures 3 and 4; Table 1). Focal mechanism 10 is for a m6 5.8 earthquake in the central Juan de Fuca plate. This 1973 l
i I
earthquake has a north-trending P axis. The P-axis of this highly reliable focal mechanism 2
is consistent with the P axes for most previously discussed focal mechanisms. Focal mech-anism 11 is anomalous. It was determined for the smallest earthquake (ma 4.6) of Figures 3 and 4, because of its location in the central Juan de Fuca plate. Although this is the least certain mechanism in this study, the P-wave data on which it is bued are internally P
consistent and the P-axis of this earthquake appears to be nearly parallel to the 500,000 yr old absolute plate motion vector for the Juan de Fuca plate (Riddihough,1984). While
(
this earthquakes's P axis could be interpreted as due to plate compression due to the ridge push force (analogous to earthquakes studied by hiendiguren,1971, and Christensen and l
Ruff,1983), this P axis orientation may be fortuitous. Bratt et al. (1955) modeled stresses to match the main characteristics of near ridge earthquakes and concluded that stresses of thermoelastic origin dominate over stresses of ridge push origin in causing earthquakes l
1 in oceanic lithosphere younger than 15 m.y. On the other hand,if the Juan de Fuca plate is strongly coupled to the North American plate, then compressional stresses arising from e
the southwestward motion of the North American plate may be sufficient to have produced this earthquake. In general, the focal mechanisms of coastal and offshore earthquakes of the Gorda block and the Juan de Fuca plate are inconsistent with subduction of the Juan de Fuca plate but are consistent with stresses originating with the Pacific plate's driving the hiendocino and Blanco fracture rones northward.
Stresses in and near the Explorer subplate
(
1 l
12 i
i i
The Explorer'subplate's active tectonics (Davis and RI'ddihough,1982) and seismicity
(
(Figure 3; Milne et al.,1978; Hyndman et al.,1979) ludicate that it is undergoing inten-sive deformation. Although Figure 3 shows numerous large earthquakes in the Explorer subplate, their P-wave first motions generally are so inconsistent that reliable focal mech-anism solutions could not be obtained. Such earthquakes include the Als 6.8 earthquakes of Dec. 20,1976 and Dec. 17,1980. Figure 1 !ndicates the complicated tectonic evolution of the Explorer subplate, showing the recent tendency for sections of the Explorer ridge to jump to the northwest (Davis and Lister,1977), and the development of new fracture zones. Reconstructions of earlier plate positions (such as by Atwater,1970) indicate that, prior to this local plate reorganization, the Vancouver triple junction essentially was in a fixed position. Riddihough (1984) shows that the Explorer subplate began to act inde-pendently of the Juan de Fuca plate about 4 m.y. ago and that the Explorer subplate's absolute pole of rotation has moved to near that subplate. Thus the Explorer subplate is not strongly subducting (due to in-plate forces) but still may be overridden by the North
(\\
American plate (Riddihough,1984). The 15-20 km wide Nootka fault zone (Hyndman et al.,1979) separates the Explorer subplate from the Juan de Fuca plate (Figures 1 and 3).
The southwest boundary of the Explorer subplate is the Sovanco fracture zone. Cowan et al. (1986), using SEABEAM bathymetry data, interpreted the Sovanco fracture zone to be a 15 km wide zone of right-lateral shear. Since the Explorer subplate has low absolute
{
velocity relative to the overriding plate, the north stepping jumps of the Explorer ridge
{
and the right lateral shear of the Sovanco fracture zone must be related to the motion of l
1 the adjoining Pacific plate. Hyndman et al. (1979) suggest that the Nootka fault zone has been pushed northwestward along the continental margin, in response to the broad, northward movement of the nearby oceanic plate system.
NORTH SOUTil COMPRESSION IN THE OVERRIDING PLATE Focal mechanisms 16-18 (Figure 3, Table 1) are for the three largest of six shallow
(
13
v earthquakes in the region of Vancouver 1. for which Rogers (1979) determined focal mech.
1 r
anism solutions. While these three are strih slip earthquakes (the largest is the Ms 7.3 event of 1946), two of the smaller orthquakes are thrust events (Rogers,1979). The average trend of P-axa m all these earthquakes, and for focal mechanism 19 (Figures r
3 and 4; Table 1), is just east of north. Thus, in the Vancouver I. region, aarthquake i
focal mechanisms require northward compression that extends well east of the Explorer subplate.
.I l
North-trending compression also is the dominant stress in the shallow crust of much of Washington State. Specific data points are shown by convergent arrows in Figure 3, beginning east of Vancouver I., continuing through Puget Sound, across the Cascades and southwards past the Washington Oregon border. In a comprehensive study of Puget i
i Sound focal mechanisms, Yelin (1982) finds that 19 out of 21 reliable solutions for crustal earthquakes occurring during 1976-1981 had P axes trending N S or slightly east of north.
(
The five largest of the earthquakes studied by Yelin (1982) had magnitudes in the range 4.0-4.6.
In the region of the Hanford Site, southern Washingto.1, Kim and hicCabe's (1984) hydraulic fracturing data indicate north south compression,with ratios of maximum compression to vertical stress in the range 2.12.7. The location of this data point is i
the easternmost compression symbol of Figure 3. These high horizontal compression to
(
lithostat ratios suggest a tectonic origin of these stresses. Bued on data from a USGS seismic network operated at the Hanford Site, Malone et al. (1975) found that typical earthquakes were very shallow and had thrust faulting focal mechanisms, with N trending r
P axes. M. Pitt (personal communication,1986) noted that one Hanford area earthquake l
with a N trending P nxis had a focal depth of 28 km. These data on north directed l
compression in much of Vancouver and Washington are unlike that expected from active l
subduction of either the Explorer subplate or the Juan de Fuca plate. Similarly, this northerly compression is unlike that associable with the local southwest rnotion of the 4
I (
North American plate (NE-SW compression is found adjacent to the San Andreas fault 14 i
l i
t i
i and largely is attributed to the southwest motion of the No'rth American plate lZoback et al.,1987]).
MODELING OF STRESS TRAJECTORIES The Pacific plate is being driven into the Gorda block and the Juan de Fuca plate (at the hiendocino and Blanco fracture zones), causing northernly compression in the offshore plate system. Strong coupling at the interface contact between the Cascadian oceanic plate system and the overriding plate should allow offshore compression to be transferred into I
the adjacent overriding plate. This general concept is tested by a comparison of the stress trajectories resulting from a discrete element modeling (h!UDEC code) of the offshore plate I
systern with the axes of compression observed for earthquakes there.
1 This discrete element application models the compre sion in the offshore plate system, due to the Pacific plate's motion, relative to fixed Juan de Fuca and North American j
plates. The solid continuum elements of the discrete element mesh are shown in Figure
]
Sa. The Gorda, Juan de Fuca, and Explorer ridges are defined to be free of traction, the blendocino and Blanco fracture zones are velocity boundarles, and the offshore and overriding plate system are taken to be elastically deformable. The boundary between the offshore plate system and the overriding plate is the inferred mid line of a locked interface 4
thrust zone. The latter boundary is modeled as a joint with shear strength one half that of the surrounding plates. The components of motion of the Pacific plate into the hiendocino and Blanco fracture zones are used to calculate stresses in the offshore plate system. The resulting stress state is then used as a starting model for another motion input from the i
Pacific plate. This process conserves linearity with the failure criteria of the htUDEC code 1
l and is repested until a coherent pleture of stresses emerges, as shown in Figure 5b.
The north trending lines in Figure 5b chart the trajectories of maximum compressive i
j stress, and the approximately east trending lines chart the trajectories ofleast compressive
{
stress. The offshore trajectories of maximum compressive stress are strongly influenced by l
the orientation of the overriding plate. These trajectories tend to parallel the contact
'\\
15 i
e
(
between the ocean'ic and centinental lithospheres at Orego'n and Washington, but at the northwest trending contact of northern Washington and Vancouver Island the trajectories cross this contact into the overriding plate. Thus the offshore system acts like a element j
of shear relative to the overriding plate, until blocked by the westward shift of the plate 1
contact at northern Washington and Vancouver Island.
Comparison of the trajectories of maximum compression in Figure 5b closely match the orientations of P-axes for earthquake focal mechanisms. The calculated trajectories for the offshore plate system are superposed on the seismicity and focal mechanism niap in Figure 6. The pocket of prominent least compressive stress above the Blanco fracture
)
zone corresponds to the observed extension discussed earlier for the Blanco fracture zone.
Also, results of this modeling are consistent with the observed north trending compression in the overriding plate. The success of this simple modeling confirms that the motion of the Pacific plate causes the northerly compression in the offshore Cascadian plate system I
and that this compression is strongly coupled into the overriding plate. This coupling l
produces the crustalnorth tre.nding compression observed throughout much of Washington f
and. Vancouver Island.
STRESSES DUE TO THE SUBDUCTED JUAN de FUCA PLATE i
1 Vertical sections of selsmicity at northern Washington indicate subducted oceanic lithosphere to depths of at least 80 km (Crosson,1983; Taber and Smith,1985). Because volcanic arcs at subduction zones are usociated with plate subducted to depths of about 100 km (Isacks and Bararangi,1977; Gill,1981), the presence of active Cascade volcanoes, from landward of mid-Vancouver (htcager hit, and hits. Cayley and Garibaldi) to east i
l of Cape hiendocino (Lassen Peak), Indicates that subducted plate extends at least to that depth throughout Cucadia. An inversion of teleselsmic P wave delays at seismic 1
i stations in Wuhington and northern Oregon (hilchaelson and Weaver,1986) indicates j
that the Juan de Fuca plate hu subducted to depths of 200 300 km. The P wave travel-
[
time studies of Solomon and Butler (1974) and hilchaelson and Weaver (1966) show a
16
)
1 L
-)
the subducted pla'te's P wave velocity, compared with the' surrounding mantle's P-wave velocity, is highest beneath Washington, Velocity-density systematics suggest that greater plate density is associated.with plate of higher P-wave velocity. Seismicity within the subducted Juan de Fuca plate is most active for the plate section of h!gher P-wave velocity and presumed higher density.
The primary cause of motions of subducting plates is the slab pull force and thus 4
the slab pull force is the primary source of stress at interface thrust zones and the main cause of the great earthquakes that occur there (Spence,1987). The absence of modern interface thrust carthquakes of any magnitude at Cascadia (Taber and Smith,1985; Heaton and Hartzell,1986) has several possible explanations: (1) there is a quiescence anomaly (possibly preceding a significant interface thrust earthquake), (2) present day subduction is aseismic, or (3) the stresses acting at the interface thrust zone presectly are insufficient I
to produce interface thrust earthquakes.
Evidence for slab pullforces in the subducted Juan de Fuca plate I
(
Figure 7 is a cross section of seismicity of western Washingtun, based on results from a high quality, regional seismograph network (Taber and Smith,1985) Focal mechanisms for the shallow seismicity beneath Puget Sound reflect N S compression (Crosson,1972; Crosson,1983; Yelin,1982). The deeper, east dipping trend of seismicity is within the subducted Juan de Fuca plate. The dip of the probthle interface thrust zone is about 11'E (Taber and Smith,1985) and the dip of the deeper plate is 20 - 45'E (Taber and f
Smith,1985; Michaelson and Weaver,1986; Weaver and Baker,1987). The dip increase corresponds to the slab bend feature observed for most subduction zones. For typical I
subduction zones, great interface thrust earthquakes generally nue.eate just updip of the slab bend and the associated ruptures then then propagate updip and laterally (Ruff and Kanamori,1983). Downdip from the slab hend, earthquakes generally occur within the subducted plate. Focal mechanisms for earthquakes in the 11'E dipping Juan de Fuca plate show an average tension axis that is downdip; no mechanisms exist that indicate a
17 l
I 4
1
~
thrusting at the interface thrust zone (Taber and Smith,1985), which is probably near the hachured zone in Figure 7. It appears that the slab pull force has been transmitte1 updip past the slab bend to the zone beneath the locked interface thrust zone, giving the downdip tension axes observed by Taber and Smith (1985).
The earthquakes of 1949, (m6 7.0),1965 (m6 6.5), and 1976 (m6 5.2) (earthquakes 13, 14, and 15 of Table 1 and Figure 3) occurred below the slab bend, downdip of the 11'E-dipping seismicity observed by Taber and Smith (1985) and Weaver and Baker (1988).
[
]
These large extensional earthquakes, with downdip T-axes, re3ect the slab pull force of more deeply subducted plate, and are interpreted as due to plate extension, as the sinking of the Juan de Fuca plate is resisted at a locked interface thrust zone. Rogers (1983)
{
notes other significant earthquakes probably below the slab bend to have occurred east I
t of Victoria, Vancouver in 1909 (magnitude about 6) and at southern Puget Sound in
{
1946 (magnitude 6.3; Figure 3). Such earthquakes often are observed in other subduction zones (Isacks and Molnar,1971; Fujita and Kanamori,1981), and subsets of these normal-j j
faulting earthquakes appear to be indicative of forthcoming great subduction earthquakes (Spence,19S7; Dawwska at aj.,1988). However, at a final stage of subduction it is possible 4
that slab pull stresses may be sufficient to cause in plate normal faulting earthquakes, but insufficient to overcome resistance at the interface thrust zone.
Velocity density systematics and slab pull at Cascadia l
Indications of the shape and velocity structure of the subducted Juan de Fuca plate, between 45 - 49'N, have been obtained from Inversion of teleseismic P waves that have traversed that plate (Michaelson and Weaver,1980). They suggest that the subducted Juan l
de Fuca plate is comprised of three sections, extending to depths of 200 300 km. Between i
l j
48.5-49.0* is a small, steeply dipping section of plate. The central section of plate extends I
southwards to about 47'N, dips 30-45' and its P wave velocity is significantly higher than a
that of the surrounding mantle. The third section section extends southwards from 47*N, j
dips steeply, and has a P wave velocity that is only slightly greater than the surrounding i
18
)
i
mantle. These results on the character of the subducted plate are not well-resolved (C.
Weaver, personal communication,1987).
Velocity-density systematics, such as Birch's Law, imply that the greater the plate's seismic velocity contrast with the stirrounding mantle, the greater the plate's density con-trast and thus the greater the plate's slab pull force. At Cascadia, the subducted central segment by far has the greatest amount of seismicity. This high level of seismicity is con-sistent with the moderate density contrast observed for that segment and the condition of a locked interface thrust zone. Conversely, the condition of very few earthquakes in the plate subducted beneath Oregon (Weaver and Baker,1998) is consistent with the low I
3 density contrast observed for the most southern segment, and a small slab pull force there.
The variation of plate densities (and seismicity rates) for the segments of plate sub-ducted beneath Washington and Oregon may be explained by the recent plate motion history of the Juan de*Fuca plate (Riddihough,1984). For the about the last 3 m.y. the
(
absolute and relative poles of rotation of the Juan de Fuca plate have had rapid northward j
migratlons-(the absolute pole had moved into northern California by-0.5 m.y. ago). The i
sicnvest rate of penetration of the Juan de Fuca plate into the mantle is for the southern l
segment (Figure 1), giving it the greatest time to be warmed by the mantle, with a corre-sponding decrease in plate density. This decreasing density then leads to a corresponding t
decrease of the slab pull force. The consequent slowing of subduction perpetuates yet l
1 further slowing of subduction. This process may ev'olve to the point such that the plate j
driving forces are insufficient to overcome the resistance to subduction at the interface l
thrust zone. The general lack of subduction related seismicity at the southern segment Implies that this circumstance may have occurred. If the subduction rate of the southern segment has slowed to near zern, then this would greatly slow the subduction rate for the I
1 remainder of Cascadia.
The boundary between the Cascadian central and southern segments is near 47"N
_ (
(Michaelson and Weaver,1980), north of the St. Helens seismic zone. Analogous to the i
19 i
i 4
increase in dip frorn the central to the southern seg'ments, a i:hange in dip between adjacent plate segments has been observed in southern Peru (Hasegawa and Sacks,1981; Schneider i
and Sacks,1987). Detailed seismicity indicates that the Nazca plate was contorted and f
j t
1 stretched, but not torn. Similarly the seismicity results of Weaver and Baker (1988) 7 l
Indicate that the contortions of the subducted Juan de Fuca plate are not discontinuous.
j The 1949, Als 7.0, left lateral, strike-slip earthquake is near the boundary of the southern I
and central segments of the Juan de Fuca plate (Baker and Langston,1987). The 54 km i
l-focal depth for this earthquake clearly places it within the subducting plate (earthquake I
j 13 of Figure 3 and Table 1). The T axis trends southeast and dips parallel to the downdip trend of local seismicity (Weaver and Baker,1988). This suggests that the 1949 earthquake
[
is caused by the slab pull force, almilar to most other earthquakes in this depth range, I
I j
worldwide. Baker and Langston (1987) interpreted the left lateral, strike-slip faulting of the 1949 earthquake as evidence for downdip motion of the southern segment. Differential sinking of the southern segment beneath a locked interface may be consistent with the l
(
l t
j lack of seismicity at the subduction interface and with the St. Helens seismic zone, where i
4 i
shallow seismicity in the overriding plate indicates compression there (focal mechanism 12). However, there is virtually no independent evidence for differential downdip motion of the southern segment. An alternate explanation for the 1949 earthquake appeals to the f
greater density of the central segment. If subduction at Cascadia nearly is stopped, then i
J l
preferential seaward sinking of the more dense central segment could produce a couple i
near the 1949 earthquake that results in a left lateral shear.
l j
The central segment, beneath the State of Washington, has experienced significant I
subduction in the last few million years. The relatively strong slab pull force beneath i
Washington would set the subduction vector at N50*E observed for 0.5 m.y. ago, and
)
i earlier. The presently acting slab pull force is great enough to cause in. plate seismicity 1
there. If a significant interface thrust earthquake is possible at Cascadia, the most likely
(
rupture sone would be in the sector betweer,47'N 49'N. The slowing of subduction of
)
l i
i
)
the Gorda-Juan de Fuca-Explorer plate system over the las't 2.5 m.y., interpreted in terms
(
of the ridge push and slab pull forces, indicates that if subduction earthquakes can occur at the central segment, that they are becoming increasingly rare.
Slab pull as the primary cause of tilt and strain at subduction zones
]
Crustal strain near Seattle, determined from geodetic measurements, shows low but i
t j
significant compression in the direction N71'E (Savage et al.,1981). Because the theoret-t leal convergence direction is about N50*E (Riddihough,1984) the observed compression i
H is interpreted as due to the subducted plate's motion being resisted at the locked inter-l face thrust zone. Contraction parallel to the direction of plate convergence commonly is i
observed prior to subduction earthquakes along the Pacific coast of Japan (Shimazaki, 1
1974),
l At the coast of Washington and Oregon, there are many observations of landward 1
tilt. Precise leveling over a 70 year period shows a landward tilt of western Washington l
a (coastal uplift accompanied by subsidence in the region of the Puget Sound depression)
I
[Ando and Balazs,1979). Ando and Balazs (1979) combined the observations of landward tilt in Japan being associated with the coselsmic phase of subduction and lack of interface thrust earthquakes at Cascadia, to interpret the Washington tilt data as reflecting aselsmic subduction. Reilinger and Adams (1982) used leveling routes that extended south through i
j Oregon and found landward tilt for the entire zone. These short term rates of tilt substan-i tlally are the same as rates of tilt over the lut 100,000 500,000 years, Indicated by uplift of
]
contal marine terraces (Rellinger and Adams,1982; Adams,1984). The pervasive uplift
)
of the Washington Oregon coast appears to imply aseismic subduction,in contradiction I
j to the locked subduction interface implied by the crustal shortening near the continental 4
l margin.
1 Oceanward tilting of the Japanese coast (coastal depression)is observed both before 4
j and after great interface, thrust earthquakes. Exactly the opposite sense of tilting ae-
)(
companies the rupture of an interface thrust earthquake (Shimazaki,1974). The elastic 4
21 i
1
)
d
coupling of the subducting plate to the overriding plate (whi h produces a ' drag') causes c
(-
the preseismic and postseismic oceanward tilting, whereas rebound of the overriding plate causes the coseismic landward tilting (coastal uplift). Partially successful modelings of these observations of tilt have been achieved by Savage (1983) and Thatcher and Rundle (1984), who assumed steady state preseismic and postseismic slip of the subducted plate along a shallow, subduction interface with a fixed dip of 30'. The results of such modeling for Japanese earthquakes possibly could be improved by incorporating the more realistic j
dip of about 10' for the interface thrust zone and plate dip of about 50' just downdip l
of the lower tip of the locked Interface (He.segawa et al.,1978; Yoshil,1979; Kawakatsu and Seno,1983) and by including some vertical deformation just downdip of the slab bend (Kato,1979). Kato's (1979) results not only fit deformation above the laterface thrust rone but also fit the vertical deformation observed inland. The slab pull load applied at the downdip tip of an interface thrust zone will have one component delivered through the stress-guide of subducted plate, causing localized compression at a locked interface
(
thrust zone. Another portion of the slab pull load will produce a downward moment on the oceanic plate at the slab bend, causing depression centered above the slab bend.
6 If the slab pull force, acting through the slah bend, is the primary cause of vertical deformation in the region of a locked interface thrust zone, then the distance from the slab bend to the oceanic trench physically is more meaningful than the distance from the oceanic trench to the coast. Figure 8 chows six well defined seismicity cross sections, aligned at the downdip tips of the interface thrust zones (equivalently, at the updip ends J
of the slab bends). The slab pull force, Tse,is indicated in Figure 8, with the component parallel to the plate, Fg, and the component perpendicular to the plate's surface, Fu.
The fu is guided updip in the plate and causes a drag on the overriding plate. The fu l
tends to make the subducted plate propagate seaward and,in the case of Casdadia, helps 3
the subducted Juan <
E<
plate retain its dip even though the North American plate is J
moving southwestward.
22 I
The numbers on four of the sections are beneath the corresponding coasts. This shows k
that the coasts of Peru and NE Japan are landward of their slab bends, whereas the coasts of Cascadia and C. Chile are seaward of their slab bends. Spence (1987) noted that the seaward propagation of slab bends and the downdip mass transfer that occurs at slab bends should lead to surface depressions above those features. At Olympic Mountains section of Cascadia, the slab bend is at the transition from the shallow,11*E-dipping thrust (Figure
- 7) to the deeper,30 - 45'E-dipping plate. South of the Olympic Mountains the slab bend is smoothly shifted westward and then assumes a southerly trend (Weaver and Baker, l
1958). The Puget Sound depression (where Ando and Balasz [1979} observed subsidence of 12 mm/yr) and its Oregon exte ision, the Willamette Valley, lie directly above the slab bend of the subducted plate. The Central Valley of C. Chile is above the slab bend there (Kadinsky Cade,1985). When slab bends occur offshore, forearc basins actively davelop.
The Java outer are and Lombok basins (Itamilton,1979) are above the corresponding slab bend (profile 5) and the subsiding deep-sea terrace off NE Japan (von Huene et al.,1978)
(
is above the slab bend there (profile 3). The Lima basin, offshore of central Peru,is above
,i the local slab bend (Langer and Spence, in press), and has subsided >1100 m within the last 0.93-C.98 m.y. (Kulm et al.,1981). The long term subsidence rate for Puget Sound is similar to that of the Lima basin. Of the seven sections in Figure 8, the greatest downdip extent of interface thrust zone is that beneath the Olympic Mountains.
Figure 91s a schematic showing a downward bending moment applied to the subducted Juan de Fuca plate at the downd'p end of the interface thrust zone beneath the Olympics.
This vertical moment would be analogous to the downward bending moment that causes plate bending beneath an oceanic trench (Chapple and Forsyth,1979; Caldwell et al.,
l 1976). Updip of a slab bend, analogous to the rise occurring seaward of an oceanic trench, the bending plate should have an upward flexure. In Figure 9, the upward flexure of this 1
bending beam is scaled to correspond to the observed vertical deformation at western
[
Washington This model qualitatively explains the depression at the Puget Sound region i
23 i
and the uplift at the Wuhington coast. Thus the coastal up' lift at Cascadia is ar alogous to k
the preseismic or postseismic coastal depression at Japan, and is consistent with a locked interface thrust zone at Cascadia.
DISCUSSION A major theme of this paper is to provide seismological evidence to show that recent subduction of the Juan de Fuca plate is anomalous. A continuation of the dominance of tectonics at Cascadia by the Pacific plate may lead to increased earthquake activity within the horizontal Juan de Fuca plate, and ultimatly, further fragmentation and northward motion of these plate fragments. A more complete and systematic survey of stresses in
< nd offshore of Oregon and Washington than has been achieved is important to better define the regional stress field, and thus permit more complete modeling of the origins of the regional stresses.
Seismic or aseismic subduction at Cascadia?
\\
Figure 2 indicates that seismic coupling at an interface thrust zone is inversely related to the age of the subducti.ig oceanic plate. Because the Juan de Fuca plate is one of the youngest or.canic plates at a subduction zone it is implied that the seismic coupling a' the Cucadia subductior, zone is very gres.t (Ruff and Sanamori,19S0; Spence,1987). Iloweve,
Kanamori and Astiz (1985), using data from subduction of young oceanic lithosphere at Mexico, tentatively concluded that the majority of slip at the Cascadia subduction zone is ueismic. The observation that the Gorda block hu ceued subduction indicates that the local plate-driving forces are insufficient to cause subduction there. This lack of subduction (I the Gorda block thus is a regional analog to contradict the hypothesis that subduction at the remainder of Cascadia is occurring aseismically. This conclusion is supported by earthquakes at the St. lielens seistnic zone (in the crust of the overriding plate), indicating that the local interface thrust zone is locked (Weaver and Smith,1983). The normal.
g faulting earthquakes within the subducted Juan de Fuca plate (Table 1; Taber and Smith, 24 j
~.
l I
j i
f 1985) also suggest that the shallow subduction interface l's locked (Spence,1987), Also i
4 l
(^
j indicative of a locked interface thrust zone are strain neasurements at western Washington i
and southwestern British Columbia that show crustal shortening about parallel to the j
theoretical direction of plate convergence, (Savage et al.,1981; Lisowski et al.,1987) and
[
]
observed shortening across the coast of Oregon (Adams,1984).
Subduction that is nearly aseismic, with only small to moderate interface thrust earth.
2 l
quakes, usually is associated with very old oceanic lithosphere. A r,ood example exists for l
1 i
the eastern Sunda arc, where subducting oceanic lithosphere is about 145 m.y. old. The l
great slab pull force of this very old plate largely has decoupled the local interface thrust f
I sone and much of this slab pull force is transmitted updip to cause normal faulting earth-
]
quakes near the eastern Sunda trench (Spence,1986a). If the interface thrust zone at i
l j
Cascadia were decoupled, then we should observe normal faulting earthquakes near the l
Washington Oregon trench. No such earthquakes are known. It is concluded that aseismic subduction is not significant at the Cascadia subduction zone and that theinterface thrust i(
l zone essentially is locked.
l Have Holocene earthquakes xcurred at Cascadia?
3 l
j No seismological evidence exists that requires subduction to be active at Cascadia's t
j subduction interface. It is clear that the subduction interface presently is locked, if great l
9 subduction earthquakes have occurred periodically in the Holocene, evidence for this must i
be from geological data. Such evidence must show, beyond a reasonable doubt, causality with an earthquake origin or the debate of attive subduction at Cascadia will remain f
I
{
unresolved.
i
}
i q
The works of Atwater (1M7) and Atwater et al. (1988) show that, in the last 5000 i
l years, there have been eight sudden burials of coastal vegetation at western Washington.
I I
I i
These studies suggests that coseismic subsidence due to great subduction zone earthquakes, 1
and seem to be able to discount other possible explanations such as major storm surges or tsunamis from distant sources. The finding of a regional coherence of burial events and a 3
j 25 I
1
O t
systematic timing'of these burials would help substantiate' Atwaters's hypothesis. Other
(
~
i information on possible subduction events in the Holocene may result from detailed studies of the uplifted marine terraces at the Washington Oregon coast and the rates and timing 1
of uplift and depression at Puget Sound it would be further suggestive if uplifts in Puget I
Sound corresponded in time to the inferred coastal depressions.
Hypothetical consequences of the anomalous subduction at Cascadia
(
)
The stopping of subduction of the Explorer subplate and Gorda block probably are
]
due to the buoyancy of these very young units and the consequent high coupling at their t
]
subduction interfaces. That is, the resistance at the subduction interface is greater than I
the ridge push and slab pull forces can overcome. Subduction will be most persistent with
(
the plate element with the highest density contrast, apparently beneath Washington at i
about 47 - 49'N. Subduction beneath Oregon may be very slow er even have ceased (the f
l absolute pole of the Juan de Fuca plate at 500,000 years ago was in northern California).
Th:refore the stronger slab pull force beneath Washington would tend to torque the Juan
\\'
de Fuca plate clockwise. If the slab pull force becomes insuRicient to accomplish this, then 3
1 subduction bereath Washington also would tend to slow to a halt.
i l
l The northerly compression that exists throughout much of Cascadia will be super-t 1
posed on the stresses resulting from the slab pull force. During an interface thrust earth-l quake, when the interface momentarily is decoupled, it is likely that the dip-slip motion I
1 will be accompanied by a right lateral translation of the offshore plate system. Because i
j subduction primarily is occurring beneath Washington, such a right lateral offset would 1'
be most pronounced at the subduction interface there. Conversely, because subduction beneath Oregon is much slower the right-lateral deformation would tend to be distributed j
throughout the plates that are offshore of Oregon. Such a repeated scenario could lead to t
the further loading of northerly compresssion at northern Washington and Vancouver.
i Because the North American plate is moving southwestward at about 23 km/my and I
j it is implied that when the subduction i *rface is lo^ked, then the subducted and offshore i
l t
26 I
l l
i i
\\
r components of Ju'an de Fuca plate system would accompany the motion of the North American plate. Riddihough (1984) estimates that the area cf the Juan de Fuca plate 1
has decreased by about 50% in the last 7 m.y. Ilowever, the Juan de Fuca ridge now is l
moving westward in the absolute / hot spot reference frame at about the same speed as the North American plate (R. Riddihough, written communication,1987). The recent development of ridge jumps west of the Explorer subplate and the tectonic characteristics of the Gorda block indicate that these units have a component of motion with the North American plate. Some seaward propagation of the Juan de Fuca spreading centers is required to explain the pseudofaults and pattern of magnetic anomalies there (licy.1977; i
j Wilcon et al.,1984). Such translation of the Juan de Fuca plate can occur witheut much j
southwestward compression because the Pacific plate is pulling away from the west side of the Cascadian spreading centers, j
The subducted plate, acting under the slab pull force, sinks and propagates seaward
[
(Garfunkel et al.,1986; Spence,19S7), allowing the subducted plate to maintain its dip
{
relative to the advancing North American plate. If the subducted plate were sinking seaward futer than the rate of abance of the North American plate (with the volcanic arc accompanying the westward propagation of the subducted plate), then suHicient conditions exist for the opening of a back are buin. Carlson and liart (1087) describe how mantle j
flow, resulting from rapidly changing convergence rates of the Farallon plate system our the lut is my, may have led to the volcanic flows of the Oregon Plateau and also suggest
{
possibly similar histories for the Buln and Range province an ' the Colombia Plateau. The a
j young age of the subducting lithosphere at the Cascadia subduction zone makes it unlike
)
typleal subduction zones that exhibit back arc spreading (Uyeda and Kanamori,1979; I
Garfunkel et al.,1986) (see Figure 2]. Carlson and Hart (1987) summarize similarities and differences between the Oregon Plateau and back are basins of the Western P#cific.
4 It appears that the young age of Farallon plate subducted beneath Cascadia hu led to a form of back arc development, but less well developed than associated with the subduction l
27 I
J l
I
O of much older ocea'nic lithospheres in the Western Pacific.
~
The Lake Chelan earthquake of Decembe-14,1872 liistorically, this may be the largest earthquake to occur in Washington and Oregon, and it is a key earthquake in seismic hazard studies for the U. S. Pacific Northwest (liopper et al.,1975). The primary aftershocks of the Lake Chelan earthquake continued for about 21/2 years and often caused aquifer disturbances at Lake Chelan, implying a shallow focal depth for these earthquakes. Rasmussen's (1967) seismic history of Washington indicates numerous felt earthquakes near Lake Che!an. The shallow depth implied for the Lake Chelan earthquake difiers from suggested focal depths of about 60 km (hf alone and Bor,1979) and of 100-150 km (hfichaelson and Weaver,1986), who assumed that the 1872 earthquake occurred in the subducted Juan de Fuca plate. Neither the 1949 nor the 1965 earthquakes, the largest known to occur within the subducted Juan de Fuca plate, had extensive aftershock series (Algtrmissen and liarding,1905) and, moreover, no
[
microcarthquakes are known in the subducted Joan de Fuca plate with depths as great as 100 km (Crosson,1983; Taber and Smith,1985). These data, combined with analysis of the intensity attenuation pattern for the main shock, indicate a shallow, crustal focal i
depth (llopper et al.,1988).
j Given that the focal mechanisms of shallow earthquakes (mostly strike slip) through-out much of Vancouver and Washington have N trending compression axes and given the l
l high, shallow horizontal stresses in central Washington, also with N trending compression i
2 axes, it is reasonable that the shallow Lake Chelan earthquake occurred on a fault that was actived by this regional north trending compression.
I CONCLUSIONS The collision of the northwestward moving Pacific plate with the blendocino and Blanco fracture zones causes compression northwards to Vancouver Island. The north-ward compression in the offshore plate system is coupled into the overriding plate, causing 28
northward compression there. The observed high levels of' northerly compression indicate
(
i significant rist. from large earthquakes in both the offshore and overriding plates. Addi-A tional sources of lithospheric stress at the Cascadia subduction zone are the slab pull force of subducted Juan de Fuca plate, the relatively minor ' ridge push' force of the horizontal j
Juan de Fuca, plate, and force due to the southwestward motion of the overriding plate, The stress at any part of the Cascadia subduction zone is a superposition of stresses from a
t i
these sources. Speel5e regions around the Cascadia subduction zone have distinct modes i
of deformation, usually because of the proximity of the region to one of the primary stress sources or to a major resistance to plate movement.
/
The slab pull force, due to the excess mass of already subducted Juan de Fuca plate,
(
)
{
is the primary cause of subduction processes at the Cascadia subduction zone. This force causes platt extension, reflected by earthquakes with downdia T axes. Bc:ause plate motion is resisted at the interface thrust zone, the slab pull force and the wea'4er ridge push force cause localized compression there. The subducting plate dips at about 11'E
(
i 3
for a distance of about 180 220 km, and then the plate dip steepens to 20 - 45'E. The j
j slab pull force acting at this slab bend causa surface depression above the bend (at Puget j
Sound), and an upwarp of plate updip from ths bend, leading to coastal uplift, r
The abrolute subduction rate of the Juan de Fuca plate has slowed at least by 60%
l since 6.5 m.y. ago (Riddihough,1984), and the present absolute subduction rate is un-known. During this slowing, the Gorda block and possibly the Explorer subplate have 1
j ceased subduction. The subduction of the Juan de Fuca plate beneath Oregon has greatly slowed and earthquakes related to the subducted plate virtually are absent there. There 1
are no known thrust faulting earthquakes at the interface thrust :ane of Cascadia and i
j aseismic subduction seems unlikely, it is concluded that the great resistance to subduction j
provided by this interface and the progressive weakening of the slab pull force is leading 1
to a long term cessation of subduction at Cascadia, it is not known whether one or more
{
l great subduction earthquakes can yet occur at Cucadia. if a subduction earthquake can l
(
I 29 i
_ _ =. -
t i
occur at Cascadia,' the most likely location would be in the' mone 47 - 49'N.
i l (.
ACKNOWLEDGEMENTS l
i Mikko Ahola (U. S. Bureau hiines) helped with the discrete element modeling of stresses, using the code MUDEC developed by Itasca Consulting Group, Minneapolis.
[
Russ Needham helped in the determination of the new focal mechanisms. I thank Robin i
i Riddihough, Craig Weaver, and Jim Dewey for their thorough and thoughtful reviews and i
i
)
Bob Embley, Bill Savage, Margaret Ilopper, and Mitch Pitt for discussions.
[
}
I l
I i
i i
i i
j
}
d
(
i 1
\\
[
{
i i
i I
I
)
I t
l i
t l
\\
\\
4 1
)
i t
I I
i i
/
t
\\
l i
1 30 4
i
-l i
f 4
i
REFERENCES Adams, J., Active deformation of the Pacifie Northwest continental margin, Tectonies,3, l
449-472, 1984.
I
]
Algermissen, S. T., and S. T. Harding, The Puget Sound, Washington earthquake of April l
18,1965, U. S. Government Printing Office, Washington, D.C.,51 p.,1965.
f j
Ando, h1, and E. I. Balazs, Geodetic evidence for aseismic subduction of the Juan do Fuca plate, J. Ceophys. Res., 84, 3023 3028, 1979.
[
Atwater, B. F., Evidence for great Holoc<ne earthquakes along the outer cout of Wash-4 ington State, Science, 236,942 944, 1987.
I Atwater, B. F., S. P. Nishenko,1987). A. G. Hull, and J. B. Phipps, Extent and timing of f
i probably coseismic subsidence in Westernmost Washington State during the last 5000 i
t
(
years,.1988.
I Atwater, T. h!. Implications of plate tectonics for the Cenozoic tectonic evolution of western North America, Geol. Soc. Am. Bull., 81, 3513-3536, 1970.
J t
i Atwater, T. hl., and J. D. hfudie, Block faulting on the Gorda rise, Science, 159, 729-731, j
1968.
i i
Baker, G. E. and C. A. Langston, Source parameters of the 1949 magnitude 7.1 South 1
Puget. Sound, Washington, earthquake as determined from long period body waves
[
i 1
]
and strong ground motions, Bull. Seismol. Soc. Am., 77, 1530 1557, 1987.
Bolt, B. A., C. Lomnitz, and T. V hicEvilly, Seismological evidence on the tectonics of i
a j
central and northern California and the hiendocino escarpment,11ull. Seismol. Soc.
1 A m., 58,1725-h767,1968, I
t f
l l
Bratt, S. R., E. A. Bergman, and S. C. Solomon, Thermoelastic stress: How important as
[
l(
i 31 1
l l
i 2
a cause of eaithquakes in young oceanic lithosphere?,'J. Ceophys. Res., 90, 10,249-10,260, 1985.
Byerly, P., The earthquake of July 6,1934: amplitudes and first motion, Bull. Seismol.
Soc. im., 28,1 13,1938.
Caldwell, J. G., W. F. Haxby, D. E. Karis, and D. L. Turcotte, On the applicability of a universal trench profile, Earth Planet. Sci. Lett., 31, 239-246, 1976.
Carlson, R. L., Plate motions, boundary forces, and horizontal temperature gradients:
Implications for the driving mechanism, Tectonophysics, 99, 149-164, 1983.
Carlson, R. W. and W. K. Hart, Crustal genesis on the Colorado plateau, J. Ceophys.
Res., 92, 6191 6206,1987.
Carver, G. A., R. h1. Burke, and H. 51. Kelsey, Late Quaternary tectonics of coastal Cal.
ifornal north of the hiendocine triple junction (abstract), EOS Trans. Am. Ceophys.
(
U., 67,1224,1956.
Chapple, W. hl. and D. L. Forsyth, Earthquakes and bending of plates at trenches, J.
Geophys. Res., 84, 6729-6749, 1979.
Chinnery, bl. A., Secondary faulting 11. Geological aspects Can. J. Earth Sci., 3,1966.
Christensen, D. H. and L. J. Ruff, Outer rise earthquakes and seismic coupling, Ceophys.
Res. Lett., 10, 697 700,1983.
Cockerham, R. S., Evidence for a 180 km long subducted slab beneath northern California, Bull. Seismol. Soc. A m., 74, 569 576, 1984.
Coudert, E. B., B. L. Isacks, bl. Barazangi, R. Lount, R. Caldwell, A. Chen, J. Dubois, G.
Latham, and P. Pontoise, Spatial distribution and mechanisms of earthquakes in the southern New Hebrides are from a temporary land and ocean bottom seismic network and from worldwide observations, J. Ceophys. Res., 86,5905-5925,1981.
32
L Cowan, D. S., bl. Botros, and H. P. Johnson, Bookshelf tectonics: Rotated crustal blocks within the Sovanco fracture zone, Geophys. Res. Lett., 13, 995-998, 1986.
t Crosson, R. S., Small earthquakes, structure, and tectonics of the Puget Sound region, Bull. Seism. Soc. Am., 62, 1133-1171, 1972.
I f
Crosson, R. S., Review of seismicity in the Puget Sound region from 1970 through 1978, l
U.S. Geol. Sur. Open./ile Rept.,8319, Earthquake Hazards of the Puget Sound Re-i gion, Washington, ed by J. C. Yount and R. S. Crosson, 6-18, 1983.
i i
]
Davis, E. E. and R. P. Riddihough, The Winona basin: structure and tectonics, Can. J.
j Earth Sci., lo, 767-788, 1982, l
4 J
l j
Dickinson, W. R. and W. S-Snyder, Geometry of triple junctions related to San Andreas
(
j transform, J. Geophys. Res., 84, 561-572.
1 l
Dmowska, R., J. R. Rice, L. C. Lovison, and D. Josell, Stress transfer and seismic phe-i
[
nomena in coupled subduction zones during the earthquake cycle, J. Ceophys. Res.
I submitted.
i Eaton, J. P., Detailed study of the November 8,1980 Eureka, Calif., earthquake and its i
aftershocks (abstract), EOS Trans. Am. Ceophys. U., 62,959,1981, 5
i l
)
Elvers, D., S. P. Srivastuva, K. Potter, J. htorley, and D. Seidel, Asymmetric spreading across the Juan de Fuca and Gorda Rises as obtained from a detailed magnetic survey, 1
l Earth Planet. Sci. Letf., 20, 211 219, 1973.
Embley, R. W., L. D. Kulm, G. hiassoth, D. Abbott, and bl. Holmes, htorphology, structure I
l and resource potential of the Blanco transform fault zone, in D. W. Scholl, A. Grantz, 5
and J. G. Vedder, eds., Geology and Resource Potential of the Continental Afargin of 1
i Western North America and Adjacent Ocean Basins - Beaufort Sea to Baja California:
{
4 Cirewm Pacific Councilfor Energy and Mineral Resources, Circum-Pacific Council for
!(
Energy and hiineral Resources, Houston, Texas,in press.
1 i
33 i
Forsyth, D., and S', Uyeda, on the relative importance of driving forces of plate motions,
(
Ceophys. J. Roy. Astron. Soc., 43, 163-200, 1975.
Fox, Jr., K. F., R. J. Fleck, G. H. Curtis, and C. E. Meyer, Implications of the northwesterly younger age of the volvanic rocks of west central California, Geol. Soc. Am. Bull.,00, 647 654, 1985.
Fujita, K., and H. Kanamori, Double seismic zones and stresses of intermediate depth earthquakes, Geophys. J. Roy. Astron. Soc., 66, 131-156, 1981.
Garfunkel, A., C. A. Anderson, and G. Schubert, Mantle circulation and the lateral mi.
gration of subducted slabs, J. Ceophys. Res., 91, 7205-7223, 1986.
Gill, J. B. Orogenic andesites and plate lectonies, Springer-Verlag, New York, 390 pp.,
1981.
Gutenberg, B. and C. F. Richter, Seismicity of the Earth and Associated Phenomena, Hafner Pub. Co., New York,310 pp,,1954.
\\
Hamillion, W., Tectonics of the indonesia region, U. S. Geol. Sur. Prof. Pap. 1075,345 pp.,1979.
Hasegawa, A. and I. S. Sacks, Subduction of the Nazca plate beneath Peru as determined i
from seismic observations, J. Geophys. Res., 80,1971 4980, 1981.
Hasegawa, A., N. Umino, and A. Takagi, Double-planed seismicity patterns of the deep seismic zone in the northeastern Japan arc, Tectonophysics, 47,43 58, 197S.
Hauksson, E., Structure of the BenioR zone beneath the Shumagin Islands, Alaska: R e.
location of local earthquakes using three-dimensional ray tracing, J. Ceophys. Res.,
00, 635-649, 19S5.
Heaton, T. H. and H. Kanamori, Seismic potential associated with subduction in the northwestern United States, Bull. Seismol. Soc. Am., 74, 933-944, 1984.
l
\\
34 l
Heaton, T. H. and S. H. Hartiell, Source characteristics of hypothetical earthquakes in
(
northwestern United States, Bull. Seismol. Soc. Am., 76, 675-708, 1986.
q Heaton, T. H. and S. H. Hartzell, Earthquake hazards on the Cascadia subduction zone, Science, 236, 102 168, 1987.
I i
i Hey, R., A new class of"pseudofaults" and their bearing on plate tectonics: A propagating l
1 j
rift model, Earth Planet. Sci. Letti, 37, 321 325, 1977.
Hopper, ht. G., C. J. Langer, W. J. Spence, A. hl. Rogers, and S. T. Algermissen, A study of earthquake losses in the Puget Sound, Washington, area, U. S. Geological Survey Open file Report,75-375,298 pp.,1975.
j Hopper, hi. G., S. T. Algermissen, D. 51. Perkins, S. R. Brockman, and E. P. Arnold, The December 14,1872 earthquake in the Pacific Northwest,, in press.
i i
f l
Hyndman, R. D and D. H. Weichert, Seismicity and rates of relative motion on the plate
[
(
boundaries of vestern North America, Ceophys. J. R. astron. Soc., 72,59 82,1983.
r Hyndman, R. D., R. P. Riddihough, and R. Herzer, The Nootka fault none - a new plate l
I boundary off western Canada, Geophys. J. R. astron. Soc., 58, 667 683, 1979, i
Ibach, D., The structure and tectonics of the Blanco fracture zone, bl.S. thesis, Oregon State U., Corvallis, OR,1981.
f I
Isacks, B. L. and h1. Barazangi, Geometry of Benloff mones: Lateral segmentaticn and downwards tending of the subducted lithosphere,in Isalnd Ares, Deep Sco Trenches, 1
and Back Are Basins, Maurice Euing Ser,1, edited by h1. Talwani and W. C. Pitman i
111, pp. 99114, AGU, Washington, D. C.,1977 l
Isacks, B. L. and P. hiolnar, Distribution of stresses in the descending lithosphere from a j
l 1
l global survey of focal mechanisms of mantle earthquakes, Rev. Ceophys., 9, 103-174,
)
1971.
I(
3s i
i I
f 1
Jachens, R. C. anli A. Griscom, Three-dimensional geometry of the Gorda plate beneath
(
Northern California, J. Geophys. Res., 88, 9375-9392, 1983.
Kadinsky Cade, K., Seismotectoniet of the Chile margin and the 1977 Caucete earthquake of Western Argentina (Ph.D. thesis), Cornell Univ., Ithaca, N.Y.,253 pp,1985.
Kanamori, H. and L. Astiz, The 1983 Akita Oki earthquake (h!w 7.8) and its implict.tions
'or systematics of subduction earthquakes, Earthg. Predict. Res., 3, 305 317, 1985.
Kanamori, H. and K. C. hicNally, Variable rupture mode of the subduction zone along the Ecuador Colombia coast, Bull Seism. Soc. Am., 72, 1241 1253, 1982.
Kato, T., Crustal movements in the Tohoku District, Japan, during the period 1904 1975, and their tectonic Impilcations, Tectonophysics, 60, 141 167, 1979.
Kawakatsu, H. and T. Seno, Triple seismic zone and the regional variation of seismicity along the ne'thma tionsha arc, J. Geophys. Res., 88,4215-4230, 1953.
Kelleher, J. and W. NicCann, Buoyant zones, great earthquakes, and unstable boundaries of subduction, J. Geophys. Res., 81,4885-4896, 1976.
Kim, K., and W. h1 hicCabe, Geomechanics characterization of a proposed nuclear waste repository site in basalt, in Rock Mechanies in Productivity and Protection, Proc.
Tu enty fifth Sym on Rock Mechs, Soc. Niining Engrs.of Am. Inst. hiining, hietallurgy, and Pet. Engrs, New York,112G1135,1984.
l Kulm, L. D., T. hl. Thornburg, H. Schrader, A. h!asias, J. M. Resis, and L. Johnson, Late Cenozoic carbonates on the Peru continental margin: Lithostratigraphy, biostratigra-phy, and tectonic history,in Nazca Plate: Crustal Formation and Andean Convergence, GSA h!em.154, ed. by L. D. Kulm et al., Geol. Soc. Am., Boulder, CO, 469 507, 1981.
Lajoie, J. R., G. L. Kennedy, S. A. hiathieson, A. ht. Sarna Wojicki, S. A. hiorrison, and bl. K. Tobish, Emergent Holocene marine terraces at Cape hiendocino and Ven-f 36
A turs, Ccilfernin, U.S.A., Proc. Internatienal Symposidn en Devel pment of Holocene i
(
Shorelines, Tokyo, Japan,1983.
1 Langer, C. J. and W. Spence, Deformation of the hasca plate related to the gap-filling earthquake series of October November,1974, J. Ceophys. Res.,, submitted, i
1 Lisowski, bl., J. C. Savage, W. H. Prescott, and H. Dragert, Strain accumulation along the Cascadia subduction zone in western Washington, Eos Trarts. Am. Ceophys. U., 68, l
1240,1987.
I I
h!alone, S. D. and S. Bor, Attenuation patterns in the Pacific Northwest and the location i
j of the 1872 North Cascades earthquake, Bull. Seismol. Soc. Am., 69, 531 546, 1979.
[
I h!alone, S. D., G. H. Rothe, and S. W. Smith, Details of microcarthquake. swarms in the i
Colombia Basin, Washington, Bull. Seism. Soc. Am., 65, 855 664, 1975.
i l
I hienard, H. W., Fragmentation of the Farallon plate by pivoting subduction, J. Geol., 80, i
99 110, 1978.
l i
hiendiguren, J. A., Focal mech'anism of a shock in the middle of the Nazca plate, J.
Geophys. Res., 76, 3S61 3879,1971.
I i
hiichaelson, C. A. and C. S. Weaver, Upper mantle structure from teleseismic P wave I
j arrivals in Washington and northern Oregon, J. Ceophys. Res., 91, 2077 2094, 1956.
t j
hiilne, W. C., G. C. Rogers, R. P. Riddihough, G. A. hichtechan, and R. D. Hyndman, i
l i
Seismicity of western Canada, Can. J. Earth Sci., 15, 1170-1193, 1978.
[
i i
hiinster, J. B., T. H. Jordan, P. Afolnar, and E. Haines, Numerical modeling of instanta-i neous plate tectonics, Geophys. J. Roy. Astron. Soc., 36, 541 576, 1974..
Nakamura, K. and S. Uyeda, Stress gradient in arc. back are regions and plate subduction.
J. Geophys. Res., 85, 6419-6428, 1980.
37
j 2
Nishimura, C., D~.'S. Wilson, and R. N. Hey, Pole of rotation analysis of present day Juan l
(
de Fuca plate motion, J. Ceophys. Res., 89, 10,283-10,290, 1984.
4 L
Nuttll, O. W., The western Washington earthquake of April 13, 1949, Bull. Seism. Soc.
A m., 42, 21 28,1952.
i Peter, G. and R. Lattimore, Magnetic structure of the Juan de Fuca Gorda Ridge area, i
f J. Ceophys. Res., 74, 586 593, 1969.
l l
i j
. RaR, A. D. and'R. G. Mason, Magnetic survey oR the west coast of North America,40*N i
i latitude to 50*N latitude, Geol. Soc. Am. Bull., 72, 1267 1270, 1961.
i I
i Rumussen, N., Washington State earthquakes,1840 through 1965, Bull. Seism. Soc. Am.,
l 1
i j
57,1967.
1 j
Reilinger, R., and J. Adams, Geodetic evidence for anive landward tilting of the Oregon a
1 and Washington coastal ranges, Ceophys. Res. Lett., 9, 401-403, 195' 1
I I" (
Riddihough, R. P., A model for recent plate interactions oR Canada's west coast, Can. J.
l i
l Earth Sci., 14, 354 396, 1977.
i l
Riddihough, R., Gorda plate motions from magnetic anomaly analysis, Earth Planet. Sci.
l 1
i Lef t., 51,163 170,1980.
t t
l
{
l Riddihough, R., Contemporary movements and tectonics on Canada's West Coast: A dis.
cussion, Tectonophysics, 86,319 341, 1982.
l I
Riddihough, R., Recent movements of the Juan de Fuca plate system, J. Ceophys. Res.,
l a
89, 6950 6994, 1984, i
i Rogers, G. C. Earthquake fault plane solutions near Vancouver Island, Can. J. Earth Sci.,
/
c
{
16, 523-531, 1979.
i f
l Rogers, G. C., Some comments on the seismicity of the northern Puget Sound - southern f
YanCouver Island region, in Earthquake Hazards of the Puget Sound Region, Wash-l(
i 38 3
6
}
f I
9 Insten, U. S. Geol. Sur. Open. File Rpt. 83-19, ed by J. C. Yount and R. S. Crosson, l
i l (
19 30, 1983.
)
'~
Ruff, L. and H. Kanamori, Seismicity and the subduction process, Phys. Earth Planet.
l f
Int., 23, 240-252,1980.
i l
Ruff, L. and H. Kanamori, Seismic coupling and uncoupling at subduction zones. Tr: tono-i physics, 99,99-117, 1983, i
I Schneider, J. F. and I. S. Sacks, Stress in the contorted Nuca plate beneath southern Peru j
from local earthquakes, J. Ceophys. Res., 92, 13,887 13,902, 1987.
1 Sarna Wojcicki, A. 51., C. E. hieyer, and J. L. Slate, Displacement of a ca. 6 Af a tuff j
acron the San Andreu Fault System, northern California (abstract), EOS Trans.
Am. Geophys. U.,,67,1224,1986.
i Savage, J. C., A dislocation model of strain accumulation and relcue at a subduction zone,
(
J. Geophys. Res., 88, 4984 4996, 1933.
i a
j Savage, J. C.,51. Lisowski, and W. H. Prescott, Geodetic strain meuurements in Wuh-l ington, J. Ceophys. Res., 86, 1929 4940, 1981.
l; Scholl, D., Sedimentary sequences in north Pacific trenches,in The Ceology of Continental
't j
Margins, ed. by C. Burk and L. Drake, Springer, New York, 493 504, 1974.
L Seeber, L, h1. Bararangl, and A. Nowroori, hticrocarthquake seismicity and tectonics of
)
contal Northern California, Bull. Seism. Soc. 4m.,60,16691699,1970.
Shimaraki, K., Pre seismic crustal deformation caused by an underthrusting oceanic plate, in Eutern Hokkaldo, Japan, Phys. Eart A Planet. Lett., 8, 148 157, 1974, 1
J j
Silver E. A., Tectonics of the hiendocino triple junction Geol. Soc. Am. Bull., 82,2965-J 2978, 1971.
1 5
3
)
i k 39 i
i 1
l I
i
Simila, G. W., W. h. Peppin, and T. V. hicEvilly, Seismotectonics of the Cape h!endocino,
(
California, area, Geol. Soc. Am. Bull., 80, 1399-1406, 1075.
Singh, S. K., L. Ponce, and S. P. Nishenko, The great Jalisco, hiexico earthquakes of 1932:
subduction of the Rivera plate, Bull. Seismol. Soc. Am., 75, 1301 1313, 1985.
Solomon, S. C. and R. G. Dutler, Prospecting for dead slabs, Earth Planet. Sci. Lett., 21, 421-430,1974.
Spence, W., The 1977 Sumba earthquake series: evidence for slab pull force acting at a subduction zone, J. Geophys. Res., 01,7225 7239,1956a.
1 Spence, W., Oris., ins of stresses at the Cascadia subduction zone, U.S. Pacific Northwest (abstract), EOS Trans. Am. Ceophys. U., 67,1115,1986b.
Spence, W., Slab pull pnd the seismotectonics of subducting lithosphere, Rec. Geophys.,
25, 5 5 69,1957.
(
Stoddard, P. R., A kinematic model for the evolution of the Gorda plate, J. Geophys. Res.,
1 02, 11,524 11,532, 1987 Taber, J. J. and S. W. Smith, Seismicity and focal mechanisms associated with the sub-duction of the Juan de Fuca plate beneath the Olympic Peninsula, Washington, Bull.
{
Seismol. Soc. Am., 75, 237 249, 1985.
l Thatcher, W. and J. B. Rundle, A viscoelutic coupling model for the cyclic deformation due to periodically repeated earthquakes at subduction zones, J. Geophys. Res., 80, 7631 7640, 1954.
l Tobin, D. G. and L. R. Sykes, Seismicity and tectonics of the norther.st Pacific Ocean, J.
Geophys. Res., 73,3S213S45,196S.
Ukawa, bl., Lateral stretching of the Philippine Sea plate subducting along the Nankal-Suruga trough, Tectonies, 1, 543-571, 1982.
k 40
i Uyeds, S. and H. kr.ntm:rl, Back cre cpsning end th2 moda cf subducti:n, J. Ce: phys.
{
Res., 84,10491061,1979.
{
Vine, F. J., hiagnetic anomalies associated with mid ocean ridges, in The History of the l
Earth's Crust, edited by R. A. Phinney, Princeton University Press, Princeton, NJ, 73-89, 1968, f
t von Huene R., and L. D. Kulm, Tectonic summary of Leg 18, in Initial Reports of the Deep Sea Drilling Project,18, U.S. Gov't Printing Office, Wash ngton, D.C.,961976, 1973.
von Huene R., N. Nasu, et al., Japan trench transected on Les 57 Geotimes, 23,1G.21, i
i 1978.
l Walter, S. R., Intermediate focus earthquakes associated with Gorda plate subduction in northern California, Bull. Seism. Soc. Am., 76, 583 558, 1956.
l Weaver, C. S. and G. E. Baker, Geometry of the Juan de Fuca plate beneath Washington j
and northern Oregon from selamicity Bull. Seism. Soc. Am., 78, 264-175, 1968.
Weaver, C. S. and C. A. hiichaelson, Seismicity and volcanism in the Pacific northwest:
i evidence for the segmentation of the Juan de Fuca plate, Ceophys. Res. I.ett.,12,
[
215 218, 1955.
f i
Weaver, C. S. and S. W. Smith, Regional tectonic and earthquake hazard implications of a crustal fault zone in southwestern Washington, J. Geophys. Res., 88,10,371 10,3s3, 1983.
[
Wilson, D. S., A kinematic model for the Gorda deformation zone as a diffuse southern boundary of the Juan de Fuca piste, J. Ccophys. Res., 91, 10,259-10269, 1986.
Wilson, D. S., R N. Hey, and C. Nishimura, Propagation as a mechanism of reorientation of the Juan de Fuca ridge J. Geophys. Res., 89, 9215 9225,1984.
41
i i
4 Yelin, T. S., The Seattle earthquake swarms and Pug'et basin focal mechanisms and 1
(
their tector.lc implications, M.S. Thesis, Univ. of Washington, Seattle,1982.
Yoshii, T., /., detailed cross-section of the deep seismic zone beneath northeastern Honshu, 2
i Japan, Tectonophysics, 55,349 360, 1979.
i i
{
Zoback, h!
L., S. P. Nishenko, R. 51. Richardson, H. S. Hasegawa, and 51. D. Zoback, hild plate stress, deformation, and seismicity, in The Geology of North America, vol.
t 1
l i
hi, The Western North Atlantic Region: Geol. Soc. Am., edited 6y P. R. Vogt and B.
l 3
t E. Tucholke, 297 312, 1967.
i i
i t
1 i
k l
4 i
- i,
(
i I
1, i
i 1
I i
j i
! (-
42 l
4 j
1 6
3 i
t i
l i
i i
FIGURE CAPTIONS l
- 1. Evolution of absolute motions (km/my) of the Juan de Fuca plate over the last 6.5 m.y., determined from detailed analysis of geomagnetic anomalim (Riddihough 1984).
[
l/
This sequence shows the slowing of subduction of the Juan de Fuca plate, the beginnings of l
Independent motions by the Gorda and Explorer subplates, and the recent northwestward f
5 jumps of the northern spreading centers. The absolute relative plate motions for the 0.5
[
i m.y.B.P. frame are the latest plate motions that are resolvable from geomagnetic anomaly l
)
1 data.
I l
l
?
'l
- 2. Characteristic maximum earthquake, Mw, for most subduction zones, plotted as l
i j
functions of convergence rate and age of subducted plate (adapted from Ruff and Kanamori, l
l 1950). Lower right hand corner shows the slowing of convergence of the Juan de Fuca and l
j North American plates; this slowing entirely is due to the 6&c slowing of the absloute i
a I
i motion of the Juan de Fuca plate over the last 6.5 hia. Above upper dashed line are l
I
(
j f
youngest plates, which show strongest coupling at interface thrust nones (due to weak slab j
s j
pull forces and the resistance to bend.iog by the subducting plate) and which are usociated
)
with the seaward advance of overriding plate. Below lower duhed line are oldest plates, l
l i
which show weakest coupling at shallow, interface thrust zones (becauee of strong slab pull l
1 l
t j
forces) and which are usociated with marginal basin developut (Garfenkel et al.,1956).
l l
]
- 3. Cascadia seismicity,1964. June 1956, for earthquakes of magnitude > 5.1. Additional 1
key earthquakes are shown by dates or by focal mechanism solutions for older earthquakes.
1 Focal mechanism solutions correspond to earthquakes in Table 1. For earthquake focal 1
J mechanisms, directions of greatest compressional stress indicated by convergent arrows 4
i for strike. slip earthquakes; directions of least compressional stress indicated by divergent l
]
arrows for normal-faulting earthquakes. Additional data indicating N.S compression in i
I
]
central Washington and eut of Vancouver are shown by arrows that meet (Weaver and Smith,1983; Rogers,1979; Kim and hicCabe,1984) hfost earthquake; deeper than 30 km l
1 i
4j(
43 t
t
. - ~ _.
1 i
}
l are within the bracketed zone. Volcanoes indicated by triangles. Geometry of ridges and
- (
fracture zones, and absolute plate motions for time frame ending 0.5 m.y. ago are from f
j Riddihough (1984).
i
- 4. P wave first motion data for the eight new focal mechanism solutions of this study, t
}
plotted on stereographic projections of the lower focal hemispheres. Nurnber ir. upper left-1 j
hand corner of each frame corresponds to that earthquake in Table 1. Shown are nodal l
j planes with the mies of the x. and y-planes, and the pressure (P), tensional (T), and null l
(B) axes. Except for earthquake 11, these new focal mecharilsms indies.te a general N S l
compression.
- 5. a. Discrete element mesh for the Cascadia plate system; b. Results of discrete ele-j ment modeling. North trending lines describe trajectories of me.ximum compressive stress:
j i
1 t
j east trending lines describe trajectories of minimum compressive stress.
- 6. Heavy, solid lines in offshore area are compressional stress trajectories inferred from i
1 j
j-modeling and provide excellent fit to focal mechanism and tectonic data. Observed stresses
[
t l
sugge.st that the offshore stress pattern results from the Pacific plate's collision with the f
1 j
Mendocino and Blanco fracture zones and the resistance by Vancouver I, to northward 1
movement of the offuhore plate system. Jagged lines and dashed motion trajectories cor.
l respond to the section of the subducted Juan de Fuca plate that is seismically active; this i
motion was not modeled by the discrete element method.
- 7. Cross section of microcarthquakes occurring beneath the Olympic Mts., including
(
)
most Cascadian earthquakes deeper than 30 km, indicated in Figure 3. Larger earthquakes l
j in lower sone have average T axes downdip, and are probably within subducted Juan de c
{
Fuca plate (Taber and Smith,1985),
f l
j
- 8. Cross sections of well resolved plate dips, plotted as distance from downdip end of l
}
interface thrust none (equivalently, near updip end of slab bend), positive to trench. Num.
I i
)'(.
bcrs on profiles correspond to positions of coasts. Distance from slab bend to Washington f
u i
l i
t
t trench is anomalously large. Inset shows slab pull force, /sp, the component parallel to
(
subducted plate, En, and the component normal to subducted slab, F.v. These components l
of the slab pull force cause stress to accumulate at the shallow, interface thrust zone and I
4 i
lead to the strain and vertical deformation observed at the surface. Data sources for pro.
l files are (1) Taber and Smith,1985; Michaelson and Weaver,1986; (2) Coudert al.,1981;
{
(3) Huegawa et sl.,1979; (4) Kad!nsky.Cade (5) Spence,1986; (0) Langer and Spence (in i
i i
press); (7) Hauksson,1985.
l
- 9. A. Observed urtical deformation above subducted Juan de Fuca plate (Rellinger i
and Adams,1982). Surface depression is above slab bend none, whereas surface uplift is
}
above upwardly flexed plate. Plate flexure in (B) is drawn to align with observed vertical
[
deformation. B. Solid line is top surface of Juan de Fuca plate at western Washington.
}
I Bending moment (-M,) due to slab pull force is applied at downdip tip of interface thrust I
mone, causing plate flexure as shown. Tu is the component of the slab pull force that is j
transmitted through the subducted plate but is resisted at the interface contact (see inset f
t
\\
in Figure 8).
j r
h i
h J
l l
i t
f l
45 i
i
e I
I 6.5 Ma 4.5 Ma
\\[
\\
N
~
- N e
_e g
',3 2
\\,/
,/ J,33 f
,t'
,s l 5, l
y
\\
M q
/
\\
1
/ \\,,
i i
f/,*
l,,,s'
, 1,',
i m,
.w-
,A./,
i
/s i
/
/
s $t 3-
\\/
3 l
s
,,e
\\.
\\
I_
A i
J#
p I
\\
e' I
-k I
\\
{
i f
f f.
2.5 Ma 0.5 Ma s
i I
\\
i y,.-s
\\
\\
c4x
<n v s.
- 3. -
en-i uq %,X!. *gj,,,_
j y
a-wk l
uN
- \\
s l /
L '.
i
~ l
,. X,
/.-
Tc' l, ' /sa
~ ~,$f',. ^ " o ss~~
I.(. +y y.. -
t g,,,
_k__
4 a '
+s f
,7
\\
l
\\
r, ~.:,
\\
t
\\
l l
(
12 12 85 9.5 C. Chile e S. Chile 8.2
\\
8.2 NE Japan
\\
9.0 Perue 10 10 pO e
Kuriles Komehotko
'S\\
S N
8.1 Tongo N N
C. Americo 8'8 8.3
\\
\\
Aleuti ns
' 8 8
7,3 g
g b
e i
Jovo 3
Ker adec h*
N 5 e
4 a b.0 o
c r
yus O I:u eBonin 7.8
\\
8 %
e 4.5my g-6 d
7.2 S New Zealand
\\
N Atoska g
T
\\ 9.2 e
g
\\(g
\\
0.5 my @
4-e g
4
'i 0 8 New Hebride\\s s\\
7.9 SW Japane w
7.2 8.6 7.5 e
Corribbean 40 e
Rg escotic s y,xic o,
2 2
7.0 8.2 0
O 16 0 14 0 12 0 100 80 60 40 20 0
Age, m y.
C,p r<
1 i
t
)
4 173 131 178 127 1{$
1{3, 1{1 1]9 t'
_gy
$u\\
L
.t s
qA S0
- 4 ge m agg o
-50 44 3t L.
4 of%
$ $l
<s s
f,.,,
48-e
\\
a\\t,,,
~48 Q,7 CHELAN u--
so 3 ei
.... A g
\\#
N'"
,3 g g
" 4
\\
g g
o 46-
\\
is 10 a N
-46 g
\\
s a
ne g
4' i
t
't
~
p*./.
,I t
c l
).
l
- dj 8
a
~44 i
g i
~
I 42-y 4
j
_-42
[I,.
{
, _..o. i.
40-1 4* e "
40 a
e
,pt.,4 q,
+
gg g
j 38, S44 ancagag saw 38 g
133 ist 129 127 125 123 III W AG M T L'O E S: G 4.6
= 8. l-8.5 C l.6-8.0 C 4.1-8,5 Q 4.0-7.4 b
]
l 3
l(*DL( )
J
l 1
I r
(
l d
]
t l
1 2
3 i
P T
'n;.
- .;;,<\\
\\
\\
y.n
- ...p.
\\
/
\\
i P
i.
l 7
=
l 8
l
=
i j
I 4
,C 4..ga.
Cow..t84)C4 l
. e '.. /
- 8..e,..t he.4 C c u.. g 8 3.c.,
.* /. ' g
,(Qg r
y jq
.j y.i.
q..
N. j
. x.
1
(
1
}
p P
i j
...re..,n.s.....s k
i 6..es..,...s......
10 11 i
19 A
j P
l l
.~.g.
i e \\~ '.y
- /, *.
- i i
, I.
i eO
.s'.
o.
t' Y
i
........e..............
e......<.............,..
1 i
85 gO i
d i
s 1
i 4
l 1
J i
1 a
4
t-(
\\^
\\/
A//
/
E
- X(/
\\
, ' i s,
53 '
/
53
s f, #
s vx
,.i s,,,
4 '
5,
's g
A
,i
$A_
4 1
i j\\ ' /
- gD 4
t i
fMN O
6 8 i
g/\\./
's i
aX1N
,' i i
y2 hh
(
\\
2
'\\
MENOOCINO 5
/
e A
l l
l i
(
F t wn s'
~
6 m
(
iv. i i
. in, iv. i s iv iv, m e
.# \\'
q 50-ti casa
-s0 m'O.. m
,ff3 a
3- )
<cyd,o ' 'I. *' ',*,T' k \\
f"~'
ii. <
-u j f ein
/.
\\//f';#,e'
~
~
h'.h sy
/
8
.s.
..e
!/ h q' \\
-(
v g
\\[
4,\\
/
~
se,'./k,kd.,,
g q
f.
I
^
,s i
l
).
t i
t OI' 4I 8 **' r p
Y/
Y l,._ c (d.
a x,
.0-ra
-.0 i
f..,
- _orp S
S a
4 ss.
.se 133 131 129 127 125 123 121 W AO NIT UD E S: G..e
. s.i-s.s c s.s-e.0 0 s,i-e.s O s.s-7..
cis set C
l i
WEST EAST 01.Y W PIC S CASCADES Otm C O A{-- f-X PMTj0]D 3;),,
eh
""'E,,,,,,,,,,,,,,
- c
,,, r t,...,,.....,, N. em. y,s c g#
[
.% at M~ g.,
30..
83 t:
100 a m i
FIG u CC
)
i l
r k
f
^
\\
l e
i I
DISTANCE (KM) FROM DOWNDIP TIP OF THRUST INTERFACE
-100
-50 0
50 100 150 200 250 0
n
= a..a". 2.fd.N ** * a -
.ek.p u.
- 743 j) 2 gQ A k' a M 50 g o js
( 1)
CASCAOIA Age A
(6)
(2) a a a a S. NEW HEBRIDES ARC
[
Y d 'f "(3) c.
(3)sesa NE JAPAN (TOHOKU) 100ff 4
( 4 ) ---- CHILE (37.5'S)
Fn a
a Fp
( 51 - - -
E. SUNDA ARC 3
(6)eeee C. PERU
[
s a
F
( 7 ) ---
C. ALEUTI AN ARC g
a FORCES DUE TO 15 0 EXCESS MASS I
J FIG 4Cd.
8
'l i
i I
(
(
A' VERTICAL 4
mf DEFORMATION f
-1b0
-5'O
- 'O 100 5
7
-2 I
y
~4" l
B.
I I
NCH
-15 0
-50 0
50 100 COAST 150 200 250 g
l l
I i
i i
i
~
V, #, " t "
l l
I l
<Fiij
_a-
/
p 2
50 t
M
/
/
i
/
/
?
/
a.
Lu
-Mo
/
e, O 100 JUAN de FUCA PLATE 150 i
ptGu ef_
9
(
l
m
-._.--._e.__.-._
m
._____r.
w TASti 1.--Earthquak e date for focal mechselses shown to FI ure 3 S
Extet pionee Strese Asee Tlee latitude Im sttede I
2 P
T Ise.
Lacettee es/ deft)..
UT MeSei t ende
- M
- W St Dp 58 St Dp St Tr PI Tr F1 Seferences 8.
seredecine fr. s.
82/20/83 to 48 N
S.6 (ISC) 40.33 125.82 1
90
-to 98 80 -180 386 7
46 7
g 2.
Needecise f r. a.
09/10/84 03 84 M 6.7 (MOS) 40.50 126.83 278 85 -178 187 88
-5 142 S 233 2
g 3.
E. Corde embetese 18/08/80 80 27 N 7.3 (ISC) 41.1$
124.30 60 90
-3 150 87 -180 IS 2
105 2
g 4.
Cent. Corde 07/06/34 22 48 6.5 48.4 125.4 323 84 -IF2 232 82
-6 to 188 98 i
Syerly (1938) embplete S.
Off Celtforete-08/23/62 39 29 S. 6 48.85 324.3 116 85 tuo 26 90
-5 343 4
78 4
Salt et al. (8 %8)
OreSee her.
6.
Cerde rid 8e 04/88/65 06 33 a 5.4 (ASC) 48.44 127.38 327 65 -887 198 36
-45 196 60 77 16 Tot.te med Sykes (1968) b 6.2 (ISC)' 43.55 127.69 289 05 356 308 86 885 873 6
8) 1 7.
Slance fr. a.
II/03/88 I) 47 Ng l
S.
Stence fr. a.
03/83/85 19 34 M 6.3 (USGS) 43.51 I27.56 302 90 -873 212 83 -360 867 5
77 5
g l
9.
Ceest of OreSon 03/07/63 23 53 S.4 44.88 123.74 318 90 880
-48 90 0
3 0
93 0
Solt et al. (1968) 30.
Cent. Juese de 06/I6/73 84 43 m,S.8 (ASC) 44.98 825.86 300 90 875 30 85 360 345 3 251 3
Fece plate M
5.7. (12 obe) g 88.
Cent. Jwee de 82/88/68 13 09 a 4.6 (ISC) 45.73 127.30 109 84
-19 201
- 78. -174 63 18 158 9
b Fece plate e
e-,
.*Wumer--e-wmwws-w--"w-me, ww,=w www ur mw- - - -
w 1-----
r-'--t'---
t'
+9' e--
y'--y r-w e
e-u--
,g-w-m e
<ww
~w-T
-y-
--m r
~-
.. ~.-
... ~ _.
e C
I i
I i
i i
TAttf. l.-Earthquake date ter f ocet mecheatsee sheen in Figure 3--Cent tamed i
Wodal planee Stress Ames Tier lat t tiede IAegitude i
2 P
T po.
tacettee mo/de/89..
IFT Megat tende
- N
- W St op 51 5t Dr St Tr F4 Tr F1 Referencee 82.
So. Weentegtee 02/44/03 06 09 g S.S 46.42 122.09 SFS 84 -872 84 82
-6 40 10 310 2 weaver and Setth (1983)
- 83. Subdected Jese 04/33/49 38 SS M 7.0 (CR) 4 F.t 122.7 275 45
-10 12 83 -13S 244 36 133 24 Ihattti (8932)I taker de Foca plate (h=S4 he) and Leageton (1987) 34.
Sebdiscred Jaime 04/29/65 IS 28 a 6.5 47.41 822.29 344 72
-76 325 23 -127 274 61 63 26 Algermiseen and b
de Fece plate (h='9 he)
Nerdte.g (t%3)I neecke and Me!aar (I97I) 1 %.
Subdected Jose OS/36/76 08 35 e 5.2 (ISC) 48.92 R23.10 3% 75
-85 137 I6 -808 259 60 68 28 angers (1983),
b de Fue plate
)
(h=60 he) 16.
Teoce wer I.
06/23/46 SF 13 N 7.3 (CR) 49.8 (23.3 228 85
-82 339 78 -IFS 389 82 274 S
Rogere (1979)
IF. T ceveer t.
$2/t6/S7 IF 27 6.0 49.8 324.5 248 78 17 IS4 73 167 21 3
112 20 angere (1979)
{
j
- 18. Meet et 07/03/F2 to I6 98 6.0 (MOS) 49.43 427d8 154 89 163 244 73 8
200 It 308 33 7.ogere (1979) vence oer I.
3 19.
W. Esplever 07/23/72 39 83 M 6.3 (Mos) 50.90 129.29 34 5 90 ITF FS 81 360 30 2 300 2
esbytete g
Iplaternettemet Seteentegical Cuesteetoe yMeece. Observetery g
.S. CeetogIc I Surwey C
Cetemberg and 9tchter (1954) a cJ