GEOLOGY, February 2008 127

ABSTRACTIn 2004, two clusters of earthquakes occurred in the central part

of the Cascadia forearc, which displays several characteristics indica-tive of along-strike and downdip variations in plate coupling. Moment tensor analysis for the main shock in each cluster indicates that both events, which had magnitudes of 4.9 and 4.8, were compatible with low-angle thrust motion on fault planes dipping 6°–15° to the east, consistent with the plate boundary dip of ~12°. By tracing rays through a high-resolution two-dimensional crustal velocity model to match arrival times of secondary arrivals (pP and PmP), we estimate that the source depth was 9–11 km for the M4.9 event and 15–17 km for the M4.8 event. We conclude that these earthquakes probably represent seismogenic thrust motion on the nominally locked or transitional part of the Cascadia megathrust, updip of where episodic tremor and slip (ETS) have been documented. Continued high-resolution observations of seismicity and improved crustal models are needed to confi rm whether apparent tem-poral correlations between ETS and continued seismic activity in these clusters indicate stress transfer along the megathrust.

Keywords: Cascadia, subduction thrust earthquake, subduction zone, epi-sodic tremor, slip.

INTRODUCTIONThe Cascadia subduction zone represents a global end member in

which a large accretionary complex is developing as a young hot plate is subducting. Compared to other subduction zones, it has been relatively aseismic since the advent of modern seismic instrumentation. While earthquakes occur frequently in the upper and lower plates (McCrory et al., 2003), including several damaging events in the past two decades, no instrumentally recorded low-angle thrust earthquakes have been doc-umented on the plate boundary during historic times, with the possible exception of an earthquake near the Mendocino Triple Junction in 1992 (Oppenheimer et al., 1993). There is, however, considerable paleoseismic evidence for very large interplate earthquakes (Goldfi nger et al., 2003; Kelsey et al., 2005), including a magnitude 9 plate boundary event on 26 January 1700, that appears to have ruptured the entire subduction zone and was recorded as a tsunami in Japan (Satake et al., 1996).

On 12 July 2004, an earthquake with moment magnitude 4.9 occurred offshore Newport, Oregon, followed on 19 August by an event with moment magnitude 4.8 (see GSA Data Repository Table DR11). Although these events did not cause any signifi cant damage onshore, they were widely felt. The epicenters of these events (Fig. 1A) indicate that they occurred in a part of the forearc where the predicted temperature at the plate boundary is

between 350 and 450 °C (Fig. 1A). Based on studies of the frictional prop-erties of metamorphosed sediments, this temperature range is considered to be transitional between where the plate boundary is locked and where plate motion is accommodated aseismically (Hyndman and Wang, 1995). Several other earthquakes have occurred in this region since 1989 (Figs. 1B, 1C), whereas very few earthquakes have occurred elsewhere within the nominally locked or transitional zone. The adjacent Juan de Fuca plate west of the Cascadia deformation front has also been anomalously active. This region of relatively elevated seismic activity coincides with the north-ern end of a buried ridge on the subducted plate (Fleming and Tréhu, 1999) and an apparent widening of the partially coupled zone, as determined from geodetic observations (McCaffrey et al., 2007), suggesting that structures on the subducted plate may affect plate coupling.

The 2004 events occurred in two well-defi ned clusters (Fig. 1D) with several aftershocks in the days following the events (Fig. 1E). Several small events also occurred near both main shocks in the years preceding the 2004 events and again in 2007 (Table DR1). Seismograms recorded at seismic station COR for a foreshock and three aftershocks are nearly identi-cal to those of the 19 August earthquake in spite of nearly three orders of magnitude variation in amplitude (Fig. 2), indicating that they all had the same mechanism and hypocenter. Similar repeating earthquake sequences have been reported from many fault zones around the world, including in fault segments that are characterized by a large amount of aseismic creep (Nadeau and McEvilly, 1999; Uchida et al., 2003). Waveforms from events in the July cluster show more variability, although P-to-S time dif-ferences indicate that the earthquakes within this cluster also originated within 1 km of each other. Much of the scatter in the reported locations of events within these clusters may be due to location uncertainty.

MOMENT TENSOR INVERSIONThe fault mechanisms and depths of the two 2004 main shocks were

determined through moment tensor inversion of long-period regional 3-component seismograms in the 25–50 s pass band (Fig. 3; Nabelek and Xia, 1995). For each event, the source mechanism thus derived is well constrained and indicates a thrust fault with an eastward dip of 6°–15°. The slip vector of 66°–73° for the 19 August event is very similar to the expected slip vector for Juan de Fuca–Oregon block motion at this lati-tude (34.8 mm/yr with an azimuth of 75°; McCaffrey et al., 2007); for the July 12 event, the slip vector is rotated ~25° clockwise. Uncertainties in the moment tensor inversion and depth estimates from teleseismic pP arrivals are discussed in Figures DR1–DR7 in the GSA Data Repository.

HYPOCENTER RELOCATION USING 2D RAYTRACINGTo obtain an independent, high-resolution estimate of source depth

for the 19 August earthquake, we modeled traveltimes picked from short-period data (0.1–1 s) by tracing rays through the crustal model of Fig-ure 4A. The dip of the subducting Juan de Fuca plate (12° at km 185) is well constrained by controlled-source seismic data (Fig. 1D). In the upper plate, the Siletz terrane, a thick, basaltic terrane that underlies the Coast Range and Willamette Valley, forms the backstop for the Cascadia accre-tionary complex (Snavely and Wells, 1996; Fleming and Tréhu, 1999). It is underlain by a low-velocity wedge that thickens toward the east (Li and Nabelek, 1999) and is interpreted to indicate subducted oceanic crust and sediments and/or serpentinized upper-plate mantle (Tréhu et al., 1994;

Geology, February 2008; v. 36; no. 2; p. 127–130; doi: 10.1130/G24145A.1; 4 fi gures; Data Repository item 2008032.© 2008 The Geological Society of America. For permission to copy, contact Copyright Permissions, GSA, or editing@geosociety.org.

1GSA Data Repository item 2008032, Table DR1 (hypocentral parameters from the Advanced National Seismic System [ANSS] bulletin for earthquakes in the two clusters discussed in this paper), Table DR2 and Figures DR1–DR4 ( moment tensor inversion results for the two main shocks), Figures DR5 and DR6 (summary of constraints on the source depth of the main shocks from analysis of pP arrivals), and Figure DR7 (velocity models used for the various calculations documented in Figs. DR1–DR6), is available online at www.geosociety.org/pubs/ft2008.htm, or on request from editing@geosociety.org or Documents Secretary, GSA, P.O. Box 9140, Boulder, CO 80301, USA.

*E-mail: trehu@coas.oregonstate.edu.

Probable low-angle thrust earthquakes on the Juan de Fuca–North America plate boundaryAnne M. Tréhu*Jochen Braunmiller College of Oceanic and Atmospheric Sciences, Oregon State University, Corvallis, Oregon 97331-5503, USAJohn L. Nabelek

128 GEOLOGY, February 2008

Bostock et al., 2002). A narrow band of strong refl ectivity beneath the continental shelf is interpreted to be the plate boundary on this profi le and on a seismic profi le located ~40 km to the north (Tréhu et al., 1995). The strong refl ectivity likely results from high fl uid pressure. Similar strong refl ectivity has been observed offshore Vancouver Island (Nedimovic et al., 2003) and in the Nankai Trough (Kodaira et al., 2004). Whether such observations of high refl ectivity correspond to regions where the plates are locked or are slipping aseismically remains controversial; both end-member models have been proposed.

Key to our approach to decreasing the uncertainty in hypocentral parameters is a secondary arrival that follows the direct P wave by ~0.85 s

1990

1994

1998

2002

2006

1

3

5

Mag

nitu

de

125

45

44

126 124° W124° W40

42

-2000

0

126128

46

44

48°N

08/04

07/04

Mag. > 3.5 since 1989 Magnitude > 1 since 198946°N

B.D.0808

07

000

8/008/0

7/07

88

7

D Mag. > 1 2003-2007

125.0 124.5 124.044.0

44.5

45.0°N

Tréhu et al., 1994

Gerdom et al., 2000

Buried ridge Base of Siletz

C

L7

L8L9

500

500

500

500

500

500

500500500

500

500

500

500

500

500

123.5° W

TOLOCOR

A B

CTZ

500

Fleming and Tréhu, 1999

4122032

Figure 1. A: Seismicity of northeastern Pacifi c Ocean. All earthquakes in the Advanced National Seismic System (ANSS) composite catalog with magnitudes >3.5 since 1989 are included. Dark gray shading shows where thermal models indicate that temperature at megathrust is below 350 °C, and light gray indicates megathrust temperatures between 350 and 450 °C (Hyndman and Wang, 1995): these have been inter-preted to indicate regions where the plate boundary is fully locked or transitional between being locked and freely slipping, respectively. Megathrust earthquakes are expected to rupture shaded portion of plate boundary. Note absence of seismic activity within this region, with the exception of central Cascadia and northern California. The two earthquakes that are the focus of this paper are labeled by their month and year of occurrence. Bathymetric and topographic contour intervals are 1000 m. Box outlines the region shown in B. B: All earthquakes from ANSS composite catalog since 1989 with magnitude >1. Symbol size is proportional to magnitude. Clusters of seismic activity within subduction zone and in adjacent oceanic plate are apparent. Bathymetric contour interval is 250 m. Box outlines region included in D. C: Time history of seismicity since 1989 from ANSS composite catalog. Color coding in C and D shows events before (green dots) and after (blue dots) the summer 2004 earthquakes; violet indicates events in these clusters in 2007. D: Earthquakes since 2003 from ANSS com-posite catalog. Color coding shows clustering of events before (green dots) and after (blue dots) the earthquake on 12 July 2004. Violet indicates events in these clusters in 2007. Red and blue lines show locations of published two-dimensional crustal velocity models based on controlled-source seismic data. Orange segments show extent of strong lower-crustal refl ection interpreted to indicate the presence of fl uids at plate boundary. Dotted red line labeled CTZ (crustal transition zone) indicates abrupt north-south increase in seismic velocity beneath continental shelf and Coast Range. Dashed red lines indicate contours of slip defi cit (in mm/yr) relative to expected local plate motion rate of 34.8 mm/yr (McCaffrey et al., 2007); these show a transition from a narrow zone of intermediate interplate coupling north of 44.8°N to a wide zone south of 44.5°N. Background shading offshore represents aeromagnetic data illuminated from the west. Subtle ridge in magnetic fi eld is interpreted as buried basement ridge (Fleming and Tréhu, 1999). Bathymetric contour interval is 250 m.

Figure 2. Velocity seismograms for foreshock, main shock, and three aftershocks of the 19 August 2004 earthquake at seismic station COR (source-receiver distance ~80 km) and for main shock at TOLO (~30 km). Waveforms for the 5 earthquakes at COR are remarkably similar in spite of nearly 3 orders of magnitude difference in ampli-tude (scale shown in counts). Predicted relative arrival times for Pg, PmP, Sg, and SmS are for a source at x = 185 km and z = 16 km and a Poisson’s ratio of 0.29 for the model of Figure 4A. Broadband seis-mograms (20 samples/s) from the Incorporated Research Institutions for Seismology (IRIS) Data Management Center are shown with no additional fi ltering. Coda magnitudes (Mc) from the Pacifi c Northwest Seismic Network (PNSN) catalog are given for the smaller events.

GEOLOGY, February 2008 129

at COR (Fig. 2). Based on the controlled-source seismic data, which indi-cate that the amplitude of refl ections from the Moho of the subducting plate (PmP) is similar to that of the direct P-wave arrival (Pg) at source-receiver offsets of 60–110 km for sources near the epicenter of the August earth-quake (Gerdom et al., 2000), we interpret the secondary arrival observed at COR as PmP and used this observation to constrain hypocentral depth. Direct S-wave (Sg) arrivals observed at seismic stations COR and TOLO were used to decrease uncertainty in the earthquake epicenter, which can be quite large in the direction perpendicular to the coastline for offshore earthquakes (Braunmiller et al., 1997).

Observed (PmP – Pg) and (Sg – Pg) times at seismic stations COR and TOLO were compared to predicted times for a wide range of pos-sible source positions and Poisson’s ratios, as shown in Figures 4B–4D. This analysis indicates that the earthquake occurred within 1 km of the interpreted plate boundary at a depth of ~16 km. It also confi rms that the

Figure 3. Source mechanisms of the summer 2004 main shocks from full waveform inversion of 25–50 s data. Azimuthal coverage, shown by triangles around the edge of the focal sphere plots, is excellent for an offshore earthquake.

July 12(projected)

coastDeformation front

Siletz Terrane

Juan de Fuca mantle

Distance along model (km)

Dep

th (

km)

Low- velocity zone

“Bright spot”

COR

Vel

ocity

(km

/s)

A

Juan de Fuca crustAugust 19

vertical exaggeration 2.7X

200 250Distance along model (km)

COR

Pg,Sg

PmP,SmS

+ -

Plate boundary

B ETOLO0

30

10

40

20

Dep

th (

km)

5 5.5

0

5

10

15

20

160 170 180 190 200

Source position (km)

Sou

rce

dept

h (k

m)

2 s1 s

C0 0.5 1 1.5 2

0

5

10

15

20

25

PmP - Pg (s)

Sou

rce

dept

h (k

m)

S -

P ti

me

(s)

D

0.6

Moho

4

6

8

10

120.25 0.27 0.29 0.31

Poisson’s ratio

TOLO

COR

km 180 185 190

Critical angleCritical angle

Plate boundary

Figure 4. A: P-wave velocity model for modeling arrivals identifi ed in short-period data (Gerdom et al., 2000). Model was developed from multi channel seismic refl ection data and large-aperture data from offshore airgun shots recorded on 17 ocean bottom and 22 onshore seis-mometers using a hybrid inverse and forward-modeling approach based on RAYINVR software (Zelt and Smith, 1992). White dashed line delimits region imaged by controlled-source data. Base of the Juan de Fuca crust is well constrained everywhere within this region. Red dashed line shows plate boundary, which is observed within ~20 km of deformation front and at strong pre-critical refl ection (thick orange line), and inferred to be ~7 km above base of Juan de Fuca plate crust elsewhere. Siletz terrane is indicated by diagonal lines. B: Predicted time difference between Pg and PmP as a function of source depth for a source at x = 185 km, which corresponds to the projection of the 19 August 2004 epicenter onto seismic profi le. Blue dot shows time difference for shots near sea surface observed for the same source-receiver offset. Time difference (Δt) of 0.85 ± 0.05 s observed in seismograms for seismic station COR in Figure 2 corresponds to a depth of 16 ± 0.5 km. C: Contours of predicted Δt between PmP and Pg. Source position was varied in 5 km intervals horizontally and in 2 km intervals in depth. Solid red line shows interpreted position of plate boundary. Contour for Δt = 0.85 s has its closest approach to plate boundary at km 185 and moves away gradually as source moves east. To constrain the horizontal position, time differences between Sg and Pg at stations COR and TOLO were modeled for range of Poisson’s ratio (see D); dashed red line shows predicted contours that correspond to the observa-tion at TOLO of 5.25 ± 0.05 s assuming the preferred Poisson’s ratio of 0.29. D: Predicted S-P time differences at COR and TOLO as a func-tion of Poisson’s ratio for three different source positions along the 0.85 s contour in C: (x = 180 km, z = 14 km), dotted lines; (185, 16), solid lines; (190, 16), dashed lines. Observed time differences of 5.25 ± 0.05 s at TOLO and 10.90 ± 0.15 at COR are shown as solid horizontal lines. Near-vertical dashed lines join solutions for same source position and indicate best-fi tting Poisson’s ratio for each position. Observed S-P times at both stations are the same as predicted times for source at (185,16) if Poisson’s ratio is 0.285–0.29, which is consistent with labora tory measurements of Poisson’s ratio in mafi c rocks. Sources at km 180 or 190 are consistent with observed S-P times only if Poisson’s ratio is different for paths to the two stations, and in both cases required Poisson’s ratio is either too low or too high for mafi c rocks. E: Ray diagram for a source at x = 185 km, z = 16 km. Ray paths for Pg, Sg, PmP, and SmS to seismic stations COR and TOLO are highlighted. Star indicates critical angle, defi ned by Snell’s law, for this source depth and velocity structure; beyond critical angle, all energy is refl ected, consistent with large PmP amplitude observed at COR (refl ection coeffi cient 1.0) and with the absence of a clear PmP phase at TOLO (refl ection coeffi -cient ~0.1). Radiation pattern for moment tensor solution of Figure 2 is also shown and indicates that Pg and PmP originated in quadrants with opposite polarity, consistent with observed phase shift of ~180°.

130 GEOLOGY, February 2008

epicentral parameters published by the Pacifi c Northwest Seismic Net-work (PNSN) for the 19 August event are accurate to within 5 km, even though the PNSN depth was overestimated by ~12 km. The Poisson’s ratio of 0.285–0.29 obtained for the path between the earthquake and stations TOLO and COR is consistent with the mafi c nature of the Siletz terrane (Christensen, 1996). The ray-tracing analysis is also consistent with the time difference between Sg and SmS at COR (Fig. 2) and with the obser-vation that Pg and PmP have opposite polarity (Figs. 2 and 4E). Ray trac-ing through this velocity model to match traveltimes of teleseismic pP arrivals supports this depth estimate (Fig. DR5).

No PmP arrival is observed for the 12 July 2004 earthquake. The range of source-receiver offsets within which the amplitude of PmP rises above the background noise level is small, and observations of PmP arrivals on tem-porary stations deployed onshore to record shots along L8 and L9 (Fig. 1D) suggest that stations TOLO and COR were not near the critical distance for PmP arrivals for this event. By modifying the model of Figure 4A to incor-porate along-strike changes in the density and velocity of the accretionary complex (Flemings and Tréhu, 1999; Gerdom et al., 2000) and modeling pP arrivals for this event (Fig. DR6), we conclude that this event occurred at a depth of 9–12 km, also close to the interpreted plate boundary. Proximity to where the subducted ridge abuts the Siletz backstop may explain the appar-ent rotation of the slip vector for this event.

DISCUSSIONThe close correspondence between the preferred hypocentral depths

and the plate boundary as interpreted from the high-resolution crustal model suggests that the two main shocks (and other events with similar waveforms) resulted from slip on the plate boundary. This is the fi rst evi-dence for seismogenic slip in the locked or transitional zone (Fig. 1A) since the great Cascadia earthquake of A.D. 1700 (Satake et al., 1996). A time delay of ~6 months between these earthquakes and the previous episode of episodic tremor and slip (ETS) along this segment of Cascadia (Brudzinski and Allen, 2007) suggests that slip downdip of the transition zone, where ETS originates, may trigger slip at shallower depth (Mazzotti and Adams, 2004). That seismic activity occurred in both clusters in 2004 and again in 2007 (Figs. 1C, 1D) suggests external triggering by the same process. While it may be premature to conclude whether these earthquakes represent repeated slip on small asperities surrounded by a region of stable sliding or slip on a weak portion of the plate boundary embedded within a locked fault, confi rmation of apparent correlations between ETS episodes and clustered seismicity by future observations would support the latter model. This study also underlines the importance of having an accurate model for the crustal velocity structure in order to accurately locate earth-quakes and understand their tectonic implications.

ACKNOWLEDGMENTSControlled-source seismic data were acquired with support from the National

Earthquake Hazards Reduction Program (NEHRP), managed by the U.S. Geologi-cal Survey. Earthquake seismograms for this analysis were obtained from the Data Management Center operated by the Incorporated Research Institutions for Seis-mology (IRIS) and from the Geological Survey of Canada. Data analysis was sup-ported by U.S. National Science Foundation grant OCE-0550402. We thank Rob McCaffrey for a preprint of his paper and Garry Rogers for a thoughtful review and for suggesting that we look for pP.

REFERENCES CITEDBostock, M.G., Hyndman, R.D., Rondenay, S., and Peacock, S.M., 2002, An

inverted continental Moho and serpentinization of the forearc mantle: Nature, v. 417, p. 536–538, doi: 10.1038/417536a.

Braunmiller, J., Leitner, B., Nabelek, J.L., and Tréhu, A.M., 1997, Location and source parameters of the June 19, 1994 (Mw = 5.0) offshore Petrolia, CA, earthquake: Seismological Society of America Bulletin, v. 87, p. 272–276.

Brudzinski, M.R., and Allen, R.M., 2007, Segmentation in episodic tremor and slip all along Cascadia: Geology, v. 35, p. 907–910, doi: 10.1130/G23740A.

Christensen, N.I., 1996, Poisson’s ratio and crustal seismology: Journal of Geo-physical Research, v. 101, p. 3139–3156, doi: 10.1029/95JB03446.

Fleming, S.W., and Tréhu, A.M., 1999, Crustal structure beneath the central Oregon convergent margin from potential-fi eld modeling: Evidence for a buried base-ment ridge in local contact with a seaward dipping backstop: Journal of Geo-physical Research, v. 104, p. 20,431–20,447, doi: 10.1029/1999JB900159.

Gerdom, M., Tréhu, A.M., Flueh, E.R., and Klaeschen, D., 2000, The continen-tal margin off Oregon from seismic investigations: Tectonophysics, v. 329, p. 79–97, doi: 10.1016/S0040–1951(00)00190–6.

Goldfi nger, C., Nelson, C.H., and Johnson, J.E., 2003, Holocene earthquake records from the Cascadia subduction zone and northern San Andreas fault based on precise data of offshore turbidites: Annual Reviews of Geo physics, v. 31, p. 555–577.

Hyndman, R.D., and Wang, K., 1995, The rupture zone of Cascadia great earth-quakes from current deformation and the thermal regime: Journal of Geo-physical Research, v. 100, p. 22,133–22,154, doi: 10.1029/95JB01970.

Kelsey, H.M., Nelson, A.R., Hemphill-Haley, E., and Witter, R., 2005, Tsunami history of an Oregon coastal lake reveals a 4,600 year record of great earth-quakes on the Cascadia subduction zone: Geological Society of America Bulletin, v. 117, p. 1009–1032, doi: 10.1130/B25452.1.

Kodaira, S., Iidaka, T., Kato, A., Park, J., Iwasaki, T., and Kaneda, Y., 2004, High pore fl uid pressure may cause silent slip in the Nankai Trough: Science, v. 304, p. 1295–1298, doi: 10.1126/science.1096535.

Li, X., and Nabelek, J.L., 1999, Deconvolution of teleseismic body waves for enhancing structure beneath a seismometer array: Seismological Society of America Bulletin, v. 89, p. 190–201.

Mazzotti, S., and Adams, J., 2004, Variability of near-term probability for the next great earthquake on the Cascadia subduction zone: Seismological Society of America Bulletin, v. 94, p. 1954–1959, doi: 10.1785/012004032.

McCaffrey, R., Qamar, A.I., King, R.W., Wells, R., Khazaradze, G., Williams, C.A., Stevens, C.W., Vollick, J.J., and Zwick, P.C., 2007, Fault locking, block rotation and crustal deformation in the Pacifi c Northwest: Geophysi-cal Journal International, v. 169, p. 1315–1340.

McCrory, P.A., Blair, J.L., Oppenheimer, D.H., and Walter, S.R., 2003, Depth to the Juan de Fuca slab beneath the Cascadia subduction margin—A 3D model for sorting earthquakes: U.S. Geological Survey Digital Data Series, 1 CD-ROM.

Nabelek, J.L., and Xia, G., 1995, Moment-tensor analysis using regional data: Application to the 25 March, 1993, Scotts Mills, Oregon, earthquake: Geo-physical Research Letters, v. 22, p. 13–16, doi: 10.1029/94GL02760.

Nadeau, R.M., and McEvilly, T.V., 1999, Fault slip rates at depth from recur-rence intervals of repeating earthquakes: Science, v. 285, p. 718–721, doi: 10.1126/science.285.5428.718.

Nedimovic, M.R., Hyndman, R.D., Ramachandran, K., and Spence, G.D., 2003, Refl ection signature of seismic and aseismic slip on the northern Cascadia subduction thrust: Nature, v. 424, p. 416–420, doi: 10.1038/nature01840.

Oppenheimer, D.H., Eaton, J., Jayko, A., Lisowski, M., Marshall, G., Murray, M., Simpson, R., Stein, R., Beroza, G., Magee, M., Carver, G., Dengler, L., McPherson, R., Gee, L., Romanowicz, B., Gonzalez, F., Li, W.H., Satake, K., Somerville, P., and Valentine, D., 1993, The Cape Mendocino, Califor-nia, earthquakes of April 1992: Subduction at the Triple Junction: Science, v. 261, p. 433, doi: 10.1126/science.261.5120.433.

Satake, K., Shimazaki, K., Tsuji, Y., and Ueda, K., 1996, Time and size of a giant earthquake in Cascadia inferred from Japanese tsunami records of January, 1700: Nature, v. 379, p. 246–249, doi: 10.1038/379246a0.

Snavely, P.D., and Wells, R.E., 1996, Cenozoic evolution of the continental margin of Oregon and Washington, in Rogers, A., et al., eds., Earthquake hazards in the Pacifi c Northwest of the United States: U.S. Geological Sur-vey Professional Paper 1560, p. 161–182.

Tréhu, A.M., Asudeh, T.M., Brocher, T.M., Luetgert, J.H., Mooney, W.D., Nabelek, J.L., and Nakamura, Y., 1994, Crustal architecture of the Cascadia forearc: Science, v. 266, p. 237–243, doi: 10.1126/science.266.5183.237.

Tréhu, A.M., Lin, G., Maxwell, E., and Goldfi nger, C., 1995, A seismic refl ec-tion profi le across the Cascadia subduction zone offshore central Oregon: New constraints on the deep crustal structure and on the distribution of methane in the accretionary prism: Journal of Geophysical Research, v. 100, p. 15,101–15,116, doi: 10.1029/95JB00240.

Uchida, N., Matsuzawa, T., Igarashi, T., and Hasagawa, A., 2003, Interplate quasi-static slip off Sanriku, NE Japan, estimated from repeating earthquakes: Geo-physical Research Letters, v. 30, no. 15, 1801, doi: 10.1029/2003GL017452.

Zelt, C.A., and Smith, R.B., 1992, Seismic traveltime inversion for 2D crustal structure: Geophysical Journal International, v. 108, p. 16–34, doi: 10.1111/j.1365–246X.1992.tb00836.x.

Manuscript received 4 June 2007Revised manuscript received 1 October 2007Manuscript accepted 2 October 2007

Printed in USA