morphology of the distorted subducted paciﬁc slab beneath...

Physics of the Earth and Planetary Interiors 156 (2006) 1–11

Morphology of the distorted subducted Pacific slabbeneath the Hokkaido corner, Japan

M.S. Miller a,∗, B.L.N. Kennett a,1, A. Gorbatov b,2

a Research School of Earth Sciences, The Australian National University, Mills Road, Building 61, Canberra, ACT 0200, Australiab Geoscience Australia, GPO Box 378, Canberra, ACT 2601, Australia

Received 10 May 2005; received in revised form 1 January 2006; accepted 17 January 2006

Abstract

The intersection of the Japan and Kurile arcs is expressed as a cuspate feature at the trench, a bend in the Japanese islands,and a complex lithospheric structure and is known as the Hokkaido corner. The Pacific plate is subducting beneath the two arcsin the northwest Pacific at different velocities, which has resulted in an arc–arc collision and distortion of the subducting oceaniclithosphere. Using P and S wave tomographic images and seismicity the distorted shape of the subducted Pacific plate beneaththe Hokkaido corner can be interpreted in three dimensions. The Pacific plate is imaged as continuous along the Japan–Kurile arcwith dip angle increasing from south to north, but at the arc–arc junction the geometry is complex and appears crumpled. Beneaththe Hokkaido corner there are two features in the tomography that are clearly imaged: a region of low velocity at approximately50–150 km depth between 142.5–145◦E and 41–43◦N and a distorted, buckled lower slab boundary. These two unusual characteristicsin the slab morphology are likely to be related to deformation of the subducted Pacific plate at the arc–arc junction.©

K

1

itePt

f

bA

0

2006 Elsevier B.V. All rights reserved.

eywords: Subducted slab; Japan and Kurile arcs; Upper mantle; Tomography; 3D imaging; Arc–arc junction

. Introduction

Along the northwest Pacific margin the Pacific plates converging at a rate of approximately 8.2–9.2 cm/yearo the northwest beneath the Kurile and Japan arcs (Senot al., 1993; DeMets et al., 1994). At the Japan arc theacific plate is subducting approximately perpendicular

o the arc, but is subducting obliquely at the Kurile arc,

∗ Corresponding author. Tel.: +61 2 6125 3803 (O);ax: +61 2 6125 8253.

E-mail addresses: [email protected] (M.S. Miller),[email protected] (B.L.N. Kennett),[email protected] (A. Gorbatov).1 Tel. +61 2 6125 4621 (O); fax: +61 2 6257 2737.2 Tel. +61 2 6249 9376 (O); fax: +61 2 6249 9999.

which results in an arc–arc collision. The junction of thetwo island arcs is expressed as a sharp cuspate featureand bend in the northern Japanese islands known as theHokkaido corner (Fig. 1). The slab geometry beneaththe Hokkaido corner is very complex and distorted asa result of the collision of the two arcs. The deforma-tion of subducting plate at the arc–arc junction may beassociated with interplate and lithospheric earthquakesthat occur in the region (Suzuki and Kasahara, 1996;Katsumata et al., 2003) and expressed in the uplift ofthe Hidaka Mountains (Okada, 1983; Miyamachi andMoriya, 1984, 1987; Miyamachi et al., 1994; Kimura,1996; Moriya et al., 1998). The unique setting and geo-logical characteristics of the region results in a com-plicated model, but we present a new interpretation ofthe subducting slab morphology beneath the Hokkaido

031-9201/$ – see front matter © 2006 Elsevier B.V. All rights reserved.doi:10.1016/j.pepi.2006.01.007

mailto:[email protected]



dx.doi.org/10.1016/j.pepi.2006.01.007

2 M.S. Miller et al. / Physics of the Earth and Planetary Interiors 156 (2006) 1–11

Fig. 1. Map of the major features in northwest Pacific including ETOPO5 bathymetry (1998) contoured in 1 km increments, volcano locationsfrom the Smithsonian Institute volcano index convergence rates of the Pacific plate (Seno et al., 1993 and Zang et al., 2002), and location of thecross-sections in Figs. 3 and 4.

corner using 3D ray tracing tomographic inversions andseismicity.

2. Previous studies

Initial studies of subduction zone structures alongareas of sharp bends, such as the junction of the Japanand Kurile arcs, suggested the slab would be contortedor discontinuous (Isacks and Molnar, 1971). Later inves-tigation of the subducting slab morphology beneath thenorthwestern Pacific integrated seismicity and seismictomography data to define the geometry (Burbach andFrohlich, 1986; Zhou et al., 1990; Chiu et al., 1991; vander Hilst et al., 1991, 1993; Fukao et al., 1992; Zhao etal., 1992; Zhao and Hasegawa, 1993; Miyamachi et al.,1994; Gudmundsson and Sambridge, 1998; Moriya etal., 1998). Many of these studies suggested the Pacific

plate changes dip near the Hokkaido corner from 40◦to 50◦ at the Kurile arc to approximately 30◦ beneaththe Japan arc and that the slab varies in thickness alongthe arc. Previous research has investigated the geome-try of the slab beneath the northwest Pacific margin, butfew studies have discussed the slab morphology at theHokkaido corner in detail.

Chiu et al. (1991) were pioneers in three-dimensionalimaging of subducting slabs in the Western Pacific. Theydefined the top of the subducting Pacific plate as anenvelope of selected hypocenters and used this data toillustrate the slab surface in three dimensions. Theirmodels for the Japan and Kurile arcs did not illustrateany distortion in the slab, but detailed the contrast insubduction angle of the gently dipping central Japanzone and the steeply dipping Kurile zone in the north.The transition was featured as a smooth surface with

M.S. Miller et al. / Physics of the Earth and Planetary Interiors 156 (2006) 1–11 3

curved, fan shaped contours in their models. They foundthat shallow to intermediate depth seismicity was notdisrupted at the Hokkaido corner and noted that seis-mic waves propagated efficiently from events in theKurile and Hokkaido areas to the stations on the east-ern coast of Hokkaido, therefore suggested a continuousslab.

An inverse method was applied to P and S wavedata from local events by Miyamachi et al. (1994) tomodel the three-dimensional velocity structure beneathnorthern Japan. This inversion proposed the top of thePacific plate ranged in depth from 50 to 170 km andthe subduction angle was greater beneath Hokkaidothan along northeastern Japan. The upper slab bound-ary in their model exhibited a gentle curved shapethat approximately mirrored the trench shape, but with-out a sharp bend often shown at the Hokkaido cor-ner. The estimated thickness of the subducting slabranged from 90 km along the Kurile arc and 110 kmalong the northern Japan arc from the local earthquakedata.

An alternative approach was used to create a three-dimensional model of subducting slabs along theWestern Pacific by contouring slab-related seismicityto approximate the top of the slab and the bottomdefined as a constant 200 km thick oceanic lithosphere(Gudmundsson and Sambridge, 1998). The morphologywas simplified since the slab extent was based on earth-quakes and a projection of constant thickness, yet slabcomplexity and general characteristics of the geometryatcsmm

irpmeurmbeceove

3. Tomographic controls

The models presented in this study are based onregional body wave joint tomographic inversion ofGorbatov and Kennett (2003) together with a newP-wave tomographic inversion were produced fromthe same arrival-time data. The detailed joint inver-sion datasets used an inversion algorithm introducedby Kennett et al. (1998) that was then adapted byGorbatov and Kennett (2003) to include 3D raytracing. In this method the trajectory of seismic-raypropagation between source and receiver through thethree-dimensional structure of the earth is included,which improves the resolution of gradients and strongvariations in wave speeds. The joint inversion usedP and S arrival-time data with the same source andreceiver from the global catalog of Engdahl et al. (1998),which are plotted in Fig. 2. In order to optimize datacoverage and ensure the final images can be directlycompared, single rays with both P and S readings werepicked for events and stations within the study area.The surrounding Western Pacific mantle structure wasparameterized into a non-overlapping grid of 5◦ × 5◦with 16 layers ranging from 35 to 200 km down to atotal depth of 1600 km. The study area consisted of agrid of 19 layers with cells 0.5◦ × 0.5◦ in the uppermostmantle, 1◦ × 1◦ in the transition zone and lower mantle,and 2◦ × 2◦ beneath continents to a depth of 1500 km.The final dataset contained 900,000 pairs of P and S raypaths and used the ak135 model as reference (Kennett

long the arc were well defined. The model depictedhe bend at the Japan–Kurile arc junction and the asso-iated change in dip along the plate boundary as amooth curved shape in their regionalized upper mantleodel and provided another method to model the slaborphology.Early investigations were limited by lower resolution

mages making smaller scale structures more difficult toesolve and made assumptions to simplify the modelingrocess. As technology and data quality has improvedore recent studies have been able to define the geom-

try of the Japan–Kurile arc–arc junction in more detailsing enhanced methodology. Katsumata et al. (2003)ecently investigated the geometry of the deep seis-ic zone associated with the subducting Pacific plate

eneath the Hokkaido corner using hypocenters fromvents between July 1999 and July 2001. Their workonfirmed the change in dip of the subducting plate onach side of the arc–arc junction, but the lateral transitionf the subduction angle was more significant than in pre-ious models by Miyamachi et al. (1994) and Hasegawat al. (1994).

et al., 1995). This nested iterative approach resultedin P and S models that were then used in a non-linearjoint tomographic inversion for bulk-sound and shearwave speed inversions (Gorbatov and Kennett, 2003).Then a separate P-wave tomographic inversion wasobtained with a similar non-linear scheme using thesame dataset as in the joint tomography and with thesame inversion parameters, but using a standard tomo-graphic formulation. Although the inversion schemesused for joint tomography and the P-wave tomographyare different, the three inversions can be comparedwith careful consideration of the smoothing (damping)parameters. We present a subset of the previouslypublished shear wave-speed model and the new P-wavemodel to illustrate the structure of the subducting Pacificplate beneath the junction of the Japan and Kurile arcs.

4. Tomographic images

The tomographic images presented have a maxi-mum depth of 900 km and over the region between136◦–149◦E and 37◦–48◦N. The large number of seis-


Fig. 2. World maps illustrating (A) the distribution of seismic stations from which arrival times were recorded as triangles and (B) epicenterdistribution (as dots) of earthquakes selected from Engdahl et al. (1998) catalog and that were used in the inversions.

mic stations and earthquake events in the area, recordedteleseismic events, and small cell size allow for thestructure of the subducting slab to be well resolved.A convenient way to measure the data coverage is thesum of the lengths of the ray segments traversing eachcell. This is illustrated in Fig. 3 with ray density plots,which show very consistent high ray density across theregions where subduction zone features are expectedbeneath the Japan and Kurile arcs. Detailed resolu-tion tests using a synthetic model, in Gorbatov andKennett (2003), concluded the image of the subductedslab could be recovered along its complete extensionusing the same inversion parameters as in the observeddata.

Fast wavespeed anomalies used to define the subduct-ing slab in the P-wave inversion and shear wave-speed

images are clearly imaged down to at depth of least700 km (Figs. 4–6). The dip of the Pacific plate decreasesfrom north to south along the Japan–Kurile arc, consis-tent with previous studies (Chiu et al., 1991; van derHilst et al., 1993; Miyamachi et al., 1994; Katsumata etal., 2003), and is illustrated in the cross sections throughthe P-wave and shear wave-speed tomographic models(Figs. 4 and 5). Beneath the Kurile arc the subductingslab is dipping at approximately 50◦ and the seismicityfalls within the high velocity zone (Figs. 4 (K–K′) and 5(B–B′ and C–C′)). Beneath the Japan arc the Pacific slabis well defined as dipping at approximately 30◦ and thengradually transitions into a horizontal structure above the660 km discontinuity (Fig. 4 (J–J′)).

An interesting low velocity feature is present in thetomographic images, which suggests a complex shape


Fig. 3. Maps and cross-sections, respectively, representing density of ray-length segments in log 10 km. (A) P-wave and (B) S-wave seismic raysillustrate the ray density in plan view sliced along 275 km depth and (C) P-wave and (D) S-wave are cross-sections along 42◦N latitude.

in the neighborhood of 41–43◦N, 145◦–147◦E, and50–150 km depth. This feature lies approximately belowthe Hokkaido corner, where the Japan and Kurile arcsmeet at a point (Fig. 1). A series of slices through theP-wave and shear wave-speed models at the Hokkaidocorner show the anomalous feature in the upper mantle(Figs. 5 and 6). The P-wave images have lower ampli-tudes in comparison to the shear wave-speed images,but both illustrate the prominent low velocity featurein the upper 150 km to the west of the slab. The shearwave speed images have stronger fast velocity pertur-bations that illustrate an approximately linear top ofthe subducting plate, but a discontinuous lower bound-ary. The bottom of the subducting slab appears crum-pled in the upper mantle in both P and S inversions,unlike the top slab surface, which is a more smoothand coherent feature (Fig. 6). It appears that the sub-ducting Pacific plate at the Hokkaido corner becomesvery distorted and does not bend smoothly at the arc–arcjunction.

5. Three-dimensional morphology

The complex morphology of the subducting Pacificplate beneath northern Japan and the Kurile islands hasbeen interpreted from a combination of seismicity andseismic tomography. A three-dimensional visualizationof the slab geometry beneath the Hokkaido corner wascreated by systematically picking the top and bottom ofthe slab from the P-wave inversion in Figs. 4–6. The topand bottom of the slab was defined by picking the max-imum gradient of velocity in the P-wave model, whichcompensates for the lower resolution in the northern por-tion of the model where there are fewer seismic stations.The high velocity anomalies that were used to define theslab geometry are subjective, but with care in pickingthe maximum gradient rather than a constant value con-tour the slab structure can be extracted irrespective ofamplitude of the velocity perturbations.

The interpreted points were then used to createa schematic model of the subducting Pacific plate


Fig. 4. Cross-sections of the P-wave velocity perturbation from loca-tions shown in Fig. 1 with hypocenters of events during 1967–2002with magnitude greater than 4.5 from the NEIC/USGS catalog as yel-low dots. Cyan colors represent fast velocity perturbations and Browncolors represent slow velocity perturbations relative to ak135 (Kennettet al., 1995). Cross-section K–K′ illustrates the relatively steep dippingPacific plate beneath the Kurile arc and J–J′ shows the more shallowdipping slab beneath the Japan arc. The Z units are in km depth. (Forinterpretation of the references to colour in this figure caption, thereader is referred to the web version of the article.)

(Fig. 7(A–C)). The interpreted top of the slab picks wereused to form a surface that illustrates the top of the sub-ducting Pacific plate and the bottom of the slab pickswere used to model the bottom of the slab. Using the twosets of interpreted points surfaces representing the topand bottom of the subducting slab were created. The twosurfaces were used to define the solid three-dimensionalshape and extent of the slab. The three-dimensional inter-preted models illustrate the change in dip of the slabbeneath the Japan and Kurile arcs in Fig. 7(A and B)and intra-plate seismicity (Fig. 7(C)) when the slab isviewed transparently. This new model agrees with pre-vious models in which the subducting Pacific plate dipangle is between 40–50◦ beneath the Kurile arc andapproximately 30◦ beneath the Japan arc. The subduct-ing slab appears to be continuous and have a coherenttransition between the two subduction angles near thearc–arc junction at the Hokkaido corner.

Fig. 5. Three-dimensional (A) P-wave tomographic model and (B)shear wave-speed model sliced along the cross-sections shown in Fig. 1(A–A′ to F–F′) to illustrate the Pacific plate subducting beneath theJapan and Kurile arcs. Fast perturbations are in cyan and slow per-turbations are in brown relative to ak135 (Kennett et al., 1995). TheAsian coastline is in purple, plate boundary is a broken black line, andQuaternary volcanoes from the Smithsonian volcano index are plot-ted as yellow triangles. (For interpretation of the references to colourin this figure caption, the reader is referred to the web version of thearticle.)

The detailed morphology of the slab at the Hokkaidocorner is illustrated in Fig. 8(A and B) using con-tour maps of the top and bottom of the interpretedslab surfaces. The depth of the top and bottom ofthe slab is contoured in 10 km increments down toa maximum depth of 260 km. The slab appears tohave a generally smooth surface except along the bot-tom of the slab near the Hokkaido corner. The steep-ening of the subduction angle from south to northalong the arc and the complex shape is apparent in


Fig. 6. Three-dimensional P-wave and shear wave speed tomographic images sliced approximately perpendicular to the trench near the Hokkaidocorner to illustrate the anomalous features in the subducting Pacific plate and the mantle wedge. The locations of the cross-sections are shown inFig. 1 as dotted lines. The volcano locations are projected on to the profiles. The Z units are in km with a maximum depth of 450 km.

the rapid change in dip in the slab surface (Fig. 6B).The bottom of the slab has a more contorted morphol-ogy, which illustrates the complex nature of the junc-tion of the Japan and Kurile arcs. The contour mapspresent a more complex and detailed image of the

slab morphology than those contour maps produced byMiyamachi et al. (1994) and Katsumata et al. (2003). Themost distorted region is at the Hokkaido corner, whichis located at approximately 42.5◦N and 145–146◦E(Figs. 5–8B).


Fig. 7. (A–D) Interpreted three-dimensional model of the morphology and geometry of the subducting Pacific plate beneath the Hokkaido corner.The Japanese coastline is in purple, current plate boundaries are represented by black dashed lines. (A) Represents the method of picking the topof the slab (green dots) and the bottom of the slab (yellow dots) from the P-wave images. (B and C) Illustrate the interpreted model and (D) showsthe earthquake hypocenters for events greater than 4.5 during 1967–2002 from the USGS/NEIC catalog and are color coded with depth. (Note: theslab does not converge at a point in the NW. It is an effect of the perspective.) (For interpretation of the references to color in this figure caption, thereader is referred to the web version of the article.)

6. Discussion

Hokkaido is a unique region because it lies at the junc-tion of two subduction zones that are both seismicallyand volcanically active. The Hidaka Mountains, whichstrike approximately NE–SW on the island of Hokkaidoare an expression of the collision of the northern Japanand southern Kurile arcs. Many of the studies done inthis region have focused on the crustal structure and itsrelationship to the arc–arc junction, but few have inves-tigated the slab morphology beneath this region.

Moriya (1986), Moriya et al. (1998) and Miyamachiand Moriya (1984, 1987) investigated the structurebeneath the Hidaka Mountains using seismic observa-tions and their models suggest the complex velocitystructure of the area is due to the oblique collision of

the northern Japan (Honshu) and southern Kurile arcs.It was proposed that the variation of rate and angle ofconvergence along the arc (Fig. 1) resulted in a complexoverlap structure in the crust and in the transition zone.Miyamachi and Moriya (1984) imaged an inclined lowvelocity zone at a depth to 70 km beneath the HidakdaMountains. Miyamachi et al. (1994) suggested that thisdepth coincided with the location of the top of the slaband could be interacting; they speculated on the thicknessof the Pacific plate based on seismic velocity distribution(range of 90–110 km), but the lower plate boundary wasan unknown parameter in their models.

Katsumata et al. (2003) also studied the seismicity inthe region to image the shape of deep seismic zone atthe Japan–Kurile arc junction. In the process of model-ing the seismic zone they found an unusual cluster of


Fig. 8. (A–B) Contour maps of the top and bottom of the interpreted Pacific slab surfaces with contours in 10 km increments with Hokkaido coastlinein blue, the Pacific plate boundary in red, and earthquake events with magnitude greater than 4.5 between 1967 and 2002 from the USGS/NEICcatalog. (For interpretation of the references to color in this figure caption, the reader is referred to the web version of the article.)

intra-plate earthquakes where the dip of the Pacific platechanges at the Hokkaido corner. Focal mechanisms forthis near vertical linear cluster were interpreted to havepredominately extensional characteristics. From theseresults they suggest this cluster is a slab-cracking zonethat could develop into a slab tear due to the rapid lat-eral change in subduction angle at the arc–arc junction(Katsumata et al., 2003). This near vertical cluster islocated at a depth range of 50–110 km near 42.5◦N and143.5◦E. Although the new model presented here doesnot image a tear, a low velocity anomaly in the slab asseen in the tomographic images in Figs. 5 and 6 is locatedin the same region.

A different explanation for the shallow low veloc-ity zone and earthquake cluster found by Katsumata(2003) near the Hokkaido corner can be explained byits position beneath a cluster of Quaternary volcanoesin central Hokkaido (Figs. 1, 6, and 9). The low veloc-ity zone in the upper mantle extends from approxi-mately 125 km to near the surface beneath the Tokachi,Daisetsu, Nipesotsu-Upepesanke, and Shikaribetsu vol-canic groups along B–B′. These volcanoes in centralHokkaido (Figs. 1, 5, 6, and 9) are composed of a seriesof overlapping andesitic to dacitic lava domes and strato-volcanoes trending perpendicular to the plate boundary.The correlation between this series of volcanoes thattrend along a line running NW–SE, which is unlike theother volcano groups in Hokkaido, and the presence of

a low velocity anomaly (LVZ) in the mantle wedge isevidence for magma generation and conduits to the vol-canoes (Fig. 9).

Similar low velocity zones interpreted as the sourceand path of magma for volcanoes have been imaged inother regions in Japan and South America (Nakajima etal., 2001; Wyss et al., 2001; Tamura et al., 2002; Husenet al., 2003), but not in this specific region. The lowvelocity in both P-wave and shear wave speed images donot appear continuously along the Japan arc, as noted in

Fig. 9. Map of the Hokkaido corner illustrating the relationship of thelow velocity zone (LVZ) and the active volcanic front.


previous research (Tamura et al., 2002), but there appearsto be a basic correlation between volcanic groups andlow velocity zones in the mantle wedge. Cross-sectionB–B′ in Fig. 6 images the localized low velocity zoneextending from about 125 km depth up to the near surfacewhere the volcanic group is located, which contrasts themore typical mantle wedge geometry features such asthose observed in E–E′. This low-velocity feature is alsopresent in the recent work of Wang and Zhao (2005) (intheir Fig. 9 sections E–E′ and F–F′) using a different dataset, but is neither commented upon nor interpreted.

The position of the cluster of volcanoes in centralHokkaido lie above the most crumpled and distorted por-tion of the subducting Pacific plate. The contour mapsin Fig. 6 illustrate the position of the volcanoes rela-tive to the top of the slab, which are at approximatelythe same depth as the other active volcanoes in the area(∼150 km depth). The position of these volcanoes rela-tive to the main volcanic front and is similar to location ofthe Klyuchevskoy and Sheveluch volcanic groups northof ∼55◦N on the Kamchatka peninsula (Gorbatov et al.,1997). The Kamchatka volcanoes are shifted west of themain axis of the volcanic front and lie above a portionof the slab that has changed dip and become shallower(Figs. 8 and 9). West of this uplifted area beneath centralHokkaido, where the cluster of volcanoes are located,the depth of the top of the slab is deeper and thicker dueto the change in dip and thickness of the slab. Abovethis thickened oceanic lithosphere there appears to be agap in the volcanic front. Beyond the gap there is a clus-

which appears to have retreated to its present positionover the last 15 million years.

References

Burbach, G.V., Frohlich, C., 1986. Intermediate and deep seismicityand lateral structure of subducted lithosphere in the Circum-Pacificregion. Res. Geophys. 24, 833–874.

Chiu, J.-M., Isacks, B.L., Cardwell, R.K., 1991. 3-D configuration ofsubducted lithosphere in the western Pacific. Geophys. J. Int. 106,99–112.

DeMets, C., Gordon, R.G., Argus, D.D., Stein, S., 1994. Effect ofrecent revisions to the geomagnetic reversal time scale on estimatesof current plate motions. Geophys. Res. Lett. 21, 2191–2194.

Engdahl, E.R., van der Hilst, R.D., Buland, R., 1998. Global tele-seismic earthquake relocation with improved travel times andprocedures for depth determination. Bull. Seismol. Soc. Am. 88,722–743.

Fukao, Y., Obayashi, M., Inoue, H., Nenbai, M., 1992. Subductingslabs stagnant in the mantle transition zone. J. Geophys. Res. 97(B4), 4809–4822.

Gorbatov, A., Kennett, B.L.N., 2003. Joint bulk-sound and sheartomography for Western Pacific subduction zones. Earth Planet.Sci. Lett. 210, 527–543.

Gorbatov, A., Kostoglodov, V., Saurez, G., Gordeev, E., 1997. Seismic-ity and structure of the Kamchatka subduction zone. J. Geophys.Res. 102 (B8), 17,883–17,898.

Gudmundsson, O., Sambridge, M., 1998. A regionalized upper mantleseismic model. J. Geophys. Res. 103 (B4), 7121–7136.

Hasegawa, A., Horiuchi, S., Umino, N., 1994. Seismic structure of thenortheastern Japan convergent margin: a synthesis. J. Geophys.Res. 99, 22,195–22,311.

Husen, S., Quintero, R., Kissling, E., Hacker, B., 2003. Subduction-zone structure and magmatic processes beneath Costa Rica con-

ter of stratovolcanoes in southwest corner of Hokkaido,including Usu, Shikotsu, Komaga-take, among others,which are at least 200 km away from the closest vol-cano in central Hokkaido, Tokachi (Figs. 1, 6, and 9).It is possible that the gap in the volcanic front maybe due to the thickened slab beneath this region,which is due to the collision of the Japan and Kurilearcs.

The subducting slab morphology beneath theHokkaido corner is very complex, but through the com-bination of seismicity, tomography, and plate veloci-ties the complexity can be understood in terms of thecollision of the Japan and Kurile arcs. The subductedplate beneath the northwestern Pacific margin is contin-uous, but changes geometry as the slab subduction angledecreases from north to south as the slab thins. Directlybeneath the Hokkaido corner a low velocity zone in theupper mantle (70–125 km depth) is imaged, which maybe related to volcanism in the area or another result of thecollision of the two arcs. The present complex state of theKurile and northern Japan arcs with apparent crumplingmay be a result of the evolution of the arc–arc junction,

strained by local earthquake tomography and petrological mod-elling. Geophys. J. Int. 155, 11–32.

Isacks, B.L., Molnar, P., 1971. Distribution of stresses in the descendinglithosphere from a global survey of focal mechanism solutions ofmantle earthquakes. Res. Geophys. 9, 103–174.

Katsumata, K., Wada, N., Kasahara, M., 2003. Newly imaged shape ofthe deep seismic zone within the subducting Pacific plate beneaththe Hokkaido corner, Japan–Kurile arc–arc junction. J. Geophys.Res. 108 (B12), doi:10.1029/2002JB002175.

Kennett, B.L.N., Engdahl, E.R., Buland, R., 1995. Constraints on seis-mic velocities in the Earth from traveltimes. Geophys. J. Int. 122,108–124.

Kennett, B.L.N., Widiyantoro, S., van der Hilst, R.D., 1998. Joint seis-mic tomography for bulk sound and shear wave speed in the Earth’smantle. J. Geophys. Res. 103 (B6), 12,469–12,493.

Kimura, G., 1996. Collision orogeny of arc–arc junctions in theJapanese Islands. Isl. Arc 5, 262–275.

Miyamachi, H., Kasahara, M., Suzuki, S., Tanaka, K., Hawegawa,A., 1994. Seismic velocity structure in the crust and upper mantlebeneath Northern Japan. J. Phys. Earth 42 (4), 269–301.

Miyamachi, H., Moriya, T., 1984. Velocity structure beneath theHidaka Mountains in Hokkaido. Japan. J. Phys. Earth 32, 13–42.

Miyamachi, H., Moriya, T., 1987. Velocity structure and aftershockdistribution of the Urakawa-oki earthquake. J. Phys. Earth 35,309–326.

Moriya, T., 1986. Collision of forearcs and overlapped deepseismic zone in the transitional zone between the Northern


Honshu and Kurile arcs. J. Phys. Earth 34 (Suppl.), S175–S192.

Moriya, T., Okada, H., Matsushima, T., Asano, S., Yoshii, T., Ikami, A.,1998. Collision structure in the upper crust beneath the southwest-ern foot of the Hidaka Mountains, Hokkaido, Japan as derived fromexplosion seismic observations. Tectonophysics 290, 181–196.

Nakajima, J., Matsuzawa, T., Hasegawa, A., Zhao, D., 2001. Seismicimaging of arc magma and fluids under the central part of north-eastern Japan. Tectonophysics 341, 1–17.

Okada, H., 1983. Collision orogenesis and sedimentation in HokkaidoJapan. In: Hashimoto, M., Uyeda, S. (Eds.), Accretion Tectonics inthe Circum-Pacific Regions. Terra Scientific Publishing Company,Tokyo, pp. 91–105.

Seno, T., Stein, S., Gripp, A.E., 1993. A model for the motion of thePhilippine Sea plate consistent with NUVEL-1 and geological data.J. Geophys. Res. 98 (B10), 17941–17948.

Suzuki, S., Kasahara, M., 1996. Unbending and horizontal fracture ofthe subducting Pacific plate, as evidenced by the 1993 Kushiro-oki and the 1981 and 1987 intermediate-depth earthquakes inHokkaido. Phys. Earth Planet. Inter. 93, 91–104.

Tamura, Y., Tatsumi, Y., Zhao, D., Kido, Y., Shukuno, H., 2002. Hotfingers in the mantle wedge: new insights into magma genesis insubduction zones. Earth Planet. Sci. Lett. 197, 105–116.

van der Hilst, R.D., Engdahl, E.R., Spakman, W., 1993. Tomographicinversion of P and pP data for aspherical mantle structure belowthe Northwest Pacific region. Geophys. J. Int. 115 (1), 264–302.

van der Hilst, R.D., Engdahl, E.R., Spakman, W., Nolet, G., 1991.Tomographic imaging of subducted lithosphere below NorthwestPacific island arcs. Nature 353 (6339), 37–43.

Wang, Z., Zhao, D., 2005. Seismic imaging of the entire arc of Tohokuand Hokkaido in Japan using P-wave, S-wave and sP depth-phasedata. Phys. Earth Planet. Inter. 152 (2), 144–162.

Wyss, M., Hasegawa, A., Nakajima, J., 2001. Source and path ofmagma for volcanoes in the subduction zone of northeastern Japan.Geophys. Res. Lett. 28 (9), 1819–1822.

Zhao, D., Hasegawa, A., 1993. P wave tomographic imaging of thecrust and upper mantle beneath the Japan Islands. J. Geophys. Res.98, 4333–4353.

Zhao, D., Horiuchi, S., Hasegawa, A., 1992. Seismic velocity struc-ture of the crust beneath the Japan Islands. Tectonophysics 212,289–301.

Zhou, H.-W., Anderson, D.L., Clayton, R.W., 1990. Modeling of resid-ual spheres for subduction zone earthquakes: apparent slab pene-tration signatures in the NW Pacific caused by deep diffuse mantleanomalies. J. Geophys. Res. 95 (B5), 6799–6827.

morphology of the distorted subducted paciﬁc slab beneath...

Documents