Upper plate deformation associated with seamount …As subduction progresses, a shadow zone forms in...

ELSEVIER Tectonophysics 293 (1998) 207–224

Upper plate deformation associated with seamount subduction

S. Dominguez a,Ł, S.E. Lallemand a, J. Malavieille a, R. von Huene b

a Laboratoire de Geophysique et Tectonique, UMR 5573, Universite de Montpellier II, 34095 Montpellier, Franceb GEOMAR, Christian-Albrechts-Universitat, 24148 Kiel, Germany

Received 4 March 1997; accepted 24 March 1998

Abstract

In many active margins, severe deformation is observed at the front of the overriding plate where seamounts or aseismicridges subduct. Such deformation appears to be a main tectonic feature of these areas which influences the morphologyand the seismicity of the margin. To better understand the different stages of seamount subduction, we have performedsandbox experiments to study in detail the evolution of deformation both in space and time and thus complement seismicimages and bathymetry interpretation. We focus, in this paper, on the surface deformation directly comparable with seafloormorphology. Two types of subducting seamounts were modelled: relatively small conical seamounts, and larger flat-toppedseamounts. The indentation of the margin by the seamount inhibits frontal accretion and produces a re-entrant. The marginuplift includes displacement along backthrusts which propagate from the base of the seamount, and out-of-sequenceforethrusts which define a shadow zone located on the landward flank of the seamount. When the seamount is totally buriedbeneath the margin, this landward shielded zone disappears and a larger one is created in the wake of the asperity due tothe elevated position of the decollement. As a consequence, a section of the margin front follows behind the seamountto greater depth. A ‘slip-line’ network develops concurrently above the subducting seamount flanks from the transtensionalong the boundaries of the shadow zone. In a final stage, normal faults, controlled by the shape of the seamount, develop inthe subsiding wake of the asperity. Swath-bathymetric data from the Costa Rica margin reveal detailed surface deformationof the margin above three subducting seamounts. Shaded perspective views highlight the detailed structure of the seafloorand compare well with surface deformation in the sandbox experiments. The good correlation between the marine dataand experimental results strengthen a structural interpretation of the Costa Rican seamount subduction. 1998 ElsevierScience B.V. All rights reserved.

Keywords: seamount subduction; sandbox experiment; slip-line fracture network; backthrust

1. Introduction

The subduction of oceanic highs, seamounts, vol-canic chains or oceanic plateaus, involves a vigorousdeformation of the upper plate (Cloos, 1992; Scholtzand Small, 1997). Numerous bathymetric and seismicrecords show that deformation depends on several pa-

Ł Corresponding author. E-mail: [email protected]

rameters such as the nature of the asperity, the struc-ture and the dip of the oceanic plate, and also the ge-ology and the tectonic regime of the overriding plate.

Many authors have studied these problems, es-pecially in the western Pacific Ocean. These stud-ies of large seamount subduction, like the Daiichi–Kashima in the Japan Trench (Mogi and Nishizawa,1980; Tomoda and Fujimoto, 1980; Oshima et al.,1985; Kobayashi et al., 1987; Lallemand et al., 1989;

0040-1951/98/$19.00 1998 Elsevier Science B.V. All rights reserved.PII S 0 0 4 0 - 1 9 5 1 ( 9 8 ) 0 0 0 8 6 - 9

208 S. Dominguez et al. / Tectonophysics 293 (1998) 207–224

Fig. 1. Morphological analysis of seamounts located close to active margins (vertical exaggeration is 5).

Dominguez et al., 1995a), the Erimo at the Japan–Kuril Trench junction (Cadet et al., 1985; Lallemandand Chamot-Rooke, 1986; Lallemand and Le Pichon,1987; Dubois and Deplus, 1989), the Bougainville inthe New Hebrides Trench (Collot and Fisher, 1989;Collot et al., 1992) and several seamounts in the Mid-dle America Trench (Von Huene and Flueh, 1993;Von Huene et al., 1995), have played a major rolein the evaluation of deformation from a subductingasperity. Lallemand et al. (1990, 1992, 1994) haveproposed a ridge and seamount subduction sequencebased on marine data and sandbox modelling. Scholtzand Small (1997) have studied the influence of sub-ducting seamount on seismic coupling and earthquakeactivity across the subduction interface, mainly inthe Izu–Bonin and Tonga–Kermadec Trench. Cloos(1992) has also studied seismic slip and rupture areaof large earthquakes in the region of seamount sub-duction. He suggests that these earthquakes are re-lated with shearing off of the seamounts.

Generally, the first indication of a subductingseamount is given by an anomaly in the morphologyof the margin like a re-entrant, a scarp or an up-lifted bulge. In the Mediterranean ridge (Von Huene,1997), it is even possible to observe directly the topof the Bannock seamount due to the high dissolutionrate, around the subducting seamount, of the evapor-itic sediments forming the accretionary wedge. Anal-yses of magnetic anomalies (Dubois and Deplus,1989) or seismic profiles (Hinz et al., 1996; Ye et al.,1996) may help to localize the subducted asperity.

However, morphological and geophysical evi-dence is less clear when the seamount is subducted

deeply. To improve detection and interpretation, ana-logue modelling is used for additional informationon morphology and structure associated with sub-ducted seamounts.

With this objective, we will present in the firstsection of this paper, the results of experimentaldata based on sandbox experiments performed usinggranular materials and realistic boundary conditionspatterned after marine data. To better simulate fielddata, we have analyzed the morphology of volcanicseamounts at various stages of subduction. It appearsfrom this work that we can divide them into twogroups according to their size and shape. The firstgroup corresponds to seamounts less than 3 km highwith a conical shape and the second corresponds toflat-topped seamounts higher than 3 km (Fig. 1).

In the second section, the results of these generalexperimental studies are compared with subductingconical seamounts beneath the Costa Rica margin(Von Huene et al., 1995) and flat-topped seamountssubducting into the Japan Trench. These areas havebeen chosen because several seamounts are subduct-ing and swath-mapping, with high resolution and100% coverage, is available. Furthermore, we havebeen able to compare our models with high resolu-tion 2-D and 3-D shaded views of the Costa Rica andJapan margin bathymetry.

2. Experimental apparatus and procedure

Each sandbox experiment was performed undernormal gravity using an apparatus (Fig. 2) which

S. Dominguez et al. / Tectonophysics 293 (1998) 207–224 209

Fig. 2. Experimental set-up used to perform the sandbox experiments of seamount subduction.

simulates the basic geometry and the main mecha-nisms of a subduction zone. This experimental set-upis similar to the one used by Malavieille (1984) andMalavieille et al. (1991). A 1.2 m wide mylar sheet(analogue of the oceanic plate), lying on a horizon-tal plate, is pulled beneath a rigid backstop. Theuse of an inclined basal plate does not change themechanisms of the deformation. Indeed, we havealready observed in previous experiments that theonly significant effect is a slight modification oftopographical slopes.

This device allows more than 1.5 m of conver-gence for each experiment. In front of the backstop,a wedge was constructed that represents the structurecommonly observed in seismic records. This wedgewas constructed of a cohesive material (compactedrock powder or mortar) to simulate a deformable

Fig. 3. Schematic cross-section of the experimental set-up showing the internal structure at initial stage.

buttress. Then, a low cohesive sand wedge was builtto simulate an accretionary wedge in front of thebuttress. On the mylar sheet, successive colouredsand layers simulated the oceanic sediments (Fig. 3).The seamount consisted of a rigid core attached tothe mylar sheet and was covered with low cohesivesand. The aeolian sand used in this study is roundedwith a grain size of less than 300 µm and a density² D 1600 kg=m3. The internal coefficient of frictionis ¼ D 0:6 and the cohesion C0 D 20 Pa.

The granular materials used to build the marginand the oceanic sediments have frictional propertiessatisfying the Coulomb theory (Dahlen et al., 1984;Dahlen, 1984) and thus, they are fair analogues ofthe rocks of the upper crust (Krantz, 1991). Scalingis discussed in Lallemand et al. (1992), and 1 cm inour experiment is equivalent to 1 km in nature. Two


cameras and one video recorder recorded each stageof deformation. Structural interpretations are basedon planar views recorded at the same time as themorphological perspective views presented.

3. First experiment: Subduction of a conicalseamount

During the first stage of subduction, the seamountindents the base of the inner trench slope (Fig. 4aand Fig. 5a). The front of the margin is uplifted andslides on the subducted flank of the seamount. At thisstage, some thrusts reactivation can be observed dueto the increase of the dip angle of the decollement,which is now located on the landward flank of theseamount.

As subduction progresses, a shadow zone formsin the wake of the seamount (Fig. 4b and Fig. 5b)because the decollement is deviated upwards abovethe seamount landward flank. The detachment risesto a higher level in the front of the margin andthe sediments located underneath subduct, then, withlittle deformation.

In the wake of the seamount, frontal accretionstops and the oceanic sedimentary sequence subductswithout any deformation. On both sides of the re-entrant, new imbricates are added to the accretionarywedge. Close to each side of the re-entrant, thesenew imbricates are curved in the direction of theconvergence and connect with the landward flank ofthe subducted seamount.

The scarp, associated with the re-entrant, is af-fected by intense mass-sliding. Slumped masses pro-ceeding from the uplifted part of the margin, depositon the subducting sequence located in the shadowzone (Fig. 4b and Fig. 5b). They are subducted inthe wake of the seamount as long as the decollementcrops out in the upper part of the frontal margin.Such processes allow material from the margin to besubducted to greater depth.

After a certain amount of seamount subduction(15 cm), the decollement switches to a deeper level,located at the top of the subducting sequence. Thenfinally, the decollement returns to its basal level(at the base of the subducting sequence) causingunderplating of the unit previously dragged behindthe seamount.

Contraction in front of the seamount is largelyaccommodated by backthrusts. The backthrusts arecurved landward and have a small lateral extension.Offsets are about 1 to 2 mm which can be comparedwith 100 m in nature.

Associated with the differential shortening, atranstensive fracture network develops (Dominguezet al., 1995b, 1996). These fractures are directly in-duced by the indentation of the margin (Fig. 4b andFig. 5b). The fracture network consists of subverticalstrike-slip faults that follow the slip-line orientation.They grow above the top of the subducting seamountand diverge slightly on both sides from the directionof the convergence (Fig. 4c and Fig. 5c). The geome-try of the fractured zone can be compared with a fanstructure, developing intermittently as the seamountsubducts. The length of these faults is a few centime-tres in our experiment (several kilometres in nature)and the offset is about 1 to 2 mm (100–200 m innature). The root zone of the slip-lines is intensivelyfractured by subvertical faults and forms a weak zonewhere mass-sliding (and canyon erosion in nature)is active. When the seamount is further subducted(Fig. 4c and Fig. 5c), frontal accretion resumes. Thetrench fill, in the former shadow zone located in thewake of the seamount, accretes because the decolle-ment moves from the top of the subducting sedimentsequence to its base (Dominguez et al., 1994).

Above the seamount, local uplift of the marginforms a circular knoll whose geometry and volumeare controlled by the shape and size of the asperity.The point of maximum uplift is located landward ofthe top of the subducting seamount. In this region,small horsts and grabens (few millimetres offsets)trending parallel to the direction of the convergence,develop above the subducting asperity. The bendingof the margin above the subducted seamount pro-duces a normal fault network. These faults are super-imposed on the ‘slip-line’ fracture network (Fig. 4dand Fig. 5d).

As the whole seamount subducts beneath the co-hesive part of the margin, the backthrust spacingincreases (Fig. 4d and Fig. 5d). These are cut bylandward-dipping normal faults that develop sub-sequently in the wake of the subducted seamount.Displacements on these normal faults are only a fewmillimetres. The subsided area, observed behind thecrest of the seamount, is elongated in the direction

S.Dom

inguezet

al./Tectonophysics293

(1998)207–224

211

Fig. 4. (a through d) First experiment. Subduction of a conical seamount.


Fig. 5. (a through d) Tectonic interpretation of the first experiment. The circle outlines the location of the subducted seamount.

of the convergence. It could be filled by sedimentsand then form a basin. It develops when frontal ac-cretion resumed and corresponds to the underplatingof the sediments situated in the shadow zone. As the

seamount subducts, extension, expressed by normalfaulting, occurs in the wake.

After 50 cm of convergence, deformation re-lated to the seamount subduction ends because the


seamount is deeply subducted and located fartherfrom the front of the accretionary wedge. Only agentle uplift without any fracture network can bedetected.

To better study the scarps associated with the frac-ture network, we performed the same experiment us-ing a very fine granular material to favour fracturingand a larger seamount to increase fault displacement(Fig. 6a,b). The ‘slip-line fracture network’ is clearlyexpressed in surface morphology and we observe theright- and the left-lateral displacement of conjugatedstrike-slip faults accommodating the indentation ofthe margin by the seamount. The normal faults arealso clearly observable in this experiment above thetrailing flank of the seamount (Fig. 6a,b).

We have also performed sandbox experiments tostudy the influence of seamounts slope angles. Withthis objective, we have used three seamounts withthe same diameters but different heights (Fig. 6c).It appears from these experiments that the shape ofthe subducting seamount directly controls the devel-opment of the backthrust and the slip-line network.Seamounts with flat slopes (less than 5º) do notgenerate slip-line faults and the backthrust spacingis very short. Seamounts with steeper slopes (morethan 5º) produce a well developed slip-line networkand widely spaced backthrusts (Fig. 6d).

4. Second experiment: Subduction of aflat-topped seamount

The experiment was performed using the samematerials and parameters as previously. The maindifferences are the shape of the asperity (flat-topped,higher summit and lower slope angles) and its largersize compared with the accretionary wedge and themargin (Fig. 7a).

The main processes, obtained during guyot sub-duction, are comparable to those produced during themain stages of a conical seamount subduction. How-ever, due to the size and the shape of the asperity, thedeformation of the margin differs slightly.

The indentation of the margin by the guyot rapidlyaffects the accretionary wedge. Shortening is accom-modated by reactivation of landward-dipping out-of-sequence thrusts. A shielded zone develops behind thetrailing flank of the seamount (Fig. 7b). A very large

uplifted zone cut by numerous backthrusts devel-ops over the seamount leading flank and over-criticalslopes with gravity slides develop on the trailing flank.

On both sides of the re-entrant, the wedge isdragged upslope and two lateral conjugated shearzones, composed of several strike-slip faults with ver-tical displacement, develop (Fig. 7c,d). These faultsbound a triangular-shielded zone in the wake of theguyot which allows a great part of the margin front tobe dragged landward behind the subducting guyot.

When the seamount is totally subducted beneaththe front of the margin, the large uplifted area pro-duced by the guyot becomes subcircular and thebackthrust spacing increases rapidly. At the sametime, the height of the bulge decreases. Frontal ac-cretion resumes in the wake of the seamount andthe scarp associated with guyot subduction becomesinactive. The two lateral shear-bands continue prop-agating landward and become subparallel (Fig. 7d).

Superimposed on the backthrusts is a dense net-work of small slip-lines, subparallel to the directionof the convergence. This network is not strongly di-vergent as in the case of a conical seamount which isprobably related to the flat top of the seamount andthe low angle of its slopes. The root zone of thesefaults is intensely fractured, displacement is smallbut they extend laterally over more than 20 cm (20 to30 km in nature).

In the very large subsided area (Fig. 7d), trans-verse normal faults and late-stage strike-slip faults onboth sides of the re-entrant cut early-stage structures.These faults are similar to those observed in theconical seamount experiment. They produce a largesubcircular basin bounded seaward by the upliftedaccretionary wedge dragged upslope in the wake ofthe guyot (Fig. 7c).

At the end of the experiment, the main morpho-tectonic features were: a re-entrant in the front ofthe margin about three times the diameter of theguyot, an elongated trough in its wake and an areahighly deformed by backthrusts and transverse verti-cal faults above the subducting guyot.

5. Swath-bathymetry reprocessing

The experimental results have been compared toreprocessed swath-mapping (Hydrosweep) morphol-

214S.D

ominguez

etal./Tectonophysics

293(1998)

207–224

Fig. 6. (a) Perspective views and tectonic interpretation of an experiment of conical seamount subduction. The fractures following the slip-lines are well observable. (b)Tectonic interpretation of a conical seamount subduction showing the relations between the different fracture networks. (c, d) Perspective views of a sandbox experimentshowing the relations between slip-line, backthrust development and subducting seamount shapes.


ogy in areas of seamount subduction. It is commonpractice to smooth bathymetric data to eliminate badsoundings and other artefacts. Reprocessing of rawdata and unfiltered soundings are presented to re-solve the very fine morphotectonic features whichwould normally have been discarded. Perspectiveimages with artificial illumination generated with theGMT software (Wessel and Smith, 1991), allowed usto significantly improve resolution. Examples fromthe Costa Rica margin (Fig. 8) and the Japan Trench(Fig. 12) are compared with the sandbox experimentspreviously described to allow a detailed structural in-terpretation.

6. Effect of conical seamount subduction asillustrated in Costa Rica

Along this portion of Middle America, the Co-cos plate, of Late Miocene age subducts beneath theCaribbean plate at a rate of about 9 cm=year (Demetset al., 1990). On the basis of seismic data, Bour-gois et al. (1984), Crowe and Buffler (1985), haveproposed that the Costa Rica margin is mainly con-stituted by the seaward prolongation of the Nicoyaultramafic rocks complex. They have interpreted thegeomorphology of the margin offshore as the resultof an extensional tectonic regime affecting the wholeoverriding plate.

Most recent works have shown that this inter-pretation cannot explain the margin deformationsobserved. The first description of seamount subduc-tion beneath the Costa Rica margin (Von Huene andFlueh, 1993; Lallemand et al., 1994) identified threeseamount tracks across the margin. The high resolu-tion morphology indicates that the tectonics of theCosta Rica margin is significantly affected by thesubduction of seamounts on the Cocos plate (Fig. 8).Several local uplifted zones, associated seaward withsedimentary slides, contrast with the smoothed mor-phology of the rest of the margin.

To improve the resolution, different angles ofillumination were viewed to derive a high resolutiontectonic map (resolution of about 50 m to 100 m,depending on depth). After processing, the tectonicsof subducting seamounts was at a scale similar tothe sandbox model and the structural map comparedbetter with the sandbox experiments.

The studied area (Fig. 8) corresponds to a verydeformed part of the Costa Rica margin. In thisarea, three well-developed seamount tracks illustratethree stages of seamount subduction. Fig. 9 showsdifferent angles of illumination of the third re-entrantto reveal the very fine fracture network above thesupposed position of the subducted seamount.

The first re-entrant (9º000N, �84º400E) in thesoutheast shows morphotectonic features typical ofthe preliminary stages of subduction observed in thesandbox experiments (Fig. 9). The subcircular scarpassociated to this re-entrant is nearly 1.5 km high andshows recent gravity collapse. The wedge is indentedand shows a re-entrant more than 8 km long. Frontalaccretion is inhibited in a shielded zone and occursonly on both sides of the re-entrant. The decollementcomes to the surface at the top of the oceanic sed-iment sequence. The circular bulge associated withthe re-entrant (dashed line on Fig. 10) is 13 km long,15 km wide and 1 km high. All these features sug-gest a conical shape for the subducted volcanic reliefas seen from other seamounts exposed on the oceanicplate (Fig. 8). The fractures due to the indentationof the margin by the seamount are not very well ex-pressed at this stage, but some evidence of strike-slipfaults (slip-lines oriented fractures network) can beobserved at the toe of the scarp. The vertical dis-placement on these faults could be about 20–50 mwhich is near the limit of resolution in the bathymet-ric data. Furthermore, since sediment deposition israpid on the Costa Rican margin, there is little timefor fracture development. Upslope from the bulge,some backthrusts and a faint divergent pattern oc-curs. We can also observe two diverging canyonswhich shape is controlled by the landward propa-gation of the bulge associated with the subductedseamount.

The second re-entrant (9º050N, �84º500E) in themiddle of the area corresponds to a more advancedstage of subduction. This re-entrant has two lateralscarps, about 1 km high, trending parallel to thedirection of convergence. Its landward headwall isformed by active landslides located more than 25km from the trench. The geometry of this re-entrantsuggests subduction of a seamount comparable insize with the previous one.

The present front of the margin at the seamounttrail is not indented because the seamount has mi-

216S.D

ominguez

etal./Tectonophysics

293(1998)

207–224

Fig. 7. (a, b, c) Second experiment. Subduction of a flat-topped seamount (guyot).


Fig. 7 (continued). (d) Structural interpretation of the secondexperiment corresponding to the final stage viewed from above.

grated sufficiently far to allow the frontal accretionto resume. The initial V-shape indentation remainsobservable and is comparable to the one produced bythe first seamount.

Immediately seaward of the steepest part of thescarp is a small flat area that may correspond to thebeginning of the subsidence observed in the sandboxexperiments as the seamount was deeply subducted.Subsidence along previous strike-slip faults whichdeveloped behind the seamount forms the seamounttrail across the accretionary wedge and backstop.There are no normal faults perpendicular to theconvergence. Perhaps the cohesive part of the margin(the relative deformable backstop) was just recentlyaffected by the asperity. The sediments in the wakeof the subducted seamount probably belong to thepre-existing accretionary wedge and slope depositsand are not cohesive enough to produce these faults.As shown in the sandbox experiments, the normalfaults develop when the seamount starts subductingbeneath the cohesive part of the margin.

Above the position of the subducted asperity out-lined by the magnetic anomalies (Von Huene et al.,1995) is a circular bulge, 15 km in diameter and500 m high. This uplifted zone is cut by many smallfaults. High resolution imagery shows several curvedbackthrusts dipping seaward with little displacement(<30 m) close to the landward boundary of thebulge.

A dense network of faults oriented along slip-linestrends is now observable in the reprocessed data. Thefan-shape distribution of the faults is comparable tothe experiment in which the conical seamount wassubducted (Fig. 6c). Their lengths are about 5 to10 km and the maximum offset associated to thesefaults is about 100 m.

The third seamount trail observed in the areais probably a large surface scar resulting fromseamount subduction even earlier than the previ-ous one. The long trail between the trench and theeroded scarp indicates a seamount deeply buried be-neath the margin. There is no associated bulge onthe margin but the contours show a wide elongatedremnant basin, suggesting that the asperity whichcaused this trail was probably larger than those thatcaused the two previous ones.

7. Case of a flat-topped seamount as illustratedby the Daiichi–Kashima seamount in the JapanTrench

Very few flat-topped seamounts have been de-scribed in detail (with swath-bathymetry available)in subduction zones. The two most famous arethe Daiichi–Kashima in the Japan Trench and theBougainville in the New Hebrides Trench. Unfortu-nately these guyots are in the first stages of sub-duction and there is no clear evidence in any activemargin that a guyot is presently subducting. How-ever, the subduction of the Daiichi–Kashima, despitethe fact that a huge normal fault affects the guyot(Lallemand et al., 1989; Dominguez et al., 1995a),deforms significantly the Japan margin.

First, its western flank is completely subductedbeneath the front of the Japan margin. A bulge witha curved shape can be clearly observed, as well asfractures, using 3-D shaded views (Fig. 11). Thesefractures probably correspond to the incipient de-


Fig. 8. General 2-D illuminated view of the Costa Rica margin showing several re-entrants caused by conical seamount subductions.

velopment of a slip-line fracture network associatedwith the beginning of the margin indentation.

Second, the Daiichi–Kashima belongs to a vol-canic chain of big seamounts and, as noticed byLallemand et al. (1989), there is a possibility thatsome of them have been previously subducted. Inthe upper part of the margin, we can observe anuplifted area. The presence of two diverging canyons(like in the Costa Rica margin), curved ridges (back-thrusts) and an eroded area (remnant scarp of there-entrant), suggests that a big seamount is deeplyburied beneath the Japan margin (Fig. 11).

Nevertheless, the average quality of swath-bathymetry (Seabeam), the incomplete survey of thisregion and the complex morphology of the marginmake a more detailed interpretation difficult.

8. Discussion

The good fit between sandbox experiments andnatural examples (Costa Rica and Japan seafloormorphology), suggests a fairly good replication ofseamount subduction. Based on this evidence andreprocessed morphology, we have studied the me-chanics of the deformation in the sandbox modelsto better understand mechanisms in active margins.The subducting seamounts have shaped the segmentof the Costa Rica margin, at least during Plio–Quaternary times if we assume that the third imprinton Fig. 10 was caused by a huge subducting asperity.

They were responsible for: (1) short interrup-tions of frontal accretion and re-entrants into theaccretionary wedge; (2) net removal of accretionary


Fig. 9. Zooms of the second re-entrant, with different angles of illumination, showing the fan-shape fault network following the slip-lines.


Fig. 10. Tectonic map of the studied area in the Costa Rica margin.

wedge material from the front of the margin; (3)transverse sedimentary traps in the wake of subduct-ing seamounts; (4) weak zones along the seamounttrails. These weak zones are characterized by thedense network of subvertical faults as the top of theseamount migrates beneath the wedge that are laterreactivated as normal faults.

The decollement moulds the subducted seamount(and laterally, it tends to close again in its wake(Fig. 12a).

A continuing progression of backthrusts and frac-tures develops across the margin. First, backthrustsnucleate as the compression increases in front ofthe indenter. Second, a first generation of fan-shaped

S.Dom

inguezet

al./Tectonophysics293

(1998)207–224

221

Fig. 11. 3-D shaded view of the Japan margin in the area of the Daiichi–Kashima guyot. Vertical exaggeration is about 5.


Fig. 12. (a) Sketch showing the shape of the decollement in the area of a subducted seamount. Gridding represents the maximum andminimum stress directions. (b) Sketch showing the geometry of the fault planes on the landward part of the subducting seamount at thetime of their nucleation. Only one set of strike-slip faults and one backthrust is represented on this cartoon. During seamount subduction,other sets of faults superimpose on the first one.

subvertical fractures develops (Fig. 12b) whose ge-ometry is controlled by the asperity shape. Then,both backthrusts and subvertical fractures are reacti-

vated during the next stages of seamount subductionuntil compression reaches again the yield limit of thematerial and another generation of backthrusts and


then subvertical fractures develops. The final stageinvolves the interaction between pre-existing faultsand newly activated ones.

9. Conclusion

The impact of the subducting seamounts or vol-canic ridges on the tectonics of active margins isgenerally well documented in the literature. Naturalexamples have been described but the kinematics andthe fault networks associated with the subduction ofsuch oceanic relief is poorly understood. As we haveseen in a comparison between experimental resultsand the Costa Rica margin, the sandbox experimentsof seamount subduction are a useful aid to improveinterpretation of the geomorphology associated withthis process. We describe the weak zone above andin the wake of the asperity. The indentation of themargin by the seamount causes a fan-like fault net-work associated with backthrusts. The margin re-covers its initial slope as the seamount is deeplysubducted. These experiments allow us to observethe kinematics and the timing of deformation andthe interactions between concurrent compressionaland extensional deformations above the subductingseamount.

Results of the analogue modelling andswath-mapping, show that the subduction of vol-canic asperities involves very strong deformationand intense fracturing of the margin. They affectsignificantly the structure of the overriding plate,allowing material from the front of the margin tobe subducted to greater depth. As a result of thisintense deformation of the margin, they also proba-bly modify significantly the fluid circulation and theseismic coupling. As suggested by Scholtz and Small(1997) and considering the deformation mechanismsoutlined in our study, we agree that large subductingseamounts increase locally the normal stress acrossthe subduction interface. This could explain the largeinterplate earthquakes events recorded in areas ofseamount subduction. The other explanation, as sug-gested by Cloos (1992), is that the seamounts arefinally shearing off when they are deeply subductedbeneath the inner part of the overriding plate. Webelieve that sandbox experiments could clarify thispoint.

Acknowledgements

The authors wish to thank D. Davis, H. Kerr,J. Bourgois and C. Scholz for their reviews of themanuscript. We also thank B. Sanche for his tech-nical support in performing sandbox experiments,Niita is acknowledged for the grammatical correc-tions of the manuscript.

References

Bourgois, J., Azema, J., Baumgartner, P.O., Tournon, J., Desmet,A., Aubouin, J., 1984. The geologic history of the Caribbean–Cocos plate boundary with special reference to the Nicoyaophiolite complex (Costa Rica) and D.S.D.P. results (Legs 67and 84 off Guatemala): a synthesis. Tectonophysics 108, 1–32.

Cadet, J.P., Kobayashi, K., Aubouin, J., Boulegue, J., Dubois,J., von Huene, R., Jolivet, L., Kanazawa, T., Kasahara, J.,Koizumi, K., Lallemand, S.E., Nakamura, Y., Pautot, G.,Suyehiro, K., Tani, S., Tokuyama, H., Yamazaki, T., 1985.De la fosse des Kouriles: Premiers resultats de la campagneoceanographique franco–japonaise Kaiko (Leg 3). C. R. Acad.Sci. Paris 301, 287–296.

Cloos, M., 1992. Thrust-type subduction-zone earthquake andseamount asperities: A physical model for seismic rupture.Geology 20, 601–604.

Collot, J.Y., Fisher, M.A., 1989. Formation of forearc basinsby collision between seamounts and accretionary wedges: anexample from the New-Hebrides subduction zone. Geology17, 930–933.

Collot, J.Y., Lallemand, S., Pelletier, B., Essein, J.P., Glacon, G.,Fisher, M.A., Greene, H.G., Boulin, J., Daniel, J., Monzier,M., 1992. Geology of the d’entrecasteaux New-HebridesArc collision zone: results from a deep submersible survey.Tectonophysics 212, 213–241.

Crowe, J.C., Buffler, R.T., 1985. Multichannel seismic recordsacross the Middle America trench and Costa Rica–Nicaraguaconvergent margin, NCY-7 and NCI-1. In: Middle AmericaTrench off western Central America, Ocean Drilling ProgramRegional Data Synthesis Series, Atlas 7, pp. 6–11.

Dahlen, F.A., 1984. Noncohesive critical Coulomb wedges: anexact solution. J. Geophys. Res. 89, 10125–10133.

Dahlen, F.A., Suppe, J., Davis, D., 1984. Mechanics of fold-and-thrust belts and accretionary wedges: cohesive Coulomb the-ory. J. Geophys. Res. 89, 10087–10101.

Demets, C., Gordon, R.G., Argus, D.F., Stein, S., 1990. Currentplate motions. Geophys. J. R. Astron. Soc. 101, 425–478.

Dominguez, S., Lallemand, S.E., Malavieille, J., 1994. New re-sults from sandbox modeling of seamount subduction and pos-sible applications (abstr.). AGU Fall Meeting, San Francisco,EOS 75 (44), p. 671.

Dominguez, S., Lallemand, S.E., Malavieille, J., Pouclet, A.,1995a. Nouvelle interpretation structurale du mont sous-marin


Daiichi–Kashima et essai de reconstitution geodynamique. C.R. Acad. Sci. Paris 320, 403–409.

Dominguez, S., Lallemand, S.E., Malavieille, J., 1995. Physicalsandbox modelling of seamount subduction, effects on thetectonic of accretionary wedges. Abstr. EUG 8, Terra Abstr.,p. 245.

Dominguez, S., Lallemand, S.E., Malavieille, J., 1996. Effetsde la subduction de reliefs oceaniques sur la deformation desmarges actives: Approche experimentale, 16ieme Reunion desSciences de la Terre-Orleans. Soc. Geol. Edit. Paris, 23 pp.

Dubois, J., Deplus, C., 1989. Gravimetry on the Erimo seamountJapan, subduction zones: the KAIKO project. Tectonophysics160, 267–275.

Hinz, K., von Huene, R., Ranero, C.R., 1996. Tectonic structureof the convergent Pacific margin offshore Costa-Rica frommultichannel seismic reflection data. Tectonics 15, 54–66.

Kobayashi, K., Cadet, J.P., Aubouin, J., Boulegue, J., Dubois,J., von Huene, R., Jolivet, L., Kanazawa, T., Kasahara, J.,Koizumi, K., Lallemand, S.E., Nakamura, Y., Pautot, G., Suye-hiro, K., Tani, S., Tokuyama, H., Yamazaki, T., 1987. Normalfaulting of Daiichi–Kashima seamount in the Japan Trenchrevealed by the Kaiko 1 cruise, Leg 3. Earth Planet. Sci. Lett.83, 267–284.

Krantz, R.W., 1991. Measurements of friction coefficients andcohesion for faulting and fault reactivation in laboratory mod-els, using sand and san mixtures. Tectonophysics 188, 203–207.

Lallemand, S.E., Chamot-Rooke, N., 1986. Sur la cause dudecrochement senestre entre les fosses du Japon et desKouriles: subduction-collision d’un ancien volcan sous-marin.C. R. Acad. Sci. Paris 303 II 16, 1443–1448.

Lallemand, S.E., Le Pichon, X., 1987. Coulomb wedge modelapplied to the subduction of seamounts in the Japan Trench.Geology 15, 1065–1069.

Lallemand, S.E., Culotta, R., von Huene, R., 1989. Subduc-tion of the Daiichi–Kashima seamount in the Japan Trench.Tectonophysics 160, 231–247.

Lallemand, S.E., Collot, J.Y., Pelletier, B., Rangin, C., Cadet,J.P., 1990. Impact of oceanic asperities on the tectogenesis ofmodern convergent margins. Oceanol. Acta 10, 17–30.

Lallemand, S.E., Malavieille, J., Calassou, S., 1992. Effects ofoceanic ridge subduction on accretionary wedges: Experimen-tal modeling and marine observation. Tectonics 11, 1301–1313.

Lallemand, S.E., Schnurle, P., Malavieille, J., 1994. Coulombtheory applied to accretionary and nonaccretionary wedges:Possible causes for tectonic erosion and=or frontal accretion.J. Geophys. Res. 99, 12033–12055.

Malavieille, J., 1984. Modelisation experimentale des chevauche-ments imbriques: application aux chaınes de montagnes. Bull.Soc. Geol. Fr. 7, 129–138.

Malavieille, J., Larroque, C., Lallemand, S.E., Stephan, J.F.,1991. Experimental modelling of accretionary wedges. TerraAbstr. 3, 367.

Mogi, A., Nishizawa, K., 1980. Breakdown of a seamount on theslope of the Japan trench. Proc. Jpn. Acad. 56, 257–259.

Oshima, S., Ogino, T., Katsura, T., Ikeda, K., Uchida, M.,Nagano, M., Hayashida, M., Muneda, K., Kasuga, S., 1985.Subduction of Daiiti–Kasima seamount into de landward slopeof the Japan trench. Rep. Hydrol. Res. 20, 25–46.

Scholtz, C.H., Small, C., 1997. The effect of seamount subduc-tion on seismic coupling. Geology 25 (6), 487–490.

Tomoda, Y., Fujimoto, H., 1980. Interaction between trench andseamount in case of Kasima Seamount Nı1. Hydrol. Dep.,Maritime Safety Agency 26, pp. 493–499.

Von Huene, R., 1997. Mediterranean ridge structure: results fromIMERSE. Eos, April 15.

Von Huene, R., Flueh, E.R., 1993. A review of marine geophys-ical studies along the Middle America trench off Costa-Ricaand the problematic seaward terminus of continental crust.Profil 7, 143–159.

von Huene, R., et al., 1995. Morphotectonic features of the CostaRican Pacific margin surveyed during the Sonne 76 cruise.Geol. Soc. Am. Spec. Pap. 295, 291–307.

Wessel, P., Smith, W.H.F., 1991. Free software helps map anddisplay data. Eos 72, 441.

Ye, S., Bialas, J., Flueh, E.R., Stavenhagen, A., von Huene, R.,1996. Crustal structure of the Middle American Trench offCosta-Rica from wide-angle seismic data. Tectonics 15, 1006–1021.

Upper plate deformation associated with seamount …As subduction progresses, a shadow zone forms in...

Documents

Transcript of Upper plate deformation associated with seamount …As subduction progresses, a shadow zone forms in...