<<

Deformation of accretionary wedges in response to seamount : Insights from sandbox experiments S Dominguez, J Malavieille, Serge E. Lallemand

To cite this version:

S Dominguez, J Malavieille, Serge E. Lallemand. Deformation of accretionary wedges in response to seamount subduction: Insights from sandbox experiments. , American Geophysical Union (AGU), 2010, 19 (1), pp.192-196. ￿10.1029/1999TC900055￿. ￿hal-01261523￿

HAL Id: hal-01261523 https://hal.archives-ouvertes.fr/hal-01261523 Submitted on 28 Jan 2016

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in France or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés. TECTONICS, VOL. 19, NO. 1, PAGES 182-196 FEBRUARY 2000

Deformation of accretionary wedgesin responseto seamount subduction- Insights from sandbox experiments

S. Dominguez,J. Malavieille, and S. E. Lallemand Laboratoirede G6ophysique,Tectonique et S6dimentologie,UMR 5573, CNRS Institutdes Sciences de la Terre, de l'Eau et de I'Espacede Montpe!!ier Universit6de Montpellier II, Montpellier,France

Abstract. Sandboxexperiments, using a two-dimensional and al., 1998, 1999; Park et al., 1999], the subduction of these a three-dimensional approach, are used to study the volcanic highs strongly deformsthe overriding plate margin deformation of margins in response to seamount subduction. and also influences the seismicity across the subduction Successivemechanisms of deformationare activated during interfaceand the magmaticactivity of the volcanic arc [Nut the subduction of conical seamounts. First, reactivation of the and Ben-Avraham, 1981; McGeary, 1985; Tatchef, 1989]. frontal thrusts and compactionof the accretionarywedge is Recent works propose that when these volcanic highs reach observed. Then, back thrusting and, conjugate strike-slip the seismogeniczone, they modify,the seismiccoupling and faulting develops above the leading slope of the subducted favor the nucleation of strong earthquakes [Christensen and seamount.The basal d6collementis deflectedupward in the Lay; 1988; Cloos, 1992; Scholtz and Small, 1997]. Small wake of the subducting high, and a large shadow zone and Abbott[1998] evensuggest that the subductionof very develops behind the seamounttrailing slope. Consequently, largeseamounts (e.g., the Louisvilleridge in the Tonga- frontal accretionis inhibited and part of the frontal margin is KermadecTrench) could fracturethe surroundingoceanic dragged into the subduction zone. When the main crust. d6collement returns to its basal level in the wake of the Geophysicaldata recordedduring oceanographiccruises seamount,the margin records a rapid subsidence and a new allow detailed studiesof the morphologyand structureof accretionarywedge develops, closing the margin reentrant. convergent margins. Nevertheless,while seismic profiles The underthrusted in the wake of the seamount into succeedin resolvingthe geologicstructure of the upperplate the shadowzone, are underplated beneaththe rear part of the in relativelyundeformed regions of a margin,interpretation accretionarywedge. Substantialshortening and thickening of becomes difficult when the subduction of a seamount occurs. the deformable seaward termination of the upper plate In suchareas the accretionarywedge and the fi'ontalmargin , associated with basal erosion is observed. areso stronglydeformed that the internalstructure is poorly Seamount subduction induces significant material transfer resolved. within the , favors large tectonic erosion Sandboxexperiments already have been used to study of the frontal margin and thickening of the rear part of the sedimentaryaccretionary wedges and seamountsubduction margin. The subduction and underplating of relatively [Davis et al., 1983; Malavieille, 1984; Lallemand et al., undeformed, water-ladden sediments, associated with fluid 1992; Kukowskiet al., 1994; Gutschef et al., 1998l. It expulsion along the fractures affecting the margin could appearsfrom thesestudies that analogexperiments aflbrd good modify the fluid pressure along the basal ddcollement. tools to complementthe marine observations. Consequently, significant variations of the effective basal We have usedsimilar sandbox experiments to investigate friction and local mechanicalcoupling betweenthe two plates the structuralevolution of the overridingplate marginin could be expected around the subductingseamount. responseto seamountsubduction. We analyze structuresin two andthree dimensions, the evolution in time and spaceof 1. Introduction the stressfield induced by the seamount,and we estimatethe amount of material transfer within the margin. The main 1.1. Geologic Setting purpose of this experimentalstudy is to improve our understandingof the deIbrmationmechanisms occurring Large domainsof oceanicplates (Figure 1) are characterized during seamount subduction and to propose some by a very rough seafloor covered by numerous seamounts, quantitativeresults. Indeed, since fluids and pore pressureare aseismicridges, and volcanic plateaus [Batiza, 1982; Smith not considered,angles of reposeand angles of taperare greater and Jordan, 1988; Wessel and Lyons, 1997]. As revealed by than those observed in nature. Nevertheless, as demonstrated marine data and large-scale observations [Ballance et al., by morphologiccomparisons between previous experiments 1989; Collot and Fisher, 1989, 1991; Fisher et al., 1991; and natural cases[Dominguez e! al., 1998], we believe that Lallemand et al., 1990; von Huene and Scholl, 1991; von the mechanismsobserved in such sandboxexperiments are Huene et al., 1995; Moore and Sender, 1995; Dominguez et analog to those occurringin subductionzones. Figure2 summarizesthe main morphologicti:atures induced Copyright2000 by theAmerican Geophysical Union. by seamount subduction, presented in previous works [Dominguezet al., 1994, 1998]. The accretionarywedge is Papernumber 1999TC900055. indented,and a large reentrantdevelop associatedwith a 0278-7407/00/1999TC00055512.00 critical frontal slope, where massivesedimentary sliding

182 DOMINGUEZET AL.: SANDBOXEXPERIMENTS OF SEAMOUNTSUBDUCTION 183

120' 130' 140' 150' 160' 170'40' 40'

30'

20' 20'

10'

120' 130' 140' 150' 160' 170' Figure1. Shadedview of the Pacific and Philippine Plates showing themorphologic features ofthe seafloor. Areasof high density ofseamounts areoutlined aswell as regions ofseamount orridge subduction (circles).

occurs.Landward of the reentrant,the margin is uplifted sedimentaryrocks and marinesediments. The basal friction above the subductedseamount and shows a characteristic along the sand-mylar interfaceis intermediate(Ob=20ø). faultnetwork. Seaward dipping back thrusts and conjugated Scaling is such as I cm in the sandboxexperiment is strike-slipfaults, related to the indentationof the margin, equivalentto I km in nature [Lallemandet al., 1992]. The propagatelandward as the seamountsubduct. cohesion of the sand and the cohesion of the rock powder, used in our experiments,are 20 Pa and about 100 Pa, 1.2. Experimental Setup respectively,and scaleto 2 and 10 MPa [Gutschef eta!., Thetwo experimental devices used to performthe sandbox 1998]. Thesecohesion values are very close to values of experiments(Figure 2)are comparableto those used by unconsolidated marine sediments and lithified sedimentary Malavieilleet al. [1991] andDominguez et al. [1994; 1998]. rocks [Hoshinoet al., 1972]. Both consistof an inclinedrigid plate on which a mylar sheet, The shapeof the subductinghigh is conicalwith a slopeof simulatingthe subductingoceanic plate slides,is located. 20ø. It consistsof a rigid corecovered by a 2 cmthick layer of This sheet is pulled beneath a rigid backstopwhich cohesive rock powder with a thin (< I cm)sand cover, representsthe undeformablepart of the overridingplate. A simulating a volcanoclasticcover. The friction along the cohesivesand wedge is built in front of this backstopto interfacebetween the rigid core and the cohesive layer is simulatethe morecompacted inner part of the accretionary moderateto high (closeto •b=25ø). ,wedgeand the morecohesive seaward termination of the The first device, used to study in cross sections the overridingplate [yonHuene and Scholl,1991]. The oceanic differentstages of a seamountsubduction, has two lateral sedimentscarried by the subductingplate are built by glasspanes. They allow the deformationto be observedin real sprinklinghorizontal layers of coloredsand on the mylarfilm. time on both sides of the experiment.The second device, The sandlayers and backstop wedge are composedof granular dedicated to the study of the margin structure in three material(sand and rock powder). The sand used in our dimensions,is very muchlarger to avoid boundaryeffects. On experimentsis well-sortedeolian quartz sand with a grain thisapparatus the modelcan be impregnatedwith waterat the sizeof about250 ILtm.Its physical•properties (angle of internal endof the experimentto increasethe sandcohesion and to cut frictionOs of 30ø anda very low cohesion)satisfy the Mohr- vertical and also horizontal cross sections. Coulombcriteria [Dahlen, 1984, 1990; Dahlen and Suppe, The results of three selected sandbox experiments are 1984]. They can be consideredas good analogsof natural describedin this paper.The first one showsin crosssection 184 DOMINGUEZ ET AL.: SANDBOX EXPERIMENTS OF SEAMOUNT SUBDUCTION

Camera Rigidbackstop Accretionary wedge Seamounts Glasses 30 cm

Mylar sheet

Mylar sheet

Seamount Deformablebackstop Oceanic sediments (cohesiveand compactedrock powder) (Sand layers)

Back thrusts Cohesivemargin

Diverging strike-slipfaults followingthe slip lines

Seamount Accretionary location __ wedge

scar

Frontal margin reentrant

I sedimentsOceanic Figure 2. Principles of the experimentalsetup used to performthe sandboxexperiments. Specific apparatus, based on the sameprinciples, allow study of the deformationin two dimensions (cross section) or in three dimensions. The main morphotectonic features associated with the subduction of a conical seamount are summarized [see Dominguez et al., 1998].

the progressivedefornl.,ation of the margin during the subducting seamount,a half cone is used. Lateral efl:ectsare subduction of a conical topographic high. The second not discussed in this 2-D experimentand will be studied in presents horizontal cross sections performed in the the next two experiments,using the 3-D device. accretionarywedge above the subductedhigh. The internal Duringthe first 20 cm of convergence,an accretionary structureand deformationsof the marginare then digitized wedgedevelops by imbricationof fi'ontalthrust slices against the cohesivewedge (Figure 3a). Someminor back thrusts, and analyzed to determine the field around the relatedto the growthof the accretionarywedge, slightly subductedseamount. The third and final experimentconcerns deformthe frontof the cohesivewedge. the subductionof threeconical seamounts trending obliquely comparedto the convergencevector. Vertical cross sections, 2.1. First Stage:Uplift of the Frontal Margin parallel to the direction of the convergence,are Used to As the seamountstarts to subduct,the frontalmargin is estimatematerial transtarinside the deformedmargin at three uplifted. The basal d6collementis deflected into the thin sand different stages. cover (analog to the volcanoclasticcover) of the seamount flank(Figure 3b). Part of this coveris locallyoff-scraped and 2. Geometry, Kinematics, and Mechanics of accretedbeneath the frontalthrust unit (Figure4a). Becauseof Deformation Induced by Seamount the shapeof the subductinghigh, the d6collementslope Subduction abruptlychanges to bypassthe seamount.As a consequence, the frontaltaper of the marginincreases as well as the friction The deformational stages of seamount subduction are alongthe basald6collement, which now crossesthrough the observedalong a vertical crosssection on the 2-D device. To seamountsand layer. The frontal part of the accretionary study the section located directly above the top of the wedge deformsto reachthe steepercritical slope that is DOMINGUEZ ET AL.' SANDBOX EXPERIMENTS OF SEAMOUNT SUBDUCTION 185

5cm Trailing flank Leading flank .... Cohesive margin / •ccreuonary weage

...... •...... :::::::::::•:•.•...... •::•':•:" .:,:•,•-'...... •. ;•'•.. ,•-•L•"_•.,•' ..;..... •.•:'...... ,: ...... ::•-'?•i•;::::•::' ß. •...... •...... _--•:•----:...... ½--•.....:..• ...... ::::•:::•:•---•ii!•i ..... Seamount ' ' ....' ' rigidcore 10 cm

5 cm Upliftof the frontalmargin • ...... I•. •ealmenIaaon

...... •:•::.•:•::•'"•-"-'-'-'•'•:..:.• ...... •j;• -- ,• ...... -•--•:•:-•- •- •. • ...... •r.:•...... •:• -.. •..:,::,:+•:•.:•.•a½-..•::,•---<•:•-. -

. .. ..:...... ?:..:.. ...::...;.;:..'. :•:::• ---•--•------•••---••••••..•

5 cm seamountAccretion coverofthe top t backthrustIncipient ...... '•• %, • ...... ••••...• :...... :...... :..:•"". -.. •...... •:g,•.,.....,• • . - ...... ". ....:...... •.... .,,,::Z ...... ,...... ,...... •...... ,...... ,:::...... : ...... :.•. ;;.;.--:,::::::,**:..•. :•.;;;.

lO cm

Upliftof the cohesive• marginfront T Backthrusts Shadowzone •,,,'•'•"•,"f,'•/•'"'••.--.....•.**,,•.,••1•'• • ;-;...... ; ...... : ....:...... :::::•"...:.:.•,,.:;:..•'•,; ....

Incipientsubsidence Frontalaccretion resumes

lO cm Old backthrusts reactivated

5 cm Slope Newaccretionary wedge • break ...... ß..•.•..•:,...•.-•.::.:...... i"""•;• ...... ;;::..:,•. .•. . - •,½•,, ...... •..,. ½•:•...:¾'•: .•::.•....-•?•:? :•. ,.-'<%::4;;% '•••• ...... •: ß ...... - ß ß ß;•.•;.;;•:.:½: •.'..;•.5,1• ..:.;; ...." ..':.' ...... •••,•, ••• •:,•. ... .-;•.?•...... ß ..... • ...... '"••%•;;::.:':." "•?"•:5•:;;'•"•'•'•:•:Z;•.,.;:...,,,,.•,...,•...... •.•:.;;'•'*% ';•.- ':::' ...... '"" ":'<'-•.":"" ...... d.•.: ";::..":. ':'::•":'"'"" •4:::.•:.•:::::•:•½•::::::::,• '• ;.•-&•'".::•:•:::•:':? ...... :.":•':'.".'::.":::'":.:.:•.."•:•:,:.,::, ......

.. • ...... _...... •.••••••••••. Remnant ß '.' ß:".:":':•:;•:.:':'•:h:. :. shadow zone

Figure 3. (a-f) Photographsand structural interpretation of the firstexperiment showing the differentstages of conicalseamount subduction in crosssection. The experimentalset-up consists of a glass-sidedbox so that the deformationcan be observedin realtime. The trenchfill(1.2 cm thick)consists of six layersof coloredsand, and the seamountis composedof a rigid corecovered by a cohesivelayer of rock powderplus a thin sand cover. The cohesive sand wedge dips 5ø , and its initial length is 65 cm. Black arrows show the relative verticalmovements affecting the accretionarywedge during seamountsubduction. The greatestpart of the seamountcover is underplatedjust beneaththe front of the margin. 186 DOMINGUEZ ET AL.: SANDBOX EXPERIMENTS OF SEAMOUNT SUBDUCTION

Reactivationthrustofunits the frontal t Uplift compressionalBackthrustrelatedstress toinducedthe • • • ,• bythe subducting seamount

5c ....•• •'• '""'•

Normalfaulting cm Lowdipping normalfaults Slopebreak

cover (suturezone)

Figure4. Enlargementsof a conicaltopographic high showing details of margindeformation: (a) initial stage and (b) final stage.

associated with the new geometry and basal friction of this cohesivelayer at which friction is less.The cohesivelayer of portion of the d6collement.At this stage, reactivation of the the seamount'sleading slope is then accretedbeneath the thrusts is observed, inducing a general shearing of the frontalmargin, whereas its trailing slope remainsrelatively accretionary wedge, coupled with a significant undeformed.The seamountsubduction strongly disturbs the (6% betweenFigure 3a and 3b). These mechanismsin the very stressfield, and the basal d6collementis deflectedupward in early stages ofsubduction accommodatethe shortening and its wake. The d6collementcrops out in the middle part of the thickening of the accretionary wedge. A ponded basin frontal slope margin,which developsan overcriticalslope develops on the backside of the thickened wedge, which is where sedimentarymass sliding is active. A large shadow filled with sediments(sand). zone forms in the wake of the seamount. The trenchfill sedimentsand part of the frontal marginare draggedinto the 2.2. Second Stage: Indentation of the Accretionary Wedge subductionzone along with the seamount(Figure 3d). After further convergence,the d•collement .jumps back After 10 cm of seamountsubduction, half of the topographic down to the top of the subductingsedimentary sequence, in high is now buried beneath the accretionary wedge (Figure the wake of the seamount.Shearing of the material located in 3c). The top of the subducting seamountcover is off-scraped and accreted to the frontal margin. Because of the the shadowzone suggeststhat a smallpart of the convergence is still accommodatedalong the previous position of the compressionalstress induced by the subducting seamount, d&ollement. Only the trenchfill sediments continue compactionof the accretionary wedge increases(5% between subducting in the wake of the seamountwithout significant Figure 3b and 3c). Finally, when thrust reactivation and deformation. compactionhave reachedtheir limits, back thrusts initiate at the base of the subducted seamount slope and accommodate ßthe uplift of the accretionary wedge and part of the 2.4. Fourth stage: Subsidenceof the Frontal Margin convergence(Figure 4a). The basal d6collement,located on After 50 cm of convergencethe seamountsubducts beneath the leading slope of the subducted seamount, still the cohesivemargin. This cohesivepart of the model simulates accommodatesthe greatest part of the convergence.At this older accreted sediments and part of the detbrmable stagethe frontal margin recordsa rapid uplift correlatedwith termination of the upper plate basement (Figure 3e). the height of the subductingseamount. Reactivation of remnant back thrusts, as well as initiation of new ones generatedby the seamount,is observed. 2.3. Third Stage: Initiation of a Shadow Zone and Becauseof the increasing load on the seamount'strailing Indentation of the Cohesive Margin slope and the increase of the friction surface between the After 40 cm of convergencethe seamount is completely shadow zone and the rest of the margin, a new basal subductedand startsuplifting the rear part of the accretionary d6collement initiates at the base of the subducting trench fill wedge and the seaward end of the cohesive wedge (Figure sequence.Frontal resumes,and a new accretionary 3d). As the seamount subducts beneath the margin, the load wedgedevelops at the front of the margin. on its leading slope increasesrapidly. The basal d6collement During a certainamount of convergenceboth d6collements .jumpsdown to the interfacebetween the rigid core and the are active, and an intense shearing of the margin material from DOMINGUEZ ET AL.' SANDBOX EXPERIMENTS OF SEAMOUNT SUBDUCTION 187

Point 1 Deformationfront/ Po!nt2 • / Accretionary wedge /

Rigid butress (Backstop) ...... _,i•,•\ •,. •" - • Deformablebackstop

cm cm lOO/ i . I i .•• i i i 5 95 -'1 Uplift...... of the frontal part of:•,' •. /•,. Subsidenceofthe seaward end of 90•/ tne precolllSlOn'Pint1' •# ', %t •# • '•. [ne..... conerove weege 85/ •o., / ; y, • (.o•.t•) 80/ wedgefront

60/

50

40 / © © © © Cohesive margin 35 front (B) •--- Vertical motion 30 I I I cm 0 10 ' 20 30 '40 • 60 70 80 90 100 i ! ! i ! ' Frontal accretion resumes Shortening i ! ' in the wake of the The seamount starts i subducting ,, subductingseamount The seamount reaches the cohesivemargin front Figure5. Diagramshowing vertical and horizontal motion of thedeformation front and the seaward end of the cohesivemargin with respect to therigid backstop. The different stages of retreatof the marginand the effects on the wedgegrowth induced by the seamountsubduction are quantifiedby the two curves.

the shadow zone is observed. Finally, the upper d•collement A remnant shadow zone, composedmainly of part of the dies, and the basal d•collement accommodates the whole seamountcover and material scraped from the base of the convergence.As a consequence,the greatestpart of the margin cohesivemargin, is still observedin the wake of the seamount and trench fill sediments,dragged into the subduction zone (Figure 4b). behind the seamount, are underplated beneath the Other sandbox experiments,not presentedin this study, accretionarywedge. The size of the shadow zone reducesto show that the size of this remnant shadow zone depends on approximatelyhalf of the seamountvolume. The underplating the frictionalong the basald•collement level (herealong the of large volumes of sedimentgenerates a rapid subsidencein mylar-sand interface). High basal friction favors the the wake of the seamount,located beneath the rear part of the developmentof a muchlarger shadow zone, because there is accretionarywedge (Figure 3e). The old back thrusts, now no lower friction level at the base of the subducting located above the seamounttrailing slope are reactivated with sedimentarysequence. The decollementremains in an upper a normal component. Conjugate normal faults dipping position muchlonger. The greatestpart of the accretionary landward, and seaward dipping listric normal faults affecting sequence is then definitively subducted or underplated the accretionarywedge are also observed (Figure 3e). At this beneaththe cohesivepart of the margin. stagethe frontal slopeexhibits a typical slope break (suture), which separatesparts of the accretionarywedge developed 2.6. Kinematics before and after the seamount subduction. Duringthe experimentthe seawardend of the accretionary wedge (the deformationfront)and the termination of the 2.5. Fifth Stage cohesive wedge are recorded with respect to the rigid After 60 cm of convergencethe topographic high subducts backstop,as well as their vertical motions (Figure 5). The beneaththe cohesive part of the overriding plate (Figure 30. seawardgrowth of the accretionarywedge (Figure 5, stage 1) 188 DOMINGUEZ ET AL.: SANDBOX EXPERIMENTS OF SEAMOUNT SUBDUCTION occurs at a rate of about 0.75 cm/cc (where cc reft:rs to increasethe sand cohesion (Figure 6c). This method permits centimetersof convergence).During this stage the cohesive cutting horizontal cross sections, showing the internal marginfront recordsonly a very little landward displacement structure of the accretionary wedge around the subducted (a total of 1 cm) related to the initiation of minor back thrusts. seamount(Figure 7). When the seamount starts subducting, frontal accretion ends, and the deformationfront stops propagating seaward 3.1. Description of the Cross Sections (Figure 5, stage 2). The back thrusts generated by the The first horizontal section cuts the circular bulge located seamountinduce a modestretreat of the accretionarywedge at above the seamount (Figure 8a). Minor out-of-sequence a rateof about 0.1 cm/ccand a rapid uplift close to 0.2 cm/cc thrustsare observed in the cohesive part of the margin.They (Figure 4a). The subductionof the seamountstrongly deforms initiate during the first stages of the experiment and are the cohesivewedge. The shorteningof this part of the margin associatedwith the growth of the accretionarywedge. Curved acceleratesto 0.3. cm/cc, and the wedge front records a retreat back thrusts generated by the seamountcan also be observed at the same rate (Figure 5, stage3). above and landward of the subducted seamountposition. At After 35 cm of seamount subduction (50 cm of this stage they only affectthe accretionary wedge, which is convergence),frontal accretion resumes,and the seaward end uplifted and thickened. of the cohesive margin starts to record a rapid subsidence(- The second section is located just above the top of the 0.05 cm/cc). The effects of the seamounton the evolutionof the subducting seamount (Figure 8b). The accretionary thrust accretionary wedge structure progressively decrease(Figure units are clearly deformed. They curve landward in the 5, stage 4). The deformation front again propagates seaward vicinity of the seamount.Close to the subducting seamount, but at a lower rate, because the shortening of the cohesive out-of-sequence thrusts affect the middle part of the margin still continues (Figure 5, stage 1). After 80 cm of accretionarywedge. They develop from above the top of the seamountsubduction the total retreat of the cohesive margin seamountand propagateseaward on each side. Laterally, they front is close to 20 cm. At this stage the landward motion of connectwith the preexisting thrust units of the accretionary the cohesive margin tends to decreasesignificantly, but the wedge. A large part of the frontal margin is thus underthrust seamount still slightly disturbs the development of the beneaththe rear part of the accretionarywedge. In the wake of accretionarywedge in its wake. the seamounta stronglydeformed seaward dipping thrust unit is observed. It correspondsto the frontal part of the margin, 3. The 3-D Structure of the Margin in the indented during the first stage of the seamount subduction Vicinity of a Subducted Seamount and now subsidingin the wake of the seamount. The third section cuts the top of the subducting conical This experimentis performedwith experimentalparameters high (Figure 8c). The thrust units of the accretionarywedge comparableto the first 2-D experiment(Figures 6a and 6b). located above the seamount's leading flank are strongly When the seamount is subducted beneath the accretionary deformed and stretched.Some conjugatestrike-slip faults with wedge, the model is carefully impregnatedwith water to small offsetscan be observeddespite the intense crushing of

":::::7;::;::..:•:•':..•:%•:...½d'•.'.•a•.•½':'..',.. ß •wr:':. ' ..-. ß...... :;:::-'x'---.:-'ß,,;':•....::•...:o::•-,4::.,- .;:'::':'i-::5.: .:..;.77:-':•.•:.:':-...:.'..•.7':::7-•"- ' :"' .."". •:..======:..:::::::::::::::::::::::::!:.,.:::.-:.::::::::.:..::.:::-'::.::.:::::::::::::::::::::::::::::::::::::::::::::::::::::::::.•.. :'...... "' '.....,.•• :...-.: •,•.,..:•....:..-.,:...... :...:.-.-: .:,.. ,:•....::...... •.,...:::.,...... ;:::.5,:•:.:•?..,'•:$,•,•: .•e,--.....•. w, •:.?..:..• .•; .:,..:::,-::::-:, .:::.:::,...... ::.::.: :•:::.: .::....: :,..::::•:' ...... :::"•2,.....•x•/:;:'..::..•..-: .:•-:...... •...-''.,

... -.:.. :-..... ':•:::.,:•.':•. ß .....ßG:. '::•:• ...... •.• .'•:•}•,½ ...... 'm•

ß'•'.' ...... •'•:--•:'"'":"':.....-..'.:-...... '*' -s?.'.. "'•..%•.i...... :- -.....•:.-,..,,•,...•.,.-....,•,•--.-ß ...-.,...':•.t.M':':-'..' .--•.-..,,•..•.%...•--...... :'•,:.:.::•-:%;.,'..•'ß...... x.,. -'..

ß ' ::;;:::4'•[:5'.'-.-:"--•'. •'.•:' "> ..,.,...... ,.-.,-',..,-'-', '•i':.?[[..:.i':..i::•:<,-,•.d,.:. :.:•..5.'.,:..'"• ' :: d.•,•,•:•:,-:Z.:•..,...•.72•"•

, ,•,.• ...... >...,4,...'"-' • ...... :...... 7• ...... - ...... :.::,<,".-..• •<•'<,,•.. '.,..;.-:->:.;--":::::"q::'::::":'"...... :.:'":5-'.'"i•.>.... .<.z::,:-:::-.<:'.•'•-:i.,' ...... ' ...... ::'======':"::::::•S'::.'". *:.•' "'..'::...... " ..':•. ":" .,....?: :*"''"'-½:.:.::.:-½ :' '>-'' '•..."':'-•::::: '-:::'•:-•: :.-'::::::-.: :<•••••?'-'*•"'"•,•--,-'"'•"'•:}'•" '•"- '....-..- -. ,. :?::?: :'•:,• .'",'4•...-'•:'?:: :':'..:' ...... :::-.: ...... :..'....'-:. ::::::::::::::::::::::::::::::::::....======,/. .:.:.:a?':•:....•. • .,.,,• •.,,..::::•_.. "*'•,.*:,..,••:c:.•''<'"-- • <, •,•,,;:. • ..>.

ß ....,::.,:...*,...::,:•:;:½..:.::,...,:.:::,s, :•:.f½:?•$:.<;::,,.-::..::..;{•:•.-,.,,:...':'. -*•:y;.%•. ,:• ::'%*..--,- .,..,

Figure 6. (a,b) Planar photographsof the secondexperiment. (c) A perspectiveview of the second experiment. When the topographichigh is completely subducted beneath the accretionarywedge, the model is carefully impregnatedwith water to increasethe sand cohesion. Sections are then cut to study the deformationof the wedge structure. DOMINGUEZ ET AL.: SANDBOX EXPERIMENTS OF SEAMOUNT SUBDUCTION 189

Backgroundßmargin uplift REFERENCECROSS SECTION Cohesivemargin associatedwiththe subducted / j relief ?"''"'•';•:'•":'• ...... ; '"':::•ii•"'.'" ...... '•'Z:..,•*•:t --- I Accrebonarywedge' ,,•:,,½.•'• ß ,-'•:'..'½;,• ...... :--:-...... ----•r•.,.:' ;•5-z:::•.:.•: ...... •:•.... :.:.;•:.:•.•::. . ,,•.• :.-.;.:•:;:•;•:.:;.zz::,,...... Oceanic- sedimentsß ß•-.".. "*. ß%:- "'?..'...... ,•,";•;' :'•""•" ß- ' " "•..•½•::•7':'"""'-......

. ., - - .--, .:,:: '•½., •,,,-,•.•,., ß •½ ß,-•..•.,•::::;•:•½•.•, .,•.,•...,ß '4:.•,i...... , ß'.' ,4:..•.....• .'.,;.-.,•.:•;-".•..:'.:;•...... ,.=..,,.- .... - ' ,:::".' ...... :•...... ,•...... '• ...... , ...... :½:,,. •...... ,...... '. ;;:•,: •4 ½•:½b4.4:.•,,•½"•' •'"..... ½ ...'•;•'" • r.. .,½ * :.'*&•... '" '.... ,•'.'*'ß ..*•=;, -•::.•..-,.•.'""'•'•' ' •'-' ß"..' '":'.,•..,'",' ''"...... •'""¾'>. .•.... ;4a...... ½ ß,'..::','"'..,..•;",;-•;'½%;-;,;..%:L:-½•v" ß..':,.;&•' • ½•-'½•:%-;'.•"'.' ".. "::"-• ":• ..... ". ":'•,. t'. ,A.....*.,•r'"%.• '•' ß '•'•...... ,:•.•.•-½,,.:"•'•:• ' ...... ' ß : . ': .',.--..,.: :-• "...... ,j...... :...... •...... ,...... g.::..?.:.:...•::;•:•:•;•:7•::;•:•::•;•:•;•:•:•::•:•;•;•4•;;•:;;:•;•;•;•:•:•::•:•.•.•;:•:::`•:•;;•;:•:::::•:•:•:•:...:(•;•:::.•::;.:..•:.;:.:.:•:::•;.:•:•;•:•:•:.(•:.•;:.;•:;•:•:•::•:•;•;•;•....;•.•::..•.:•..•:•:•.•..:•.:.•:.:.::::::::::::::::::::::::::::::

STRUCTURAL INTERPRETATION Out-of-sequence thrust Pop-up structure \ Highly deformed zone • Thrustunits Deformationfront _---- , , \

....-:::..-:.::::...... •-:• .---..-.--..... -- Basald6collement level

Figure 7. Photographand structuralinterpretation of the referencesection, trending parallel to the direction of convergence.These depictions show the structure of the accretionarywedge and the seaward end of the cohesivewedge in a region situatedoutside the deformationgenerated by the seamount.Dashed lines show the location of the horizontal cross sectionspresented.

the accretionary wedge in this area. Seaward of the seamount 3.2. Deduced From the Network top, two faults trending parallel to the direction of convergencebound the upperpart of the shadow zones which We roughly estimatedthe shape of the global stress field is located in the wake of the seamount. around the subducted topographic high (Figure I0). This The last cross section cuts the margin at the top of the reconstruction is based on the interpretation of the previous subducting sedimentary sequence (Figure 8d). This section horizontal cross sections, the study of the fault network, and provides good constraints on the margin structure in the morphologic studies(Figure 2 and Dominguez et al. [1998]). wake of the seamountand on the shapeof the shadowzone. Above the leading slope of the seamountthe bending of the Seaward of the subductinghigh, part of the seamountcover accretionarythrust units associatedwith curved back thrusts and the trenchfill sequencelocated in the shadow zone can be and conjugated strike-slip faults (better observed in the observed and appearsto be undeformed.Two transfer zones, morphology) indicatesthat the major componentof the stress which connect to the base of the leading seamount flank, field (CSl)diverges landward. The indentationof the !nargin bound the shadow zone. These faults have a strike-slip occursabove the leading slope of the seamountand does not component close to the trailing seamount flank. Farther extend far away on both sides. Laterally, the thrust units seaward,they also exhibit a superimposedthrust component, appear to abruptly strike normal to the direction of which suggests that the shadow zone is progressively convergence. reduced in width. A reconstruction of the d6collement surface Conjugate zones associated with strike-slip faults geometry is shown in Figure 9. The main out-of-sequence trendingoblique to the directionof convergencesuggest that plane is also represented. CSl tends to convergein the wake of the seamount. The d6collement molds the shadow zone in the wake of the Furthermore, subsidence and extension observed in this area seamount.At this stage the volume of the shadow zone is confirmthat CS3tends to parallelthe relativeplate motion in comparableto the size of the seamountif we consider the the wake of the seamount. Finally, the successive shapes of trench fill sediments located in the margin reentrant. the deformation front during closing of the reentrant when 190 DOMINGUEZ ET AL.' SANDBOX EXPERIMENTS OF SEAMOUNT SUBDUCTION

•:... ' ..... •:,...... -•.,.- -*;?¾•4•:,:::::;"-::* ':•7:i'*'•"•;½•...... :aa•.,%.:;•.• ...... ;:-:•:::::'-;:;'i:;:•';:•E;f;:•;(?;-•(½-:,:•:•,:;•& 9'**½':":':':':""7C :t•... -.-•:•":• P"•:•a•*::i:•a•4i½:::::::,:..:a?•½,:---:--..•.-...... ,...... a:::::ii::';-•;.•.::.-;'.<- :'-.'•*:½•%•{a•:::½:E:•::f ::•::::...... '-': ....'.:-....'.::'.:::::.'.'....:::::;:.•.::.':'- ;..:...... • ' "•:::---•:½;i;•;:.:;•(;•.;:..;,.... ::::::::::::::::::::::: '...... "•::4,::...... :.&.:;:::.: .....•-:::•:::*'?';..-:?:L. - -'•" '...... '...... ::...... •m•}::•?.::,.. ::.;;...... : ,..;;-;;•*•:as½-•-:..•-...... ' ...... ;:;,•;;--::•:';.:?½;.}-:. --::'::•"'-:-:•:--•...; ...... '3;:;;.':'../:-' . %" . :::}•'":...... -.. ..

......

...... *.. .. :: ...... ' "::::"?½:"' ..... •;•-'"::'x•::,.,+;2L'.-'.' -- ...... ½

....

::"';7'; ...... ;...... ,.•;2:.'...... •,•:•½ ......

..:-/:?:!:•:.::•:•..::.

,?•i'-';:.•."...... :...... "":"":""'::•*• .-:..:-::,•.?.%,•...... '"'•'::'*':':":':'...... ":"-;;;" ...... '::'i,i •;•'*::•:::•:•'•'• .••.:•.%•:...... "••••••,...... • ...... -' ..-.:,,-.--;'";.L.: 'i...... :'•?:/":•'-?:';':':'"":":" ...... ':::.;:c...... ;:...:;...::::.::'"""::::•'•;i ...... -'.:.:.' ...... ß..&:y.'**:4;..• ...:,.. :'"'"%.:;;;;*• "•C:>:-::*½:"/"' ...... •:::...... :.{.....,'---?. ,.,.:...•:•......

.

::,..:..:.....:.' ...... •,. ':' ...... ½.:::.-'..!..;½. ½......

...... -:¾'-fa:::':':.?::::::.... '*** :•-'•::**' ?-:-w*

Figure 8. (a-d) Photographsof the horizontalcross sections discussed in the text.

frontal accretion resumes indicates that (•1 becomes margin at three different stagesof seamountsubduction. At the increasinglyparallel to the convergence. end of the experimentthe model is impregnatedwith water, as for the second experiment, and cut by 30 vertical cross 4. Deformation and Material Transfer sections,all parallel to the direction of convergence(Figure •). Within the Accretionary Wedge The subduction of three conical topographic highs, 4.1. Description of the Cross Sections trending obliquely with respect to the direction of A reference cross section is cut far from the effect of convergence,is performedto studymaterial transtarinside the seamount subduction and shows the structure of the normal DOMINGUEZ ET AL.' SANDBOX EXPERIMENTS OF SEAMOUNT SUBDUCTION 191

Subducting Topographic IShadow zone I seamount surface Deformation front /

Margin reentrant

D•collement surface Out-of-sequence thrust

Figure 9. Blockdiagram showing 3-D geometryof the mainthrust surt•tces(d{Scollement. shadow zone, and out-of-sequencethrust).

accretionarywedge (Figure 11. See referencecross section). Back thrustingof the rear part of the accretionary wedge onto the cohesive margin generates an "outer arc high" trending parallel to the backstop. It bounds a narrow morphological trough correspondingto the basin, which is filled with (sand) during the experiment. Section A shows the structure of the margin above seamount 1, which is in the first stages of subduction and uplifts the frontal margin (Figure 11, section A). The basal d{Scollementis deflected upward and emergesat the surfacein the middle part of the frontal margin. Above the seamount trailing slope, a small shadow zone develops, mainly composedof parts of the seamountcover and some frontal margin sediments. Section B shows the margin structure above seamount 2 ,,•Shadow buried at the boundary between the accretionary wedge and ' t zone the most compactedpart of the margin (Figure 11, section B). At this stage,frontal accretionhas already resumed,and a new , ; accretionary wedge develops. A volume of frontal margin . material, comparableto the size of the subducting high, is underplated beneath the rear part of the wedge. As a consequence,the accretionarywedge exhibits a steep slope (19ø) in this area, whereasthe frontal slope dips only 5ø. The Frontal margin slope break marks the suturezone. reentrant Section C is located through the seamount 3, which is deeply buried beneaththe cohesivewedge (Figure 11, section C). Here the effectsof seamountsubduction on the margin Figure 10. Shape of the stress field around the subducted structureare most important. seamount. The main thrusts and fault network, deduced from In the middle part of the accretionarywedge, near the suture the experiments presented and previous morphological area, the lower parts of the thrust units are missing. A very studies are superimposed. 192 DOMINGUEZET AL.: SANDBOXEXPERIMENTS OF SEAMOUNT$UBDUCTION

o c

o

o DOMINGUEZ ET AL.' SANDBOX EXPERIMENTS OF SEAMOUNT SUBDUCTION 193

0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 ca I • • I I I I I I I I ! I I I I I

'.';'iiiiiiii •...... :i' :•::

:.

U.S.

20

25

30

35 E3 )

40

45

5O

55

6O

cm

Thickenedarea R ßFrontal margin reentrant U.S. - Underplatedsediments Erodedarea •' Directionofconvergence E ß Eroded part of the margin Subductedseamount / ..•-• Difference the(cm)deformed betweenmarginthe undeformed surface and Figure 12. Map illustrating the transferand underplating of frontal margin material to the rear part of the margin. Dark shading with positive contours correspondsto the region thickened at the end of the experiment.Light shadingand white areaswith negativecontours correspond to regionsof the margin eroded during the seamoutsubductions.

long underplatedunit is observed instead, which to the size of the shadowzone when the seamountwas correspondsto the part of thetrench fill sequencepreviously subductingbeneath the accretionarywedge. draggedinto the subductionzone behind the seamount.In this region the lower part of the accretionarywedge is 4.2. Material Transfer in Three Dimensions strongly erodedand has beentransported into the seamount shadowzone and finally underplatedlandward. The 30 crosssections are processedto reconstructthe 3-D Part of the accretionary wedge sediments appears to structureof the margin and to calculate the material transfer subductwith the seamount beneath the cohesive' wedge in a (Figure12). Thetopography of the deformedmargin is reduced shadow zone. The intense de/brmation of the digitized using the vertical crosssections. The referencecross sedimentslocated in this remnantshadow zone suggests section is used to reconstruct, by interpolation, an internalshearing induced by a verticalvelocity gradient undeformedmargin topographyat the same stage of inside this kind of subductionchannel. At this stage the convergence.Figure 12 shows in map prqiection the amountof material dragged into subduction appearsto be difference(calculated in centimeters)between the deformed greatlyreduced (about half of the seamountvolume) compared and undeformedsurfaces of the margin.Assuming that the 194 DOMINGUEZ ET AL.: SANDBOX EXPERIMENTS OF SEAMOUNT SUBDUCTION marginmaterial cannot cross the rigid backstop(no output), now underplated at the boundary between the accretionary the topographicdifferences between these two surfacesare wedge and the cohesivemargin front. directlylinked to materialtransfer inside the margin. To avoid A thickened area can also be observed above the seamount artifactsinduced by the seamountshapes, the topographic leading slope. It correspondsto the active deformationof the effects of the seamount have been removed from the deformed seaward end of the cohesivemargin which recordsshortening. topographicsurface. Seamount 3 shows similar features. Nevertheless, because it The negativecontours (light shadingand white areasin represents a more advanced stage of deformation, some Figure12) show the regionof the marginwhere a material differencesare observed.The deformationfront is still slightly deficit exists(sediment has been removed),and the positive indented,but its morphologyis nearly normal, becausefrontal contours(dark shadingin Figure12) showthe regionof the accretion has resumed and a new accretionary wedge has marginwhere a materialsurplus exists (sedimenthas been developed.The eroded area of the margin is still observable added). and can be comparedin size and volume to the one associated Seamount1, which is the first stageof subduction,slightly with seamount 2.'The cohesive margin front is strongly deformsthe frontalmargin. A small thickenedregion (l cm) indented, and the._underplatedsediment proceeding from the located above the seamount'sleading edge is observed. It frontal margin lodges in this area. The volume of the margin correspondsto shorteningassociated with the reactivationof affectedby underplating of material and thickening of the the accretionarythrust units. cohesivemargin appearsto be very large (about 8 to 10 times Seamount2 is more deeply subductedand has strongly the seamount volume). This last observation leads to the disturbedthe accretionarywedge in its wake. The frontal conclusion that the seamountsubduction deformsthe margin marginappears to be erodedover a largearea, about 3 times not only along its track but also regionally. largerthan the seamount's basal surface. The amountof eroded sediment is estimated to be twice the seamount volume and is 5. Discussion and Conclusion alsocomparable to the volumeof excesssediment observed directlyseaward of the seamountposition. This thickened The use of analog modelingcan makea great contribution area,which extendslaterally over morethan 20 cm,reveals toward understandingthe evolution of the deformationand thatthe materialpreviously eroded from the frontalmargin is quantifying the resultant material transfer.The sandbox

Accretionary Forearc wedge Basement Coastaluplift 9ß basin Seamount Volcanic/ arc Trench / /

FLUID SEEPAGE UPLIFT "LIMITED'•ACCR'ETION OF PART OF THE SEAMOUNT ?

'LARGE.EARTHQUAKES

UNDERTHRUSTING OF FRONTAL MARGIN i i FRACTURATION OF MATERIAL 20 km THE MARGIN Vertical Exaggeration = 2 FLUID EXPULSION TECTONIC EROSION

UNDERPLATING OF LARGE POSSIBLE CHANGE IN VOLUME OF SEDIMENTS THE MECHANICAL COUPLING

MATERIAL TRANSFER

Figure 13. Schematiccross section illustrating material transfers inside the accretionarywedge induced by seamount subduction. DOMINGUEZ ET AL.: SANDBOX EXPERIMENTS OF SEAMOUNT SUBDUCTION 195

experimentsdescribed here show that the topographichighs volume of the subducting highs. Some materials from the carried by the subducting plate strongly interact with the accretionary wedge are also removed and deeply buried growth of the accretionarywedge and greatly influencethe beneaththe more resistantpart of the overriding plate margin. tectonicregime of the margin (Figure 13). Unfortunately, this estimation is only approximatesince Material transfer is expressed by different tectonic we do not consider the effects of fluids in our experiments, processesdepending on the part of the margin being affi:cted which necessarily limits an accurate quantification of the by the subductinghigh. The accretionarywedge and the inner slope angles and material volumes. It is important to note, compactedpart of the margin undergo significant deformation however, that our experimentsclearly show that basal friction includingfrontal marginerosion, removal and underplatingof and margin rheology can significantly modify the evolution of large volumes of sediment, indentation, and basal erosion of the shadow zone during seamount subduction. the cohesivepart of the margin (Figure 13). On the basis of the observationsreported in this paper, we Regions of active margins where large propose that the underplating of large volumes of relatively featuresare subducting, such as Costa Rica [Gardner et al., undeformed,water-laden sediments beneath the rear part of the 1992; Hinz et al., 1996; Fisher et al., 1998] and the Tonga- margin could result in local variation of pore fluid pressure in KermadecTrench [Ye et al., 1996; Ballance et al., 1989], the wake of the seamount. The dense network show an intense tectonic erosion. If the mechanisms generated by seamount subduction would favor fluid responsiblefor the deformationin our experimentsare similar expulsion and induce a decreasein the fluid pressure and to those occurring at active convergent margins, seamount effective basal friction. It is possible, then, that such a subductionappears to be responsiblefor an importantpart of mechanismcontributes to regional variation of the mechanical the tectonicerosion observed in natfiralsettings. We suggest and seismic coupling around subductingasperity. that in such regions, seamount subduction is the main Acknowledgments. The authorswish to thankC. Small and Z. Ben- tectonic processcontrolling the geodynamicevolution of the Avrahamfor their constructivecomments that improvedthe manuscript. whole margin and should be taken into consideration when The authorsare grateful to M. Jolivet, B. Sanche, C. Y. Lu, and A. calculatingsedimentary mass balance. Deschampsfor their help in performingthe sandboxexperiments and for technical support.M. A. Gutscher,G. W. Moore and P. Cormoily Sandbox experimentsreveal that the volume of sediments are greatly acknowledged for grammatical corrections of the transfered from the frontal margin to the rear part of the manuscript.Some figures presented in this paper have been done using accretionary wedge could be as much as 2 or 3 times the GMT software [Wesseland Smith, 1991].

References

Ballance, P., D.W. Scholl, T.L. Va!lier, A.J. New results from sandbox modeling of beneath the Peruvian continental margin at Stevenson, H. Ryan, and R.H. Herzer, seamount subduction and possible 12øS as simulatedwith sandbox •nodeling, Subduction of a Late seamount of applications,Eos Tran.v.A(iU. 75, 671, 1994. Geol. Rundsch., 83, 822-831, 1994. the Louisville Ridge at the Tonga Trench: A Dominguez, S., S.E. Lallemand, J. Malavieille, and Lallemand, S.E., J.Y. Collot, B. Pelletlet, C. model of normal and accelerated tectonic R. von Huene, Upper plate deformations Rangin, and J.P., Cadet, Impact of oceanic erosion, Tectontcx, 8, 953-962, 1989. associated with seamount subduction, asperities on the tectogenesis of •nodern Batiza, R, Abundances, distribution and sizes of Tectonophyxtcs, 293, 207-224, 1998. convergentmargins, Oceanol. Acta, 10, 17-30, volcanoesin the PacificOcean and i•nplications Dominguez S., S. Lallemand, and J. Malaviei!le, 1990. for the origin of non-hotspot volcanoes, Earth The oblique subduction of the Gagua Ridge Lallemand, S.E., J. Malavieille and S. Calassou, Planet. Sct. Left., 60, 195-206, 1982. beneath the Ryukyu accretionary wedge Effects of oceanic ridge subduction on Christensen,D.H., and T. Lay, Large earthquakesin system,Mar. Geophys.Res., in press, 1999. accretionarywedges: Experimentalmodeling the Tonga region associatedwith subduction of Fisher, M.A., J.Y. Collot, and E.L. Geist, Structure and marine observations, Tectontcs, 11, 1301- the Louisville Ridge, .I. Geophys. Rex., 93, of the collision zone between Bougainville 1313, 1992. 13,367-13,389, 1988. Gnyot and the accretionarywedge of the New Malavieille, J., Mod61isationexp6rimentale des Cloos, M., Thrust type subductionzone earthquake Hebrides , southwest Pacific, chevauchementsimbriqu•s: Application aux and seamount asperities:A model for seismic Tectonic.Y, 1 O, 887-903, 1991. cha•nesde montagnes,Bull. Soc. (idol. Fr., 7, rupture, Geology, 20, 601-604, 1992. Fisher,D.M., T.W Gardner,J.S Marshall, P.B. Sak, 129-138, 1984. Co!lot, J.Y., and M.A. Fisher, Formation of forearc and M. Protti, Effect of subducting sea-floor Malavieille, J., C. Larroque,S.E. Lallemandand J.F. basins by collision between seamounts and roughnesson fore-arckinematics, Pacific coast, Stephan, Experimental modelling of accretionarywedges: An example from the New Costa-Rica, Geology, 26, 467-470, 1998. accretionarywedges, Terra Nova Ab.¾tr.,3, Hebrides subduction zone, Geology, 17, 930- Gardner, T., et al., Quaternary uplift astride the 367, 1991. 933, 1989. aseismicCocos Ridge, Pacific coast of Costa McGeary,S., A. Nur, and Z. Ben-Avraham,Spatial Collot, J.Y., and M. Fisher, The collision zone Rica, Geol. Soc. Am. Bull., 104, 219-232, gapsin arc volcanism:The effect of collision or between the North d'EntrecasteauxRidge and 1992. subduction of oceanic plateaus, the New HebridesIslm•d Arc, .I. Geophyx.Res., Gutschef, M.A., N. Kukowski, J. Malavieille, and 7'ectonophysicx,119, 195-221, 1985. 96, 4457-4478, 1991. S. Lallemand, Material transferin accretionary Moore, G.F., and K.L. Sender, Fracture zone Dahlen, F.A., Noncohesive critical Coulomb wedges from analysis of a systematicseries of collisionalong the southPanama margin, Geol. wedges: An exact solution, .I. Geophy.¾.Res., analog experiments,.I. Struct. Geol., 20, 407- Soc. Am., ,h•ec. l'ap., 295, 201-212, 1995. 89, 10,125-10,133, 1984. 416, 1998. Nut, A., and Z. Ben-Avraham,Volcanic gaps and Dahlen, F.A., model of -and-thrust Hinz, K., R. von Huene, and C.R. Ranero, Tectonic the consumption of aseismicridges in South belts and accretionarywedges, Ann. Rev. Earth structure of the convergent Pacific margin America,Mere. Geol. Soc. Am., 154, 729-740, Planet. Sc•., 18, 55-99, 1990. offshore Costa-Rica from multichannel seismic 1981. Dahlen, F.A., and J. Suppe,Mechanics of fold-and- reflection data, Tectontcx,15, 54-66, 1996. Park, J. O., T. Tsuru, Y. Kaneda, Y. Kono, S. thrust belts and accretionarywedges: Cohesive Hoshino,K., H. Koide, K. Inmni,S. lwamura,and Kodaira, N. Takahashi and H. Kinoshita, A Coulomb theory, .I. Geophys. Res., 89, 10,087- S. Mitsui, Mechanicalproperties of Japanese subducting seamount beneath the Nankai 10,101, 1984. Tertiary sedimentary rocks under high accretionaryprism off Shikoku, southwestern Davis, D., J. Suppe and F.A. Dahlen, Mechanics of confiningpressures, Rep. 244, 200 pp., Geol. Japan,Geophyx. Res. Lett., 26, 931-934, 1999. fold-and-thrustbelts and accretionarywedges, Surv. of.lpn, Kawasaki,Japan, 1972. Scholtz, C.H., and C. Small, The effect of seamount .I. Geophys. Res., 88, 1153-1172, 1983. Kukowski,N.,R. vøn/Huene, J.Malavieille, andS. subductionon seismiccoupling, Geology, 25, Dominguez, S., S.E. Lallemand and J. Malavieille, Lallemand,Sediment accretion against a buttress 487-490, 1997 196 DOMINGUEZ ET AL.: SANDBOX EXPERIMENTS OF SEAMOUNT SUBDUCTION

Small, C. and D. Abbott, Subduction obstruction Pacific convergent margin of Costa Rica, in Middle Aanerican Trench off Costa-Rica frown andthe crack-up of thePacific plate 7, Geology, (leologic and ?kctonic Development o.f the wide-angleseismic data, Tectontcs,15, 1006- 26, 795-798, 1998. ('artbbean Plate Boundary in Southern 1021. Smith, D.K. and T.H. Jordan,Seamount statistics in ('entral Amerwa, edited by P. Mann, Spec. the Pacific Ocean,.1. (leophys. Res., 93, 2899- Pap. Geol. Soc. Am., 295, 291-308, 1995. S. Dominguez,S. Lallelnand,and J. Malavieille, 2918, 1988. Wessei, P., and S. Lyons, Distribution of large Tatcher, W., Earthquake recurrence and risk Pacific seamounts from Geosat/ERS- 1: Laboratoire de G6ophysique, Tectonique et assessmentin circmn-Pacificseismic gaps, hnplications for the history of mtraplate S6dimentologie,ISTEEM, UMR 5573, CNRS, Nature, 341, 432-434, 1989. volcanism, .I. Geophys. Res., 102, 22,459- Universit6 de Montpellier II, c.c. 060, 34095 von Huene, R., and D.W., Scholl, Observationsat 22,475, 1997. Montpellier, France. (stephdoln•dstu.univ- convergent margins concerning sediment Wessel, P., and W.H.F. Smith, Free softwarehelps montp2.fr) subduction, subduction erosion, and the map and display data, L:o.¾Tram'. A(iU, 72: growthof continentalcrust, Rev. Geophy.v., 29, 441, 1991. (Received December 9, 1998; 279-316, 1991. Ye, S., J. Bialas, E.R. Flueh, A. Stavenbagen, and revisedMay 17, 1999; yon Huene, R., et al., Morphotectonicsof the Huene, R., 1996. Crustal structure of the acceptedJuly 29, 1999.)