______O_ C_E_A_N_O_L_O_G_'C_A__ A_CT _A_,_'_9_6_' ,_N_O_S_P__ ~e------

' POO Convergenl margÎns A summary of results Japan Mariana Middle America

IPOD from the IPOD transects Marges convergentes Japon MarÎannes across the Japan, Mariana, Amérique Centrale and Middle-America convergent margins

R. von Huene · , S. Uyeda b • us GeOlogie;!] S urvey, MaÎISlo p 9'J, 345. Middleficld Road . Mellio Park. California 94025, USA. . b Earthquake Research Institute. Unive rsity of Tokyo, Bunkyo-ku , Tokyo 113. Japan.

ABSTRA CT Investigations of convergent ma rgins along the IPOD transects support the concept of occan floor s preading in bad·arc basins and the concept of tcclonically accreled sediment al the front of convergenl margins. However. not ail convergent margins have large accreted complexes, and other Jess frequcntl y used concepts arc required in the interpretations of the ~ e convergent margins . Ir the pre ~e nt rates of plate convergence are accepted, then much sediment that entered the trenches studied is presumably subd ucted rather than 3ccreted. In some instances, the continental framework i ~ truncated and somehow removed by tectonic erosion. Some convergent margins have s ubsided significantly during subduction. The crus t above the Be ni off zone appears to have been thinned by subcrustal erosion or the configura ti on of the Wadati-Benioff has changed. However. since only pari of the subsidence can be explained by erosion, the change in thermal structure res ulting from e hanging rates of s ubduction may be the major cause. To diHerentiate s ubduction-related processes eommonly requires time·stratigraphie information from drill samples to ma ke a kinematic reconstruction of. the margin. Ocemrol. A cta, 1981 . Proceedings 26,n International Geological Congress. Geology of continental margins symposium. Paris, July 7·17, 1980, 233·239,

RÉSUMÉ Résumé des résultats des forages IPOD à travers les marges convergentes du Japon, des Mariannes et de l'Amérique centrale. Le s recherches s ur les ma rges convergentes le long des transects IPOD confirment tes concepts d 'expansion océanique dans les bassins d'arrière-arc. et d'accréti on tectonique de sédiments a u front de ces marges. Cependant, toutes les marges convergentes n'ont p'IS de grands complexes accrétés et l'on doit utiliser d' autres concepts moins classiques pour interpréter ces dernières. Si les taux actuels de convergence de plaques sont acceptés, on doit alors considérer que la majeure partie des sédiments arrivant dans les fosses élUdiées est probablement subd uctée a u lieu de s'accréter. Dans quelques cas, le bâti continental est tronqué et q uelque pe u enlevé pa r l'érosion tectonique. Quelques marges convergentes ont connu une importante subsidence durant la s ubd uction. La croûte a u·dessus du plan de Benioff semble a voir été amincie par érosion so us·crustale. à moins que la confi guration de la zone de Wadati· Be nioff ait changé. Cependa nt . on ne pe ut expliquer qu'une partie de la s ubsidence par l'érosion : le cha ngement dans la structure thermique rés ultant des taux variables de s ubducti on peut en être la cause majeure. Po ur reconnaître les processus dus à ln subduction, il fa ut des données chronologiques et s tratigraphi ques obtenues à partir des forages, afi n de faire une reconstruction cinématique de la marge. Ocecmo/. Ac/a, 1981. Actes 26· Congrès International de Géologie, colloque Géologie des marges continentales, Paris, 7- 17 lu il. 1980, 233·239.

233 A. VON HUENE, S. UYEDA

INTRODUCTION The staff of Leg 59, working farther south, concluded from a sequence of basement ages that back-arc spreading in the The IPOO transecls of geophysical data and drill cores south part of the west Philippine Basin may have been across the Japan, Mariana and middle-America col'lvergent simullaneous with volcanism along the Palau-Kyushu margins el!:pand knowledge of active margins into areas not Ridge. Initial spreading and arc volcanism were followed by touched by geophysical data alone. The eombined modern subsequent episodes of spreading and volcanism along the geophysics and drill data provide a time-stratigraphie sec­ Mariana Ridge and Arc. The staffs of Legs 59 and 60 tion from which the rates of tectonÎC and sedimentary documented more thoroughly than before that new volcanic processes ean be esti mated . Samples from beneath the arcs developed sequentially. leaving behind a series of ocean floor indieate the depth of past depositional environ­ rem nant volcanie ridges and assocÎated back·arc basins. ments and Ihus, by comparing their past and present Although the back-arc basin crust appears 10 develop position, sorne sense of vertical tectonism can be obtained simultaneous with island-arc volcanism during long periods . in a geologic sening dominated by the horizontal motion of there is sorne indication that for short periods one process two convergent lithospheric plates. occurs without the other. Therefore it is slili not clear how mechanisms generating back-arc magma and the associated The information provided by IPOD drilling on active tensional forces are assocÎated with converging lithospheric margins supports many of the general concepts that the plates. program was designed 10 tesi. The accretion of sediment at the front of the trenen slope was persuasively supported by dala from the Oaxaca transect (Moore et al.. 1979), but accreted deposils were not ro und everywhere. T he develop­ MAG MATIC ROC KS IN THE FORE-A RC REGION ment of back-arc basins by crustal spread ing was convincin­ gly confirmed (Klein el al. , 1978 ; Kroenke, Scon, 1978 ; Along the Japan and Mariana tra nsects. arc-relaled volcanic Hussong et 01., 1978), and sorne insight was obtained as to and intrusive rocks were recovered at sites near the trench the complexity of this proeess. Allhough new discoveries slope, far seaward of the presenll y active arc (Fig. 1). Off were el!:pected, few anti ci pated their proportions. There is Japan, the recovered island-arc roc ks are of rhyolitic to evidence along ail transects Ihat a large amount of sediment dacitic composition and from intrusive and el!: lrusive envi­ is subducted rather than accreted. that continental li thos­ ronments (Fujioka, 1980). Refraction data (Nagumo el al., phere may be subducted, and thal lithofacies of the treneh 1980) indicate that another igneous body occurs nearby, and floor , treneh slope. anç! shelf cannot be predicted wit h only other plUlons are suggested by low-amplitude magnetic simple models of sediment distribution. T he information anomalies (Oshima et 01. , 1975). The crystallization age of from each leg has been summarized at this colloquium by these rocks determined by isotopic dating is 22-24 MY , the scientific staffs of each of the [POO active-margin legs. whieh may be about the same age as the initial volcanism ln this summary we can refer only to the preliminary results along the present arc (Yanagasawa el al.. 1980). Whe re of the legs because many of the Initial Reports of the Deep these rocks occur. the upper surface of the subduclina Sea Drilling Project on IPOD active margin legs have nOI yet ocean lithosphere is shallow, and if the present temperature been published. We have el!:tracted the main resul ts from gradient measured downhole is eXlended to the base of the the IPO D active margins program in this summary of the upper plate. the temperature is about one-quarter of that first of thTee colloquia on active margins at the needed 10 produce island-arc magma (Fig. 1). 26'h International Geological Congress. ln the Mariana transect. arc-related volcanic and plutonie rock was recovered by dredging from as far seaward as the lower slope o f the Irench (Dietrich el al. , 1978). The dredged and drilled Eocene and Oligocene rocks are sorne of the THE DEVELOPMENT OF BACK-ARC BASINS BY oldest of the Mariana Arc (Hussong et al. , 1978). The SEA-FLOOR SPREA DI NG altered state and their low melting temperatures indicate Karig (1971) proposed Ihat back-arc basÎns in the Philippine Sea (for el!:a mple. the Mariana Trough) formed by spread ing of ocean crust, and the results of Qlomar Cha/lengerdri1ling ...... 1 ...:~: :;"". . . ~ on Leg 31 are ge nerally consistent with the spreading hypothesis (KarÎg. 1975). There has always been a question '''- _...... , of how back-arc spreading was related to plate convergence " --- and subduction. ..: 1~~="~'~~~~~~~'~"~"~'~'~"~"~'~"~"i"~~;;~~~~~~~ ln the complex Oiato Ridge and Basin province of the Contln.nt.1 Cru" c.,~., northern Philippine Sea. back-arc spreading SÎmultancous " '''nll. with arc magmatism was confirmed by the studies of the -_... - scientific staff on Leg 58 (Klein et al. , 1980). Some fo rmer .. ~=:~~~~O~'~.. ._~______~~:::::..._.~::::.... ."::.~ magmatic ridges are now subsided remnant arcs ; the basins ! presently contain volcanic rock formed during arc and Figure 1 post-arc vo[canÎsm. The evolution of Ihis complex are

234 IPOO TRANSECTS ACROSS CONVERGENT MARGINS that the)' were formed in the presence of abundant water. Sediment recovered from the landward slopes of the tre nd Perhaps this water was derived from the dehydration of the on ail transects is mainly hemipelagic mud and interbedded lower plate during subduction (Anderson et al .. 1916) as thin volcanic ash. No pelagie sediment characteristic o f the suggested 10 us by J . Natland (wrillen comm.). deep ocean basin was recovered. This observation does not The positi on of ancient a rc-related rocks. much closer to the in ilself rule o ut the possibility that oHscraped oceanie trench than the presently active arc. suggests thatthe depth material was recovered because thick hemipelagie sediment to the top of the subducting oceanic lithosphere was once WHS recovered in holes as much as 20 km seaward of . and deeper or that the base of the continent may formerly ha ve 2.000 m above. the trench fl ocr of the Middle America and been farther seaward unless forearc magmas are produced Japan Trenches. If only the top of the deep oçean section by some other process than the melting of subducted and trench floor deposits are accreted (Moore. 1975) they materials. would be difficult to distinguish from sections deposited on the landward slope of the trench. NonetheJess. benthic foraminiferal faunas on the Japan and Guatemala landward slopes seem to indicate that o nly slope deposits were SUBDUCTION ACCRETION penetrated beçause they contain a mixture of transported forms and not the faunal assemblages Cound at trench Prior to drilling , multichannel seismic- reflection records depths. were interpreted as showing tectonic accretio n related to Deformation of shallow sedime nt on t.he landward slope by subduction, or subductio n accretion. off Japan and Middle miero fracturing was commonly observed beginning at a America. The clearest pre-drilling geophysical e vidence was depth of about 200 m and continuing to more than 600 m. from the middle·America Trench transect off Oaxaca (Shi. Microfracturing may be a dominant mode o f tectonic pley et al., 1980). Seismic dala from Ih is area show a ddormalion under the trench slope. There is e vidence for sequence of strong landward dipping renectors below a thin geopressured seclions from the logging in holes o n the lower layer oC reflections paralleling the seafloor (Fig. 2). Drilling slope of the trench oH Japan (von Hue ne et al .. 1978 : at three sites on Leg 66 indieates that the age at the 10p of Arthur et al .• 1980). We speculate that : 1) perhaps the the landward-dipping renectors progresses from Pleistocene grealest dislocations wilhin sorne underlhrust margin! are at the foot of the trench slope to upper Miocene near the concenlraled along narrow zones that are saturated with middle of the slope. This age progression is e vi dence for the overpressured water: 2) Ihat the absence of teleseismie stacking of thrust slices of material from the subducting earthquakes al the seaward end of the BenioH zone is ocean crust (Moore et al., 1979). However. despite the explained by water lubricated faults and abundant micro­ increase of deformation with depth observed in every core. fracturing ; and 3) that a smaU frictional resÎstance along no major faults were recognized from repeated ages or the lubricated faults could explain lerranes tha t are coherent on ph ysical character of cored sediment (Moore ef al .• 1979). a large sçale despite the thousands of kilomelers of oceanic Penetration (maxi mum 542 m) was less than expecled a nd crust Ihat have been subducled. perhaps deeper drilling woutd have provided evidence o f the thrust faults interpreted from the seismic records and dating of cores. SEDIMENT SUBDUCTION

Sofl sediment entering a subduction zone is commo nly presumed to be scraped off the more competent ocean crust and to be mechanicatty attached to the margin's existing framework. Some authors argue that sediment may be entrained in graben of the ocean plate that is subducted beneath the Jandward slope of the trençh (Lister. 1971 ; Hilde. Sharman. 1978 : Schwell er. Kulm, 1978). Accretion of oceanic sediment was stud ied along the active margin

Figure 2 Composi/t seclion along tht Middle Aml!'riCD TTl!'n ch ITllnSI!'Ç/ 011 Oaxaca. Muico. sho wing lond ... urd·dippinll refltclions and Ih l!' ITuncaltd conljntntol Iruml!'work. . AI Iht bollom 01 holt 492 Ihl!' Sl!'dimtnf Is aboul 12 MY old ... hrrtQs ollhl!' bOl/am al ho/' 488 it j, al Quo/trnory 08llr. From Moorl!' el al.. 1979.

Off Japan. the Marianas and Guatemala small-scale defor· v,. ,u. ' o. c. , .. . vo ~ or ".''''N' 'N"'" , ~ ' " ' matio n increases rapidl y with depth in slope sediment but no accreted subduction complex or major thrull t faulls were recognized in the core : however. driUÎng probably did not penetrate deepl)' into the section inferred from seismic data to be an accreted sequence. On the Japan Trench transect Figure 3 an interval of repeated ages at the bottom of hole 434 coutd DlalJTam sha ..· jng ho.., t h~ l' olum~ 01 subdaCfl!'d Sl!'diml!'l1/ " 'IIS signify thrust faulting . slumping. or possibly environmental deri~l!'d {rom drill and gtophysical data. Auumptialls hal·t bun mad~ al Ih~ limils Ihal nsalt III 0 minimum dt/icir al ndimtnt. changes recorded by the mierofossil fJora. Thus. an imbri· Drill/lIll obo~t Ih ~ occf~ud complu ho! gttlually IIQ / p~n ~ "altd cate thrust sequence might have been penetrated but this 1/1/('. or l' ~ ry lar /n/O. Ih t /l c c r~ /l an complu.... hich /s commanl)' interpretation is not unequi vocal. Inlured lrom la/ldwofd-dipplng rtf/tclioru in "bmic dala.

235 R. VON HUENE , S. UYEDA

transeclS by comparing the potential input. as estÎmated dominates Neogene vertical movement in the Mariana (rom sediment thickness on the ocean crust and forearc area. Off Oaxaca. the shelf first subsided 2: 10 3 km . convergence rates. with the corresponding sediment volume and then rose during the latest episode of subduction of the accreted body at the front o( the subduction zone. In accretion to the present depths of 1 10 1.5 km. Off Guate· ail cases the vol ume of accreted sediment was inferred from mala, subsidence since early Miocene time was indicated by interpretation of seismic records and its age was limited Ihe study of benthic fauna in the Neogene slope sediment. frorn drilling (Fig. 3). Despite the inabîlity of the Gloma/ and either subsidence or a large ri se in sea level are reported Chal/cogerto drill completely through the landward dipping from the Guatemalan shelf (Seely, 1979). Ir subsidence of sediment sections and recover the oldest sediment in a the margin involves some isostacy. thinning of low-density subduction complex. the oldest rocks penetrated at sites on continental crust and replacement wilh higher density the lower slope of the trench can be used as the minimum oceanic crust would seem to be required. An alternative age of the accreled subduction complex (Fig. 3). This mechanism, reporled by Langseth and hi s colleagues. procedure would give the largesl possible volume of sedi­ accomplishes subsidence by changes in the rate of subduc­ ment scraped off the ocean crust in a given lime, 10 compare tion with its consequent change in thermal conditions 10 wilh the smallest vol ume of sediment cOnlTibuled from the provide cooler rock of greater density. Such exchanges of oceanic plate: the ca1c ulalions produce results in an esti­ less dense for denser crust would probably occur al the mate of the maximum amounl of sediment that has been crustal contact formed by Ihe present subduction zone. accreted. Such an estimate on the Japan Trench transect The subsided paleolandmass off Japan had il Miocene indicates that d uring the past 6 MY only about 20 % of the Pacific shoreline that is now beneath the landward slope of sediment input has been accreted and the rest is presumably Ihe trench within 20 km of the present trench axis. A subducled. Off Guatemala there i5 vÎTluall y no sediment reconslruction of the margin with a shoreline within 20 km younger than Cretaceous at the foot of the trench slope of the trench axis requires that the axis was formerJy further requiring a period of no accretion or removal of the east and suggests that during the Neogene it relreated west previously accreted sedimenl. Along the Mariana Trench relative to Honshu by erosion of the trenc h slope. Off most of the rocks are igneous. and any sediment scraped Oaxaca. subduction accretion began about 10 MY ago from Ihe ocean crust is rare. Even along the Oaxaca margin. against a Iruncated margin o f PaJeozoic and older cratoni c where subduction accretion is best demonstrated, a maxi­ rock. Here the lTuncation may have been caused by mum of 66 % o f the incoming sediment is estimated to he slrlke-slip faullÎng. Off Guatemala. truncation by subduc­ accreted. The Imprecision of these estimales is insufficient tion erosion is perhaps the most attractive o f three expIa na­ to affect significantly the large amount of sediment sub· tions for the Cretaceous section recovered by drilling at the ducled. fool of Ihe slope. Truncated margins are common. but the Watkins. Moore and their coJleagues (this volume) suggest type of tectonic erosion that has caused truncation is not that of{ Oaxaca much of the sediment passing beneath the always dear. Considerable evidence exists for subduClion fronl of the margin ls accreted by being attached to the erosion. especiatly off Japan (von Huene Cl al .• 1978; underside of the accretionary mass. T he evidence for thls MU niuchi. Ludwi~. 1980: Lana~elh el (JI.. 19R1 . Thi~ process. which they term underplating. is that the wedging vol ume). o f successively accreted thrust slices to uplift and tilt strata Perhaps the conee pt of subduction erosion has been applied landward occurs only at the very toe of the trench slope. infrequently by geoscienlists becausc it requires processes Yet the paleontological evidence shows uplift of the whole that seem unlikel y. such as stuffing young rock beneath slope. Underplating of <; ubducted sediment Îs one expia na­ older rock of greater density or large-scale abrasion of Ihe tion of the uplift. continental framework. The concept of tectonic erosion by lateral faulting has been more frequenlly invoked toexplain truncated margins. and indeed il is much less difficult to SUBDUCTION EROSION imagine (sec. for instance::. Karig. 197~ ). Howevcr. if the plate convergence vector al some of the margins truncated Tectonic erosion. which Îs the relreat or subsidence of a in the past 10 MY has not changed appreciably. then the margin through removal of mlllerial from Il land ward slope lateral component of fault motion, and hence the lateml of Ihe trench . can be caused either by strike-slip faulting or transport of crustal fragments. should be smlitl . Near the by subduction erosion. Subductio n eroslon (Scholl el a/.. Japan and Guatemala lTansects the Neogene crustal 1980) is li term equÎvalent to consumption (Kulm et al.. fnlgments missing from the fool of the margin have not been 1977). to sorne uses of tectonic erosion (Karig el al.. 1978). located in adjacent seismic records for hundreds of kilome­ or ju~t ~i mply crosion (Hussong el al.. 1976). Subduction ters along strike. Therefore. subduction erosion deserves erosion can occur eit her al the leading edge of li margin. consideration as one cause of structural truncation at thus leaving a truncated margin. or along the subsurface underthrust marl!ins despite the present conceptual pro­ base of the lithosphere above the Benioff zone. thus causing blems. sub~tdence as the overlying crust is thinned. Subduction erosion has been suggested infrequenlly in the last few decades (e.g. Hussong el al.• 1976: Kulm et al., 1977: SEDIM ENT FACIES DISTRIBUTION Scholl el al .. 1981). However. arc magmatism in the present forearc atea. evidence of massive subsidence in the forearc Most o f the sediment recovered by DSDP driJling in and trench slope areas during subduction. and abrupt deep·water forearc areas and trench slopes is hemipelagic truncation of the conti nental framework on the landward sil! and clay rich in volcanic ash (including sites on Le!!~ 18. slope of the lTench ail sugaest that sorne trench slopes ha\'e 19. and 31). Clean sand in dÎscrele beds is a relatively minor periods of retreat. constituent in the core material recovered : however. il is Subsidence dominate~ the Neogene lectonic history o f the al 50 more diHicult to recover thlin silt and clay. On the forearc area off Japan. and net subsidence rather than uplift middle America transect off Oaxaca. caving sand stopped

236 IPOO TRANSECTS ACROSS CONVERGENT MARGINS drilling in the trench and on the Irench lower slope. On the The number of cores and other samples is obvio usly middle America lransecl orr Guatemala the trench fill insufficientto clearly show the smalliithofacies differenees (sampled in nine holes) i~ predominently silt and clay. Il is be.twee n depth environment along a modern conve rgent also in the Aleutian T rench ort Kodiak (Kulm el al., 1973). margin. Underwood et al. (1980) conclude Ihat modern Conventional ~ampling (piston coring and dredging) of the trench floor. slope basi n, and slope deposits can range from shallow strala along the Japan Trench margin. the Aleutian fan complexes through hemipelagic and pelagic deposits. A T rench margin. and the middle America Trench margin has similar conclusion results Crom ail DSDP studies except the produced mai niy silt and clay similar to that in the OSOP Oaxaca trll nsecl. Off the Japan and Middle America Tren· core. Off Oaxaca near a major canyon aerou the margin , a ches coarse· to fine'arained sediment is transported in signiricant amount of sand was recovered in piston cores. channels trom near shore across the margin and ultÎmately and along the other canyons coarse sediment was cored or tO Ihe trench (Arthur et al., 1980 : Moore et al., 1980). This dredged. T hus. ahhough sand is perhaps underrepresented bypassing of the shelf prod uees the similarity of lit hologies in recovered material because it impeded the drilling or across the margin. The similarity of lithofacies is consislent coring, hemipelagic silt and clay are the main materials with observations of morpho-tecto nic and sedimentary recovered from the trench and lower slope. a conclusion features. The morphologies and structures in seismic also confirmed by downhole logging. records across any part of the Japan margin reveal sl ump. channel, overbank. ponded, draped. a nd fan deposits, only in different relative proportions (von Huene. Arthur. in ". press). There are no definite links belween the sedimenlary • ,. . l structures seen in geophysical records and the correspon. l'i h . \: ding sedimentary facies associations: however. the trend of one should be refleCled in the other. We speculate that no single type of facies association 15 unique to a particular i e nvironment, but sorne types may be more frequent in one ...... 0- Ihan in another. Thus, for instance. slumpi ng or soft ·sedi­ j. "< '. f ment deformation might be more frequenl on trench slopes •f .n or in slope basins as compared to a basin on the shel!, and if .. 1 ,1 the sediment is sufficiently sandy the re is probably a .Y, ,1 • relative abundance of certain facies associations that could ... ! 1 aid in distinguishing one envi ronment Irom another. The .. i..i,.;.j ._. - . key 10 identifyina teclOnically displaced sedimentary envi· o .... - --.~. ,'H" -,,- •'- 'H .. .~"" ! ronments in ancient rock may be a relative comparison of the fac ies across the whole margin. In converge nI margin Fisure 4 environments. the sediment distribUied across the margin in l..IlIwlogic Jte/ion ptne/rO/td 01 Ji/u 4JS ond "J9/rom 0'0/1 Hut llt et over a sho rl period 01 lime should be sludied because al.. 1978. sediment pathways and basins change as the trench slope morphology is re·scul ptured by the nlpid tectonic move­ menl associaled with subduction. Hemipelagic sediment like that deposiled on continental arc·trench systems is also deposited beyond the trench in the deep ocean basin. The hemi pelagic sections seaward of the trenches are generally of laIe Miocene age. although CONCLUSIONS near the Aleutian Trench (site 178) and Kamchatka Trench (site 192). the deepest hemipelagic material is of early or Perhaps we have biased this summary by discussing the perhaps mi ddle Miocene age. When the sites of deposition unanticipated results at grealer length Ihan the resulls are reslO red to their position in Miocene time by applying consistent wi th previous concepts. Wi thout much do ubt . the global plate motion history. these sites were apparently accrelionary teetonic model for conve rgent margins has several hundreds of kilometers out 10 sea when they been prominent in the evolution of plate-tectonic concepts. received terrigenous sediment. One consequence of the Duri ng the past decade, this model was applied in a majorit y thick hemipelagic sediment lar out on the ocean plaIe is that of interpre lations for bath ancient lold belts ilnd modern considerabl y more hemipelagic sediment is available 10 be margins, seemingly to the exclusio n of other mechanisms accreted Ihan if Ihe whole oceanic section were Iypical that had been suggested in the pasto For instance . so pelagic ooze. This may be one explan3tion for the spMse convincingly has the accretionary tectonic model of conti· oceanic sediment in sorne a ncient deposits thal are interpre. nued Ihrust wedges seemed 10 explain pllrts of the Francis­ ted as teclOnically accreted complexes. From the cores and can complex of California or the modern soulheast Asian surface samples. it appears that in modern trenches o nly are-trench systems that other tectoni c concepts seemed ve ry sub tle li lhologic differences distinguish environments overshadowed. T hus. th rough the diseoveryof features thal of the shel f. slope, trench, and perhaps even the trench cannOI be explil ined by accre tion, the lPOD results may seaward slope. T he conu ast of muddy sediment sampled have renewed curiosity about other tectonic mechanisms. Irom modern trenches with sandy sediment from presumed Nearly 20 yeaTS ago, Coats (1962) struggled with the ancient analogues is puzzli ng. concept of sediment subduction in his analysis of Aleutian Biofacies are the best indicator of paleodepth : however, island arc milgmalism. His concept was furthered by Gill ul y those interpre talions of paleodepth referenced to oceano­ (1969 ; 197J) and extended by Sc holl and Marlow (1974 a. b) araphic depth boundaries 5uch as the carbonate compensa· and by Karig (1974) and Moore (1975) relative 10 subduction tion dept h are difficult to project back in li me . The depths of pelagic deposits. Perhaps what is needed 10 further of the carbonate compensation depth have fl ucluated in the knowledge of the sed imenl subduction concept are samples pasto especially along margins. from an active Beniorr zone to help conslrain some physical

237 A. VON HUENE , S. UYEOA

paramc:ters such as geopressuring, sediment consolidation. and the massive subsidence of continental crust above a n and stress 50 that the process can he modeled. A more active subduction zone are probably s ufficient evidence for convincing argument for sediment subduction requires a sorne form of subduclion erosion. phys ical explanation fo r the undenhrus ting of soft sedi· ment. Subduction erosion is also nOI a new idea ; it wa:. Acknowlc:dgeme:nts sugges tcd more than a decade ago by Van Bemmelen (1966) and has been Ireated infrequently and in a spec ulative way for some time in studies of a ncient margins. To make Much of what is contained in this paper wa!> disc ussed at persuasive arguments fo r subduction erosion is difficult length in meetings of the !POD Active Margins Panel. We becaus e the proces$ is recognized by the absence of rock arc ver y grateful our col1eagues on the original and o n the and is difficult to observe in the geologic record. Noncthe· present panels. and a lso to Michael Fis her a nd Terry Bruns less. the relreat of the leading edge o f a convergent margin for critically reading our mtlnuscript.

REFERENCES

AodC:t'$C)n R . N., U)'eda S .. Mlytihlro A., 1976. Geophysical and Karlt D. E., Ingll: J. C, Jr. el I l .. 1975. Site 299. in ; 'nirial R ~PQrlS geochemical constraint 5 at con"'erginll plate boundaries _ Part 1 : 01 /ht Dup Sta Drillinl: Projtc/. "III. J I . Apra. Guam to HukodHle Dehydration in the downgoing slab, Gtophys. J. Rrs. A S/roll , Soc .. Japan - J une lu August. 1973. Nat. Sci. Found, Washington U.C., 44, J3J·3H. 10. Arthur M. A., ,on llueoe H., Adetwck G. Jr, 1980. Sedimenlary Klfln G. dc:V ., I\:obayashl K.. 1978. l'hilippine Sea drillcd. Gto,i· evolution of the lapan rorearc region off Northern Honshu, Legs 56 mts. 23, 5. 23·25. and 57, Dc:ep Sea Drilling Project. ' nitial Rtp()r/S 01 the Detp Sto Drilling Projtel, '·01.51 , Wus hington. US Gov. Prin!. Off., l , Kltln G. deV., Koba)·tihl K .. Cham~ y H., Curlu D., DÏ(k H., 521·S6!I. fA,bolJi D. J., Founlaln 0 . 1\1 .. Kinoshlta H .. Manh N. G .. MllU' COIIU R. H., 1962. Magma I)'peand crustal struclure in the Aleutian nU A., /I01s1c:rc:nko G. V., Okada H. , SIOIID J . R., Wl plt$ D., Whl· Arc. in : The crust of the Pacifi e basin, cdilC~d oy G. A. Macdomlld le S. '\1., 1978. Philippine Sea drilled (Leg 51). Geotimu. 23, 23·25. and H . Kuno. Am. Gtophys. Union , Groph)'s. Mflllog r .. 6, n·1()9. I\:II:/n G. 0.. KnhaYlshl K., Chllmlty H.. Curtis D., Dkk H .. O~trkh V. , ErmaDD R .. OberhllMlI R., Purcbelt H. , 1978. Geo­ fA, hab D. J ., Founaln D. M., I\: Jnoshlta H .. Manh N. G., MlzLi' c hemistr)' of basaltic and gabbroie rocks from the we~t Mariana no A., NISIc:r c: nko G. V., Okada If., SILIli n J . H., WlY'ks D. W. , Basin and the Mariana Trench, E(1fI1r Planet. Sei. I.elt .. J9, White S. M., 1980. In i,iul Reports oltht Dttp Sta Drilling Pro/tel. 127·144 . "01 • .58, Washington. US Gov. Print. Off. Fujloh K., 1980. Conglomerales of volcanic rocks of the: Dc:ep Sea Kroc:nke t ., Seau R., 1918. Old qUbtions answered _ and new Drilling Projcci. site 439, lni/jol RtfWru 1,/lIrt Dup Sea Drilfing ones askcd (Leg59). GeorimtS. 2J, 7. 2G-2). ProjeCI. "01.51, Wushington. US Gov. Prin!. orr .. 2, I075· I(}Ij2 . Glll ul y J ., 1969. Oceanic sediment volumes and conti nental drirt , Kulm L. D.. '-on Huene H., Duncan J. R .. In,lt l. C" Klinll S. A., Scienet, 166, 992·993. MU:li kh L. f.. Pi ptr D. J. W., Pralt R. M., Sc:hrader H. J ., W~r O. , Wise S. W., 1973. Jllilial R~ponlollht Dt~p Sea Drilling Gilluly J ., 1973 . Stead)' plate motion and episodic oragen)' and Projett, "01. 18, Wa shington. US Gov. l'rint. Off.. 1017 p. mllgmatism, ~o l. Sec. Am. Bull .. 84 , 499·514. "ulm L. D., Sch.ntJfr W. J ., Muais A .. 1917. A prc:liminary ana· llilde T. W. C., Shllnnan C. F .. 1978. Fault ~truc tur e of Ihe du· Jy$is of the: subduction procc:ss along the Andian continental cending plllte and its influence of the subduclion process (abslr.), margi n. 6' to 45' S. in ; Island ures. deep,sM trenches. und buck·of(' Proe. Am. Geophys. Union Metl .. Dcc .. 1978. busiu, edited by M. Talwani and W. C. Pitman III . Am. Geophys. HUMO ns D. M .. U)edll S. , 1978. Leg 60 end~ in Guam, GeOiimes, Union. Washington D.C .. 285·302. 2J, 10. 19·22. 1.lIng~lb M., Okada H .. 1978. Ne.. r the Japan Trrnch trun,~ cts H~ng D. M .. Ed .. anb" l'. H., Johnson S. Il .. Cllmpbell J. R" bc:gun, Gel/limes. 23, J. 22·26. Sulta n G. Il ., 1916. Crustal structure of the Peru-C hile Trench ; 8·12·S latitude. in : The gc:oph)'sics of the Pacific Ocean Basin and Lanveth M. , ,'on H ~IW R., /IOtiU 1'01 .. Oklldll H., 1981. Subsidence ilS marlin, e(lite(l by G. H. SUIIOn, M, H. Manghnani. R. Mobc:rl)' of tlle Japan Trencll tortare rt,lion of nOrlhern Honshu. (keullal. and E. U. McAfee, Am. Gtophyl. Union. Gtt/phys. Monog,_. 19, ACIII. Thi s ",olume. 113·179. 11·85 . Lister C. R. H., 1971 . Tectonic mo"'ement in the Chile Trench, Hussone D. M., ll)'eda S., HJa llCh" R., H~II tJ., Ell is C. Il .. Fran· Scitnce. I7J, 719·722. cu T. J . G., .·ryc:r P. , Ho ...' K., Kllnl S .. Meljc:r A., Nlbmu· n 1\: ., Natllnd J ., Pllckham G. , Shllfukln A. , 1918. Leg 60ends in Mizuno A. , Okuda V., Tlmakl K. , Klnoshlla V. , /IOoharl M., Vua . Guam, GtOlime~ , 23 , 10. 19·22. I II M., NM hJlma M., Munakaml F., Tua§h!ml S., isblbashl K., 1915. Marine: lIeo[ogy and geoloaical history of the Dailo Rid ge bhlwllda Y., Oellwa K., 1976. Pc:troleum geoloe)' of the: off shore arc:a. northweSTcrn Philippine: Sell, Mo '. Sei.. 7. 4&4-491 and areas around the Japanese 151ands. U.N. ESCAP. CCOP Tteh. 54)·548, Bull .. 10, 23.34. Miluno A., Watklns J . 5.. Shlpley T. H. , Bachman S. H.. JOlDES Acth'C: Mllrglns l'aneJ, 1971:1. Acti",e rnar.:in~ acti "'ely B~l tl F. W., Huit A .. Dldyk H. M .. IAUtU J. K., Lllndberll N.. considered. G~Qlim~s. 23, 2. 21·29. ~kMiilf,D K. J ., /IOl1lsuma N., Shephlrd L. E.. Sirphan J . F., Klrle D. E .. 1971. Origin and de"'elopmcnt of marginal basins in St ...dner H., 1979. Prollfessi",e accretion in the Middle America the western Pacifie, J. Gtuphys. Rts., 87, 1259-1268. Trcnch. southern Mexico, NalU'~. 281 . 63J-642. Karl, O. E .• 1974. Tectonic eroslon atlrenches. Eanh l'faner. Sei. Moore J. C" 1975. Selective subduction. Gm/OR)' (Bouldu). J, Lm., 21 , 2()9.212. 53O-SJ2. KllrI, O. E .. 1915. Basin "enesis in the Philippine Sea. I"ilial Repnrts 01 the {)tep StU Drilling Pro/ter, '-al. J I , Washington. US Moore J . C., Watklll!l J . S., 1979. OH Mexico; Middle Arnerieu Gov. Prin!. Off.. 857·879. Trench. Gent/mts. 24, 9, 20.22. I\:arla D. E .. Cnd..-ell R. K .. Moo~ G. F., Moorr O. G., 1978. Moore J . C .. W"klns J .. Bat hman S. H. , B-tahttl Jo' . W., 8uII A .. Latc: Cenozoic )ubduction and conlinental mar"in trunclttion along DId)'k B. M .. UlUlell J ., Lundber, N .. McMilJrn J. J .. Nlluu. the "orthern Middle Ameriea Trench. Geul. S"c. Am. Bul/ .. 89. 2. ma N .. Shephlrd L. E .. Shlplty T. Il .. SIc:phlin J . •'., Stradntr Il .. 265·276. 1979. Middle America Trench, Geotimts. 24 , 9, 20-22.

238 IPOO TRANSECTS ACROSS CONVERGENT MARGINS

Moore J. C., Watklns J. 5., McMlllen K. J., Bllcnmlln S. B., Stely D. R., 1979. The evolution of structural highs bordering major Lqgeu J. K., Lundberc N., Shlpltoy T. H., Stephll n J. F., forearc basins, in: Geological and geophysical investigation of ~ htel F. W., Buu A., Dldyk B. A. , NlllS um li N., Shepard L. E., continental margins, edited by J. S. Watkins el al .. Am. Assac. 51rltdner H., 1980. Facies belts of the Middle America Trench a nd Pe/ral. Geai. Mem" 29 , Tulsa. Oklahoma. 245-260. forearc region. Southern Mexico results from Leg 66 DSDP. Geol. S«Jy D. R .. VIOU P. R. , Wallon G. G., 1975. Trench slope mode! , Soc. London. in press. in : Geo/tJgy 0/ CQn/inenlal m(/l·gi,.,s . edited by C. A. Burk and MUnluchl S .. LudwlC W. J ., 1980. Crustal structure of the l apan C. L. Drake. Springer-Verlag, New York. 249·260. Trench : the effee! of subducti on of ocean crust. InUial Reports 0/ SlIlpley T. H., McMUlen K. J., Watklns J . 5., Moo~ J. C., Sando­ the Detp Sta Drilling PrQject . ~ul. 57, Washington. US Go~. Print . "al-Ochoa J . H., Wonel J . L., 1980. Continental margin and lower Off .. 1,463-469. slope structures of the Middle America Trench near AeapuJeu NlIllwno 5., Kasahllnl J ., Koresawli 5., 1980. OBS airgun seismic (Meltieo). Mar. Gtal., 35, 113. 65-82. rtfraetion survey near Sites 441 and 434 (I-IA). 438 and 439 (1-12), Underwood M. B., Bachman 5 . 8., Sc hwt ller W. J .. 1980. Sedi­ and proposed site J-28. Legs 56 and .57. Deep Seu Drilling Project. mentary processes and facies associations wi thin trench and Initial Reports 0/ Ihe Detp Sta Dri/ling l'mjut. vol. 57, trench-slope senings, in : QUalemary deposi/ional en"ironmems on Washington. US Guv. Pri nt. OH .. l , 459-462. Ihe Pacifie continental margin. edilcd by M. E. Field. Oshlmll 5., Kondu T., Tsuk .. moto T., Ooodera K .• 197.5. Magnetic A. H . Bouma and 1. Colborn. Soc. Eeon. PaleontoJ. Miner .. Paci­ anomalies at sea around the northern part of lapan. Rep . Hydrogr. fie Section. 211·229. Res .. 10, Hydrographie Office of Japan. Tokyo. Van Ikmmelen R. W. , 1966. On mega· undulations _ a new model Scholl D. W., Mllrlow M. S .. 1974 a. Sedimentary sequences in for the earth's evolulion . TtcfOnophysics. 3. 83-127. modern Pacific trenches and the deformed circum-Pacific eugeo­ Von Huent R. , NHU N. , 1978. On Leg 57 : Japan Trench transec_ syncline. in: Modern and ancient geosynclin;Ll sediment.'1tion. ted. GtOlimes. 23, 4. 16-21- edited by R. H . Doit Jr. and R. H. Shaver. Soc. Econ. Pulmn/o/. MineT. Sptc. Publ .. 19, 193-211. Vo n Huene R., Arthur M. A., 1981. Sedimentation !lcro~s the Japan Trench off Northern HOrlshu Island . Geol. SOt'. LOlldon. Scholl D. W., Mllrlow M. S., 1974 b. Global tectonics and the Spec. Publ. (in press). sediments of modern and ancient trenches - some different inlerpretations. in: Plate teetonics. assessments and reasses­ Von H~n e R., Aubouln J . el 111., 1980. Leg67. The Deep Sea smenl;;. edited by C. F. Kahle, Am. A ssoe. Pe/rol. Geol. Mlm .. 23, Orilling Project Middle America Treneh transect off Guatemala. 255-272. Geol. Soc. Am. Bull .. 91. 421-432. Von Huene R., NHII 1'1., Arthur M. , Cadel J. P .. CIiNOn B., Moo­ Sc holl D. W. t von Huene R.t Vallier T. L., HO"'eli D. G .. 1981. Sedimentary masses and concepts about tectonie processes at ~ G. W. , Honu E., Fujloka K., Barron J . A., Keller G., Reynol­ underihrusi ocean margins, Gc%gy. 8, 564-568. ds R., Suffer B. L., SalO 5 .. Bell G .. 1978. Japan Trench transee­ tcd on Leg57, GeOlimes. 23. 4.16-21- Schwelle W. J ., Kulm L. D., 1978. Eltle nsional rupture of oceanie crust in the Chile Trench. Mar. Geol.. 28 , 271-291. Yallll§\s.IIwa M. , Tllklgaml Y .• Odma M., Kaneoka 1., 1980.

239 240