Subduction and Accretion in Trenches

DANIEL E. KARIG Department of Geological Sciences, Cornell University, Ithaca, New York 14850 GEORGE F. SHARMAN III Scripps Institution of Oceanography, P.O. Box 109, La Jolla, California 92037

ABSTRACT INTRODUCTION the associated trench, and plotted at similar horizontal and vertical scales. These Although the reality of subduction has One of the more controversial aspects of profiles, integrated with available seismic been greatly strengthened by recent inves- plate tectonics has been the assumption that reflection profiles and other geophysical tigations, there is little information dealing large-scale subduction of oceanic litho- data, provide insights to the variations in with the mechanisms by which material is sphère takes place along trenches. Early in- subduction-zone structures which were not subducted or accreted to the upper plate. terpretations of gravity data (Worzel and obvious from analysis of scattered separate An attempt to determine the gross evolu- Shurbet, 1955) and of seismic reflection data sets presented at different scales and tion of subduction zones has been made, as- profiles (Scholl and others, 1968; von vertical exaggerations. The basic assump- suming that geographic variations in mor- Huene and Shor, 1969) have been cited as tion utilized in our analysis is that morphologic and geophysical characteristics of evidence precluding underthrusting in phologic variations among the arc systems trenches can be transformed into temporal trenches (Beloussov, 1970; Carey, 1970; can be interpreted in terms of evolutionary trends. Deformation associated with sub- Meyerhoff and Meyerhoff, 1972). How- trends. duction extends across the lower trench ever, these interpretations have been shown slope, from the trench axis to the trench- to be erroneous by more recent results using TECTONIC FRAMEWORK OF slope break. This region is a rising tectonic these same techniques (Grow, 1973a; SUBDUCTION ZONES element, but the upper slope is a subsiding Holmes and others, 1972; Beck, 1972) and region of sediment accumulation. An upper by data from the Deep Sea Drilling Project The uniformity of geological and slope discontinuity separates this zone of (Kulm and others, 1973b; Ingle and others, geophysical features among the various subsidence from the rising frontal-arc 1973, 1975), as well as by the results of island-arc systems has often been noted block. Examination of very young trenches earthquake seismological studies (Isacks (Hess, 1948; Umbgrove, 1947) and has indicates that the upper-slope discontinuity and Molnar, 1971; Barazangi and others, since been amplified (Karig, 1971a; Karig marks the upper section of the continental 1972). and Mammerickx, 1972; Grow, 1973a, or insular slope that existed before a sub- It is now apparent that the oceanic 1973b; Dickinson, 1973b). A generalized duction pulse began. As material is fed to lithosphere is subjected to large-scale un- terminology (Karig, 1970, 1974b) based on the subduction zone, the distance between derthrusting at trench axis and that some of this uniformity is utilized, for the most part, the upper slope discontinuity and the trench the uppermost material is transferred from in this paper (Fig. 2). Modification is re- increases, and an accretionary prism de- the lower to the upper plate in the process. quired chiefly in the slope area between the velops, but its shape depends on the relative In some trenches, tectonic erosion of the crest of the frontal arc and the trench, rates of sediment feed from the arc and upper plate during subduction has been where the greatest variability within the arc from the offshore basin. suggested (Scholl and others, 1970); but system is observed. The lower boundary of the accretionary this process is neither well documented nor The slope between the trench and frontal prism is the upper section of the seismic necessary (Karig, 1974a). arc, in all but very young trenches, consists zone, which apparently widens and flattens Accretion of oceanic sediment and deeper of two sections (Fig. 3). The upper section is as one or more accretionary prisms ac- crustal material onto the upper plate during relatively smooth, reflecting a little- cumulate. The sediment cover on the subduction has been postulated using sev- deformed sediment cover, and contrasts downgoing plate and some of the igneous eral different approaches (Dewey and Bird, with a steeper, less regular lower section crust appears to be stripped off the plate be- 1970; Gilluly, 1972; Moore, 1973; Burk, where sediments are either deformed, fore it reaches a point beneath the volcanic 1965; Hamilton, 1969; von Huene, 1972), acoustically unresolvable, or absent. The chain. Turbiditic sediments deposited in the but as yet the variations in the accretionary two slope sections are separated by a ridge, trench axis are preferentially sheared off the process and in the resultant geologic struc- bench, or slope break, which, because of its underlying pelagic sediments and are ac- tures remain almost completely unknown. variable aspect, has received a number of creted to the lower trench wall. The pelagic In large part, this void results from a lack of designations. In the Indonesian arc system, sediments and crustal material are probably data from the inner trench wall where where it forms a ridge that occasionally accreted at deeper structural levels. accretion is suspected to occur. Acoustic breaks sea level, this boundary has been Where turbidites overlie pelagic sedi- methods cannot adequately resolve the termed the tectonic arc (Vening-Meinesz, ments in the trench axis, the turbidites are geomorphic detail or the internal structure 1964), outer arc (Umbgrove, 1947), or stripped off in fold packets with axial sur- of the constituent rocks, and the conven- nonvolcanic outer arc (Van Bemmelen, faces having very low dips. These dewa- tional bottom sampling too often collects 1949). The basin thus formed in the upper tered and rigidified structural units move up only recent sediment overlying the accreted slope area has been referred to as the inter- the lower slope, as subsequent packets are material. deep (Van Bemmelen, 1949). In some areas, accreted. In trenches that subduct litho- This paper reports the results of a sufficient sediment has been deposited be- sphere carrying very thin pelagic sediment broader scale approach to the problem of hind the boundary to fill the upper slope covers, accretion and uplift of crustal slabs accretion at subduction zones. Digitized basin and to produce topographic benches, seem to occur as topographic irregularities bathymétrie profiles across almost all which have been called terraces (Gates and enter the trench. Key words: marine tec- Pacific and Indonesian trenches (Fig. 1) Gibson, 1956) or deep-sea terraces tonics, island arcs, subduction, trenches. were computer-reoriented, perpendicular to (Hoshino, 1969; Tayama, 1950). Recogni-

Geological Society of America Bulletin, v. 86, p. 377-389, 8 figs., March 1975, Doc. no. 50314.

377

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/86/3/377/3433679/i0016-7606-86-3-377.pdf by guest on 25 September 2021 Figure 1. Index map of the Indo-Pacific region, showing profiles used in this study. Those used and Navy digitized sounding lines deposited in the S.I.O. data bank, along with a few hand re- in other figures in this paper are shown by heavy lines. These are oriented perpendicular to the oriented profiles from references cited in the text. Sediment thicknesses are from the same sources trench and at 20x vertical exaggeration, and are located on Figure 1 by small letters. Other sym- and from Initial Reports of the Deep Sea Drilling Project. bols are shown on the explanation. Profiles are from Scripps Institution of Oceanography (S.I.O.)

i Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/86/3/377/3433679/i0016-7606-86-3-377.pdf by guest on 25 September 2021 SUBDUCTION AND ACCRETION IN TRENCHES 379

Inactive Remnant Marginal Basin Arc

x Outer Swel Z Trench Wedge

Upper Slope -REAR Discontinuity FRONT V. E.= 5x

-500 -400 -300 -200 -100 -50 • 50 f 100 • 200 KILOMETERS Figure 2. Cross section of a typical island-arc system, showing tectonic units and terminology used in this paper (revised after Karig, 1970, 1971a).

tion of the variability of the morphology in and others, 1973b; Creager and others, Japan and eastern Aleutian arc systems, the this two-part slope has led to the more 1973). This zone of deformation between upper-slope discontinuity is a fault zone generalized definitions of midslope base- the trench-slope break and the trench axis (von Huene and others, 1971) or a steep ment high (Karig, 1971a) and "slope limits and probably defines the surface ex- contact without significant morphologic break" (Dickinson, 1971). A nongenetic pression of the subduction zone. expression (Fig. 3). Arc systems having a term, "trench-slope break," has been gen- The rocks dredged or cored from beneath deeper upper-slope trough or apron (such erally agreed upon to cover all aspects of the surficial sediments of the inner trench as the central Aleutians, Marianas, and the boundary between slope sections (Dick- wall are identical or similar in lithology to Luzon systems) display the upper-slope dis- inson, 1973 a) and is used here. those at the trench axis or on the oceanic continuity as the downward extension of a crust (Kulm and others, 1973b; von Huene, steeper upper-slope section, along which Trench Slope Break and Lower Slope 1972; Fisher and Engel, 1969; Hawkins the sediments are faulted, strongly flexed, and others, 1972; Ingle and others, 1973). or lapped against the frontal arc (Fig. 3; Direct observations of the trench-slope Basement rocks on the Mentawai Islands Grow, 1973a; Ludwig and others, 1967). break are restricted to the few areas where and on Barbados can also best be inter- The upper-slope discontinuity in the New the feature emerges as islands. The best ex- preted as having originated in those areas. Hebrides, Solomons, and similar arc sys- amples are Barbados (Caribbean arc sys- This implied upward displacement of the tems without upper slope sediments (Fig. 3) tem; Baadsgaard, 1960; Mesolella and deformed material continues to the crest of is placed at the upper section of the steep others, 1970; Steinen and others, 1973; the trench-slope break. However, down- trench slope, where it forms the forward Lohman, 1974), the Mentawai Islands to-the-rear faulting and folding, and rear- boundary to the frontal-arc platform. (Sumatra arc system; Van Bemmelen, 1949; ward tilting of the upper slope sediments The upper-slope discontinuity in most Hamilton, 1973; D. E. Karig, unpub. field near the trench slope break (Fig. 3; see also arc systems describes a smooth map trace work data), and Middleton Island (Aleutian Figs. 4, 16, and 18 of Ross and Shor, 1965; about 75 km in front of the volcanic chain. arc system; Miller, 1953; Plafker, 1969). von Huene and others, 1971) suggest that In the eastern Aleutians and Hikurangi All show rapid tectonic uplift of Quater- the upper slope area is not rising as rapidly areas, where sediment influx to the subduc- nary reefs or wave-cut terraces. The bed- as the trench-slope break and may even be tion zone is very high and the distance rock beneath the surficial carapace of reef subsiding relative to sea level. from the volcanos to the trench broadens, or terrace deposits consists of either highly The sediments of the upper slope, be- the volcanic-chain-upper-slope discon- deformed, steeply dipping clastics which tween the trench-slope break and the fron- tinuity separation also increases. In the may contain slices of basalt or ultramafic tal arc, are usually synformal or show a Bonin arc system, where the frontal arc is rocks, or less steeply dipping, nondisrupted homoclinal tilt toward the frontal arc. composed of en echelon ridges (Karig, clastics or carbonates. On the Mentawai Is- These sediments may reach a thickness of 1971a, and unpub. charts), the upper-slope lands and Barbados (Lohman, 1974), slope several kilometers (Ludwig and others, discontinuity follows the frontal-arc offsets deposits overlie more highly deformed 1966, 1967; Grow, 1973a; Den and others, and has an irregular trace. rocks. This relationship is amplified by 1968). In contrast, the frontal arc is The fundamental structure of this impor- seismic reflection profiling in many arc sys- strongly emergent, especially along its sea- tant feature is not apparent. Small normal tems that show that upper slope sediments ward edge (Fitch and Scholz, 1971; Plafker, faults on the frontal arc of the Tonga arc are openly folded over an opaque basement 1969, 1972; Mitchell, 1969). To a close system, interpreted as synthetic (Karig, core at the trench slope break (Hilde and approximation, the frontal arc is a rising 1970), may instead be antithetic, gravity- others, 1969; Ludwig and others, 1967; and rearward-tilting or rotating block, with driven features. Grow (1973a) and Mur- Grow, 1973a, 1973b; Karig, 1973a). little internal deformation. There is, there- dock (1969) discuss the possibility of high- Deformation near the trench slope break fore, a strong tectonic discontinuity be- angle reverse faulting along this zone in the apparently marks the upper boundary of a tween the frontal arc and the upper-slope Aleutian system. In this paper, only near- zone of active deformation that extends to sediment pile. vertical basement displacement is assumed the base of the inner trench slope. Available and the upper-slope discontinuity will be data indicate that the maximum rate of de- Upper Slope Discontinuity used to mark the forward edge of the fron- formation occurs very close to the foot of tal arc. That volume of material between the lower trench slope (Ingle and others, The boundary between the frontal arc the discontinuity and the trench will be re- 1973; J. C. Moore and Karig, in prep.) and and the upper slope takes several different ferred to as the accretionary prism (Fig. 2). continues at a decreasing rate up the slope forms, all of which will be referred to as the Use of this term will hopefully be justified in to the crest of the trench-slope break (Kulm upper-slope discontinuity (Figs. 2, 3). In the the following discussions.

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/86/3/377/3433679/i0016-7606-86-3-377.pdf by guest on 25 September 2021 380 KARIG AND SHARMAN

BOUNDARY CONDITIONS IN nately, lack of exposure and the difficulty in conditions are those of the Shikoku, East ISLAND-ARC SYSTEMS resolving deformation ages of rocks in- Luzon, New Hebrides, Solomons, and New volved in the subduction zone prevent di- Britain arc systems. All of these are newly Determination of the quantity of material rect dating of subduction activity. occupied or strongly rejuvenated crustal in- accreted, the rate at which accretion occurs, Major discontinuities can often be de- terfaces of Miocene or younger age. and the disposition of accreted material duced from the geologic history along plate The west Melanesian trenches (New within the arc system are important in any margins. These may result from regional ef- Hebrides, Solomons, and New Britain) are study of subduction zones. However, there fects, as the shift in subduction zones from all similar in character. Instead of a distinct are several fundamental problems which one small plate boundary to another (for upper slope, these arc systems have a single must be attacked before these more specific example, the Shikoku system; Ingle and steep slope leading from the edge of the questions can be answered. These concern others, 1975) or arc polarity reversal with- frontal arc to the trench axis (Fig. 3; see the tectonic boundaries of the subduction out collision (for example, New Hebrides; also Hayes and Ewing, 1970, Fig. 11). zone and their stability over time, the initial Karig and Mammerickx, 1972). Igneous rocks have been dredged from most condition of a subduction zone, and the ex- In addition to the reorganization of tec- of these slopes (Petelin, 1964). The 90- to tent of temporal continuity at consuming tonism within arc systems and among a 130-km separations between the trench plate margins. complex of downgoing plates, discon- axis and the volcanic chain are the smallest tinuities in subduction might also be attrib- observed in contemporary arc systems. Temporal Continuity of uted to more general changes in plate mo- The East Luzon and Shikoku arc systems Subduction tions (Coney, 1972; Larsen and Chase, represent a second style of initial morphol- 1972). For instance, steady-state extension ogy, with a small but well-defined trench- Although it has generally been assumed along the East Pacific Rise could be coupled slope break and upper-slope sediment that tectonic activity in orogenic zones is with sharp changes in subduction rates trough (Figs. 4, 5B). The characteristics of discontinuous, the persistence of crustal ex- along the two sides of the Pacific basin. trench rejuvenation are especially well tension along individual spreading ridges shown along the East Luzon system, which over periods of time often exceeding 100 Initial Configurations of resumed subduction in the Quaternary m.y. requires that the assumption of discon- Subduction Zones (Fitch, 1972; Karig, 1973b and unpub. tinuous subduction be examined. Indirect data). Because subduction has not yet evidence supporting episodic subduction is Newly created or rejuvenated subduction propagated along the entire Luzon coast, supplied by the discrete pulses of arc vol- zones should most nearly display the initial pre-existing conditions can be observed off canism (Dott, 1969; Mitchell and Bell, morphology and structure of the inner northeast Luzon (Karig and Wageman, 1970) and inter-arc basin extension (Karig, trench wall. The young trenches that ap- 1975). This boundary became inactive in 1971b) along many arc systems. Unfortu- pear to most closely approximate initial the mid-Tertiary and was subsequently the site of a moderate-sized continental rise wedge. Renewed underthrusting seems to have deformed this wedge and generated a small but well-formed accretionary prism (Fig. 4). Sediments on the lower slope or uppermost rise have been tilted westward toward the frontal arc (Luzon). More recent sediments have collected behind the accreted material to form an incipient upper-slope basin. In this case, the upper- slope discontinuity is clearly a remnant of the pre-existing continental or insular slope. A similar but more advanced situation can be observed along the Shikoku or southwest Japan system (Karig, 1975). Following the rapid accretion of a continental or insular rise wedge, the rate of accretion would drop to a lower, more nearly steady state rate. Thus, the difference in initial configuration between the Shikoku-Luzon and west Melanesian types is explained by the greater sediment build- up at the slope bases of the larger continental masses. Different sediment feed rates to the subduction zone from the offshore basin and trench could also be contribu- tory, but even several hundred meters of sediment subducted at a rate of 10 cm/yr or more since the end of the Miocene or longer has not begun to build an appreciable trench-slope break in the New Hebrides arc system. These two types of young arcs may reasonably be used as initial configurations the wide range in width and shape of the accretionary prism. The axis of the volcanic chain (v) is also of subduction zones, and they suggest that indicated. Profile e displays relative subsidence and rearward tilting of the upper slope sediments. accretion causes the enlargement of the area Profile d (from Menard, 1964) was enhanced with reflection data from Hayes and Ewing (1970). between the upper-slope discontinuity and

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/86/3/377/3433679/i0016-7606-86-3-377.pdf by guest on 25 September 2021 SUBDUCTION AND ACCRETION IN TRENCHES 381

the trench. The accreted material also forms tion of upper-slope sediments against upper-slope discontinuity for periods of 40 a basement upon which the upper-slope frontal-arc basement is thus in part an m.y. or more. sediments can accumulate. onlap relationship and in part a result of Upper-slope sediments appear to be in- structural displacement. corporated into the continent within broad The Upper-Slope Discontinuity Although the upper-slope discontinuity prisms of subduction materials. In the east- as a Tectonic Reference quite clearly appears to originate as an ern Aleutians, Burk (1965, 1973) has de- upper continental slope, it is not at all clear scribed several such prisms, separated by Analyses of the youngest subduction that this boundary remains spatially fixed steep faults that rotate the prisms away zones leads to the indentification of the during subduction, or that there is a one- from the trench. These prisms become suc- upper-slope discontinuity as the upper sec- to-one correlation between pulses of sub- cessively older toward the north and are at- tion of the continental slope which existed duction and positions of the upper-slope tributed to progressive subduction, but data before the subduction pulse began. Mor- discontinuity. are lacking by which to interpret the phologic relief between the frontal-arc plat- If there were a steady trenchward migra- significance of the individual prisms. Mat- form and the upper-slope sediments in very tion of the upper-slope discontinuity, a suda and Uyeda (1971) suggest a similar young subduction zones is greater than 2 thick section of upper-slope sediments stepwise progression of subduction in km and generally decreases with increasing should be exposed along the leading edge of Japan, in which the area between the vol- age and growth of the accretionary prism the frontal arc. Such deposits have not been canic front and the trench becomes the (Figs. 4, 5). Low relief on older zones ap- observed. Rather, the sediments overlying frontal arc during successive subduction. pears to strongly reflect high sedimentation the frontal-arc basement are generally thin, In the Mariana system, the upper-slope rates on the upper slope and growth of the shallow-water strata, and, together with discontinuity has remained fixed probably trench slope break as demonstrated along reflection profiles which show no uplift of since the Eocene and perhaps longer (Karig, the eastern Aleutian, southern Kermadec, the upper-slope sediments relative to the 1971a), but three pulses of subduction are and Japan-Bonin systems. The juxtaposi- frontal arc, they indicate stability of the suggested by the volcanic activity and marginal basin history. Along the Tonga- Kermadec system, and perhaps in other areas, the situation appears similar. The apparent contradiction in correlation of the 1.0—1 upper-slope position with pulses might be explained either by the reactivation of the x same trench interface in arc systems where x the accretion rate is slow, or more likely, by K Q. 2.0 — attributing the pulses of marginal basin and UJ Q volcanic activity to discontinuities in sub- Ct duction rates rather than to absolute sub- UJ duction halts. 1 3.0— In this paper, the upper-slope discontinuity is used as a stable rear boundary to / the accretionary prism that is presently being developed. An upper limit to the time spanned by this accretion is given by the age of the youngest subducted and accreted rocks exposed on the frontal arc or by the oldest shelf sediments overlying them.

The Volcanic Chain as a Tectonic Reference

A second possible tectonic reference for the growth of arc systems is the volcanic chain. This marker would be useful over longer time spans than the upper slope discontinuity would be and might provide information concerning the average rates of accretion. The volcanic chain or front describes a remarkably smooth trace in most arc systems and generally lies above the 125- to 175-km-deep earthquakes of the seismic zone. Migration of the volcanic chain both toward (Matsuda and Uyeda, 1971) and away from (Dickinson, 1973a; Martin- Kaye, 1969) the subduction zone relative to the crust beneath has been proposed. Where arc systems have most clearly been accreting material and have not suffered in- Figure 4. Seismic reflection profile and interpretation across the inner slope of a nascent trench east ternal tectonic disruption, the axis of max- of Luzon. Notice the very great relief of the upper slope area. The small accretionary prism is built of imum igneous activity has either remained deformed insular rise and lower slope deposits (unpublished core data from S.I.O. Cruise Tasaday). fixed or has shifted rearward at average Younger deposits are ponded behind the small trench slope break. rates of up to 2 km/m.y. (Dickinson, 1973a;

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/86/3/377/3433679/i0016-7606-86-3-377.pdf by guest on 25 September 2021 te £ 4

I 1 1 1 1 I i i i i 0 50 100 "150 -100 - 50 0 50 100 150 200 250 300 Distance in kilometers Distance in kilometers

• o

2 1 1 5' Figure 5. Comparative bathymetric profiles across inner-trench slopes displaying a variety of morphologic styles and the independence between style and size (from 1). Group A are those trenches accreting high density, high velocity material. Group B are those where the sediment feed into the trench exceeds sedimentation of the upper slope. Group C are those where sediments from the frontal arc fill the upper slope and augment the sediment feed to the trench. Profiles are aligned on the upper slope discontinuity and the volcanic chain (V) is noted where present. Profile A-b is from Fisher and Engel (1969) and A-d from Menard (1964).

Distance in kilometers

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/86/3/377/3433679/i0016-7606-86-3-377.pdf by guest on 25 September 2021 SUBDUCTION AND ACCRETION IN TRENCHES 383

Kawano and Uyeda, 1967; Peck and others, 1969; Petelin, 1964) to consist largely of the availability of sediment. The upper- 1964). basalt and ultramafic rocks. Although slope discontinuity-trench separation and The cause of the apparent rearward mi- ocean-floor pelagics have not been de- other gauges of accretion generally increase gration of the volcanic chain could be either scribed from dredge hauls, these may be with parameters that reflect increasing du- a trenchward shifting of the frontal arc present but perhaps have been washed from ration of subduction, but the great scatter crust, relative to a volcanic source fixed rel- the dredge during retrieval. On several in these relationships (Dickinson, 1973a) ative to the mantle and Benioff zone, or a profiles where the effects of down-slope indicates large variations in rates of accre- rearward shift of the seismic zone relative sediment redeposition have not masked the tion. These rates of accretion depend on to the mantle and crust above. Dragging of internal structure of the apron, prograda- rates of sediment influx, which are only the upper plate into the subduction zone is tion of the sediment pile over an irregular partly related to subduction rates. For in- precluded by accretion rather than erosion but shallow-dipping basement is observed stance, the southward widening of ac- on the inner trench slope. Some form of (Fig. 3; Fischer and others, 1971). The up- cretionary prisms in the Kermadec and crustal extension along the volcanic chain is permost sediment unit spills over the Caribbean arc systems are clearly the effects another alternative. Extension within and trench-slope break, suggesting that the of high sediment feed rates from the adja- behind the volcanic chain, which results in apron can extend trenchward only as the cent continents rather than of changes in the development of marginal basins, drives basement platform grows by accretion. the duration or rates of subduction along the entire arc system forward as a unit and In a second major type of configuration, the arcs. The morphologic variations of the does not appear applicable. Crustal exten- the trench-slope break is a ridge, behind accretionary prism seem to be more depen- sion resulting from plutonism beneath the which a sediment-filled trough develops dent upon relative rates of sedimentation volcanic chain does occur, but in the better (Fig. 5B). The Sumatra, Java, Luzon, cen- from the various sources feeding the sub- mapped volcanic zones is not of the neces- tral Aleutian, and Shikoku arc systems dis- duction zone than on rate or duration of sary scale (Peck and others, 1964; C. A. play this morphology. Seismic refraction subduction. Sediment that passes into the Hopson, 1974, personal commun.). Rear- studies (Ludwig, 1970; Raitt, 1967; Den accretionary prism at the base of the inner- ward shift of the 125- to 175-km-deep part and others, 1968) and the geology of is- trench wall can be derived from either side of the seismic zone remains a possible solu- lands on the trench-slope break indicate of the trench and by pelagic, hemipelagic, tion but is not easily tested. that the material underlying this form of and density-flow mechanisms. In general, volcanic chains can be fairly accretionary prism is largely sedimentary The basal section of the sediment pile en- stable or tend to shift rearward during and contains only minor amounts of igne- tering the subduction zone consists of periods of as much as 50 m.y. or more. The ous rock. pelagic or terrigenous sediments deposited increase of the volcanic-chain—trench sep- A third variety of inner trench slope dis- on the downgoing plate before it begins to aration is a function of both the migration plays a trench-slope break along the edge of descend. These vary in thickness with the of the volcanic chain and the growth of the a broad continental shelf, at depths up to age of the plate, with the productivity of the inner-trench slope (Dickinson, 1973 a). It is several kilometers (Fig. 5C). The eastern waters beneath which it has traveled, and thus a poor reference by which to measure Aleutian, Japan, and sections of the Middle with available terrigenous sources. Sedi- accretion. America arc systems are in this category. A ment thickness varies from less than 200 m structural high at the shelf edge, forming on the young crust subducted into the Mid- GEOMETRIC VARIATIONS OF the trench-slope break, is sometimes ex- dle America Trench (Frazer and others, THE ACCRETIONARY PRISM pressed morphologically but more often is 1972; Riedel and others, 1961) to more discernable only on seismic-reflection than 3 km where the Bengal fan is sub- The accretionary prism, bounded by the profiles (von Huene and others, 1971; Ross ducted in the Sumatra Trench (Curray and upper-slope discontinuity behind and by and Shor, 1965). Moore, 1971). The rate at which this sedi- the seismic zone beneath, is discussed most There is no correlation between the type ment is fed into the subduction zone is a easily in two parts. The upper section, of prism morphology and the distance from function of its thickness and the component above the trench floor, is analyzed using the trench to the upper slope discontinuity of the subduction rate acting perpendicular morphologic, seismic reflection and refrac- (Fig. 5). The upper slopes in the Skikoku, to the trench axis. tion profiles, and geologic data. Analysis of Aleutian, and Sumatra systems all have As the downgoing plate, with its sedi- the deeper section, for which far fewer data ridge and trough morphologies, but with ment cover, descends into the trench, it be- are available, centers on the behavior of the very different widths. The upper slopes of comes covered by a wedge of clastic sedi- seismic zone beneath the prism during sub- the Mariana and Tonga arcs, although ments that is fed to the trench from shallow duction. seemingly primitive, are wider than the depths through canyon systems and by ridge-trough upper slopes of many arc sys- other downslope mechanisms (Piper and Morphology and Shallow Structure tems. The relief of the upper part of the ac- others, 1973; Ross 1971a, 1971b; If the New Hebrides and East Luzon cretionary prism above the trench floor is as Anikouchine and Ling, 1967). In all but the trenches approximate initial conditions of important as the width in determining its most slowly subducting trenches, the clastic subduction zones, accretionary enlarge- bulk. This relief varies from about 2 km in wedge is less than a few tens of kilometers ment can assume one of three general the Shikoku system to more than 6 km in wide and a kilometer or so thick. The ma- configurations, or some morphologic tran- the Mariana system. Relief increases jority of trenches have either no such sition between these (Fig. 5). The simplest roughly with the width of the upper slope wedge, or more likely, wedges so narrow type, represented by the Tonga, Mariana, area, but not in a simple way. The Tonga that they are unresolvable by surface acous- and parts of the Bonin and Kermadec arc sys- and Mariana systems have an accretionary tic techniques (Fisher and Hess, 1963). Al- tems, displays a two-part slope, consisting prism whose upper section is relatively though the size of the clastic wedge may be of an upper slope apron that is separated thick and narrow, whereas that of the quite small at any instant in time, the resi- from the steeper lower slope by a simple Sumatra system is broad and thin. dence time of any piece of sediment in the slope break (Fig. 5A). The acoustically wedge is quite short, so "that the total quan- opaque material constituting the basement Factors Affecting the Size and Shape tity of sediment fed to the subduction zone on which the upper slope apron rests ap- of the Accretionary Prism may be large. A direct measure of the rela- pears from available refraction studies tive amounts of sediment supplied to the (Murauchi and others, 1968) and from The most obvious factors affecting the subduction zone by the wedge and by the dredge hauls (Yagi, 1960; Fisher and Engel, amount and style of accretion are time and downgoing plate cover is given by the

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/86/3/377/3433679/i0016-7606-86-3-377.pdf by guest on 25 September 2021 384 KARIG AND SHARMAN

thickness of each type of sediment at the the lower, highly porous section to deeper eastern end of the Aleutian system receives inner-trench wall. levels (Karig, 1974b). Such a décollement a much greater sediment influx than do The wedge will remain the same size if has been observed in DSDP Site 298 in the most other arc systems (Plafker, 1972), the sediment influx (sedimentation rate x Nankai Trough off Shikoku (Ingle and more deposition on the lower wall and wedge width) is balanced by the outflow others, 1975). more slumping of this material is expected. (component of subduction rate in horizon- In oceanic arc systems, where the turbi- Near-bottom investigations in the central tal plane and perpendicular to trench X dite influx is relatively small, the accretion- part of the Aleutian arc system (Grow, wedge thickness corrected for compaction). ary prism should be constructed largely of 1973a) and a detailed survey in the Nankai Changes in either sedimentation or subduc- pelagic sediments and igneous rocks, but Trough (Ingle and others, 1975) found tion rate will be reflected by a change in the accreted materials are very seldom exposed linear ridges on the lower trench wall rather wedge size, but simple calculations show in these arc systems. Where collision proc- than the equi-dimension topography sug- that the new equilibrium size will be esses have exposed accretionary prisms of gested by von Huene (1972). logarithmically approached within a few oceanic arcs, as in eastern Mindanao The ponded and tilted sediments seem million years in most cases. (Melendres and Comsti, 1951; Irving, more likely to have been deposited between The few available sedimentation rates in 1950, 1952; Hamilton, 1973; Karig, 1975), active thrust units (Fig. 7A). In both the trenches (Table 1) indicate that there is no they do contain a large proportion of Nankai Trough (Hilde and others, 1969; inconsistency between the size of the wedge pelagic and igneous material. Ingle and others, 1975) and the eastern and the rate of sedimentation, even assum- The Franciscan complex, in which turbi- Aleutian Trench, the pond size and general ing a steady-state condition. The conclusion dites predominate over pelagic sediments slope sediment thickness increase upslope, that the wedge in the Peru-Chile Trench and basaltic rocks (Bailey and others, 1964), which is to be expected if they develop in represented a time span much greater than is still best attributed to subduction in increasingly older thrust units upslope. It is that allowed by calculated subduction trenches. The nearly ubiquitous association assumed here that slumping of material rates (Scholl and others, 1970) was based of pelagic shales or chert with pillow from the inner wall is generally restricted to on less direct methods, which are felt to be basalts (Bailey and others, 1964) implies the unconsolidated cover and that it forms less reliable. that these sediments were deposited in an one end of the range of mechanisms by Scholl and Marlow (1972, 1974) have ocean basin rather than on the upper trench which shallow-water sediments are fed into suggested that the turbidite wedges are a re- or continental slope. The fact that these the trench wedge. Other sources of feed sult of lowered Pleistocene sea levels and of diagnostic pelagic-shale and pillow-basalt into the trench, including pelagic and glacial erosion. They claim that, without sequences are intercalated throughout a hemipelagic biogenous and detrital sedi- these effects, most Pacific trenches would be complex of highly deformed rocks consti- ments, are generally less important than devoid of turbidite fills. Although it is un- tutes a strong argument that the associated those already discussed. doubtedly true that Pleistocene glaciation turbidites originated in a trench. The amounts and relative proportions of increased the supply of terrigenous sedi- Another unresolved problem concerning sediments fed into the subduction zone and ment to the trenches, it does not necessarily turbidite sedimentation and deformation is to the upper-slope area from various follow that the observed abundance of ter- the importance of slumping and gravity sources determine the shape of the ac- rigenous and volcanic-rich turbidites in sliding on the inner trench slope (Chase and cretionary prism. Arc systems in which the mélange complexes of all ages requires a Bunce, 1969; Scholl and von Huene, 1970; accretionary prism is subdued and contains site other than the trench wedge for their von Huene, 1972). Shear strength tests a large fraction of oceanic basement rocks origin. (Ross, 1971b) have suggested that the (type A of Fig. 5) result from the subduction Trenches with substantial turbidite pelagic and hemipelagic material deposited of lithosphere with a thin sediment cover, wedges are not restricted to glacial regions on the inner trench wall may be unstable together with a low rate of turbidite influx or to those with broad shallow shelves, al- and susceptible to slumping. Small mud to both the trench and the upper slope. In though such wedges are more common in balls and semi-lithified clasts incorporated arc systems where the feed rate of sediments those areas. Trench wedges occur in areas in the turbidites cored from several trenches on the downgoing plate is higher and the where there was little or no glaciation and (Anikouchine and Ling, 1967; J. C. Moore turbidite influx from the frontal arc is rela- where there are very narrow shelves (Fig. and Karig, in prep.) are evidence that such tively low, the trench slope break develops 1). A turbidite wedge more than 5 km wide displacement occurs. In sequences thought into a high ridge of deformed sediment and 400 m thick occupies sections of the to represent trench-wedge sediments, slump (type B of Fig. 5). The deep trough behind axis of the southern New Hebrides Trench folds have been described (Moore, 1973), reflects the inability of land-derived sedi- where the only terrigenous sources are the but few instances of mass gravity move- ments to fill this trap. Those arc systems small recent volcanos of Matthew and ment of previously accreted material have with high terrigenous feed rates pour Hunter Islands. been documented. enough sediment into the trench wedge to The predominance of turbidites in most Von Huene (1972) has attributed irregu- produce large accretionary ridges and also accretionary prisms exposed along conti- lar topography and tilted sediment ponds to keep the upper slope basin filled so that a nental margins is most probably the result on the inner slope of the eastern Aleutian shelf or terrace develops (type C of Fig. 5). of preferential accretion of the more rigid Trench to large-scale slumping and trench wedge turbidites and subduction of backward-rotating slide blocks. Because the Configuration of the Seismic Zone and Deep Structure of the TABLE 1. SEDIMENTATION RATE AND TURBIDITE WEDGE SIZE IN TRENCHES Accretionary Prism

Example Aleutian Middle America Shikoku (von Huene, 1972) (Ross, 1971) (Ingle and The seismic zone, which forms the lower others, 1975) boundary of the accretionary prism, is as-

Width of wedge (km) 30 13 17 sumed, in its upper section, to mark the Max. thickness of zone of shearing between the two litho- wedge (km) 0.8 1.25 0.7 spheric plates (Isacks and Molnar, 1971). Subduction rate (km/m.y.) 60 65 (Le Pichon 20 The trench marks the intersection of this in- and others, 1973) Sedimentation rate terplate boundary v/ith the ocean floor. (km/m.y.) 1 to 2 1 (observed) 0.8 Trench depth and position are functions of (observed) (calc) a surprisingly few variables. Rate of sub-

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/86/3/377/3433679/i0016-7606-86-3-377.pdf by guest on 25 September 2021 SUBDUCTION AND ACCRETION IN TRENCHES 385

duction does not appear to strongly affect or front. Between the volcanic chain and tween plates. Some additional control for trench depths. A large southward decrease the upper-slope discontinuity, the seismic the position of the upper section of the in subduction rate is not expressed in trench zone begins to flatten sharply (Mitronovas seismic zone is obtained from seismic re- depths along the Bonin and Mariana-Yap- and others, 1969; Karig, 1971b; Fedotov, fraction profiles and from focal Palau system or along the Tonga-Kermadec 1968), but the range over which the mechanisms. system beyond the influence of sedimenta- flattening occurs and the shapes of this The small depth range to the seismic zone tion from New Zealand. Shallow trenches, upper zone vary widely (Fig. 6). In the New beneath the volcanic chain or volcanic front such as the Middle America and Java, are Hebrides, the change in seismic zone dip implies that the upper end of the deeper among more rapidly subducting examples. from 70° at depth (Dubois, 1971; G. Pascal, part of the Benioff zone remains stationary The most clear-cut factor controlling 1974, personal commun.) to 8° at the or migrates slowly rearward over periods of trench depth is the general depth of the top trench axis (Fig. 6) occurs rapidly in hori- up to 200 m.y. The width of the upper sec- of the downgoing plate before it descends zontal distances of less than 100 km. In tion increases directly as the separation be- into the trench. Thus, along one or more contrast, the seismic zones beneath the tween the trench and volcanic chain (Fig. arc systems where the shape of the inner eastern Aleutian (Plafker, 1972; Lahr and 8). In part, this reflects the rearward move- and outer trench walls are similar, the Page, 1972; Lahr, 1973, written commun.) ment of the volcanic centers, but it is pre- depth to the oceanic basement at the trench and southern Kermadec (Hamilton and dominantly a correlation with the amount axis increases directly as the water depth in Gale, 1968) systems begin to flatten sharply of accreted material. There is also a correla- the basin being subducted (Le Pichon and under the frontal arc and continue at dips tion between the width of the accretionary others, 1973). Profiles along a given arc sys- of less than 20° for several hundred prism and the width of the flattened zone, tem, normalized so that the upper sections kilometers before surfacing at the trench but, as illustrated in the eastern Aleutian of the downgoing plate are superimposed, with dips of 2° to 4° (Fig. 6). Some systems, system, the flattened section of the seismic demonstrate that the trench depth and posi- such as the Mariana (Katsumata and Sykes, zone spans several active and inactive ac- tion are also functions of local differences 1969) and Kermadec (Sykes, 1966) arcs, cretionary prisms representing more than in accretion along the inner-trench wall have very steep, deeper seismic zones and 150 m.y. of subduction (Burk, 1965). The (Karig and others, in prep.). broad upper sections, but others with correlation of the flattened upper section The shape of the upper part of the down- equally steep lower sections (New Heb- with the width of accreted material and going plate, however, is not constant rides; Dubois, 1971; and Solomons; Den- with the outward migration of the gravity among the various arc systems (Fig. 6). ham, 1969) have narrow upper sections. minimum related to the depressed slab Mechanical properties of the subducted There appears to be no relationship be- (Karig and others, unpub. data) is the lithosphere (Le Pichon and others, 1973; tween the dip of the lower part of the zone strongest argument favoring depression of Watts and Talwani, 1974) are probably re- and the configuration of the shallower sec- the oceanic lithosphere by the weight of ac- sponsible for a part of this variation in tion. creted material and upper slope sediments shape. In addition, young plates and those The shape of the upper section of the (Fig. 6). of marginal basin origin have smaller radii seismic zone is constrained by the slope of The amount of material fed to the sub- of curvature than do older oceanic plates. the downgoing plate at the trench axis and duction zone that enters the accretionary The shape of the downgoing plate also ap- by the position of the deeper seismic zone prism can be determined, in principle, by pears to be strongly controlled by the load- where hypocenters can be fairly well lo- comparing the amount of sediment with the ing due to the overlying accretionary prism cated. Shallow hypocenters seldom aid in increase in size of the accretionary prism (Karig and others, in prep.). defining the upper seismic zone because of from an initial configuration. The volume In nearly all arc systems, the seismic scatter resulting from inhomogenous shal- of the prism must be corrrected for sedi- zone, regardless of its shape, is at depths of low velocity structure and from the lack of ment compaction during accretion and 125 to 175 km beneath the volcanic chain restriction of seismicity to the interface be- must be reduced by the amount of sediment

Volcanic South East Chain New Hebrides Trench Tonga Trench Aleutian Trench J\ i i S. L. _ I East. Aleutian^

/V/ r /Mf

i 1 1 i i i i i i i 0 50 100 150 200 250 300 350 400 450 KILOMETERS Figure 6. Configurations of the upper sections of three better-known Benioff zones, normalized to the volcanic chain, demonstrating the pronounced flattening with the growth of the accretionary prism. Control is provided by earthquake hypocenters (dotted areas) and by the dip of the downgoing plate at the trench axis and focal mechanism (thin lines). Data sources are, for the New Hebrides: Dubois (1971), Johnson and Molnar (1972); for the Tonga: Mitronovas and others (1969), Isacks and others (1968); and for the Aleutian: Lahr (1972, personal commun.); G. G. Shor, in von Huene (1972).

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/86/3/377/3433679/i0016-7606-86-3-377.pdf by guest on 25 September 2021 386 KARIG AND SHARMAN

TABLE 2. ACCRETIONARY PRISM VOLUME CALCULATIONS* Mode of Deformation within the

1 Trench E. Luzon Shikoku S. New Hebrides Tonga Mariana E. Aleutian Accretionary Prism

Duration of subduction pulse «1 m.y. 10 m.y. 3 m.y. 45 m.y. 45 m.y. 40 m.y.? Reflection profiles across active subduc- Average subduction tion zones and mapping in older exposed rate ? 1.5 cm/yr 5 cm/yr 7 cm/yr 6 cm/yr 5 cm/yr subduction terranes suggest a wide varia- Average sediment tion in the structural style and lithologie thickness 2 km 1.5 km 0.6 km <0.4 km 0.4 km 0.5 km Total sediment feed content of accretionary complexes. Some, (40% sediment such as much of the southern part of the porosity reduction assumed during Franciscan complex, show relatively little 2 2 2 2 2 2 subduction) 30 km 125 km 50 km 700 km 600 km 600 km folding and much shearing and are com- Accretionary prism area (minus slope prised dominantly of turbidites and secon- 2 2 sediments) 50 km2? 550 km2 500 km2 1,200 km2 2,000 km 1,200 km darily of pelagic sediments and basement * Per unit length of arc. rocks (Bailey and others, 1964; Hsii, 1969). t After von Huene, 1972. Other examples show more intensely folded turbidites with only a small percentage of deposited on the inner slope. Unfortu- velocity material below the inner slopes of igneous rocks (Moore, 1973). Still others nately, the present resolution of sedimenta- the Shikoku (Den and others, 1968) and seem to be comprised predominantly of tion and subduction rates and of the eastern Aleutian (G. G. Shor, in von Huene, pelagics (Melendres and Comsti, 1951). seismic zone configuration permits only 1972) and the magnetic sources beneath the This variation appears to reflect differences approximate calculations. central Aleutian slope (Grow, 1973a). It is in sediment cover on the downgoing plate, The most reliable data show, however, very unlikely, therefore, that sedimentary in rates of subduction, and perhaps other that the accreted volume on the upper plate material is carried down the seismic zone to factors, such as obliquity of subduction. significantly exceeds the volume of sedi- the area feeding the volcanic chain. Effects Processed multi-channel seismic ment fed to the trenches (Table 2). Accre- of sediment on magma composition are reflection profiles over trenches that sub- tion of at least a part of the oceanic second more likely accomplished by interaction of duct thick sediment sections (Beck, 1972; layer, required by these calculations, would the magma with earlier subducted meta- Beck and Lehner, 1974) reveal that the sedi- explain the thickened sections of high- sediments at shallow levels. ments are stripped off the oceanic basement and support the idea that accretion occurs in fold or thrust packets (Fig. 7A; Chase and Bunce, 1969; Karig, 1974b). Deep-sea drilling and related investigations in the Nankai Trough (Shikoku system) Trtnch Slop« Brtak demonstrated that trench-wedge turbidites are accreted in nearly recumbent fold units that are expressed on the lower slope as linear ridges (Ingle and others, 1975; Moore and Karig, 1975). Dewatering and rigidification of the accreted sediments begins during the very Vert Eiogg«2.5> early stages of deformation, even before the unit is effectively removed from the downgoing plate, in a proto fold region (Moore and Karig, 1975). This response, in which 15 20 25 30 35 40 45 50 55 60 65 70 75 the well-developed folds sometimes face Kilometers away from the trench (Silver, 1972), seems typical of the slower subducting trenches with thick sediment fills (for example, the Lesser Antilles, Shikoku, West Luzon, and Cascades). In more rapidly subducting trenches with high sediment feed, dewatering might not occur until more deformation had occurred, and then a different structural response could be expected. By the time the deformed sediments are exposed on the trench-slope break, they form steeply dipping, generally isoclinal folds (Moore, 1973). In some cases (Tabor and others, 1970; Tabor, 1970, personal commun.; Stewart, 1970), the fold units, although internally deformed, are separated by zones of much more intense deforma- B Kilometers tion. These shear zones may form during the initial décollement of trench-wedge tur- Figure 7. Speculative modes of deformation of accreted material on the lower branch wall (Karig, 1974b). A. Accretion of thick sediment cover or thick trench-wedge section. The upper turbidite sec- bidites or may reflect the decreasing inten- tion tends to be sheared off along the weak, high-porosity uppermost pelagic section and rides over the • sity of deformation in increasingly compe- trench wedge, probably aided by high pore pressures. B. Accretion of thin pelagic sediment cover and tent material as it moves toward the oceanic crust. Slabs of the upper oceanic crust are intermittently sheared off when topographic ir- trench-slope break. Subsequent accretion regularities enter the trench. beneath and horizontal compressive stress

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/86/3/377/3433679/i0016-7606-86-3-377.pdf by guest on 25 September 2021 SUBDUCTION AND ACCRETION IN TRENCHES 387

Benioff zones with very broad and flattened upper sections. In the eastern Aleutian system, no permanent seaward migration of the volcanic chain has occurred for more than 150 m.y., and the distance from volcanos to trench now exceeds 400 km. This displays the extent to which this process can proceed. If the overstepping of the low P/T thermal regime (associated with the volcanic chain, over the high P/T zone, which probably ex- ists beneath the trench slope break area) does occur, it must happen over a greater time span. Simple accretion has not per- sisted for such great periods of time along most plate edges. Rather, tectonic disruption by collision, longitudinal shifting of accreted zones during oblique subduction, and polarity reversals overprint the simple picture.

ACKNOWLEDGMENTS

0 10 20 30 40 50 60 70 60 90 100 110 I2Q 130 140 Kilometers The ideas presented here have been mel- Figure 8. Hypothetical "building out" of accretionary prism, showing migration of the morphol- lowed by discussions with J. C. Crowell, C. ogy over the accreted material. The original trench slope break becomes part of the subsiding base- A. Hopson, W. R. Dickinson, P. J. Coney, ment of the upper slope area. Numbers mark arbitrary successive locations of the trench-slope break. and colleagues at Cornell University since first presented in May 1972 at a U.S. Geological Survey "Pick and Hammer" talk. Special acknowledgment is due Xavier can explain both the marked rotation of the until the accretionary prism becomes quite Le Pichon, with whom most of the profiles fold units and the rearward tilt of sediments wide and is best ascribed to the loading ef- were collected and initially analyzed. This in the ponds on the lower trench slope. fects by the prism on the downgoing plate. study was supported by National Science The style of deformation in arc systems At this stage, the trench-slope break ap- Foundation Grants GA-35990 and which subduct plates with thin sediment pears to mark the point where thickening of GA-38107. covers may be similar to that displayed in the accretionary prism is balanced by de- the southern part of the Franciscan com- pression of the seismic zone. The cause of REFERENCES CITED plex, with relatively unfolded and unde- displacement along the upper-slope discon- Anikouchine, W. A., and Ling, Hsin-yi, 1967, formed slabs of sediment and basement tinuity is not obvious, but it is suggested Evidence for turbidite accumulation in separated by shear or mélange zones (D. L. that this boundary marks the mechanical trenches in the Indo-Pacific region: Marine Jones, 1973, personal commun.; Karig, leading edge of the upper plate and that the Geology, v. 5, p. 141-154. unpub. field mapping). Igneous rocks, accretionary prism is nearly mechanically Baadsgaard, P. H., 1960, Barbados, W. I.: Ex- dredged from the inner walls of this type uncoupled. ploration results, 1950-58: Internat. Geol. trench, suggest that thrust slices which in- Cong., 21st, Copenhagen, Rept., pt. 18, p. clude crustal and sometimes mantle mate- CONCLUSIONS 21-27. rial are successively tucked under the inner Bailey, E. H., Irwin, W. P., and Jones, D. L., wall (Fig. 7B). A linear ridge, more than 50 1964, Franciscan and related rocks, and Systematic profiles across the inner slopes their significance in the geology of western km long, which divides the Mariana Trench of active trenches demonstrate the wide var- California: California Div. Mines and into a double axial trough near 13.5° N., iation in size and style of accretionary Geology Bull. 183, 177 p. 146.5° E. (Karig, 1971a) and another in the prisms. When these morphologic variations Barazangi, M., Isacks, B., and Oliver, J., 1972, Peru Trench (Kulm and others, 1973a, are viewed together with other evidence, a Propagation of seismic waves through and 1974) may represent such thrust slices. Ir- logical dependence between morphological beneath the lithosphere that descends under regular basaltic bodies, some of which and structural factors and the type and rate the Tonga island arc: Jour. Geophys. Re- show columnar jointing, are imbedded in of material feed is observed. The resulting search, v. 77, p. 952-958. pelagic sediments in one southern Francis- rock mass, which preserves the record of Beck, R. H., 1972, The oceans, the new frontier in exploration: APEA Jour., v. 12, p. 5-29. can locality (Karig, unpub. mapping), sug- subduction, will carry the effects of these Beck, R. H., and Lehner, P., 1974, Oceans, new gesting that large-scale topographic ir- variations, and, unless the correct relation- frontier in exploration: Am. Assoc. Petro- regularities might also be sheared off and ship is known, correlation between con- leum Geologists Bull., v. 58, p. 376-395. accreted (Karig, 1974b). temporary arc systems and deeply eroded Beloussov, V. V., 1970, Against the hypothesis of The zone of uplift due to accretion does orogenic zones will be impossible. ocean-floor spreading: Tectonophysics, v. not extend far behind the trench slope The variations in contemporary arc sys- 9, p. 489-511. break, where a progressively enlarging sedi- tems also supply information bearing on Burk, C. A., 1965, Geology of the Alaska ment body accumulates (Fig. 8). In the ear- problems of continental growth. Rather Peninsula-island arc and continental margin: Geol. Soc. America Mem. 99, 250 p. lier stages of accretion, the upper-slope sedi- surprisingly, over time intervals of up to 50 1973, Uplifted eugeosynclines and continen- ment body enlarges as the accreted mass m.y. or more, the volcanic chain does not tal margins: Geol. Soc. America Mem. 132, builds upward and outward (Figs. 4, 5). migrate seaward, as does the trench. In- p. 75-86. Subsidence of the upper slope is only rela- stead, it often migrates to the rear (Dickin- Carey, S. W., 1970, Australia, New Guinea, tive to the trench slope break. Subsidence son, 1973a) and, together with the and Melanesia in the current revolution in relative to sea level does not seem to occur dominating effects of accretion, results in concepts of the evolution of the earth:

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/86/3/377/3433679/i0016-7606-86-3-377.pdf by guest on 25 September 2021 388 KARIG AND SHARMAN

Search, v. 1, p. 178-189. of underthrusting in southwest Japan: A 1975, Initial reports of the Deep Sea Drilling Chase, R. L., and Bunce, E. T., 1969, Under- model of convergent plate interactions: Project, Vol. 31: Washington, D.C., U.S. thrusting of the eastern margin of the Antil- Jour. Geophys. Research, v. 76, p. Govt. Printing Office (in press). les by the floor of the western North Atlan- 7260-7292. Irving, E. M., 1950, Review of Philippine base- tic Ocean, and origin of the Barbados ridge: Frazer, J. Z., Hawkins, D. L., and Arrhenius, G., ment geology and its problems: Philippine Jour. Geophys. Research, v. 74, p. 1972, Surface sediments and topography of Jour. Sci., v. 79, p. 267-307. 1413-1420. the northern Pacific: California Univ. Inst. 1952, Geologic history and petroleum pos- Coney, P. J., 1972, Cordilleran tectonics and Marine Research Tech. Rept. Ser., TR 30, sibilities of the Philippines: Am. Assoc. Pe- North America plate motion: Am. Jour. 10 maps. troleum Geologists Bull., v. 36, p. 437—476. Sci., v. 272, p. 603-628. Gates, Olcott, and Gibson, William, 1956, In- Isacks, B., and Molnar, P., 1971, Distribution of Creager, J. S., Scholl, D. W., and others, 1973, terpretation of the Aleutian Ridge: Geol. stresses in the descending lithosphere from Initial reports of the Deep Sea Drilling Soc. America Bull., v. 67, p. 127-146. a global survey of local-mechanism solu- Project, Vol. 19: Washington, D.C., U.S. Gilluly, James, 1972, Tectonics involved in the tions of mantle earthquakes: Rev. Govt. Printing Office, 913 p. evolution of mountain ranges, in Geophysics and Space Physics, v. 9, p. Curray, J. R., and Moore, D. G., 1971, Growth Robertson, E. C., ed., Nature of the solid 103-125. of the Bengal deep-sea fan and denudation earth: New York, McGraw-Hill, p. Isacks, B., Oliver, J., and Sykes, L., 1968, Seis- in the Himalayas: Geol. Soc. America Bull., 406-439. mology and the new global tectonics: Jour. v. 82, p. 563-572. Grow, J. A., 1973a, Crustal and upper mantle Geophys. Research, v. 73, p. 5855-5900. Den, N., Murauchi, S., Hotta, H., Asunama, T., structure of the central Aleutian arc: Geol. Johnson, T., and Molnar, P., 1972, Focal and Hagiwara, K., 1968, A seismic refrac- Soc. America Bull., v. 84, p. 2169-2192. mechanisms and plate tectonics of the tion exploration of Tosa deep-sea terrace 1973b, Implications of deep sea drilling, southwestern Pacific: Jour. Geophys. Re- off Sikoku: Jour. Physics Earth, v. 16, p. Sites 186 and 187, on island arc structure, search, v. 77, p. 5000-5032. 7-10. in Creager, J. S., Scholl, D. W., and others, Karig, D. E., 1970, Ridges and basins of the Denham, D., 1969, Distribution of earthquakes Initial reports of the Deep Sea Drilling Pro- Tonga-Kermadec island arc system: Jour. in the New Guinea-Solomon islands re- ject, Vol. 19: Washington, D.C., U.S. Govt. Geophys. Research, v. 75, p. 239-255. gion: Jour. Geophys. Research, v. 74, p. Printing Office. ——1971a, Structural history of the Mariana is- 4290-4299. Hamilton, R. M., and Gale, A. W., 1968, Seis- land arc system: Geol. Soc. America Bull., Dewey, J. F., and Bird, J. M., 1970, Mountain micity and structure of North Island, New v. 82, p. 323-344. belts and the new global tectonics: Jour. Zealand: Jour. Geophys. Research, v. 73, p. 1971b, Origin and development of marginal Geophys. Research, v. 75, p. 2625-2647. 3859-3876. basins in the western Pacific: Jour. Dickinson, W. R., 1971, Clastic sedimentary se- Hamilton, Warren, 1969, Mesozoic California Geophys. Research, v. 76, p. 2542-2561. quences deposited in shelf, slope, and and the underflow of the Pacific mantle: 1973a, Comparison of island arc - marginal trough settings between magmatic arcs and Geol. Soc. America Bull., v. 80, p. basin complexes in the northwest and associated trenches: Pacific Geology, v. 3, p. 2409-2430. southwest Pacific, in Coleman, P. J., ed., 15-30. 1973, Tectonic map of Indonesia — A prog- The western Pacific: Island arcs, marginal 1973a, Widths of modern arc-trench gaps ress report: U.S. Geol. Survey Open-File seas, geochemistry: Nedlands, Univ. West- proportional to past duration of igneous ac- Rept., 20 p. ern Australia Press, p. 355-364. tivity in associated magmatic arcs: Jour. Hawkins, J. W., Fisher, R. L., and Engel, C. G., 1973b, Plate convergence between the Geophys. Research, v. 78, p. 3376-3389. 1972, Ultramafic and mafic rock suites ex- Philippines and the Ryukyu Islands: Marine 1973b, Reconstruction of past arc-trench posed on the deep flanks of Tonga trench: Geology, v. 14, p. 153-168. systems from petrotectonic assemblages of Geol. Soc. America, Abs. with Programs 1974a, Tectonic erosion at trenches: Earth the western Pacific; in Coleman, P. J., ed., (Cordilleran Sec.), v. 4, no. 3, p. 167. and Planetary Sci. Letters, v. 21, p. The western Pacific island arcs, marginal Hayes, D. E., and Ewing, M., 1970, Pacific 209-212. seas, geochemistry: Nedlands, Univ. West- boundary structure, in Maxwell, A. E., ed., 1974b, Evolution of arc systems in the western Australia Press, p. 569—601. The sea, Vol. 4, Pt. 2: New York, John ern Pacific: Ann. Rev. Earth and Planetary Dott, R. H., Jr., 1969, Circum-Pacific late Wiley & Sons, Inc., p. 29-72. Sciences, v. 2, p. 51-76. Cenozoic structural rejuvenation: Implica- Hayes, D. E., Talwani, M., Houtz, R., and Pit- 1975, Origin of the Philippine Sea basin tions for sea-floor spreading: Science, v. man, W. C., 1970, Preliminary report of complex, in Ingle, J. C., Jr., Karig, D. E., 166, p. 874-879. Volume 22, U.S.N.S. Elianin, Cruises and others, Initial reports of the Deep Sea Dubois, Jacques, 1971, Propagation of P waves 28-32, pt. C; Seismic reflection profiles, p. Drilling Project, Vol. 31: Washington, and Rayleigh waves in Melanesia: Struc- 174-200. Lamont-Doherty Geol. Observ., D.C., U.S. Govt. Printing Office (in press). tural implications: Jour. Geophys. Re- Tech. Rept. no. CU-1-72. Karig, D. E., and Mammerickx, J., 1972, Tec- search, v. 76, p. 7217-7240. Hess, H. H., 1948, Major structural features of tonic framework of New Hebrides island Fedotov, S. A., 1968, On deep structure, proper- the western north Pacific, and interpreta- arc system: Marine Geology, v. 12, p. ties of the upper mantle, and volcanism of tion of H.O. 5985, Bathymetric chart, 187-205. the Kuril-Kamchatka island arc according Korea to New Guinea: Geol. Soc. America Karig, D. E., and Wageman, J. H., 1975, Struc- to seismic data, in Crust and upper mantle Bull., v. 59, p. 417-446. ture and sediment distribution in the of the Pacific area: Am. Geophys. Union Hilde, T.W.C., Wageman, J. H., and Hammond, northwestern corner of the West Philippine Geophys. Mem. 12, p. 131-139. W. T., 1969, The structure of Tosa terrace Basin, in Ingle, J. C., Jr., Karig, D. E., and Fischer, A. G., Heezen, B. C., and others, 1971, and Nankai trough off southeastern Japan: others, Initial reports of the Deep Sea Initial reports of the Deep Sea Drilling Deep-Sea Research, v. 16, p. 67-76. Drilling Project, Vol. 31: Washington, Project, Vol. 6: Washinton, D.C., U.S. Holmes, M. L., von Huene, R., and McManus, D.C., U.S. Govt. Printing Office (in press). Govt. Printing Office, 1329 p. D. A., 1972, Seismic reflection evidence Katsumata, Mamoru, and Sykes, L. R., 1969, Fisher, R. L., and Engel, C. G., 1969, Ultramafic supporting underthrusting beneath the Seismicity and tectonics of the western and basaltic rocks dredged from the near- Aleutian Arc near Amchitka Island: Jour. Pacific: Izu-Mariana Caroline and shore flank of the Tonga Trench: Geol. Soc. Geophys. Research, v. 77, p. 959-964. Ryukyu-Taiwan regions: Jour. Geophys. America Bull., v. 80, p. 1373-1378. Hoshino, M., 1969, On the deep sea terrace: La Research, v. 74, p. 5923-5948. Fisher, R. L., and Hess, H. H., 1963, Trenches, in Mer, v. 3, p. 222-224. Kawano, Y., and Uyeda, Y., 1967, Periods of the Hill, M. N., ed., The sea, Vol. 3: New York, Hsu, K. J., 1969, Preliminary report and geologic igneous activities of the granite rocks in Interscience, p. 411-436. guide to Franciscan melanges of the Morro Japan by K-Ar dating method: Tec- Fitch, T. J., 1972, Plate convergence, transcur- Bay-San Simeon area, California: Califor- tonophysics, v. 4, p. 523-530. rent faults and internal deformation adja- nia Div. Mines and Geology Spec. Pub. 35, Kulm, L. D., Scheidegger, K. F., Prince, R. A., cent to southeast Asia and the western 46 p. Dymond, J., Moore, T. C., and Hussong, Pacific: Jour. Geophys. Research, v. 77, p. Ingle, J. C., Jr., Karig, D. E., and others, 1973, D. M., 1973a, Tholeiitic basalt ridge in the 4432-4460. Leg 31 — Western Pacific floor: Geotimes, Peru Trench: Geology, v. 1, p. 11—14. Fitch, T. J., and Scholz, C. H., 1971, Mechanism v. 18, p. 22-25. Kulm, L. D., von Huene, R., and others, 1973b,

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/86/3/377/3433679/i0016-7606-86-3-377.pdf by guest on 25 September 2021 SUBDUCTION AND ACCRETION IN TRENCHES 389

Initial reports of the Deep Sea Drilling propagation in the upper 250 km of the the Barbados Ridge: Jour. Geophys. Re- Project, Vol. 18: Washington, D.C., U.S. Tonga island arc: Seismol. Soc. America search, v. 75, p. 488-490. Govt. Printing Office, 1077 p. Bull., v. 59, p. 1115-1137. Scholl, D. W., von Huene, R., and Ridlon, J. B., Kulm, L. D., Resig, J. M., Moore, T. C., Jr., Moore, J. G, 1973, Complex deformation of 1968, Undeformed sediments in the Peru- Rosato, V. J., 1074, Transfer of Nazca Cretaceous trench deposits, southwestern Chile trench: Science, v. 159, p. 869-871. Ridge pelagic sediments to the Peru conti- Alaska: Geol. Soc. America Bull., v. 84, p. Scholl, D. W., Christiansen, M. N., von Huene, nental margin: Geol. Soc. America Bull., v. 2005-2020. R., and Marlow, M. S., 1970, Peru-Chile 85, p. 769-780. Murauchi, S., Den, N., Asano, S., Hotta, H., trench sediment and sea-floor spreading: Lahr, J. C., and Page, R. A., 1972, Hypocentral Yoshii, T., Asanuma, T., Hagiwara, K., Geol. Soc. America Bull., v. 81, p. locations in the Cook inlet region of Ichikawa, K., Solo, T., Ludwig, W. J., 1339-1360. Alaska: EOS (Am. Geophys. Union Trans.), Edgar, N. T., and Houtz, R. E., 1968, Silver, E. A., 1972, Pleistocene tectonic accretion v. 53, p. 1042. Crustal structure of the Philippine Sea: of the continental slope off Washington: Larsen, R. L., and Chase, C. G., 1972, Late Jour. Geophys. Research, v. 73, p. Marine Geology, v. 13, p. 239-249. Mesozoic evolution of the western Pacific 3143-3171. Steinen, R. P., Harrison, R. S., and Matthews, Ocean: Geol. Soc. America Bull., v. 83, p. Murdock, J. N., 1969, Crust-mantle system in R. K., 1973, Eustatic low stand of sea level 3627-3644. the central Aleutian region: Seismol. Soc. between 125,000 and 105,000 years B.P. Le Pichon, X., Francheteau, J., and Bonnin, J., America Bull., v. 59, p. 1543-1558. Evidence from the subsurface of Barbados, 1973, Plate tectonics: Amsterdam, Elsevier, Peck, D. L., Griggs, A. B., Schlicker, H. G., West Indies: Geol. Soc. America Bull., v. 300 p. Wells, F. G., and Dole, H. M., 1964, Geol- 84, p. 63-70. Lohman, G. P., 1974, Paleo-oceanography of the ogy of the central and northern parts of the Stewart, R. J., 1970, Petrology, metamorphism Oceanic Formation (Eocene-Oligocene), western Cascade Range in Oregon: U.S. and structural relations of graywackes in Barbados, West Indies [Ph.D. dissert.]: Geol. Survey Prof. Paper 449, 56 p. the western Olympic Peninsula, Washing- Providence, R.I., Brown Univ. Petelin, V. P., 1964, Hard rocks in the deep- ton [Ph.D. thesis]: Stanford, Calif., Stan- Ludwig, W. J., 1970, The Manila Trench and water trenches of the southwestern part of ford Univ., 107 p. West Luzon Trough—III: Seismic refraction the Pacific ocean: Internat. Geol. Cong., Sykes, L. R., 1966, The seismicity and deep struc- measurement: Deep-Sea Research, v. 17, p. 22d, Doklady by the Soviet Geologists, p. ture of island arcs: Jour. Geophys. Re- 553-571. 78-86. search, v. 71, p. 2781-3006. Ludwig, W. J., Ewing, J. I., Murauchi, S., Den, Piper, D.J.W., von Heune, R., and Duncan, J. R., Tabor, R., Cady, W. D., and Yeats, R. S., 1970, N., Asano, S., Hotta, H., Hayakawa, M., 1973, Late Quaternary sedimentation in the Broken formations and thrust faulting in Asanuma, T., Ichikawa, K., and Nagurchi, active eastern Aleutian trench: Geology, v. the northeastern Olympic mountains, I., 1966, Sediments and structure of the 1, p. 19-22. Washington: Geol. Soc. America, Abs. with Japan Trench: Jour. Geophys. Research, v. Plafker, G., 1969, Tectonics of the March 1964 Programs (Cordilleran Sec.), v. 2, no. 2, p. 71, p. 2121-2137. Alaska earthquake: U.S. Geol. Survey Prof. 152. Ludwig, W. J., Hayes, D. E., and Ewing, J. I., Paper 543-1, 74 p. Tayama, R., 1950, The submarine configuration 1967, The Manila Trench and West Luzon 1972, Alaskan earthquake of 1964 and off Shikoku, especially the continental slope Trough—I: Bathymetry and sediment dis- Chilean earthquake of 1960: Implications [in Japanese]: Japan Hydrog. Office Bull., tribution: Deep-Sea Research, v. 14, p. for arc tectonics: Jour. Geophys. Research, Spec. Pub., v. 7, p. 54-82. 533-544. v. 77, p. 901-925. Umbgrove, J.H.F., 1947 (1949), The pulse of the Martin-Kaye, P.H.A., 1969, A summary of the Raitt, R. W., 1967, Marine seismic studies of the earth: The Hague, Martinus Nijhoff, 358 p. geology of the Lesser Antilles: Overseas Indonesian island arc: Am. Geophys. Union Van Bemmelen, R. W., 1949, The geology of In- Geology and Mineral Resources, v. 10, p. Trans, [abs.], v. 48, p. 217. donesia, Vol. 14: The Hague, Govt. Print- 172-206. Riedel, W. R., Ladd, H. S., Tracey, J. I., Jr., and ing Office, 732 p. Matsuda, T., and Uyeda, S., 1971, On the Pacific Bramlette, M. N., 1961, Preliminary drill- Vening-Meinesz, F. A., 1964, The earth's crust type orogeny and its model-extension of the ing phase of Mohole Project, 2: Summary and mantle: Amsterdam, Elsevier Pub. Co., paired belts concept and possible origin of of coring operations: Am. Assoc. Petroleum 124 p. marginal seas: Tectonophysics, v. 11, p. Geologists Bull., v. 45, p. 1793-1798. von Huene, R., 1972, Structure of the continen- 5-27. Ross, D. A., 1971a, Sediments of the northern tal margin and tectonism at the eastern Melendres, M. M., and Comsti, F„ 1951, Re- Middle America trench: Geol. Soc. America Aleutian trench: Geol. Soc. America Bull., connaissance geology of southeastern Bull., v. 82, p. 303-322. v. 83, p. 3613-3626. Davao: Philippine Geologist, v. 5, no. 2, p. 1971b, Mass physical properties and slope von Huene, R., and Shor, G. G., Jr., 1969, The 38-46. stability of the sediments of the northern structure and tectonic history of the eastern Menard, H. W., 1964, Marine geology of the Middle America Trench: Jour. Geophys. Aleutian Trench: Geol. Soc. America Bull., Pacific: New York, McGraw-Hill Book Research, v. 76, p. 704-712. v. 80, p. 1889-1902. Co., 271 p. Ross, D. A., and Shor, G. G., Jr., 1965, von Huene, R., Lathram, E. H., and Reimnitz, E., Mesolella, K. H., Sealy, H. A., and Matthews, Reflection profiles across the Middle 1971, Possible petroleum resources of R. K., 1970, Facies geometries within Pleis- . America Trench: Jour. Geophys. Research, offshore Pacific margin Tertiary basin, tocene reefs of Barbados, West Indies: Am. v. 70, p. 5551-5572. Alaska: Am. Assoc. Petroleum Geologists Assoc. Petroleum Geologists Bull., v. 54, p. Scholl, D. W., and Marlow, M. S., 1972, The Mem. 15, p. 136-151. 1899-1917. sedimentary sequence in modern Pacific Watts, A. B., and Talwani, M., 1974, Gravity Meyerhoff, A. A., and Meyerhoff, H. A., 1972, trenches, and the apparent rarity of similar anomalies seaward of deep-sea trenches and The new global tectonics: Major inconsis- sequences in deformed circumpacific their tectonic implications: Royal Astron. tencies: Am. Assoc. Petroleum Geologists eugeosynclines, in Modern and ancient Soc. Geophys. Jour., v. 36, p. 57-90. Bull, v. 56, p. 269-336. geosynclinal sedimentation: Proc. of a con- Worzel, J. L., and Shurbet, G. L., 1955, Gravity Miller, D. J., 1953, Late Cenozoic marine glacial ference held at Wisconsin Univ., Nov. interpretations from standard oceanic and sediments and marine terraces of Middleton 10-11, 1972, R. H. Dott, Jr., convener, p. continental crustal sections, in Crust of the Island, Alaska: Jour. Geology, v. 61, p. 36-37. Earth: Geol. Soc. America Spec. Paper 62, 17-40. 1974, Sedimentary sequence in modern - p. 87-100. Mitchell, A.H.G., 1969, Raised reef-capped ter- Pacific trenches and the deformed Pacific Yagi, K., 1960, A dolerite block dredged from races and Plio-Pleistocene sea level changes, eugeosyncline: Soc. Econ. Paleontologists the bottom of the Vityaz Deep, Mariana north Malekula, New Hebrides: Jour. and Mineralogists Spec. Paper no. 19, p. trench: Japan Acad. Proc., v. 36, p. Geology, v. 77, p. 56-67. 193-211. 213-216. Mitchell, A.H.G., and Bell, J. D., 1970, Volcanic Scholl, D. W., and von Huene, R., 1970, Com- MANUSCRIPT RECEIVED BY THE SOCIETY episodes in island arcs: Geol. Soc. London ments on paper by R. L. Chase and E. T. DECEMBER 26, 1973 Proc., no. 1662, p. 9-12. Bunce: Underthrusting of the eastern mar- REVISED MANUSCRIPT RECEIVED JUNE 17, 1974 Mitronovas, W., Isacks, B., and Seeber, L., 1969, gin of the Antilles by the floor of the west- Earthquake locations and seismic wave ern North Atlantic Ocean, and origin of Printed in U.S.A.

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/86/3/377/3433679/i0016-7606-86-3-377.pdf by guest on 25 September 2021