Back-Arc Opening and the Mode of Subduction

VOL.84, NO. B3 JOURNALOF GEOPHYSICALRESEARCH MARCH10, 1979

Back-ArcOpening and the Mode of Subduction

SEIYA UYEDA

EarthquakeResearch Institute, University of Tokyo,Tokyo, Japan

HIROO K ANAMORI

SeismologicalLaboratory, California Institute of Technology,Pasadena, California 91125

Trench-arcsystems (subduction zones) can be classified into two types depending onwhether or not activelyopening back-arc basins are associated with them. This suggests that subduction of an oceanic plateis not a sufficientcondition for back-arc opening, though it may be necessary one. Mechanisms that causethe distinction between the two types have been investigated. Earthquake studies suggest that there isa significantdifference inthe mode of plate motion at interplate boundaries between the two types of trench-arcsystems. Extreme cases are Chile, where plate motion is seismic, and the M arianas arc, where it is aseismic.This difference seems to indicatethat the stress state in theback-arc area differs between the twotypes: compression in the Chilean type and tension in theMarianas type. This difference in thestress stateis alsomanifested in other tectonic features, such as topography, gravity, volcanic activity, and crustalmovement. Two possible mechanisms forthe difference between the two types are suggested: (1) Thenature of thecontact zone between upper and lower plates chahges from tight coupling (Chile) to decoupling(the Marianas) through the evolutionary process ofsubduction. The decoupling results inan oceanwardretreat of the trench and back-arc opening. (2) The downgoing slab is anchored tothe mantle, sothat the position of a trenchis also fixed with respect to themantle. Since the motion in themantle is slowcompared tothat of surface plates, it is the motion of the landward plate which controls the opening and nonopeningof back-arcs.

INTRODUCTION closelycorrelated with eachother because, from geometrical In platetectonics, the fundamental process occurring under considerations,it is clear that a subductionboundary can thetrench-arc systems is the subduction of an oceanicplate. easilychange toga 'leaky' transform boundary upon a collision Smalloceanic basins, such as the Japan Sea, Okhotsk Sea, and of a spreadingcenter with the trench.In fact, the openingof Philippine Sea are often found landward of arcs.The intent of the Gulf of Californiamay be an exampleof the combined the presentpaper is to discussthe genesisof thesebasins, effectsof 3 and 4. Until 20 m.y.B.P. the East PacificRise was calledback-arc basins or marginalseas. situatedto the westof Baja California[Atwater, 1970] while Wefirst note the fact that not all thetrench-arc systems have now it is in the gulf in the form of a seriesof ridgesand back-arc basins.The Peruvian and Chilean arcs do not have transforms.The ridgepossibly died duringits collisionwith back-arc basins.The processof back-arcbasin formation the trenchthat existedto the westof Bajaand reappearedin appearsto be inactiveat presentin somearcs as in the caseof theformer back-arc region by jumping rather than migrating theJapan Sea, whereas others are believed to havean actively underthe Baja. The AndamanSea may be anotherexample openingback-arc basin, for example,the M arianaTrough [Eguchiet al., 1979].Possibility 5 seemsto haveno petrolog- [Karig, 1971a].Any theoryfor the genesisof back-arcbasins ical support. must accountfor this fact. Clearly, the subductionof an oce- Possibility2 is our mainconcern in this paper.As we men- anic plate is not a sufficientcondition for the formation of tioned already, not all the subductionboundaries have back- back-arcbasins, although it maybe a necessaryone. arc basins, and not all subduction boundaries that have back- Severalpossibilities have been suggested to explainthe ori- arc basinsare presentlyactively opening. We will firstclassify gin of back-arcbasins: (1) entrapmentof a marginalpart of the subductionboundaries by back-arctype and then demon- preexistingocean by the formation of an island arc: Aleutian stratethat the differencesin back-arctype correlate with other Basin[Cooper et al., 1976]and west Philippine Sea [Uyeda and phenomenawhich are presumedto be relatedto the tectonic Ben-Avraham,1972], (2) back-arcspreading caused by or re- stressstate. We will next dicussthe possibilitythat the differ- latedto subduction:Lau Basinand M arianaTrough [Karig, encein the tectonicstress state in back-arcregions may be 1970,197 la], ScotiaSea [Barker, 1970], and Japan Sea [Karig, causedby a differencein themode of subductionand attempt 1971b;Matsuda and Uyeda, 1971], (3) openingrelated to to find out why such different modes of subductionoccur. 'leaky'•transform fault: Gulf of California[Wilson, 1965] and CLASSIFICATIONOF SUBDUCTIONBOUNDARIES AndamanSea [Curray et al., 1979],(4) openingrelated to the subductionof a ridge:Japan Sea [Uyeda and Miyashiro, 1974], Karig[1971b] classified marginal basins into active marginal and (5) subsidencecaused by oceanizationof continental basins(type 1), inactivemarginal basins with highheat flow crust:Japan and Okhotsk seas [Beloussov and Ruditch, 1961 ]. (type 2), and inactivemarginal basinswith normal heat flow Among the abovepossibilities, 1 is probablythe likeliest (type 3). He postulatedthat types2 and 3 wereactive in the mechanismprovided that theage of thebasin is olderthan that past,type 2 in the morerecent past than type3. We useideas of the arc. The Mesozoicsequence of magneticlineations in more or less similar to Karig's to classifythe subduction the AleutianBasin is a notableexample provided that the boundariesin Table1. First,subduction boundaries (trench- identifications.are correct. Possibilities3 and 4 seem to be arc-back-arcsystems) are groupedinto the continentalarcs and islandarcs. Continental arcs, by definition,have no back- Copyright¸ 1979by theAmerican Geophysical Union. arc basins.The back-arc area of the Sumatra and Java arc is Paper number 8B1006. 1049 0148-0227/79/008 B- 1006501.00 1050 UYEDAAND KANAMORI: BACK-ARc OPENING AND $UBDUCTION

TABLE 1. Classification of Trench-Arc-Back-Arc Systems

Classification SubductionBoundary Location

Continental ,4 rc Continents Peru-Chile Andes Middle America Middle America (semi-active?) Proto-North America Basin and Range Province(semi-active?) Alaska Alaska Range Sumatra-Java Sunda Shelf

Island Arc Back-arc active Back-arcspreading Mariana Mariana Trough Ryukyu Ryukyu Trough Tonga Lau Basin Scotia Scotia Sea New Hebrides 9 Bonin 9 Philippine 9 Leaky transform Proto-BajaCalifornia Gulf of California Andaman-Nicobar Andaman Sea Melanesian Borderlands Bismarck Sea, New Guinea New Hebrides Fiji plateau? Back-arc inactive Inactivated Antilles Grenada Trough? West Aleutian Kamchatka Basin Kurile Kurile Basin Japan JapanBasin, Tsushima Basin Proto-Izu-Bonin Shikoku Basin Proto-Mariana Parece Vela Basin Trapped Aleutian Bering Sea Proto-Mariana West PhilippineSea? Proto-Tonga South Fiji Basin? Middle America Caribbean Plate

water covered,but this arc is groupedwith the continentalarcs posed to have existedbefore the collisionof the East Pacific because most of the back-arc area has only shallow water Ri•e with the North American continent [McKenzie and Mor- depth (about 100 m) and presumablya continentalcrust. gan, 1969;Scholz et al., 1971]. Island arcsare classifiedinto Conversely,the Middle Americanarc, thoughincluded in the those with active back-arc basins and those with inactive back- continentalarcs, might be better groupedwith the back-arc arc basins.Active and inactiverefer to whetherthe openingof spreadingtype becauseof the extensionalvolcanic grabens the basin is presentlyin progressor not. A back-arc basin found in Middle America [Plafker, 1976].A similar argument which has rift topographyand/or fault structuresuggestive of may hold for the Proto-North American arc, which is sup- tensional tectonicsis consideredto be actively opening if it showsone or more of the following characteristics:(1) thin

+90' +120' +150' -+180' -150' -120' -90* -60' -30* sedimentcover of very young age over rugged volcanic base- ment relief, (2) shallow water depth comparable to that of

+60 +60 active midoceanicridges, (3) high heat flow or highly variable heat flow [Watanabe et al., 1977], or (4) magneticlineations with zero age.

• k •o C•-o1(14) The actual classificationin Table 1 was made using the referencesindicated in the legendof Figure 1. Figure 1 summa- 0 rizesthis information regardingthe time and rate of opening. (19 ' '' Typical examplesof the actively spreadingback-arc basins are the M ariana Trough and the Lau Basin.Among thosethat are supposedto be actively spreading,magnetic lineations of very young age have beendocumented for the Scotia Sea. On the whole, the magneticlineations in the back-arc basinsare I I • • { I { I I I [ I I I I { [ I I I • ] I { I I much less well developedthan those in oceans [Watts and +90* +120' +150' -+180' -150* -120* -90* -60' -30' Weissel,1975; Weisseland Watts, 1975;Kobayashi and lsezaki, Fig. 1. Summary of back-arc openingactivities. Reference' num- 1977; Lawyer and Hawkins, 1978], suggestingthat either the betsare as follows: 1, Barkerand Griffiths [1972]; 2, Malfaitand modeof spreadingin the two cases is different or theback-arc Dinkelman [1972]; 3, Plafker [1976]; 4, Larson [1972]; 5, Scholzet al. [1971]; 6, Christiansenand Lipman [1972], Smith [1977]; 7, Cormier spreading correspondsto the very early stage of oceanic [1975]; 8, Beloussoo[1968]; 9, HYde and Wageman[19731;-10, Uyeda spreading. and Miyashiro [1974]; l 1, Tomodaet al. [ 1975]; 12, Watts and Weissel The Fiji Plateau,behind the New Hebridesarc, satisfiesthe [1975]; 13, Herman et al. [1979]; 14, Karig [1971a]; 15, Heezen et al. abovecriteria for an active back-arcbasin, but there the polar- [ 1971]; 16,Louden [1976]; 17,Shih [1978]; 18, Ben-,4orahamandUyeda ity of the subductionis anomalous;polarity reversaloccurred [1973]; 19, Curray et a/,0[1979]; 20, Connelly [1976]; 21, Karig and Mammerickx [1972]; 22, Chase [1971]; 23, Karig [1970]; 24, Karig in the late Miocene (10 m.y.B.P.), when the Ontong Java [1971b]; and 25, Weisseland Watts [1975]. Plateau collided with the now extinct trench [Chase, 1971; UYEDA AND KANAMORI: BACK-ARC OPENING AND SUBDUCTION 1051

+90o +120ø +150ø _+180ø -150o -120o -900 -600 -30 o Typical of the 'inactivated'back-arc basins are the Japan, Kurile, Shikoku, and Parece Vela Basins. Those basins are supposedto havebeen actively spreading in the recentpast but do not appear to be active any more. It is important to note that some of these inactivated back-arc basins are still associated with active subductionboundaries. This appliesalso to the Bering Sea.

STRESS STATE IN BACK-ARC PLATE It seemsnatural to conjecturethat the tectonicstress in the upper plate at a subduction boundary may be dominantly compressionalbecause two plates are converging at the boundary. In fact, earthquake source mechanisms,geodetic +a0• +1•0• +1•0• •180• -I•0 • -IZ0• -a0 • -•0 • -30• observations,and faults and folds [Sugimuraand Uyeda, 1973] Fig. 2. Stressstate estimated from source mechanism ofintraplate suggestthat the stressesin the Japaneseislands are generally earthquakesin back-arcregions; mo is magnitude.Reference numbers compressionalin the direction of the plate convergence.It are as follows:l, Slauder[1973]; 2, Stauder[197:]; 3, Molnar and seems,however, that the upper layer of the upper plate must Sykes [1969]: 4, Malumolo and Lal•am [1976]: :, Sykes [1968]; 6, be under a tensional stressregime for the back-arc basin to •cholz et al. [1971];7, Gedney[1970]; 8, •y•es and •bar [1974];9, Na•amura et al. [1977]; 10, Cormier[197:]; 11, •eith [1974];12, spread.Thus, all the models so far proposedto explain the Sla•der and M•alcMn [1976]; 1•, Honda el al. [1967]; 14, IcM•awa spreadingin back-arc basinswere devisedto generatea ten- [1969]; 1:, •ukao and •ummoto [197:]; 1:, •veryanova [1973]: 17, sional stressfield in the upper part of the upper plate at a Katsumataand •ykes [1969]; 18, Fitc• [1972]; 19, Fitc• [1970]: 20, converging plate boundary [Karig, 1971b; McKenzie, 1969; •gucM el al. [1979];21, Chungand Kanamori[1978]; and 22, HawMns Hasebeet al., 1970;Sleep and Toksoz, 1971;Sleep, 1975]. In this paper, we concentrateon the fact that some subduction boundaries have an active back-arc and others do not and attempt to find a clue to the problem of the origin of back-arc Karig and Mammerickx,1972]. Conflciting ideas have been spreadingby clarifying the differencesin these two types of proposedfor the activeback-arc spreading in the beltimmedi- subductionboundaries. For this purpose,we first investigate ately behindthe New Hebridesarc [Karigand Mammerickx, whether thesetwo typesof back-arcsare indeed under differ- 1972; Luyendyket al., 1974]. ent stressconditions using published source mechanism solu- The entirelength from Fiji to New Guinea,called the Mela- tions of shallowearthquakes occurring in the upper plates. nesia Borderland, forms the complex megashearboundary At the earlier stage of plate tectonics,mainly interplate betweenthe two large Pacificand India-Australiaplates. The earthquakes were intensively investigated to determine the BismarckSea is said to be spreadingIConnelly, 1976] but is relative plate motions. More recently, the mechanismsof in- regardedas atypical[Coleman and Packham, 1976], due to the traplate earthquakeshave beenstudied to determinethe stress reverselypolarized Solomon Trench. The Gulf of California statesin the plates, which bear directly on the driving mecha- and A ndaman Sea are also located on megashearboundaries nism of plate motions [Mendiguren,1971; Forsyth, 1973;Sykes betweentwo largeplates and are activelyopening by the leaky and Sbar, 1973, 1974; Richardsonet al., 1976]. transform fault mode. In both cases,some magnetic lineations Unfortunately, few mechanismsolutions are available for with youngages have beenreported [Larson, 1972; Curray et use in this paper becauseintraplate earthquakes(especially al., 1979].Common to both casesare the collisionof a ridge with magnitude large enough to be recordedat a sufficiently and a trenchin the recentgeologic time (Miocene) [Atwater, large number of stations)are very much fewer than those at 1970;Eguchi et al., 1979].The Gulf of Californiais not a back- plate boundaries,and initial interestin the intraplatestress has arc at present,but it wasso in the pastuntil the time of the ridge-trenchcollision. The Basinand Range Province may be similarin someways [Thompsonand Burke, 1974;Scholz et al., 1971;Smith, 1977]. Thus, it now seemsthat active back-arc +90o +1200 +150ø _+1800 -1500 -1200 -90o -600 -300 • - basinsmay be classified as a subduction-relatedtype or a leaky transform type. +60'_ +600 The Middle American arc is classifiedas having a trapped back-arcalso because it hasa largearea of back-arcbasin, the +40• +400

Caribbean.The presentCaribbean Sea in the modelof Malfait +20o and Dinkelman[1972] was a part of the PacificOcean during the Late Cretaceousto early Oligocene,just as the present 0o _ ,5. 00 ScotiaSea is, and wastrapped by the emergenceof the Middle -20 o

Americaarc in the earlyOligocene time. Portions of the Carib- -40 ø -400 bean Sea may insteadbe regardedas a back-arcbasin of the _

Antilles arc. Probably, the presentGrenada Basin was pro- _ -60 o Conver•7ence )- -60o ducedby spreadingbehind the Antillesarc in the past, but rate at ,sabduct/on zones I I ] I I I I I I I I ] I I I I I I I I [ I I I I I I whetherit is activelyspreading at presenthas not beenascer- +90ø +120o +150ø ._+180o -150ø -1200 -90 o -600 -30o tained. Thus, the Antilles arc is classifiedas having an inacti- Fig. 3. Convergencerates at subductionboundaries in centime- rated back-arc.The West PhilippineBasin [Uyedaand Ben- ters per year (after AM I of Minster et al. [1974]and Fitch [1972]). Avraham, 1972]and the South Fiji Basin[Hilde et al., 1977] Arrows indicatethe componentof the absolutevelocity vector of each may alsobe the trappedtype, but thisis highlyconjectural. plate in the directionof plate convergence. 1052 UYEDA AND KANAMORI.'BACK-ARC OPENINGAND SUBDUCTION beenconcentrated on the central portionsof platesrather than TABLE 2. Ten Greatest Earthquakes During 1904-1976 on peripheral areas. Earthquake Year M8 Mw Wo A compilationof the publishedresults of the few available sourcemechanism solutions is shownin Figure 2. In the Chil- Chile 1960 8.3 9.5 1.1 X 10 ø'6 ean and Peruvianarcs, the mechanismsof upperplate shallow Alaska 1964 8.4 9.2 4.0 X 10 ø'5 earthquakesobtained by Stauder [1973, 1975] are mostly Aleutian Islands 1957 8.25 9.1 2.8 X 10 ø'5 Kamchatka 1952 8.25 9.0 2.0 X 1025 thrust and strike slip, but the directionof the compressional Ecuador 1906 8.6 8.8 1.0 X 10 ø'5 stressin both casesis perpendicularto the strike of the arc Aleutian Islands 1965 7.75 8.7 7.1 X 10 ø'4 (Figure 2). The situationis the samefor the Kurile and Japan Assam 1950 8.6 8.6 5.0 X 10 ø'4 arcs, which have inactivated back-arcs.On the other hand, Kurile Islands 1963 8.1 8.5 3.5 X 1024 upperplate shallowearthquakes in the typicallyactive back- Chile 1922 8.3 8.5 3.5 X 10 ø'4 Banda Sea 1938 8.2 8.5 3.5 X 10 ø'4 arc basins,such as Mariana Trough, Lau Basin,and Andaman Sea, and in similar types of regions, such as the Gulf of California, the Basin and Range Province,and Middle Amer- ica, show normal faulting. Here again, there are strike slip suitable as the meaningfulscale of the size of earthquakes. earthquakesin theseregions, but the directionsof compression This happenswhen the size of the rupture surfacebecomes and tension are consistent with those inferred from the normal comparableor largerthan the wavelengthof the seismicwave fault events.Immediately behind the New Hebridesarc, the usedfor M determination.For sucha great earthquake,since one available solution is a reversefault [Chungand Kanamori, moreenergy is releasedby lower-frequencywaves, the 'energy' 1978] and arguesagainst the active openingof this area. In meaningfulto tectonicsmay better be estimatedfrom static view of generalevidence for extensionalfeatures in the back- sourceparameters such as seismicmoment. Kanamori [1977a] arc region[Dugas et al., 1977],the New Hebridesarc is an used this approach for the world's great earthquakessince enigmafor us at this moment.The stressarrows immediately 1904.The seismic moment (M0 = #/5S,where # isrigidity,/5 is behind the Aleutians and in the middle of the Bering Sea are averagedisplacement on the fault, and S is fault area) was estimates,based on the distributionof flank eruptions,which evaluatedfor each great earthquake from long-periodbody can be consideredas natural hydrofractures[Nakarnura et al., wave, surfacewave, free oscillation,and geodeticdata, and the 1977].The stressis compressionalin the immediateback-arc, energyrelease W0 was obtainedby asexpected for an inactiveback-arc, but that in the BeringSea W0 = (Aa/2#) M0 is tensional.A tensionalearthquake mechanism was reported in the westerntip of Alaska [Sykesand Sbar, 1974].This may where /xa is the stressdrop. W0 is consideredto be a good mean that the compressivestress in the volcanicarc is not estimateof the energyreleased in seismicwaves. Using W0,a transmitted several hundred kilometers inward, or there is an new magnitudeMw, which is meaningfuleven in the range intrinsic mechanismthat generatesa tensional tectonic stress M • 8.0, was calculatedusing Gutenberg and Richter's for- regime behind the arcs [Scholzet al., 1971;K. Nakamura, mula: private communication,1977]. Although the amountof infor- log W0 = 1.5Mw + 11.8 mation is small and there are arcs, like the Scotia and Sunda arcs, where no information is available, it appears that ten- Table 2 (from Table 2 of Kanatnori [1977a]) lists the 10 sionaland compressionalstress regimes dominate in the active largestearthquakes that occurredbetween 1904 and 1976.Two and inactiveback-arc regions,respectively. In the following, points should be noted: (1) These 10 large earthquakesre- we will call thesetwo typesof back-arcregions 'Mariana' type leased more than 90% of the world's total seismicwave energy and 'Chilean' type for simplicity. releasedduring 1904-1976, and (2) none of thesereally great earthquakesoccurred at the M arianatype boundaries.This is INTERPLATEEARTHQUAKES AND MODE OF SUBDUCTION true for the 40 largestearthquakes listed by Kanamori[1977a], Why are there differentstress regimes behind subduction which cover more than 95% of the total seismicenergy (see boundaries?Could this difference be causedby some impor- Figure 4). Althoughthere are a few great earthquakesin the tant difference in the subduction processitself?. Interplate Assamand Mongoliaareas, all the truly greatthrust interplate thrust type earthquakes,that are causedby interactionbe- earthquakesoccur exclusively at the Chileantype subduction tweenthe subductingand overthrustingplates, should provide boundaries.At the Mariana type boundaries,subduction must pertinentinformation on thispoint. As iswell known,a major- be going on essentiallyaseismically if the convergencerate ity of the earthquakesthat occurare interplatethrust events at shown in Figure 3 representsthe present-daymotion [Kana- subduction boundaries. An examination of world seismicity mori, 1977b]. Figure 4 showsthat Chilean type boundaries maps[e.g., Barazangiand Dorman,1969, Tarr, 1974]shows have great earthquakes. that epicentersare more or lessevenly distributed along sub- Thus, in general,at a Chilean type marginthe slabappears duction boundaries.No systematicdifference is obviousbe- to force itself into the mantle overcomingthe strong resistive tween the Mariana and Chilean type boundaries.The sameis interactionwith the upper plate, whereasin the Mariana type true for the convergencerate at the boundaries(Figure 3): margin the slab falls into the mantle relatively freely. The Although the convergencerate seemshigher at Chileantype gradationfrom the former to the latter may be observedin boundariesthan at M ariana type ones,the distinctionis not Figure 4 from Alaska to the Marianas via the K urile, Japan, completelyclear. and Izu-Bonin arcs. The distinctionis, however, very clear in the rate of earth- OTHER PHENOMENA POSSIBLY RELATED TO THE quakeenergy release. To demonstratethis, we usethe seismic Two MODES OF SUBDUCTION energyreleased in great earthquakes.The conventionalearth- quake magnitude,M, [Gutenbergand Richter, 1954]suffers The differencesin the mode of subductionsuggested in the saturationfor great earthquakes(M > 7.5) and becomesun- previoussection have other possibleconsequences. We will UYEDA AND KANAMORI: BACK-ARC OPENING AND SUBDUCTION 1053

,./1906(8.6) ...... r -1970(7.8) •___•1966(73 o -• 3•9•, -20, ••1906 -2O

-40 1960 -4O

-60 MS _> 8.0 PlusAlaska, 1958(7.9) Kurile,1969 (7.8) Azores,1975 (7.9) Peru, 1966 (7.5) Peru, 1970 (7.8) Tokochi-Oki,1968 (7.9) Peru, 1974 (7.6) 0 I1904-1976 I ' I ½0I I I I I I ,50I I' I I I I -,50I I I I I I -40 I I ' " 0

Fig. 4. Great earthquakesduring 1904-1976 [after Kanamori, 1978]. Conventional magnitudes are in parentheses,and new magnitudesare in brackets.Black areasare rupture zones. examine someof them in this section.They are, in our view, horizontally beforesubduction. One featureto be noted is that promisingbut need further elaboration to be conclusive. both gravity and topographic highs are wide and prominent off the Aleutian and Kurile trenchesbut becomegradually less Outer Gravity High and Outer wide and less prominent toward the south from the Japan TopographicRise trench to the Marianas trench [Watts and Talwani, 1974, Fig- It is well known that often there exist a positive gravity ure 1]. The previously noted similarity [Kanamori, 1977b] to anomaly belt and topographic rise just outside of trenches. the trend of diminishinggreat earthquakesis interesting.Outer Hanks [1971] and Watts and Talwani [1974] interpretedthese gravity highsaround the circum-Pacificregion are summarized features as a result of the oceanic plate being compressed in Table 3 and Figure 5 [Getts, 1975; Watts et al., 1976]. Figure

TABLE 3. Outer Gravity High

Trench References* Outer Gravity High Ag, mGal Width, km

Continental Arc Peru-Chile 1, 2 Well developedsouth of 18øS 40-60 300-400 Middle America 1 Exists but not well defined Alaska 1, 3 Well developed 50 600 Sumatra-Java 1 Well developed 50 200 Nankai 1 Not clear ......

Trapped Back-Arc Aleutian 1, 3 Well developed 50-60 350-500

Inactivated Back-A rc Kurile 1, 3, 4 Well developed 40 300-400 Japan 1, 3 Well developed 30 300

Leaky Transform Andaman-Nicobar 1 Not developed

Back-Arc Spreading Mariana 1, 3 Not developed Tonga 1, 3 Not developednorth of 24øS 40-20 70 Kermadec 1, 3 Developed Ryukyu 1, 3 Not developed - 30 1O0

Unidentified Bonin 1, 3 North, developed;south, not 40 150 developed Philippine 1, 3 Well developed 60 400

*Referencesare 1, Wattsand Talwani [ 1974];2, Getts[1975]; 3, Wattset al. [ 1976];and 4, Hanks [ 1971]. 1054 UYEDA AND KANAMORI: BAcK-ARC OPENING AND SUBDUCTION

+90" +{20ø +{50ø +{80ø -{50ø -{20ø -900 -60 ø -50 ø tians, Alaska, and LesserAntilles, featuresshallower than 100 /''''''''''' c km are shown, but the problems related to the shallower features[Karig et al., 1976]will be touchedupon later. Figure 6 testifiesthat the dip of the Benioff-Wadatizone is definitely •0o•_ •.-"; + shallow for the Peru-Chile arc and Sumatra arc and deep for •/- , _ • •., -+20• the Mariana, Kermadec, and Scotia (not well constrained) arcs. Here again, the grouping typified by the Chilean and +20-• ' •- .••• M ariana arcs seems to hold. Some arcs, that are consideredto have inactive back-arcs, alsohave steepBenioff-Wadati zones: for example,Java, New Britain, Philippine,Solomon and possiblyNew Hebridesarcs. -•oø• • • , •11•ve/op• •-•o• However, theseBenioff-Wadati zonestend to be discontinuous I - poo,o, in the directionof dip as indicatedby broken linesin Figure 6. ] I I ] I I I I I • • •1 • I I I I I I I I I I I I I I +90• +120' +150' •180' -150' -120• -90 • -60 • -50• This may mean that the downgoingslab is disruptedin these subduction zones. Fig. 5. Distribution of trench outer gravity high [after Watts and Talwani, 1974; Getts, 1975; Watts et al., 1976]. Nature of Volcanism 5 shows substantial correspondencewith the two modes of The compositionsof volcanic rocks are different for differ- subduction. ent tectonicsettings [Miyashiro, 1975].Generally speaking,the We realize that there are problemsin both observationsand representativevolcanics are midoceanicridge type tholeiitesat modelsof the outer gravity and topographichighs [e.g., Cald- divergingboundaries, alkali basaltsand tholeiitesin midplate well et al., 1976]. Therefore,although the first-ordereffect may volcanism,and andesitesin addition to tholeiitesat converging be meaningful,final conclusionsmust await further investiga- boundaries.The origin of andesiteitself is one of the major tion. problems in igneouspetrology, but here we attempt to com- pare the abundance of andesitesfor different subduction Dip of Benioff-WadatiZones boundaries. The reason for such a comparison is that, for The dip of Benioff-Wadatizones may be shallowerfor the example, in westernNorth America, andesiticvolcanism was Chilean type subduction than for the M ariana type sub- dominant in early Tertiary, when the west coast of North duction. SimplifiedBenioff-Wadati zonesfrom varioussources America was a subductionboundary; the volcanismchanged including Isacks and Barazangi [1977], are shown in Table 4 to the basalt-rhyolite type after the collision of the Pacific- and Figure 6, in which mainly featuresdeeper than 100 km are Farallon ridge with the then existingtrench [Dickinson,1970]. represented.When the Benioff-Wadati zone is shallow, as in $cholz et al. [1971] suggestedthat the stressfield changed the case of the Andaman-Nicobar, Middle America, Aleu- from compressionalto tensional in western North America

TABLE 4. Dip of Benioff-Wadati Zones

Approximate Position of Plate Strike of Maximum Arc Boundary,deg Section,deg Depth, km Dip, deg References

Scotia 58os, 24ow EW 180 70 E.R. Engdahl, personalcommunication (1977) South Chile 40os, 72ow N100øE 160 30 Isacks and Barazangi [ 1977] Central Chile 30os, 70ow N100øE 150 I 1 Isacksand Barazangi[ 1977] North Chile 20os, 70ow N70øE 29• 30 Isacksand Barazangi[1977] Peru 10os, 78ow N60øE 150 8 Isacksand Barazangi[1977] Middle America 12ON, 87ow N 45 ø E 200 65 Deweyand Algermissen [1974] Lesser Antilles 15ON, 60ow EW 230 65 Sykesand Ewing [ 1965] Alaska 59.5ON, 150ow N73øW 150 55 Lahr[1975] Central Aleutians 51ON, 180OE NS 260 65 Engdahlet al. [1977] Kuriles 45øN, 153øE N40øW 650 50 Engdahlet al. [ 1977] Northeast Japan 40ON, 144OE EW 600 40 Ishida [ 1970] Ryukyu 24ON, 126øE N30 ow 300 45 Katsumataand Sykes[1969] Philippine(Mindanao) 5ON, 128OE EW 600 55 Hamilton [1974] Izu-Bonin 30øN, 143øE N 100 øW 500 65 Katsumataand Sykes [1969] Marianas 18øN, 148øE EW 700 90 Katsumataand Sykes [1969] Tonga 21os, 173ow N70øW 650 53 Isacksand Barazangi[1977] Kermadec 33øS, 178øW N70øW 550 65 Isacksand Barazangi[ 1977] New Zealand 40.5øS, 177øE N48øW 300 67 AnsellandSmith [1975] New Hebrides 17os, 167OE N70øE 300, 530 70 Isacksand Barazangi[ 1977] Dubois [ 1971] New Britain 6os, 147OE N40øW 200, 600 65 Isacks and Molnar [ 1971] Solomons 6os, 154OE N40øE 150; 550 in 70 Isacks and Molnar [ 1971] northwest corner Andaman 12ON, 92OE EW 100 50 Eguchi et al. [1979] Sumatra 5os, 100OE N60øE 200 30 Fitch [ 1970] Java 10øS, 113øE NS 300, 600 55 Fitch [ 1970] Banda Sea 7os, 130OE N30øW 400 55 Fitch [1970] UYEDA AND KANAMORI: BACK-ARC OPENING AND $UBDUCTION 1055

probably of terrigenousorigin or are scrapedoff pelagicsedi- o ments from the oceanic plate. The prism rises due to the

km compressionand sometimesemerges above sea level to form the so-called outer arc such as Mentawasi Island off Sumatra and Middleton Island off the Aleutian Arc. Landward of the

200 outer arc, sedimentarybasins called deep seaterraces or fore- arc basinsoften develop.The developmentof thosestructures is controlledby various factors suchas the supply of terrige- 300 nous material, the duration of subduction, and the history before the onsetof subduction.Thus one would expectthat the 400 regional difference in the mode of subduction significantly affectsthe structureof the accretiqnaryprism. In fact, the .500 growth of accretionaryprism appearsmore pronouncedin our Chilean type boundariesthan M ariana type ones [Karig and 6o0 Sharman, 1975, Figure 5]. Often, one encountersexposures of basementin the wall of the latter type trenches.As a result of

700 the growth of the sedimentary structure, the oceanic plate tends to be bent downward by the sediment load and thus Fig. 6. Dip of Benioff-Wadati zones (after the referenceslisted in createsthe long shallowdipping slopeat a shallowpart (depth Table 4). of <30 km) of the Benioff-Wadati zone [Karig et al., 1976]. when the Pacific-Farallonridge collided with the subduction Karig and Sharman [1975] attribute the difference in the zone. The changein the natureof volcanismmay be explained growth of fore-arc sedimentarystructures to that of the sedi- as follows (K. Nakamura, personal communication, 1977): ment supply(supply rate timesduration). Here we suggestthat The parental basaltic magmas can ascendand erupt more the differencein the mode of subductionmay be an additional easily in a tensionalstress field than in a compressionalenvi- factor which givesrise to the differencein the structureof the ronment. In the latter situation, the magma, while forcing its accretionaryprism. In the Chilean type boundaries,the strong way to the surface,tends to staylonger in the crustand is more mechanicalcoupling causes an extensivecontortion and devel- proneto reactwith the surroundingrocks to produceandesitic opment of the accretionaryprism, whereasin the M ariana rocks. type boundaries,because of the relativelyweak coupling,sub- According to the synthesisof Miyashiro [1975], andesites duction takes place without strong resistancefrom the upper are abundantin continentalarcs and mature islandarcs (island plate, resulting in less contorted, or lack of, accretionary arcswith substantialcontinental crust, e.g., Japan), and rare in prisms.However, more detailedinterpretation must await fur- immature island arcs (island arcs without continental crust, ther investigationson the origin of the sedimentsthat form the e.g., the Marianas). Here we note that thesecorrespond with accretionaryprism. The differencein the supplyitself would, our Chilean type and M ariana type arcs.This seemsto suggest to a certain degree,ultimately be causedby the differencein that the mode of developmentof arcs may also be controlled the subduction mode as mentioned earlier, because the M a- by the mode of subduction.In the Mariana type subduction riana type arcs always tend to be at somedistance from the boundaries,back-arc spreading will soonstart to break the arc continent. longitudinally[Karig, 1972], so that the arc becomesthin and As stated above, the outer arcs or frontal arcs are probably pushedaway from the continent,the sourceof terrigeneous the raised sediments. Such a rise would have the same mecha- sediments,inhibiting the rapid growth of the accretionary nism as that of the coastal uplifts estimated from coastal prism (discussedlater). terracesand the actual coastal uplifts associatedwith major Sugisaki[1972, 1976] demonstratedthat the abundanceof thrust earthquakes [Yonekura, 1975]. Such uplifts are also andesiteis controlledby the convergencerate: andesiteis more expectedto be much more evident at the Chilean type sub- abundantwhere the convergencerate is high. However,in his duction boundariesthan the M ariana type. Although further observation, the Tonga and Izu-Mariana arcs constituteno- investigationsare required, preliminary literature surveysin- table exceptionswhere the convergencerate is high and ande- dicate that Quaternary coastaluplifts are prominent in Japan, site is scarce.Sugisaki's observation seems to be better ex- Taiwan, Indonesia, western South America, Alaska, New plained by our interpretation. Guinea, and New Hebrides IOta and Naruse, 1977]. Further, accordingto Nakamura et al. [1977], the structure Although terracesare observedin Mariana and Tonga, the of volcanoesis also dependenton the tectonic stress:large rate of uplift appearsto be almost an order of magnitude stratovolanic edificesgrow in the Chilean type environment, smaller than in the Chilean type coasts [Taylor and Bloom, whereassmall many monogenicvolcanoes are found in the 1977; N. Yonekura, personalcommunication, 1978]. tensionalenvironment. The seriesof monogeniccones in the Middle America arc are examplesof the latter. Again, the POSSIBLE MODELS OF THE Two MODES differenceis causedby the general ease of magma eruption OF SUBDUCTION under a tensional regime. These observationsappear to con- firm our notion of the two typesof subduction,but certainly Summarizing the argument developedin the previous sec- more thorough investigationis needed. tions, there are two different modes of subduction. The extreme casesof each mode are representedby the Chilean arc SedimentaryStructure on the and M ariana arc. At the Chilean type subductionboundaries, Landward Wall of Trenches the mechanicalcoupling between the upper and lower platesis It is now believedthat an accretiona•'yprism of sediments strong,whereas at the M ariana type boundariesthe couplingis developson the landward wall of trenches.The sedimentsare either much weaker or effectivelynonexistent. The two types 1056 UYEDAAND KANAMORI: BACK-ARC OPENING AND SUBDUCTION

Fig. 7. Two kindsof subductionboundaries. may be illustratedas shownin Figure7. The differencein the in piecesinto the mantle. It may also be possiblethat under subductionmode seemsto be important not only for the such circumstances,the oceanicplate losesits supportat the problemof back-arcspreading but alsofor moregeneral prob- end and startssinking well beforeit reachesthe trench.Such a lems of orogeny. collapseof the oceanic plate may lead to the formation of marginalseas (Figures 8d-8f). It has beensuggested that the EvolutionaryModel decouplingof the slab took place at the Ryukyu Trench at This model views the different modesof subductionas rep- sometime in the Tertiary and then the basinsin the Philippine resentingthe different stagesof a singleevolutionary process Sea were formed as the slab foundered and the trench re- of subduction[Kanamori, 1971a, 1977b].As seenin Figure 8, treated, as in the fashion suggestedby Elsasser [1969] and the processof subductionstarts with low-anglethrusting as in Moberly [1972]. the case of the Alaska Trench. Here a wide contact zone In suchan evolutionarymodel, someexplanation is needed betweenthe upperand lower platesprovides a strongcoupling to explainwhy progressivelyadvanced stages are found succes- betweenthe two plates, and great earthquakesoccur at the sivelyfrom Alaska to the Ryukyu arc. One explanationis that interface. This is the mode that we call in this paper the as the subductionrate increasesprogressively from north to Chileantype and that we regardas the firststage of subduction southwestalong the northwesternPacific margin, the evolu- (Figure 8a). As the processgoes on, the oceanicplate subducts tionary processhas proceededmore rapidly along the margin. to a greaterdepth in the mantle, and the contactinterface is This explanation,however, may not applyto the SouthAmeri- heavilyfractured by repeatingfault movements.This repeated can margin, where a relativelyearly evolutionarystage is sug- fracturing may result in reducedcoupling of the two plates gestedin our model, yet the plate has been subductingfor a through various processessuch as formation of fault gouge, long time with a high rate. It is possiblethat as Molnar and increasedpore pressure, and possiblepartial melting. This may Atwater [1978] suggested,the age of the subductedplate is be the state of the Kurile arc (Figure 8b). If the process another factor that affects the mode of subduction. continues,the upper and lower plates will be almost com- Anchored Slab Model pletely decoupledso that there will be no more great earthquakes. The elongatedtongue of the high-densityslab will This model is an alternativeto the evolutionarymodel. The exert strongtension in the upperpart of the slab to causegreat basic idea is very simpleand similar to the one proposedby normal fault intraplate earthquake(Figure 8c), like the 1933 Moberly [1972] and Hyndman [1972]. It even resemblesone of Sanriku earthquake[Kanamori, 197lb]. Once alecoupled,the Wegener's[ 1924]postulates that marginalseas form in the rear slabmay break by thistype of intraplateearthquakes and sink of advancing continent. UYEDA AND KANAMORI: B^cK-Agc OPENING AND SUBDUCTION 1057

Cc•ntinental'10t•o2•._ Oceanic (3 LitI•osp•.•r•• , Lithosphere•(Alaska)Chiletype Breakoff •

Kurile N. Japan type

•c Sanriku type

t Litosphere) f

Fig. 8. Evolutionarymodel [Kanamori,1977b]. (a) Strongcoupling between the oceanicand continentallithospheres resultsin great earthquakesand breakoffof the subductionlithosphere at shallowdepths. (b) Partial decouplingresults in smallerearthquakes and continuoussubduction. (c) Further decouplingresults in aseismicevents and intraplatetensional events.(d) Sinkingplate resultsin retreatingsubduction and formation of a new thin lithosphere.(e) Episodicretreat and formation of ridges.(f) Deceleratedretreat and commencementof new subduction.

Referring to Figure 9, three velocities,VL, Vu, and Vr, are to open is Vv < Vt. In this model, V• is not important except defined. VL and Vu are the velocities of the lower and the that Vu must exceedV• for the boundary to be convergent.At upperplates, respectively. Vr is definedas the velocityof point this point, we assumethat the position of the trench is fixed to T where the lower plate starts deflectingdownward, i.e., the the mantle. This assumption seems reasonable because it positionof trench.The back-arcspreading can be formulated would be very difficult for the slab to move in the direction in terms of Vu and Vt. Namely, the condition for the back-arc perpendicularto the strike of the trench. Thus the slab acts as an anchor [Talwani, 1969; Tullis, 1972].In a frame of reference fixed to the mantle, we may put Vr = O. In this case, the vrr criterion for opening of back-arc basin would become Opening 7•la_(•. Vu<0

Not opening Vu>0 This criterion is testableif we know the plate motions relative to the mantle, or 'absolute motions.' Figure 10 showsthe absolute motions of upper plates at subductionzones com- •X puted from model AM1 of Minster et aL [1974]. Absolute Fig. 9. Anchored slab model. motion of the Philipine Sea plate is not well determined, 1058 UYEDA AND KANAMORI: BACK-ARC OPENING AND SUBDUCTION

+900 +1200 +150ø _+1800 -150o -120o -90 o -60 o -300 At the Chileantype boundaries,the upperplate overridesthe trench so that the trench is forcedto move with the speed +6 +600 of the upper plate. It can easily be seen,in suchcases, the Benioff-Wadatizone would dip at a shallow angle. In our model, openingby back-arcspreading and openingby leaky

+ +200 transform faults are consideredas being causedby the same ; _ ultimate mechanism,the motions of large plates.Since ordi- .o,.-. 0ø nary ocean floor spreadingis also consideredto be caused by the divergentmotions of platesin a similarway, we may even be able to generalizethat all theseprocesses are essen- tially alike. Finally, it shouldbe noticedthat the presentmodel is quite differentfrom that of Wilsonand Burke [1972],in whichthe I I ] I I I I I I I I [ I I I I I I I I [ I I • • I +900 +120o +lSOø •180ø -ISOø -120o -900 -60 o -300 two typesof subductiontectonics are assumedto be causedby Fig. 10. Absolutevelocity vectors of landward(upper) plates at whetherthe continental plate or the oceanicplate is advanci'ng subductionboundaries (after AM I of Minsteret al. [1974]and Fitch at the convergentboundaries. It might be pointedout that the [1972]). Units are centimetersper year. Nazca plate is advancingtoward South America as fast as the Pacificplate is advancingtoward Asia. although severalattempts have beenmade [Fitch, 1972;Mor- ModifiedAnchor Model (Lithospheric gan, 1972; Seno, 1977]. Here we used one of Fitch's models. Seno's motion is not to be used here because it relies on the RotationModel or CounterFlow Model) relativemotion at the Izu-Mariana boundaryand assumption One notablefeature of trench-arc-back-arcsystems around of nonspreadingMariana Trough. However, in fact, sinceall the Pacific,the east-westasymmetry, has been pointed out by the modelsindicate substantial westward motion of the Philip- severalauthors [e.g., Nelson and Temple,1972; Moore, 1973]. pine Sea plate, our discussionis not affectedby the choiceof Arcswith an activeback-arc basin tend to havea westdipping the model. It appearsthat the abovecriterion for 'active' back- high-angledowngoing slab, whereasarcs with inactiveback- arc is almost perfectly valid. At the Chilean arc, the South arc havean eastdipping low-angle slab. Someof the models American continent is moving westwardand overridingthe [Nelsonand Temple,1972; Moore, 1973]imply that thereis an trench, while at the Marianas arc, the upper plate is fairly eastwardmotion of asthenosphererelative to the lithosphere rapidly retreating from the trench, and most of the inter- causingthe situationillustrated in Figure 11. The absolute mediate casesare also quite supportiveof the model. In the plate motionmodel AM1 of Minsteret al. [1974]is basedon Andaman and Tonga areas,though marginal, the upper plate the hot spot frame of referenceand in fact has a net almost is moving away from the trench, whereasin the areas of the westwardlithospheric rotation (129øE, 74øS as the Euler pole) Aleutian, Kurile, and Japan trenchesthe upper plates are of 0.1 lø/m.y. If the lithospherehas a netwestward rotation, it moving toward the trenches,but their velocitiesare small. The would be in someway relatedto the earth'srotation and tidal only casewhere the model is not supportedis the Ryukyu arc. effects[Bostrom, 1971; Moore, 1973].This point will be the As the absoluteplate motionsare not accurateto 1 cm/yr, the matter of future investigationboth observationallyand theo- Ryukyu situation may not be too serious.Since the trench- retically. ward velocity of the upper plate (Eurasia plate) at the Kurile In our model, so far, the downgoingslab hasbeen consid- and Japan trenchesis also of the order of 1 cm/yr, it may be eredfixed to the deepmantle. However, it may be fixedto the quite possiblethat slight changesin the motion of the Eurasia asthenosphere.If this were the case, the movement of the plate could have made its velocitylandward in the past. This trenchwould be controlledby the flow in the asthenosphere. would explain how the Kurile and Japan basinswere formed We are unableto assessthe velocityof suchasthenospheric at some past time and are now inactivated. In the caseof the flow, but if we simplyassume that the flow is the counterflow New Hebridesarc, the motion of the upperplate (Fiji Plateau) to balancethe westwardlithospheric rotation and the thick- is unclear. If it were fixed to the Pacific Plate as in the case nessof the asthenosphereis comparableto that of the litho- shownin Figure 10, the New Hebridesback-arc should not be sphere,the velocitymay be similarto 0.1lø/m.y. This would spreading. meanthat the trenchesare movingeastward with this speed In this model, tensiontectonics in the back-arcregion are relativeto the deepmantle. The relative velocities of theupper not a directresult of the subductionof a slab.The subducting plateswith respect to thecorresponding trenches computed on slab--the anchor--only fixesthe trenchposition to the mantle, this model are as shownin Figure 12. The generalsituation and the developmentof tensionaltectonics is determinedby doesnot differgreatly from that in Figure10, but the Tonga the motion of the upper plate. Of course,not all the trenches and Ryukyu areasbecome more conformablewith our ideas. are fixed to the mantle; the slab plays the role of the anchor effectivelyonly at the Mariana type boundaries. In the presentmodel, the mantleis consideredstationary for simplicity. If some flow exists in the mantle, the slab and trench would be movedalong with the flow [Hager and O'Con- nell, 1978]. When the downgoingslab is short, as in the case of Scotia arc, flow in the asthenospheremay have a large effect.In fact,although the Scdtia plate is probablymoving westwardwith a small velocity, the Sandwichplate and the trench may be moving eastwardif the Scotia ridge is really spreadingas fast as 8 cm/yr [Barker, 1970;Forsyth, 1975]. Fig. 11. An interpretationfor the east-westasymmetry. UYEDA AND KANAMORI: BACK-ARC OPENING AND SUBDUCTION 1059

+90ø +120ø +150ø +180ø -150o -120o -90 o -60 o -30 ø Bostrom, R. C., Westward displacementof the lithosphere,Nature, 234, 356-538, 1971.

+6 +60ø Caldwell,J. G., W. F. Haxby, D. E. Karig, and D. L. Turcotte,On the applicabilityof a universalelastic trench profile, Earth Planet.Sci. Lett., 31, 239-246, 1976. Chase, C. G., Tectonic history of the Fiji Plateau, Geol.Soc. Amer. _

_ Bull., 82, 3087-3110, 1971. Christiansen,R. L., and P. W. Lipman,Cenozoic volcanism and plate 0 o tectonicevolution of the westernUnited States,II, Late Cenozoic, Phil. Trans. Roy. Soc. London,271, 249-284, 1972. -20 --- 20 ø Chung, W. Y., and H. Kanamori, Subductionprocess of a fracture zone and aseismicridges--The focal mechanismand sourcecharac- -40ø _-40ø teristicsof the New Hebridesearthquakes of Jan. 19, 1969and some related events,Geophys. J. Roy. Astron. Soc., 54, 221-240, 1978. -60 -60 o Coleman, P. J., and G. H. Packham, The Melanesianborderlands and

I i i , i i • i i i • I i i i I I I I i I I I i i i L_ India-Pacificplates' boundary, Earth. Sci.Rev., 12, 197-233, 1976. +90o +120ø +150ø _+180ø -150ø -120o -90 o -600 -30o Connelly, J. B., Tectonic developmentof the BismarckSea basedon gravity and magneticmodeling, Geophys. J. Roy. Astron.Soc., 46, Fig. 12. Velocity vectors of landward (upper) plates relative to 23-40, 1976. trenchesin counterflow model. Units are centimetersper year. Cooper,A. K., M. S. Marlow, and D. W. Scholl,Mesozoic magnetic lineationsin the BeringSea marginalbasins, J. Geophys.Res., 81, 1916-1934, 1976. CONCLUSIONS Cormlet, V. F., Tectonics near the junction of the Aleutian and Subduction plate boundaries can be classified into two Kurile-Kamchatkaarcs and a mechanismfor middleTertiary mag- matism in the Kamchatka Basin, J. Geophys.Res., 80, 443-453, groups according to whether their back-arc regions are ac- 1975. tively spreadingor not. The typical casesare the M ariana and Curray, J. R., D. G. Moore, L. Lawyer, F. Enamel, R. W. Raitt, and the Chile boundaries.Intraplate earthquakemechanisms show M. Henry, Tectonics of the Andaman Sea and Burma, Amer. Ass. that these two classesof back-arc regionsare actually under Petrol. Geol. Bull., in press, 1979. tensional and compressionaltectonic stresses.Great thrust Dewey, J. W., and S. T. Algermissen,Seismicity of the middle Amer- ica arc-trenchsystem near Managua, Nicaragua,Bull. Seismol.Soc. type interplate earthquakesare occurringonly along the latter Amer., 64, 1033-1048, 1974. type of subductionboundaries, indicating that there is a signif- Dickinson, W. R., Relations of andesites,granites and derivative icant differencein the degree of coupling of the upper and sandstonesto arc-trenchtectonics, Rev. Geophys.Space Phys., 8, lower plates betweenthe M ariana type and the Chilean type 813-860, 1970. boundaries. This in turn indicates that the mode of subduction Dubois, J., Propagationof P wavesand Rayleighwaves in Melanesia: Structural implications,J. Geophys.Res., 76, 7217-7240, 1971. is differentin the two cases.The differentcases are interpreted Dugas, F., J. Dubois, A. Lapoui!le, R. Louat, and C. Revenne,Struc- as caused by either differencesin the stage of subduction tural characteristics and tectonics of an active island arc--The process(evolutionary model) or differencesin the motion of New Hebrides, in Proceedingsof the International Symposiumof Geodynarnicsof the SouthwestPacific, pp. 79-90, Technip, Paris, the landward plate (anchored slab model). 1977. Eguchi, T., S. Uyeda, and T. Maki, Seismotectonicsand spreading •lcknowledgments.This studywas conductedwhile Seiya Uyeda was at the SeismologicalLaboratory, California Institute of Tech- history of Andaman Sea, Tectonophysics,in press, 1979. nology,as a ShermanFairchild DistinguishedScholar. We thank Seth E!sasser,W. M., Convection and stresspropagation in the upper mantle, in The Application of Modern Physics to the Earth and Stein, Robert Geller, and Don Anderson for helpful commentson the manuscript.We also thank Seth Stein for his assistancein computing Planetary Interiors, edited by S. K. Runcorn, pp. 223-246, Inter- science, New York, 1969. plate motions.This work was partially supportedby a grant from the National Science Foundation, EAR76-14262. Contribution 3012 of Engdahl, E. R., N.H. Sleep, and M. T. Lin, Plate effectsin north the Division of Geological and PlanetarySciences, California Institute Pacific subductionzones, Tectonophysics,37, 95-116, 1977. of Technology,Pasadena, California, 91125. Fitch, T., Earthquake mechanismsand island arc tectonics in the Indonesian-Philippineregion, Bull. Seismol.Soc. Amer., 60, 565- REFERENCES 591, 1970. Fitch, T., Plate convergence,transcurrent faults and internal deforma- Ansell, J. H., and E.G. C. Smith, Detailed structure of a mantle tion adjacentto southeastAsia and the westernPacific, J. Geophys. seismiczone using the homogeneousstation method, Nature, 253, Res., 77, 4432-4460, 1972. 518-519, 1975. Forsyth, D., An observationof compressirestress between two mid- Atwater, T., Implicationsof plate tectonicsfor the Cenozoictectonic ocean ridges,Nature, 243, 78, 1973. evolution of western North America, Geol. Soc. •lmer. Bull., 81, Forsyth, D. W., Fault plane solutions and tectonics of the south 3513-3535, 1970. Atlantic and Scotia Sea, J. Geophys.Res., 80, 1429-1443, 1975. Averyanova, V. N., SeismicFoci in the Far East, p. 207, K etter Press, Fukao, Y., and M. Furumoto, Mechanismof large earthquakesalong Jerusalem, 1973. the easternmargin of the Japan Sea, Tectonophysics,25, 247-266, Barazangi,N., and J. Dorman, World seismicitymap compiledfrom 1975. ESSA Coast and Geodetic Survey epicenterdata, 1961-1967, Bull. Gedney, L., Tectonic stressesin southern Alaska in relationshipto Seismol. Soc. •tmer., 59, 369-380, 1969. regional seismicityand the new global tectonics,Bull. Seismol.Soc. Barker, P. F., A spreadingcenter in the East Scotia Sea, Earth Planet. Amer., 60, 1789-1802, 1970. $ci. Lett., 15, 123-132, 1970. Getts, T. R., Gravity and tectonics of the Peru-Chile Trench and Barker, P. F., and D. H. Griffiths,The evolutionof the ScotiaRidge eastern Nazca Plate 0-33ø30'S, Ph.D. thesis, Univ. of Hawaii, and ScotiaSea, Phil. Trans.Roy. Soc.London, 271, 151-183, 1972. Honolulu, 1975. Beloussov,V. V., Some problemsof developmentof the earth's crust Gutenberg,B., and C. F. Richter,Seismicity of the Earth andAssoci- and upper mantle of oceans,in The Crustand UpperMantle of the ated Phenomena,310 pp., Princeton University Press, Princeton, Pacific•lrea, Geophys.Monogr. Ser., vol. 12, edited by L. Knopoff, N.J., 1954. and P. Hart, pp. 449-459, AGU, Washington,D.C., 1968. Hager, B. H., and R. J. O'Connell, Subductionzone dip anglesand Beloussov,V. V., and E. M. Ruditch, Island arcsin the development flow driven by plate motion, Tectonophysics,50, 111-133, 1978. of the earth's structure, J. Geol., 69, 647-658, 1961. Ben-Avraham, Z., and S. Uyeda, The evolution of the China Basin Hamilton, W., Earthquake map of the Indonesianregion, Map 1-875- and the Mesozoic paleogeographyof Borneo, Earth Planet. Sci. C, U.S. Geol. Surv., Reston, Va., 1974. Lett., 18, 365-376, 1973. Hanks, T. C., The Kurile Trench-Hokkaido risesystem: Large shallow 1060 UYEDAAND KANAMORI: BACK-ARC OPENING AND $UBDUCTION

earthquakesand simplemodels of deformation,Geophys. J. Roy. Lahr, J. C., Detailed seismicinvestigation of Pacific-North American Astron. $oc., 23, 173-189, 1971. plate interaction in South Alaska, Ph.D. thesis, Columbia Univ., Hasebe,K., N. Fujii, and S. Uyeda, Thermal processesunder island New York, 1975. arcs, Tectonophysics,10, 335-355, 1970. Larson, R. L., Bathymetry, magnetic anomalies and plate tectonic Hawkins,J. W., Jr., Geologyof the Lau Basin,a marginalsea behind history of the mouth of the Gulf of California, Geol. $oc. elmer. the Tonga arc, in The Geologyof ContinentalMargins, edited by Bull., 83, 3345-3360, 1972. C. A. Burk and C. L. Drake, pp. 505-520, Springer,New York, Lawyer, L. A., and J. W. Hawkins, Diffuse magnetic anomalies in 1974. marginal basins:Their possibletectonic and petrologicsignificance, Heezen,B., et al., Initial Reportof theDeep Sea DrillingProject, vol. 6, Tectonophysics,45, 323-339, 1978. pp. 293-388, U.S. GovernmentPrinting Office, Washington, D.C., Louden, K., Magnetic anomaliesin the westernPhilippine Basin, in 1971. The Geophysicsof the Pacific Basin and Its Margin, Geophys. Herman, B. M., R. N. Anderson, and M. Truchan, Extensional tec- Monogr. $er., vol. 19, edited by G. Sutton, M. H. Manghnani, and tonicsin OkinawaTrough, Amer. Ass. Petrol. Geol. Bull., in press, R. Moberly, pp. 253-267, AGU, Washington, D.C., 1976. 1979. Luyendyk, B. P., W. F. Brgan, and P. A. Jezek, Shallow structureof Hilde, T. W. C., and J. M. Wageman,Structure and origin of Japan New Hebrides Island Arc, Geol. Soc. elmer. Bull., 85, 1287-1300, Sea, in The WesternPacific, edited by P. J. Coleman,pp. 415-434, 1974. University of Western Atistralia Press,Nedlands, 1973. M alfait, B. T., and M. G. Dinkelman, Circum-Caribbean tectonic and Hilde, T. W. C., S. Uyeda, and L. Kroenke, Evolution of the western igneousactivity and the evolutionof the Caribbeanplate, Geol.Soc. Pacific and its margin, Tectonophysics,38, 145-165, 1977. elmer. Bull., 83, 251-272, 1972. Honda, H., A. Masatsuka, and M. Ichikawa, On the mechanism of Matsuda, T., and S. Uyeda, On the Pacific-type orogeny and its earthquakesand the stresses•roducing them in Japan and its model-extensionof the paired belts concept and possible origin vicinity, 3, Geophys. Mag., 33, 271-279, 1967. of marginal seas,Tectonophysics, 11, 5-27, 1971. Hyndman, R., Plate motions relative to the deep mantle and the Matumoto, T., and G. V. Latham, Aftershocks of the Guatemalan developmentof subductionzones, Nature, 238, 1972. earthquake of February 2, 1976, Geophys.Res. Lett., 3, 599-603, Ichikawa, M., Mechanism of earthquakesin and near Japan and 1976. related problems, in The Earth's Crust and Upper Mantle, Geophys. McKenzie, D. P., Speculationson the consequencesand causesof Monogr. $er., vol. 13, edited by P. Hart, AGU, Washington,D.C., plate motions, Geophys.J. Roy. elstron.$oc., 18, 1-32, 1969. 1969. McKenzie, D. 1•., and W. J. Morgan, Evolution of triple junctions, Isacks, B. L., and M. Barazangi, Geometry of Benioff zones, lateral Nature, 224, 125-133, 1969. segmentationand downwardsbending of the subductedlithosphere, Mendiguren, J. A., Focal mechanismof a shock in the middle of the in Island Arcs, Deep Sea Trenchesand Back-Arc Basins,Maurice Nazca plate, J. Geophys.Res., 76, 3861-3879, 1971. Ewing$er., vol. 1, edited by M. Talwani and W. C. Pitman III, pp. Minster, J. B., T. H. Jordan, P. Molnar, and E. Haines, Numerical 99-114, AGU, Washington, D.C., 1977. modelingof instantaneousplate tectonics,Geophys. J. Roy. ,4stron. Isacks,B., and P. Molnar, Distribution of stressesin the descending Soc., 36, 541-576, 1974. lithdspherefrom a global survey of focal mechanismsolution of Miyashiro, A., Volcanic rock seriesand tectonic setting,elnnu. Rev. mantle earthquakes,Rev. Geophys.Space Phys., 9, 103-174, 1971. Earth Planet. Sci., 3, 251-269, 1975. Ishida, M., Seismicityand travel-timeanomaly in and aroundJapan, Moberly, R., Origin of lithospherebehind island arcs, with reference Bull. EarthquakeRes. Inst. Tokyo Univ., 48, 1032-1051, 1970. to the western Pacific, Geol. Soc. elmer. Mern., 132, 35-55, 1972. Kanamori, H., Great earthquakesat islandarcs and the lithosphere, M olnar,P., andT. Atwater, Interarc spreading and cordilleran tec- Tectonophysics,12, 187-198, 197la. tonics as alternatesrelated to the age of subductedoceanic litho- Kanamori, H., Seismologicalevidence for a lithosphericnormal fault- sphere,Earth Planet. $ci. Lett., 41, 330-340, 1978. ing--The Sanrikuearthquake of 1933,Phys. Earth Planet.Interiors, Molnar, P., and L. R. Sykes,Tectonics of the Caribbean and Middle 4, 289-300, 1971b. America regionsfrom focal mechanismand seismicity,Geol. Soc. Kanamori, H., The energyrelease in great earthquakes,J. Geophys. elmer. Bull., 80, 1639-1684, 1969. Res., 82, 2981-2987, 1977a. Moore, G. W., Westward tidal lag as the driving force of plate tecton- Kanamori, H., Seismicand aseismicslip along subductionzones and ics, Geology,1, 99-100, 1973. their tectonicimplications, in IslandArcs, Deep Sea Trenchesand Morgan, W. J., Plate motions and deep mantle convection,Geol. Soc. Back-Arc Basins,Maurice Ewing Ser., vol. 1, edited by M. Talwani elmer. Mem., 132, 7-22, 1972. and W. C. Pitman III, pp. 162-174, AGU, Washington, D.C., Nakamura, K., Arrangement of parasitic conesas a possiblekey to 1977b. regional stressfield, Bull. Volcanol.Soc. Jap. (in Japanesewith Kanamori, H., Quantification of earthquakes,Nature, 271,411-414, English abstract), 14, 8-20, 1969. 1978. Nakamura, K., K. H. Jacobs,and J. N. Davies, Volcanoesas possible Karig, D. E., Ridges and basinsof the Tonga-Kermadecisland-arc indicators of tectonic stress orientations--Aleutian and Alaska, system,J. Geophys.Res., 75, 239-254, 1970. Pure elppl. Geophys.,115, 87-112, 1977. Karig, D. E., Structural history of the Mariana Island Arc system, Nelson,T. H., and P. G. Temple, Mainstreammantle convection:A Geol. Soc. Amer. Bull., 82, 323-344, 1971a. geologicanalysis of plate motion,elmer. elss. Petrol. Geol. Bull., 56, Karig, D. E., Origin and developmentof marginal basins in the 226-246, 1972. westernPacific, J. Geophys.Res., 76, 2542-2561, 1971b. Ota, Y., and Y. Naruse, Coastal terraces of Japan (in Japanese), Karig, D. E., Remnant arcs, Geol. $oc. Amer. Bull., 83, 1057-1068, Kagaku,47, 281-292, 1977. 1972. Plafker, G., Tectonicaspects of the Guatemalaearthquake of 4 Feb. Karig, D. E., and J. Mammerickx, Tectonic framework of the New 1976, Science, 193, 1201-1208, 1976. Hebrides Island arc, Mar. Geol., 12, 187-205, 1972. Richardson,R. M., S. Solomon, and N. Sleep,Intraplate stressas an Karig, D. E., and G. F. Sharman III, Subduction and accretion in indicator of plate tectonics,J. Geophys.Res., 81, 1847-1856, 1976. trenches, Geol. Soc. Amer. Bull., 86, 377-389, 1975. Scholz,C. H., M. Barazangi,and M. L. Sbar, Late Cenozoicevolution Karig, D. E., J. G. Caldwell, and E. M. Parmentier, Effectsof accre- of the Great Basin, western United States, as an ensialic inter-arc tion on the geometry of the descendinglithosphere, J. Geophys. basin, Geol. Soc. elmer. Bull., 82, 2979-2990, 1971. Res., 81, 6281-6291, 1976. Shih, T., Marine magnetic anomaliesand origin of the western and Katsumata, M., and L. R. Sykes,Seismicity and tectonicsof the west- northern Philippine Sea, paper presentedat the Joint Annual meet- ern Pacific: Izu-Mariana-Caroline and Ryukyu-Taiwan regions, ing, Geol. Soc. of Amer., Geol. Ass. of Can., and Mineral. Ass. of J. Geophys.Res., 74, 5923-5948, 1969. Can., Toronto, Ont., Oct. 23-26, 1978. Kobayashi, K2 and N. Isezaki, Magnetic anomaliesin Japan Sea and Shikoku Basin and their possibletectonic implications, in The Seno, T., The instantaneousrotation vector of the Philippine Sea Geophysicsof the Pacific Ocean Basin and Its Margin, Geophys. plate relative to the Eurasian plate, Tectonophysics,42, 209-226, Monogr. $er., vol. 19, edited by G. H. Sutton, M. H. Manghnani, 1977. and R. Moberly, pp. 235-251, AGU, Washington,D.C., 1976. Sleep, N., and M. N. T oksoz, Evolution of marginal basins,Nature, Katsumata, M., and L. R. Sykes, Seismicityand tectonics of the 233, 548-550, 1971. westernPacific: Izu-Mariana-Caroline and Ryukyu-Taiwanregions, Sleep,N.H., Stressand flow beneath island arcs, Geophys. J. Roy. J. Geophys.Res., 74, 5923-5948, 1969. elstron. $oc., 42, 827-857, 1975. UYEDA AND KANAMORI: BACK-ARc OPENING AND SUBDUCTION 1061

Smith, R. B., Intra-plate tectonicsof the westernNorth American Tomoda,Y., K. Kobayashi,J. Segawa,M. Nomura,K. Kimura, and plate, Tectonophysics,?7, 323-336, 1977. T. Saki, Linear magneticanomalies in the ShikokuBasin, north- Stauder,W., Mechanismand partial distributionof Chilean earth- easternPhilippine Sea, J. GeomagmGeoelec., 28, 47-56, 1975. quakeswith relationto subductionof oceanicplate, J. Geophys. Tullis, T. E., Evidencethat lithosphereslabs act as anchors(abstract), Res., 78, 5033-5061, 1973. Eos Trans. AGU, 53, 522, 1972. Stauder,W., Subductionof the Nazca plate underPeru as evidenced Uyeda, S., ,and Z. Ben-Avraham,Origin and developmentof the by focalmechanism and by seismicity,J. Geophys.Res., 80, 1053- PhilippineSea, Nature, 240, 176-178, 1972. 1064, 1975. Uyeda,S., andA. Miyashiro,Plate tectonics and the Japanese Islands: Stauder,W., and L. Mualchin, Fault motionin the largerearthquakes A synthesis,Geol. Soc. Amer. Bull., 85, 1159-1170,1974. of the Kurile-Kamchatka arc and of the Kurile-Hokkaido corner, J. Veith, K. F., The relationshipof islandarc seismicity to platetectonics Geophys.Res., 81,297-308, 1976. (abstract),Eos Trans.AGU, 55, 349, 1974. Sugimura,A., and S. Uyeda,Island Arcs, Japan and Its Environs,247 Watanabe,T., M. G. Langseth,and R. N. Anderson,Heat floor in pp., Elsevier,New York, 1973. back-arc basins of the western Pacific, in Island Arcs, Deep Sea Sugisaki,R., Tectonicaspects of Andesitcline, Nature Phys.Sci., 240, Trenchesand Back-Arc Basins,Maurice Ewing Ser., vol. l, edited 109-111, 1972. by M. Talwaniand W. C. PitmanIll, pp. 137-161,AGU, Washing- Sugisaki,R., Chemicalcharacteristics of volcanicrocks: Relation to ton, D. C., 1977. plate movements,Lithos, 9, 17-30, 1976. Watts, A. B., and M. Talwani, Gravity anomaliesseaward of deep-sea Sykes,L. R., Seismologicalevidence for transformfaults, sea floor trenchesand their tectonic implications, Geophys.J. Roy. Astron. spreadingand continental drift, in Proceedingsof NASA Sym- Soc., 36, 57-90, 1974. posium,History of the Earth's Crust,edited by R. A. Phinney,pp. Watts, A. B., and J. K. Weissel,Tectonic history of the Shikoku 120-150, PrincetonUniversity Press,Princeton, N.J., 1968. marginalbasin, Earth Planet.Sci. Lett., 25, 239-250, 1975. Sykes,L. R., and M. Ewing,The seismicityof the Caribbeanregion, J. Watts, A. B., M. Talwani, and J. R. Cochran, Gravity field of the Geopbys.Res., 70, 5065-5074, 1965. northwestPacific Ocean Basin and its margin, in The Geophysicsof Sykes, L. R., and M. Sbar, Intra-plate earthquakes,lithospheric the PacificOcean and Its Margin, Geophys.Monogr. Ser., vol. 19, stressesand the driving mechanismof plate tectonics,Nature, 245, pp. 17-34, AGU, Washington,D.C., 1976. 298-302, 1973. Wegener,A., TheOrigin of Continentsand Oceans, Methuen, London, Sykes,L. R., and M. L. Sbar,Focal mechanismsolutions of intraplate 1924. earthquakesand stressesin the lithosphere,in Geodynamicsof Ice- Weissel,J. K., and A. B. Watts, Tectoniccomplexities in the South Fiji land and North Atlantic Area, edited by L. Kristjansson,pp. 207- marginalbasin, Earth Planet.Sci. Lett., 28, 121-126, 1975. 224, D. Riedel, Hingham, Mass., 1974. Wilson, J. T., A new classof faults and their bearing on continental Talwani, M., Plate tectonics and deep-seatrenches (abstract), Eos drift, Nature, 207, 343-347, 1965. Trans. A G U, 50, 180, 1969. Wilson, J. T., and K. Burke, Two typesof mountainbuilding, Nature, Tarr, A. C., World seismicitymap, U.S. Geol. Surv., Reston, Va., 239, 448-449, 1972. 1974. Y onekura, N., Quaternary tectonic movementsin the outer arc of Taylor, F. W., and A. L. Bloom,Coral reefson tectonicblocks, Tonga southwestJapan with specialreference to seismiccrustal deforma- IslandArc, in Proceedingsof the ThirdInternational Coral Reef tiona, Bull. 7, pp. 19-71, Dep. of Geogr.,Univ. of Tokyo, Tokyo, Symposium,pp. 275-281,Rosenstiel School of Marineand Atmo- Japan, 1975. sphericScience, University of Miami, Miami, Fla., 1977. Thompson,G. A., andD. B. Burke,Regional geophysics ofthe Basin (Received February 7, 1978; and RangeProvince, Annu. Rev. Earth Planet. Sci., 2, 213-238, revised August 21, 1978; 1974. acceptedSeptember 25, 1978.)