80 Physics of the Earth and Planetary Interiors, 53 (1988) 80—166 Elsevier Science Publishers By., Amsterdam — Printed in The Netherlands Large intermediate-depth earthquakes and the subduction process Luciana Astiz ~, Thorne Lay 2 and Hiroo Kanamori ~ ‘Seismological Laboratory, California Institute of Technology, Pasadena, CA (U.S.A.) 2 Department of Geological Sciences, University of Michigan, Ann Arbor, MI (USA.) (Received September 22, 1987; accepted October 21, 1987) Astiz, L., Lay, T. and Kanamori, H., 1988. Large intermediate-depth earthquakes and the subduction process. Phys. Earth Planet. Inter., 53: 80—166. This study provides an overview of intermediate-depth earthquake phenomena, placing emphasis on the larger, tectonically significant events, and exploring the relation of intermediate-depth earthquakes to shallower seismicity. Especially, we examine whether intermediate-depth events reflect the state of interplate coupling at subduction zones. and whether this activity exhibits temporal changes associated with the occurrence of large underthrusting earthquakes. Historic record of large intraplate earthquakes (m B 7.0) in this century shows that the New Hebrides and Tonga subduction zones have the largest number of large intraplate events. Regions associated with bends in the subducted lithosphere also have many large events (e.g. Altiplano and New Ireland). We compiled a catalog of focal mechanisms for events that occurred between 1960 and 1984 with M> 6 and depth between 40 and 200 km. The final catalog includes 335 events with 47 new focal mechanisms, and is probably complete for earthquakes with mB 6.5. For events with M 6.5, nearly 48% of the events had no aftershocks and only 15% of the events had more than five aftershocks within one week of the mainshock. Events with more than ten aftershocks are located in regions associated with bends in the subducted slab. Focal mechanism solutions for intermediate-depth earthquakes with M> 6.8 can be grouped into four categories: (1) Normal-fault events (44%), and (2) reverse-fault events (33%), both with a strike nearly parallel to the trench axis. (3) Normal or reverse-fault events with a strike significantly oblique to the trench axis (10%), and (4) tear-faulting events (13%). The focal mechanisms of type I events occur mainly along strongly or moderately coupled subduction zones where a down-dip extensional stress prevails in a gently dipping plate. In contrast, along decoupled subduction zones great normal-fault earthquakes occur at shallow depths (e.g., the 1977 Sumbawa earthquake in Indonesia). Type 2 events, with strike subparallel to the subduction zone, and most of them with a near vertical tension axis, occur mainly in regions that have partially coupled or uncoupled subduction zones and the observed continuous seismicity is deeper than 300 km. The increased dip of the downgoing slab associated with weakly coupled subduction zones and the weight of the slab may be responsible for the near vertical tensional stress at intermediate depth and, consequently, the change in focal mechanism from type 1 to type 2 events. Events of type 3 occur where the trench axis bends sharply causing horizontal (parallel to the trench strike) extensional or compressional intraplate stress. Type 4 are hinge-faulting events. For strongly coupled zones we observed temporal changes of intermediate-depth earthquake activity associated with the occurrence of a large underthrusting event. After the occurrence of a large underthrusting event, the stress axis orientation of intermediate-depth earthquakes changes from down-dip tensional to down-dip compressional (e.g., 1960 Chile, 1974 Peru, 1982 Tonga and 1952 Kamchatka earthquakes), or the number of large intermediate events decreases for a few years (e.g., 1964 Alaska and 1985 Valparaiso earthquakes). We conclude that even though the stress changes induced by slab pull and slab distortion control the general pattern of intermediate-depth seismicity, spatial and temporal variations of the intraplate stress associated with interplate coupling are important in controlling the global occurrence of large intermediate-depth events. I. Introduction contact zone between subducting and overriding plates being the locus of large underthrusting In the theory of plate tectonics, subduction of earthquakes. Such large interplate thrust earth- oceanic lithosphere plays a primary role, with the quakes have been extensively studied, and are
87
Embed
Large intermediate-depth earthquakes and the subduction ...
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
80 PhysicsoftheEarth and PlanetaryInteriors, 53 (1988)80—166ElsevierSciencePublishersBy., Amsterdam— Printed in The Netherlands
Astiz, L., Lay, T. andKanamori, H., 1988. Large intermediate-depthearthquakesand the subductionprocess.Phys.Earth Planet. Inter., 53: 80—166.
This study provides an overview of intermediate-depthearthquakephenomena,placing emphasison the larger,tectonically significant events,and exploring the relation of intermediate-depthearthquakesto shallower seismicity.Especially,we examinewhetherintermediate-deptheventsreflect the stateof interplatecoupling at subductionzones.andwhetherthis activity exhibits temporalchangesassociatedwith theoccurrenceof largeunderthrustingearthquakes.Historic recordof large intraplateearthquakes(mB � 7.0) in this centuryshows that the New Hebridesand Tongasubductionzoneshave thelargest numberof large intraplateevents.Regionsassociatedwith bendsin thesubductedlithospherealso havemanylargeevents(e.g. Altiplano andNew Ireland). We compileda catalogof focalmechanismsfor events that occurredbetween1960 and 1984 with M> 6 and depthbetween40 and200 km. The final catalogincludes 335 events with 47 new focal mechanisms,and is probably complete for earthquakeswith mB � 6.5. Foreventswith M � 6.5, nearly 48% of theevents had no aftershocksandonly 15% of the eventshad more than fiveaftershockswithin oneweek of themainshock.Eventswith morethan tenaftershocksarelocatedin regionsassociatedwith bendsin thesubductedslab. Focalmechanismsolutionsfor intermediate-depthearthquakeswith M> 6.8 canbegrouped into four categories:(1) Normal-fault events(44%), and (2) reverse-faultevents(33%), both with a strikenearlyparallelto the trenchaxis. (3) Normalorreverse-faulteventswith a strikesignificantly oblique to thetrenchaxis(10%), and (4) tear-faultingevents (13%). The focal mechanismsof type I events occur mainly along strongly ormoderatelycoupled subductionzones where a down-dip extensional stressprevails in a gently dipping plate. Incontrast,along decoupledsubductionzones great normal-faultearthquakesoccur at shallow depths(e.g., the 1977Sumbawaearthquakein Indonesia).Type 2 events,with strike subparallelto the subductionzone, and mostof themwith a nearvertical tensionaxis, occur mainly in regionsthat have partially coupledor uncoupledsubductionzonesandtheobservedcontinuousseismicityis deeperthan300 km. The increaseddip of thedowngoingslabassociatedwithweakly coupledsubductionzonesand theweightof theslabmaybe responsiblefor thenearvertical tensionalstressatintermediatedepthand,consequently,the changein focalmechanismfrom type 1 to type2 events.Eventsof type 3occurwherethetrenchaxisbendssharplycausinghorizontal(parallelto thetrenchstrike)extensionalor compressionalintraplate stress.Type 4 are hinge-faulting events. For strongly coupled zones we observed temporal changesofintermediate-depthearthquakeactivity associatedwith the occurrenceof a largeunderthrustingevent. After theoccurrenceof a largeunderthrustingevent, thestressaxis orientationof intermediate-depthearthquakeschangesfromdown-dip tensionalto down-dip compressional(e.g., 1960 Chile, 1974 Peru, 1982 Tonga and 1952 Kamchatkaearthquakes),or the number of large intermediateevents decreasesfor a few years (e.g., 1964 Alaska and 1985Valparaisoearthquakes).We concludethat even though the stresschangesinducedby slabpull and slabdistortioncontrol thegeneralpatternof intermediate-depthseismicity, spatialand temporalvariations of the intraplatestressassociatedwith interplatecoupling are important in controlling the global occurrenceof large intermediate-depthevents.
I. Introduction contactzone betweensubductingand overridingplates being the locus of large underthrusting
In the theory of plate tectonics,subductionof earthquakes.Such large interplate thrust earth-oceaniclithosphereplays a primary role, with the quakes have been extensively studied, and are
81
00 km Trench of the focal mechanismsof recent large inter-mediate-depthevents.We explore the interactionF ~ ________ between the shallow seismically coupled region,/ / wherethrust events occur, andthe deeperuncou-
/ / pled regionin subductionzones.This analysiswas/ Event B motivated in part by the recent suggestionthat/ e. g. 933 Sonriku outer-riseintraplateeventsmay, undersomecon-
977 S~rnbo~a ditions, respond to variations in interplate cou-
Event c pling (ChristensenandRuff, 1983).A comprehen-g. 982 El Salvador sive catalogof focal mechanismsof large(M> 6)
Event D data base. This is used in the developmentofg. 965 Nozco Plote qualitative models relatinggross tectonicproper-
Fig. 1. A schematicfigure showingthe location of the seismic ties to the intraplateearthquakeactivity. Achiev-eventsneara subductionboundary.EventA: thrusteventson .ing an understandingof the relationshipbetweentheinterplateboundary.EventB: trenchandouter-nseevents.EventC: intermediate-deptheventsoccurringin thesubducted interplateand intermediate-depthseismic activityplate. Event D: oceanicintraplateevents. Also indicated are is particularly important for seismic gap regionsexamplesof eventsfor someof thesecategories, and for zonesof uncertainseismicpotential such
as the Juan de Fuca plate, where large inter-mediate-depth earthquakes have occurred re-
basicallywell understoodwithin the contextof the cently.tectonicconvergenceprocess.However,while most We divide this paper into several relativelylargeearthquakesin subductionzonesinvolve in- self-containedsections.Section2 is a generalre-terplatethrusting(eventA in Fig. 1), a significant view of intermediateand deepseismicity. Sectionnumberare intraplateevents within the subduct- 3 concentrateson the spatio-temporaldistributioning lithosphere.Theseintraplaterupturesoccur as of large intermediateand deepfocusearthquakesshallow events near the trench axis (e.g., 1933 that occurredduring this century. Section 4 pre-Sanriku, 1977 Sumba)or in the outer-riseregion sents the catalog of focal mechanismsof recent(e.g., 1981 Chile, event B in Fig. 1), or they occur largeand moderateintermediate-deptheventsbutas deepereventslocateddown-dip from the plate focuseson the relationof the largereventsto theinterfacewithin the subductingslab (e.g., 1982 El interplate seismiccoupling. Section 5 studiestheSalvador,eventC in Fig. 1). Intraplateeventsalso aftershock characteristicsof intermediate-depthoccur well removedfrom subductionzones,in the events with M � 6.5 in the catalog. Section 6interior of oceanicplates(e.g., 1965 Nazca,event shows the stress-axisdistribution of intermediate-D in Fig. 1). In general,our understandingof the depth eventswhereassection 7 is a regionalpre-processproducing the intraplate activity is very sentationof all the eventsin the catalogandtheirlimited, principally becausethe tectoniccontextof relation to the recent large thrust earthquakes.intraplateevents is much more difficult to inter- Section8 presentsthe discussionandconclusions.pret than that of interplateevents.
In order to improve our understandingof thenatureof intermediate-depthseismicity,this paper 2. Reviewpresentsa review of the spatio-temporaldistribu-tion of large(M � 7) intermediateanddeepfocus To provide a context for the analysisof theearthquakesthat occurred during this century. intraplateactivity we will briefly review our un-Given our relative ignorance of the causesof derstandingof shallow interplateearthquakephe-intermediate-depthevents,as well as their hazard nomena.After recognizinga systematicvariationpotential, we also conducta global investigation in the size of interplateearthquakesin the north-
Fig. 2. Location of large(Ms� 7.6) shallowearthquakesfrom 1904 to 1985. M5 is in parenthesesand M~in brackets.Aftershockareasof recentgreatearthquakesare shaded.Note the unevendistribution of great earthquakes(Mw = 8.5), suggestinga globalvariation of interplatecoupling at subductionzones(from Kanamori,1986).
west Pacific, Kanamon (1977b) introduced the In an attempt to understandthe subductionterm seismic coupling to describethe interaction process,several investigators(e.g., Isacks et al.,betweenthe two plates in subductionzones.He 1968;Vlaar andWortel, 1976;Molnar et al., 1979)noticed that in some regions, such as Chile and havestudiedthe interrelationamongthe physicalAlaska, greatearthquakesoccur along the entire andgeometriccharacteristicsof subductionzonestrench, indicating strongcoupling, whereasin the (Table I). For example, Uyeda and KanamoriMarianaIslandsthereis no recordof greatevents, (1979)characterizedsubductionzonesby whetherindicating seismic decoupling.Similarly, Kelleher or not active back-arcspreadingis taking place.et al. (1974) observeda correlationbetweenearth- They constructedan evolutionarymodel basedonquakesize andwidth of the contactzonebetween the degreeof interplatecoupling, where the endthe two plates at shallow depth in subduction membersare the Chile andMananatype subduc-zones. Figure 2 shows the epicentersof earth- tion zones that are respectivelyunder regionalquakeswith M~>8.0 that occurredbetween1904 compressionaland regional tensional tectonicand 1985, with known aftershock areas being stressesat shallow depth. Ruff and Kanamonshaded.M5 is given in parenthesesand M~ in (1983) found a strong correlation between thebrackets. M~ is a measureof seismic energy maximumearthquakesizeobservedin subduction(Kanamori, 1977a), which remains unsaturated zonesand the convergencerate and age of theeven for greatearthquakes.Note the unevendis- subductingplate; great subductionearthquakestribution of great(M~= 8.5) earthquakesin Fig. occur where young oceanic lithosphere is being2 that, together with estimatesof seismic and subductedat high convergencerate,while smalleraseismicslip (Kanamori,1977b; Sykesand Quitt- events are associatedwith old plates with slowmeyer,1981), suggesta globalvariation of seismic convergencerates.The aboveobservationssuggestcoupling at plateboundaries, that the physical characteristicsof the contact
83
zonebetweenthe subductedand overridingplates sphereby relating the orientationof the compres-can be related to the degreeof seismiccoupling. sional and tensional axes of earthquakefocalRecently, Jarrard (1986) presentedan extensive mechanismsto the orientationof theseismic zone.reviewof the relationsamongsubductionparame- Basedon a globalstudyof individual focalmecha-ters, againconcluding that interplatecoupling is nism solutions of deep and intermediate-depthmainly regulated by convergencerate and slab earthquakes,IsacksandMolnar (1969, 1971) con-age. In summary,stronglycoupledplatesproduce cluded that the descendingslab acts like a stress-greatearthquakeslike the 1960 Chile(M~= 9.5), guide, in which compressionalstressesare domi-1964 Alaska (M~= 9.1) and 1952 Kamchatka nantbelow 300 km. andeitherdown-dip tensional(M~= 9.0) earthquakes.Weakly coupled zones or compressionalstressesare observedbetween70do not produce large earthquakes,as is the case and300 km depth.They attributeddown-dip ten-for the Mariana,Bonin andPhilippine arcs.Mod- sional events to extensional stressesinduced byerately coupled zones do produce large earth- the slab’s own negativebuoyancy,but they alsoquakes,but with maximum ruptures <500 km argue that the slab encounters more resistantlong, as is true for the Peru, Middle America, material as it sinks into the mantle so that theAleutian and Kurile trenches(Fig. 2). stressesat intermediatedepth becomecompres-
Most earthquakesoccurringbelow40 km depth sional for plates that extend to 650 km depth.near convergentplate margins are locatedwithin Oike (1971)madesimilar observationsby contour-the subductingplate.Theseeventsform inclined ing pressureandtensionaxesfor eventsin variouszones of seismicity known as Wadati—Benioff subductionzones. Moment tensor inversions ofzones, that presumablydelineatethe location of recentearthquakes(Vassiliou, 1984) also indicatethe colder subductinglithosphereto as deep as that down-dip compressionalstressesare domi-650 km depth (e.g., Isacks et al., 1968; Sleep, nant below 300 km depth,while the behaviorof1973; Richter, 1979). It is generallybelievedthat intermediate-depthevents is much more com-intraplate subductionzoneeventsare infrequent plicated, as discussedbelow.and less important than interplate thrust events Richter(1979) suggestedthat the occurrenceoffor evaluating regional seismic hazard. However, down-dip compressionaleventsbelow 300 km isrecent studiesby severalinvestigators(e.g., Abe, evidenceof theinability of the subductedmaterial1972a, b; Malgrange et al., 1981) indicate that to penetratebelow 700 km depth.Vassiliou et al.intermediate-deptheventsoccurfrequentlyin some (1984) conducted numerical modeling to testsubductionzones, and are often very damaging whether the observed stress pattern within the(e.g., 1965 Puget Sound, Washington, 1973 subductedslab (transition from down-dip tensionOrizaba,Mexico; 1979 Colombiaearthquakes).It to compressionwith increasingdepth)is the resultis clearly a difficult task to appraisethe seismic of the slab encounteringa penetrableviscosityhazardposedby intraplateactivity since we can- increaseor an impenetrablechemicaldiscontinu-not evenreliably assumethat long-termplate mo- ity in the mantle. Either hypothesiscan explaintions will drive repeatedfailure on the samefault; the grosscharacteristicsof the stressorientation.in other wordswe cannotrely on the conceptof Vassiliou et al. (1984) further suggestedthat theseismic gapswhich has beensuccessfulfor fore- observeddecreasein seismic activity between250casting interplatebehavior. However, the intrap- and400 km depthmaybethe resultof the olivinelate events must be controlled by the regional to spinelphasetransitionat 400 km depth(Ander-stress environment, thus many studieshave at- son, 1967) and that the lack of seismicity belowtemptedto extractinformation aboutthe stateof 670 km depthindicatesthat eitherthe slab cannotstress within the subductingplate by analyzing penetratebelow thisdepthor thata changeof theintraplatefocal mechanisms. subducted lithosphere rheology occurs, which
Most analysesof deep andintermediate-depth somehowpreventsbrittle earthquakefailure.earthquakeshave focused on understandingthe The non-uniformstresspatternobservedat in-stress distribution within the subducting litho- termediatedepthsmay berelatedto the existence
84
TABLE I
Characteristicsof subductionzones
Region Subduction Age Dip TAz TLg Depth Plate ConvergenceSP-OP Zone Name Rate Azimuth VN
of double seismic zones, which are observedin and thermo-elastic stresses(House and Jacob,several regions.Well-located hypocentersin the 1982; Molnar et al., 1979; Goto et al., 1985).Kanto district in Japan(Tsumura,1973) demon- Fujita and Kanamori (1981) interpretedpub-strate the existence of a double Benioff zone. lished focal mechanismsolutionsof intermediate-Composite focal mechanismsof small events in depth earthquakesas a function of the slab agethe region indicatethat the upperseismiczone is andconvergencerateas follows. Stressesat inter-underin-platecompressionwhereasthe lowerlayer mediatedepth in old and slow slabs are tensileis underin-plate tensionalstresses(Hasegawaand and induced by the slab’s own weight, whereasUmino, 1978). The two layers of seismicity are those in old and fast slabs are generally mixed,separatedby 35 km andextendin depth from resulting in double seismic zones. Intermediate-60 to 190 km. Double seismic zoneshave been depth stressesin young and slow slabs are alsoobservedbeneathmost of the Japanesearc (e.g., mixed, but insteadexhibit stress-segmentedzones.SuzukiandMotoya, 1978;Yoshii, 1979; Ishikawa, Youngandfast slabsare tensile, specificallyunder1985) andalso in the Kuriles—Kamchatkasubduc- South America, probably owing to continentaltion zone(Sykes,1966; Fedotov,1968;Fedotovet loading.Their resultsrepresentthe first attempttoal., 1971; Veith, 1974,1977;StauderandMaulchin, interpret intermediate-depthearthquakesin the1976). However, double seismic zones are ap- context of the global variation of interplatecou-parently not found in every subduction zone pling at subductionzones,which is predominantly(Fujita and Kanamori, 1981). Some regions ex- controlledby slab ageandconvergencerate(Ruffhibit stress-segmentedseismiczoneswhereregions and Kanamori,1980).of compressionalandtensionaleventsare in close RecentlyYamaokaet al. (1986) studiedscaledproximity, potentially giving the impressionof a physical modelsof the downgoing plate in eachdouble-seismiczonein crosssection,as is the case subductionzone. They showedthat the shapeoffor the central Aleutian arc (Engdahland Scholz, the descendinglithosphere,as inferred from the1977; Fujita and Kanamori,1981). Some of the observedintraplateseismicityin most regions,canmechanismsproposedto explain double seismic beexplainedby modelingthelithosphereby bend-zones include stresses associated with phase ing an inextensible spherical shell, which showschanges(Veith, 1974), unbending of the litho- little deformationundermoderatestress.Regionssphere(Engdahland Scholz,1977; Samowitzand with a poor fit improved greatly by tearing theForsyth, 1981; Kawakatsu,1986a,b), saggingof spherical shell in the region where the lowestthe subductedplate (Yoshii, 1977; Sleep, 1979), seismicity is observed.Their experimentssuggest
NA-North America, Pc-Pacific, Ph-Philippine, Ri-Rivera, SA-SouthAmerica, Sw-Sandwich, So-Solomon)Age of the SP from Sciatter and Parsons, 1981; and Moore, 1982DIP of the SP at intermediate depthTAz Trench Azimuth from Moore, 1982TLg ApproximateTrenchLengthfrom Moore, 1982Depth of maximum extent of continuousobservedseismicity. Note that there are deep eventsin South America but the seismicity
is not continuousPlate Convergence(Rate, Azimuth) calculated from RM2 model of Minster and Jordan, 1978; and Chase,1978.VN convergencevelocity normal to the trench axis
References:1. Barazangi andIsacks, 1976; 2. Frankel and McCann, 1979; 3. Barker, 1972; 4. Molnar and Sykes, 1969, LeFevre andMcNally, 1985; 5. Dewey and Algerniissen, 1974, Dean and Drake, 1978; 6. Frankel et al., 1980; 7. McCann and Sykes, 1984; 8.Cloweset al., 1983; 9. Nishimura et al., 1984; 10. Enghdal ci al., 1977; 11. Ishida, 1970; 12. Shiono; et al., 1980; 13. Seno, 1977; 14.Katsumata and Sykes, 1969; 15. Kanamon and Tsumura, 1971; 16. Cardwell et al., 1980; 17. Fitch, 1970b; 18. Eguchi ci al., 1979;19.Fitch, 1970a;20. McCaffrey et al., 1985, Cardwell and Isacks, 1978; 21. Pascal,1979; 22. Johnsonand Molnar, 1972;23. Pascalet al.,1978; 24. Billington, 1980, Giardini and Woodhouse, 1984; 25. Ansell and Smith, 1975; 26. Reid, 1976.
86
that lateral constraint and bending of the sub- Yet this staticstressregimedoesnot explain allducted lithosphere are important factors in de- intraplateactivity. Mogi (1973) studied the rela-termining the shapeand the strain concentration tionship betweenshallow and deep seismicity inin local regions. Similarly, numerical models of the western Pacific region and concluded thatthe downgoing plate calculatedby Burbachand temporalchangesin the seismicactivity at depthFrohlich (1986)suggestthat the subductinglitho- associatedwith the occurrence of large thrustsphereis remarkablycohesive,only rarely break- events are not fortuitous, but may indicate aning or stretching.Thesestudiesimply that down- integralpart of the earthquakecycle. Mogi’s (1973)dip tensionalstressesdue to slab pull are domi- observationssuggest that the descendinglitho-nant for most subductionzones, with the excep- spheremay act as a viscoelasticbody undergradu-tion of the Tonga trench, where down-dip com- ally increasingload, rather thanas a rigid elasticpressionalstressesare observed(e.g., Isacksand plate (Kawakatsu, 1986a). Seno and PongswatMolnar, 1969, Giardini andWoodhouse,1984). (1981) andKawakatsuand Seno(1983)defineda
Figure 3a shows the static stressdistribution third seismic zone in northern Japan,and sug-nearthe interplateboundaryfor the intermediate- gestedthat the focal mechanismsof the eventsindepthand outer-riseregions.At intermediatede- this zonechangein responseto stresschangesonpththe upperseismiczoneis in-platecompression the interplateboundary.Recently,McNally et al.stress,whereasthe lower seismiczoneis in tension (1986) and González-Ruiz(1986) constructeda(Hasegawaet al., 1979).Note that the stressesdue composite spatio-temporalcross section of shal-to bending of the lithospherein the outer-rise low and intermediate-depthevents that occurredregion are tensionalat shallow depthsand corn- along the Mexican subductionzone, which sug-pressionalat largerdepth(Stauder,1968; Chapple gests that stress transferbetweenthe shallow in-and Forsyth, 1979; Christensenand Ruff, 1987). terplateboundaryandthe intraplateseismicactiv-This simple 2-D model does not considereither ity takes place.Furthermore,temporalvariationslateral changesin the subductionzone such as in the focal mechanismsof intraplateearthquakesbendsin the trenchaxes,contortionor segmenta- in relationto the occurrenceof largethrustearth-tion of the subductedlithosphere,or subduction quakes have been observed in southern Chile,of topographicirregularities (e.g., ridges or oc- both in the outer-rise region (Christensenandeanic platforms), which can presumablychange Ruff, 1983,1987) andat intermediatedepths(Astizthe strengthor orientationof the stressfield in a and Kanamori,1986).givenregion(BurbachandFrohlich, 1986). Figure 3b shows a dynamicmodel for regions
STATIC DYNAMICBefore Thrust
Bending.Detachment ~-~----~
4--,
~
Pull After Thrust
Intermediate—depth ~—..
a b
Fig. 3. (a) A 2-D model of thestressdistributionwithin thedowngoinglithosphereresultingfrom bending,unbendingandtheslab’spull in theouter-riseandintermediate-depthregion.(b) A dynamicstressmodel that is believed to apply to stronglycoupledregionsneartheinterplateboundary.The changein the observedstressdistribution is relatedto displacementof the interplateboundaryaftera largethrustevent.
87
in which, before a major underthrustingearth- moderateandlarge(mb � 6.0) intermediate-depthquake, the interplate boundary is strongly cou- (40—200km) earthquakesthat occurredfrom 1960pled, the down-dip slab is under tensionat inter- to 1984 in relation to both static conditionsandmediatedepths,andthe outer-riseregionis under local variationsof the strengthof interplatecou-compression.After displacementat the interplate pling in a region. Temporal changesof focalboundary, the outer-riseregion is under tension, mechanismsof intermediate-deptheventsowingwhereasthe down-dip slab may be either under to large thrusteventsare examined.We study thecompression or under diminished tensional spatial and temporal variations for different re-stresses.Examples of observationsthat support gionsandinterpretthem within the contextof thethis model are the temporal variation of focal two modelsin Fig. 3.mechanismsat intermediatedepths in southernChile with respect to the occurrenceof the 1960Chile (M~= 9.5) earthquake(Astiz andKanamo- 3. Large intennediate and deep focus earthquakes:ri, 1986) and the temporal variations of focal 1904—1984mechanismsin the outer-rise region in stronglycoupled zones documentedby Christensenand To provide a general context for a detailedRuff (1987). Recentelastic lithospheremodelsby investigationof recent intermediate-depthearth-Dmowska et al. (1987) suggest that the stress quakes,we first considerthe historic recordof thechangewithin an earthquakecycle in theouter-rise largesteventswithin the subductingslabs.Figureand intermediate-depthregionswould be similar 4 showsthe world-wide distribution of large(M �
to thoseshownin Fig. 3b. 7) intermediateand deep focus earthquakesthatIn this study,we examinethe historic recordof occurredfrom 1904 to 1984. Most intermediate
large (mB> 7.0) intermediate and deep focus and deep focusearthquakesare locatedat majorevents as well as the focal mechanismsof 335 active plate boundaries,with only a few events
I ~5i~ ‘~%‘~..J~~J—L))!~~II lii III iiJ~i II) 1)1))! iTh.)
90 120 150 180 -ISO -120 -90 —60 -30 0 30 60
Fig. 4. Globaldistributionof large (M � 7) intermediateanddeep focus earthquakesthat occurredbetween1904 and1984. Symbolsizesareproportionalto theearthquakemagnitude,which are mB for eventsthat occurredbefore 1975 and Mw for later events.
Boxesencompassregions(TableI) in thedepth—magnitudeprofiles in Appendix I.
88
apparentlybeing unrelatedto presentlysubduct- duction process(e.g., the shallow interplateearth-ing plates.Examplesof thelatter include the 1954 quake cycle) or to global changessuch as thedeep Spanishevent (Hodgson and Cook, 1956; Chandlerwobble (Abe and Kanamori,1979).Richter,1958; ChungandKanamori,1976) or the Appendix 1 shows the location of large(mb �
1915 event located offshore western Australia, 7.0) intermediateand deepfocusearthquakesto-which may be mislocated.Theseevents are listed getherwith large (M~� 7.5) shallow events in aby regionin TableA (Appendix 1). It is believed projection along strike with respect to time forthat this catalogis relativelycompletefor m
11 � 7 eachregionin Fig.4. Although no clearrelation isand depth greaterthan 70 km, from 1904 to the observedbetween the shallow interplate activitypresent. Appendix 1 shows the depth and time and theintermediate-deptheventsin mostregions,distribution for each region in Fig. 4. sequencesof large intermediateand deep focus
The most active regionat intermediatedepths events tend to occur at periods when no largeis the New Hebrides(35). Tonga(36) is the most earthquakesoccur at the interplateboundary(seeactive below 400 km depth,but also hasrelatively figuresin Appendix1).abundant intermediate-depth seismic activity. Vassiliouet al. (1984)determinedthe variationRegionswith largenumbersof intermediate-depth of the numberof mb � 5 eventsbetween1964andearthquakeswith mB� 7.0 are the Altiplano (4), 1980 from the NOAA catalog.They observedthatTimor (30), Sulawesi (25), Scotia (8), and New for the global seismicity, the numberof intraplateIreland (33), all of which havesubstantialdistor- eventsdecreaseswith depthexponentiallyto — 300tions of the subductedlithosphere.Otherregions km, then remainsconstantto 500 km depth,with a large number of large intermediate-depth followed by an increaseto 600 km depth.Theyeventsare North Chile (5), Kuriles (17), Ryukyu assumedthat this distribution reflects the level of(21), and the Philippines (24). The Nankai and stress within the plate with increasing depth.Rivera regions, where the subductinglithosphere Giardini (1987) determinedthe b-valuefor deepis young and the observedseismicityis all shal- seismic regions and found large regional varia-lower than 100 km, do not have large intraplate tions, from 0.33 in South America to 1.22 inearthquakes.Someregionssuchas Luzon (23)and Tonga. The b-value for most regions is — 0.6.New Zealand(38) hadvery few intraplateevents ChungandKanamori(1980) investigatedthe van-in this century. ation of earthquakesourceparameterswith depth
By comparing the time distribution of large andconcludedthat the stressdrop of intermediateearthquakesin this centurywith the time distribu- anddeepeventsrangesfrom 20 bar to — 4.6 kbar,tion of theseeventsafter1960 (seeAppendix 1 for with a generalincreasewith depth;however,eventsdetails),we cangetan ideaof how well seismicity close to bendsin the subductinglithospherehaveafter 1960 representsthe intraplate activity of higher stress drops than events at comparabledifferent subduction zones. Regions where depthlocatedin undistortedsubductingslabs.Theseismicity within the subductingplate has been exact relation betweenabsolutestresswithin therelatively constant in the last 80 years are the plateandearthquakestressdrop is unclear.How-Altiplano (4), North Chile (5), Kuriles (17), ever, sincemost of the principal seismicenergyisNortheastJapan(18), Timor (30), New Hebrides releasedin large earthquakes,the profiles shown(35), Tonga(36) and the Hindu—Kush(39). How- in Appendix 1 should reflect the seismic energyever, seismic activity in some other regionshas releasedwithin the platefor eachsubductionzonechanged during this century. For example the during this century.Ryukyu (21) and Sulawesi(23) regionswereespe- We divide the largeintraplateearthquakesintocially active in the early part of the century, three depth ranges(40 � d � 260; 260<d � 400;whereasotherregionssuchas Peru(3), Philippines and d> 400 km) that roughly correspondto dif-(24) andNew Guinea(31) seemto be moreactive ferent stress regimeswithin the downgoing slabrecently. Thesetemporalchangesin activity may (Vassiliou et al., 1984), and assign a maximumbe relatedeither to changesin the regional sub- observed earthquakemagnitude, m~,to each
89
TABLE II
Largestearthquakemagnitudeby seismicdepth
Subduction Maximum Observed Magnitude SeismicZone Name M~ m
M~ are mostlytakenfrom Ruff and Kanamori(1980); valuesin parenthesesareM~from Abe (1989).~values arebody-wavemagnitudeat T 8s(Table A)
90
depthrange.Observethat the maximum mB de- suggeststhat the ageand depthof the slab alonecreaseswith depthfor mostregions(TableII) with do not play a major role in determiningthe maxi-the exceptionof the Sunda (28), Java(29) and mum size of intraplateearthquakes.The correla-Tonga(36) regions,for which the observedmagni- tion coefficientbetweenseismicityandtheconver-tude mB increasesfor the deepestrange,and the gence nate at the trench increaseswith depth,Philippines (24) for which the middle range (260 suggestingthat the rate of influx of subducted<d� 400 km) has the largestobservedmB. The material controls to someextent the strengthbe-deepestrange (d> 400 km) correspondsto the tween compressionalstresswithin the plate andregion with down-dip compressionalevents (e.g., ultimately the size of intraplate earthquakesatIsacks and Molnar, 1971). In the middle depth depth. Finally, the correlation coefficient of therange, down-dip tensional events predominate slab dip and the seismicity is small at the shal-(Vassiliou et al., 1984)with theexceptionof Tonga. lowerdepthrange(40—260km), high for the depthIn the shallowestrange(40—260km) many events range 260—400 km, and then it decreasesat thehavenearlydown-dip tensionalaxes;however,the deepestrange.If we considerthat the slab couldfocal mechanismsof these eventsexhibit a large be modeled by a bendingplate, then the torquevariation as shown in the following sections.The exertedby the slab’s weight is larger for a shal-aboveobservationssuggestthat the occurrenceof lowly dipping slab than it is for steeplydippinglargeintraplateeventsmay be influencednot only plate. Then the large negative value of theby the slab’s rheologybut also by other factors dip—seismicitycorrelationcoefficient for the mid-suchas lateral thermalvariationsin the surround- dle depth rangesuggeststhat the size of earth-ing mantle. quakesat this range may be controlled by the
In an attemptto understandthe regionalvaria- slab’s bendingmoment.tion in maximum earthquakesize for intraplateevents,we testthe correlation betweenmaximumobserved mB of the intraplate events (Table II) 4. Catalog of focal mechanisms of recent inter-and the variousphysicalcharacteristicsof subduc- mediate-depthearthquakestion zones (Table I) for the three depth rangesmentioned aboveas well as the classicaldefini- There is evidence that subducted plates aretionsof intermediate(70—300km) anddeepevents uncoupledfrom the overridingplatesbelow — 40(d> 300 km). The linear correlation coefficients, km depth,wherea steepeningof the Benioff zonewhich vary between0 and ±1 for increasingcor- is generally observed(e.g., Engdahl, 1977; Isacksrelation,are given in TableIII. andBarazangi,1977; Billington, 1980). This bend
The correlationsbetweenmaximum observed in the subductingplate has been related to theseismicity and plate age are small for all depth phasechangein the oceaniccrust from basalt toranges.Similarly, the correlationsbetweenthe de- eclogite that initiates at — 35 km depth,and re-pth of the continuousseismicity and the maxi- sults in an increasein the averagedensityof themum size of intraplate events are small. This slab (Wortel, 1982; Pennington, 1983; Ruff and
TABLE III constrainedto depths <40 km (e.g., Schwartz
Linearcorrelationcoefficients andRuff, 1987).Thus,mostearthquakes> 40 kmdepthare believedto occur within the subducting
Seismicity Age Depth Rate Dip . . .plate, since it is no longer in contact with the40� d � 260 0.12 0.08 0.17 —0.15 overriding plate.An analysis of intraplateevents
260<d � 400 —0.30 —0.14 0.52 —0.70 .
d> 400 km —0.38 —0.35 0.55 —0.44 in the depthrange 40—200km shouldprovideuswith a better understandingof the interaction
70� d � 300 0.22 0.15 0.05 0.14 betweenthe stateof stress in the deeperuncou-d> 300km —0.37 —0.23 0.52 —0.51 -pled region and the degreeof coupling of the
91
interplateboundaryat subductionzones. Such a catalogwerevery small, and reliablefocal mecha-studyrequiresa detailedcatalogof eventsthat are nism could not be determinedusing WWSSNreliably identifiedas being intraplate. seismograms.Most of theseeventsoccurredin the
We gathereda catalogof intermediate-depth early 1960sand their magnitudesmay havebeenearthquakes.First, all earthquakeswith mB � 6.0 overestimated.Theseeventswere removedfromthat occurredfrom 1960 to 1984 between40 and the catalogas well. After timing depth phasesin200 km depthwereidentifiedby sortingtheNOAA short-periodWWSSN records,we found that 40and ISC (International SeismicCentre) catalogs, events were clearly shallow, and thesewere notThen,a searchfor publishedfocal mechanismsfor included in the intermediate-depthearthquakethese earthquakeswas undertaken.Some events catalogeither. Someof theseeventsare outer-risehave been studied by several groups of re- events, for which a comprehensivecatalog hassearchers.However, only the focal mechanism been compiled by Christensenand Ruff (1987).solution that was consideredthe most reliableis Theseeventswere eliminated from our catalog.listed in Table IV. Reliability was determinedon Finally, we determined40 new focal mechanismthe basisof the most recent study(that generally solutionsof intermediate-depthearthquakes,whichreferredto earlierwork) or on the basis of studies are shown in Appendix 2. First-motiondatawerein which the authorspicked the first-motion and insufficient to constrainboth fault planesfor halfS-polarizationdata themselves.Although there is of theseevents,but we determinedthe depthanda considerablevariation in the confidence at- oneof the focal mechanismparametersby model-tachedto individual focalmechanisms,thegeneral ing long-periodbody waves. Only a few criticallycharacteristicsof the solutionsare reliable. If the oriented stationswere modeledfor eachof thesestudiesincluded epicentralrelocation thenwe list events,but generally this providedadequatereso-thesehypocentersin TableIV insteadof the loca- lution of the source depth and mechanism.Thetion reportedby NOAA. In surveying the litera- asterisk(*) in TableIV indicateswhich parameterture, numerousintermediate-depthevents given was constrainedby the synthetic seismogramsordepths <40 km or missing in the NOAA andISC by timing of pP—P phasesin short periodrecordslistings were identified and includedin Table IV. to determinethe eventdepth.Published focal mechanism solutions were not The final catalog, listed in Table IV, containsavailablefor nearly140 of the earthquakesin the 335 hypocenters,magnitudesandfocal mechanisminitial list, solutionsof earthquakesthat occurredbetween40
A preliminary focal mechanismsolution was and200 km depth.Eventnumbersarechronologi-determinedfrom first-motion datain the ISC bul- cal, hypocentersare given in degrees(latitudeletin for thoseeventslackinga publishedsolution. north and longitude east are positive), depth isMost of these mechanismswere not very well given in km. The magnitude M listed is mh forconstrained.If the depth of the eventwas shal- mosteventsthat occurredbefore 1977 except forlower than 60 km, with no depth phasesbeing the largereventsfor which mB from Abe (1981) isreported in the ISC bulletin, and if the pre- given. For events that occurredafter 1977, M~liminary focal mechanismsolution was relatively was determinedfrom Kanamori’s (1977a) rela-reliable and consistentwith an interplate thrust tion: M~= (2/3)/logM0 — 10.73. The magni-orientation in the direction of local subduction tudesgiven in TableIV are intendedto reflectthe(Table I), we assumedthat the eventwasprobably real size of the event, for thereis a large scatterinterplate, and it was removed from the initial betweenmb and eitherM~or mB for the inter-catalog.In this way, we eliminated — 30 events, mediate-deptheventsin TableIV, indicating thatThen we picked first-motion data from short and mb is not a veryreliablemeasureof large(M � 7)
long-period records of the World-Wide Seismo- intermediate-depthevents. RelationsbetweenmBgraph StandardNetwork (WWSSN) for over a and seismic moment, M0, or M~ are given byhundredevents to determinebetter constrained Kanamori (1983). The fault parameters,azimuthfocal mechanisms.Recordingsfor 25 eventsof the (4), dip (6), and rake (X), of oneof the possible
92
TABLE IV
Catalogof intermediate-depthearthquakes
No. Date Lat. Long. Depth M Fault Plane P T B Ref.
N+,S- E+,w- ~/j° ~5O >,,O Az° P1° Az 0 P1° Az ° P1°
Stauder,1972;S73 - Stauder,1973; S75 - Stauder,1975; SB - Stauder and Bollinger, 1966; SMa - Stauderand Maulchin, 1976; 5 & al
- Stein, Engeln, Wiens,Speedand Fujita, 1982.
90 120 50 180 -Ii —120 -90 -60 —30 I 30 60
60— I I I I I I I ~ ~ I I ‘‘~~
- 2 78 -
40- 393 19 .1 -40
20:280 ~ 0 ~
-60 - 263 __~__~_~__//VI I~~l 11111 111111 IJ~I ~ Ii ~I I
90 120 50 180 -ISi -Ill -90 -60 -30 0 30 60
Fig. 5. Focal mechanismsof type 1 eventsthat indicate normal faulting within the subducted plate and strike subparallel to thetrench axis. The eventsshown are large (M
5> 6.8) intermediate-depth earthquakes that occurred from 1960 to 1984. The mechanismdiagramsarelower hemisphereequalareaprojectionsthat areproportional in sizeto theearthquake magnitudes.Dark quadrants arecompressional.Numbersidentify the eventsin TableIV.
101
90 120 ISO 180 -ISO -120 -90 -60 -30 0 30 60
60 1~’ I I 1111’ I I • I I I I I I I~:~~.21~ I I I
16~40-’~-~ 99 .~ -40
20- 244 291 • 20
-20 // 25 18111 ~ .~ ~ i96O-1~~/t-4:
-60 - 121 8~ ~ - -60I Ir—I..——’”f’—..l’———....l,———...l———.l..... I I I . I I I 1.1. I I I I I I I I I
90 120 150 180 -150 —120 -90 —60 -30 0 30 60
Fig. 6. Focal mechanismsof type 2 eventsthat showreversefaultingwithin the subductedplate. Thestrike of theseeventsis also
subparallelto the trench axis. Numbers identify the eventsin Table IV. Dark quadrantsarecompressional.The diagramsare lower
hemisphereequalarea projections that have sizesproportionalto the magnitudeof theearthquakes.
90 120 150 80 -150 -120 -90 -60 -30 0 30 60
60... I I ~2 I I I I I I I I I I I ~ ~ I I I I’ç~~—~’I
- 15
40 7 ~196 184 1 .~ 40
20. ~8 20
- 156
j:: ~ j~33o 36 1960-1:84
256—60 — —60
lc_’fl’’J~8—Ll I I I I I I I I I I Ls4’I I I I I I I I I Y~i
90 120 150 180 -150 -120 -90 -60 -30 0 30 60
Fig. 7. Focalmechanismsof type 3 and 4 eventsareshown.Theseeventsindicate hinge faultingor contorsionswithin thesubductedslab and tear faulting, respectively. Compressionalquadrants are darkened. The mechanismdiagrams are plotted in a lowerhemisphereequalareaprojection. Their sizeis proportional to the earthquake’smagnitude.Numbers identify the eventsin TableIV.
102
fault planesare given as well as the pressure(F), Kanamori, 1980; Wesnouskyet al., 1986) thattension (T) and intermediate (B) stress axes. suggestmoderateinterplatecoupling, althoughnoAzimuth is measuredclockwise from north. Ab- singleearthquakewith M> 7.7 hasoccurredthere.breviations are referencesthat are listed at the Events 122, 209, and237 in the Tongatrenchmaybottom of Table IV. be related to subductionof the Louisville ridge,
This catalog probably includes most inter- which may causea localizedincreasein interplatemediate-depthintraplate earthquakeswith M � coupling, in an otherwiseuncoupledregion.Note6.5 that occurredwithin the subductingslab be- thatevent209 is down-diptensionalandthelargesttween 1960 and 1984. However, some inter- event(M~= 8.0) in the catalog.This eventmaymediate-deptheventsthat wereoriginally listedby havefracturedanddetachedthe subductedslabatNOAA and ISC as shallow may be missing,par- the leading edge of the coupled interplateticularly if they havecompressionalmechanisms, boundary (Given and Kanamori, 1980). EventsAlso, a few of the eventslisted in the catalogmay 122 and 237, on the other hand, are down-dipactuallybe interplateeventswith slightly anoma- compressionaland deeperthan event209 (Tablebusdepthsandmechanisms.This is a significant IV). Theseevents are discussedin more detailexpansionof the databaseof 210 events in the below.Simple modelsof platecoupling andgeom-sameperiod compiled by Fujita and Kanamori etry suggestthat most type 1 events occur at(1981); 33 of their mechanismsare for eventsprior stronglycoupledplate boundaries,wherea down-to 1960,most of which are basedon ISS reports, dip extensionalstressprevailsin a gently dippingandonly 75 of their othereventsare largeenough plate.Various explanationsfor the occurrenceofto be includedin Table IV. theseevents include (1) bending or unbending
If we consideroniy thelargestof the eventsin stresses,(2) continentalloading and (3) down-dipthis catalogue(M> 6.8), a regionalpatternin the gravitational loading by the leading edge of thetypes of mechanismsemerges,as shown in Figs. plate. The dominant mechanismmay vary from5—7. Theseeventscanbe groupedas follows: region to region, as we will explore in a later
(1) normal-faultevents(44%)with a strike sub- section. One possible interpretation is that atparallelto the trenchaxis (Fig. 5); strongly coupled subductionzones the negative
(2) reverse-faultevents(33%)with a strike sub- buoyancyof the subductingplate tends to con-parallelto the trenchaxis (Fig. 6); centratedown-dip tensionalstressesin the down-
(3) normal or reversefault eventswith a strike going slab near the lower edge of the coupledsignificantly oblique to the trenchaxis (10%, Fig. thrustplane, as shownin Fig. 8a.Thus, the occur-7); rence of type 1 events may be interpreted as
(4) tearfaulting events(13%, Fig. 7). evidenceof strong coupling betweenplates. Al-Fault planesolutionsof type I events(Fig. 5) thoughthe coupling of the Pacific Northwestis
indicate a steeplydipping normal fault, usually being debated,note that the 1965 and1949 Pugetwith the ‘continental’ side down-dropped.These Soundearthquakesare type 1 events,which wouldeventsoccur mainly along the South and Middle suggeststrongcoupling betweenthe Juande FucaAmerica trenches,buttheyalsooccurin thePacific andNorth Americanplates.This studymay pro-Northwest,Alaska, and Kamchatka.All of these vide an additionalclue to the strengthof couplingzonesare stronglyor moderatelycoupledsubduc- in this region.tion zones (Kanamori, 1977b; Uyeda and In contrast,if the interplateboundaryis weaklyKanamoni, 1979; Ruff and Kanamoni, 1983) and coupled, the stressdue to negativeslab buoyancythe focal mechanismsof the intermediate-depth is transferredto even shallower depth, causingeventsare down-dip tensional.However, the type large normal-fault eventsnear the trench whereI events that occur along the Philippine and the curvatureof the plate is largest(Fig. 8b). TheSolomon Island regions(335 and 111) are down- great1933 Sannikuearthquakein northeastJapandip compressional.The SolomonIslandsregionis (Kanamoni, 1970; Mogi, 1973), and the 1977thesite of numerouslargedoubletevents(Lay and Sumbawa,Javaearthquake(GivenandKanamori,
103
young—fast stresses.We will discuss these three events in
O -. detail in the regionaldescriptionslater.
I coupled Events of types 3 and 4 are shown in Fig. 7.100 a Type 3 events (7, 149, 152, 156, 184, 294) occur2 0 - where the trench axis bends sharply, causing0 horizontal extensionalor compressionalintraplatekm old-slow stresses.Theseeventsindicatehinge faulting within
O ~ the subductingslab. Type 4 events(8, 13, 36, 196,100 - 217, 223, 256, 330) include all those that do not
- b fall in the abovecategories.Theseeventsoccur at200 - plate boundarieswith complex featuresand maykm be relatedto tearfaulting.
0partially coupled
100- C 5. Aftershock characteristics for intermediate-
200 depth eventskm
Fig. 8. A schematicfigure showing(a) the possiblemechanism Intermediate-depthearthquakestend to haveof type1 eventsata stronglycoupledplate boundaryand (c)type 2 events at a moderatelycoupledboundary. Figure 8b far fewer aftershocksthan shallow subductionshows the casewhere the interplate boundary is decoupled, eventswith similar magnitudes.We determinetheallowing slab pull stressesto producelarge normal faulting numberof aftershocksof the eventswith M � 6.5eventsat shallow depthsbelowthetrench, in Table V, reported by the NOAA and ISC
catalogs,within one week after the main event.About 48% of the eventshad no aftershocksand
1980; Silver and Jordan, 1983; Spence, 1986), 37% of the eventshadbetween1 and5 aftershocks.which occurred at weakly coupled subduction If we considera one-monthperiod, thenumberofzones,are examplesof this type. aftershockson various characteristicsof the sub-
Type 2 eventshavereverse-faultsolutionsand ducting slabs. Little correlation with slab age,strike subparallelto the subductionzone(Fig. 6). maximum depthof seismicityor convergencerateThe events that occur in the Philippines, north- was found.west Solomon Islands, New Hebrides, and Figure 9 shows the variation of the numberofKermadecregionsshow nearvertical tensionand one-week aftershocks, N, with the mainshockhorizontalcompressionaxes.Notethat all of these magnitudefor eventswith M> 6.5 in Table IV.regionsare consideredpartially coupledor uncou- Note that mosteventshaveno aftershocksor onlypled subductionzones(Kanamori, 1977b;Uyeda a few detected. As expected, there is a slightand Kanamori,1979; Ruff and Kanamoni,1983), positivecorrelationbetweenmainshockmagnitudewherethecontinuousseismicityis deeperthan300 and numberof aftershocksoccurring within onekm (see Table I). In terms of our simplemodel, week that could be attributed to detectionthe increaseddip angle of the downgoing slab threshold. However, thereare some eventswithassociatedwith weakly coupledsubductionzones, particularly large numbersof aftershocksshowntogether with the weight of the relatively long on themap in Fig. 9. Opencircles indicateeventssubductedplate,inducesvertical tensionalstresses with N> 5, filled circles N> 10 andstarsN> 25at intermediatedepth, which are responsiblefor aftershocks.Theseeventsare generallylargeinter-the changein focalmechanismfrom type 1 to type mediate-depthevents that are associatedwith2 events(Fig. 8c). Exceptionsare events114 in bendsin the subductedslab, or with moderatelyNew Guinea, 205 in Rumania, and 244 in the coupledor uncoupledregions like the DecemberPhilippines that show down-dip compressional 25, 1969 (event 119, Stein et al., 1982) and the
104
Magnitude6.5 7.0 7.5 8.0
I I , 80
60.
N 40
S
20. S S
S S • •~ S
.1 k~sas.i.tI •.• :.1 .‘ s.~ 0
-~ ~ 6
i i....—J...f’1~”—.i’_k’’_1’’—3. I I I I I I I I I L.i~1’ I I I I I I I
90 12i 50 190 -ISO -120 -90 -60 -30 i 30 90
Fig. 9. The distributionof numberof aftershockswith M> 3.5 that occurredwithin oneweekof an intermediate-deptheventand themagnitude of the mainshock is shown on the top diagram. Mainshocks are from 1960 to 1984. Note that most events have noaftershocksor only a few. Location of events(Table IV) with largernumberof aftershocksare indicatedin the map.Open circlesindicate N> 5, filled circlesN>10, andstarsN> 25 aftershocks.
1972 Izu-Bonin event(168), which hadthe largest spondingdepth.This approachhasbeenusedbynumberof aftershockswithin a week. many investigators(e.g., Oike, 1971; Fujita and
Kanamori, 1981; Vassiliou et al., 1984; BurbachandFrohlich, 1986).Sinceour databasehas been
6. Stress-axisdistribution of intermediate-depth expanded from previousstudieswe usethis sameearthquakes method not only to determinethe general char-
acteristicsof the stress field in each region butIn principle, we can determinethe generalin- also to help usdiscriminate‘anomalousevents’.
traplate stress field orientation for a region by Figure 10 shows equal arealower hemisphereplotting the P and T axesof focal mechanismsof projectionsof the distribution of compression(Peventsin that region in a focal spheretogether = filled symbols)and tension(T = opensymbols)with the downgoingplateorientationat the corre- axes for different subductionzones. Circles are
105
used for events locatedbetween40 and 100 km GreaterAntilles eventshavenearly vertical ten-depthanddiamondsfor events100—200km deep. sion axesconsistentwith theweaklycoupledmodelRegion names are defined in Fig. 4 and open described earlier. In contrast, the tension andarrows indicate the convergencedirection of the compressionaxesof eventsin the LesserAntillesplatesin eachregion(TableI). The curveindicates are alignedwith the trenchaxes.Theseeventsarethe trenchazimuth and dip. The dip angleshown discussedin moredetail in the following section.in Fig. 10 is that of the seismic zoneat 100 km The 1965 PugetSoundevent, that occurredindepth.It is commonto observean abruptincrease the Juande Fucaregion, hasa down-dip tensionalof the seismicitydip angleat 40 km depthandthe stressaxis. The dip anglein this regionis shallowdip of the seismiczone increasesgently to 100 (22°), consistent with a strongly coupled inter-km, remainingapproximatelyconstantthereafter plateboundary.TheNorth Pacific regionsare also(Spence,1987). We definein-plate tensionor com- consideredstronglycoupledregionsandhad onlypressional events on the basis of whether the a few events at intermediatedepth. Alaska hasrespectivestressaxesare locatedwithin 200 of the mostly down-dip tensionalevents,but eventsthatdown-dip slab location. This method ignores occur in the Aleutians and Kamchatka regionsearthquakemagnitudes;but we believe that the have their stress axes widely distributed in thelargereventsusedherereflect the first-order stress focal sphere.distribution in a region, whereassmaller events The Kurile subductionzoneis divided into tworeflect only secondaryor local effects. Inspection sectionsaccording to differencesin the dip andof Fig. 10 indicatesthat most regionshaveeither depthextentof seismicity(see TableI). The stressdominantdown-dip tensionalstressesor a mixed axis distribution is mixed in both Kurile and thepatternof down-dip tensionaland compressional NortheastJapanregions.Thesetwo regionshavestressesat intermediatedepth. Only Tonga has beenvery active. However, only a few eventshavedominantdown-dipcompressionalstresses, occurredwithin the oldest(> 150 Ma) subducting
Most intermediate-depthevents that occur in oceaniclithospherein the Izu-Bonin and MarianaSouthAmericaregions(from Colombia to Central regions.The deepereventsin the Izu-Bonin regionChile) havenearin-plate tensionalaxes.Notealso have in-plate compressionaxes; however, thethat the tensionaxis of all eventsin the Altiplano events with depth < 100 km are mixed. TheandNorth Chile regionsare especiallywell aligned Mariana region was divided into northern andwith the subductedplate and the convergence southernsegmentsbecauseof the abrupt changedirection. Theseresultsagreewith thoseof previ- in the trenchaxis.ous investigators(e.g., Stauder,1973, 1975). The Events in the Ryukyu region have a mixedSouth Chile region has only one down-dip com- stressaxis pattern,but it is alignedwith the direc-pressioneventthat occurredon May 8, 1971.This tion of convergenceof the Eurasiaand Philippineeventhas been discussedin detail by Astiz and plates. The North Taiwan region events haveKanamori (1986) who infer that the greatunder- mostly down-dip tensionalaxeswhich are nearlythrusting event in 1960 temporally induced in- vertical (SenoandKurita, 1978). Thus,this regiontraplatecompressiondown-dip.A largenumberof is consistentwith a rather weakly coupledinter-eventsin the Scotia regionhavenearly down-dip plateboundary.In the subductingplate in Luzontensional axes. This result is consistentwith the only three eventshave occurred,but the stressmodelof weakcoupling at theinterplateboundary axesare distributed randomlywhich may reflectin which nearvertical tensionalstressis induced the complexity of this area. The Philippine sub-as a result of the negativebuoyancyof the slab. duction zone is among the more active seismic
The Mexicoand Central Americaregionshave regions at intermediatedepth (see Appendix 1).mostly shallow-dipping tensional axes,consistent Most events in the northernsectionshow nearlywith the observed seismicity dip angle. In the vertical tension axesthat align with the steeplyCaribbean region only a few intermediate-depth dipping seismic zone. In the southernsegment,eventshaveoccurredbetween1960 and 1984.The however,the tensionaxesof the eventsare mostly
106
Colombia Ecuador Peru Altiplano N. Chile
•• -i-::• •
: - 8 - .. .~ - - ,~. - I
C Chile S Chile Scolia Mexico C AmerIca~
.91. /• - , S S• 4. S
- - - - is:> - • <C1 •.• /- •
G Antilles L Antilles Juan de Fuca ,~Alaska Aleutian
9~
- ,.-. • 9 -
•~
Kamchatka N Kuriles S Kuriles NE Japan Tzu—Bonin
S ~ •~• - cT~. S S
- ..1.9 .~ . •
• s:-. - .‘~ -•~:1
N Mariana S Mariona T P
. d<lOOkm
• • d>lOOkm
Ryukyu N. Taiwan Luzon N. Philippines S. Philippines
• 5’S
S • - •• . , , • I ~ ~n/-’ ~ •
S S I.4~-- I. ,-.~
Sulawesi Burma Andaman a Sunda /7 Java
S • • - S
S .‘
S • CI., ~ •, I 5-,• . S.• %• ~I: I
• 1 •, . •
/7 /7 0Timor /7 New Guinea New ~Britoin New Ireland Solomon
..i•~ ~
- • .s ~- ~- 5, ~-k-9
• I• r — •~ • , x ~‘ ~
• , c7 4 •5 S
-~ .
•4
/7 KermadecN. New Hebrides S. New Hebrides N. Tango S Tonga New Zealand
-9.5 .4 -
•~5 4 4 ~ S 0:5 t -, ~ cT. -~ -
45• sl• ,~•, ~-
55s: .• -• ,• S - •~ S
107
vertical but are not aligned with the subducted zontal compressionalaxes; the tensionalaxes ofplate. the events located below 100 km depth (open
The Sulawesi region is a complex region with diamonds)are aligned with the plate. The stresssparselarge seismicity. In the Burma subduction axis distribution in the southernNew Hebrideszone continentalcollision is occurring; however, segmentis mixed.intermediate-deptheventsmay be locatedwithin The Tongasubductionzone is also very activethe attached oceanic lithosphere. Tension and at intermediatedepths.The area is divided intocompressionaxesare separatedin the diagrambut two regions that separatethe events which oc-are not easily relatedto eitherthe seismiczoneor curred near the northernbend from those in thethe convergencedirection in the region. Events more linear section of the trench. Most of thethatoccurredin theAndamanregionare all <100 southerneventshavedown-dipcompressionalaxeskm depth,but the stressaxisdistribution is mixed, with the exception of a few shallower events,Most of the events in the Sunda trench have which are discussedin detail below. The Kerma-down-diptensionalaxes,which are nearlyvertical, dec—NewZealanddiagramshows a mixed distri-This observationis consistentwith a weakly cou- bution of the stress.pled interplateboundary.The stressaxesof inter-mediate-deptheventsin the Javaregion aremixed.In the Timor subductionregionmostearthquakes 7. Regional examination of intermediate-depthhave nearly down-dip tensional axes that are eventsalignedin the direction perpendicularto the con-vergencedirection of the subductingand overrid- Although the regional stress field of inter-ing plates. mediate-depth events inferred from the focal
The New Guinearegionhasmixed stressaxes, mechanismcatalogin TableIV showsthat a largeMost events shallowerthan 100 km that occur in numberof eventshavedown-dipor nearly verticalthe New Britain regionhavenear vertical tension tension axes,as discussedabove (Fig. 10), manyaxes; however, the tension axes of eventswith regions show a mixed or complex distribution ofdepthsbetween100 and 200 km are nearly hori- the tensionand compressionaxes.zontal,This patternreflectsthe complexity of this We wish to evaluatethesevariationsin greaterregion. The New Irelandregion is small and very detail, to appraisethe relativeimportanceof staticactive (seeAppendix 1), which may be relatedto versusdynamic stressfields. Figure 11 shows thehigh stress levels due to lateral bendingof the location of the intermediate-depthearthquakesSolomon plate beneaththis arc. The stressaxis with M � 6.0, which occurredfrom 1960 to 1984.distribution in the New Ireland region is mixed. Boxes enclosethe regionsshown in Figs. 12—19.Intermediate-depthevents(M � 6) in the Solomon On the regionalmapsthe plate names,bathymet-subductionzoneare shallowerthan 100 km and nc featuresand relativemotion betweenadjacentexhibit a mixed distribution of the stressaxis, platesareindicated,Holocenevolcanoesareshown
The New Hebrides region is the most active as open triangles and filled triangles representareaat intermediatedepth(seeAppendix 1). The volcanoesactive in the last 1000 years (Moore,trench azimuth changesfrom north to south, so 1982). The fault parametersof the eventsin eachthe events havebeendivided accordingly. Most regional figure and the eventnumber,shownnexteventsin the northernsegmenthavenearly hori- to the lower hemisphereprojection of the focal
Fig. 10. Equal area lower hemisphereprojection showing P (filled symbols)and T (open symbols) for eventsin the regionslisted inTable 1.3. The curve indicates the trench azimuth and the dip of the seismic zone at intermediate depth. Open arrows show thedirection of convergenceacrossplate boundaries for each region. Diamonds are symbols for eventsdeeper than 100 km and circlesfor shallower events.
108
90 120 150 180 -150 -120 —90 —60 -30 0 30 60
60 — I I I I I ~ ~ I I I ,I I ~_f~J~l — 60
~ /~Y ~ -,
-60 - /7~. / Fig. 13 ‘ - -80
I I~ I I I I I I I l~ I I I I90 120 50 180 -150 —120 -90 —60 -30 0 30 60
Fig. 11. Locationof intermediate-depthearthquakes(40—200 km) with M � 6.0 that occurredfrom 1960 to 1984. Boxesenclosethe
regionsshownin Figs. 12—19.
sphere,are given in TableIV, Dark quadrantsare (10—20°)beneaththe North America plate (Fig.compressional,and numberswithin parentheses 12). The seismicity extendseastwardto ‘~ 150 kmindicatethe eventdepth. depth. The chain of active volcanoesforms an
In the region by regiondiscussionthat follows anglewith the trenchandis locatedfarther inlandwe describethe characteristicsof the downgoing than in most subductionzones. Large shallowplate as well as those of the interplateboundary. underthrustingearthquakeswith rupture lengthsThen we examineindividual events or groupsof of 100 km andshortrecurrenceintervals,30—80nearby events relative to their location to local years (Singh et al,, 1981) are characteristicof thephysical changesin the subductionzone, for ex- Mexican subductionzone, suggestingstrong toample,bends or tear of the downgoing plate or moderatecoupling along this interplateboundary.subductionof topographichighs.Detailedseismic- Three large intermediate-depthevents occurredity profilesandmodelingby Yamaokaet al. (1986) from 1960 to 1984 down-dip of future large sub-andBurbachandFrohlich(1986) for eachsubdue- duction events, The 1964 event (39) occurredtion zone help us locate the regions where the down-dip of the 1979 Petatlán earthquakeplate is bending or tearing. Then we explore the aftershock area. Similarly, the 1980 Huajuapanrelativelocation, both in spaceand time, of inter- event (248) occurred down-dip of the 1982mediate-depthevents that may be relatedto the Ometepecdoublet, and the 1973 Orizaba earth-occurrenceof large thrusteventsat the interplate quake (178) was located down-dip of the 1978boundary. Oaxacaaftershockzone. Although theseeventsare
consistentwith the dynamicmodel for stronglyor7.1. Middle America moderatelycoupledregionspresentedearlier, no
largecompressiveintraplateeventshaveoccurredNorth of the TehuantepecRidge the Rivera after theselargesubductionearthquakes.This sug-
and Cocos plate subduct with a shallow dip gests that the displacement at the interplate
109
100 80 60
\~ \ I NORTH s:u~- — 248(0w ~ 228~,n , IERICA PLATE
Fig. 12. Location and focalmechanismsfor events with M � 6.0 that occurredin the Middle America subductionzone and in theGreaterand LesserAntilles in the Caribbeanduring the period of 1960—1984.Numbersrefer to eventsin Table IV and thoseinparenthesesindicate eventdepthin km. Plate names,bathymetrichighsand trenchnamesareindicated in thefigure. Opentrianglesarelocationsof Holocenevolcanoesandfilled trianglesarerecentactivevolcanoes(Moore,1982).Openarrowsareconvergencerates(TableI) and doublearrowsindicatestrike-slip motion acrossplateboundaries.The focalspheresarelower hemisphereprojections,with compressionalquadrantsdarkened.Thesizesof the focalmechanismsareproportionalto themagnitude.
boundarymay not be largeenoughto changethe Southof the Tehuantepecridge, the Cocosplatestress characteristicsat depth or that any such subductsbeneaththe Caribbeanplatewith a con-down-dip intraplatecompressionaleventsmay be siderablylargerdip angle, which variesfrom 30°too small to identify. González-Ruiz(1986)docu- beneath Guatemalato 65 ° beneath Nicaragua.mentedthe occurrenceof normal faulting events The seismicityis continuousto a depthof 250 kmdown-dip of the Ometepecregion within a few along Central America, but shallows toward theyearsbefore four consecutivethrusting episodes. southeast,where a shallowerdip is also observed.Hesuggestedthat the occurrenceof normal fault- All eventsin this region are down-dip tensional,ing events at intermediatedepth may be an in- However,events98 and 313 (M 6) havereversetegral part of the earthquakecycle in the Omete- focal mechanismsthat suggest,under the classifi-peeregion. However, the 1931 normal fault event cation of large eventsdiscussedearlier, an uncou-studied by Singh et al, (1985) was located at 45 pled interplateboundary. McNally and Minsterkm depth down-dip of the region brokenin 1928 (1981) observed low seismic slip along Centralby four large shallow thrust events offshore America,indicating eitherweak couplingor longerOaxaca. Singh et al. (1985) suggestedthat this recurrenceintervals along this plate boundary.event may have broken the lithosphere,decou- The largest intermediate-depthevent is the 1982pling the Cocos plate from the overriding con- El Salvador earthquake(event 274), which oc-tinental plate. Events 228 and 286 may be associ- curred down-dip of a region where large eventsated with subductionof the Tehuantepecridge, lastoccurredat the turn of the century(Astiz andwherethearc junction is identified by Yamaokaet Kanamori, 1984). If this region is at least mod-al. (1986). It is not known whetherlargeinterplate eratelycoupled, this eventmay indicatedown-dipeventsoccur where the Tehuantepecridge enters loading preparatoryto a large subductionearth-the trench, quake; however, the mechanismsof the historic
110
eventsare not known, so the seismic potential is smaller magnitudeevents in this region are alsodifficult to evaluate, down-dip tensional,
7.2.2. Scotiaarc7.2. Atlantic island arcs
Tectonicstudiesof the Scotia arc region (Fig.13) indicate that back-arcspreadingis occurring
7.2.1. Caribbean (Barker, 1970, 1972) and that the 70-Ma-old, oc-Subductionof old Atlantic oceaniclithosphere eanic lithosphereof the SouthAmericaplate gets
beneaththe Caribbeanalong the Lesser Antilles younger to the south (<10 Ma) as it subductstrenchoccurs at 2 cm a1 in a westerlydirec- beneaththe oceanic floor of the recently formedlion. However, as the trenchcurves to an E—W Sandwichplate(Frankeland McCann,1979). Thetrend, along the Greater Antilles, the North southernportionof the plate is less activeandtheAmericaandCaribbeanplaterelative motionsbe- focal mechanismof event93 indicatesdown-dipcome very oblique and convergencereduces to compression;however, in the northern segment
0.2 cm a3 (Fig. 12). Severalstudies(e.g., Stein mosteventsare down-dip tensional(see Fig. 13).et al., 1982; Sykeset al,, 1982; Yamaokaet al., The change from predominantly down-dip ten-1986) indicate that the downgoing slab curves sional to down-dip compressionalstressesalongcontinuouslybeneaththe Caribbean plate. Mc- the Scotiaarc was also observedfor focal mecha-Cann and Sykes(1984), on the other hand, Sug- nisms of smaller magnitude events by Forsythgested tearing of the plate basedon the focal (1975), who explains this changeas the result ofmechanismof the 1974 earthquake(M~= ~ reducednegativebuoyancyforces in the younger,event184); however, this event is also consistent southernhalf of the subductedplate. Evidenceofwith in-plate tensional stressescaused by the tearingof the downgoing plate in the northernlateral bending of the lithosphere. The 1969 segmentis given by someearthquakefocal mecha-ChristmasDay earthquake(M~= 7.5, event119) nisms(e.g., events 159, 328; Forsyth, 1975) andalocation nearthe trenchaxis indicateslithospheric gap in seismicityis observedat intermediatedepthfaulting in responseto the slab pull in a weakly (Yamaokaet al., 1986). Events266 and269, whichcoupled region (McCann and Sykes, 1984), in occurredwithin threemonthsof eachother, haveagreementwith the model shown in Fig. 8. Note very different focal mechanismsand may reflectalso that this eventhad a relatively largenumber internal deformation of the subducting plate.of aftershockswith respectto other intermediate- Severalvery large intraplateeventshaveoccurreddepth events. Theseindicate faulting along the in the Scotia arc in this century (Table A). Thesteeply dipping plane(Stein et al., 1982). Event most recent event on May 26, 1964 (36), was106 indicatestear faulting. Severallargeintraplate studied in detail by Abe (1972a), who obtainedearthquakesare locatedalong the LesserAntilles M~= 7.8 (from surface-waveseismic moment ofarc (TableA); however,no largesubductionevents 6.2 X 1027 dyn cm). The 1964 event, which is thehaveoccurredthereduring this century(McCann largest recent event in this region, is consistentandSykes,1984). with the model describedabovefor weakly cou-
In the GreaterAntilles region the 120-Ma-old pled regions.No large(M> 7.5) subductionthrustoceanic plate subducts along the Puerto Rico eventshaveoccurredin this region.trench, At intermediate depth the downgoinglithosphereis nearly vertical (Frankelet al,, 1980).
7.3. SouthAmericaModerateintraplateearthquakesin this area(147,232, 260) are down-dip tensional events, con-sistentwith the model for a weakly coupledre- The Nazcaplate subductseastwardalong thegion. In this model thedip of the slab is steepand curving margin of western South America forthe slab’s weight inducestensionalstressesat in- 5000 km. The major topographicfeaturesbeingtermediate depth. Frankel (1982) reports that subducted along this oceanic—continentalplate
Fig. 13. Epicentersand focal mechanismsof intermediate-depthearthquakeswith M � 6.0 that occurredfrom 1960 to 1984 (TableIV) in South America and the Scotia arc. For symbols, seeFig. 12.
112
boundaryare the Chile rise and the Nazcaand occurredin this region during this century haveCarnegieridges (Fig. 13). The characteristicsof beendiscussedby Astiz andKanamori(1986).Wethe downgoinglithospherevary along the trench, should remark that the down-dip compressionaldefining several segments(Table I, Fig. 2). The event (145) that occurredon May 8, 1971 is theabsenceor presenceof the volcanic chain cor- only earthquakewith M � 6 to haveoccurredtorelateswith dip of the intermediate-depthseismic- date in this region at intermediatedepth. Astizity (BarazangiandIsacks,1976).Although intrap- and Kanamori (1986) interpret this event, alonglate seismicity is high in South America, mac- with the extensivesequenceof outer-risetensionalcurate earthquakehypocenters (owing to poor events following the 1960 rupture (Christensenstation coverage) allow different interpretations and Ruff, 1987) as a temporal change in theof the configurationof the downgoingNazcaplate intraplateStressenvironmentinducedby the inter-(Chin and Isacks,1983). For example,Barazangi plateevent. Large down-dip tensional eventspre-and Isacks (1976) and Yamaoka et al. (1986) ceded the 1960 thrust event in 1934 and 1949,proposea tearin the subductedlithosphere,where indicating that the down-dip edge of the futurethe slab dip changesabruptly from 8° to 28° thrust zone was loadedby tensional stressesin-along the Peru—Altiplanoborderline. In contrast, ducedby slab pull prior to the interplatefailure.HasegawaandSacks(1981) suggestthat the Nazca The remaining 250-km-long section in Southernplate is contortedratherthan torn in this region, Chileto the northof the 1960 rupturewasbrokenbased on observationsfrom local seismic data. by the 1939 (M
5 = 7.8) and the 1928 (M5 = 8.0)The observedcontinuousseismicityis as deep as interplate earthquakes.From 1960 to 1984 no300 km for all regions except Southern Chile; intermediate-depthearthquakeswith M � 6 havehowever,eventsdeeperthan 500 km are observed occurredthere.only in Colombia, Peru and Northern Chile re-gions(Fig. 13). Thesedeepeventsprobably occur 7.3.2. Central Chilewithin a detachedsegmentof the subductedlitho- This region(29°to 34°S)is characterizedby asphere(Isacksand Molnar, 1971; Stauder, 1975). very shallow-dipping slab (100) at intermediateMost intermediate-deptheventsin SouthAmerica depth, the absenceof active volcanismand occur-(from Colombia to Central Chile, Fig. 10) have renceof large underthrustingeventswith rupturenearly down-dip tensional axis induced by the lengthsof 300 km. The northernhalf is definedlithosphere’s negative buoyancy, or possibly in by the aftershock area of the 1906 Valparaisoresponseto continentalloading. (M5 = 8.4) thrust earthquake,which has rerup-
We divide the following discussion into five tured over 75% of its length, in two largesub-sections, corresponding to variations of inter- duction eventsthat occurredon July 9, 1971 (M5mediate-depthseismicityandoccurrenceof recent = 7.5) andMarch 3, 1985 (M~= 7.8). The south-large subduction earthquakes along the em-most segmentappears to be a seismic gapNazca—SouthAmerica interplateboundary. (Korrat and Madariaga,1986).
Stressaxesof intermediate-deptheventsin this7.3.1. South Chile region are aligned with the convergencedirectionSeismicity in this 1250-km-longsectionextends (Fig. 10); however,the tensionaxesof mostevents
to a depth of 160 km with a shallow dip. The (except events 245, 125 and 84) dip at — 30°
shallow activity is dominatedby the aftershocksof instead of 10°, as inferred by the intraplatethe 1960 Chile (M~= 9.5) earthquake.This re- seismicityin the region (Fig. 13). Events 125 andgion, from 340 to 45°S,is consideredto be the 245 (M 6) occurrednear the northernlimit ofend membercaseof a very strongly coupledsub- the 1943 aftershockzone.Their focal mechanismsduction zone(UyedaandKanamori, 1979). Rela- indicateinternaldeformationof the plateandslablively young,0—35 Ma, oceaniclithosphereis being segmentationassociatedwith the increasein platesubductedat — 9.5 cm a~.Large intraplate(in- dip (30°)to the north,as proposedby Isacksandtermediate-depthand outer-rise)earthquakesthat Barazangi(1977). Event 9 (mh = 6.3) occurredin
113
1963, down-dip of the 1943 thrust event rupture throughoutthis century(Appendix 1). All of thearea, suggestingthat by this time the interplate eventsin Fig. 13, from event278 to the northernboundarywas at least moderatelycoupled, and boundaryof the Altiplano region at event220 tonormal faulting eventscould occur at the baseof the south boundaryof the North Chile region,the coupledzone,Furtherinvestigationis required have down-dip tensional axes. The compressionto determinewhethera changein the stressfield axesare orientedperpendicularto the downgoingorientationof intermediate-deptheventsoccurred platewith a nearlyvertical plunge,suggestingthatdown-dipof the 1943 rupturearea, continental loading may play an important role
Before the occurrenceof the recent 1971 and in the stress distribution at intermediatedepth,1985 Valparaisoearthquakes,two normal faulting given that the Andes are much broader in theeventsoccurreddown-dipat intermediatedepthin Altiplano—North Chile region. Note also the1965 (55) andin 1983 (315), respectively,in agree- azimuthalvariation of the tensionaxis along thement with the dynamicmodel for coupledregions gently bending Nazcaplate, from — N40°E for(Fig. 3b). Malgrangeet al. (1981)studiedthe 1965 event278 to — N120°E for event220. The focaland 1971 Aconcaguaevents in detail. In 1981, a mechanismsof intermediate-depthevents in thislarge normal faulting event (263, M~= 7.0) oc- region also agreewith the model for a coupledcurreddown-dip of the 1971 aftershockzone. Al- interplateboundary,since normal faulting eventsthoughthe occurrenceof smaller (M < 6) down- occurat thebaseof the coupledregion in responsedip compressionaleventscannotbe ruled out, the to down-dippull of the subductinglithosphere.occurrenceof the 1981 event (263) suggeststhat Christensenand Ruff (1987) report two com-displacementat the intraplate boundarywas not pressional outer-rise events offshore the 1922largeenougj-i to changethe stressfield orientation aftershockarea(260 to 29°S)in 1964 and 1969.at depth.Note also that in 1983 a compressional The earlier event was followed by a M
5 = 7.4outer-riseeventoccurredoffshore the future 1985 thrust event that occurredon October 4, 1983.Valparaisoearthquakerupturezone(Christensen Malgrangeand Madariaga(1983)documentedtheand Ruff, 1983).Thus, this regionagreeswith the variation of focal mechanismsof moderateinter-dynamicmodel shownin Fig. 3b in that an outer- mediate-depthevents associatedwith the largerisecompressionaleventanda down-dip tensional 1966 (M~= 7.7) thrust event (250). A largeeventoccurredbefore the thrustevent, However, down-dip tensionalevent precededthe thrust onneithera largetensionalouter-risenor a down-dip February 23, 1965 (51), and a small down-dipcompressionalintermediate-depthevent has oc- compressionalaftershockoccurredat the northerncurredto date. Onepossibilityis that the displace- end of the rupturezoneon June 21, 1967. Theyment at the interplateboundarymay notbe large interpretedtheir observationsas possibleevidenceenough to changethe stressorientation,but may of the presenceof a double seismic zone in thebe sufficient to decreasetemporarilythe tensional region. An alternative interpretationis that theseintraplateseismicity in the region down-dip from eventsreflect the temporal changein coupling atthe largethrustingearthquake. the thrustboundaryas shown in Fig. 3b.
ing gently along the coastsof southernPeru and to — 16° S1 is boundedto the north and southbynorthernChile. Note that the volcanic chainreap- the Carnegie and Nazca ridges (Fig. 13). Thispears (Fig. 13) and that since at least 1922 no region, albeit complex, offers a good exampleoflarge thrustearthquakeshaveoccurredalong the the relationship between the coupled interplateentirelength of the interplateboundary(Kelleher boundaryand intraplateseismic activity. Seismic-et al., 1974; McCann et al., 1979). Theseregions ity along this region dips at — 300 at shallowhave beenvery active at intermediatedepth, not depth,while at — 100 km depthit becomesalmostonly during the 1960—1980 time period but horizontal, extending laterally almost 300 km
114
(Barazangi and Isacks, 1979). The shallow dip alternateinterpretationof continuousslab con-(~10°) of the subductinglithosphere at inter- tortion in this region.Also, no largeevents(M � 6)mediatedepthhasbeenassociatedwith thelack of haveoccurredin this sametime perioddown-dipvolcanism(Fig. 13) by BarazangiandIsacks(1976). of the 1940 and 1942 aftershockzones.Two regionscan be defined in terms of the ab- Before the large thrust earthquakeof Octobersence or presenceof large thrust events at the 1974 (M~= 8.1) two down-dip tensional eventsinterplateboundary.The northernsegmenthas no occurred.In 1968 (107) and in January1974 (183,known historic large subductionevents and ex- mh= 6.6) both eventswere located down-dip oftendsfrom 0°to 10°S(McCannet al,, 1979).The the future1974 thrustrupturezone. In 1982 (270),oceaniclithospherein this region is heterogeneous a down-dip compressionalaftershocktook placewith the Carnegieridge subductingfrom 0° to near the earlier 1968 tensional event. Theseob-30 s, and numerousfracturesare presentfarther servationsagreewith the dynamic model showninsouth, seawardfrom the trench. The secondme- Fig. 3b. However, no large outer-rise eventsoc-gion, extending from 10° to 16°S, has active curredeitherbefore or after the large 1974 thrustintraplatethrustactivity, earthquake(Christensenand Ruff, 1987). Down-
A clusterof events (247, 67, 28, 146) at the dip of the 1966 (M~= 8.1) earthquakeaftershocknorthern edgeof the subductingCarnegieridge areatwo down-dip tensionaleventswere recordedindicatestear faulting, with one nodal planeal- in 1963 (20, 21), and no large intraplateactivitymost perpendicularto the trench that may be has occurredsincethe 1966 event.associatedwith subductionof thepresumablymore The largestintraplateearthquakein this regionbuoyantlithosphere(BarazangiandIsacks, 1976). occurredon May 31, 1970 (128, M~= 7.8), andEvents 151, 46, 188, 296 and 12 are down-dip this eventhasbeenstudiedby severalinvestigatorstensional,consistentwith a stronglycoupledinter- (Abe, 1972b; Stauder,1975). Thiseventis unusualplate. Christensen and Ruff (1987) reported in many ways. Forinstance,it is locatedbetweenouter-rise compressionalevents in this region. the regions of active thrusting and apparentEvent 316 (M 6) may be associatedwith inter- aseismic subduction along Peru. Also, a largenal deformationof theNazcaplate, resultingfrom numberof small aftershocks(seeFig. 10) followedthe presenceof the Carnegieridge to the north. this event, defining a steeplydipping fault planeThe compressionaxes of event 276 are almost that perhapsbroke the entire subducting litho-horizontal, suggesting that high compressive sphere(Abe, 1972b). Someof theseeventshavestressesinducedby the subductingridge may be normal faulting mechanismswhile othersexhibitdominantto — 100 km depth.However,this event compressionalmechanisms(Dewey and Spence,could be explainedas a result of bendingof the 1979). Christensenand Ruff (1987) reported thatdowngoingslab as it flattensat intermediatedepth nearthe northernedgeof the 1966 thtust eventan(seefig. 8 in IsacksandBarazangi,1977). outer-rise compressionalevent occurredin Sep-
Thesouthernsegment,from 10° to — 16°S,is tember1967. Note also that two normal-faultingcharacterizedby large thrust eventswith rupture events (78, 105) occurreddown-dip prior to thelengthsof — 150 km. The most recentsequence large 1970 normal-fault earthquake.The occur-occurredfrom south to north in 1942, 1974, 1940 rence of these events, along with the tensionaland 1966 (e.g., McCann et al., 1979). Along the 1970 (Mw = 7.8)earthquake,mayreflect the pres-southernboundaryof this region several studies ence of a strongly coupledinterplate boundary.have proposeda major tear of the subducting Abe (1972b)suggeststhat the 1970 eventoccurredlithosphere(IsacksandBarazangi,1977;Yamaoka in responseto the gravitationalpull exertedby theet al., 1986) to accommodatethe sharpincreasein densersinking slab.the dip angleof thedowngoingplate farthersouth.Note that from 1960 to 1984 no large tearing 7.3.5. Colombiaeventsoccurredin this regionat intermediatede- The stronglycoupledinterplateboundarynorthpth, which favors Hasegawaand Sacks’ (1981) of the Carnegieridge, from 5°Nto 0°,hasvery
115
young oceanic lithospheresubductingalong the North Americaplate (Fig. 14). However, seismic-shallow Colombia trench at — 8 cm a~.The ity at intermediatedepth defines a plate dippingseismicity dip is small at shallow depth and in- at — 22° extending to 100 km depth (Crosson,creasesto 35°at intermediatedepth.The volcanic 1980). Heaton and Kanamori (1984) infer achain parallels the coast with numerousactive strongly coupled interplate boundary along thevolcanoes(Fig. 13). Theentirelength of the earlier Juande Fuca plate. The occurrenceof the 19651906 Colombia (M~= 8.7) earthquakehas been (57, mb= 6.9) and 1949 (mh= 7.1) PugetSoundrerupturedby the recent thrustevents,which pro- earthquakes,both down-dip tensional normal-gressedfrom south to north, in 1942, 1958 and faulting events (Langstonand Blum, 1977) at in-1979 (Kelleher, 1972; Kanamori and McNally, termediatedepth,also suggesta stronglycoupled1982; Beck and Ruff, 1984). interplateboundaryaccordingto the classification
Events 109, 179 and94 are locatedin northern in Fig. 8.Colombia where the trench is not well defined.These eventsare consistent with down-dip ten- 7.4.2. Alaska—Aleutianssional stressesthat are probably inducedby the Convergencebetweenthe North Americanandnegativebuoyancyof the sinking slab, No large Pacificplates takesplacealong the Alaska—Aleu-(M � 6) intermediate-depthevents are located tian trenchin a northwestdirection (Fig. 14). Thedown-dip of the 1942 and 1958 aftershockzones oceanicplate becomesprogressivelyolder to theduring the period of 1960 to 1984. However,near west along the trench, ranging from 40 to 65 Ma.the northern edge of the 1979 aftershockzone Seismicity definesthe downgoing slab to a maxi-severaleventshaveoccurred,The shallowestevent mum depth of 280 km underneaththe Aleutian(174),which took placein 1973, is locatednearthe Arc. The dip of the subductinglithosphereshal-trenchaxis, It indicatesouter-risetensional stress lows progressivelyto the east from 65° under-in the segmentadjacentto the 1979 rupture,per- neaththe Aleutian Islandsto — 250 beneaththehaps indicative of lateral loading of the locked Cook Inlet, whereseismicity extendsto only 150zoneby aseismicslip to the north. Down-dip of km depth. Underthe Rat Island arc, west of thethenorthernendof the December1979(M~= 8.2) Bowers ridge, the maximum seismicity depth isthrusteventrupturezone,two normal-faultevents only 100 km, and the arc turns into a transform(173, 234) at intermediatedepth precededthe in- boundaryas it approachesthe Kamchatkacoast.terplaterupture. Event 234 took placein Novem- The volcanicchain that parallelsthe trenchfromher 1979 (mb = 7.2), a few weeksbefore the large Alaska to the Rat Island arc is very active (Fig.thrust event. Although event242, which followed 14).the 1979 event, is down-dip tensional, the focal The 1964 Alaska (M~= 9.2), 1957 Aleutianmechanismindicatestearfaulting, perhapsdueto (M~= 9.1) and 1965 Rat Island (M~= 8.7)
slab segmentation.An outer-risetensional event earthquakes,with rupturelengthsof over500 km,seawardof the 1979 rupture zone occurred on are amongthe largestthrusteventsin this centuryJanuary1981 (Christensenand Ruff, 1987) mdi- (Sykes,1971).However,only a few large(mB> 7)
eating that the thrust eventwas large enoughto intermediate-depthevents have occurred in theput the outer rise into tension. No precursory Aleutian—Alaskaregion (seeTableA). Event 306outer-risecompressionaleventswere recorded. (mh= 6.3) is locatednearthe northernedgeof the
1964Alaskaearthquakeaftershockarea,Thisevent
shows in-plate horizontal tension and a vertical7.4. North Pacific , . .
compressionaxis, which suggeststhat it may beassociatedwith eitherdeformationin the subduct-
7,4.1, Cascades ing plate or with continentalloading. Only oneShallowseismicityis almostnon-existentalong event with M � 6 has occurreddown-dip of the
the 500-km-longplateboundarybetweenthe Juan 1964 Alaskaaftershockzone, on December1968de Fucaplate’syoung oceaniclithosphereand the (110, mb = 6.5). This eventhad many aftershocks
016
N
___________________________________ andmay beassociatedwith the tearin this region/ / ~ proposedby Burbach and Frohlich (1986). Only
01 /___._—~\ 9)0 ~_—\ 5)
- — one outer-risetensionaleventwas detectedafterthe 1964 rupture (Christensenand Ruff, 1987),559.1] go
0 —
o andno outer-risecompressionaleventsare known2
ID to haveoccurredin the region.o Events 5, 166 and 227 occurreddown-dip of50
~ the northernedgeof the 1938 Alaskaearthquakeaftershockzone (McCann et al., 1979). Hudnut
~ 0~ ° and Taber (1987), using local seismicity data inNIS this region (known as the Shumagingap), ob-~~lZi ) Z
I -, serveda changein the intraplateseismicity pat-7 . 0 tern, from a single seismiczone, where events 5,
7) 0 166 and227 occur,to a doubleseismiczoneto the80
- ~ west where no larger intermediate-depthevents‘0
0 0 have been observedafter 1960. They suggested7) - that this differencemay reflect a stresschangeat
05
~ ‘ the interplateboundarybetweenthesetwo regions5 9)o that haveeither different aseismic-to-seismicslip
F ./~ ratios, or are at different stagesin the earthquake‘0 cycle.
Two down-dip tensionalevents, 61 and 334,~ ~ ~ have occurred down-dip of the 1957 Aleutian
~ 7) .~ earthquakeaftershockzone, Note that event334,~—G C which occurredin 1984 was located down-dip of
~ Al the 1986 Andreanof Islands (M~= 8.0) earth-quake aftershockzone, Before the 1957 Aleutian
.07) event, three large (mB> 7) intermediate-depth
~ ~ ~ earthquakesoccurredwithin the 20 years prior to~ thisevent,whereasonly onelargeevent(61, mB =
10 3.
-~ o, ... 6.9) has occurredthereafter(Table A). This sug-I ,...
N ~ 5)
~bwo05 1UJ~ ‘~ geststhat displacementfrom the 1957 eventat the0
sequenceof outer-rise tensional events that fol-~ ChristensenandRuff (1987)discussthe extensive
~ 0I1~/ ~ lowed the 1957 and1965 thrusts.-- o 7.5. WesternPacific.~‘
/E1~~ $~~: ~. .~ The 95-Ma-oldPacificplatesubductswestwardunder the Eurasiaplate at — 9 cm a~along the
~ .~ Kamchatka—Kurile—Japantrench(Fig. 14). To the
117
south, as the Pacificplate increasesin age to 135 the interplate boundary has allowed compres-Ma, it subducts along the Izu—Bonin—Mariana sional stressesto accumulateagain.trenchsystembeneaththe Philippineplatewith aslower convergencerateandseismicity,defining a 7.5.2. Kuriles—Japannearlyverticalplate(KatsumataandSykes,1969). The interplateboundaryalong northeastJapanThe chain of activevolcanoesparallelsthe trench and the southernKurile Island arc has rupturedaxis for most of its length as shown in Figs. 14 during a recent sequenceof large thrust earth-and 15. The dip angleand extentof the intraplate quakes,which are distributedfrom southwesttoseismicity vary along the trench as indicated in northeastalong the trench, in 1952 (M~= 8.1),Table I. Double seismic zonesat intermediate 1973 (M~= 7.8), 1969 (Mw = 8.2), 1958 (Mw =
depthhavebeenreportedin the Kurile—Kamchat- 8.3) and 1963 (M~= 8.5) (Beckand Ruff, 1987;ka region (see fig. 5 of Stauderand Maulchin, Schwartz and Ruff, 1987). However, along the1976) and in Japan(Hasegawaet al., 1979). The northern 500 km of the Kurile trench, the laststrengthof coupling at the interplateboundary,as shallow earthquakeoccurredin 1915 (M
5 = 8.0),inferred from earthquakemagnitudesof thrust but it is unknown whether this eventbroke theevents(Ruff and Kanamori, 1980),changesalong Kurile gap or was possibly an intraplate eventthe trench strike, from strong coupling offshore (McCannet al., 1979). ChristensenandRuff (1987)the KamchatkaPeninsulato seismic decoupling observedcompressionaleventsoffshore from thisalong the Marianatrench. gap which they interpret as evidencefor strong
coupling at the interplateboundaryin this region.Large intermediate-depthevents(m h> 6) down-
7.5.1. Kamchatka dip of the Kurile gap are only locatedclose to itsThe 1952 Kamchatka(M~= 9.0) earthquake northernandsouthernboundaries,so theseevents
broke the southern 400 km of the interplate may be associated,respectively,with the 1952boundaryalong the peninsula.To the north, the Kamchatka(M~= 9.0) or the 1963 Kurile Island1959 (M5 = 7.8) and the 1923 (Mw = 8.5) events (M~= 8.5) thrustearthquakes.rupturedthe remaining150 km. Intermediate-de- At the southernend of the 1952 Kamchatkapth earthquakesunderKamchatkacanbe divided rupture zone, event 2 has down-dip tensionalinto two depthgroups.Events deeperthan 125 km mechanismsat a depthof 50 km. Event 6 has a(231, 301) have down-dip compressionalfocal down-dip tensionalaxis, whereasevent165 has amechanismssimilar to deepereventsin this region down-dip compressionalaxis. The relative loca-(Stauder and Mualchin, 1976). The shallower tion of theseevents is consistentwith a doubleevents show a temporal change in stress axes’ seismic zoneat intermediatedepth (Stauderandorientation, as shown in Fig. 3b. A down-dip Maulchin, 1976). Event 333 (M~= 5.6) indicatescompressionalevent (1’, mb= 6.9) occurred on intraplatedeformationnear the outer rise, EventJuly 1960, indicating that displacementat the 3, which occurredin 1961 at a depth of 160 kminterplateboundaryduring the 1952 (M~= 9.0) down-dip of the 1963 interplate event, has aand 1959 (M5 = 7.8) events induced compres- down-dipcompressionalmechanism.This eventissional stressesat intermediatedepth. After July notconsistentwith the temporalmodel in Fig. 3b.1969, threedown-dip tensionalevents (116, 157, Large intermediate-deptheventsthat occurred304) with mb� 6.5) occurred at similar depth, from 1960 to 1984 down-dip of the Japan—south-suggestingthat the interplateboundaryis at least em Kurile trench reflect the complexity of themoderatelycoupledonceagain. Moreover,Chris- subductinglithospherein this region(e.g.,Burbachtensen and Ruff (1987) remarked that in the and Frohlich, 1986). Most intermediate-depthouter-riseregion, just after the 1952 Kamchatka earthquakesin this region, events223 (mb = 7.8),earthquake,tensional eventsoccurred.However, 176 (mb = 6.0), 123 (mb = 6.7), 164 (mh= 6.7)compressionaleventshavebeenobservedoffshore and 264 (mb= 6.3), have hinge-faulting mecha-southernKamchatka,indicating that coupling at nisms, Note also that events223 and 123 had
118
more aftershockswithin one week than similar tion would be consistentwith a shallow-thrustmagnitudeintermediate-depthevents(see Fig. 9). eventin this region. ThepP—P phasesconstrainedEvent 38, with a down-dip tensionalmechanism, the depthof this eventon theassumptionthat theoccurredin 1968 at the baseof the 1952 (M~= eventis not a multiple sourceshallow earthquake.8.1) aftershockzone, indicating a coupled inter- A more detailed analysis of this event may beplate boundary.Events 69 and 249 are down-dip requiredto resolveits natureconfidently.tensional. Event 152 (mb= 7.0) has a verticalcompressionaxis and a horizontal tensionalaxis, 7.5.4. Marianasreflecting lateralbendingof the downgoingplate, In the Mariana subduction zone, back-arcconsistentwith the orientationof stressesinduced spreadingis taking place, the interplateboundaryby a buckling plate (Sasatani,1976). This corn- is uncoupled,and shallow-thrustearthquakesareplexity suggests that the magnitude of inter- infrequent,with maximum magnitudesof — 7.0mediate-depthstressesunderneathnortheastJapan (Ruff and Kanamori, 1980). In 1902 a large(M
5owing to the distortion of the downgoing slab is = 7.9) earthquakeoccurredin the MarianaIslandgreater than those induced by the interplate arc, Although this event is considereda shallowboundary.However, farther southalong Honshu, event, there was no tsunami associatedwith it,the interplate boundary is uncoupled allowing perhapsindicating that this eventwas an inter-bending stressesto concentratenear the trench mediate-depthevent (McCann Ct al., 1979). Theaxis, resulting in the great normal-faulting 1933 focal mechanism of event 268 shows in-plateSanriku(M~= 8.4) earthquake(Kanamori,1970). horizontal compression,while events 49 and 307Many large intermediate-depthevents have oc- havein-plate horizontal tension.Thesevariationscurreddown-dipof the Kurile—Japantrenchprior are consistentwith lateral bending of the slabto 1960(seeAppendix1), but the mechanismsare along the trench strike, from convex to concavenot known for many of these. (BurbachandFrohlich, 1986).Seismicity is poorly
definedin the southernMananaarc, whereevents7.5.3. Izu—Bonin 196 (mh = 7.1) and 289 (mh = 6.1) indicate tearThrust earthquakesalong the Izu—Bonin trench faulting below 100 km depth. Event 96 shows
(Fig. 14) are infrequentandhavemaximummag- vertical compressivestressesand horizontal ten-nitudesof — 7.4 (McCannet al., 1979). However, sion in the convergencedirection.manylarge(mB> 7.0) intraplateearthquakeshaveoccurredduring this century(TableA). Event 168 7.6. Philippine Sea(mB= 7.4) has a horizontal compressionaxisalignedwith theplateconvergencedirectionanda The Philippine Sea(Fig. 15) is one of the mostdown-dip tension axis, consistent with stresses complex tectonicareasbecausethe Philippineandinducedby the negativebuoyancyof the downgo- Eurasiaplates subduct in different directions ining plate. However, event70 (mb= 6.8) has near this region. The Philippine plate subductsto thevertical compressivestressesandahorizontal ten- northwest along the Nankai Trough and thesion axis. Theseeventshave similar depths,but Ryukyu trench and to the southwestalong theevent 168 is located closer to the trench axis, Philippine trench. The Eurasia plate subductssuggestingthe presenceof a double seismic zone eastwardalong the Luzon trench and along thein the region. Moreover, the focal mechanismsof westcoastof NegrosIsland.The Celebessea-floortheseeventsare consistentwith unbendingof the subductseastwardalongthe Cotabatotrench,westdowngoingplate.Earthquakesthatoccur below 80 of Mindanao Island,and to the southalong thekm depth are down-dip compressionaleventsor Sulawesi trench(Cardwell et al., 1980). Seismicitytearfaults (28, 235, 277). in the Nankai Trough area is limited to shallow
Figure 9 shows that event168 had more than depths(<60 km) and the interplateboundaryin25 aftershockswith mb> 3 within a week. If this this region is strongly coupled. In contrast, allevent was shallower, its focal mechanism’ssolu- other subduction zones in the Philippine Sea
119
110 120 130 140 150
ii192rnel~) ~AA O302)~6o) PLAT~CS - ~-, : -
243)1k ~08fl60) ~ 196~~~ - ~‘ - 1/
1(1-) ~ .~ - .
I2548~ . ‘l~” r-. - fr’~ <
~94(8oi-. V - 201)60)
lot- ~ 81)6’) 96~.4
22~~O ~ (1~~216)1 ~ II 1 -
-~ ~ 2
Fig. 15. Focal mechanismsand location of intermediate-depthearthquakes(M � 6.0) associatedwith the Philippine plate thatoccurred between1960 and 1984. For symbols, seeFig. 12.
region are weakly coupled at the interplate and stateof stressof intermediate-depthseismicityboundary (Ruff and Kanamori, 1980) and only is observedalong the Ryukyu Islandarc. North oftensional outer-rise events are observed there the Tokaora channel, the slab dips at — 70°,
(ChristensenandRuff, 1987). active volcanism is present(Fig. 15), and inter-mediate-deptheventswith mh 5 havedown-dip
7.6.1. Ryukyu—NorthTaiwan tensionalmechanismsin responseto the negativeMany large (mB > 7) intermediate-depth buoyancyof thedowngoingplate.Note thatevents
earthquakesoccurredin theRyukyu—northTaiwan 219 (mb = 6.7) and 305 (mb= 6.5) are down-dipregionbefore1960 (seeAppendix1), including the tensional. South of the Tokaora channel, thelargestknownintermediate-depthevent, mB = 8.1, downgoing plate dips at — 45°, volcanism isthat occurredin 1911 under the Ryukyu Island scarce, and focal mechanismsof mh 5 eventsarc, However, after 1960 this region has been indicatedown-dip compression,as is true for therelatively quiet at intermediatedepth (Fig. 15). largeevent68 in Fig. 15. Thesechangescould beSeismicityextendsto 200 km deepalong mostof explainedin termsof lateraldifferencesin temper-the trench. aturein the surroundingmantle, as reflectedby
Event 104 indicateshinge faulting betweenthe volcanism, which can produceless or more resis-youngeroceanfloor that subductsin the Nankai tanceto the subductinglithosphere(Shionoet al.,Trough and that of the older Philippine seafloor 1980).to the west of the Kyushu—Palauridge. Shionoet A cluster of intermediate-deptheventsunder-al. (1980) indicatethat a markedchangein the dip neath north Taiwan delineatesthe edge of the
120
downgoingPhilippine plate.The larger events(7, -~
mh = 7.2, and 79, mb = 6.7) in this region have ~.
nearvertical tensionalaxesthat reflect the weakly ~‘~‘
coupled interplate boundary and the negative ~ % - -~\ \ ~ -~
buoyancyforces exertedby the downgoingplate. ~ 00/7~~ ,~.
Events 99, 221 and283 are hinge-faultingevents.
7.6.2. LuzonShallow and intermediate-depthseismicity in 4’
this region is scarceand not well defined(Card- \ ~well et al., 1980). During this centuryonly a few ~ ~ ~largeearthquakeshaveoccurredin this region(seeAppendix 1). Intermediate-depthevents(58, 302, ~
208) in the Luzon subductionzone may reflect Q~ s ~ ~only in-plate deformation. However, note that .~- S Q
4~, C
event 302 is down-dip compressional,suggesting ~ .. ~,
resistanceto subductionof the plateat depth. ~ -~ / ~‘
7.6.3. Philippines ~ H/<~~~ ~ - .
Cardwell et al. (1980) discussedin detail the 6 ~ } ~‘W )41’sk - .0
very complex tectonic setting of the Philippine ~ ~ r-’~~ ~‘ I ‘0
Island arc, which is boundedto the west by the —‘f ~
shallow Negros and Cotabatosubductionzones )~ \- ~ Cand to the eastby the Philippine trench, where ~ - - o> ~earthquakesoccur as deep as 640 km. Frequent L>-.-.~ ‘~ -
large shallow and intermediate-depthearthquakes ‘ 1-
havetakenplacein this region during thiscentury ~ ~‘ - I 4’
(see,e.g., McCannet al. (1979) andAppendix 1).In the northernsectionof the Philippinetrenchall ./
intermediate-depthevents,from events126 to 225 - Li
in Fig. 15, haveeither near down-dip tensional - .
mechanismsor vertical tension axes consistent Li 0 1) N
with the model for partially coupledregions(Fig. q. / ~ / ~‘ 0” .~‘~
8) in which vertical tensionalstressesare induced -i
by thenegativebuoyancyof the subductingplate. “ ‘~
Similarly, eventsalong the southPhilippine trench - a / ~ Chave nearly down-dip tensional axes, with the ~ 5~, 3..
exceptionof the deepestevent, 335 Mw = 7,5, __ ___ .~~
which hasa down-dipcompressionalaxis. ‘~ ~, ~ ~
7.6.4. Sulawesi °_3 . ‘~
The Sulawesi subductionzone (Fig. 16) was ~ .~~very active at intermediatedepths pnor to 1960 1> ~
(GutenbergandRichter,1954). Sequencesof large - I,>-(mB = 7.0—7.8) earthquakesoccurred mainly in ~1905—1907and1939—1942.However, seismicityat .~0 ~
shallow depthis limited. 8Stressaxes of intermediate-depthevents that
occurredafter 1960 do not havea preferredorien-
121
tation (see Fig. 10). Event 171 has an in-plate at — 7 cm a~.The volcanic chain parallelsthehorizontal axis, in contrast to events309, 34, 271 trench and seismicity occurs as deepas 300 km.and77 for which the compressionaxesare normal Historic large shallow earthquakes(M~ 8.7)
to the slab. Events 190 and 309 are down-dip haveoccurredalong the Sundatrench(Newcombtensional events. Interpretationof events in this and McCann, 1987), suggestingat least a mod-regionis difficult sincethe geometryof the down- erately coupled interplate boundary. All inter-going plate as well as the tectonicsof the region mediate-depthevents,from event327 to the northare poorly understood. to event 23 to the south, are down-dip tensional
events, and are consistentwith the model shown7.7. Indonesia in Fig. 8 for intermediate-deptheventsthat occur
down-dip of a partially coupled interplateThe Indo-Australianplate subductsand trans- boundary.Compressionalaxesfor mosteventsare
lates along the Sunda arc from Burma to the normal to the subductingplate; the exceptionisBandaSea(Fig. 16). Various segmentsare defined event204, whichhas horizontalin-platecompres-by differencesin the subductinglithosphereage, sion.BurbachandFrohlich(1986)predictedlateralthe convergencedirectionbetweenadjacentplates, compressionstressesin the regionwherethis eventthe maximum seismicity depth(see TablesI and occurs.II), and the strengthof coupling at the interplate
7.7.3. Javaboundary. Southof the Sunda Strait the maximum depth
of seismicity increasesabruptly to 650 km depth.7.7.1. Burma—AndamanSea The subductingseafloor is 135-Ma-old and theObliqueconvergencebetweenIndiaandEurasia
active volcanic chain parallels the Java trenchoccursin the Burmaregion, wherestrongcoupling (Fig. 16). No large historic shallow earthquakesis reflectedby several large shallow events that
are known in this region (McCann et al,, 1979;occurredin 1897(M~= 8.5), 1912(M
5 = 8.0), 1946 Newcomb and McCann, 1987), indicating that(M5 = 7.8) and1950 (Ms = 8.5). The largestinter- subductionis mostly aseismicin this 1700-km-longmediate-depthevent in the region occurred in
trench. In addition, the 1977 Sumbawanormal-1954 (mB = 7,4, Table A). Most intermediate-de-
fault earthquake(M~= 8.3), which occurredun-pth earthquakeshaveoblique normal-faultmecha-
der the trenchaxis in responseto bendingof thenisms (down-dip tensional), consistentwith de- lithosphereinducedby theslab pull (Spence,1986),tachmentof the denseroceaniclithospherefrom is also consistent with an uncoupledinterplatethe more buoyant continental plate. Event 117 boundary(see Fig. 8). Several large (mB� 7) in-indicatestearingof the subductingplateand may
be relatedto somepre-existingstructure. termediate-deptheventsoccurredprior to 1950 inthis region(TableA).
In the AndamanSea,oblique subductionof the Event48 hasanin-plate tensionaxis that could55-Ma-old seafloor is accompaniedby complex
be induced by the lateral bendingof the Indianback-arcspreadingin this region (Eguchi et al., plate as it subductsin this region (Burbachand1979;Banghar,1987). Seismicityextendsto — 100 Frohlich, 1986). Events88 and 134 are down-dipkm depth, with most intermediate-depthevents
tensional, whereas the shallower event, 14, is(287,44, 294)being down-dip tensional.This is in
down-dip compressional.Note the relative loca-agreementwith the model for moderatelycoupled tion of this eventwith respectto the trenchaxissubductionzones in which down-dip tensional and the stressorientationof theseeventssuggeststressesare inducedat intermediatedepthby the the possiblepresenceof a double seismic zone.negativebuoyancyof the downgoingplate.
Event 310 may be a tearfault.
7. 7.2. Sunda 7.7.4. TimorConvergenceof the subductingIndian plate to In the Banda Sea, the subductinglithosphere
the northeastalong the Sundatrenchtakesplace bendssharplyalong the Timor trench, where the
122
intermediateand deepseismicity define a highly and Lay, 1987). The compressionaxes of thesecontortedplate(Cardwell and Isacks, 1978). The events rotate systematicallyaround the bend incomplexity of the tectonics of the surrounding the downgoing lithosphere. The largest eventsregioncontributesto the different interpretations which occurredrecently in the BandaSeaare theof the shapeof the downgoing plate under the 1963 (25, mB = 7.8) andthe 1983 (312, M~= 7.4)Banda Sea (e.g., Burbach and Frohlich, 1986). earthquakes.Theseeventsare studiedin detail byDuring this century, many large intraplate Osadaand Abe (1981), Welc and Lay (1987)andearthquakes(Table A) haveoccurredin this re- Michael-Leiba(1984).The1963 eventmay involvegion, but only a few large shallow eventshave detachmentof the oceaniclithosphereat the lead-beendetected. ing edgeof the Australiancontinentalshelf.
All intermediate-deptheventsin the BandaSea,from event314, to the eastemnmostevent253, have 7.8. SouthwestPacificnearly down-dip tensional focal mechanismsin-duced by the downgoing plate. Note that the The southwest Pacific region is characterizedcompressionalaxes of the Banda Seaevents in by subductionalong deeptrencheswhich parallelFig. 16, rotate from a nearly N—S direction to a a systemof island arcs and marginalbasinsfrommore easterly direction as the subducted plate New Guineato New Zealand(Fig. 17). The inter-bendsnorthward(CardwellandIsacks,1978;Welc action betweenthe relatively old seafloorof the
Fig. 17. Intermediate-depthearthquakes(M � 6.0) in theNew Britain, Solomon,andNew Hebridesregions, wherethe Indian andSolomonplatesinteractwith thePacific plate. Eventsoccurredfrom 1960 to 1984. SeeFig. 12 for symbols.
123
PacificandIndo-Australianplatesandtheyounger BurbachandFrohlich, 1986).Largeshallow earth-SolomonandBismarkplatesdefinesdifferent sub- quakesin this region occur as multiplet events,ductionzonesin this complextectonicregion.The with the most recent sequencesoccurring afterrupturelengthsof mostlarge shallow earthquakes 1966 (Lay andKanamori,1980;Wesnouskyet al.,in the southwestPacificregionare <150km long 1986). The occurrenceof large shallow events(McCann et al., 1979), indicating moderatecou- locatedclose in time and spaceindicatesthat thispling in the region. Christensenand Ruff (1987) region has moderate coupling at the interplateobservedthat in this region both compressional boundary.andtensionaleventsoccur in the outer-riseregion. Most intermediate-deptheventsin the New Bri-This contrasts with the uncoupled subduction tam region havenearly vertical tensionaxes andzonesalong Indonesiaand aroundthe Philippine compressionaxes normal to the plate, consistentSea in which only tensional outer-rise events with the orientation of stressesinduced by theoccur. negativebuoyancyof the downgoingplate.Excep-
tions areevents76 and288, which indicatein-plate7.8.1. New Guinea deformation.Events52 and 158, which aredeeperSubductionalong the northeastcoast of New and nearthe edgesof the New Britain subduction
Guineadefines a downgoing plate to — 500 km zone, have in-platecompressionalaxes.Christen-depth along the westernend of the New Guinea sen and Ruff (1987) reported compressionaltrench(Pascal,1979). During this centuryseveral
outer-rise eventsoffshore from the western seg-large (M5> 7.5) shallow events have occurred ment of the New Britain trench, on the basis of
along the New Guinea trenchbut had very small which they infer that the interplateboundaryisrupturezones(McCannet al., 1979). A few inter- undercompression.Large thrust events occurredmediate-deptheventswith mb � 7.2 haveoccurred
therein 1945 and 1946. Note that event47, within this region.
focal depthof 65 km hadmanyaftershockswithinSince the shapeof the subductinglithospherea week. If this eventwas actually shallower, its
underNew Guineais poorly understoodand thefocal mechanismwould also be consistent with
stress axis distribution of intermediate-depth thrustingon the interplateboundary.eventsin this region does not show a consistent The New Ireland region includesthoseeventspattern, it is difficult to interpret the intraplate located near the corner joining the New Britainseismicityin this area. Nevertheless,we may em- andSolomon trenches.This small region, betweenphasize that focal mechanismsof intermediate- 152.5°and 155°Elongitude and 2° and 7°Sdepth earthquakesvary rapidly westward along latitude,hasmanylargeintermediate-deptheventsthe regionfrom event102 to event73, andthat the
_ during this century (Table A). The interplatelargestevents(8, 244) with M — 7.3 may indicate
boundary breaks in doublet events (Lay andhinge faulting. Mislocation of intermediate-depth Kanamori,1980) that occur aboutevery 30 years,events that occur betweenthe New Guineaand the most recent of which occurred on July 14New Britain subductionzonesmay also confuse (M~= 8.0) and July 26, 1971 (M~= 8.1). Mostthe interpretationof the tectonicsand the seismic
eventsat intermediatedepth in the New Irelandhistory of this region.
region havedown-dip tensionalmechanismscon-7.8.2. NewBritain—NewIreland—Solomon sistent with stressesinduced by the pull of theThe Solomonplate subductsnorthwardwith a slab.Event 138 indicateshinge faulting within the
steepdip angleto depthsof — 500 km along the plate. However, event 298 has a down-dip com-New Britain trench.The chainof volcanoesparal- pressionalmechanism.This eventoccurreddown-lels the trench(Fig. 17) alongNew Britain andthe dip, andafter, the July 14, 1971 thrustearthquake,Solomon Islands. The Solomon plate bends and is consistentwith the dynamicmodel shownabruptly under the New Ireland region and sub- in Fig. 3b. Although the down-dip tensionaleventductsin a northeasterlydirectionunderthe Pacific 150 (mb = 7.1) that occurredon July 19, 1971 isplate (Johnsonand Molnar, 1972; Pascal,1979; locateddown-dipof the July 14 event,it may have
124
helpedtrigger the July 26 thrust earthquakethat al., 1973; Isackset al., 1979;McCannet al,, 1979).brokethe adjacentsegmentnearthe cornerof the Events 19 and 89, located at the corner of theNew Britain trench (see fig. 3 in Lay and Solomon and New Hebridestrenches,haveverti-Kanamori, 1980). Christensenand Ruff (1987) cal compressivestressesand horizontal tensionalreported tensionalouter-rise eventsafter the oc- stressesaligned with the plate, which reflect thecurrenceof the 1971 New Ireland doublet. The lateralbendingof the downgoing Indo-AustralianNew Britain—New Ireland region, albeit com- plate (Burbachand Frohlich, 1986). All other in-plicated, agreeswith the simplemodel of partially termediate-depthevents in the northern segmentcoupledregionsshownin Fig. 8, in whichvertical of theNew Hebridestrench,from event161 to thetensional stresses are induced at intermediate north to event41, have a down-dip tensionaxisdepth from the slab’s weight. Also, temporal and a compressionaxis normal to the plate,con-changesof focal mechanismsare observedin the sistentwith the orientationof stressesinducedbyouter-rise and intermediate-depthregions associ- the negativebuoyancyof the slab. Events 41 andatedwith the occurrenceof large, underthrusting 181, which may be associatedwith subductionofearthquakes. the Torres rise, had many aftershockswithin a
The Solomon trenchis the boundarybetween week(see Fig. 9).the Pacific plate and the Solomon and Indo- Back-arcspreadingoccurs in the southernseg-Australia plates. The Woodlark Ridge- subducts ment of the New Hebridestrench. Intermediate-along the central part of the Solomon trench, depth events that occur in this region show awherefew largesubductioneventsoccur (McCann complexpatternin the stressdistribution (seeFig.et al., 1979). However to the north and southof 10). Events (324,10, 13) in this southernsegmentthis topographichigh, the interplateboundaryhas of the New Hebrides trenchhavedown-dip ten-beenvery active. It is characterizedby the occur- sional axes and arc-normal compressionalaxes,renceof multiplet events with rupturelengths of similar to eventsto the north, as discussedabove.— 100 km (Lay and Kanamori, 1980; Wesnousky Event 160 has a down-dip compressionand aet a!., 1986). Seismicity extends200 km deepin horizontaltensionaxis within the plate.Thiseventthis region. At intermediatedepth,events115, 132 is locatedin the regionwhere Burbachand Froh-and 111, located near the northern edge and lich (1986) computedlateralextensionto be pre-down-dipof the Woodlarkridge, indicate a corn- sent. Event 246 shows intraplate deformation.plex stressstate,possibly along the boundaryof Events 256, 83, 217 and 29, near the southernthe Solomonand Indo-Australianplatesat depth. bend where the boundary involves left-lateralEvents 175, 139 and162 havevertical tensionaxes strike-slip motion, are hinge-faultingevents.but variable compressionaxes induced by slab Large, shallow events occurred in the Newpull and deformationof the plate. Theseevents Hebrides region in 1966 (M
5 = 7.9), 1985 (M5 =
occurred before the recent 1978—1979 thrust 7.6, 7.6), 1973 (M5 = 7.9), 1965 (M~= 7.7), andearthquakesequencesin this region(Wesnouskyet 1980 (M~= 7.8) from north to south along theal., 1986). interplateboundary.However,the relationshipbe-
tweenthe large intermediate-deptheventsand the7.8.3. New Hebrides shallow underthrustingeventsis unclear.TheNewHebridessubductionzonedipssteeply
eastwardanddelineatesthe boundarybetweenthe 7.8.4. Tonga-Kermadec—NewZealandIndo-Australianplate and the Pacificplates.The The Tonga—Kermadectrenchsystemis a fairlyvolcanicchain parallelsthe New Hebridestrench linearislandarc with seismicityoccurringas deepfor most of its length (Fig. 17). The interplate as 650 km (Fig. 18). The Pacificplatesubductstoboundaryhasbeenvery activeduring this century the west underneaththe Indian plate,delineatingand it is the most active region at intermediate a smoothslab to — 400 km depth.However, deepdepth (Table A). Seismicity is continuousto a eventsindicate a highly contortedplatebelow 400depthof 300 km in the New Hebrides.(Pascalet km depth (Billington, 1980; Giardini and
125
~118(121).-’-~ 3235)o1...~241019I along the Kermadecregion is thus partially cou-326(16800 ~ ~~~285C621 pled,asindicatedby themaximumrupturelengths
- -~ of large shallow thrust eventsof — 150 km (Mc-- .- ~-— - 330000k Cann et al., 1979) while the Tongaregionappears
~ A 92 211)8 to be uncoupled Most of the deep seisrrucity- 273) ~ 4-.~ occursin the Tongatrench(Vassiliou et al., 1984;
.~ ~30(223~93)6~,~ ~71~,)~62ow — -20 Giardini, 1987),andthis regionis also very active
- ~%..~800)o) 65)15) at intermediatedepth.
~ ~ Intermediatedepth events near the northern
237(~’~~ ~ boundary of Tonga, where the Pacific plate is112)62 ~ ~ strongly contorted,are characterizedby in-plate-.. ~ 8T90) 209(00) .
4)1005 ~ PACIFIC from the deep seismic region to nearly all depths~ // PLATE in the subductingTonga lithosphere. The few
I down-dip tensionaleventsin this region are asso-322(7 II - ciated with subductionof the Louisville ridge,
.~ ~o which appearsto increasethe coupling of the
- _...~Ise.ys...TT”~ I interplate boundary locally. Events 262 and 27- I ~ with mb = 6.5 are locateddown-dipof the Louis-
170 180 -170 -160 yule ridge at its inferred northern and southernFig. 18. Focalmechanismsandepicentersof eventsoccurredin edges(Giardini andWoodhouse,1984). The largethe period 1960 to 1984 (with M � 6.0) in the Tonga,Kerma- normal-fault 1977 Tonga earthquake(209) wasdec and New Zealand regions, where the Pacific plate underlies located down-dip and to the north of the largethe Indian plate. For symbols seeFig. 12.
1982 (M5 = 7.7) thrust earthquake.An outer-nsecompressionalevent occurred before 1982 from
Woodhouse,1984). The slab dip is 55° under the thrust rupture zone (Christensenand Ruff,Tonga and increasesto 700 farther southunder 1987). Following the 1977 eventdeeperdown-dipthe Kermadec and New Zealand trenches.The compressionalevents occurred (230, 237, 267).volcanic chain, which parallelsthe Tonga trench, Thissequenceis consistentwith the dynamicmodeldisappearsin the regionwherethe Louisville ridge shownin Fig. 3b.is being subducted(Fig. 18). Most events in the Kermadec region have a
Large, shallow eventsalong the northernsee- down-dip tensionalaxis and arc-normalcompres-tion of the Tonga trench occurred before 1950 sion. Event 163, the deepesteventin Fig. 18, has(McCann et al., 1979), although the mechanisms down-dipcompressionalstressessimilar to deeperof these events are unknown. No large under- events in the region (e.g., Isacks and Molnar,thrustingmechanismsare known to haveoccurred 1969). In 1963 a down-dip tensionalevent (103,along the Tongaarc northof the Louisville ridge. mb= 6.4) occurred down-dip of the 1976 after-However,outer-risecornpressionaleventshaveoc- shock zone. Event 255 (mb= 6.1) which occurredcurred near 22°S(Christensenand Ruff, 1987). down-dipand after the two 1976 events, is down-Along the northern boundary of the Kermadec dip compressional.Furthermore,Christensenandtrench several thrust eventshave occurred after Ruff (1987) reported that an outer-risecompres-1960 with a largedoublet in 1976 (M5 = 7.7, 8.0). sional eventoccurredprior to the 1976 KermadecA large thrust eventoccurredin 1982 (M5 = 8.2) doubletasdid a smallerouter-risetensionalevent,at the location where the Louisville ridge inter- This sequenceis generally consistent with thesects the Tonga trench. The interplate boundary dynamicmodel in Fig. 3b.
126
o 20 40 60 80 100
V / I
205~~ EURASIAN
PLATE
~° (~\~ r ‘. —
O-
~ j 297(si~~
ARABIAN ~ /INDIAN
20~ ~ PLATE - -, PLATE 20
AFRICAN NPLATE
/1~ -
~ 15/
I I --. /~20 40 60 80
Fig. 19. Eventswith M � 6.0 that occurredat intermediatedepthbetween1960 and 1984 in regionswhere continentalcoalitionsaretaking place. See Fig. 12 for symbols. The P and T axis (closed and opensymbols, respectively)of eventsthat occurredin theIndu—Kush and Hellenic arc regions are shown in a lower hemisphere projection. Circles are events shallower than 100 km and
diamonds indicate deeper events.
In New Zealand,the intermediate-depthevents 7.9.2. Iran(74, 322, 169) havedown-dip tensionalaxesresult- The seismotectonicsof the Iran—Pakistanre-ing from the negativebuoyancyof the downgoing gion have been studied by several investigatorslithosphere. (e.g., Jacksonand Fitch, 1979, 1981; Jacoband
Quittmeyer,1979). Intermediate-deptheventsoc-7.9. Alpine—HimalayaBelt cur along the Makran subductionzone, wherethe
Arabian plate is subductingbeneathEurasiaat aFigure 19 shows the intermediate-depthevents shallow dip to — 80 km depth.Event 297 (M~=
that occurredbetween1960 and 1984 in the Al- 6.7) hasa nearly vertical compressionaxis and apine—Himalayabelt. tensionaxis nearlyhorizontalandin the direction
of convergence.At greaterdepth,JacobandQuit-7.9.1. Hindu-Kush tmeyer(1979) reporteddown-dip tensionaleventsSeismicity concentratesat — 220 km depth in in this area.
the Hindu-Kushregion, wherelargemB> 7 eventsare frequent(Table A). Intermediate-deptheventshavenear vertical tensional axes,suggestingthat 7.9.3. Rumaniathe denseroceaniclithospherewhich subducteda Although the detailedstructureof the subduct-few million years ago is now being detachedfrom ing lithosphereunderRumaniais poorly defined,the morebuoyantcontinentallithosphere, it appearsthat thereis an almostvertical slab that
127
strikes NE—SW and dips to the northwestunder ble, According to this catalogthe New Hebridesthis region (Isacksand Molnar, 1969). Seismicity region is the most active region at intermediateconcentratesbetween100 and150 km depth.Two depths,while Tongais the mostactivebelow 400large events occurred in this region recently, in km depth.Other regionswith a largenumber of1940 (mB = 7.3) andin 1977(205, M~= 7.4). The intermediate-depthevents are the Altiplano,latter eventindicatesnearvertical tensionstresses Timor, Sulawesi,Scotia and New Ireland, all ofsimilarly to eventsin the Hindu-Kushregion. which have strong contortions of the subducted
lithosphere.Other active regions include North7.9.4. Greece Chile, Kuriles, Ryukyus and Philippines. TimeThe Mediterraneanseaflooris subductingeast- sequencesof the occurrenceof large interplate
ward along the Hellenic arc underEurasiaat an subductionzone events for all regions do notangle of 60° to — 200 km depth (McKenzie, reveal any simple pattern of stress migration1978).Event 56 (mh = 6.7) hasdown-dip tensional down-dip; however, there is a marked tendencystresses probably induced by the negative for sequencesof large events within the plate tobuoyancyof the subductingplate.A largerearth- occur at periodswhen no largeinterplateactivityquake, mB = 7.7, occurredin Greece in 1926 at is taking place. Severalregions exhibit long time100 km depth(Table A). scale temporal variations in intraplate activity,
such as Ryukyu, Sulawesi,Peru, Philippines andNew Guinea,
8. Discussion and conclusions We testedthe dependenceof the maximumsizeintermediateand deepfocus earthquakeson van-
In this studywe havegiven an overview of the ous parametersof the subductionzones.Conver-global occurrence of large intermediate-depth gencerateappearsto influence the maximumsizeearthquakes.By considering only the largest of deepearthquakes,while the angleof descentofevents, we have restricted our attention to the the subductedslab is inversely proportional togross spatial and temporalcharacteristicsof the maximumeventsize. No dependenceon ageof thedeformationoccurring in each region. A funda- slab or maximum depth of penetrationwas de-mental assumptionin the study of intermediate- tected.deptheventsis that the earthquakesserveas stress In order to examinethe spatial and temporalgaugesfor the intraplateenvironment,reflecting variations in stress orientation of large inter-static and dynamic stress fields in a competent, mediate-depthearthquakes,a secondcatalog isquasi-elasticsubducting slab. Many interesting compiled. This catalog comprises focal mecha-aspects of intermediate-depthearthquakephe- nismsof all intermediatedepth(40—200km) eventsnomena,suchas double Benioff zones,platecon- with M � 6.5 for the period 1960—1984. We de-tortions, and shearzone lineamentsare only re- termined40 newfocal mechanismsand confirmedvealed by consideringall earthquakesin each re- the focal depth and focal parametersfor manygion. Our purposeis to investigatethe spatialand additionalevents.A catalogwith 335 events,whichtemporal relationship between the intermediate- includespublishedfocal mechanismsof the eventsdepthactivity and interplateeventsby using two with M> 6 in the samedepth range and timerelativelycompletecatalogsof largeintermediate- period.deptheventsratherthan to unveil all of the com- We first consideredthe general spatial char-plexity of eachsubductingslab. acteristicsof the focal mechanismsof the larger
The first catalogwe consideredis the historic events(M> 6.8) in this catalog. Four categoriesrecordof large(mB> 7) eventsthat haveoccurred were defined: (1) normal-faultevents(44%), andwithin subductingslabsin this century.Thiscata- (2) reverse-faultevents(33%), both with a strikelog gives earthquakelocationsonly, for few of the nearly parallel to the trench axis; (3) normal orfocal mechanismsare known. Only for eventsin reverse-faulteventswith a strike significantly ob-the last 25 years are the depthsconsideredrelia- lique to thetrenchaxis (10%)and(4) tearfaulting
128
events (13%). The focal mechanismsof type 1 seismic zones and may exhibit double seismiceventsoccurat the baseof stronglyor moderately zonesin responseto other factorssuchas unbend-coupledsubductionzones; similar type eventsoc- ing of the downgoing lithosphere. In the uncou-cur nearthe trenchaxis in uncoupledzones.Type pled regions, no temporal changesin the focal2 eventshavenearvertical tensionaxesandoccur mechanismsof either outer-rise or intermediate-mainly in regions that have partially coupledor depthearthquakesare observedwhile such tern-uncoupledsubductionzones, and where the ob- poral behavioris sometimesobservedin the par-servedcontinuousseismicityextendsdeeperthan tially coupled regions, as is the case for the300 km. We advanceda simple model, in which SolomonIslands.the increaseddip of the downgoingslab associated Intermediate-depthevents down-dip of mod-with weakly coupled subductionzones induces erately or strongly coupledregionshave normalnearlyvertical tensionalstressat intermediatede- faulting mechanismsnear the baseof the inter-pthand,consequently,the changein focal mecha- plate boundaryin responseto the down-dip pullnism from type 1 to type 2 events,Events of type of the subductedlithosphere.In contrast,regions3 occurwherethe trenchaxisbendssharply,caus- that are uncoupledproducelargenormal faultinging horizontal(parallel to the trenchstrike) exten- eventsnear the trenchaxis, wherestressesdue tosional or compressionalintraplatestress.Type 4 bendingof the lithosphereare largest. Temporalare hinge-faulting eventsassociatedwith lateral changeseitherin the focal mechanismsof inter-segmentationof the subductingslab. mediate-depthevents,from down-dip tensionalto
We determinedthe numberof aftershocksthat down-dip compressionalevents,or a reductioninoccurredwithin one week of the mainshock re- the seismic activity at intermediatedepth are fre-ported by the NOAA and ISC catalogs for the quently observed after the occurrenceof largeintermediate-depthevents with M � 6.5 that oc- thrustearthquakesalong stronglycoupledregions.curred between1960 and 1984.About 48% of the Along Middle and South America, where theeventshadno aftershocks,37% of the eventshad interplate boundary varies from moderate tobetweenone and five aftershocksand only 15% strongly coupled, intermediate-depthearthquakesmorethan five aftershocks.Thereis a slight corre- aregenerallynormal-faulteventsthatoccur beforelation betweenmainshockmagnitudeand a num- and down-dip of future large subductionearth-her of aftershocks.However,all eventswith more quakes.In most regionsalong Middle and Souththan ten aftershocksare located in the region America,we observea reductiononly in the num-associatedwith bendsin the subductedslab. ber of large intermediate-depthevents after large
Detailedregionalobservations,on the basis of subductioneventsoccur in the region. However,all the events (M> 6) in our focal mechanism down-dip compressionalevents have been ob-catalog,support theideathat thesubductinglitho- servedafter the 1960 Chile andthe 1974 Peruviansphereactsas a stressguide in responseto spatial thrust earthquakes.After the 1964 Alaska earth-or temporal changesof the strength of interplate quakeonly one large event, a tear fault, has oc-coupling. curred at intermediatedepth. Down-dip of the
Intermediate-depthearthquakesfor uncoupled, large1957 Aleutian earthquakea few eventsmdi-or partially coupled, regions involve predomi- cate coupling at the interplate boundary. Thisnantly reversefaulting, with nearly vertical ten- region was very active at intermediate-depthbe-sion axes. Regionsthat are partially or weakly fore 1957 when the interplateboundarybroke. Incoupledare the Lesser and GreaterAntilles, the addition, oneof theseeventsoccurredbefore andScotia arc, Ryukyu, Philippines, Timor, New down-dipof thelarge1986 Andreanofearthquake.Hebrides,and Solomonsubductionzones.Uncou- Down-dip of the 1952 and 1959 Kamchatkapled regions include central Japan, Izu—Bonin, earthquakesaftershock zones, a down-dip corn-MarianaandJava.Theeventsoccur in responseto pressionaleventoccurredin 1960.After 1969 onlythenegativebuoyancyof thesubductedslab.These events with down-dip tensional axes haveregions have relatively long and steeply dipping occurred,which indicatesthat this regionis again
129
moderately coupled. Large intermediate-depth figuresin this appendix.Theseeventsare listedbyeventsunderJapanand the South Kurile trench regionsin TableA. Earthquakehypocenterswerereflect the complexity of the subducting litho- taken from Abe (1981) for most eventsprior tospherein this region, which has beenvery active 1960. A few eventswere taken from Abe (1984,throughoutthis century. 1985). The parametersfor intermediate-depth
Along the Tongatrenchmostly down-dipcom- events(40—200 km deep)are takenfrom TableV,pressionaleventsare observed;however,the large while data for deepereventsare taken from the1977 event located down-dip of the Louisville NOAA catalog. Magnitudes are mB for eventsridge had a down-dip tensionalmechanismand that occurredbefore1975. M~ is given for eventsprecededthe large1982 Tonga thrustevent.Tern- that occurred thereafter, and mB values areporal changesin the focal mechanismsare also calculated from the relation mB = 0.63Mw + 2.5observedwith the occurrenceof the 1976 Kerma- (Kanarnori, 1983). In the figureswe show the mB
decdoublet events, valuefor all events.The seismicmoment,M0, can
In conclusion,this overview hasconfirmed the be computedfrom the relation log M0 =
2.4mB +
general complexity of the spatio-temporaloccur- 10.1 given by Kanamori (1983) for intermediaterenceof intermediate-depthearthquakes,with slab anddeepfocusevents.pull forces and lateral slab deformationplaying Figure Al showsmagnitude—depthdistributionthe principal role in causingtheearthquakeoccur- on the left-hand side. Diamonds indicate eventsrence. However, we havepresentednumerousin- that occurredafter 1960, presumablywith morestancesin which temporalvariationsin the level of reliable locations,The arrows indicate the extentactivity or stressorientationin the interplateen- of continuousseismicityfor eachregion(Table I).vironment are associatedwith large interplate Observationsmadeof the variation of magnitudethrust events.Given that such variations are ob- of this century, large, intraplateearthquakesandservedfor the very largestevents,futureinvestiga- depth are strongly dependentnot only on thetions of all intraplateactivity within the subducted accuracyof mB to measureearthquakesize, butslabs promise to reveal additional featuresof the also on the accuracyof the event location. Weseismiccycle, assumethat they are approximatelycorrect. Al-
thoughsomeregionsin SouthAmerica (Colombia(1), Ecuador(2), Peru (3) and North Chile (5))
Acknowledgements havedeepevents,it is unclearif a continuousslabis presentor if the shallow and deepseismicityis
We thank Bob F. Svendsenfor his computer unrelated (Stauder, 1975). The seismic activityexpertise.This researchwas supportedby u.s. stops at — 250 km and reappearsat — 550 kmGeological Survey Grants14-08-000l-G1170and depthat aboutthe samehorizontaldistancefrom14-08-002-G1277(H.K.) and NSF Grant EAR- the trench. Note that the Rivera and the Nankai8451715 (T.L.) and a Shell Faculty Careermi- Troughregionsare not includedin TableA. Thesetiation Grantto T.L. ContributionNo. 4504 Divi- regionshaveyoung,short subductingslabs(Tablesion of Geological and PlanetarySciences,Cali- I) and no large event deeper than 40 km hasfornia Institute of Technology Pasadena,Cali- occurredtheredunng this century.fornia The figure on the right-handside displaysthe
time—distancedistribution for regions1—38. Dis-tance along strike for each subductionzone isgiven in km, the approximatedirection of the
Appendix 1 projection is also indicated. Asterisks show thelocation of large shallow earthquakeswith M~�
The depth andtime distribution of large inter- 7.5 and the approximateextent of the aftershockmediateanddeepfocusearthquakesthat occurred areafor theseeventsis indicatedwith a horizontalbetween1904 and 1984 is shown by a series of line. Cross-diamondsindicate the location of
130
TABLE A
Large intermediateanddeep focusevents
NEV REGION
DATE TIME LATITUDE LONGITUDE DEPTH MAGNITUDEYear Mo Dy HrMn Sec (+
68 7.2 7.6 8.0 0 200 400 600 800 1000o F I i992 p I
100 — 1975 >K
200 — ____ 1960 4 AA ~ A
4 A300 b 1945 ~
400 E—~ 1930
500 1915 A
600 1900
700 F F 1885 F F F
ALTIPLANO(4)m Distance(km)
6.8 7.2 7.6 8.0 0 SE 250 500 750 1000NW12500 F F F 1990 F F F
100 — 1975 4 $___ 4
200 ____________ . i 960 $~ 4 4
300 ~ ~ i 945a) \i~4
400 . ~ 1930 ‘~ 4• $
500 1915 +44
600 . 1900
700 F F 1885 F F
Fig. A1-2.
events shown in the left-hand side figure with earthquakes,and(b) shalloweventswith M~� 7.5.
d� 300 km andtrianglesare deepevents(d> 300 Deepfocusearthquakesare indicatedby a closedkm). Symbol sizes are proportionalto the magni- diamondin thesefigures.Note that the occurrencetude of the events. For regions39—42, the right- of intraplatelargeeventsin someregionssuch ashandfigure displaysthe time distribution for each the Altiplano (4), North Chile (5), Kuriles (17),region of large (a) intermediateand deep focus NortheastJapan(18), Timor (30), New Hebrides
140
NORTH-CHILE(5)mB Distance(km) N
6.8 7.2 7.6 8.0 0 150 300 450 600 7500 F F F 1990 F F F F
ioç—~~ 1975 •200 1960
A 4300 4—~ ~ 1945 $
.0 a)• 4
400 E~ 1930 41\j/ ~9
500 1915
600 ~. 1900
700 F I F 1885 F F I F
CENTRAL-CHILE(6)Distance (km) N
6.8 7.2 7.6 8.0 0 100 200 300
0 F F F 1990 I
100 1975
9200 ~ 1960
E
300 .,~i 1945 4400 ~:; 1930
500 1915
600 1900
700 F F I 1885 I I
Fig. A1-3.
(35), Tonga(36), andthe Hindu-Kush(39) regions large events in the early part of the century,havebeenrelativelyconstantin the last 80 years. whereasother regionssuch as Peru(3) and NewHowever, someotherregionssuch as the Ryukyu Guinea(31) appearmoreactivein recentdecades.(21) and Sulawesi (23) were especiallyactive in This text replacesfigure captions.
6.8 7.2 7.6 8.0 0 250 500 750 1000 1250 15000 I I I 1990 I I__ I~I
__ •4 ~1Xi100 ~ 1975
20~___ 1960 $ 9 +
~ -4~) $ $44 5.300 ..~) 194543
400 ~ 1930
4* 4500 1915 $ ~,$ ~ $
600 1900 * * *
700 I I I 1885 F F I F F
TONGA(36)mB Distance (km) N
68 7.2 7.6 8.0 0 300 600 900 1200 1500 1800
0 i99C I I I I
100 1975
$200 1960 + $ $ ~ A
-4 5.
-p 300 ~ 1945 ~ $ $=
400— . i~ 1930 A AA
A *500 1915 ~ A
600 _C 1900 * *-
700 I I F 1885 I F I P I
Fig. Al-b8.
156
KERMADEC(37)
m8 Distance (km) N
6.8 7.2 7.6 8.0 0 250 500 750 1000 1250
0 F 199c F I P
100 1975 ~~E?4EH9
200 1960 *-4 5.‘—P- 300 — ..~) 194543
400 ~ 1930
500 1915 +
600 .—.~ 1900
701’ I I I 1885 P F I F
NEW—ZEALAND(38)
mB Distance (km)6.8 7.2 7.6 8.0 0 200 400 N 600
0 F I F 1990 P I
100 1975
200 1960
.4
~— 300 ..b 1945
-p a) +400 1 1930* *
500 1915
600 1900
700 I I . 1885 P
Fig. A1-b9.
157
M w7.4 8.0 8.6 9.2
HINDU-KUSH(39) 1885~ p pmB mB
6.8 7 2 7.6 8.0 6.8 7.2 7.6 8.00 P ‘. 190C p p
100 1915
200 ___________ ,_.. 1930 —____________ 5.
-4300 1945
40p.
400 1960
500 1975 —
a b600 i99C F P P P P
700 I F F Mw7.4 8.0 8.6 9.2IRAN(40) 1885 p p
mB mB
6.8 7.2 7.6 8.0 6.8 7.2 7.6 8.00 I F 1901’ I
100 — 1915 -
200 _.. 1930 -
-4300 1945 -
.0-pp. 1-~
400 . 1960
500 1975
a __________
600 1990 P P P I P
700 P I PFig. A1-20.
158
Mw
7.4 8.0 8.6 9.2GREECE(41) 1885 P F
mB mB
6.8 7.2 7.6 8.0 6.8 7.2 7.6 8.00 I F 190C’ P F
100 1915
200 ,_. 19305.
-4300 ~ 1945
.0-pp. 1
400 1960 —
500 1975
b600 1990 P p I P P P
700 F P F Mw
7.4 8.0 8.6 9.2RUMANL&(42) 1885 P P
mB m8
6.8 7.2 7.6 8.0 6.8 7.2 7.6 8.00 P 1900 P
100 —~‘ 1915 —
200 19305.
-4‘— 300 ~ 1945
p-pp.
400 1960
500 1975
a b600 1990 P P P P P
700 p F p
Fig. A1-2b.
159
SPAIN(43)m
8 mB6.8 7.2 7.6 8.0 6.8 7.2 7.6 8.0
0 P I I 1901’ F F F
100 1915
200 ._-.. 19305.
-~ 1945300
431-
3.400 1960
500 1975
a600 1990 P P F
700 I I P
RIVERA(9’) NANKAI(2 1’)
M Mw w
7.4 8.0 8.6 9.2 7.4 8.0 8.6 9.21885 I 1885
___ Ii
1900 — 1900 —
1915 ‘ 1915
..~ 1930 ‘ .,~ 1930
11) II)Ei~ 1945 1945 ______
1960 1960
1975 1975
1990L I P b 1990L__.~. I p bFig. Ab-22.
160
Appendix 2 The mechanismdiagramsin Figs. A2.1 and A2.2are shown on an equal angle projection of the
New focal mechanismsof intermediate-depth lower focal hemisphere.Closed circles are corn-earthquakeswere determinedfrom first-motion pressionsand open circles, dilatations. Numbersdata of long and short-periodWWSSN records, correspondto eventsin TableA. The asterisknext
~
JAN. 28,1964 AUG. 5,1968 NOV. 7, I968~ DEC. 7, 1968*
JAN 5, 1969* SEP 29, 970 FEB 21,1971 APR. 8,1971
MAY B, 1971* JUNE11, l97~* JUNE 17, I97~ JULYB, 971
JAN 8, 972* JAN 28,1972* FEB 14, I972~ DEC 4,1972*
Fig. A2.1. New focal mechanismsfor intermediate-depthearthquakesthat occurred from 1964 to 1972. Seetext for symbols.
161
APR. 3, 1973 APR 24, 1973 JUNE 9, 1973* AUG. 1,973
AUG 30, 1973 JAN 2,974 JAN. 5,1974 OCT. 9,1974*
OCT 29, 1974 MAR. 8, 1975 APR 9, 1975 AUG 10,1975
OCT 17, 1975* NOV. I, 975~ DEC. 25, 1975* MAR.4, 976
~ ~22O
JUNE 3,1976* NOV 30, 1976. MAR. 4, 977 AUG. 3,1978
Fig. A2.2. New focal mechanismfor intermediate-deptheventsthat occurred betweenb973 and b978. See text for symbols.
to the eventdate indicatesthat oneor two of the in short-periodrecordsto constrainthe eventde-focal parametershavebeenconstrainedby model- pth. Figure A2.1 showseventsthat occurredfrom,ing of a few long-period P-waveformsfor these 1964 to 1972 and Figure A2.2 thoseevents thateventsor by timing of (pP—)Pand (sp—P) phases occurredbetween1973 and1978.
162
References Cardwell, R.K., Isacks, B.L. and Karig, D.E., 1980.The spatialdistribution of earthquakes,focal mechanismsolutions and
subducted lithosphere in the Philippine and NorthAbe, K., 1972a. Focal processof the South Sandwich Island Indonesia Island. In: D.E. Hayes (Editor), The Tectonic
earthquakeof May 26, 1964. Phys. Earth Planet. Inter., 5: and Geologic Evolution of the SoutheastAsian Seasand110—122. Islands. Geophys. Monogram 23, Washington, DC. pp.
Abe, K., 1972b. Mechanismsand tectonic implications of the 1—35.
1966 and 1970 Peruearthquakes.Phys.Earth Planet. Inter., Chase,C.G., 1978. Extension behind island arcs and motion5: 367—379. relative to hot spots. J. Geophys.Res., 83: 5385—5387.
Abe, K., 1981. Magnitudes of large shallow earthquakes from Chapple, W.M. and Forsyth, D.M., 1979. Earthquakesand1904 to 1980. Phys. Earth Planet. Inter., 27: 72—92. bending of plates at trenches. J. Geophys. Res., 84:
Abe, K., 1984.Complements to “Magnitude of largeshallow 6729—6749.
Abe, K., 1985. Re-evaluationof the largedeep earthquakeof America andin theNew HebridesIslandarc. Tectonics,2:January21, 1906. Phys. Earth Planet. Inter., 39: 157—166. 529—563.
Abe, K. and Kanamori, H., 1979. Temporal variation of the Christensen,D. and Ruff, L.J., 1983. Outer-rise earthquakesactivity of intermediate and deep focus earthquakes.J. and seismic coupling.Geophys.Res. Lett., 10: 697—700.Geophys.Res., 84: 3589—3595. Christensen, D. and Ruff, L.J., 1987. Outer-riseearthquakes
Anderson, D.L., 1967. Phasechangesin the upper Mantle, and their relationship to seismiccoupling (preprint).Science,157: 1165—1173. Chung, WY. and Kanamori, H., 1976. Source processand
Ansell, J.H. and Smith, E.G.C., 1975. Detailed structure of a tectonic implications of the Spanishdeep-focusearthquake
mantleseismiczoneusing thehomogeneousstation method, of March 29, 1954. Phys. Earth Planet. Inter., 13: 85—96.
Nature,253: 518—519. Chung,WY. and Kanamori, H., 1978. Subduction processof aAstiz, L. and Kanamori, H., 1984. An earthquakedoublet in fracture zone and aseismic ridges. The focal mechanism
Ometepec, Guerrero, Mexico. Phys. Earth Planet. Inter., and sourcecharacteristics of the New Hebrides earthquake34: 24—45. of 1969, January 19 and some related events. Geophys. J.
Astiz, L. and Kanamori, H., 1986. Interplate coupling and R. Astron. Soc., 54: 221—240.
temporal variation of mechanisms of intermediate-depth Chung, WY. and Kanamori, H., 1980. Variation of seismicearthquakes in Chile. Bull. Seismol. Soc. Am., 76: sourceparametersand stressdrop within a descendingslab
1614—1622. and its implications in platemechanics.Phys. Earth Planet.Banghar, AR., 1987. Seismo-tectonicsof theAndaman—Nico- Inter., 23: 134—159.
bar Islands.Tectonophysics,133: 95—104. Clowes, R.M., Ellis, R.M., Jajnal, Z. and Jones, IF., 1983.Barazangi,M. and Isacks,B.L., 1976. Spatial distributionof Seismic reflections from subductinglithosphere?Nature,
earthquakesand subductionof the Nazcaplate beneath 303: 668—670.SouthAmerica. Geology,4: 686—692. Crosson,R.S., 1980. Review of seismicity in the Puget Sound
Barazangi,M. and Isacks, B.L., 1979. Subduction of the Nazca region from 1970 to 1978. In: Proceedingsof Workshopplate beneathPeru; evidencefrom the spatialdistnbution XIV, EarthquakeHazardsof the Puget Sound Region,of earthquakes. Geophys.JR. Astron. Soc.,57: 537—555. Washington.U.S. Geol. Surv., Open-FileReport 83-19, pp.
Barker, P.F., 1970. Plate tectonics of the Scotia Sea region. 6—18.Nature, 228: 1293—1296. Dean, B.W. and Drake, C.L., 1978. Focal mechanismsolutions
Barker, P.F., 1972.A spreading centre in the East Scotia Sea, and tectonics of the Middle Amenca arc. J. Geology, 86:Earth Planet. Sci. Lett., 15: 123—132. 111—128.
Beck, S.L. and Ruff, Li., 1984. The rupture processof the Denham, D., 1977. Summaryof earthquake focal mechamsmsgreat 1979 Colombia earthquake: evidencefor the asperity for the western Pacific—Indonesia region 1929—1973.model. J. Geophys. Res., 89: 9281—9291. NOAA, U.S. Dept. of Commerce Report SE-3.
Beck, S.L. and Ruff, L.J., 1987. Sourceprocessof the 1963 Dewey, J.W. and Algermissen, ST., 1974. Seismicity of the
Kurile Islands earthquake. J. Geophys.Res. (submitted). Middle America arc-trench systems near Managua,Billington, S., 1980. The morphology and tectonics of the Nicaragua. Bull. Seismol. Soc.Am., 64: 1033—1048.
subducted lithosphere in the Tonga—Fiji—Kermadec region Dewey, J.W. and Spence,W., 1979. Seismicgaps and sourcefrom seismicity and focal mechanismsolutions. Ph.D. the- zonesof recentlarge earthquakesin coastalPeru. Pageoph,
sis, University of Cornell, 220 pp. 117: 1148—1171.
Burbach,G.V. and Frohlich, C., 1986. Intermediateand deep Dmowoska,R., Rice, J.R., Lovison, L.C. and Josell, D., 1987.
seismicity and lateral structure of subducted lithosphere in Stresstransfer and phenomenain coupled subduction zonesthe Circum-Pacificregion. Rev. Geophys.,24: 833—874. during the earthquake cycle (preprint).
Cardwell, R.K. and Isacks, B.L., 1978. Geometry of subducted Dziewonski,A.M. and Woodhouse,J.H., 1983. An experimentlithosphere beneath the Banda Sea in eastern Indonesia in the systematic study of global seismicity; centroid-mo-from seismicity and fault plane solutions. J. Geophys. Res., ment tensor solutions for 201 moderate and large earth-83: 2825—2838. quakesof 1981.J. Geophys. Res., 88: 3247—3271.
163
Dziewonski,AM., Friedman,A., Giardini,D. andWoodhouse, Fitch, T.J. and Molnar, P., 1970. Focal mechanismsalongJ.H., 1983a. Global seismicity of 1982: centroid-moment inclined earthquakezonesin the Indonesia—Philippinere-tensor solutions for 308 earthquakes.Phys. Earth Planet. gion. J. Geophys.Res.,75: 1431—1444.Inter., 33: 76—90. Forsyth, D.W., 1975. Fault planesolutionsandtectonicsof the
Dziewonski,AM., Friedman,A. andWoodhouse,J.H., b983b. south Atlantic and Scotia Sea. J. Geophys. Res., 80:Centroid-momenttensorsolutionsfor January—March1983. 1429—1443.Phys. Earth Planet. Inter., 33: 71—75. Frankel, A., 1982. A compositefocal mechanismfor micro-
Dziewonski,AM., Franzen, J.E. and Woodhouse,J.H., b983c. earthquakes along the northeastern border of the Carib-Centroid-moment tensor solutions for April—June 1983. bean plate. Geophys. Res.Lett., 9: 511—514.Phys.Earth Planet. Inter., 33: 243—249. Frankel, A. and McCann, W.R., 1979. Moderate and large
Dziewonski, AM., Franzen,J.E. andWoodhouse,J.H., b984a. earthquakesin the South Sandwich arc: Indications ofCentroid-momenttensor solutions for July—September tectonic variation along a subduction zone. J. Geophys.1983.Phys. Earth Planet. Inter., 34: 1—8. Res., 84: 5571—5577,
Dziewonski, AM., Franzen,J.E. and Woodhouse,J.H., b984b. Frankel, A., McCann, W.R. and Murphy, A.J., 1980. Observa-Centroid-moment tensor solutions for October—December tions from a seismic network in the Virgin Island region:
1983. Phys. Earth Planet. Inter., 34: 129—136. Tectonic structuresand earthquakeswarms. J. Geophys.Dziewonski, AM., Franzen, i.E. and Woodhouse,J.H., b984c. Res., 85: 2669—2678.
Centroid-momenttensorsolutionsfor January—March1984. Fujita, K. and Kanamori, H., 1981.Double seismiczonesandPhys. Earth Planet. Inter., 34: 209—219. stressesof intermediate depthearthquakes.Geophys.J. R.
Dziewonski, AM., Franzen,J.E. andWoodhouse,J.H.,b985a. Astron.Soc.,66: 131—156.Centroid-moment tensor solutions for April—June 1984. Fukao, Y., Hori, S. and Ukawa, M., 1983. A seismologicalPhys.Earth Planet. Inter., 37: 87—96. constrainton the depth of basalt—eclogite transition in a
Dziewonski,AM., Franzen,i.E. andWoodhouse,J.H., 1985b. subductingoceaniccrust. Nature,303: 413—415.Centroid-momenttensor solutions for July—September Giardini, D., 1984. Systematicanalysisof deep seismicity: 200
1984. Phys.Earth Planet. Inter., 38: 203—213. centroid momenttensorsolutionsfor earthquakesbetweenEarthquakeDataReport,U.S.G.S.,1984. 1977 and 1980. Geophys.J. R. Astron. Soc., 77: 883—914.Eguchi,T., Uyeda, S. and Maki, T., 1979.Seismotectonicsand Giardini, D., 1987. Quantification and frequency distribution
tectonic history of the AndamanSea.Tectonophysics,57: of deep earthquakes.J. Geophys. Res. (submitted).35—51. Giardini, D. and Woodhouse,J.H., 1984.Deep seismicity and
Engdahl, E.R., 1977. Seismicity and plate subduction in the modesof deformationin Tongs subductionzone. Nature,Central Aleutians. In: M. Talwani and W.C. Pitman III 307: 505—509.(Editors), Island Arcs, Deep Sea Trenches and Backarc Giardini, D., Dziewonski, AM. and Woodhouse, J.H., 1985.Basins.Am. GeophysicalUnion, Washington,DC. Centroid-moment tensor solutions for 113 large earth-
Engdahl, E.R. and Scholz,C.H., 1977. A double Bemoff zone quakes in 1977—1980. Phys. Earth Planet. Inter.. 40:beneath the central Aleutians: An unbending of the litho- 259—272.
sphere.Geophys.Res. Lett., 4: 473—476. Given, J.W. and Kanamori,H., 1980. The depth extent of theEngdahl, ER., Sleep,N.H. and Lin, MT., 1977. Plate effects 1977 Sumbawa, Indonesiaearthquake. EOS, 61: 1044.
in North Pacific subduction zones. Tectonophysics, 37: Gonzalez,J., McNally, K. and Rial, J., 1984. Recent normal95—116. faulting in southcentral Mexico: A study of sourceprocesses
Fedotov, S.A., 1968.On deep structure, propertiesof the upper and regional tectonic implications (preprint).mantle,and volcanismof the Kui-il—Kamchatka island arc González-Ruiz,JR., 1986. Earthquakesourcemechanicsandaccordingto seismicdata.Am. Geophys.Union, Geophys. tectonophysicsof the Middle America subductionzone inMonogram, 12: 131—139. Mexico. Ph. D. thesis,University of California, Santa Cruz.
Fedotov, S.A., Bagdasarova,SM., Kuzin, I.P. and Tarakanov, Goto, K., Hamaguchi, H. and Suzuki. Z.. 1985. EarthquakeR.Z., 1971. Earthquakesand Deep Structureof the South generating stressesin a descendingslab. Tectonophysics,
Kurile Island Arc. Israel Program for Scientific Transla- 112: 111—128.tions, Jerusalem,vi + 249 pp. Gutenberg, B. and Richter, C.F., 1954.Seismicity of the Earth,
Fitch, T.J., 1970a. Earthquake mechanisms and island arc 2nd Edn. Princeton University Press, Princeton, NJ, 330tectonics in the Indonesian—Philippine region. Bull. Seis- pp.mol. Soc.Am., 60: 565—591. Hasegawa,A. and Sacks, S., 1981. Subduction of the Nazca
Fitch, T.J., 1970b. Earthquake mechanismsin the Himalayan, plate beneath Peru as determined from seismic observa-Burmese, and Andaman regions and continental tectonics tions. J. Geophys. Res., 86: 4971—4980.
in central Asia. J. Geophys.Res., 75: 2699—2709. Hasegawa,A. and Umino, N., 1978. Focal mechanismsand theFitch, T.J., 1972. Plate convergence, transcurrent faults and distribution of seismicity in northeast Japan. Progr. Abstr.
internal deformation adjacent to southeast Asia and the Seismol. Soc.Jpn., 1: 34 (in Japanese).westernPacific. J. Geophys. Res., 77: 4432—4460. Hasegawa, A., Umino, N., Takagi, A. and Suzuki, Z., 1979.
164
Double-planed deep seismic zoneand anomalousstructure normal faulting: the Sanriku earthquakeof 1933. Phys.in the upper mantle beneath northeastern Honshu (Japan). Earth Planet. Inter., 4: 289—300.Tectonophysics, 57: 1—6. Kanamori, H., 1977a.The energyreleasein greatearthquakes.
Heaton, T.H. and Kanamori, H., 1984. Seismicpotential asso- J. Geophys.Res., 82: 2981—2987.ciated with subduction in the Northwestern United States. Kanamori, H., 1977b. Seismicand aseismicslip along subduc-Bull. Seismol.Soc. Am., 74: 933—941. tion zones and their tectonic implications. Am. Geophys.
House, L.S. and Jacob, K.H., 1983. Earthquakes, platesubduc- earthquakes.Ann. Rev. Earth Planet. Sci., 14: 293—322.tion and stress reversals in the eastern Aleutian arc. J. Kanamori, H. and McNally, K.C., 1982. Variable rupture
Geophys. Res., 88: 9347—9373. . mode of the subduction zone along the Ecuador—ColombiaHudnut, K.W. and Taber, J.J., 1987. Transition from double to coast. Bull. Seismol. Soc. Am., 72: 1241—1253.
single Wadati—Benioff seismic zone in the Shumagin Is- Kanamori, H. and Tsumura,K., 1971. Spatial distribution oflands, Alaska. Geophys.Res. Lett., 14: 143—146. earthquakes in the Ku Peninsula, Japan, south of the
Isacks, B.L. and Barazangi, M., 1977. Geometry of Benioff MedianTectonicline. Tectonophysics,12: 327—342.zones: lateral segmentation and downward bending of the Katsumata, M. and Sykes,L.R., 1969. Seismicity and tectonicssubducting lithosphere. Am. Geophys. Union, Maurice of the western Pacific: Izu—Mariana—Caroline andEwing Series, 1: 99—114. Ryuku—Taiwan Region. J. Geophys.Res., 74: 5923—5948.
Isacks, B.L. and Molnar, P., 1969. Mantleearthquakemecha- Kawakatsu, H., 1985. Double seismic zones: A first ordernisms and the sinking of the lithosphere. Nature, 223: featureof plate tectonics.Ph.D. thesis,StanfordUniversity,1121—1124. California.
Isacks,B.L. and Molnar, P., 1971. Distributionof stressesin Kawakatsu, H., 1986a. Double seismic zones: kinematics. J.
the descending lithosphere from a global survey of Geophys.Res., 91: 4811—4825.focal-mechanismssolutions of mantle earthquakes.Rev. Kawakatsu, H., 1986b. Downdip tensionalearthquakesbe-Geophys.SpacePhys.,9: 103—174. neath theTonga arc: A doubleseismiczone?J. Geophys.
Isacks, B.L., Oliver, J. and Sykes, L.R., 1968. Seismology and Res., 91: 6432—6440.
thenew global tectonics.J. Geophys.Res., 73: 5855—5899. Kawakatsu,H. andSeno,T., 1983. Triple seismiczone and theIsacks, B.L., Sykes,L.R. and Oliver, J., 1969. Focal mechanism regional variation of seismicity along the northern Honshu
of deep and shallowearthquakesin the Tonga—Kermadec arc. J. Geophys.Res., 88: 4215—4230.region and the tectonics of island arcs. Geol. Soc. Am. Kelleher, J., 1972. Rupture zones of large South AmericanBull., 80: 1443—1470. earthquakesand somepredictions. J. Geophys. Res., 77:
Isacks, B.L., Cardwell, R.K., Chatelain, J.L., Barazangi, M., 2087—2103.Marthelot, J.M., Chim, D. and Loutat, R., 1979. Seismicity Kelleher, J., Savino, J., Rowlett, H. and McCann, W., 1974.and tectonics of the central New Hebrides island arc. Am. Why and wheregreat thrust earthquakesoccuralong island
Geophys. Union, Maurice Ewing Series,4: 93—116. arcs. J. Geophys. Res.,79: 4889—4899.Ishida, M., 1970. Seismicity and travel-time anomalyin and Korrat, I. andMadariaga,R., 1986. Ruptureof theValparaiso
around Japan. Bull. Earth Res. Inst., 48: 1023—1051. (Chile) gap from 1971 to 1985. In: S. Das, J. BoatwrightIshikawa, Y., 1985. Double seismic zone beneath Kysushu, and C.H. Scholtz (Editors), Earthquakesourcemechanics.
Japan. Zisin, 38: 265—269. Am. Geophys.Union, pp. 247—258.Jackson, J. and Fitch, TJ., 1979. Seismotectonic implications Langston, CA. and Blum, D.E., 1977. The April 29, 1965
of relocated aftershock sequencesin Iran and Turkey. Puget Sound earthquake and the crustal and upper mantleGeophys. J. R. Astron. Soc., 57: 209—229. structure of western Washington. Bull. Seismol. Soc. Am.,
Jackson, J. and Fitch, T., 1981. Basement faulting and the 67: 693—711.focal depthsof the larger earthquakes in the Zagros moun- Lay, T. and Kanamori, H., 1980. Earthquake doublets in thetains (Iran). Geophys.J. R. Astron. Soc., 64: 561—586. Solomon Islands. Phys. Earth Planet. Inter., 21: 283—304.
Jacob, K.H. and Quittmeyer, R.L., 1979. The Makron region LeDain, A.Y., Tapponier, P. and Molnar, P., 1984. Activeof Pakistan and Iran: trench—arc system with active plate faulting and tectonics of Burma and surrounding regions. J.subduction. In: Geodynamicsof Pakistan. Geophys. Res., 89: 453—472.
Jarrard, RD., 1986. Relations among subduction parameters. LeFevre, L.V. and McNally, K.C., 1985. Stress distributionRev. Geophys.,24: 217—284. and subduction of aseismic ridges in the Middle America
Johnson, T. and Molnar, P., 1972. Focal mechanism and plate subduction zone.J. Geophys. Res.,90: 4495—4510.tectonics of the southwestPacific. J. Geophys. Res., 77: Malgrange, M. and Madariaga, R., 1983. Complex distribution5000—5032. of large thrust and normal fault earthquakesin the Chilean
Malgrange,M., Deschamps,A. andMadariaga,R., 1981. Thrust Pascal,G., Dubois, J., Barazangi, M., Isacks, B.L. and Oliver,and extensional faulting under the Chilean coast, 1965, J., 1973. Seismic velocity anomalies beneath the New1971 Aconcagua earthquakes. Geophys. J. R. Astron. Soc., Hebrides Island arc: evidence for a detached slab in the66: 313—331. Upper-mantle. J. Geophys. Res., 78: 6998—7004.
McCaffrey, R., Molnar, P., Roecker, SW. and Joyodiwiryo, Y., Pascal, G., Isacks, B.L., Barazangi, M. and Dubois, R., 1978.1985. Microearthquake seismicity and fault plane solutions Precise relocations of earthquakes and seismotectonicsofrelated to arc—continent collision in the eastern Sunda arc, the New Hebrides island arc. J. Geophys. Res., 83:Indonesia. J. Geophys. Res., 90: 4511—4528. 4957—4973.
McCann, W.R. and Sykes,L.R., 1984.Subduction of aseismic Pennington,W.D., 1983.Role of shallow phase changes in theridges beneath the Caribbean plate: implications for the subduction of oceaniccrust. Science,220: 1045—1047.tectonic and seismic potential of the northeastern Carib- Reid, ID., 1976. The Rivera plate: A study in seismologyandbean. J. Geophys. Res., 89: 4493—4519. plate tectonics. Ph.D. thesis, Universityof California, San
McCann, W.R., Nishenko, S.P., Sykes, L.R. and Krause, J., Diego.
1979. Seismic gaps and plate tectonics: seismic potential for Richter, CF., 1958. Elementary Seismology. Freeman, Sanmajor boundaries. Pageoph., 117: 1082—1147. Francisco, CA, 768 pp.
McKenzie, D., 1978. Active tectonics of the Alpine—Himalaya Richter, F.M., 1979. Focal Mechanisms and seismic energy
belt: the Aegean Sea and surrounding regions. Geophys. J. release of deep and intermediate earthquakes in theR. Astron. Soc., 55: 217—254. Tonga—Kermadec region and their bearing on the depth
McNally, K.C. and Minster, B., 1981. Nonuniform seismic slip extent of mantle flow. J. Geophys. Res., 84: 6783—6795.rates along the Middle America trench. J. Geophys. Res., Ripper, I.D., 1974. Some earthquake focal mechanismsin the
86: 4949—4959. New Guinea—Solomon Island region, 1963—1968. Aust.,McNally, K.C., Gonzalez-Ruiz, J.R. and Stolte, C., 1986. Bur. Miner. Resour. Geol. Geophys. Rep. 178.
Seismogenesis of the 1985 Great (M5 = 8.1) Michoacan, Ripper, I.D., 1975. Some earthquake focal mechanismsin the
Mexico earthquake. Geophys. Res. Lett., 13: 585—588. New Guinea—Solomon Island region, 1969—1971. Aust.,Michael-Leiba, M.O., 1984. The Bands Sea earthquake of 24 Bur. Miner. Geol. Geophys. Resour. Rep. 192.
November 1983: evidence for intermediate-depth thrust Ruff, L. and Kanamon, H., 1980. Seismicity and the subdue-faulting in the Benioff zone. Phys. Earth Planet. Inter., 36: tion process. Phys. Earth Planet. Inter., 23: 240—252.95—98. Ruff, L. and Kanamori, H., 1983. Seismic coupling and un-
Minster, J.B. and Jordan, T.H., 1978. Present day plate mo- coupling at subduction zones. Tectonophysics, 99: 99—117.
tions. J. Geophys. Res., 83: 5331—5354. Samowitz, I.R. and Forsyth, D.W., 1981. Double seismic zoneMogi, K., 1973. Relationship between shallow and deep beneath the Mariana Island arc. J. Geophys. Res., 86:
seismicity in the western Pacific region. Tectonophysics,17: 7013—7021.1—22. Sasatani, T., 1976. Mechanisms of mantle earthquakesnear the
Molnar, P. and Sykes,L.R., 1969. Tectonics of the Caribbean junction of the Kurile and northern Honshu arcs. J. Phys.and Middle-America regions from focal mechanisms and Earth, 24: 341—354.seismicity. Geol. Soc. Am. Bull., 80: 1639—1684. Schwartz, S. and Ruff, L.J., 1987. Asperity distribution and
Molnar, P., Freedman, D. and Shih, J., 1979. Lengths of earthquake occurrence in the Southern Kurile Islands Arc.
intermediate and deep seismic zones and temperatures in Phys. Earth Planet. Inter. (submitted).downgoing slabs of lithosphere. Geophys. J. R. Astron. Seno, T., 1977. The instantaneous rotation vector of the Philip-
Soc., 56: 41—54. pine Seaplate relative to the Eurasian plate. Tectonophysics,Moore, G.W., 1982. Plate tectonic map of the circum Pacific 42: 209—226.
region, explanatory notes:Tulsa, OK. AmericanAssocia- Seno,T. and Kurita, K., 1978. Focalmechanismandtectonicstion of Petroleum Geologist, 14 pp. in the Taiwan—Philippine region. J. Phys. Earth, 26 Suppl.,
Newcomb, K.R. and McCann,W.R., 1987.Seismic history and S249—S263.seismotectonicsof the Sunda arc. J. Geophys. Res., 92: Seno, T. and Pongswat, B., 1981.A triple-planed structure of421—440. seismicity and earthquakemechanismsat the subduction
Nishimura, C., Wilson, D.S. and Hey, R.N., 1984. Pole of zoneoff Miyagi Prefecture, northern Honshu, Japan. Earthrotation analysisof present-day Juan de Fuca plate motion. Planet. Sci. Lett., 55: 25—36.
J. Geophys. Res., 89: 10283—10290. Shiono, K., Mikumo, T. and Ishikawa, Y., 1980. Tectonics ofOike, K., 1971.On the nature and occurrenceof intermediate the Kyushu—Ryukyuarc as evidenced from seismicity and
and deep earthquakes. 1. The world-wide distribution of focal mechanismsof shallow to intermediate depth earth-the earthquake generating stress. Bull. Disas. Prey. Res. quakes. J. Phys. Earth, 28: 17—43.
Inst., Kyoto U., 20: 145—182. Silver, P.G. and Jordan, T.H., 1983. Total moment spectra ofOsada, M. and Abe, K., 1981. Mechanism and tectonic im- fourteen large earthquakes. J. Geophys. Res., 88:
plications of the great Banda Sea earthquake of November 3273—3293.4, 1963. Phys. Earth Planet. Inter., 25: 129—139. Singh, S.K., Astiz, L. and Havskov, J., 1981.Seismicgapsand
Pascal, G., 1979. Seismotectonics of the Papua-New recurrence periods of large earthquakesalong the MexicanGuinea—Solomon Island region. Tectonophysics,57: 7—34.
166
subduction zone: A reexamination. Bull. Seismol.Soc.Am., the Aleutians. J. Geophys. Res., 76: 8021—8041.71: 827—843. Sykes, L.R. and Quittmeyer, R.C., 1981.Repeat times of great
Singh, S.K., Suarez,G. and Dominguez,T., 1985. The Oaxaca, earthquakesalong simple plate boundaries. In: D.W. Simp-Mexico, earthquake of 1931: lithosphencnormal faulting in son and P.G. Richards (Editors), Earthquake Prediction, anthe subducted Cocosplate. Nature, 317: 56—58. International Review. Am. Geophys. Union, Washington,
Sleep, N.H., 1973. Teleseismic P-wavetransmissionthrough DC, pp. 217—247.slabs.Bull. Seismol. Soc.Am., 63: 1349—1373. Sykes,L.R., McCann, W.R. and Kafka, A.L., 1982. Motion of
Sleep,N.H., 1979. The double seismiczonein downgoing slabs the Caribbean plate during the last 7 my. and implicationsand the viscosity of the mesosphere.J. Geophys. Res., 84: for earlier Cenozoic movements. J. Geophys. Res., 87:4565—4571. 10656—10676.
Soedarmo,P.D., 1973. Focal mechanismsof earthquake gener- Tsumara, K., 1973. Microearthquake activity in the Kantoating stress system in the Indonesia region. Presented at district. In: Special Publication for the 50th Anniversary ofSymposium on RecentCrustal Movements, Bandung. the Great Kanto Earthquake, Earthquake Research In-
Spence, W., 1977. The Aleutian arc: Tectonic blocks, episodic stitute, Tokyo, pp. 67—87.subduction, strain diffusion, and magma generation. J. Uyeda, S. and Kanamori, H., 1979. Back-arc opening and the
Geophys. Res., 82: 213—230. mode of subduction. J. Geophys. Res., 80: 1053—1064.Spence,W., 1986. The 1977 Sumbaearthquake series:evidence Vassiliou, M.S., 1984. The state of stress in subducting slabsas
for slab pull force acting at a subduction zone.J. Geophys. revealed by earthquakesanalyzed by moment tensor inver-Res., 91: 7693—7701. sion. Earth Planet. Sci. Lett., 69: 195—202.
Spence, W., 1987. Slab pull and the seismotectonicsof sub- Vassiliou, MS., Hager, B.H. and Raefsky,A., 1984.Thedistri-ducting lithosphere. Rev. Geophys.,25: 55—70. bution of earthquakes with depth and stressin subducting
Stauder, W., 1968. Tensional character of earthquake foci slabs.J. Geodynamics,1: 11—28.
beneath the Aleutian trench with relation to sea-floor Veith, K., 1974.The relationship of island and seismicity tospreading. J. Geophys. Res.,73: 7693—7701. plate tectonics, Ph.D. thesis, South Methodist University,
Stauder, W., 1972. Fault motion and spatially bounded char- Dallas, TX.acter of earthquakes in Amchitka Pass and the Delarof Veith, K., 1977. The nature of the dual zone of seismicity inislands.J. Geophys. Res., 77: 2072—2080. the Kuriles arc. EOS, 58: 1232.
Stauder, W., 1973. Mechanisms and spatial distribution of the Vlaar, N.J. and Wortel, M.J.R., 1976. Lithospheric aging,Chilean earthquakes with relation to subduction of the instability, and subduction. Tectonophysics,32: 331—351.oceanic plate. J. Geophys. Res., 78: 5033—5061. Welc, J.L. and Lay, T., 1987. The sourcerupture processof the
Stauder, W., 1975. Subduction of the Nazca plate under Peru great Banda Sea earthquake of November 4, 1963. Phys.as evidenced by focal mechanismsand seismicity. J. Geo- Earth Planet. Inter., 5: 242—254.phys. Res., 80: 1053—1064. Wesnousky, 5G., Astiz, L. and Kanamon, H., 1986. Earth-
Stauder, W. and Bollinger, GA., 1966. The S-waveproject for quake multiplets in the Southeastern Solomon Islands. Phys.focal mechanism studies: Earthquakes of 1963. Bull. Seis- Earth Planet. Inter., 44: 304—318.mol. Soc.Am., 56: 1363—1371. Wickens,A.J. and Hodgson, J.H., 1967. Computer re-evalua-
Stauder, W. and Maulchin, L., 1976.Fault motion in the larger tion of earthquake mechanism solutions 1922—1962. PubI.
earthquakes of the Kunle—Kamchatka arc and the Dom. Obs., Ottawa, 33: 1—560.Kunle—Hokkaidocorner.J. Geophys. Res.,81: 2297—2308. Wortel, R., 1982. Seismicity and rheology of subducted slabs.
Stein, S., Engeln,iF., Wiens, D.A., Speed,R.C. and Fujita, K., Nature, 296: 553—555.1982. Subduction, seismicity and tectonics in the Lesser Yamaoka, K., Fukao, Y. and Kumazawa, M., 1986. Spherical
Antilles arc. J. Geophys. Res.,87: 8642—8664. shell tectonics: effects of sphencity and inextensibility onSuzuki, S. and Motoya, Y., 1978.Microearthquake activity and the geometryof the descendinglithosphere. Rev. Geophys.,
a two-layered deep seismic zone in the Hokkaido region. 24: 27—54.Abstr. Progr. Seismol. Soc. Jpn., 1: 32 (in Japanese). Yoshii, T., 1977. Structure of the Earth’s crust and mantle in
Sykes,L.R., 1966. Seismicityand deep structure of island arcs, northeast Japan. Kagaku, 47: 170—176.J. Geophys. Res., 71: 2981—3006. Yoshii, T., 1979. A detailed cross section of the deep seismic
Sykes, L.R., 1971. Aftershock zones of great earthquakes, zone beneathJapan. Tectonophysics,55: 349—360.seismicity, gaps and earthquake prediction for Alaska and