Cenozoic tectonics of SE Asia and Australasia 47

From: Petroleum Systems of SE Asia and Australasia. pp. 47-62.Edited by J. V. C. Howes & R. A. Noble1997, Indonesian Petroleum Association, Jakarta.

INDONESIAN PETROLEUM ASSOCIATIONProceedings of the Petroleum Systems of SE Asia and Australasia Conference, May 1997

CENOZOIC TECTONICS OF SE ASIA AND AUSTRALASIA

Robert HallSE Asia Research Group, University of London

quency and rapidity of changes in regional tecton-ics.

INTRODUCTION

The region of SE Asia and Australasia includesexamples of almost every plate tectonic configura-tion at different stages in the Wilson cycle betweenrifting and continental collision. It is the only placeon Earth where we can observe arcs in collision,one of the few places where an oceanic spreadingcentre is actively propagating into continentalcrust, and includes areas with the highest globalrates of plate convergence and separation. But howuseful is plate tectonics in describing the evolutionof the region? It is good at describing interactionbetween slowly moving, large plates with rela-tively simple geometries but its application to theSE Asia-Australasian region is more difficult be-cause of the number of small plates required to ac-count for the region’s development. Furthermore,many ideas about orogenic evolution through theWilson cycle are based on the early, simple andpowerful plate tectonic concepts reproduced intext-books but some of the axioms have changed oradvances in knowledge have made them invalid.For example, ophiolites are rarely formed in nor-mal ocean basins, continental crust is now knownto be subductable, convergence may continue longafter continent-continent collision, and much de-formation of continents may be distributed ratherthan concentrated at plate margins.

In discussing regional tectonics it is necessary tobe aware of the many problems. How reliably andfor how long can present plate motions, for exam-ple from GPS measurements, be projected into thepast? How much of the record is missing? Subduc-tion leads to destruction and formation of marginal

ABSTRACT

A plate tectonic model for the development of theregion of SE Asia and Australasia is presented andits implications are summarised. The complexityof the present-day tectonics of the region and theobservable rates of plate motions indicate that ma-jor oceans, or multiple small oceans, have closedduring the Cenozoic, and that the configuration ofthe region has changed significantly during thistime. Despite the long-term convergence there hasbeen frequent opening of marginal basins, and ex-tension related to strike-slip faults resulting frompartitioning of oblique convergence at plateboundaries. Present-day plate motions, based forexample on GPS measurements and seismicity, il-lustrate the complexity of processes but appear tohave little relevance in understanding the long-term kinematic development of the region. Thereare three important periods in regional develop-ment: at about 45 Ma, 25 Ma and 5 Ma. At thesetimes plate boundaries and motions changed, prob-ably as a result of major collision events. Indenta-tion of Asia by India may have modified the Eura-sian continent but there is little indication that In-dia has been the driving force of tectonics in SEAsia. The movement of Australia northwards hascaused rotations of SE Asia blocks and accretionof microcontinental fragments to SE Asia. Since25 Ma the oceanic region east of Eurasia has beendriven by motion of the Pacific plate. To improveour tectonic models, detail is needed which can becompiled from proprietary data, such as coastline,shelf edge, age and lithofacies information, heldby companies. Improved dating of events is re-quired in order to identify regional events and theirconsequences and identify the processes that causethe effects. Few sedimentary basins in the regionwill fit into simple basin models because of the fre-

48 R. Hall

basins at the major plate boundaries during shorttime intervals and these basins may be impossibleto reconstruct. There are considerable difficultiesin identifying the importance of strike-slip mo-tions. What is the relevance of plate tectonics tosmaller (e.g. basin) scale reconstructions? Distrib-uted deformation, for example of Sundaland, maynot be amenable to simple plate tectonic analysis.There are important problems in identifying verti-cal axis rotations on a regional scale.

A logical parsimony is necessary in reconstruc-tion, limiting the number of plates and their detail,and observations of the present-day tectonics ofthe region show us that any plate model must be anover-simplification. Nonetheless, accepting itslimitations, analysis of the region as rigid platescan still give important insights. It is important towork progressively back in time, recognising theincreasing uncertainty in older reconstructions; re-alistic Mesozoic reconstructions are not currentlypossible. 2-D plate tectonic cartoons are no longeradequate descriptions or tools for understanding. Itis essential to test plate tectonic models by usingmaps which can be examined at short time inter-vals. Animations (e.g. Yan and Kroenke, 1993;Hall, 1996) can expose flaws in models, and majorgaps in our knowledge, but also help identify trulyregional events. Compilations, for example oflithofacies, do not make sense unless regional platemovements are considered. This paper attempts tosummarise the regional tectonic development ofSE Asia and Australasia based on such a plate tec-tonic model which has been animated using 1 Matime-slices. For the petroleum industry these re-constructions may help in understanding the devel-opment of sedimentary basins, and the distributionof petroleum resources, by identifying importantcontrols on their tectonic setting and the timing ofregionally important events.

Basis for the Model

The reconstructions were made using the ATLAScomputer program (Cambridge Paleomap Serv-ices, 1993). In the ATLAS model the motions ofthe major plates are defined relative to Africa andits movement is defined relative to magnetic north.There has been little Cenozoic motion of Eurasiaand it remains in a similar position in all the recon-structions, although there are small movements ofEurasia due to the plate circuit used in the ATLASmodel, particularly for the last 5 Ma. Thereforethere are minor differences compared to recon-

structions which keep Eurasia fixed in its presentposition (Rangin et al., 1990; Lee and Lawver,1994).

Reconstructions of SE Asia and Australasia (Fig.1) shown on a global projection are presented at 10Ma intervals for the period 50-10 Ma (Figs. 2 to 6).More than 100 fragments are currently used, andmost retain their current size in order that they re-main recognisable. During the 50 Ma period frag-ments represented may have changed size andshape or may not have existed, both for arc andcontinental terranes. Thus, the plate model canonly be an approximation. Some of the elements ofthe model are deliberately represented in a stylisticmanner to convey the processes inferred ratherthan display exactly what has happened, for exam-ple, the motion of the terranes of north NewGuinea. The reconstructions presented here extendthe model developed previously for SE Asia andthe reader is referred to Hall (1995, 1996) for amore complete account of the assumptions anddata used in reconstructing that region. Yan andKroenke (1993) have produced an animated recon-struction of the SW Pacific. Some important dif-ferences between this model and theirs resultsfrom the choice of reference frames; they used thehotspot frame whereas this reconstruction uses apalaeomagnetic reference frame. Other differencesresult from different interpretations of geologicaldata. The model presented here shows the move-ment of India and new interpretations of the SWPacific; a few important references are cited in thetext but in the space available here the reasons forthe differences from previous models and the justi-fication of the reconstruction must await a morecomplete account which will be published else-where. What follows below is a brief account ofthe model and its major features, with discussionof its principal implications.

RECONSTRUCTIONS

Configuration at 50 Ma

At 50 Ma (Fig. 2) India and Australia were sepa-rate plates although their motions were not greatlydifferent. Transform faults linked the slow-spread-ing Australia-Antarctic and the fast spreading In-dia-Australia spreading centres. Older parts of theEast Sulawesi ophiolite probably formed at the In-dia-Australia ridge. India collided with Asia in theearly Tertiary but there remains considerable con-troversy about the exact age of collision, and its

Cenozoic tectonics of SE Asia and Australasia 49

consequences (Packham, 1996; Rowley, 1996).The position of the Eurasian margin and the extentof Greater India are major problems. The recon-struction shown in Fig. 2 shows a conservative es-timate, and since India-Asia collision began atabout 50 Ma this implies that the Asian margin ex-tended south to at least 30°N. Many of the tectonicevents in SE Asia are commonly attributed to theeffects of Indian indentation into Asia and the sub-sequent extrusion of continental fragments east-wards along major strike-slip faults. Despite thegreat attraction of this hypothesis and the spec-tacular evidence of displacements on the RedRiver Fault (Tapponnier et al., 1990) the predic-tions of major rotations, southeastward extrusionof fragments, and the timing of events (Tapponnieret al., 1982), remain poorly supported by geologi-cal evidence in SE Asia.

The east Eurasian continental margin was orientedbroadly NE-SW. From Japan northwards Asia wasbounded by an active margin. Taiwan, Palawanand the now extended crust of the South China Seamargins formed a passive margin, established dur-ing Cretaceous times. Sundaland was separatedfrom Eurasia by a wide proto-South China Seaprobably floored by Mesozoic ocean crust. Thesouthern edge of this ocean was a passive conti-nental margin north of a continental promontoryextending from Borneo to Zamboanga. The Malaypeninsula was closer to Indochina and the Malay-Sumatra margin was closer to NNW-SSE. Becauserotation of Borneo is accepted here the reconstruc-tion differs from those of Rangin et al. (1990) andDaly et al. (1991) who infer a margin orientedcloser to E-W. I see no evidence to support the al-most E-W orientation of the Sundaland margin inthe region of Sumatra as shown on these and manyother reconstructions (e.g. Briais et al., 1993;Hutchison, 1996). Furthermore, such models havemajor difficulties in explaining the amount, timingand mechanism of rotation required to moveSumatra from an E-W to NW-SE orientation. WestSumatra includes arc and ophiolitic materialaccreted in the Cretaceous. East Borneo and WestSulawesi appear to be underlain by accreted arcand ophiolitic material as well as continental crustwhich may be early-rifted Gondwana fragments.This material had been accreted during the Creta-ceous and may have resulted in a highly thickenedcrust in this part of Sundaland, possibly sustainedby subduction.

Australia was essentially surrounded by passive

margins on all sides. To the west the passive mar-gin was formed in the Late Jurassic and Fig. 2 pos-tulates a failed rift, possibly floored by oceaniccrust on the site of the present-day Banda Sea, par-tially separating a Bird’s Head microcontinentfrom Australia. Mesozoic lithosphere was presentnorth of the Bird’s Head south of the active Indian-Australian spreading centre. Further east in the Pa-cific, Indian and Australian oceanic lithospherehad been subducting northwards beneath theSepik-Papuan arc before about 55 Ma. The NewGuinea Mesozoic passive margin had collidedwith this intra-oceanic arc in the early Eocenecausing emplacement of the Sepik, Papuan andNew Caledonia ophiolites. Subsequently, most ofthe New Guinea margin remained a passive marginduring the Paleogene but the oceanic crust to thenorth is inferred to have formed during theMesozoic in an intra-oceanic marginal basin be-hind the Sepik-Papuan arc. The position of the eastAustralia-Pacific margin is also uncertain. Tasmanand Coral Sea opening had probably been drivenby subduction but if so the site of subduction wasconsiderably east of the Australian continent, be-yond the Loyalty Rise and New Caledonia Rise.Spreading had ceased in both basins by about 60Ma (Paleocene). The history of this region remainspoorly known since it is almost entirely submarineand magnetic anomalies in this area are poorly de-fined.

Java and West Sulawesi were situated above atrench where Indian plate lithosphere wassubducting towards the north. The character of thisboundary is shown as a simple arc but may haveincluded marginal basins and both strike-slip andconvergent segments depending on its local orien-tation. Extending plate boundaries into the Pacificis very difficult. A very large area of the West Pa-cific has been eliminated by subduction since 50Ma which will continue to cause major problemsfor reconstructions. However, there is clear evi-dence that this area resembled the present-dayWest Pacific in containing marginal basins, intra-oceanic arcs and subduction zones. The Java sub-duction system is linked east into Pacific intra-oce-anic subduction zones required by the intra-oce-anic arc rocks within the Philippine Sea plate;parts of the east Philippines, the West Philippinebasin and Halmahera include arc rocks dating backat least to the Cretaceous. North of the PhilippineSea plate there was a south-dipping subductionzone at the southern edge of a Northern NewGuinea plate.

50 R. Hall

50-40 Ma

Whatever the timing of India-Asia collision, aconsequence was the slowing of the rate of plateconvergence after anomaly C21 and a majorchange in spreading systems between anomalyC20 and C19 at about 42 Ma. India and Australiabecame one plate during this period (Figs. 2 and 3)and the ridge between them became inactive.Northward subduction of Indian-Australianlithosphere continued beneath the Sunda-Java-Sulawesi arcs although the direction of conver-gence may have changed. Rift basins formedthroughout Sundaland, but the timing of their ini-tial extension is uncertain because they containcontinental clastics which are poorly dated, andtheir cause is therefore also uncertain. They mayrepresent the consequences of oblique conver-gence or extension due to relaxation in the over-riding plate in response to India-Asia collision, en-hanced by slowing of subduction, further influ-enced by older structural fabrics.

The Java-Sulawesi subduction system continuedinto the West Pacific beneath the east Philippinesand Halmahera arcs. Further east, the direction ofsubduction was southward towards Australia andthis led to the formation of a Melanesian arc sys-tem. Subduction began beneath Papua NewGuinea with major arc growth producing the olderparts of the New Britain, Solomons and Tonga-Kermadec systems, leading to development of ma-jor marginal basins in the SW Pacific whose rem-nants probably survive only in the Solomon Sea.This model postulates the initial formation of thesearcs at the Papuan-east Australian margin as previ-ously suggested by Crook and Belbin (1978) fol-lowing subduction flip, rather than by initiation ofintra-oceanic subduction within the Pacific plateoutboard of Australia as suggested by Yan andKroenke (1993). The evidence for either proposalis limited but this model has the simplicity of a sin-gle continuous Melanesian arc.

During this interval there were major changes inthe Pacific. The Pacific plate is widely said to havechanged its motion direction at 43 Ma, based onthe age of the bend in the Hawaiian-Emperorseamount chain, although this view has recentlybeen challenged by Norton (1995) who attributesthe bend to a moving hotspot which became fixedonly at 43 Ma. There is regional boniniticmagmatism, the cause of which is still not under-stood, which resulted in massive outpouring of in-

tra-oceanic magmatic products, much greater involume than typical arc magma production rates(Stern and Bloomer, 1992). Fig. 2 implies thatboninitic magmatism was linked to subduction ofthe Pacific-Northern New Guinea ridge. Thisformed the Izu-Bonin and Mariana arc systems andthe Philippine Sea plate was a recognisable entityby the end of this period. There were major rota-tions of the Philippine Sea plate between 50 and 40Ma and the motion history of this plate (Hall et al.,1995) provides an important constraint on devel-opment of the eastern part of SE Asia. The WestPhilippine Basin, Celebes Sea, and MakassarStrait opened as single basin within the PhilippineSea plate although the reconstructions probablyunderestimate the width of the Makassar Strait andCelebes Sea, which may have been partiallysubducted in the Miocene beneath west Sulawesi.

The opening of the West Philippine-Celebes Seabasin required the initiation of southward subduc-tion of the proto-South China Sea beneath Luzonand the Sulu arc. It is this subduction which causedrenewed extension along the South China margin,driven by slab-pull forces due to subduction be-tween eastern Borneo and Luzon, and later led tosea-floor spreading in the South China Sea, ratherthan indentor-driven tectonics.

40-30 Ma

In this interval (Figs. 3 and 4) the spreading of themarginal basins of the West and SW Pacific con-tinued. Indian ocean subduction continued at theSunda-Java trenches, and also beneath the arc ex-tending from Sulawesi through the east Philippinesto Halmahera. Sea floor spreading continued in theWest Philippine-Celebes Sea basin until about 34Ma. This spreading centre may been linked tobackarc spreading of the Caroline Sea whichformed from about 40 Ma due to subduction of thePacific plate. The Caroline Ridge is interpreted inpart as a remnant arc resulting from Caroline Seabackarc spreading, and the South Caroline arc ulti-mately became the north New Guinea arc terranes.By 30 Ma the Caroline Sea was widening above asubduction zone at which the newly-formed Solo-mon Sea was being destroyed as the Melanesianarc system migrated north. The backarc basins inthe SW Pacific were probably very complex, as in-dicated by the anomalies in the South Fiji Basin,and will never be completely reconstructed be-cause most of these basins have been subducted.

Cenozoic tectonics of SE Asia and Australasia 51

The Philippines-Halmahera arc remained station-ary, so spreading in the West Philippine-CelebesSea basin maintained subduction between NE Bor-neo and north of Luzon. The pull forces of thesubducting slab therefore account for stretching ofthe Eurasian margin north of Palawan, and laterdevelopment of oceanic crust in the South ChinaSea which began by 32 Ma. This model incorpo-rates the 500-600 km movement on the Red Riverfault postulated by Briais et al. (1993). In contrast,the indentor model does not account for stretchingat the leading edge of the extruded blocks, such asIndochina, or the normal faulting east of Vietnamoften shown as kinematically linked to the RedRiver Fault system. If these faults were linked, thefaults offshore Vietnam should be a restrainingbend during the phase of left-lateral movementproposed (32-15 Ma) and consequently showthrust faulting rather than the normal faulting ob-served.

The dextral Three Pagodas and Wang Chao faultsare simplified as a single fault at the north end ofthe Malay peninsula. There are a host of faultsthrough this region and a plate tectonic model canonly oversimplify the tectonics of the continentalregions by considering large and simple blockmovements and broadly predicting regional stressfields. The implication of this simple model is thatbasins such as the Malay and Gulf of Thailand ba-sins have a significant component of strike-slipmovement on faults controlling their development.However, they were initiated in a different tectonicsetting, and in a region with an older structural fab-ric (Hutchison, 1996) which influenced their de-velopment.

30-20 Ma

This period of time (Figs. 4 and 5) saw the mostimportant Cenozoic plate boundary reorganisationwithin SE Asia. At about 25 Ma, the New Guineapassive margin collided with the leading edge ofthe east Philippines-Halmahera-New Guinea arcsystem. The Australian margin, in the Bird’s Headregion, was also close to collision with the Eura-sian margin in West Sulawesi and during this inter-val ophiolite was emplaced in Sulawesi. By 30 Mathe Sulawesi margin may have been complex andincluded ocean crust components of differenttypes (MORB, backarc basins). Thus the Sulawesiophiolite probably includes material formedwithin the Indian Ocean (Mubroto et al., 1994) as

well as ocean basins marginal to Eurasia (Monnieret al., 1995).

The arrival of the Australian margin at the subduc-tion zone caused northward subduction to cease.This trapped ocean crust between Sulawesi andHalmahera which first became part of the Philip-pine Sea plate and later the Molucca Sea plate. ThePhilippine Sea plate began to rotate clockwise andthe trapped ocean crust began to subduct beneathSulawesi in the Sangihe arc.

Soon afterwards the Ontong Java plateau collidedwith the Melanesian arc. These two major colli-sions caused a major change in the character ofplate boundaries in the region between about 25and 20 Ma (Early Miocene). They also linked theisland arcs of Melanesia, the New Guinea terranesat the southern Caroline margin, and theHalmahera-Philippines arcs. This linkage seems tohave coupled the Pacific to the marginal basins ofthe West Pacific, and the Caroline and PhilippineSea plates were subsequently driven by the Pacific.They both began to rotate, almost as a single plate,and the Izu-Bonin-Mariana trench system rolledback into the Pacific. Rifting of the Palau-Kyushuridge began, leading first to opening of the PareceVela basin and later to spreading in the Shikokubasin. The change in plate boundaries led to sub-duction beneath the Asian margin.

Subduction beneath the Halmahera-Philippinesarc ceased and the New Guinea sector of the Aus-tralian margin became a strike-slip zone, theSorong Fault system, which subsequently movedterranes of the South Caroline arc along the NewGuinea margin.

Advance of the Melanesian arc system led to wid-ening of the South Fiji basin and Solomon Sea ba-sin (now mainly subducted). At the Three KingsRise subduction seems to have been initiated soonafter ocean crust was formed to the east, allowingthe rise to advance east and spreading to propagatebehind the rise into the Norfolk basin from a triplejunction to the north.

20-10 Ma

The clockwise rotation of the Philippine Sea platenecessitated changes in plate boundaries through-out SE Asia which resulted in the tectonic patternrecognisable today (Figs. 5 and 6). These changes

52 R. Hall

include the re-orientation of spreading in the SouthChina Sea, and the development of new subduc-tion zones at the eastern edge of Eurasia and in theSW Pacific. Continued northward motion of Aus-tralia caused the counter-clockwise rotation ofBorneo. Northern Borneo is much more complexthan shown. There was volcanic activity and build-out of delta and turbidite systems into the proto-South China Sea basin. Major problems includethe source of sediment in the basins surroundingcentral Borneo and the location and timing of vol-canic activity in Borneo. Some of this sedimentmay have been derived from the north across theSunda shelf. Important igneous activity promotedeconomic mineralisation but its tectonic setting isnot clear. The reconstruction exaggerates thewidth of oceanic crust remaining in the westernproto-South China Sea, and much of this area mayhave been underlain by thinned continental crustwhich was thrust beneath Borneo, thus thickeningthe crust and ultimately leading to crustal melting.

The rotation of Borneo was accompanied by coun-ter-clockwise motion of west Sulawesi, andsmaller counter-clockwise rotations of adjacentSundaland blocks. In contrast, the north Malay pe-ninsula rotated clockwise, but remained linked toboth Indochina and the south Malay peninsula.This allowed widening of basins in the Gulf ofThailand, but the rigid simple plate model overes-timates the extension in this region. This extensionwas probably more widely distributed throughoutSundaland and Indochina on many different faults.Here therefore, clockwise rotations are not attrib-uted to indentation as the limited evidence datingthe rotations also implies.

North Sumatra rotated counter-clockwise withsouth Malaya, and as the rotation proceeded theorientation of the Sumatran margin changed withrespect to the Indian plate motion vector. The con-sequent increase in the convergent component ofmotion, taken up by subduction, may have in-creased magmatic activity in the arc and weakenedthe upper plate, leading to formation of the dextralSumatran strike-slip fault system taking up the arc-parallel component of India-Eurasia plate motion.

East of Borneo, the increased rate of subductioncaused arc splitting in the Sulu arc and the SuluSea opened as a back-arc basin (Silver and Rangin,1991) south of the Cagayan ridge. The Cagayanridge then moved northwards, eliminating the east-ern proto-South China Sea, to collide with the

Palawan margin. New subduction had also begunat the west edge of the Philippine Sea plate belowthe north Sulawesi-Sangihe arc which extendednorth to south Luzon. This was a complex zone ofopposed subduction zones linked by strike-slipfaults. The Philippine islands and Halmahera werecarried with the Philippine Sea plate towards thissubduction zone. North of Luzon, sinistral strike-slip movement linked the subducting west marginof the Philippine Sea plate to subduction at theRyukyu trench. Collision of Luzon and theCagayan ridge with the Eurasian continental mar-gin in Mindoro and north Palawan resulted in ajump of subduction to the south side of the SuluSea. Southward subduction beneath the Sulu arccontinued until 10 Ma. The remainder of the Phil-ippines continued to move with the Philippine Seaplate, possibly with intra-plate strike-slip motionand subduction resulting in local volcanic activity.At the east edge of the Philippine Sea plate spread-ing terminated in the Shikoku basin.

As a result of changing plate boundaries fragmentsof continental crust were emplaced in Sulawesi onsplays at the western end of the Sorong Fault sys-tem. The earliest fragment to collide is inferred tohave been completely underthrust beneath WestSulawesi and contributed to later crustal melting.The potassic magmatism of Sulawesi (Polvé et al.,1997) is not typical of an arc setting and may bedue to extension and rifting of over-thickenedcrust. Later, the Tukang Besi platform separatedfrom the Bird’s Head and was carried west on thePhilippine Sea plate to collide with Sulawesi.Locking of splays of the Sorong fault caused sub-duction to initiate at the eastern margin of theMolucca Sea, producing the Neogene Halmaheraarc. Thus the Molucca Sea became a separate plateas the double subduction system developed.

After the collision of the Ontong Java plateau withthe Melanesian arc the Solomons became attachedto the Pacific plate. Westward subduction beganon the SW side of Solomon Sea, beneath easternNew Guinea, eliminating most of Solomon Seaand resulting in the formation of Maramuni arcsystem. As the Solomon Sea was eliminated theSouth Caroline arc began to converge on the northNew Guinea margin and the arc terranes weretranslated west in the major left-lateral shear zone,probably accompanied by rotation. In the southernpart of the Solomons Sea subduction was in theopposite direction (eastward) and created the NewHebrides arc system. Spreading ceased in the

Cenozoic tectonics of SE Asia and Australasia 53

South Fiji basin.

10-0 Ma

At the beginning of this period SE Asia was largelyrecognisable in its present form (Figs. 6 and 1).Rotation of Borneo was complete. This, with colli-sion in the central Philippines and Mindoro, andcontinued northward movement of Australia, re-sulted in reorganisation of plate boundaries and in-tra-plate deformation in the Philippines. TheLuzon arc came into collision with the Eurasianmargin in Taiwan. This may be the cause of themost recent regional change in plate motions atabout 5 Ma. The Philippine Sea plate rotation polemoved north from a position east of the plate;clockwise rotation continued but the change inmotion caused re-orientation of existing, and de-velopment of new, plate boundaries. Subductioncontinued at the Manila, Sangihe and Halmaheratrenches, and new subduction began at the Negrosand Philippine trenches. These subduction zoneswere linked by strike-slip systems active within thePhilippines and this intra-plate deformation cre-ated many very small fragments which are difficultto describe using rigid plate tectonics.

The Molucca Sea continued to close by subductionon both sides. At present the Sangihe forearc hasoverridden the northern end of the Halmahera arc,and is beginning to over-thrust west Halmahera. Inthe Sorong fault zone, accretion of Tukang Besi toSulawesi locked a strand of the fault and initiated anew splay south of the Sula platform. The Sulaplatform then collided with the east arm ofSulawesi, causing rotation of the east and northarms to their present position, leading to south-ward subduction of the Celebes Sea at the northSulawesi trench.

The Eurasia-Philippine Sea plate-Australia triplejunction was and remains a zone of microplates butwithin this contractional setting extension contin-ued in the Banda Sea. The Bird’s Head movednorth relative to Australia along a strike-slip faultat the Aru basin edge. Mesozoic ocean crust northof Timor was eliminated at the eastern end of theJava trench by continued northern motion of Aus-tralia which brought the Australian margin intothis trench as the volcanic inner Banda arc propa-gated east. Seram began to move east requiringsubduction and strike-slip motion at the edges ofthis microplate. Since 5 Ma the southern Banda

Sea has extended to its present dimensions andcontinental fragments are now found in the BandaSea ridges within young volcanic crust. The BandaSea is interpreted to be very young as suggested byHamilton (1979) and others.

In west Sundaland, partitioning of convergence inSumatra into orthogonal subduction and strike-slipmotion effectively established one or moreSumatran forearc sliver plates. Extension on thestrike-slip system linked to the spreading centre inthe Andaman Sea (Curray et al., 1979). WithinEurasia reversal of motion on the Red River sys-tem may have been one consequence of the re-gional change in plate motions.

Opening of the Ayu trough separated the Carolineplate and Philippine Sea plate, although the rate ofseparation at this spreading centre was very low.North of the Bird’s Head, and further east in NewGuinea, transpressional movements were markedby deformation of arc and ophiolite slivers sepa-rated by sedimentary basins. Progressive westwardmotion of the South Caroline arc within the left-lateral transpressional zone led to docking of thenorth New Guinea terranes. This caused cessationof southward subduction of the Solomon Sea platebut resulted in its northward subduction beneathNew Britain. The New Britain subduction led torapid spreading in Woodlark basin as a conse-quence of slab-pull forces and rapid ripping openof continental crust beneath the Papuan peninsulaaccompanied by major exhumation of the middleand lower crust to form core complexes.

Elimination of most of the remaining Solomonsmarginal basin by eastward subduction led to for-mation of the New Hebrides arc and opening of theNorth Fiji basins. There was significant local rota-tion of small plates illustrated by Fiji. This regionalso illustrates the complexity of backarc basinsprocesses: ridge jumps, propagating rifts, re-orien-tation of spreading and major rotations (90°) invery short periods (<10 Ma).

CONCLUSIONS

Studies of present-day motions, based for exampleon GPS measurements and seismicity, demon-strate the complexity of regional tectonics and theextremely high rates of plate tectonic processes. Itis a salutary lesson to discover that in many partsof the region the position of modern plate bounda-

54 R. Hall

ries is not known, or they do not connect, and thereremains considerable uncertainty about how to de-scribe the modern region in plate tectonic terms(e.g. McCaffrey, 1996). Our understanding ofmodern tectonics should therefore warn us that theregion must have changed considerably, at timesvery rapidly, during the last 50 Ma. We should beaware that much evidence has been lost during col-lision and subduction, which makes any recon-struction an over-simplification. Present-day platemotions appear to have little relevance in under-standing the long-term kinematic development ofthe region. They cannot be projected back into thepast very far, at most 5 Ma and probably much less.Major plate boundaries may have remained in thesame position but have changed character. Newboundaries have been created at different times.To illustrate these points, the evolution of the NewGuinea orogenic belt cannot be understood fromdetermination of present-day Australia-Pacificmotion and plate boundaries, and indeed an analy-sis in terms of the motions of these two plates tellsus practically nothing about its development.

There are three important periods in regional de-velopment. At about 45 Ma plate boundarieschanged, probably as a result of India-Asia colli-sion. Indentation of Asia by India may have modi-fied the Eurasian continent but much more detail isrequired of the timing of fault movements and theamounts of displacements before Sundaland canbe adequately understood. However, even after in-corporating the best documented evidence for ex-trusion (Briais et al., 1993) the reconstructionsshow little indication that India has been the driv-ing force of tectonics in SE Asia. The second ma-jor period is around 25 Ma when plate boundariesand motions changed again, perhaps due to colli-sion between the north Australian margin and arcsto the north. This, together with collision of theMelanesian arcs and the Ontong Java plateauchanged the tectonics of the oceanic-arc regioneast of Asia (Philippines, Celebes Sea, Sulu Sea,Philippine Sea, Caroline Sea, north New Guinea,New Britain, Solomons, Tonga). Since 25 Ma theregion east of Eurasia has been driven by motion ofthe Pacific plate. The rotation of regions within themarginal basins seems to be an expression of intra-Pacific plate deformation due to roll-back of theIzu-Bonin-Mariana trench. It appears that the ca-pacity for this may have reached its limit at about 5Ma, again possibly as a consequence of arc-conti-nent collision in Taiwan, and plate motions andboundaries changed again.

In the long-term a greater systematic collection ofregional data sets is required to understand tec-tonic development. Palaeomagnetic studies areone example of the type of work which only be-comes valuable when there are large data-sets forlarge areas, for example to distinguish regionalfrom local vertical axis rotations. The requirementfor conclusions from local studies in many casesprevents these regional data-sets from being ac-quired, and often leads to over-interpretation of re-sults. Dating of rocks and events, both isotopic andbiostratigraphic, is in many areas still inadequate.Most tectonic models are too restricted in area tocontribute to a regional understanding. Such lim-ited models are often not useful since they solvelocal problems by ignoring or moving larger prob-lems outside the area of interest. To go further, de-tail is needed which can be compiled from propri-etary data, such as coastline, shelf edge, age andlithofacies information, held by companies. In par-ticular the display of uplift and subsidence, andtiming of magmatic events, on reconstructionswould permit examination of the relationships be-tween vertical motions and plate movements andhelp in identifying underlying processes. The iden-tification of cause and effect is difficult or impos-sible at present because there is still a tendency toidentify an event in the area of interest as a ‘cause’.Global reconstructions put these ‘causes’ into con-text. As far as sedimentary basins are concerned itis necessary to be alert to the probability that fewbasins have a simple history. They are unlikely tohave the characteristics of a simple classical basintype because of the changes in regional tectonicsand this will apply to any basin with a life greaterthan 10 Ma.

ACKNOWLEDGEMENTS

Financial support has been provided by NERC, theRoyal Society, the London University Central Re-search Fund, and the London University SE AsiaResearch Group currently supported by Arco, Ca-nadian Occidental, Exxon, Lasmo, Mobil, UnionTexas and Unocal. Work in Indonesia has been fa-cilitated by GRDC, Bandung and Directors includ-ing H. M. S. Hartono, M. Untung, R. Sukamto andI. Bahar. I thank Kevin Hill for considerable helpand discussion during the reconstruction of the SWPacific.

Cenozoic tectonics of SE Asia and Australasia 55

REFERENCES

Briais, A., Patriat, P. and Tapponnier, P. 1993.Updated interpretation of magnetic anomalies andseafloor spreading stages in the South China Sea:implications for the Tertiary tectonics of SoutheastAsia. Journal of Geophysical Research, 98, 6299-6328.

Cambridge Paleomap Services. 1993. ATLAS ver-sion 3.3. Cambridge Paleomap Services, P.O. Box246, Cambridge, U.K.

Crook, K.A.W. and Belbin, L. 1978. The South-west Pacific area during the last 90 million years.Journal of the Geological Society of Australia, 25,23-40.

Curray, J. R., Moore, D. G., Lawver, L. A.,Emmel, F. J., Raitt, R. W., Henry, M. andKieckheffer, R. 1979. Tectonics of the AndamanSea and Burma. American Association of Petro-leum Geologists Memoir, 29, 189-198.

Daly, M. C., Cooper, M. A., Wilson, I., Smith, D.G. and Hooper, B. G. D. 1991. Cenozoic plate tec-tonics and basin evolution in Indonesia. Marineand Petroleum Geology, 8, 2-21.

Hall, R. 1995. Plate tectonic reconstructions of theIndonesian region. Proceedings of the IndonesianPetroleum Association 24th Annual Convention,71-84.

Hall, R. 1996. Reconstructing Cenozoic SE Asia.In: Hall, R. and Blundell, D. J. (eds.) Tectonic Evo-lution of SE Asia. Geological Society of LondonSpecial Publication, 106, 153-184.

Hall, R., Fuller, M., Ali, J. R. and Anderson, C. D.1995. The Philippine Sea Plate: magnetism and re-constructions. American Geophysical UnionMonograph, 88, 371-404.

Hamilton, W. 1979. Tectonics of the Indonesianregion. USGS Professional Paper, 1078, 345 pp.

Hinz, K., Block, M., Kudrass, H. R. and Meyer, H.1991. Structural elements of the Sulu Sea.Geologische Jahrbuch, A127, 883-506.

Hutchison, C. S. 1996. South-East Asian Oil, Gas,Coal and Mineral Deposits. Clarendon Press, Ox-ford. 265 pp.

Katili, J. A. 1975. Volcanism and plate tectonics inthe Indonesian island arcs. Tectonophysics, 26,165-188.Lee, T-Y. and Lawver, L. A. 1994. Cenozoic platetectonic reconstruction of the South China Sea re-gion. Tectonophysics, 235, 149-180.

McCaffrey, R. 1996. Slip Partitioning at Conver-gent Plate Boundaries of SE Asia. In: Hall, R. &Blundell, D. J. (eds.), Tectonic Evolution of SEAsia, Geological Society of London Special Publi-cation, 106, 3-18.

Monnier, C., Girardeau, J., Maury, R. C. andCotten, J. 1995. Back arc basin origin for the EastSulawesi ophiolite (eastern Indonesia). Geology,23, 851-854.

Norton, I. O. 1995. Tertiary relative plate motionsin the North Pacific: the 43 Ma non-event. Tecton-ics, 14, 1080-1094.

Packham, G. 1996. Cenozoic SE Asia: Recon-structing its aggregation and reorganisation. In:Hall, R. & Blundell, D. J. (eds.), Tectonic Evolu-tion of SE Asia, Geological Society of LondonSpecial Publication, 106, 123-152.

Polvé, M. Maury, R. C., Bellon, H., Rangin, C.,Priadi, B., Yuwono, S., Joron, J. L. & SoeriaAtmadja, R. 1997. Magmatic evolution ofSulawesi,(Indonesia): constraints on the Cenozoicgeodynamic history of the Sundaland active mar-gin. Tectonophysics, in press.

Rangin, C., Jolivet, L., Pubellier, M. 1990. A sim-ple model for the tectonic evolution of southeastAsia and Indonesia region for the past 43 m. y. Bul-letin de la Société géologique de France, 8 VI,889-905.

Rowley, D. B. 1996. Age of initiation of collisionbetween India and Asia: a review of stratigraphicdata. Earth and Planetary Science Letters, 145, 1-13.

Silver, E. A. and Rangin, C. 1991. Leg 124 tec-tonic synthesis. In: Silver, E. A., Rangin, C., vonBreymann, M. T. et al., Proceedings of the ODP,Scientific Results, 124, 3-9.

Stern, R. J. and Bloomer, S. H. 1992. Subductionzone infancy: examples from the Eocene Izu-

56 R. Hall

KEY TO FIGURE 1

Marginal Basins Tectonic features

A Japan Sea Ba Banda ArcB Okinawa Trough BH Bird’s HeadC South China Sea Ca Cagayan ArcD Sulu Sea Fj FijiE Celebes Sea Ha Halmahera ArcF Molucca Sea IB Izu-Bonin ArcG Banda Sea Ja Japan ArcH Andaman Sea Lo Loyalty IslandsJ West Philippine Basin Lu Luzon ArcK Shikoku Basin Mk Makassar StraitL Parece Vela Basin Mn Manus IslandM Mariana Trough NB New Britain ArcN Ayu Trough NC New CaledoniaP Caroline Sea NH New Hebrides ArcQ Bismarck Sea NI New IrelandR Solomon Sea Nng North New Guinea TerranesS Woodlark Basin Pa Papuan OphioliteT Coral Sea Pk Palau-Kyushu RidgeU Tasman Sea Ry Ryukyu ArcV Loyalty Basin Sa Sangihe ArcW Norfolk Basin Se Sepik ArcX North Fiji Basin So Solomons ArcY South Fiji Basin Sp Sula PlatformZ Lau Basin Su Sulu Arc

TK Three Kings RiseTo Tonga ArcTu Tukang Besi Platform

Bonin-Mariana and Jurassic California arcs. Bulle-tin of the Geological Society of America, 104,1621-1636.

Tapponnier, P., Lacassin, R., Leloup, P., Scharer,U., Dalai, Z., Haiwei, W., Xiaohan, L., Shaocheng,J., Lianshang, Z. and Jiayou, Z. 1990. The AilaoShan/Red River metamorphic belt:Tertiary left lat-eral shear between Indochina and South China.Nature, 343, 431-437.

Tapponnier, P., Peltzer, G., LeDain, A., Armijo, R.and Cobbold, P. 1982. Propagating extrusion tec-tonics in Asia: new insights from simple experi-ments with plasticine. Geology, 10, 611-616.

Yan, C. Y. and Kroenke, L. W. 1993. A plate tec-tonic reconstruction of the SW Pacific 0-100 Ma.In: Berger, T., Kroenke, L.W., Mayer, L. et al.,Proceedings of the Ocean Drilling Program, Sci-entific Results, 130, 697-709.

Cenozoic tectonics of SE Asia and Australasia 57

0 MaPresent Day

INDIA

AUSTRALIA

PACIFICPLATE

EURASIA

40oN

20oN

20oS

40oS

60oS

90oE120oE 150oE

180oE

ANTARCTICA

INDIANPLATE

H

NEWZEALAND

PHILIPPINESEA

PLATE

Figure 1. Present-day tectonic features of SE Asia and the SW Pacific. Yellow lines are selected marinemagnetic anomalies. Cyan lines outline bathymetric features. Red lines are active spreading centres. Whitelines are subduction zones and strike-slip faults. The present extent of the Pacific plate is shown in paleblue. Areas filled with green are mainly arc, ophiolitic, and accreted material formed at plate margins duringthe Cenozoic. Areas filled in cyan are submarine arc regions, hot spot volcanic products, and oceanic pla-teaus. Pale yellow areas represent submarine parts of the Eurasian continental margins. Pale and deep pinkareas represent submarine parts of the Australian continental margins. Letters represent marginal basins andtectonic features as shown on the opposite page.

58 R. Hall

50 MaEnd EarlyEocene

INDIA

AUSTRALIA

PACIFICPLATE

INDIANPLATE

AUSTRALIANPLATE

EURASIA

40oN

20oN

20oS

40oS

60oS

90oE 180oE

?

ANTARCTICA

Figure 2. Reconstruction of the region at 50 Ma. The possible extent of Greater India and the Eurasianmargin north of India are shown schematically. Shortly before 50 Ma collision between the north Australianpassive continental margin and an island arc had emplaced ophiolites on the north New Guinea margin, andin New Caledonia, eliminating ocean crust formed at the former Australian-Indian ocean spreading centre.Double black arrows indicate extension in Sundaland.

Cenozoic tectonics of SE Asia and Australasia 59

40 MaMiddle Eocene

INDIA

AUSTRALIA

PACIFICPLATE

INDIANPLATE

EURASIA

40oN

20oN

20oS

40oS

60oS

90oE 180oE

ANTARCTICA

Figure 3. Reconstruction of the region at 40 Ma. India and Australia were now parts of the same plate. Anoceanic spreading centre linked the north Makassar Strait, the Celebes Sea and the West Philippine basin.Spreading began at about this time in the Caroline Sea, separating the Caroline Ridge remnant arc from theSouth Caroline arc. Spreading also began after subduction flip in marginal basins around eastern Australa-sia producing the Solomon Sea and the island arcs of Melanesia.

60 R. Hall

30 MaMid Oligocene

INDIA

AUSTRALIA

PACIFICPLATE

INDIANPLATE

EURASIA

40oN

20oN

20oS

40oS

60oS

90oE

ANTARCTICA

180oE

Figure 4. Reconstruction of the region at 30 Ma. Indentation of Eurasia by India led to extrusion of theIndochina block by movement on the Red River Fault and Wang Chao-Three Pagodas (WC-TP) Faults. Slabpull due to southward subduction of the proto-South China Sea caused extension of the South China andIndochina continental margin and the present South China Sea began to open. A wide area of marginalbasins separated the Melanesian arc from passive margins of eastern Australasia, shown schematicallybetween the Solomon Sea and the South Fiji basin.

Cenozoic tectonics of SE Asia and Australasia 61

20 MaEarly Miocene

AUSTRALIA

PACIFICPLATE

INDIANPLATE

EURASIA

40oN

20oN

20oS

40oS

60oS

90oE 180oE

ANTARCTICA

INDIA

Figure 5. Reconstruction of the region at 20 Ma. Collision of the north Australian margin between the Bird’sHead microcontinent and eastern New Guinea occurred at about 25 Ma. The Ontong Java plateau arrived atthe Melanesian trench at about 20 Ma. These two major events caused major reorganisation of plateboundaries. Subduction of the Solomon Sea began at the eastern New Guinea margin to produce theMaramuni arc. Spreading began in the Parece Vela and Shikoku marginal basins accompanied by roll-backof the Izu-Bonin-Mariana trench. The north Australian margin became a major left-lateral strike-slip systemas the Philippine Sea-Caroline plate began to rotate clockwise. Movement on splays of the Sorong Faultsystem led to the collision of Australian continental fragments in Sulawesi. This in turn led to counter-clockwise rotation of Borneo and related Sundaland fragments, eliminating the proto-South China Sea. TheSumatra Fault system was initiated.

62 R. Hall

10 MaLate Miocene

AUSTRALIA

PACIFICPLATE

EURASIA

40oN

20oN

20oS

40oS

60oS

90oE 180oE

ANTARCTICA

INDIA

Figure 6. Reconstruction of the region at 10 Ma. The Solomon Sea was being eliminated by subductionbeneath eastern new Guinea and beneath the New Hebrides arc. However, continued subduction led todevelopment of new marginal basins within the period 10-0 Ma, including the Bismarck Sea, Woodlarkbasin, North Fiji basins, and Lau basin. The New Guinea terranes, formed in the South Caroline arc, dockedin New Guinea but continued to move in a wide left-lateral strike-slip zone. Further west, motion on strandsof the Sorong Fault system caused the arrival of the Tukang Besi and Sula fragments in Sulawesi. Collisionevents at the Eurasian continental margin in the Philippines, and subsequently between the Luzon arc andTaiwan, were accompanied by intra-plate deformation, important strike-slip faulting and complex develop-ment of opposed subduction zones. Rotation of Borneo was complete but motion of the Sumatran forearcslivers linked to new spreading in the Andaman Sea.