Geometrical problems and implications of large scale over- thrusting in the Banda Arc -Australian margin collision zone

M. G. Audley-Charles

SUMMARY: Inversion of stratigraphical sequences, the presence of high grade metamorphic rocks sitting on serpentinite rafts bounded by unmetamorphosed sedimen- tary rocks, and zones of imbrication have been cited as evidence of large scale overthrusting in the Outer Banda Arc. Earthquake data reveal a Benioff zone implying intermediate depth and deep focus underthrusting beneath the Banda Arc but there is a paucity of shallow focus shocks with underthrusting mechanisms. Multichannel seismic reflection surfeys between the Outer Banda Arc and the Australian-New Guinea shelf have been interpreted as showing imbrication in the north wall of the Timor Trough and the S wall of the Seram Trough. Quaternary steep angled block faulting post-dates and thus obscures much of the thrust faulting in the Banda Arc.

Many geometrical problems arise in trying to interpret the structural history of this collision zone between Australia and the volcanic (Inner) Banda Arc. They include:- (1) the apparent absence of a subduction trench and accretionary arc-trench gap in the Banda Arc; (2) the location of the surface trace of the Benioff zone before collision; (3) the history of the Benioff zone after the Pliocene oceanic trench was destroyed; (4) the relationship of the developing fold and thrust belt to the pre-collision geometry of the Australia-New Guinea continental margin; (5) the apparent absence of the continental slope a n d rise in the northern Australia collision zone; (6) the relationship of the crystalline basement of the Outer Banda Arc to the cover rocks and (7) the tectonic significance of the apparent continuity of the stratigraphically and structurally very different Sunda and Banda Arcs.

Continent-arc collision resulted from the convergence of Australia and the Banda volcanic arc. The associated deformation, represented by the folding and local im- brication of the Australian continental rise sediments of the Outer Banda Arcs with the emplacement of overthrust exotic sheets, was accomplished in 2 Ma. It appears to have been caused by the failure of the Benioff zone to accommodate the thick sediment prism of Australian proximal continental rise. Geometrical considerations suggest that, as a consequence, the Benioffzone and most of the approximately 200 km wide arc-trench gap were overridden by the Australian lithospheric plate during the continued plate conver- gence of the last 3 Ma. An important conclusion is that the Banda Arc fold and thrust belt developed in proximal continental rise deposits at the foot of the Australian continental slope

The Banda Arc of eas tern Indones ia comprises a volcanic ( inner) arc and an (outer) arc of strongly fo lded and thrust sed imenta ry and m e t a m o r p h i c rocks with minor igneous rocks. On the larger islands of the Ou te r Banda Arc (Timor and Seram) mounta ins rise to a lmost 3 km; the in ter -arc mar ine t roughs descend to 3 km and locally, in the W e b e r D e e p , to 7 km. The Banda Arc curves th rough 180 ~ has a str ike length of about 2400 km and a width of about 250 km (Fig. 1). In size (including relief) the Banda Arc is closely comparab le to the whole of the E u r o p e a n Alp ine moun ta in chain and the two largest islands, T imor and Seram, are each larger than the Swiss Alps.

The late Cenozoic de fo rma t ion in the Ou te r B a n d a Arc resu l ted f rom the collision be tween the no r thward drifting A u s t r a l i a - N e w Gu inea cont inent and the volcanic Inne r B a n d a Arc

system. The volcanic Inne r B a n d a Arc is re- la ted to the subduct ion of the ocean floor that lay N of no r thward moving A u s t r a l i a - N e w Gu inea dur ing the Cenozoic .

Ear l ie r geological investigators were led to in te rpre t the p resence of large scale over- thrusts, which some called nappes , in the Ou te r B a n d a Arc islands of T imor (Wanne r 1913; B e m m e l e n 1949) and Seram (Valk 1945; G e r m e r a a d 1946) by geomorpho log ica l fea tures and s t ructures reminiscent of Alp ine thrust zones. More critical ev idence in these islands was found by G r u n a u (1953, 1957), D e W a a r d (1954, 1955a, b 1956, 1957), L e m o i n e (1959), A u d l e y - C h a r l e s (1968), Car te r et al. (1976a) , B a r b e r et al. (1977), Brunnschwei le r (1978) and Aud ley -Cha r l e s et al. (1979). H o w - ever , part ly, because so few ma jo r thrust planes have been found exposed, G r a d y

Thrust and Nappe Tectonics. 1981. The Geological Society of London 407

408 M . G. A u d l e y - C h a r l e s

I / SOL A . . . . . ~ k~ ~ S E R A M ~ r R n ' ~ ' ~ . . ~ . -

"B.~,N ~. :.. . o ..WEBER?. I I ~

~6"s fl o SEA .:Tl.::./-fi- ~il/y}P'/ I U ,SE=dA :. . ~ . g / / / /

.,,,~/:'~ 7~.)fo/ / t . . I o ,, "...:):"....":-;!?:V.,d// r

i . " .. ~ . , ~ , V t , . A ! ~ z, rc /__/r~,,~,A,; , / / ~ . 1 ....... :"..'.":.:: . . . . / ' - / J l j - ~ - ~

' ~ ~Ro,//'"-.::/i::; fr "---~sH ELF

,'3eE

PA C IFIC

OCEAN

J

N E W

G U I N E A .

o. t

0 15o 300 KM I , I

B INNER BANDA ARC

INTER-ARC TROUGH

~ r ~ O U T E R BANDA ARC

A U S T R A L I A ~a'*E f

FIG. 1. Banda Arc of eastern Indonesia.

(1975), Grady & Berry (1977) and Chamalaun & Grady (1978) have argued that Plio- Pleistocene block faulting is the dominant tec- tonic style. These writers have not accepted the view of Carter et aI. (1976a) and Barber et al. (1977) that the Plio-Pleistocene block fault- ing has displaced the early to mid-Pliocene thrust faults. However, Grady & Berry (1977) admit that their investigations were limited to a small part of northern Timor and they have not checked much of the field evidence cited as evidence for overthrusting.

On the basis of marine geophysical surveys, Bowin et al. (in press) followed Cardwell & Isacks (1978), who studied the earthquakes of the Banda region, in concluding that there is evidence to indicate that the India-Australia plate has underthrust the Eurasia plate below the Timor and Seram sectors of the S and N parts (respectively) of the Banda Arc.

This paper is concerned with some problems and implications raised by the mapping of thrust sheets which appear to have moved away from the Banda Sea (4-5 km deep) to- wards the Austral ia-New Guinea continent, that is, moved northwards in Seram on the N

side of the Banda Arc and southwards in Timor on the S side. In an attempt to under- stand the history and mode of development of overthrust tectonics in a continent-island arc collision zone, the simplified geometry of the convergent zone has been reconstructed on a natural scale using recently available geophysi- cal and geological data

Implicat ions of overthrusting for col l is ion tectonics in the

B a n d a Arc

The age of overthrusting and collision in the Banda Arc

On the basis of stratigraphical data (mainly the age of unconformities) the late Cenozoic deformation of the Outer Banda Arc, indi- cated by folding and emplacement of over- thrusts, has been dated by Carter et al. (1976a) in Timor as zone N 20 or Middle Pliocene (approximately 3 Ma) and by Audley- Charles et al. (1979) in Seram as zone N 18 or late Miocene-early Pliocene (approximately

L a r g e scale over thrus t ing in the B a n d a A r c 409

5 Ma). The correlation of radiometric dates with planktonic foraminiferal zonation follows Billman & Scrutton (1976). Carter et al. (1976a) argued that the emplacement of thrust sheets in Timor must correlate with the colli- sion between the volcanic Banda Arc and the Australia-New Guinea continental margin over which the thrust sheets travelled towards the continent. Only 20 km now separates the thrust sheets of northern Timor from the vol- canic island of Atauru of the Inner Banda Arc. Abbott & Chamalaun (1978) followed Carter et al. (1976a) in regarding the emplacement of the Ocussi Volcanics thrust sheet in northern Timor as having occurred when the continental margin of Australia collided with the volcanic Banda Arc in the Middle Pliocene (3 Ma). Bowin et al. (in press) also conclude that the collision between the northward drifting Au- stralian continent and the Banda Arc occurred during the Pliocene 3-5 Ma ago.

Location of pre-Pi iocene collision Benioff zone

Between the northward drifting Australia- New Guinea continent and the pre-Pliocene volcanic Banda Arc there must have been an ocean trench which accommodated the sub- ducting ocean floor separating these converg- ing plates. The position of this trench with respect to the pre-Pliocene volcanic Banda Arc is not known but, following Abbott & Chamalaun (1978), by comparison with the Sunda Arc, we might expect the trench to have been about 200 km S of the smaller volcanic Banda Arc. The northward moving Australia-

New Guinea continental margin must have collided with this trench before it ceased underthrusting the volcanic arc about 3 Ma ago.

Benioff zone and collision geometry deduced from pre-Pl iocene volcanic arc

The evidence of contamination by continen- tal crustal material of the magmas extruded in the volcanic Banda Arc (Whitford et al. 1977) indicates that Australian continental margin rocks (probably continental rise sediments- - for reasons set out later) were subducted to depths of 100-200 km where they contami- nated the rising magmas that erupted in Wetar at least 3 M a ago (Abbott & Chamalaun 1978).

Using the 7 cm per year plate convergence rate of Bowin et al. (in press) and a pre- Pliocene Banda arc-trench gap approximately 200 km wide, the date when the magma con- taminating Australian continental rise sedi- ments reached the trench can be estimated as about 7.3 Ma or zone N 17 of late Miocene (Fig. 2). The Kolbano-Betano-Aliambata- Iliomar bathyal sediments of southern Timor are Australian continental rise deposits of late Jurassic to early Pliocene age and now form part of the imbricated para-autochthon (Carter et al. 1976a, b; Barber et al. 1977). There is no evidence of a stratigraphical break in the N 16-N 18 zone sequer~ces of this facies, but zones N 19 and N 20 are missing. One inter- pretation that can be placed on this (Fig. 2) is that bathyal sedimentation continued on the foot wall of the trench during N 17 and N 18 times and that, until the end of N 18 times, the

N W SE AUSTRALIA

Proto-Banda Arc A r c - T r e n c h G a p Trench Continental Rose Continental Continental Slope Shelf SOUTH ~4NDA SEA ~ __. j /Cablac Lmst ~

/ 7 ~ . . . . . . . . . . . . . . . . . . . . . . . . . . . ; . . . . . . . . . . . �9 . . . . . . . . . . . . . . . . ~ : ~ :::-:=" ~ : : ~:.:.', T:'::: : S~H E L F o t o ~ .~o.~o ~ ~ i , . : : : . R E.: S E: O ! u E n

. . . . . . . . . . . . " ~ / ~ ~ O a r ......... - - - - , - - , ~'L '-- '--

., ::;: . . . . . . . . . . . . p~ Maubisse ..... ....... ~ ~'~,~ . . . . . . . . . . .

" R A = I A N " ~ " ' = = ' = = = ' = " ' ~ ' ~ ' ~ " " ~ / / . . . . I . 120km =t = u o RtoO~t Z~.~le of ~ . . . . / J ~ 0 S P H E R I- . . . . . . . C . . . . . . . . . . . . . .

. . . . . . . S e ~ t ~ j ~ ~" . \ .~ y~ ~ - of Tlmor Tim= Trough

J / / I

// / ,~. ~ ~ r.. Cr,ba8 Format . . . . [ p l / / " / -- 5 ' 't~ ~" A . . . . . . ;o . . . . . . . . Or . . . . ]

/ . . . 5 7 k m a w e * 7 c ~ y r ~ LATE MIOCENE 16 MY. BP)

FIG. 2. Reconstruction of the late Miocene northern Australian margin-volcanic Banda Arc convergent zone. Note that about 300 km of Australian distal continental rise deposits must have been subducted before they could contaminate magmas erupted in the volcanic arc.

410 M. G. Audley-Charles

continental rise deposits were subducted with- out structural effects on the sediments just entering and outside the trench on the down- going plate in a manner resembling the 'under- thrusting mode' described by Helwig & Hall (1974). The distal part of this continental rise containing these sediments would have already been subducted to depths of 1 0 0 - 2 0 0 k m below the Pliocene (N 20-3 Ma) volcanic arc. The subducted deposits contaminated the ris- ing magmas about 4.3 Ma after they had ac- cumulated in the region of the trench, thus giving rise to the cordierite-bearing volcanics extruded about 3 Ma ago in the mid-Pliocene. This 3 Ma date appears to be highly significant in the Timor region of the Banda Arc, for it not only coincides with the collision of the Australian proximal continental rise with the volcanic arc-trench gap, as indicated by the emplacement of the overthrusts onto Timor, but also with the cessation of volcanism in this volcanic arc opposite the Timor region (Abbott & Chamalaun 1978). This death of volcanic activity (Fig. 3) has been related to two factors: (a) the consumption of all the oceanic crust between the converging Australian continent and the volcanic arc; most of the oceanic crust had been subducted, but the last remnants, represented perhaps by the Ocussi Volcanics, described as ocean floor tholeiites by Abbott &

Chamalaun (1978), were overthrust onto northern Timor about 3 Ma; (b) with all the ocean crust consumed or overthrust the conti- nental margin of Australia began to be sub- ducted but the buoyancy effect of the density contrast inhibited further underthrusting so that volcanism in the arc died (Fig. 4).

Overthrust stratigraphy implies Benioff zone was overridden by

Australian continental rise

The presence of Permian and Mesozoic Au- stralian continental rise deposits now only 2 0 k m from the volcanic Inner Banda Arc (Figs 1 & 4) implies that the volcanic arc- trench gap, probably about 200 km wide dur- ing the pre-Pliocene has either suffered ex- treme shortening, for which there is no evi- dence, or else it has been overridden by the converging Australian continental rise (Fig. 4). An important question related to the develop- ment of the collision margin is how to deter- mine the age of the present configuration of only 2 0 k m between the Inner and Outer Banda Arc in part of the Timor region. The overthrust sheets were emplaced on the Au- stralian margin of Timor about 3 Ma in zone N 20 of the mid-Pliocene when the subduction

FIG. 3. Reconstruction of the Banda Arc convergent zone during the mid-Pliocene collision. Failure of the Benioff zone to accommodate the thick proximal continental rise deposits led to cessation of shallow focus subduction. Continuation of plate convergence resulted in Australian plate overthrusting the Benioff zone. The upper part of the arc-trench gap, which formed the hanging wall of the subduction zone, sheared and overthrust the continental rise as it rode over the Benioff zone.

Large scale overthrusting in the B a n d a Arc 411

NW Extinct Volcanic SE Inner Banda Arc Outer Banda Arc Timor Trough Australian Shelf

" \ J s,,,,t, .. ~ . . . i 5 /v ~ '~ j!':'i:i:C: O 0 w N W A R P e D S H E L F

CRYSTALLINE BASEMENT /~e- ........ CONTINENTAL

. ~ /__~ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

E U R A S I A N "N-c"4~..,. I A U S T R A L I A N L ITHOSPHERE

P L A T E ~'.. ~ _

~ Z ~" "" ~- A S T H E N O S P H E R E ,~?:~V, ' "- ~.

d.:~: ~,, *.,..

r0 1 / / 50kin

. . . . . ~ MID PLEISTOCENE (1 MY. BP)

FIG. 4. Mid-Pleistocene reconstruction of the collision zone showing the Australian plate overrid- ing the pre-Pliocene Banda Benioff zone and much of the arc-trench gap resulting in the Australian continental rise arriving only 50 km S of the volcanic Banda Arc. Note that the detached upper part of the arc-trench gap forms the exotic thrust sheets above the continental rise deposits of the para-autochthon. Hamilton (1977, 1978) identified a new trench forming N of the Flores-Alor part of the Wetar volcanic arc initiating southward underthrusting of the Banda Sea below the arc. This accords well with this model but the reversed direction of underthrusting is much less easy to accommodate to the tectonic model of Hamilton (1978), vonder Borch (1979) and Bowin et al. (in press).

trench, estimated to be then about 2 0 0 k m S of the volcanic arc, failed to accommo- date the underthrusting Australian continental margin. At least 300 km of distal continental rise had already been underthrust to reach 200 km depth below the volcanic arc as out- lined in Fig. 2. Plate convergence, estimated at a rate of 5 cm per year (Abbott & Chamalaun 1978), continued after this mid-Pliocene (3 Ma) collision so that the collision suture became deformed by folds and overthrusting. The continuation of this northward component of drift by Austral ia-New Guinea continent after the mid-Pliocene led to the Australian continental rise riding over most of the 200 km wide volcanic arc-trench gap during the last 3 Ma. Geometrical considerations represented by natural scale cross sections (Figs 2-4) suggest that this continuing plate convergence could only be accommodated by the Australian plate overriding not only the Asian arc-trench gap but also at least 300 km of its own sub- ducted continental rise. Atlantic type rifted

continental margins have continental rise de- posits that extend up to 500 km or more across the ocean floor (Heezen 1974). The postulated subducted 300 km of northern Australian con- tinental rise would explain why the present continental rise deposits extend for only about 6 0 k m (across Timor). This allows about another 60 km width for continental slope de- posits occupying the region between southern Timor and the foot of the N wall of the Timor Trough, which is underlain by Australian shelf according to Bowin et al. (in press). These dimensions allow for both considerable shor- tening in the continental rise following colli- sion and for more than 300 km of rise to have been lost by subduction.

After the main orogenic phase in Timor there is evidence for late Pliocene folding and relatively minor thrusting (Carter et al. 1976a) and for two Pleistocene folding phases (Kenyon 1974), as well as widespread strong steep angled block faulting during the Pleisto- cene. These movements appear to be associated

412 M. G. A u d l e y - C h a r l e s

with the continuing plate convergence and the buckling response ot the collision zone in which the Banda Arc islands have emerged and continued to rise while the interarc troughs and the 'foredeep' between the Outer Banda Arc and the Australia-New Guinea continent have subsided (Veevers et al. 1978).

Audley-Charles & Hooijer (1973) argued from the evidence for dwarfing Stegodont populations having wandered back and forth between Timor (Outer Banda Arc) and Flores (volcanic Inner Banda Arc) that there was a land connection between the Inner and Outer Banda Arc during the Pleistocene and that the present Ombai Strait between them is a young subsided feature induced by faulting. The evi- dence from the stegodonts suggests that the present island configuration in the Timor reg- ion was acquired by Middle Pleistocene time (approximately 1 Ma).

Relationship of fold and thrust belt to pre-collision geometry of

the Australian continental margin

The dating of the various phases in the de- velopment of this Banda Arc orogenesis allows us to investigate the relationship of the loca- tion of the fold and thrust belt to the structure of the pre-collision continental margin (Fig. 2). Seismic surveys have revealed that the Australian shelf strata can be traced from the present shelf northwards, below the Timor Trough, as far as the foot of the N wall, where they are overlain by thick deposits lacking

coherent reflectors but which appear to be strongly folded and imbricated (Jacobson et al. 1978). This indicates that the Cenozoic, Mesozoic and possibly older shelf deposits ex- tend northwards to about 50 km from the S Timor coast. Thus, only 120 km from the de- pressed shelf below the present Timor Trough to the N coast of Timor remains to accommo- date what is left of the Cenozoic, Mesozoic and Permian Australian continental slope and rise. Taking the average width of 60k in for Atlantic-type rifted continental margin contin- ental slopes (Heezen 1974) means that the present island of Timor approximates to the proximal continental rise extending from the foot of the continental slope located at depth below the S coast of the island. This would mean that in the Timor region the fold and thrust belt of the Banda Arc coincides with what was originally the foot of the continental slope, where the greatest thickness of sediment is trapped (Heezen 1974). Apthorpe (1979) identified the Mesozoic shelf break in the Browse Basin SW of Timor which corresponds with this interpretation. In view of the excep- tionally thick sediment (10 km) on the shelf (Robertson et al. 1976) we could expect that perhaps at least 15 km of sediment might have been accumulated at the foot of the slope and in the proximal rise. All of the para-autoch- thonous Cenozoic, Mesozoic and Permian for- mations of Timor represent deep water slope deposits of the continental rise, locally charac- terized by various kinds of pelagic sediments, turbidites and sedimentary slumping, which, in the Cribas and Aitutu formations of western Timor, is on a very large scale. The slumping

N THRUST SHEETS FROM ASIAN ARC-TRENCH GAP S

- ILl: ~ T i " ~ 1 ~ { IMBRICATED 7~ ~ . . . . . - " ~ ~ * ~ - - , " ~ T Y .~KOLBANO FAClES t | ~ ~ ~ - - - ~ " - - - ~ . , . " ~ "" " " " . ~ J ~ T (JURAS-PLIOCENE) | ,,.T

POSITS

CRIBAS FM. ~ PARA--AuTOCHTHO "~ "" ATAHOC FM.

/< I 0 0 K M ~)~

FIG. 5. Diagrammatic sketch across Timor showing the exotic thrust sheets emplaced above the Australian continental rise deposits during the mid-Pliocene collision. A = radiolarian lutites and muds with siltstones of Jurassic-Miocene age containing tiny clasts of metamorphic rocks (Lolotoi Complex?) and Palelo-type volcanic rocks described by Carter et al. (1976b).

L a r g e scale over thrus t ing in the B a n d a A r c 413

and reworking with local biostratigraphical breaks in the Kolbano facies of Jurassic- Pliocene age has been described in Carter et al. (1976a, b). The term 'continental shelf unit ' applied to these para-autochthonous Permian- Pliocene Australian strata by Barber et al. (1977) and others is misleading in this respect; it should have been designated 'continental rise unit'.

Audley-Charles (1978) summarized the evi- dence for interpreting the Mesozoic-Cenozoic continental slope of the 'Sula Peninsula' (of the Australia-New Guinea continent) to have been situated between N Seram and Misool. A mul- tichannel seismic profile indicates that the Australia-New Guinea shelf deposits of Misool can be traced southwards, where they are de- pressed below the axis of the Seram Trough, as far as the S wall of the Trough where re- flections are obscured below younger deposits (Audley-Charles & Carter 1978). This implies that in the northern part of the Banda Arc, as in the southern part, the fold and thrust belt developed in the continental rise near the foot of the continental slope.

Relat ion of crystalline basement to cover rocks in the Outer

Banda Arc

The coincidence of the Banda Arc fold and thrust belt with the foot of the pre-Middle Pliocene Australian continental slope, where sedimentary accumulations achieved their maximum thickness, explains why no sub- Permian strata nor their crystalline basement have been seen in Timor. The original sedi- ment thickness (probably between 10 and 15 km) in this palaeogeographical zone and the vergence of the fold belt towards the Au- stralian continent, involving displacement of folded and sliced Permian, Mesozoic and Cenozoic strata southwards across Timor, have together provided a very thick pile of cover rocks.

According to Bowin et al. (in press) Au- stralian continental crust can be traced by re- flection seismic data northwards from the shelf below the Timor Trough where it is about 40 km thick below the trough axis. Chamalaun et al. (1976) used gravity data to construct a model that indicates that Australian continen- tal crust continues as far as the north coast of Timor. Their model densities of 2.7 and 2.9 provide for continental crustal thicknesses below Timor of 30 km. These observations and models suggest that most of the distal portion of the Australian continental rise in the Timor

region of the Banda Arc is either missing (subducted) or has been involved in con- siderable shortening to account for the 30 km crustal thickness below Timor indicated by the gravity model and by the refraction picture of the Trough.

The tectonic style of the Outer Banda Arc islands, Timor ; nd Seram, involves strong fold- ing, thrusting and imbrication in the Australian continental rise deposits. This indicates that these strata have become detached from their original crystalline (probably oceanic) base- ment; hence they are designated as para- autochthonous. They form complex though coherent fold and fault structures which can be mapped in detail if biostratigraphical techni- ques are employed.

Metamorphic rocks, ranging from granulite to greenschist facies, form a number of massifs in Timor and Seram (Barber & Audley- Charles 1976). These rocks correspond to levels varying from the base of the continental crust to epizonal depths and also include metamorphosed ocean floor materials (Barber 1979). Evidence indicates that these metamorphics have been thrust northwards in Seram and southwards in Timor (Barber & Audley-Charles 1976; Audley-Charles et al. 1979). The structural position of the migmatic gneisses in Seram is uncertain; on the available evidence they could be regarded as thrust sheets or upthrust horsts.

Identif ication of the 'root zone' of the Banda Arc exot ic thrust

sheets

The tectonic model proposed by Carter et al. (1976a) suggested the ' root zone' of the exotic thrust sheets was rifted from the Sunda cra- ton in the late Cretaceous and formed the base of the Banda arc-trench gap which migrated southwards from Sulawesi during the late Cre- taceous and Palaeocene. Their suggestion that this migration continued by back-arc spreading during the Eocene-Pliocene seems doubtful, in view of the heat flow characters of the south- ern Banda Sea (Bowin et al. in press). Hamil- ton (1979, fig. 77) followed closely the Carter et al. (1976a) model but he argued that the main back-arc spreading of the Banda Sea began about 15 Ma in the mid-Miocene. The Carter et al. (1976a) model can be extended to include the Australian continental rise over- thrusting the volcanic arc-trench gap by post- ulating that the base of the hanging wall of the subduction trench sheared and detached from

414 M . G. A u d l e y - C h a r l e s

its basement during the mid-Pliocene deforma- tion (Figs 2, 3, & 4). The failure of the under- thrusting mechanism in the upper part of the mid-Pliocene Benioff zone coincided with the arrival at the trench of the great thickness of continental rise sediment near the foot of the Australian continental slope. This composite model (Figs 3 & 4) implies that the 'root zone' of the exotic thrust sheets of the Timor region has been overridden by the Australian conti- nental rise carrying the detached exotic thrust sheets back (northwards) over their own roots.

Tectonic significance ot the ap- parent continuity ot the Sunda

and Banda Arcs

Bowin et al. (in press) argued that the position of the Java Trench relative to the volcanic arc is comparable with the Banda Arc-Timor Trough separation if allowance is made for post-collision compression. They pointed out that both the Java Trench and volcanic arc fit small circles and that the western end of the Timor Trough fits the eastward extrapolation of the Java Trench small circle. They ascribe the progressive divergence of the Timor Trough, from the small circle projection east- ward from Timor, to post-collision con- vergence. They reject the Veevers et al. (1978) model for the evolution of the Timor Trough as a downwarping of the Australian margin that migrated towards the continent after colli- sion. However, using the composite Carter et al. (1976a) and Abbott & Chamalaun (1978) models for the arc-continent collision, it is possible also to combine the Bowin et al. (in press) and Veevers et al. (1978) models for Timor Trough evolution. This new model would place the mid-Pliocene trench below the western end of the present Timor Trough. The progressive narrowing of the distance between the present volcanic Inner Banda Arc and the Timor-Tanimbar Trough (Fig. 1) could then be attributed to one or more factors: (a) The original (pre-mid Pliocene) width of the arc- trench gap being progressively narrower east- wards from Timor, reflecting the eastwards progressive diminution of the volcanic arc: (b) alternatively the pre-mid Pliocene volcanic arc may have been overridden (as well as the arc-trench gap) E of Wetar and Romang; this would imply that the present small vol- canic arc E of Romang is a younger post- mid Pliocene feature, as Bowin et al. (in press) suspect.

The small circle fit of the volcanic inner part

of both the Sunda and Banda arcs suggests that they evolved as a single tectonic element on the same plate margin associated with the sub- duction of the Tethys and Indian Oceans as Australia-New Guinea drifted northwards with the Indian Ocean. However, the stratigraphy and structure of the outer part of the Sunda Arc system exemplified by the Mentawai Is- lands (Moore & Karig in press) are very differ- ent from the geology of the Outer Banda Arc, which involves Australian continental margin rocks overlain by exotic thrust sheets.

This fundamental difference in origin of the outer part of the Sunda and Banda Arcs means that, unlike the Mentawai islands of the Sunda Arc system exemplified by the Mentawai Is- Banda Arc islands cannot be regarded as analogous to the 'trench-slope break' of the Pacific arcs, as Cardwell & Isacks (1978) suggested. Neither can the Outer Banda Arc be regarded as an 'accretionary prism' of the hanging wall sitting directly above the under- thrusting Australian crust, as Jacobson etPal. (1978), Hamilton (1979, fig. 78, 1978, yon der Borch (1979) and Bowin et al. (in press) sug- gest. The Australian continental rise deposits in the Outer Banda Arc belong to the footwall that has overridden the Benioff zone and the true accretionary prism to arrive only 20 km from the volcanic arc (Figs 2-4).

Conclusions

1. The region is too complicated and insuffi- ciently known to allow a unique interpretation of the data, nevertheless the Carter et al. (1976a) tectonic model based largely on stratigraphy of the Outer Banda Arc combines well with the Abbott & Chamalaun (1978) model based mainly on igneous stratigraphy of the volcanic arc. Furthermore, many aspects of the interpretations of Bowin et al. (in press) derived largely from marine geophysics and the Veevers et al. (1978) model for the Timor Trough can also be combined to form a new composite model that fits the available geologi- cal and geophysical data.

2. This composite model involves the ac- cretionary arc-trench gap of the mid-Pliocene volcanic arc being underthrust by about 300 km of the Australian distal continental rise. The arrival at the trench in mid-Pliocene times (3 Ma) of the proximal part of the Au- stralian rise comprising 10-15 km of low den- sity sediments led to the failure of the Benioff zone to accommodate these rocks by under- thrusting. Continuing plate convergence re-

Large scale overthrust ing in the B a n d a A r c 415

sulted in approximately 150 km of the proxi- mal part of the Australian continental rise and slope overriding the trench and the 200 km wide arc-trench gap until, by about Middle Pleistocene times (1Ma), the proximal conti- nental rise arrived only 20 km from the vol- canic arc (then extinct) in the western part of the arc (Timor region). Further E, beyond Romang, this mid-Pliocene volcanic arc seems to have been completely overridden by the Australian continental rise.

3. This model can account for some of the outstanding geological anomalies of this com- plex arc. These include:- (a) The apparent absence of the Australian continental slope and rise in the northern Australia collision zone. (b) The apparent absence of a subduc- tion trench and an accretionary arc-trench gap in the Banda Arc. (c) The presence of Au- stralian continental rise deposits now only 20 km south of the pre-mid Pliocene volcanic arc. (d) The occurrence above the Australian continental margin of a variety of exotic thrust sheets whose 'root zone' seems most likely to have been the margin of the Sunda craton. (e) The observation that the lowest thrust sheet (Kolbano-Nief) of Timor and Seram comprises Jurassic-Pliocene Australian deep-water conti- nental rise deposits overlain by thrust sheets comprising Eocene and Miocene sediments (Dartollu and Cablac) which were deposited in very shallow water.

4. Implications of this model are: (a) The 100-120 km wide fold and thrust belt of the Banda Arc developed in proximal continental rise deposits at the foot of the Australian con-

tinental slope. (b) The only parts of Australian continental crust to be subducted were the sedi- mentary deposits on the 300 km wide distal portion of the continental rise. (c) The failure of the Beniott zone to accommodate the thick proximal continental rise deposits, and the continuing plate convergence resulted in the collision of the Banda Arc and Australia. The folding, imbrication and overthrusting in this 100 km wide collision suture was accomplished between mid-Pliocene and mid-Pleistocene (about 2 Ma) in the Timor region. (d) Other arc-continent and continent-continent Tethyan collision zones such as the Himalayas may also have involved the Gondwana continental rise or even shelf deposits overridding the Asian arc-trench gap and volcanic arc, leaving little or no trace of their former presence.

5. Predictions for the future of this collision fold belt involve the now deformed and over- thrust Australian continental margin (lacking about 300 km of its subducted continental rise) riding over the volcanic Inner Banda Arc if convergence continues. Not only would this remove the volcanic arc from view but it could lead, by erosion of the overthrust volcanic rocks, to little evidence of the volcanic activity remaining.

6. This new composite tectonic model could be tested by a series of seismic refraction lines in the immediate region of the Inner and Outer Banda Arc.

ACKNOWLEDGMENTS. The writer is indebted to A. W. Bally, A. J. Barber, D. J. Carter and N. Schneidermann for discussion.

R e f e r e n c e s

ABBOTT, M. J. tf, z; CHAMALAUN, F. H. 1978. New K/Ar age data for Banda Arc volcanics. Inst. Australas. Geodynamics, Publ. 78/5, 1-34.

APTHORPE, M. C. 1979. Depositional history of the Upper Cretaceous of the north-west shelf, based upon foraminifera. APEA. J. 19, 74-89.

AUDLEY-CHARLES, M. G. 1968. The geology of Por- tuguese Timor. Mere. geol. Soc. London, 4, 1- 76.

- - 1978. Indonesian and Philippine Archipelago. In: MOULLADE, M. & NAIRN, A. M. (eds), The Phanerozoic Geology of the World H The Mesozoic A. Elsevier Publishing Co. Amster- dam. 165-207.

& CARTER, D. J. 1978. Interpretation of a regional seismic line from Misool to Seram: Implications for regional structure and pet- roleum exploration. Proc. Indones. Petrol. Assoc. (6th Ann. Cony. Jakarta, 1977), 2, 3-11.

- - , - - , BARBER, A. J., NORVICK, M. S. & TJOK-

ROSAPOETRO, S. 1979. Reinterpretation of the geology of Seram: Implications for the Banda Arc and northern Australia. J. geol. Soc. London, 136, 547-68.

HOOIJER, D. A. 1973. Relation of Pleis- tocene migrations of pygmy stegodonts to island arc tectonics in eastern Indonesia. Nature Lon- don, 241, 197-8.

BARBER, A. J. 1979. Structural interpretations of the island of Timor, eastern Indonesia. SEA- PEX, Singapore, 4, 9-21.

- - & AUDLEY-CHARLES, M. G. 1976. The signifi- cance of the metamorphic rocks of Timor in the development of the Banda Arc. Tectonophysics, 30, 119-28. , & CARTER, O. J. 1977. Thrust tectonics in Timor. J. geol. Soc. Aust. 24, 51-62.

BEMMELEN, R. W. VAN 1949. The Geology of In- donesia. Government Printing Office, The Hague.

4 1 6 M . G. A u d I e y - C h a r l e s

BILLMAN, H. G. & SCRUTTON, M. E. 1976. Strati- graphic Correlation in Indonesia. SEAPEX, Sing- apore, 9, 1-14.

BOWIN, C., PURDY, G. M., JOHNSTON, C., SHOR, C., LAWVER, L., HARTONO, H. M. S. & JEZEK, P. in press. Arc-continent collision in the Banda Sea region. Bull. Am. Assoc. Petrol. Geol.

BRUNNSCHWEILER, R. O. 1978. Notes of the geology of eastern Timor. Bull. Bur. Miner. Resour. Geol. Geophys. Melbourne. 192, 9-18.

CARDWELL, R. K. & ISACKS, B. L. 1978. Geometry of the subducted lithosphere beneath the Banda Sea in eastern Indonesia from seismicity and fault plane solutions. J. geophys. Res. 83, 2825- 38.

CARTER, D. J., AUDLEY-CHARLES, M. G. & BARBER, A. J. 1976a. Stratigraphical analysis of island arc-continental margin collision in eastern Indonesia. J. geol. Soc. London, 132, 179-98.

- - & - - 1976b. Discussion of Stratigraphi- cal analysis of island arc-continental margin col- lision in eastern Indonesia. J. geol. Soc. London, 132, 358-61.

CHAMALAUN, F. H., LOCKWOOD, K. & WHITE, A. 1976. The Bouguer gravity field and crustal structure of Timor. Tectonophysics, 30, 241-59.

- - & GRADY, A. E. 1978. The tectonic develop- ment of Timor: A new model and its implica- tions for petroleum exploration. A P E A J. 18, 102-8.

GERMERAAD, J. H. 1946. Geology of Central Seram. Amsterdam (de Bussy).

GRADY, A. E. 1975. A reinvestigation of thrusting in Portuguese Timor. J. geol. Soc. Aust. 22, 223-7.

& BERRY, R. F. 1977. Some Palaeozoic- Mesozoic stratigraphic and structural relation- ships in east Timor and their significance in the tectonics of Timor. J. geol. Soc. Aust. 24, 203- 14.

GRUNAU, H. R. 1953. Geologie yon Portugiesisch Ost-Timor. Eine kurtze Ubersicht. Eclog. geol. Helv. 46, 29-37.

1957. Nueue Daten zur geologie yon Por- tiesisch Ost-Timor. Eclog. geol. Helv. 50, 69-98.

HAMILTON, W. 1977. Subduction on the Indonesian region. Proc. Indones. Petrol. Assoc. 2-23.

- - 1978. Tectonic maps of the Indonesian region. U.S. Geol. Surv. Misc. Invest. Series M I-875-D.

- - 1979. Tectonics of the Indonesian region. U.S. Geol. Surv. Prof. Pap. 1078, 1-345.

HEEZEN, B. C. 1974. Atlantic-type continental mar- gins. In: BURK, C. A. & DRAKE, C. L. (eds), The Geology of Continental Margins, 13-24, Springer-Verlag New York.

HELWIG, J. & HALL, G. A. 1974. Steady-state trenches? Geology, 2, 309-16.

JACOBSON, R. S., SHOR, G. G., KIECKHEFER, R. M. PURDY, G. M. 1978. Seismic refraction and

reflection studies in the Timor-Aru Trough sys- tem and Australian continental shelf. Mere. Am. Assoc. Petrol. Geol. 29, 209-22.

KENYON, C. S. 1974. Stratigraphy and Sedimentology of the Late Miocene to Quaternary Deposits of Timor. Ph.D. Thesis, Univ. of London. (unpubl).

LEMOINE, M. 1959. Un example de tectonique chaoti- que, Timor: essai de coordination et d'interpre- tation. Revue Geogr. phys. Geol. dyn. Paris 2, 205-30.

MOORE, G. F. & KARIG, D. E. in press. Structural geology of Nias Island, Indonesia: Implications for subduction zone tectonics. Am. J. Sci.

ROBERTSON, G. A., POWELL, D. E. & EDMOND, G. M. 1976. Australian northwest continental shelf: Results of ten years of exploration. In: HALBOUTY, M. T., MAHER, J. C. & LIAN, H. M. (eds), Circum Pacific Energy and Mineral Re- sources, Mere. Am. Assoc. Petrol. Geol. 25, 231-8.

VALK, W. 1944. Contributions to the Geology of West Seram. Amsterdam (de Bussy).

VEEVERS, J. J., FALVEY, D. A. & ROBINS, S. 1978. Timor Trough and Australia: Facies show to- pographic wave migrated 80 km during the past 3 my. Tectonophysics, 45, 217-27.

VON DER BORCH, C. 1979. Continent-island arc colli- sion in the Banda Arc. Tectonophysics, 54, 169- 93.

WAARD, D. DE, 1954. Contributions to the geology of Timor. 1 Geological research in Timor, an introduction. Indones. J. nat. Sci. 110, 1-8.

1955a. Contributions to the geology of Timor VII. On the tectonics of the Ofu Series. Indones. J. nat. Sci. 111, 137-43.

1955b. Contributions to the geology of Timor VIII. Tectonics of the Sonnebait overthrust unit near Nikiniki and Basleo. Indones. J. nat. Sci. 111, 144-50.

1956. Contributions to the geology of Timor IX. Geology of a N-S section across western Timor. Indones. J. nat. Sci. 122, 101-14.

1957. Contribution to the geology of Timor XII. The third Timor Geological Expedition preliminary results. Indones. J. nat. Sci. 113, 7-42.

WANNER, J. 1913. Geologie von West Timor. Geol. Rndsch. 4, 136-50.

WHITFORD, D. J., COMPSTON, W., NICHOLLS, I. A. & ABBOTt, M. J. 1977. Geochemistry of late Cenozoic lavas from eastern Indonesia: Role of subducted sediments in petrogenesis. Geology, 5, 571-5.

M . G . AUDLEY-CHARLES, Department of Geological Sciences, Queen Mary College, London.