www.elsevier.com/locate/tecto

Tectonophysics 389

Ocean trench blocked and obliterated by Banda forearc collision

with Australian proximal continental slope

M.G. Audley-Charles*

Department of Earth Sciences, University College London, Gower Street, London WC1E 6BT, UK

SE Asia Research Group, Department of Geology, Royal Holloway, University of London, Egham, Surrey TW20 0EX, UK

Received 19 November 2003; accepted 9 July 2004

Available online 27 August 2004

Abstract

The bathymetry and abrupt changes in earthquake seismicity around the eastern end of the Java Trench suggest it is now

blocked south–east of Sumba by the Australian, Jurassic-rifted, continental margin forming the largely submarine Roti–Savu

Ridge. Plate reconstructions have demonstrated that from at least 45 Ma the Java Trench continued far to the east of Sumba.

From about 12 Ma the eastern part of the Java Trench (called Banda Trench) continued as the active plate boundary, located

between what was to become Timor Island, then part of the Australian proximal continental slope, and the Banda Volcanic Arc.

This Banda Trench began to be obliterated by continental margin-arc collision between about 3.5 and 2 Ma.

The present position of the defunct Banda Trench can be located by use of plate reconstructions, earthquake seismology,

deep reflection seismology, DSDP 262 results and geological mapping as being buried under the para-autochthon below the

foothills of southern Timor. Locating the former trench guides the location of the apparently missing large southern part of the

Banda forearc that was carried over the Australian continental margin during the final stage of the period of subduction of that

continental margin that lasted from about 12 Ma to about 3.5 Ma.

Tectonic collision is defined and distinguished from subduction and rollback. Collision in the southern part of the Banda Arc

was initiated when the overriding forearc basement of the upper plate reached the proximal part of the Australian continental

slope of the lower plate, and subduction stopped. Collision is characterised by fold and thrust deformation associated with the

development of structurally high decollements. This collision deformed the basement and cover of the forearc accretionary

prism of the upper plate with part of the unsubducted Australian cover rock sequences from the lower plate. Together with parts

of the forearc basement they now form the exposed Banda orogen. The conversion of the northern flank of the Timor Trough

from being the distal part of the Banda forearc accretionary prism, carried over the Australian continental margin, into a

foreland basin was initiated by the cessation of subduction and simultaneous onset of collisional tectonics.

This reinterpretation of the locked eastern end of the Java Trench proposes that, from its termination south of Sumba to at

least as far east as Timor, and probably far beyond, the Java-Banda Trench and forearc overrode the subducting Australian

proximal continental slope, locally to within 60 km of the shelf break. Part of the proximal forearc’s accretionary prism together

with part of the proximal continental slope cover sequence were detached and thrust northwards over the Java-Banda Trench

and forearc by up to 80 km along the southwards dipping Savu Thrust and Wetar Suture. These reinterpretations explain the

0040-1951/$ - s

doi:10.1016/j.tec

* La Serre,

E-mail addr

(2004) 65–79

ee front matter D 2004 Elsevier B.V. All rights reserved.

to.2004.07.048

St. Pantaleon, 46800 Montcuq, France. Tel.: +33 5 65 31 80 67; fax: +44 709 231 7196.

ess: maudley@wanadoo.fr.

M.G. Audley-Charles / Tectonophysics 389 (2004) 65–7966

present absence of any discernible subduction ocean trench in the southern Banda Arc and the narrowness of the forearc,

reduced to 30 km at Atauro, north of East Timor.

D 2004 Elsevier B.V. All rights reserved.

Keywords: Banda trench; Banda arc; Java trench; Sumba; Timor trough

1. Introduction

Geological and geophysical observations, analyses

and interpretations during the last dozen years have

advanced knowledge and understanding of the region

to a stage that encourages reconsideration of some of

the outstanding unresolved problems of the Banda Arc

(Engdahl et al., 1998; Fortuin et al., 1994, 1997;

Genrich et al., 1994; Hall, 2002; Hall and Wilson,

2000; Hall and Spakman, 2002; Hamilton, 1995;

Harris, 1991; Harris and Long, 2000; Harris et al.,

2000; Hughes et al., 1996; Kreemer et al., 2000;

McCaffrey, 1996; Richardson and Blundell, 1996;

Snyder et al., 1996).

One of the enigmatic geological features of SE

Asia (Fig. 1) is the sudden termination at 1208E of the

3000 km long, more than 6 km deep Sunda–Java

Trench (Fitch, 1970; Audley-Charles, 1975; Chama-

laun et al., 1982; McCaffrey, 1988; Engdahl et al.,

1998; Kreemer et al., 2000; Hall, 2002). A linked

problem is the absence of any clearly recognisable

subduction ocean trench (Fig. 2) that could have been

responsible for the Banda Volcanic Arc and the related

Neogene orogen represented by the islands of the

Outer Banda Arc and their linking submarine ridge

whose geology is best known from the largest islands

of Timor and Seram. A related geological mystery is

the whereabouts within the Banda orogen of the

fundamental decollement that originally separated the

now folded and thrust para-autochthonous cover rocks

from their autochthonous Australian continental base-

ment of the lower plate. This key objective was not

attained by seismic reflection surveys (Richardson

and Blundell, 1996; Snyder et al., 1996). Another

anomaly of the Banda Arc is the locally extreme

narrowness of the Banda forearc between Atauro and

Wetar, reduced to 30–50 km. These riddles are

considered as part of the discussion of the mode of

blocking and of burial of the Banda subduction

trench, and burial of a large part of the forearc

basement during the Neogene–Quaternary collision.

This paper presents no new data and does not

attempt an overall review of Banda Arc evolution. Its

principal objectives are focused on illuminating the

post-15 Ma evolution of the Java Trench eastward of

Sumba with the formation of the Banda subduction

ocean trench. This trench was intimately associated

with the initiation of the Banda Volcanic Arc. The

Banda Trench was then destroyed at the collision

stage of subduction-rollback, as part of the evolution

of the Banda orogen in the region from Sumba to

Timor. The main nine objectives of this paper may be

summarised as:

(a) To locate and define the active continuation of the

Java Trench eastward of Sumba at 5 Ma (Fig. 3).

(b) To explain the absence now of any oceanic

subduction trench in the Banda Arc (Figs. 1

and 4).

(c) To show that the Timor Trough could never have

acted as a subduction ocean trench.

(d) To explain the exceptionally small size of the

Banda forearc now having a width of only 30 to

50 km between Atauro and Wetar islands (Figs.

1, 3 and 4).

(e) To define and distinguish tectonic collision from

subduction-rollback in passive continental mar-

gin-volcanic island arc collision zones, as

exemplified by the southern part of the Banda

Arc.

(f) To explain the presence now of the distal part of

the Banda forearc accretionary prism about 150

km south of the Australian continental margin

exposed in northern Timor (Figs. 1 and 4).

(g) To explain the contribution of overthrusting of

the distal part of the Banda forearc accre-

tionary prism to the evolution of the Timor

Trough as subduction-rollback ceased and

collision commenced.

(h) To indicate the possible role of gravity and of

pore-pressure loss in slowing and stopping the

subduction process as the upper plate rode up the

Fig. 1. Physiographic and tectonic sketch map of the Timor–Sumba region to show: present-day northward dipping tectonic plate suture in the

Java Trench, and the much younger southward dipping suture now relocated north of the Banda volcanic arc; the southward dipping Savu

Thrust (Silver et al., 1983) and the Wetar Suture (Audley-Charles, 1981) now locked; selected transcurrent faults; and location of former, now

obliterated Banda subduction Trench. This assumes the Banda Arc to the north wall of Banda Trench separation dimension (Fig. 3) has not

changed by more than 10% since 5 Ma. Australian continental margin is now locked in collision with the Banda forearc from Wetar to the

Sumba Ridge. The east end of the Java Trench now blocked against the Australian margin has been overridden by the Banda forearc, which

itself has been overridden by the continental margin para-autochthon forming the Roti-Savu Ridge moving on the Savu Thrust. Bathymetry in

metres from Gebco database (IOC, IHO, BODC, 1997).

M.G. Audley-Charles / Tectonophysics 389 (2004) 65–79 67

lower plate ramp to within 60 km of the

Australian shelf break (Fig. 4).

Many of these objectives can be advanced by

recognising the geological significance of the deep

reflectors in the two profiles across the Banda forearc

in the BIRPS data (Richardson and Blundell, 1996;

Snyder et al., 1996) which can be shown to

correspond with the postulated southward dipping

Wetar Suture, and the northward overthrusting of part

of the Australian parautochthon together with part of

the proximal Banda forearc (Figs. 1 and 4).

2. The Timor Trough

2.1. The Timor Trough initiated from tectonic

emplacement of the Banda forearc’s distal part of

the accretionary prism onto the Australian continental

margin

The Timor Trough (Figs. 1 and 4) is a depression

700 km long and between about 30 and 75 km wide.

Its depth varies between 2 and 3.2 km. It is located

entirely within what is now the Australian continen-

tal slope and is underlain by Australian continental

Fig. 2. Seismicity in the Sumba to Timor region of the Banda Arc showing hypocentres below 75 km based on the dataset of Engdahl et al.

(1998). Hypocentres shallower that 75 km are omitted for clarity because they are so widely distributed they obscure deeper seismicity. The

seismic data indicate the presence below the Banda Volcanic Arc of the northward-dipping Australian lithospheric slab.

M.G. Audley-Charles / Tectonophysics 389 (2004) 65–7968

lithosphere (Richardson and Blundell, 1996; Snyder

et al., 1996). At 5 Ma the Banda Trench, located

between the Banda Volcanic Arc and proto-Timor,

was continuing to subduct the northward-dipping

Australian continental margin (Fig. 3). That active

plate boundary (Hall, 2002) was then about 250 km

north of what was to become the present axis of the

Timor Trough about 4 million years later.

The Timor Trough was formed partly as a

consequence of loading (Jordan, 1981) of the

subducting Australian continental lithospheric plate

by the Banda forearc accretionary prism and part of

its basement (Figs. 1, 3 and 4), and partly from

inelastic failure at the beginning of continental

subduction (Tandon et al., 2000). Veevers et al.

(1978) with reference to DSDP 262 showed how the

Timor Trough axis migrated away from Timor

towards the Australian continent as the northern

edge of the former shelf subsided into the trough on

its southern flank (Figs. 3 and 4). An imbricate

structural pattern was mapped by seismic reflection

by Hughes et al. (1996) in the upper part of the

cover sequence in the northern flank of the Timor

Trough. They interpreted this as indicating that these

imbricated Cenozoic deposits represent part of the

Banda forearc accretionary prism; and by implica-

tion, but not stated, of the distal part of the prism.

The southerly limit of the prism has been long

recognised as the Late Cenozoic deformation front

(Barber, 1981). The presence of the apparently

undisturbed Mesozoic–Upper Palaeozoic cover

sequence below the Mid-Base Tertiary decollement,

that marks the base of the imbricate section mapped

by Hughes et al. (1996), indicates the sub-decolle-

ment section is very likely autochthonous. How

much of the structural deformation in this accre-

tionary prism resulted from the pre-collision sub-

duction and how much from post-subduction

collision stresses is difficult to resolve. Hughes et

al. (1996) remarked on the difficulties of determining

the finer structural resolution even after reprocessing

of the seismic data.

Fig. 3. Palaeogeographical sketch map of the eastern part of the Java Trench and its continuation as the Banda Trench during active subduction

at 5 Ma. Positions of the named present-day islands follows Hall (2002) but their shapes are for reference only. Timor, Kisar, Moa, Sermata,

Babar, Roti, Savu and Sumba were entirely submarine at 5 Ma. Volcanic islands and their associated bathymetry sketched in present-day

configuration. Sumba region bathymetry partly after Fortuin et al. (1992, 1994, 1997). Schematic bathymetry in meters used to locate Banda

Trench constrained by assuming 5 km for the depth of the Banda Trench, and 50 km for its width at that depth; and by the positions of the Banda

volcanic arc islands (Hall, 2002); the Australian 5 Ma shelf edge position estimated from DSDP 262 (Veevers et al., 1978), and

micropalaeontology of Neogene part of para-autochthonous Timor Kolbano sequence (Audley-Charles, 1986b) in what was to become Timor

Island after collision at 3.5–2 Ma.

M.G. Audley-Charles / Tectonophysics 389 (2004) 65–79 69

2.2. The Timor Trough as a perisutural foreland basin

A case for the Timor Trough being a foreland basin

has already been made (Audley-Charles, 1986a;

Tandon et al., 2000). The regional geological and

tectonic setting of Neogene-Quaternary perisutural

foreland basins in Eastern Indonesia and Papua New

Guinea was discussed by Audley-Charles (1991). It is

widely recognised (Dickinson, 1974) that thrusting is

usually developed within foreland basins. Bally and

Snelson (1980) demonstrated one can expect to find

thrusting in foreland basins particularly as one

approaches their related orogen. They distinguished

Ampferer-or A-zone subduction, which they defined

as exclusively continent-based, and which they

characterised as typical of perisutural foreland basins.

They contrasted such thrust zones as fundamentally

different in location, function and scale from Wadati–

Benioff subduction zones that are always associated

with an oceanic trench. The imbricate structural

pattern in the northern flank of the Timor Trough

(Hughes et al., 1996) resembles that found in forearc

Fig. 4. Schematic cross section to illustrate present-day relationships between the Banda forearc, the former Neogene Banda Trench and the

detached and overriding Australian para-autochthon. The geological section partly after Harris et al. (2000) and Hughes et al. (1996). Tectonic

structure partly after Richardson and Blundell (1996), Hall and Wilson (2000) and Hall (2002). Location of the top of the lower plate that

functioned during subduction as the basal decollement on or near the top of the Australian autochthonous crystalline basement; and the

overriding forearc that formed the upper plate during subduction, both now hidden by burial in the Timor orogen, indicated schematically. The 5

Ma shelf edge represents its southernmost position estimated from DSDP 262. Position of buried trench linked with southern limit of overriding

forearc. Basis for locating Banda Trench as in Figs. 1 and 3. Line of section A–AVsee Fig. 1.

M.G. Audley-Charles / Tectonophysics 389 (2004) 65–7970

accretionary prisms and in many foreland basins

worldwide. Moreover, the decollement mapped by

Hughes et al. (1996) at the base of the imbricate

section is a relatively superficial structure, which they

correlated with the Mid-Base Tertiary reflector. Below

this decollement in the Timor Trough is the well

stratified autochthonous Mesozoic–Upper Palaeozoic

section that, in a highly deformed state, forms the

Timor orogen together with allochthonous nappes of

the Banda Terrane derived from the forearc basement

(Harris, 1991; Harris et al., 2000). The dominating

presence of these deformed para-autochthonous strata

and forearc basement nappes in the Timor orogen

results from their having been detached from very

much deeper decollements close to their fundamental

basements by subduction followed by collision. It

appears that the Timor Trough began to evolve from

the distal part of an overthrust forearc accretionary

prism, that had overridden the Australian proximal

continental slope, into a foreland basin when colli-

sional deformation began as subduction-rollback

ceased. The indications from reflection seismicity of

recent (Quaternary) deformation in the trough axis is

indicative of a continuation of deformation and

episodic propagation of the thrust front in the Timor

Trough (Masson et al., 1991) corresponding to the

evidence of deformation among the uplifted Quater-

nary coral-algal reefs onshore Timor (Merritts et al.,

1998).

3. The Wetar Strait and Wetar Suture

Timor Island is characterised by fold and thrust

mountains raised to almost 3 km above sea level,

with a total uplift of about 4 km of deep marine

deposits, which may have been an episodic process

initiated between about 3.5 and 2 Ma (Kenyon,

1974; Audley-Charles, 1986b; De Smet et al., 1990).

The island is bounded to both the north and south by

marine depressions (Fig. 1). To the south of the

island is the 3 km deep Timor Trough discussed

above. North of Timor is a very different deep

marine trough, the Wetar Strait, that separates East

M.G. Audley-Charles / Tectonophysics 389 (2004) 65–79 71

Timor from the islands of Wetar and Atauro. The

Strait drops precipitously to more than 3 km depth

from both the north coast of Timor and from its

northern margin along the Banda Volcanic Arc. It

varies in width between 30 km (Atauro) and 50 km

(Wetar). In both these islands of the Banda Arc most

of the volcanic activity ceased in the Pliocene

(Abbot and Chamalaun, 1981). The Wetar Strait

opens westward into a 175-km wide forearc basin

that also descends steeply from the north coast of

Timor and from the submarine margin of the Roti–

Savu Ridge and Sumba Ridge to a depth of 3 km. A

key factor in the search for the now defunct Banda

Trench is that the active Banda Trench must have

always been located south of the Banda Volcanic Arc

because it formed as the Australian continental

margin was subducted northwards into the Banda

Trench. Ocean trenches less than 7 km deep have a

commonly observed width at 5 km depth of about 50

km. The fact that the Wetar Strait now separates the

former Banda Volcanic Arc from the northern edge

of the exposed Australian continental margin, whose

deformed cover rocks crop out on the north coast of

Timor, by only 30–50 km indicates (Audley-Charles,

1981, 1986a; Harris, 1991) that the cover rocks of

the Australian continental margin, exposed in north-

ern Timor, have overridden the southern part of the

Banda forearc (Figs. 1 and 4).

The elevation difference between the bottom of the

Wetar Strait and the north Timor coastal zone,

approaches 4 km in places; and the exceptional

steepness of the gradient between them suggests

faulting on, what in this paper is interpreted as, the

Pliocene Wetar Suture (Fig. 4). Masson et al. (1991)

did not find normal marginal faults with their

reflection seismic survey. But Richardson and Blun-

dell (1996) using BIRPS seismic reflection mapped

multiple south-dipping thrusts in the Banda forearc

immediately east of Wetar and Kisar that correspond

closely with the postulated position (Audley-Charles,

1986a) of the Wetar Suture. Moreover, Masson et al.

(1991) and Snyder et al. (1996) found a south-dipping

thrust near the north coast of Kisar, but they also did

not notice any connection with the postulated funda-

mental Wetar Suture; while Breen et al. (1989), Harris

(1991) and later Prasetyadi and Harris (1996) also

found south-dipping thrusts close to the north coast of

East Timor.

4. Identification of the former Banda Trench

4.1. Implications of earthquake seismicity

The southern part of the Banda Arc is still

characterised by earthquake epicentres (McCaffrey,

1989; Engdahl et al., 1998) indicating the northward-

dipping, Australian lithospheric slab (Figs. 2 and 4)

that descends steeply from below northern Timor,

continuing at depth below the Wetar Strait and the

Banda Arc inactive volcanic islands to below 300 km.

But across Timor and south of it earthquake records

offer no support for subduction now being active south

of the Banda Volcanic Arc (McCaffrey, 1996; Kreemer

et al., 2000). McCaffrey (1996) also emphasised that

earthquake seismology indicates that the movements

between the lower and upper plates of the subduction

system in the Timor region, that had been responsible

for the strong Neogene deformation including much

thrust stacking mapped in the Timor orogenic moun-

tains (Barber, 1981; Harris, 1991; Harris et al., 2000,

Fig 4), have locked. Moreover, McCaffrey (1996),

using preliminary GPS results of Genrich et al. (1994),

added that the geophysical evidence indicated that

dTimor appears to be moving northward relative to the

Sunda Shelf at about the same rate as AustraliaT. Headded that the dnorthward motion of the extinct

volcanic islands north of Timor is revealed with GPS

to be similar to that of southern Timor (Genrich et al.,

1994); hence, little of the northward motion of

Australia is now accommodated by the internal

deformation of the Banda arc structure.T He also

observed that dThe island structure, bounded by the

Timor Trough and the Wetar backarc thrustT (also

known as the Wetar Thrust) d is rigidly thrusting over

the backarc basinT (also known as the South Banda

Sea). Kreemer et al. (2000) on the basis of GPS and

earthquake data concluded that active, southerly

directed thrusting on the Wetar and Flores backarc

thrusts suggest a jump in the locus of convergence

from the Java Trench to the backarc east of 1188E. Theentire region, comprising East Timor, Wetar and part of

the intervening Wetar Strait (Fig. 2), is without active

earthquakes at depths greater than 75 km (Engdahl et

al., 1998). This extraordinary aseismic feature may

accord with the suggestion that part of the Australian

lower plate in this region has ruptured (Price and

Audley-Charles, 1983; Tandon et al., 2000).

M.G. Audley-Charles / Tectonophysics 389 (2004) 65–7972

4.2. Wetar Strait as part of the foundered forearc

Geological mapping in northern Timor has recog-

nised that the deformation suture there includes the

roots of the thrust sheets belonging to the Banda

allochthon derived from the Banda forearc, and which

appear to continue offshore (Harris, 1991; Richardson

and Blundell, 1996; Harris et al., 2000). Moreover,

Masson et al. (1991) dredged samples from the Wetar

Strait of forearc volcanics analysed by Harris (1992).

These data indicate that the present Wetar Strait is a

subsided part of the Banda forearc, implying that the

former Banda Trench must now be south of the Strait

(Figs. 1 and 4).

5. Significance of the allochthonous Banda terrane

The metamorphic Lolotoi and Mutis Complexes

are rightly regarded by Harris (1991) as derived from

the basement of the Banda forearc on the evidence of

the rocks that rest on these metamorphic complexes.

The basement is unconformably overlain in different

places by volcanic and sedimentary rocks of the

Upper Cretaceous to Paleocene Palelo Group, neritic

Eocene limestones, and neritic Lower Miocene Cablac

Limestone. These are entirely allochthonous elements

of the Banda Terrane (Audley-Charles and Harris,

1990; Harris, 1992; Harris and Long, 2000) occupy-

ing a very high structural position. They are vastly

different from any age-equivalent deposits found in

the Australian passive margin sequences. They appear

to have been derived from the basement of the Banda

forearc, and transferred from the base of the upper

plate into the developing orogen (Fig. 4). This may

have occurred during the collision phase when the

northward thrusting Wetar Suture was active. The now

high level, relatively flat-lying thrusts of the Banda

Terrane together with the structurally underlying

folded and thrust para-autochthonous Permian to

Neogene strata are exposed widely throughout Timor.

Charlton (2002) suggested the metamorphic Lolo-

toi and Mutis Complexes were derived from the

Australian plate, but he arrived at this view by

ignoring their stratigraphic and structural relations

with the overlying volcanic and sedimentary Palelo

Group as well as with the Eocene Limestones, and his

interpretation depends on misreporting of work by

Carter et al. (1976) who had described the uncon-

formable relationship of the Cablac Limestone to the

underlying Lolotoi Complex, an observation widely

reported from throughout the island of Timor (Tap-

penbeck, 1940; De Waard, 1957; Audley-Charles and

Carter, 1972).

6. Java Trench–Banda Trench evolution

6.1. Termination of Java Trench at 1208E: bathymet-ric, earthquake and lower plate characteristics

The abrupt termination of the 5–6 km deep and 50

km wide Java Trench at 1208E (Fig. 1) is reinforced

by the observations and conclusions by geophysicists

(Fitch, 1970; Chamalaun et al., 1982; McCaffrey,

1988; Engdahl et al., 1998; Kreemer et al., 2000) of

the dramatic change in Java Trench earthquake

seismicity east of 1188E, especially the abrupt

decrease eastwards of 0–75 km depth earthquakes in

the context of a lack of indications of subduction.

Moreover, east of what is now 1208E the Australian

continental lithosphere had begun before 5 Ma to

replace the Indian Ocean lithosphere as the lower

plate entered the subduction trench (Figs. 3 and 4).

Today, between 1208E and 1218E the Australian

continental slope shows no evidence of the passage

of the Java oceanic trench, and the oblique strike

difference between the existing, well observed Java

Trench at 1208E, and the western edge of the

Australian continental slope is notable, making an

angle of 708. Harris et al. (1998) attributed the small

re-entrant in the 4000 m bathymetric contour to it

representing the southern edge of the westernmost

part of the Timor Trough, it being the distal part of the

Banda forearc accretionary prism. It is notable that,

while the location of the Java Trench with respect to

the Sunda volcanic arc west of 1208E has been

relatively fixed since the early Cenozoic (Hall, 2002),

the nearby Australian continental slope, now abruptly

cutting across the strike of the Java Trench, has only

recently arrived at its present position. At about 10

Ma, the Australian continental margin, now immedi-

ately east of the Java Trench, was located about 600

km further south (Hall, 2002). Thus, details of the

bathymetry, such as the re-entrant revealed in the

4000 m contour (Fig. 1) on the present Australian

M.G. Audley-Charles / Tectonophysics 389 (2004) 65–79 73

slope next to the end of the Java Trench, must be a

relatively recent feature. It would have developed

after the slope occupied its present position with

respect to the Java Trench. Furthermore, the Austral-

ian continental lithosphere, cutting obliquely across

the end of the Java Trench, continues north of the

4000 m re-entrant to its tectonic junction with the

volcanic Sumba Ridge, a distance that varies from 100

to 150 km, marked by the contact indicated by the

Savu Thrust (Fig. 1). This accords with the suggestion

that this submarine morphology developed after the

Australian slope arrived at the Java Trench.

6.2. Plate reconstruction palaeogeographical

implications

Plate reconstructions (Hall, 2002) show the Java

Trench continuing east of Sumba into the west Pacific

from at least 45 Ma, supported by interpretation of

seismic tomography (Hall and Spakman, 2002). This

older part of the Java Trench was eliminated by

collision in eastern Indonesia during the early

Miocene. Between 15 and 10 Ma the Java Trench

developed eastward in a new position where it

functioned as the Banda Trench (not named as such

by Hall, 2002) in association with the beginning of the

Banda Volcanic Arc. This subduction trench was

located, from at least 10 Ma to at least 5 Ma, between

the Banda volcanic arc to the north and what was to

become the island of Timor to the south of this trench.

Hall (2002) reconstruction is used as the basis for the

palaeogeographical sketch map (Fig. 3). Timor is

shown as always on the northern edge of the

Australian continental slope to the south of the Banda

Trench. Because of later northeastward movement of

the Australian continental margin, at 12 Ma the

Timor–Savu region lay some hundreds of kilometres

to the south of the position it occupied at 5 Ma. The

Australian continental margin continued its conver-

gence on the Banda Arc during the Neogene and

Quaternary.

6.3. Reconstruction of the Banda Trench at 5 Ma, and

dating cessation of subduction

The width of the Java Trench, where it is least

affected by incoming sediment from the ocean floor of

the India–Australia plate, is about 50 km. This

corresponds closely with the commonly observed

width at 5 km depth of ocean trenches that are less

than 7 km deep. Hence, 50 km is the postulated width

of the Banda Trench at about 5 Ma (Fig. 3). The Java

Trench has existed since at least the Eocene (Hall,

2002) whereas the Banda Trench cannot be much

older than the earliest volcanics in the Banda Arc

dated as 12 Ma (Abbot and Chamalaun, 1981; Barberi

et al., 1987). On the basis of biostratigraphy the

youngest deformed rocks in the Timor para-autoch-

thon are N18 pelagic lutites in the Kolbano sequence,

and the oldest autochthonous rocks resting uncon-

formably on the deformed para-autochthon belong to

the lowest part of the Batu Putih Limestone (Kenyon,

1974) determined as N18–19 which is between 5.0

and 3.5 Ma in age (Early to Mid-Pliocene). Timor

studies reveal a Late Pliocene–Quaternary regional

uplift that amounts to about 4 km; such a long-lived

and substantial regional uplift requires a fundamental

tectonic cause. Detailed studies have recognised both

local fault-related uplift and subsidence (Kenyon,

1974; De Smet et al., 1990). Uplift began with

Kenyon (1974) Fatu Laob uplift phase whose revised

date (Audley-Charles, 1986b) is Late Pliocene (N21)

ranging from 3.5 to 2.0 Ma. Beginning from between

~3.5 and 2.0 Ma, the regional uplift has continued,

with some local subsidence. The unconformable

relationships described above suggest that subduction

may have ceased in this part of the Banda Trench at

3.5 Ma. This implies that the active life of the trench

here was from ~12 to ~3.5 Ma, but if the onset of

uplift in this region was as late as 2.0 Ma and if this

uplift, which has continued throughout the Quater-

nary, is an isostatic response to cessation of sub-

duction then Banda Trench subduction may have

stopped and collision began as late as 2 Ma.

6.4. Locating the present position of the defunct

Banda trench

The defunct Banda Trench has been located in this

paper (Figs. 1 and 4) by considering the consequences

of convergence between the northward-subducting

Australian plate and the Banda Volcanic Arc from 5

Ma (Fig. 3) to the date of onset of arc–continent

collision. Geological mapping indicates collision

began between about 3.5 and 2 Ma. As explained

above, collision is correlated here with the locking of

M.G. Audley-Charles / Tectonophysics 389 (2004) 65–7974

the subduction trench. The important assumption is

that from 5 Ma to the time the trench was locked the

Banda forearc basement did not grow southwards, and

the distance from the volcanic arc to the north wall of

the Banda trench did not change significantly. Note

that in Fig. 3 the distance from the arc (at Wetar) to the

north wall of the Banda Trench was 100 km at 5 Ma. In

Figs. 1 and 4, representing present-day reconstruc-

tions, this distance has shortened by about 10%. The

consequences of the convergence of the Australian

margin with the Banda Volcanic Arc are revealed by

plotting the position of proto-Timor and the Australian

continental margin at 5 Ma in Fig. 3 in their relative

present-day positions as shown in Fig. 1. This indicates

that the present position of the Banda Trench (now

defunct) and its closely linked upper plate, comprising

the southern part of the Banda forearc basement, must

both be below Timor. The only important assumption

has been that the distance from the Banda Arc to the

north wall of the Banda Trench has not increased since

5 Ma (Figs. 1 and 4 imply a 10% shortening which

might well be expected to result from the tectonic

compression). The case for the location of the Banda

Trench at 5 Ma has already been set out above. It is

clear that there is little room for manoeuvring the

geographical positions of any feature once the width of

the Banda Trench at 5 km depth is taken to have the

value of 50 km, and that, following Hall (2002), the

position of the Java Trench in the region of Sumba has

not altered significantly during the last 5 Ma.

7. Distinguishing subduction from tectonic collision

When a passive continental margin converging on

a volcanic island arc system, whether by subduction

or by rollback, reaches the stage that all of the oceanic

lithosphere in any sector has passed below the inner

forearc wall of the trench so that the continental

lithosphere of the converging continental slope begins

to pass below this inner forearc trench wall the

condition of continental margin-arc collision may be

said to have commenced. Given that the shape of the

strike of the converging continental margins may be

irregular there maybe varying degrees of diachronism

in the tectonic collision.

Collision in the southern part of the Banda Arc was

initiated when the overriding forearc basement of the

upper plate reached the proximal part of the Austral-

ian slope whose basement was comprised of con-

tinental lithosphere, subduction probably began to

slow and stopped when the distal edge of the shelf

was at a distance of about 60 km to the south (Figs. 3

and 4). Where subduction or rollback of all the Indian

Ocean lithosphere below the Banda Arc was com-

pleted in the Sumba-Timor region the oceanic Banda

Trench began to form part of the orogenic northern

margin of the foreland basin evolving to become the

present Timor Trough.

The now missing, defunct Banda Trench can only

have been located, when it was actively subducting,

where the seismicity (Engdahl et al., 1998) implies its

former presence between the northwards subducting

Australian plate and the SE Asian Banda Volcanic

Arc. The now invisible, obliterated Banda Trench, that

probably ceased to subduct in the Timor region

between 3.5 and 2 Ma, is now located below the

islands of Timor and Roti (Figs. 1 and 4). Here it is

obscured by the thick pile of folded and thrust para-

autochthonous cover rocks from the Australian con-

tinental rise and slope now thrust northwards by the

Wetar Suture over the overriding forearc.

These now para-autochthonous rocks first began to

form the Banda forearc accretionary prism from the

time the distal Australian continental rise sediments

began to arrive in the Banda Trench, when the Banda

Volcanic Arc began to appear at about 12 Ma. This

Banda forearc accretionary prism grew until the

subduction process was halted probably after slowing

as the Australian continental slope, with its crystalline

basement and great thickness of overlying Lower

Permian to Mid-Jurassic Gondwana sequences, and

post-rift Upper Jurassic to Pliocene Australian rifted

continental margin sequences, entered the subduction

trench between about 5 and 2 Ma.

The locking of the subduction process is associated

with the gradual onset of tectonic continental margin–

volcanic arc collision. It is interpreted here as having

begun between 3.5 and 2 Ma. Thickness of the lower

plate probably increased as the slope entered the

trench. A steepening of the subduction angle may

have impeded the downward motion of the lower

plate into the trench. Subduction may then have been

halted by the increasing gravitational resistance to the

upward movement of the forearc over the Australian

margin on the incline rising from the floor of the

M.G. Audley-Charles / Tectonophysics 389 (2004) 65–79 75

Banda Trench southwards up to arrive about 60 km

from the shelf break (Fig. 4), where the sea floor could

have been at less than 1000 m depth. Another braking

effect might have been the loss of pore-fluid pressure,

from below the overriding forearc basement moving

on a decollement of Jurassic shale and escaping into

the more permeable formations of the Australian

cover sequence. Continuing or reactivated crustal

shortening after the cessation of subduction consti-

tutes tectonic collision. The collision in the southern

part of the Banda Arc resulted from compression of

the forearc accretionary wedge which consequently

became the Australian para-autochthon on both upper

and lower plates, with part of its forearc basement

becoming nappes of the allochthonous Banda Terrane.

This upper plate tectonic flake, in response to

continuing compressive stresses, produced north-

ward-verging structures concentrated along the north

coast of Timor as mapped by Harris (1991) and

Prasetyadi and Harris (1996). This northward thrust-

ing over the forearc helped close the forearc to only 30

km in the region of Atauro. Harris (1991) and

Prasetyadi and Harris (1996) also identified various

thrust slices of the Banda Terrane incorporated into

this northward-verging thrust wedge (Fig. 4) carried

on the Wetar Suture (Audley-Charles, 1986a).

Another characteristic of subduction and roll-back

related processes that distinguishes them from colli-

sion related processes is the structural level of the

principal decollement. In subduction and roll-back the

main decollement is the upper and lower plate

boundary, whereas in collision the principal decolle-

ments develop at a higher level even if they occur as

structurally low as to be within the cover rock

sequence on the autochthonous crystalline basement

of the lower plate, as in the Timor Trough (Hughes et

al., 1996). Both subduction and roll-back (Hamilton,

1988; 1995) appear to transfer much of the cover rock

sequence from the descending lower plate to the

leading edge of the upper plate as the lower plate

passes below the forearc basement that forms the

inner (arc) wall of the trench.

7.1. Main and later phases of collision

The recognition by Hughes et al. (1996) of

accretionary prism-type deformation in the northern

flank of the Timor Trough indicates they are much

simpler structures than those of the main collision

phase of structural deformation so evident in the

exposures on Timor (Fig. 4). Following the transfer of

the Australian cover rocks by subduction from the

lower plate into the forearc to form the accretionary

wedge of the upper plate, the main collision deforma-

tion involved at least three recognisable stages (i)

severe shortening of the forearc accretionary wedge to

form part of the Australian para-autochthon; (ii)

detachment of part of the underlying forearc basement

eventually to form the nappes of the allochthonous

Banda Terrane intimately associated with (iii) trans-

port northwards by the Wetar Suture and associated

uplift of the para-autochthon and slices of the forearc

basement to become the Timor orogen. In contrast, the

Timor Trough accretionary prism structures appear to

be mainly the result of some probably reduced

convergent forces (McCaffrey, 1996) towards the

end of subduction and especially after that ceased in

the Timor region of the Banda Trench between 3.5

and 2.0 Ma. A major structural feature developed

towards the end of subduction and during collision

was the Late Pliocene Wetar Suture, now locked. The

crustal shortening resulted in the deformation that

characterises that part of the para-autochthon detached

from part of the unsubducted Australian lower plate

causing the detached para-autochthon to overthrust

the forearc tectonic flake (Fig. 4) on the Wetar Suture.

The compressional strain continued into the Pleisto-

cene when the subduction polarity reversed and the

Wetar Thrust began to develop a southward-dipping

subducting plate margin (McCaffrey, 1996; Hall and

Wilson, 2000). The gentle folding of the Viqueque

Group basins (Kenyon, 1974) that affected some of

the reef terraces (Merritts et al., 1998) also indicates

that some crustal shortening continued into the late

Quaternary.

8. Sumba Ridge–Banda Trench relations

Fig. 1 shows the ENE-striking submarine Roti–

Savu Ridge has a southward-dipping thrust junction,

the Savu Thrust (Silver et al., 1983), with the southern

part of the south–east striking submarine Sumba

Ridge. Fortuin et al. (1994), Rutherford et al. (2001)

and Lytwyn et al. (2001) interpret the Sumba Ridge as

part of the Banda forearc. Comparison was made by

M.G. Audley-Charles / Tectonophysics 389 (2004) 65–7976

Audley-Charles (1985) of the allochthonous Timor

Banda Terrane, thought to have been derived from the

Banda forearc, with some sections in West Sumba

described by Van Bemmelen (1949), Chamalaun et al.

(1982) and Von der Borch et al. (1983). For example,

both Timor (Tappenbeck, 1940) and West Sumba

sections have Lower Paleogene ignimbrites overlain

by neritic limestones with Middle and Upper Eocene

large foraminifera, and both have Lower Miocene

neritic limestones. Both have an angular unconformity

at the base of the Paleogene and an unconformity at

the base of the Miocene.

Rutherford et al. (2001) have interpreted Sumba as

having been part of the Sunda volcanic arc from Late

Cretaceous to Early Miocene. If the tentative corre-

lation of the Palelo Group and associated sequences

unconformable on the allochthonous metamorphic

complexes in Timor with the Cretaceous–Lower

Miocene volcanics in Sumba is shown to be valid

then this would strengthen the correlation of the

Lolotoi and Mutis metamorphic complexes and their

unconformable Palelo cover (Banda Terrane) with the

Banda forearc basement and with the igneous base-

ment of Sumba.

The Pliocene–Quaternary uplift in East Sumba

(Fortuin et al., 1994, 1997) suggests comparison with

the Late Pliocene–Quaternary uplift of 4 km in Timor.

The Timor uplift has been tentatively attributed (Price

and Audley-Charles, 1987) to isostatic rebound of the

blocked subduction zone of the defunct Banda Trench

no longer able to subduct the Australian continental

lower plate.

The suggestion by Rutherford et al. (2001) for the

initial collision (undefined and not distinguished from

subduction) between the Australian continental mar-

gin and Banda Arc having been as early as Early

Miocene, 16–18 Ma, conflicts with the results from

numerous field studies, including apatite FT ages for

dating of exhumation in Timor (Harris et al., 2000),

geological mapping across Timor and other islands in

the southern Banda Arc (Audley-Charles, 1986a,

1986b; Harris, 1991), offshore geophysics around

the Banda Arc (Bowin et al., 1980) and the review by

Hall (2002). Keep et al. (2003) suggest that the

collision (undefined and not distinguished from

subduction) of what is here described as the sub-

marine Roti–Savu Ridge with the submarine Sumba

Ridge, that they agree represents a part of the Banda

volcanic forearc, occurred at about 8 Ma. These early

dates for Banda Arc collision in the Sumba–Timor

region cannot be reliably based on the radiometric

ages for metamorphism in the Permian Aileu Meta-

morphic Complex of Timor. Harris et al. (2000) on the

basis of vitrinite reflectance, conodont alteration and

illite crystallinity and structural mapping questioned

the interpretations of pre-Late Miocene tectonic burial

and pre-Pliocene exhumation of the metamorphism of

the Aileu Complex Australian continental margin in

Timor. They pointed out that dthe only possible

evidence of such deep burial is a thin Barrovian

metamorphic sequence along the north coast in the

Aileu ComplexT (Berry and Grady, 1981). They went

on to emphasise that dIt is also possible that the Banda

Volcanic Arc has thermally disturbed the Aileu

Complex, since it is as close as 30–40 km in placesT.Furthermore, Hall (2002) and Hall and Wilson (2000)

have argued that such metamorphic dates in the

Permian Aileu Complex (that is part of the Gondwana

cratonic basin sequence and belongs structurally with

the Australian continental basement to the post-Late

Jurassic rift sequence) are more easily attributable to

an extension episode beginning about 12 Ma and

associated with Banda Arc volcanism. There is no

independent evidence for this early collision date, and

there is much in the Timor and Sumba islands to

contradict it, as indicated above.

The Banda forearc is made exceptionally wide by

the Sumba ridge which projects about 250 km south

of the existing narrow Banda forearc (Figs. 1 and 3),

so that a collision about 5 Ma earlier than the collision

in the Timor region between the Australian continen-

tal margin forming the Roti–Savu Ridge and the

Sumba Ridge of Banda forearc seems superficially

possible. However, their suggestion (Keep et al.,

2003, Fig. 10a) that the Roti–Savu Ridge at 8 Ma

formed a narrow peninsula that protruded orthogo-

nally from the regional strike of the Australian

margin, the ridge having a rectilinear outline and

northwest strike, is difficult to reconcile with the

geological strike of the Australian margin in the

southern Banda Arc. Keep et al. (2003) appear to be

making a speculative exaggeration of the shape and

size of their projecting ridge (that they described as bafingerQ to indicate its length/width ratio) of the step-

rifted margin in the Sumba region. Their attribution of

the stepped morphology of the present rifted con-

M.G. Audley-Charles / Tectonophysics 389 (2004) 65–79 77

tinental margin to a pre-subduction age of pre-12 Ma

ignores the effects of subduction and collision-related

NNE transcurrent faulting that has affected the shape

of the present northern margin of the Australian

continent in the Banda orogen (Fig. 1). But even more

importantly it is impossible to reconcile the age of an

interpreted Late Miocene collision with the record of

uninterrupted pelagic deposition, perturbed only by

slumping in a deep water environment, at the

supposed site of the collision on Sumba itself (Fortuin

et al., 1992, 1994). The youngest pelagic sediments on

Sumba are late Early Pliocene (N19/20) in age (Roep

and Fortuin, 1996; Fortuin et al., 1997) suggest

emergence of Sumba did not occur before 3 Ma, an

age very close to that inferred from Timor for the

beginning of uplift. Thus, the evidence from

extremely well-exposed sections on Sumba support

the timing for the end of subduction and onset of

collision suggested in this paper.

Acknowledgements

In preparing this article, I have been most kindly

assisted by busy friends: Tony Barber, Graham Evans

and Paul Henderson provided timely practical help.

Robert Hall helped generously with art work for the

four figures, editing and in other practical ways. He

and Ron Harris both read and suggested improve-

ments to earlier versions of this paper, and I benefited

too from their e-mailed tutorials for a superannuated

but formerly active toiler in the Banda Arc. I am also

indebted to the helpful suggestions from Myra Keep

and an anonymous referee for improvements to this

paper.

References

Abbot, M.J., Chamalaun, F.H., 1981. Geochronology of some

Banda Arc volcanics. The Geology and Tectonics of Eastern

Indonesia. Geological Research and Development Centre,

vol. 2, pp. 253–268. Spec. Pub.

Audley-Charles, M.G., 1975. The Sumba fracture: a major

discontinuity between eastern and western Indonesia. Tectono-

physics 26, 213–228.

Audley-Charles, M.G., 1981. Geometrical problems and implication

of large scale overthrusting in the Banda Arc-Australian margin

collision zone. In: McClay, K.R., Price, N.J. (Eds.), Thrust and

Nappe Tectonics. Geological Society of London, Spec. Pub.,

vol. 9, pp. 407–416.

Audley-Charles, M.G., 1985. The Sumba Enigma: is Sumba a

diapiric nappe in process of formation? Tectonophysics 119,

435–449.

Audley-Charles, M.G., 1986a. Timor-Tanimbar Trough: the fore-

land basin to the evolving Banda orogen. Foreland Basins. Spec.

Publs. Int. Ass. Sediment, pp. 91–102.

Audley-Charles, M.G., 1986b. Rates of neogene and quaternary

tectonic movements in the southern Banda Arc based on

micropalaeontology. Journal of Geological Society (London)

143, 161–175.

Audley-Charles, M.G., 1991. Tectonics of the New Guinea Area.

Annual Review of Earth and Planetary Sciences 19, 17–41.

Audley-Charles, M.G., Carter, D.J., 1972. Palaeogeographical

significance of some aspects of Palaeogene and Early Neogene

stratigraphy and tectonics of the Timor Sea Region. Palae-

ogeography, Palaeoclimatology, Palaeoecology 11, 247–264.

Audley-Charles, M.G., Harris, R.A., 1990. Allochthonous terranes

of the South-west Pacific and Indonesia. Philosophical Trans-

actions of the Royal Society of London 331, 571–587.

Bally, A.W., Snelson, S., 1980. Facts and principles of World

petroleum occurrence: realms of subsidence. Memoir-Canadian

Society of Petroleum Geologists 6, 9–94.

Barber, A.J., 1981. Structural interpretations of the island of Timor,

Eastern Indonesia. In: Barber, A.J., Wiryosujono (Eds.), The

Geology and Tectonics of Eastern Indonesia. Geological

Research and Development Centre, vol. 2, pp. 183–198. Spec.

Pub.

Barberi, F., Bigioggero, B., Boriani, A., Cattaneo, M., Cavallin, A.,

Cioni, R., Eva, C., Gelmini, R., Giorgetti, E., Iaccarino, S.,

Innocenti, F., Marinelli, G., Slejko, D., Sudradjat, A., 1987. The

island of Sumbawa: a major structural discontinuity in the

Indonesia arc. Bulletin of the Geological Society of Italy 106,

547–620.

Berry, R.F., Grady, A.E., 1981. Deformation and metamorphism of

the Aileu Formation, north coast, East Timor; its tectonic

significance. Journal of Structural Geology 3, 143–167.

Bowin, C., Purdy, G.M., Johnson, C., Shor, G., Lawver, L.,

Hartono, H.M.S., Jezek, P., 1980. Arc-continent collision in

the Banda Sea region. American Association of Petroleum

Geologists Bulletin 64, 37–58.

Breen, N.A., Silver, E.A., Roof, S., 1989. The Wetar backthrust belt,

eastern Indonesia: the effects of accretion against an irregularly

shaped arc. Tectonics 8, 85–98.

Carter, D.J., Audley-Charles, M.G., Barber, A.J., 1976. Strati-

graphical analysis of island arc-continental margin collision in

eastern Indonesia. Journal of Geological Society (London) 132,

179–198.

Chamalaun, F.H., Grady, A.E., von der Borch, C.C., Hartono,

H.M.S., 1982. Banda Arc tectonics: the significance of the

Sumba Island (Indonesia). In: Watkins, J.S., Drake, C.L. (Eds.),

Studies in Continental Margin Geology. Amer. Assoc. Petrol.

Geologists, vol. 34, pp. 361–375.

Charlton, T.R., 2002. The structural setting and tectonic significance

of the Lolotoi, Laclubar and Aileu metamorphic massifs, East

Timor. Journal of Asian Earth Sciences 20, 851–865.

M.G. Audley-Charles / Tectonophysics 389 (2004) 65–7978

de Smet, M.E.M., Fortuin, A.R., Troelstra, S.R., van Marle, L.J.,

Karmini, M., Tjokrosapoetro, S., Hadiwasastra, S., 1990.

Detection of collision-related vertical movements in the Outer

Banda Arc (Timor, Indonesia), using micropalaeontological

data. Journal of Southeast Asian Earth Sciences 4, 337–356.

De Waard, D., 1957. The third Timor geological expedition,

preliminary results. Indonesian Journal for Natural Science

113, 7–42.

Dickinson, W.R., 1974. Plate Tectonics and sedimentation. Tec-

tonics and Sedimentation. Soc. Econ. Paleontol. Mineral. Spec.

Publ., vol. 22, pp. 1–22. Tulsa: SEPM.

Engdahl, E.R., van der Hilst, R., Buland, R., 1998. Global

teleseismic earthquake relocation and improved travel times

and procedures for depth determination. Bulletin of the

Seismological Society of America 88, 722–743.

Fitch, T.J., 1970. Earthquake mechanisms and island-arc tectonics

in the Indonesian-Philippine region. Bulletin of the Seismolog-

ical Society of America 60, 565–591.

Fortuin, A.R., Roep, T.B., Sumosusastro, P.A., Vanweering, T.C.E.,

van der Werff, W., 1992. Slumping and sliding in Miocene and

Recent developing arc basins, onshore and offshore Sumba

(Indonesia). Marine Geology 108, 345–363.

Fortuin, A.R., Roep, Th.B., Sumosusastro, P.A., 1994. The Neogene

sediments of east Sumba, Indonesia-products of a lost arc?

Journal of Southeast Asian Earth Sciences 9, 67–79.

Fortuin, A.R., van der Werff, W., Wensink, H., 1997. Neogene basin

history and paleomagnetism of a rifted and inverted forearc

region, on- and offshore Sumba, Eastern Indonesia. Journal of

Asian Earth Sciences 15, 61–88.

Genrich, J., Bock, Y., McCaffrey, R., Calais, E., Stevens, C., et al.,

1994. Kinematics of the eastern Indonesian island arc estimated

by Global positioning system measurements. EOS 75, 116.

Hall, R., 2002. Cenozoic geological and plate tectonic evolution of

SE Asia and the SW Pacific: computer-based reconstructions,

model and animations. Journal of Asian Earth Sciences 20,

353–431.

Hall, R., Spakman, W., 2002. Subducted slabs beneath the eastern

Indonesia-Tonga region: insights from tomography. Earth and

Planetary Science Letters 201, 321–336.

Hall, R., Wilson, M.E.J., 2000. Neogene sutures in eastern

Indonesia. Journal of Asian Earth Sciences 18, 781–808.

Hamilton, W., 1988. Plate Tectonics and Island Arcs. Geological

Society of America Bulletin 100, 1503–1527.

Hamilton, W., 1995. Subduction and magmatism. In: Smellie, J.L.

(Ed.), Volcanism Associated with Subduction at Consuming

Plate Margins. Geological Society of London, Special Publica-

tion, vol. 81, pp. 3–28.

Harris, R.A., 1991. Temporal distribution of strain in the active

Banda orogen: a reconciliation of rival hypotheses. Journal of

Southeast Asian Earth Sciences 6, 373–386.

Harris, R.A., 1992. Peri-collisional extension and formation of

Oman-type ophiolites in the Banda and Brooks Range. In:

Parson, L.M., Murton, B.J., Browning, P. (Eds.), Ophiolites and

their modern oceanic analogues. Geological Society of London,

Special Publication, vol. 60, pp. 301–325.

Harris, R.A., Sawyer, R.K., Audley-Charles, M.G., 1998. Colli-

sional melange development: Geologic associations of active

melange-forming processes with melange facies in the western

Banda orogen, Indonesia. Tectonics 17, 458–479.

Harris, R.A., Kaiser, J., Hurford, A., Carter, A., 2000. Thermal

history of Australian passive margin cover sequences accreted to

Timor during Late Neogene arc-continent collision, Indonesia.

Journal of Asian Earth Sciences 18, 47–69.

Harris, R.A., Long, T., 2000. The Timor ophiolite, Indonesia: Model

or myth? Geological Society of America, Special Paper 349,

321–330.

Hughes, B.D., Baxter, K., Clark, R.A., Snyder, D.B., 1996. Detailed

processing of seismic reflection data from the frontal part of the

Timor trough accretionary wedge, eastern Indonesia. In: Hall,

R., Blundell, D. (Eds.), Tectonic Evolution of Southeast Asia.

Geological Society Special Publication, vol. 106, pp. 75–83.

IOC, IHO, BODC, 1997. GEBCO-97: The 1997 Edition of the

Gebco Digital Atlas. British Oceanographic Data Centre.

Jordan, T.E., 1981. Thrust loads and foreland basin evolution,

Cretaceous, western United States. American Association of

Petroleum Geologists Bulletin 65, 291–329.

Keep, M., Jones, R., Longley, I., 2003. Sumba and its effect on

Australia’s north west margin. In: Hillis, R.R., Muller, R.D.

(Eds.), The Evolution and Dynamics of the Australian Plate.

Geological Society of Australia Special Publication, vol. 22,

pp. 303–312.

Kenyon, C.S., 1974. Stratigraphy and sedimentology of the Late

Miocene to Quaternary deposits of Timor. PhD Thesis,

University of London.

Kreemer, C., Holt, W.E., Goes, S., Govers, R., 2000. Active

deformation in eastern Indonesia and the Philippines from GPS

and seismicity data. Journal of Geophysical Research 105,

663–680.

Lytwyn, J., Rutherford, E., Burk, K., Xia, C., 2001. The geo-

chemistry of volcanic, plutonic and turbiditic rocks from Sumba,

Indonesia. Journal of Asian Earth Sciences 19, 481–500.

Masson, D.G., Milsom, J., Barber, A.J., Sikumbang, N., Dwiyanto,

B., 1991. Recent tectonics around the island of Timor, eastern

Indonesia. Marine and Petroleum Geology 8, 35–49.

McCaffrey, R., 1988. Active tectonics of the eastern Sunda and

Banda arcs. Journal of Geophysical Research 93, 15163–15182.

McCaffrey, R., 1989. Seismological constraints and speculations on

Banda Arc tectonics. Netherlands Journal of Sea Research 24,

141–152.

McCaffrey, R., 1996. Slip partitioning at convergent plate bounda-

ries of SE Asia. In: Hall, R., Blundell, D.J. (Eds.), Tectonic

evolution of Southeast Asia. Geological Society of London,

Special Publication, vol. 106, pp. 3–18.

Merritts, D., Eby, R., Harris, R.A., Edwards, R.L., Cheng, H., 1998.

Variable rates of Late Quaternary surface uplift along the Banda

Arc-Australian plate collision zone, Eastern Indonesia. In:

Stewart, I.S., Vita-Finzi, C. (Eds.), Coastal Tectonics. Geological

Society of London, Special Publications, vol. 146, pp. 213–224.

Prasetyadi, C., Harris, R.A., 1996. Hinterland structure of the active

Banda arc-continent collision, Indonesia: constraints from the

Aileu complex of East Timor. Proceedings of 25th Convention

of the Indonesian Association of Geologists, pp. 144–173.

Price, N.J., Audley-Charles, M.G., 1983. Plate rupture by hydraulic

fracture resulting in overthrusting. Nature 306, 572–575.

M.G. Audley-Charles / Tectonophysics 389 (2004) 65–79 79

Price, N.J., Audley-Charles, M.G., 1987. Tectonic collision

processes after plate rupture. Tectonophysics 140, 121–129.

Richardson, A.N., Blundell, D.J., 1996. Continental collision in the

Banda arc. In: Hall, R., Blundell, D.J. (Eds.), Tectonic Evolution

in Southeast Asia. Geological Society of London, Special

Publication, vol. 106, pp. 47–60.

Roep, T.B., Fortuin, A.R., 1996. A submarine slide scar and channel

filled with slide blocks and megarippled Globigerina sands of

possible contourite origin from the Pliocene of Sumba,

Indonesia. Sedimentary Geology 103, 145–160.

Rutherford, E., Burke, K., Lytwyn, J., 2001. Tectonic history of

Sumba Island, Indonesia, since the Late Cretaceous and its rapid

escape into the forearc in the Miocene. Journal of Asian Earth

Sciences 19, 453–479.

Silver, E.A., Reed, D., McCaffrey, R., Joyodiwiryo, Y., 1983. Back

arc thrusting in the Eastern Sunda Arc, Indonesia: a conse-

quence of arc-continent collision. Journal of Geophysical

Research 88, 7429–7448.

Snyder, D.B., Milsom, J., Prasetyo, H., 1996. Geophysical

evidence for local indentor tectonics in the Banda arc east of

Timor. In: Hall, R., Blundell, D.J. (Eds.), Tectonic Evolution

of Southeast Asia. Geological Society of London, Special

Publication, vol. 106, pp. 61–73.

Tandon, K., Lorenzo, J.M., OTBrien, G.W., 2000. Effective elastic

thickness of the northern Australian continental lithosphere

subducting beneath the Banda orogen (Indonesia): inelastic

failure at the start of continental subduction. Tectonophysics

329, 39–60.

Tappenbeck, D., 1940. Geologie des Mollogebirges und einiger

benachbarter Gebiete (Nierl7ndisch Timor). Geological Expedi-

tion to the Lesser Sunda Islands, vol. 1. University of

Amsterdam, pp. 1–105.

van Bemmelen, R.W., 1949. The Geology of Indonesia.

Government Printing Office, Nijhoff, The Hague. 2nd ed.,

1970; 732 pp.

Veevers, J.J., Falvey, D.A., Robins, S., 1978. Timor trough and

Australia: facies show topographic wave migrated 80 km during

the past 3 m.y. Tectonophysics 45, 217–227.

von der Borch, C.C., Grady, A.E., Hardjoprawiro, S., Prasetyo, H.,

Hadiwisastra, S., 1983. Mesozoic and late Tertiary submarine

fan sequences and their tectonic significance, Sumba, Indonesia.

Sedimentary Geology 37, 113–132.