IPA17-91-G

PROCEEDINGS, INDONESIAN PETROLEUM ASSOCIATION

Forty-First Annual Convention & Exhibition, May 2017

SERAM, THE SERAM TROUGH, THE ARU TROUGH, THE TANIMBAR TROUGH AND THE
WEBER DEEP: A NEW LOOK AT MAJOR STRUCTURES IN THE EASTERN BANDA ARC

Robert Hall*
Adi Patria*
Ramadhan Adhitama **
Jonathan M. Pownall ***
Lloyd T. White ****

ABSTRACT indicates active E-W extension of the Aru trough and

strike-slip movement on the Tarera–Aiduna fault
Multibeam bathymetry data, combined with 2D zone.
seismic data and studies on land, provide new
insights into the major structures in the northern and The Seram trough began to form in the Late Pliocene
eastern Banda Arc. In the west, north of Buru, is the due to loading by the Seram FTB. The Tanimbar
deep Buru basin where there is evidence of extension trough originated in the Late Miocene as an elongate
with probable Neogene oceanic crust. To the east, the extensional structure within the Australian
thrust front of the Seram fold and thrust belt (FTB) continental margin. Neither troughs are subduction
can be traced along the Seram trough which turns zones. Almost all active deformation in the Seram
from E-W to N-S to pass through the Kai Islands into FTB is on Seram, and offshore south of the trough
the Tanimbar trough. The thrust front is not east of and north of the Kumawa fault. There is almost no
the Kai Islands as often drawn. Between the Kai and seismicity at the trough itself, unlike modern
Aru Islands is the Aru trough bounded by NNE- subduction zones. South of the Kumawa fault is a
SSW-trending normal faults which converge to the wide FTB, trending roughly N-S towards the Kai
SW and can be traced southwards into the Tanimbar Islands and Tanimbar, with young mud volcanism
trough. but little seismicity. Immediately west of the FTB is
the Weber deep, the Earth’s deepest forearc basin,
An important WNW-ESE-trending, probably which is a young extensional basin bounded by a
inactive, fault bounds the northern end of the Aru major low angle detachment on its eastern side. It is
trough south of the Bird’s Head that we name the the most recent manifestation of eastward Banda
Kumawa fault. It is partly buried below disturbed subduction rollback, and may still be active.
sediments that have slumped from the shelf south of
the Lengguru FTB. This is a major tectonic feature. INTRODUCTION
To the north there is a late Neogene history of
contraction that formed the Misool–Onin–Kumawa Banda Arc (Figure 1) subduction has been known for
Ridge and Lengguru FTB. To the south, east of the many years (e.g. Hamilton, 1974, 1979; Cardwell &
thrust front, extensional structures dominate. We Isacks, 1978), as has the remarkable curve in Benioff
propose that the Kumawa fault links to the Kawa zone contours through almost 180° at the eastern end
shear zone of Seram which together form a of the arc (Figure 2). The change in orientation has
lithosphere-scale sinistral shear zone active during been interpreted in two different ways: a single slab
the Miocene and Pliocene which exhumed deep crust that was deformed in the mantle (Hamilton, 1979) or
and mantle on Seram. The Kumawa fault offsets the two separate slabs subducted from the north and
Seram thrust front close to the Kumawa peninsula. south (e.g. Cardwell & Isacks, 1978; McCaffrey,
To the east is the E-W-trending Tarera–Aiduna fault 1989; Das, 2004). This argument has now been
zone that can be traced offshore only a short distance resolved in favour of the single slab model (Spakman
from the Bird’s Head. This is a very young structure and Hall, 2010; Hall, 2012) and the geometry of the
with little displacement which does not reach the slab in the mantle is explained by the rollback of the
Seram Trough. However, present-day seismicity subduction hinge into a pre-existing U-shaped
* Royal Holloway University of London
** University of Trisakti
*** University of Australian National
**** University of Wollongong
embayment of Jurassic oceanic crust, the Banda around the Banda arc are the relationship with the
embayment, within the Australian continental Aru trough, often identified as an important region of
margin. extension (e.g. Schlüter and Fritsch, 1985; Charlton
et al., 1991) and to faults traced west from New
The islands of the Banda arc have been separated into Guinea, such as the Tarera–Aiduna fault system,
two parts; an inner volcanic arc and an outer arc in which is commonly connected to the Seram trough
which the largest islands are Timor in the south and (e.g. Hamilton, 1979; Katili, 1986; Charlton et al.,
Seram in the north (Figure 1). Timor is clearly the 1991; Harris, 2011) or traced much further west
expression of Pliocene arc–continent collision and beyond the trough (e.g. Katili, 1989; Linthout et al.,
has received by far the most attention from 1991). Teas et al. (2009) suggested an important left-
researchers in the region. Seram and the islands of lateral strike-slip shear system at the eastern end of
the northern Banda arc remain less accessible and the Seram trough is the offshore extension of the
have been the subject of much less research. Tarera–Aiduna fault zone.
Although Seram has been interpreted as effectively a
mirror image of Timor (Audley-Charles et al., 1979), The regional tectonic debate raises many questions.
the supposed arc that collided with the northern How far west can the Seram trough be traced? How
continental margin of the Banda embayment is not and where are the Seram, Tanimbar and Aru troughs
obvious and the present-day active volcanic arc connected? Are any of the troughs the trace of
terminates at Banda at the eastern end of the arc. through-going lithospheric faults connecting the
However, the shape of the subducted slab in the subducted slab to the surface? What are the major
mantle has led to the widespread assumption that the structures in the eastern and northern parts of the
system of troughs that are broadly parallel to the Banda arc and what is their role in the development
Benioff zone contours, and outside the outer arc of the collision zone? What are the implications for
islands, are subduction trenches. The troughs are hydrocarbon exploration? Detailed multibeam
commonly shown as a continuous feature, but the bathymetric maps of the seabed, and shuttle radar
position of the supposed subduction zone, notably at (SRTM) and ASTER imagery of land, display details
the eastern end of the arc, is commonly drawn in of structure offshore and onshore in unprecedented
different positions. Some authors connect the detail and allow structural and tectonic interpretation
Tanimbar trough to the Seram trough through the Kai that was previously impossible. We use 2D seismic
Islands (e.g. Honthaas et al., 1999; Milsom, 2001; and multibeam data acquired in the last 20 years
Charlton, 2002; Hinschberger et al., 2003) whereas (Pairault et al., 2003; Teas et al., 2009; Adhitama et
others trace a subduction zone from the Tanimbar al., 2017; Patria and Hall, 2017), integrated with
trough via the Aru trough to the Seram trough (e.g. recent land studies and new dating (Pownall et al.,
Hamilton, 1979; O’Sullivan et al., 1985; Silver et al., 2013, 2014, 2016, 2017; Pownall and Hall, 2014),
1985; Harris, 2011). At its western end the Seram plus insights from regional seismicity and P-wave
trough is often mapped as far as Buru. tomography (Spakman and Hall, 2010; Hall and
Spakman, 2015) to suggest answers to these
Other interpretations have been made. Linthout et al., questions.
(1991, 1997) advocated an important role for sinistral
strike-slip faulting in the Seram region, a suggestion PRESENT TECTONIC SETTING
later supported by Teas et al. (2009). Audley-Charles
et al. (1979) proposed the Seram trough was an Figure 2 shows a simplified map of the bathymetry
expression of intra-continental subduction. Audley- of the Banda arc which highlights a significant
Charles (1986) identified the Timor–Tanimbar feature of the troughs: how shallow they are. Most
trough as a foreland basin and Pairault et al. (2003) parts of the troughs are less than 3 km deep and the
made a similar suggestion for the Seram trough; the central Seram trough and Tanimbar trough are less
troughs being interpreted as a flexural response to than 2 km deep. They are shallower than almost all
thrust-loading in the outer arc islands. Spakman and subduction trenches. There are three deep areas. The
Hall (2010) accepted the trough is partly a foredeep Aru trough reaches depths of 3–4 km and is an
produced by loading, rather than a subduction trench, extensional basin within the Australian continental
but proposed it is also the limit of delamination of margin (Charlton et al., 1991; Adhitama et al., 2017).
deep lithosphere from the overlying continental At the western end of the Seram trough, north of west
crust. Pownall et al. (2013) linked delamination to Seram and Buru, the depth of the Buru basin exceeds
Neogene extension associated with subduction 5 km; this is a small rifted basin, probably formed at
rollback. Additional complexities in the the time of oceanic spreading in the North Banda Sea
interpretation of the arcuate system of troughs between c. 12 and 7 Ma (Hinschberger et al., 2003),
and likely to contain oceanic crust at its centre (Patria broadly E-W fault trends in the west changing to
and Hall, 2017). The Weber deep is a young broadly N-S trends in the east, parallel to the trace of
extensional basin in a back arc setting (Hinschberger the Seram trough, and there are relatively common
et al., 2003) interpreted to have formed in the most strike-slip fault solutions (predominantly sinistral E-
recent phase of subduction rollback which may still W trends). Normal fault solutions are uncommon.
be active (Pownall et al., 2016).
DATA AND METHODOLOGY
GPS observations (Stevens et al., 2002; Bock et al.,
2003) show that stations in the eastern Banda arc A multibeam bathymetry dataset with a cell size of
(Ambon, Banda Api, Kai Besar) are moving rapidly, 25 m was provided by TGS. Acquisition information
approximately ENE-wards, towards the Bird’s Head is given in Teas et al. (2009). This was merged with
(Figure 3) and have been used to define a Bird’s another multibeam bathymetry dataset with a cell
Head block subducting SW-wards beneath the size of 15 m provided by GeoData Ventures Pte.
northern Banda arc at the Seram trough (Bock et al., (Figure 5) and processed with ER Mapper and Global
2003). However, some caution is required in Mapper to remove small voids. The multibeam data
interpretation of the GPS measurements, even if it is were used with 2D time-migrated seismic datasets
accepted that they record movements of lithospheric provided by TGS for two projects on the Seram
blocks, rather than merely upper crust, as implied by trough (Patria and Hall, 2017) and the Aru trough
the inference of subduction. There are few stations (Adhitama et al., 2017). Figures in these two
and measurements have been made over only about accompanying papers complement the descriptions
20 years. Those from the western Bird’s Head are and illustrations in this paper. The interpretations in
consistent with a single block, but those from stations this paper are based on the results of those projects,
on Biak and Yapen suggest they are not part of this observations and interpretations from Seram
block. Furthermore, the supposed subduction trace (Pownall, 2015; Pownall et al., 2013, 2014, 2016,
drawn by Bock et al. (2003) is continued west, well 2017), other 2D seismic data, and influenced by the
beyond the Seram trough into the Buru basin, and published GPS measurements and regional
located in the wrong place in the Kai Islands–Aru seismicity discussed above. The high resolution
trough region where the station on Kai Besar should multibeam data help to distinguish between tectonic
be east, not west, of this trace (Adhitama et al., 2017; interpretations, which was not previously possible
Patria and Hall, 2017). Finally, the convergence of using lower resolution data such as global satellite
Ambon and Banda Api with a Bird’s Head block gravity-derived bathymetry or 2D seismic grids.
does not require subduction at the Seram trough, but Figure 6 summarises the major structures interpreted
simply implies contraction in the region between in the northern and eastern Banda arc which are
Ambon–Banda and the southern Bird’s Head, an area discussed below.
which includes the whole of Seram.
THE SERAM TROUGH
Earthquakes provide information on where
deformation is occurring at the present-day. Figure Patria and Hall (2017) conclude that the Seram
4A plots earthquake hypocentres from Engdahl et al. trough is not a subduction trench but a linear
(1998). There is almost no seismicity near the trough, bathymetric feature which marks the deformation
quite unlike a subduction zone, and almost all the front of a fold and thrust belt resulting from young
earthquakes between Ambon–Banda and the Bird’s oblique convergence between the Outer Banda arc
Head are beneath or close to Seram Island. The and the Bird’s Head. The trough terminates as a well-
majority are very shallow (< 50 km) and with one defined feature in the west at about 128°30’E where
exception all are < 100 km deep. There is also a there is a complex deeper zone on the sea floor before
significant diminution of seismicity between the passing west into the Buru basin. The Buru basin is
Weber deep and the Kai Islands. The CMT catalog much deeper than the Seram trough, reaching 5300
of global seismic moment tensors (Ekström et al., m, and is not a continuation of the trough as often
2012) shows the types of faults in the region interpreted (e.g. Hamilton, 1979; Silver et al., 1985;
subdivided into normal (Figure 4B), strike-slip Honthaas et al., 1999; Pairault et al., 2003). It has a
(Figure 4C) and thrusts (Figure 4D). The Aru trough relatively simple structure with E-W-trending
is the site of active E-W extension with strike-slip normal faults downthrown towards the basin axis
faults (predominantly sinistral E-W trends) at its where there is likely to be oceanic crust. We suggest
northern end. There are a few strike-slip solutions the basin formed by rifting contemporaneous with
between the Weber deep and the Kai Islands. In oceanic spreading in the North Banda Sea between
contrast, Seram is dominated by thrust faulting, with 12 and 7 Ma (Hinschberger et al., 2003).
The Seram trough differs from subduction trenches the Kai Arch (Adhitama et al., 2017; Patria and Hall,
in its shallow depth, the absence of seismicity at, and 2017) and between the islands of Kai Besar and Kai
outboard (north) of, the trough, and the character of Kecil (Figure 8) as previously suggested by several
seismicity inboard (south) of the trough. There is a authors (Charlton et al., 1991; Milsom et al., 1996).
fold and thrust belt (FTB) between Seram and the East of the Kai Arch there are NNE-SSW-trending
trough. We deliberately avoid the term accretionary normal faults which border the Aru trough.
complex for the FTB because of its connotations of Significant extension began in the Late Miocene and
subduction; similar structures in subduction and non- major subsidence occurred between the Pliocene and
subduction settings are simply the result of present day.
contraction and in our view, at least in part, the result
of gravity-driven movements. The FTB south of the The NNE-SSW normal faults can be traced
trough is narrow in the west (40 km from north to southwards where they converge. The eastern faults
south) and 2D seismic lines suggest thrusting close form the east boundary of the Tanimbar trough and
to the trough is balanced by extension close to the the western boundary fault can be traced parallel to
Seram coast (Patria and Hall, 2017, their Figure 5D) the east coast of Kai Besar (Figure 8). At its northern
suggesting gravity-driven deformation. The FTB end the Kai Arch is bounded by the narrow N-S
widens to the east where it also changes direction Seram trough on its western side, where the Seram
from E-W-trending to N-S-trending, which we FTB thrusts eastwards, and a steep east-dipping
suggest results from the former shape of the Banda normal fault on its eastern side (Adhitama et al.,
embayment margin. In the east the FTB is thrust onto 2017; Patria and Hall, 2017). Three seismic lines
the folded Misool–Onin–Kumawa ridge (MOKR) (Figures. 8 and 9) illustrate the link between
for which Fraser et al. (1993) used the acronyms thrusting and extension. Line B (Figure 9B) shows
KOMA and MOKA. This is a large open the Seram FTB front is a west-dipping thrust above
anticlinorium which has a broadly E-W axial trend an east-dipping normal fault on the west side of the
with a number of dog-leg bends mapped offshore Kai Arch and the multibeam map confirms this can
using 2D seismic lines by tracing erosional be traced southwards from the Seram trough. Further
culminations from south of Misool towards the Onin south the NNE-SSW east-dipping normal fault is
peninsula (Pairault et al., 2003). From there it can be buried beneath a large mass transport complex
mapped on land following the axial trace of folds (MTC) about 100 km in length and 40 km wide
visible on SRTM images to the Kumawa peninsula (Figure 8) which has moved downslope into deeper
(Figure 6). The ridge formed in the Late Miocene– water towards the NE. It is notable that the northern
Early Pliocene and emergence due to folding margin (Figure 9A) and the eastern margin (Figure
produced the Early Pliocene unconformity identified 9C) of the MTC are thrust zones traceable from the
by Pairault et al. (2003); thrusting of the FTB post- front of the MTC to depths of 2 to 3 seconds TWT
dates the unconformity. Between 129°30’E and below it. In the shallower areas of the Seram FTB
130°30’E (Figure 7) an extensive carbonate platform north and south of the Kai Islands there is abundant
overlies the Early Pliocene unconformity and must evidence for mud volcanism on the sea bed (Figure
have formed close to sea level. After formation of the 8) as also reported on land (Charlton et al., 1991).
platform the carbonates were tilted and subsided to
depths of between 450 m and 1400 m, and now dip Interpretations of 2D seismic lines alone would
towards the trough which is at about 2000 m depth commonly conclude that the thrust front on the west
(Figure 7). This indicates the present depth of the side of the Tanimbar trough is a subduction zone but
trough has been acquired since the Early Pliocene the multibeam imagery shows that the Seram FTB
which we interpret to be due to loading by the Seram has collapsed downslope as an MTC, moving
FTB. Carbonates capping the Kai Arch are now at towards the NE, into the pre-existing deep water
similar depths above the probable Early Pliocene Tanimbar–Aru trough. Thus, the Tanimbar trough is
unconformity (Adhitama et al., 2017) immediately not a subduction zone but a southwards-narrowing
east of the Seram FTB thrust front. extensional trough formed in the Late Miocene into
which an overpressured MTC has moved. The thrust
ARU AND TANIMBAR TROUGHS front at the northern end of the MTC (Figure 9A) is
almost identical to other parts of the thrust front
The multibeam data resolve the issue of the position along the Seram and Tanimbar troughs (Figures 9B
of the Seram trough and its link to the Tanimbar and 9C) despite being formed at the MTC front
trough. The thrust front of the Seram FTB can be orthogonal to the Banda arc–Australian margin
clearly traced from north to south on the west side of convergence direction.
At the northern end of the Kai Arch is an almost flat shear zone. The shear zone, igneous and
carbonate platform at approximately 1 km depth metamorphic rocks record multiple phases of
(Adhitama et al., 2017). This rests directly on the exhumation attributed to extension associated with
New Guinea Limestone above an erosional subduction rollback into the Banda embayment
unconformity interpreted as Early Pliocene in age (Spakman and Hall, 2010). Ages recorded on Seram
(Pairault et al., 2003). We suggest the carbonates are and interpreted from the Aru trough (Adhitama et al.,
of similar Late Pliocene to Pleistocene age to the 2017) indicate that the shear zone was active in the
carbonate platform tilted into the Seram trough Late Miocene and Pliocene.
further west (Patria and Hall, 2017) and subsidence
of the Kai Arch is attributed to loading by east- The reason why this fault zone was so important in
directed thrusting of the Seram FTB. Young rapid the Neogene is unclear. However, we note the
subsidence would have caused NE-directed orientation of the shear zone and the faults bounding
movement of the MTC and SW-directed movement the Aru and Tanimbar troughs to be parallel to the
of other MTCs from the Lengguru shelf into the Aru inferred shape of the Banda embayment. This
trough in the NE part of the Aru trough. suggests that faults were active in the Jurassic
(Figure 11) and were associated with rifting of the
Banda embayment blocks, and these have been
SERAM–KUMAWA SHEAR ZONE
reactivated during Neogene subduction rollback.

A NW-SE to WNW-ESE-trending fault system TARERA–AIDUNA FAULT ZONE

forms the northern border of the Kumawa basin
(Adhitama et al., 2017) and a fault zone with similar The Tarera–Aiduna strike-slip fault zone is usually
orientation offsets the Seram trough by about 80 km represented as a single fault trending E-W south of
at the north end of the Kai Arch (Figure 6). We name the Bird’s Head from the western end of the Central
these the Kumawa fault. The fault zone appears to be Ranges. Hamilton (1979) and Katili (1986) showed
inactive and is covered in part by the MTCs from the the fault zone has only minor expression on land,
Lengguru shelf and by the carbonate platform at the confirmed by SRTM and ASTER images, which is
north end of the Kai Arch. There is a major change much less than expected considering its commonly
across the Kumawa fault. To the north there was interpreted importance (e.g. Hinschberger et al.
significant contractional deformation during the Late (2005) show the fault as active for at least 9 Ma with
Miocene to Early Pliocene resulting in formation of several hundred kilometers of sinistral displacement
the Misool–Onin–Kumawa Ridge and the Lengguru in their Banda reconstruction). Teas et al. (2009)
fold belt. To the south there is no sign of similar mapped the southern strand of the fault zone offshore
significant contraction and all major structures, such using multibeam data on the Lengguru shelf and
as the Aru trough, are extensional. showed its sinistral character. However, they
appeared to consider it was connected to the WNW-
trending segment offset of the Seram trough (Figure
We suggest this is a structure of major importance 6) which we interpret as part of the Kumawa fault.
and we interpret it to link to the major strike-slip fault We suggest the Kumawa and Tarera–Aiduna are
zone crossing Seram (Figure 10), forming part of the different fault zones and were active at different
Kawa shear zone and including the Kobipoto times.
Mountains pop-up (Pownall et al., 2013, 2014). On
Seram this is a deep lithosphere-scale structure. In The Kumawa fault was active in the Late Miocene
the Kobipoto Mountains ultrahigh-temperature and Pliocene and as explained above links to the
(UHT) granulites record a major thermal event at 16 Kawa shear zone of Seram. Mapping by Hamilton
Ma (Pownall et al., 2014; Pownall, 2015). (1979) and Katili (1986) suggests the Tarera–Aiduna
Immediately west of the shear zone Kobipoto fault zone breaks into multiple splays typical of the
Complex rocks were exhumed beneath extensional end of fault zones. The multibeam bathymetry which
detachment faults on the Kaibobo Peninsula of includes a southern strand of the fault zone supports
western Seram at 5.7 Ma and there was deformation the interpreted sinistral strike-slip displacement and
within the Kawa shear zone at 4.5 Ma (Pownall et al., suggests a locally transtensional and transpressional
2017). Extension has exhumed mantle peridotites, character but with little offset. We suggest this fault
cordierite granites formed by melting at peridotite is very young and dies out east of the Seram trough,
contacts, and deep crustal metamorphic rocks. We although judging from seismicity (Figure 4C), it may
name the entire fault system from offshore west now be taking over from the Kumawa fault as an
Seram to the eastern Aru trough the Seram–Kumawa important fault bounding a Bird’s Head block.
THE WEBER DEEP resolve them using global satellite gravity-derived
bathymetry or with widely spaced 2D seismic grids.
The Weber deep is a spectacular 7 km deep feature The multibeam and 2D seismic data together provide
within the forearc of the eastern Banda arc and is the the basis for a much-improved interpretation of the
deepest point of the global ocean that is not a trench. structures and tectonics of the northern and eastern
Hinschberger et al. (2003) suggested it formed Banda arc. These show that there a number of
between 3.5 and 1 Ma. Pownall et al. (2016) different bathymetric features (Figure 6).
proposed that it was even younger and was formed
by forearc extension driven by eastward subduction The Buru basin is a narrow rift basin which probably
rollback since 2 Ma based on reconstructions by contains oceanic crust at its centre and formed
Spakman and Hall (2010). Lithospheric extension in between 12 and 7 Ma. It is situated at the western end
the upper plate was accommodated by a major low- of the Seram trough. The boundary between the Buru
angle normal fault system termed the Banda basin and the Seram trough is probably a sinistral
detachment. The multibeam bathymetry data show WNW-ESE fault zone that continues the Seram–
that the Banda detachment is exposed underwater Kumawa shear zone offshore but has recently been
over much of its 120 km down-dip and 450 km lateral intersected by strands of the developing Sorong fault
extent and that grooves on the submarine fault zone producing major complexity on the sea bed.
surface indicate young extension in an NW-SE This area needs more study and seismic data.
direction (Pownall et al., 2016). The detachment is
exposed in east Seram and on small islands SE of The Seram trough is a much younger aseismic
Seram (Figure 6). flexural feature formed in response to loading by the
Seram FTB. It can be traced eastwards as a narrow
There is no seismicity to indicate that the detachment linear but relatively shallow feature and turns from a
is active but this could reflect a long recurrence broadly E-W trend to a N-S trend. It then meets the
interval on a major fault, unlikely to be recorded by Kumawa fault where it is offset eastwards before
a global earthquake database that documents turning N-S again and following the western side of
earthquakes only from the late 1960s and in a remote the Kai Arch. It is significant that the Seram trough
region with a small population and little additional is not parallel to the more complex trace of the older
historical record. Low angle normal faults are MOKR (Figure 6) as observed by Pairault et al.
apparently rarely seismically active (Axen, 2007) but (2003). The change in shape of the Seram trough
their observed low frequency could indicate that probably mimics faults that formed the Banda
earthquakes on them are rare but large (Wernicke, embayment margin in the Jurassic.
1995). Liu and Harris (2014) suggested a tsunami in
the Banda Islands in 1629 may have been triggered The Seram FTB is a wide zone between southern
by a seismic event in the Seram trough. This Seram and the Seram trough in which there is both
conclusion was based partly on the assumption that folding and thrusting and strike-slip faulting. These
the Seram trough marks a subduction zone associated appear to be partly the result of partitioning of the
with a major lithospheric fault. As discussed above, convergence between the Bird’s Head and Ambon–
both these assumptions are likely to be wrong and the Banda into contraction and lateral displacement, and
present-day trough itself is marked by almost no are partly influenced from basement structures. The
seismicity. However, the Banda detachment is an FTB widens considerably where the trough turns to
obvious candidate for a major earthquake and is also a more N-S direction. However, our observations on
closer to the Banda Islands than the Seram trough. land in small islands SE of Seram suggest more
Large slumps observed at the foot of the eastern wall complexity which is not yet understood. Outcrops of
of the Weber deep, potentially triggered by an coarse breccio-conglomerates and coarse sandstones
earthquake, would be capable of generating a appear to be young alluvial fan deposits, implying a
tsunami, like submarine slumps suggested to be locally significant hinterland of metamorphic rocks
implicated in the 1998 New Guinea tsunami and possibly peridotites in the offshore FTB close to
(Synolakis et al., 2002; Okal and Synolakis, 2003) or the trace of the Kumawa fault.
the 2006 Java tsunami (Fritz et al., 2007).
On the eastern side of the Kai Arch is the Aru trough
CONCLUSIONS which is an extensional depression in which major
extension began in the Late Miocene. The amount
Part of the debate over location and significance of and rate of subsidence and extension increased with
bathymetric features has been hampered by the time. It is probable that initial trough extension was
resolution of available data – it is impossible to due to pull forces of the subducting Banda slab to the
west. These would have increased as the subduction Mesozoic faults active during formation of the Banda
hinge approached the eastern embayment margin as embayment.
the slab rolled back. In the final stages of rollback the
developing Seram FTB loaded the margin causing The implications for hydrocarbon exploration can be
subsidence that caused subsidence of Pleistocene assessed in a crude way from Figure 11. All the areas
carbonate platform on the Kai Arch. shaded in pink are underlain by continental crust
which formed part of Australia in the Early Miocene.
The Tanimbar trough originated as a SSW-narrowing Areas shown in light pink were highly stretched by
extensional zone linked to the Aru trough. This area the Early Pliocene and they were further extended
also would have been loaded by the developing since then, as were the areas south of the Seram–
Seram FTB which was thrust eastwards onto the Kumawa shear zone as far as Tanimbar. These areas
normal faulted western margin of the trough. of highly extended crust, parts of which are now part
However, in this area the FTB collapsed into the of the overthrust collision zone from the Kai Islands
trough and flowed NE-wards along the trough axis as to Timor, are unlikely to be of exploration interest.
one or more mass transport complexes. Recent On the other hand, areas often labelled accretionary,
thrusting at the MTC margins is the result of lateral such as those close to and on the north side of the
spreading of an over-pressured sequence into the Seram–Kumawa shear zone should be underlain by
trough. MTCs were also transported into the Aru Australian continental crust in fold and thrust belts
trough from the Lengguru shelf in the NE related to but are not melanges or exotic material.
young and rapid subsidence of the trough. The MTCs
are 2–3 seconds TWT thick and their boundaries ACKNOWLEDGEMENTS
resemble thrust fronts produced in other tectonic
settings. We thank TGS and GeoData Ventures Pte Ltd. for
providing datasets for our work. We thank the SE
The Weber deep is a major young extensional feature Asia Research Group and its members at Royal
with a possibly active detachment which has a long Holloway, supported over many years by a changing
recurrence interval between displacements. MTCs consortium of oil companies, for discussion and
down the eastern flank of the basin are also a assistance. AP was supported by an Indonesia
potential mechanism to have generated tsunami, if Endowment Fund for Education (LPDP) scholarship
transport was sudden and rapid. during his MSc at Royal Holloway, RA was
supported by the SEARG for his MSc at Royal
Despite some resemblances, none of the troughs are Holloway, and JMP was supported by ARC
subduction zones. Many of the contractional features fellowship DE160100128. We are especially grateful
are very young (Quaternary) and gravity-driven. to John Decker, Phil Teas, Dan Orange and
There is no support for the many models which colleagues, formerly of Niko Indonesia, for their
suggest the troughs developed during subduction support and discussion.
from the Late Miocene (e.g. O’Sullivan et al. 1985;
Hill, 2005; Bailly et al., 2009; Sapin et al., 2009). REFERENCES
They represent different types of response to
subduction rollback into the Banda embayment. Adhitama, R., Hall, R. & White, L.T. 2017.
Extension in the Kumawa block, West Papua,
The Seram–Kumawa shear zone is a newly Indonesia. Proceedings, 41st Annual Convention
recognised major structure within the Australian Indonesian Petroleum Association.
continental margin of Banda embayment. This is a
structure that has a long Neogene history. It was Audley-Charles, M.G. 1986. Timor–Tanimbar
more important and longer-lived than the Seram Trough: the foreland basin to the evolving Banda
trough, and is clearly a lithosphere-scale structure. It orogen. In: Allen, P.A. & Homewood, P. (Eds.),
too was active during subduction rollback and major Foreland Basins. International Association of
extension. Figure 11 shows a cartoon summary of the Sedimentologists, Special Publication 8, 91–102.
major structures during their development in the
Early Pliocene. More work is required on the fault Audley-Charles, M.G., Carter, D.J., Barber, A.J.,
zones. They are not single faults. This study suggests Norvick, M.S. & Tjokrosapoetro, S. 1979. Re-
that the location of strike-slip faulting has shifted interpretation of the geology of Seram: implications
within wide zones, perhaps 40–50 km wide, at for the Banda Arcs and northern Australia. Journal of
different times. Their location may be inherited from the Geological Society of London 136, 547–568.
Axen, G.J. 2007. Research Focus: Significance of Fraser, T.H., Bon, J. & Samuel, L. 1993. A new
large-displacement, low-angle normal faults. dynamic Mesozoic stratigraphy for the West Irian
Geology 35, 287–288. micro-continent Indonesia and its implications.
Proceedings, 22nd Annual Convention Indonesian
Bailly, V., Pubellier, M., Ringenbach, J.-C., de Petroleum Association, 707–761.
Sigoyer, J. & Sapin, F. 2009. Jumps of deformation
zones in a young convergent setting; the Lengguru Fritz, H.M. et al. 2007. Extreme runup from the 17
fold-and-thrust belt, New Guinea Island. Lithos 113, July 2006 Java tsunami. Geophysical Research
306–317. Letters 34, doi:10.1029/2007GL029404.

Bock, Y., Prawirodirdjo, L., Genrich, J.F., Stevens, Hall, R. 2002. Cenozoic geological and plate tectonic
C.W., McCaffrey, R., Subarya, C., Puntodewo, evolution of SE Asia and the SW Pacific: computer-
S.S.O. & Calais, E. 2003. Crustal motion in based reconstructions, model and animations.
Indonesia from Global Positioning System Journal of Asian Earth Sciences 20, 353–434.
measurements. Journal of Geophysical Research
108, 2367. Hall, R. 2012. Late Jurassic–Cenozoic
reconstructions of the Indonesian region and the
Bowin, C., Purdy, G.M., Johnston, C., Shor, G., Indian Ocean. Tectonophysics 570–571, 1–41.
Lawver, L., Hartono, H.M.S. & Jezek, P. 1980. Arc-
continent collision in the Banda Sea region. Hall, R. & Spakman, W. 2015. Mantle structure and
American Association of Petroleum Geologists tectonic history of SE Asia. Tectonophysics 658, 14–
Bulletin 64, 868–918. 45.

Cardwell, R.K. & Isacks, B.L. 1978. Geometry of the Hamilton, W. 1974. Earthquake map of the
subducted lithosphere beneath the Banda Sea in Indonesian Region. U.S. Geological Survey,
eastern Indonesia from seismicity and fault plane Miscellaneous Investigations Series Map I-875-C.
solutions. Journal of Geophysical Research 83,
2825–2838. Hamilton, W. 1979. Tectonics of the Indonesian
region, U.S. Geological Survey Professional Paper,
Charlton, T.R. 2002. The structural setting and 1078, 345 pp.
tectonic significance of the Lolotai, Laclubar and
Aileu metamorphic massifs, East Timor. Journal of Harris, R. 2011. The nature of the Banda Arc–
Asian Earth Sciences 20, 851–865. continent collision in the Timor region. In: Brown,
D. & Ryan, P.D. (Eds.), Arc-Continent Collision.
Charlton, T.R., Kaye, S.J., Samodra, H. & Sardjono Frontiers in Earth Sciences, Springer-Verlag Berlin,
1991. Geology of the Kai islands: implications for Heidelberg 163–211.
the evolution of the Aru Trough and the Weber
Basin, Banda Arc, Indonesia. Marine and Petroleum Hill, K.C. 2005. Tectonics and regional structure of
Geology 8, 63–69. Seram and the Banda Arc. Proceedings, 30th Annual
Convention Indonesian Petroleum Association, 559–
578.
Das, S. 2004. Seismicity gaps and the shape of the
seismic zone in the Banda Sea region from relocated
Hinschberger, F., Malod, J.A., Rehault, J.P. &
hypocentres. Journal of Geophysical Research 109,
Burhanuddin, S. 2003. Apport de la bathymétrie et
B12303, doi:10.1029/2004JB003192.
de la géomorphologie à la géodynamique des mers
de l'Est-indonésian. Bulletin de la Société géologique
Ekström, G., Nettles, M. & Dziewonski, A. M. 2012. de France 174, 545–560.
The global CMT project 2004-2010: Centroid-
moment tensors for 13,017 earthquakes. Physics of Hinschberger, F., Malod, J.-A., Rehault, J.-P.,
the Earth and Planetary Interiors 200–201, 1–9. Villeneuve, M., Royer, J.-Y. & Burhanuddin, S.
2005. Late Cenozoic geodynamic evolution of
Engdahl, E.R., van der Hilst, R. & Buland, R. 1998. eastern Indonesia. Tectonophysics 404, 91–118.
Global teleseismic earthquake relocation with
improved travel times and procedures for depth Honthaas, C., Maury, R.C., Priadi, B., Bellon, H. &
determination. Bulletin of the Seismological Society Cotten, J. 1999. The Plio-Quaternary Ambon arc,
of America 88, 722–743. eastern Indonesia. Tectonophysics 301, 261–281.
Katili, J.A. 1986. Geology and hydrocarbon potential Pownall, J.M. 2015. UHT metamorphism on Seram,
of the Arafura Sea. In: Halbouty, M.T. (Ed.), Future eastern Indonesia: reaction microstructures and P–T
Petroleum Provinces of the World. American evolution of spinel-bearing garnet–sillimanite
Association of Petroleum Geologists Memoir 40, granulites from the Kobipoto Complex. Journal of
487–502. Metamorphic Geology 33, 909–935.

Katili, J.A. 1989. Review of past and present Pownall, J.M. & Hall, R. 2014. Neogene extension
geotectonic concepts of eastern Indonesia. on Seram: A new tectonic model for the northern
Netherlands Journal of Sea Research 24, 103–129. Banda Arc. Proceedings, 38th Annual Convention
Indonesian Petroleum Association, IPA14-G-305 1–
Linthout, K., Helmers, H. & Andriessen, P.A.M. 17.
1991. Dextral strike-slip in Central Seram and 3–4.5
Ma Rb/Sr ages in pre-Triassic metamorphics related Pownall, J., Hall, R. & Watkinson, I. 2013. Extreme
to Early Pliocene counterclockwise rotation of the extension across Seram and Ambon, eastern
Buru-Seram microplate (E. Indonesia). Journal of Indonesia: Evidence for Banda slab rollback. Solid
Southeast Asian Earth Sciences 6, 335–342. Earth 4, 277–314.

Linthout, K., Helmers, H. & Sopaheluwakan, J. Pownall, J.M., Hall, R., Armstrong, R.A. & Forster,
1997. Late Miocene obduction and microplate M.A. 2014. Earth’s youngest known ultrahigh-
migration around the southern Banda Sea and the temperature granulites discovered on Seram, eastern
closure of the Indonesian Seaway. Tectonophysics Indonesia. Geology 42, 279–282.
281, 17–30.
Pownall, J.M., Hall, R. & Lister, G.S. 2016. Rolling
Liu, Z.Y.-C. & Harris, R.A. 2014. Discovery of open Earth's deepest forearc basin. Geology 44, 947–
possible mega-thrust earthquake along the Seram 950.
Trough from records of 1629 tsunami in eastern
Indonesian region. Natural Hazards 72, 1311–1328. Pownall, J.M., Forster, M.A., Hall, R. & Watkinson,
I.M. 2017. Tectonometamorphic evolution of Seram
and Ambon, eastern Indonesia: Insights from
McCaffrey, R. 1989. Seismological constraints and
40Ar/39Ar geochronology. Gondwana Research 44,
speculations on Banda Arc tectonics. Netherlands
35–53.
Journal of Sea Research 24, 141–152.
Sapin, F., Pubellier, M., Ringenbach, J.-C. & Bailly,
Milsom, J. 2001. Subduction in eastern Indonesia:
V. 2009. Alternating thin versus thick-skinned
how many slabs? Tectonophysics 338, 167–178.
decollements, example in a fast tectonic setting: The
Misool-Onin-Kumawa Ridge (West Papua). Journal
Okal, E.A. & Synolakis, C.E. 2003. A theoretical of Structural Geology 31, 444–459.
comparison of tsunamis from dislocations and
landslides. Pure and Applied Geophysics 160, 2177– Schlüter, H.U. & Fritsch 1985. Geology and
2188. tectonics of Banda between Tanimbar Island and Aru
Island, Indonesia. In: Results of R.V.S. Cruise SO-
O'Sullivan, T.D., Pegum, D. & Tarigan, J. 1985. 16. Geologisches Jahrbuch, Reihe E 30, 3–41.
Seram oil search, past discoveries and future oil
potential. Proceedings, 14th Annual Convention Silver, E.A., Gill, J.B., Schwartz, D., Prasetyo, H. &
Indonesian Petroleum Association, 3–20. Duncan, R.A. 1985. Evidence of submerged and
displaced continental borderland, north Banda Sea,
Pairault, A.A., Hall, R. & Elders, C.F. 2003. Indonesia. Geology 13, 687–691.
Structural styles and tectonic evolution of the Seram
Trough, Indonesia. Marine and Petroleum Geology Spakman, W. & Hall, R. 2010. Surface deformation
20, 1141–1160. and slab-mantle interaction during Banda Arc
subduction rollback. Nature Geoscience 3, 562–566.
Patria, A. & Hall, R. 2017. The origin and
significance of the Seram trough, Indonesia. Synolakis, C.E., Bardet, J.P., Borrero, J., Davies, H.,
Proceedings, 41st Annual Convention Indonesian Okal, E., Silver, E., Sweet, J. & Tappin, D. 2002.
Petroleum Association. Slump origin of the 1998 Papua New Guinea
tsunami. Proceedings of the Royal Society of London Indonesian Petroleum Association, IPA09-G-091, 1–
A458, 763–789. 18.

Teas, P.A., Decker, J., Orange, D. & Baillie, P.E.

2009. New insight into structure and tectonics of the Wernicke, B. 1995. Low-angle normal faults and
Seram trough from SEASEEPTM high resolution seismicity: A review. Journal of Geophysical
bathymetry. Proceedings, 33rd Annual Convention Research 100, 20159–20174.
Figure 1 - Tectonic elements of eastern Indonesia (modified from Hall, 2012). Seram is one of the islands of
the Outer Banda Arc. The Seram Trough is situated between Seram and Misool. Red box outlines
area shown in Figure 4 to 6. Faults are marked by solid black lines; subduction zones are indicated
by toothed black lines; dashed black lines mark the troughs around the Banda Arc.
Figure 2 - Simplified bathymetry of the Banda arc showing the very shallow depth of the troughs which curve
through almost 180° at the eastern end of the arc. Dashed black lines are Benioff zone contours in
km.

Figure 3 - Estimated surface velocities (red arrows) with error ellipses based on GPS measurements at the
sites marked showing motions relative to the Bird’s Head block outlined in green dashed line
(modified from Bock et al., 2003).
Figure 4 - Earthquakes and focal mechanisms for the Banda region. A. Well constrained hypocentres from
Engdahl et al. (1998). B, C, D. Earthquake types plotted for events in the CMT catalog of global
seismic moment tensors (Ekström et al., 2012). The blue line on each map is the trace of the Seram
trough.
Figure 5 - Merged multibeam bathymetry for the Banda region.
Figure 6 - Map of simplified major structures identified in the Banda region discussed in the text. Those of
different ages are coloured differently, but it should be noted that many of these reactivate older
structures.

Figure 7 - Area of former carbonate platform on southern Misool shelf immediately north of the Seram
trough. The Seram trough in the centre of the map is approximately 2000 m deep. The carbonate
platform is tilted southwards from a depth of approximately 450 m in the north to 1400 m at its
southern edge.
Figure 8 - Detail of sea floor bathymetry showing the southern part of the Seram FTB, the Kai Arch, Aru
trough and Tanimbar trough. Lines marked A, B and C are locations of 2D seismic lines shown in
Figure 9. Green lines mark form lines on MTC that has moved NE-wards into the Tanimbar trough
and outline mud volcanism features in shallower areas. Dashed black line is inferred position of
normal fault that formed the western boundary of the Tanimbar trough before the MTC covered it.
Figure 9 - 2D seismic lines located on Figure 8. A. Northern front of MTC which is approximately 2–3
seconds TWT thick. B. Western edge of Tanimbar trough showing the normal fault downthrown
to the east that forms the eastern boundary of the Kai Arch and western edge of the Aru trough.
The thrust front at the Seram trough further north is thrust eastwards over the normal fault. The
lower area on the east side of the section is the MTC. C. the eastern boundary of the MTC in the
Tanimbar trough thrust over the tilted Australian margin.
Figure 10 - The Kawa shear zone on Seram Island modified from Pownall et al. (2013). A. Major faults
mapped on the island. B. Detail showing slices of mantle, igneous and metamorphic rocks
exhumed in the Kobipoto pop-up, and the metamorphic rocks exhumed in the Kaibobo peninsula
of west Seram between 6 and 4 Ma.
Figure 11 - Reconstruction of the northern and eastern Banda arc in the Early Pliocene, modified from Hall
(2002, 2012), showing the inferred location of the Seram–Kumawa shear zone. The coastlines
give an indication of where various parts of the region were but this should be considered very
approximate because of major extension of the former Sula Spur. This was especially important
in the Banda ridges and the embayment margin forming west and south Seram, and areas close to
the subduction zone labelled as highly extended continental crust.