Content uploaded by Robert Holm
Author content
All content in this area was uploaded by Robert Holm on May 13, 2014
Content may be subject to copyright.
TAJE_A_824505.3d (TAJE) 09-08-2013 11:44
A re-evaluation of arc–continent collision and along-arc
variation in the Bismarck Sea region, Papua New Guinea
R. J. HOLM*AND S. W. RICHARDS
School of Earth and Environmental Sciences, James Cook University, Townsville, Queensland 4811, Australia.
The Bismarck Sea region of Papua New Guinea is marked by recent arc–continent collision giving rise to
a highly dynamic tectonic environment, characterised by complex plate interactions that are yet to be
fully understood. We present a new crustal and upper mantle crustal architecture model for northeast-
ern Papua New Guinea and western New Britain that reveals complex tectonic geometries of overprint-
ing slab subduction and partial continental subduction, resulting in a unique setting in which to
investigate along-arc magmatic variation. Earthquake hypocentre databases are combined with
detailed topography and seafloor structure together with geology and regional-scale gravity to unravel
the sub-surface structure of northeastern Papua New Guinea. These data are used in conjunction with
an updated 3-D slab map of the region to propose a new interpretation of the area whereby Australian
continental crust extends as an underthrust block beneath the accreted Finisterre Terrane. The subduct-
ing continental crust combined with slab stagnation has resulted in a complex pattern of arc-related
geochemical signatures from east to west along the Bismarck arc. In the east, where the Solomon Sea
plate is subducting beneath New Britain, the sedimentary component is low, whereas in the west, the
arc volcanics exhibit a greater sedimentary component, consistent with subduction of Australian crustal
sediments. As a result, a new plate reconstruction is provided for the region together with a forward-
looking reconstruction of the Papuan peninsula, the Solomon Sea plate and New Britain that illustrates
that the same process will likely be repeated in some 5–10 m.y.
KEY WORDS: Papua New Guinea, Bismarck arc, arc–continent collision, gravity, seismicity, arc
geochemistry.
INTRODUCTION
The boundary between the northern Australian plate
and the Pacific plate, which includes the Bismarck Sea
region of Papua New Guinea (Figure 1), comprises some
of the youngest and most active tectonic elements of the
southwest Pacific (e.g. Taylor 1979; Abbott 1995; Martinez
& Taylor 1996; Weiler & Coe 2000). Northward motion of
the Australian plate has led to a scenario where both
continental and oceanic crust is interacting along the
northern plate boundary. The complexities of present-
day crustal and mantle geometries have emerged from
new information and a reinterpretation of the mecha-
nisms leading to the tectonic amalgamation of the area
is required. Here we focus on just the latest 4 Ma or so in
northeastern Papua New Guinea where in this short
time arc–continent collision has consumed tectonic
plates and uplifted mountain ranges to more than
4000 m, neighboured by contemporary island arc mag-
matism, culminating in highly dynamic and striking
geological landscapes. Numerous workers have sought to
explain these processes of, for example, arc volcanism in
the western Bismarck Sea, the source of earthquakes, or
the timing and nature of arc–continent collision; how-
ever, these models lack an overarching geological model
with cross-disciplinary foundations that can account for
all geological phenomena.
We present a compilation and reinterpretation of
an extensive catalogue of previous data, including
topography/bathymetry, earthquake hypocentres,
regional-scale gravity, geology and geochemistry, and
models depicting the complex tectonic history of Papua
New Guinea and the southwest Pacific. From this we re-
evaluate and address gaps in our knowledge of the pres-
ent-day 3-D tectonic setting of northeast Papua New
Guinea and the Bismarck Sea. Using a new and robust
regional tectonic model, we assess the role of recent arc–
continent collision in construction of the present-day
tectonic puzzle that is Papua New Guinea.
The global importance of arc–continent collision has
been addressed by several authors related to the Africa–
Europe collision (e.g. Rosenbaum et al. 2002; Kley & Voigt
2008) or the India–Asia collision (e.g. Hendrix et al. 1994;
Sobel & Dumitru 1997; Najman et al. 2010); however, the
southwest Pacific offers the unique opportunity to
observe the process in action. Furthermore, recognition
of the subtle processes and mechanisms of collision that
are not apparent at the surface, such as crustal under-
thrusting and associated arc magmatism, will contribute
*Corresponding author: rob.holm@my.jcu.edu.au
Ó2013 Geological Society of Australia
Australian Journal of Earth Sciences (2013)
60, 605–619, http://dx.doi.org/10.1080/08120099.2013.824505
Downloaded by [JAMES COOK UNIVERSITY], [R. J. Holm] at 15:30 20 August 2013
TAJE_A_824505.3d (TAJE) 09-08-2013 11:44
Figure 1 Topography, bathymetry and major tectonic elements of the study area. (a) Major tectonic boundaries of Papua New
Guinea and the western Solomon Islands; CP, Caroline plate; MB, Manus Basin; NBP, North Bismarck plate; NBT, New Britain
trench; NGT, New Guinea trench; NST, North Solomon trench; PFTB, Papuan Fold and Thrust Belt; PT, Pocklington trough;
RMF, Ramu-Markham Fault; SBP, South Bismarck plate; SCT, San Cristobal trench; SS, Solomon Sea plate; TT, Trobriand
trough; WB, Woodlark Basin; WMT, West Melanesian trench. Study area is indicated by rectangle labelled Figure 1b; the other
inset rectangle highlights location for subsequent figures. Present day GPS motions of plates are indicated relative to the
Australian plate (from Tregoning et al. 1998,1999; Tregoning 2002; Wallace et al. 2004). (b) Detailed topography, bathymetry
and structural elements significant to the South Bismarck region (terms not in common use are referenced); AFB, Aure Fold
Belt (Davies 2012); AT, Adelbert Terrane (e.g. Wallace et al. 2004); BFZ, Bundi Fault Zone (Abbott 1995); BSSL, Bismarck Sea
Seismic Lineation; CG, Cape Gloucester; FT, Finisterre Terrane; GF, Gogol Fault (Abbott 1995); GP, Gazelle Peninsula; HP,
Huon Peninsula; MB, Manus Basin; NB, New Britain; NI, New Ireland; OSF, Owen Stanley Fault; RMF, Ramu-Markham Fault;
SS, Solomon Sea; WMR, Willaumez-Manus Rise (Johnson et al. 1979); WT, Wonga Thrust (Abbott et al. 1994); minor strike-slip
faults are shown adjacent to Huon Peninsula (Abers & McCaffrey 1994) and in east New Britain, the Gazelle Peninsula
(e.g. Madsen & Lindley 1994). Circles indicate centres of Quaternary volcanism of the Bismarck arc. Filled triangles indicate
active thrusting or subduction, empty triangles indicate extinct or negligible thrusting or subduction.
606 R. J. Holm and S. W. Richards
Downloaded by [JAMES COOK UNIVERSITY], [R. J. Holm] at 15:30 20 August 2013
TAJE_A_824505.3d (TAJE) 09-08-2013 11:44
to our understanding of collision events and terrane
accretion at ancient convergent margins.
TECTONIC SETTING
Papua New Guinea and much of the southwest Pacific
occupy a zone of oblique convergence between the Aus-
tralian and Pacific plates (Figure 1). The tectonic history
of the region is significantly more complex than other
arcs owing to the number of recognised small plates
within the region. This scenario arises from the posi-
tioning of Papua New Guinea within a regional-scale col-
lision zone between the Australian continental crust in
the south (Abbott 1995; Hall 2002; Davies 2012) and the
Ontong Java Plateau in the northeast (Petterson et al.
1999; Hall 2002; Mann & Taira 2004). The relative direc-
tion of plate convergence has resulted in development of
oblique spreading centres and the formation of numer-
ous micro-plates and associated plate boundaries. The
principal tectonic elements comprising this complex
zone are shown on Figure 1, but emphasis is placed on
the Australian plate, the Finisterre Terrane (described
in detail below), New Britain and the North and South
Bismarck plates.
Previous research has suggested that from the Upper
Oligocene to the latest Neogene, northern Papua New
Guinea is marked by a series of arc–continent collisions.
The youngest and most significant of these collisions
resulted in the accretion of the Adelbert and Finisterre
Terranes, the latter of which forms a prominent topo-
graphic high known as the Finisterre Range (Figure 1;
Abbott et al. 1994; Abbott 1995). Abbott et al. (1994) stud-
ied clastic sequences on the southern flanks of the Fini-
sterre Range and concluded that the collision must have
initiated at ca 3.7–3.0 Ma. The Adelbert and Finisterre
Terranes are largely composed of Paleogene through to
earliest Neogene volcanic arc rocks overlain by Miocene
to Plio–Pleistocene limestone (Jaques & Robinson 1977;
Weiler & Coe 2000). Collision of these terranes with
Papua New Guinea is interpreted to have resulted from
the closure of the Solomon Sea at the New Britain trench
owing to subduction-driven convergence between the
Australian and South Bismarck plates (e.g. Abbott 1995;
Hill & Raza 1999; Weiler & Coe 2000). Oblique collision
started in the west and propagated southeastwards, pro-
ducing progressive thrusting and uplift of the north
coast Adelbert and Finisterre Ranges (Johnson & Jaques
1980; Abbott 1995; Weiler & Coe 2000).
At present, the ongoing convergence between the Fin-
isterre Terrane and the Australian plate is accommo-
dated by activity along the Ramu-Markham Thrust Fault
(e.g. Cooper & Taylor 1987; Abbott et al. 1994; Pegler et al.
1995). All previous studies regarding this episode of arc–
continent collision have focused on the Finisterre Ter-
rane, the uplifted and exposed upper plate. The outstand-
ing topography of the Finisterre Range, however, only
arises as the Finisterre Terrane is thrust over the for-
mer northern coastward margin of Papua New Guinea.
This concept, and the nature or expanse of the now
underthrust Papua New Guinea margin has only been
suggested in passing by previous studies, but should be
regarded as an important, although missing piece of
Papua New Guinea. This statement is particularly signif-
icant given the prominence of major suture zones and
structures converging with the Ramu-Markham Fault
and underthrust beneath the Finisterre Terrane, for
example the Owen Stanley Fault and the Aure Fold Belt
(Figure 1).
The recent collision of the Adelbert and Finisterre
Terranes is reflected in the regional tectonics of the Bis-
marck Sea. The inferred timing of plate coupling at
3.7 Ma (Abbott et al. 1994) is coincident with the earliest
breakup and opening of the New Britain back-arc,
whichinturncreatedtwonewmicro-plates,theNorth
and South Bismarck plates (Taylor 1979). The South Bis-
marck plate is currently rotating clockwise at a rate of
8/Ma relative to Australia (Tregoning et al. 1999;
Wei le r & Coe 2000; Wallace et al. 2004) while the west-
northwest motion of the North Bismarck plate is simi-
lar to the Pacific plate (Tregoning et al. 1998;Wallace
et al. 2004) suggesting almost complete coupling
betweentheNorthBismarckandPacificplates
(Figure 1). In the eastern Bismarck Sea, the East Manus
spreading centre separates the North and South Bis-
marck Sea plates (Martinez & Taylor 1996). However, in
the wester n Bismarck Sea, the boundary becomes the
Bismarck Sea seismic lineation, defined primarily by
earthquake epicentre locations and characterised by
left-lateral transform faults and associated step-over
rifts (Denham 1969;Taylor1979). Thus, the Manus Basin
accommodates the majority of extension and rotation
in the eastern part of the Bismarck Sea while in the
west, the Bismarck Sea seismic lineation becomes a dis-
crete, east–west oriented strike-slip plate boundary
(Figure 1; Cooper & Taylor 1987;Llaneset al. 2009).
North-dipping subduction of the Solomon Sea plate
beneath New Britain, in addition to the convergence
responsible for terrane accretion, has resulted in the for-
mation of the active Bismarck volcanic arc. The arc
occupies the northern part of the island of New Britain
and extends to the west where it is present as a series of
volcanic islands off the northwest coast of New Britain
and northeast coast of the Papua New Guinea mainland
where it forms the West Bismarck Arc (Figure 1). The
composition of the volcanics centred on and around the
island of New Britain range from basalt to rhyolite with
typical low-K, island arc tholeiite signatures (Jakes &
Gill 1970). The compositions differ markedly in the West
Bismarck arc with predominantly a medium-K charac-
ter (Woodhead et al. 2010). Along-arc variation, recently
been investigated by Woodhead et al. (2010), is discussed
below.
DATA AND DATA-ANALYSIS TECHNIQUES
Earthquakes
The interpretations and reconstructions presented
below have been resolved using a variety of datasets
combined into a single 3-D tectonic map of the region
using the software GOCAD. Seismic data provide a use-
ful indicator for active tectonic structures such as faults
and subducting slabs (Figure 2). We utilise a combina-
tion of earthquake records including the EHB hypo-
centre catalogue (Engdahl et al. 1998; Engdahl 2006) and
Bismarck Sea collision tectonics 607
Downloaded by [JAMES COOK UNIVERSITY], [R. J. Holm] at 15:30 20 August 2013
TAJE_A_824505.3d (TAJE) 09-08-2013 11:44
Figure 2 Seismicity and major structure of northwest Papua New Guinea. Earthquakes are derived from the NEIC earthquake
database (1990–2010) for the 0–100 km depth bin. (a) Seismicity, structure and geology of the southwest Bismarck Sea–Huon
Peninsula region. Structures and labels follow Figure 1b; see text for discussion of geology. (b) Earthquake density distribu-
tion map for the same region. Earthquake densities are contoured from high density (white) to low density (blue). We note
there are high-density earthquake clusters adjacent to the point where the Bundi Fault Zone and Owen Stanley Fault intersect
and under-thrust the Ramu-Markham Fault.
608 R. J. Holm and S. W. Richards
Downloaded by [JAMES COOK UNIVERSITY], [R. J. Holm] at 15:30 20 August 2013
TAJE_A_824505.3d (TAJE) 09-08-2013 11:44
the USGS National Earthquake Information Center
(NEIC) database for the period between 1973 and 2010
(Figure 2). In addition, Centroid-Moment-Tensor (CMT)
earthquake solutions were derived from the Harvard
Global CMT database (1976–2010) and plotted within
ArcScene in ArcGIS using the USGS 3D Visualisations of
Earthquake Focal Mechanisms extension. The NEIC
earthquake database and CMT database include all
earthquakes with moment magnitude values (Mw)
greater than Mw 4.5; the EHB database utilises earth-
quakes greater than Mw 4.3.
The earthquake hypocentre data were scrutinised
using a variety of software and techniques. Initially, the
data were imported into the 4DEarth model (GOCAD)
where slab surface models were derived. Details of the
slab model are presented below. In order to estimate
earthquake abundance distributions and clustering (Fig-
ure 2), a simple gridding function was used to derive a
map highlighting the number of earthquakes within
grid cells with dimensions of 0.04 0.04latitude and lon-
gitude. This grid cell size allowed the minimum number
of cells to be chosen without biasing towards the genera-
tion of many separate but isolated clusters or points. In
addition, the use of equi-dimensional cell size removed
any directional bias; therefore, the trends observed in
the density distribution maps represent a true cluster
orientation. The results are presented as a series of den-
sity distribution maps which are plotted for depth bins
of 0–40 km, 40–100 km and 100–300 km (depths of 0–20,
20–40, 40–60, 60–80, 80–100, 100–120 and 120–140 km are
contained within the Supplementary Papers).
Seafloor gravity and topography
Seafloor gravity data for the Bismarck Sea (Figure 3) are
sourced from the Australian Bureau of Mineral Resour-
ces (1970). The gravity model provided has been cor-
rected for a uniform ocean thickness.
Topography and bathymetry data derived from the
National Oceanic and Atmospheric Administration
ETOPO1 1-minute global relief model (Amante & Eakins
2009) provide an additional framework for correlation
and interpretation.
Geochemical data
Woodhead et al. (1998,2010) created an extensive geo-
chemical dataset for the New Britain and West Bismarck
arcs based on new and existing geochemical data from
New Britain from Johnson & Chappell (1979) and
Woodhead & Johnson (1993). The majority of these geo-
chemical data were produced for major and trace ele-
ments by X-ray fluorescence (XRF) and limited use of
spark-source mass spectrography (SSMS; Johnson &
Chappell 1979; Woodhead & Johnson 1993; and references
therein). Woodhead & Johnson (1993) and Woodhead
et al. (1998,2010) used inductively coupled plasma mass
spectrometry (ICPMS) and Pb, Sr, Nd and Hf isotope
analyses to develop the best compilation of data for east-
ern Papua New Guinea. Woodhead et al. (2010) investi-
gated along-arc geochemical changes in Bismarck arc,
and we re-evaluate these data in the context of the new
tectonic model presented here.
Figure 3 Seafloor free-air gravity anomaly map for the southwest Bismarck–Huon Peninsula region (Data from Australian
Bureau of Mineral Resources 1970).
Bismarck Sea collision tectonics 609
Downloaded by [JAMES COOK UNIVERSITY], [R. J. Holm] at 15:30 20 August 2013
TAJE_A_824505.3d (TAJE) 09-08-2013 11:44
INTERPRETATION AND RESULTS
Subduction zone and slab architecture
Previous work in the region focused on establishing an
accepted plate boundary model using information such
as seismicity and instantaneous GPS motions (e.g.
Denham 1969; Johnson & Molnar 1972; Ripper 1982;
Abers & Roecker 1991; Pegler et al. 1995; Wallace et al.
2004). It is widely accepted that multiple subduction
zones have existed since the beginning of crustal amal-
gamation of Papua New Guinea, however the details per-
taining to the geometry and type of crust subducting
along the northeastern Papua New Guinea coast and
western New Britain are unresolved, despite its infancy.
Construction of new 3-D subducted slab models, up to
600 km depth in the mantle, build on earlier work by
O’Kane (2008). Earthquake hypocentre data (Engdahl
et al. 1998; Engdahl 2006) are primarily used to generate
the 3-D models of subducted slab with all earthquakes
below 100 km assumed to occur within the subducting
plate (Isacks et al. 1968). The method for constructing
slabs follows that outlined in Richards et al. (2007,2011).
The Global CMT database is examined in 3-D to assist in
interpreting the geometry of the slab. The final inter-
preted slab geometry of the composite Australian plate
(Solomon Sea plate, Woodlark Basin and Australian plate)
subducted at the New Britain and San Cristobal trenches,
and termed the Solomon slab, is presented in Figure 4.
Miller et al. (2006) used a similar method of analysing
slab geometries in conjunction with earthquake failure
solutions beneath the southern Mariana Arc.
Overall, the Solomon slab exhibits a moderate dip
between the surface and 100 km depth; below this depth,
the slab is steeply dipping. West of the New Britain
trench–Trobriand trough triple-junction, the Solomon
slab currently resides at a depth of 100 km, and remains
close to this depth until it terminates in the west beneath
central Papua New Guinea. Furthermore, a north-
dipping slab component is modelled in the west which
extends to 250 km depth below west New Britain and
continues to the west at shallower depths (Figure 4), con-
sistent with findings from Johnson & Molnar (1972), John-
son & Jaques (1980), and Abers & Roecker (1991).A
restricted south-dipping slab component is also imaged
but this is limited to the region adjacent to the Huon Pen-
insula, accounting for observations made by Ripper
(1982), Cooper & Taylor (1987),Pegleret al. (1995),and
Woodhead et al. (2010). The lack of a definitive modern
seismic or tomographic signature for either an extensive
slab at depth or plate interface seismicity at the trench
(Hall 2002; Hall & Spakman 2002) suggests that there is
very little evidence for substantial southward subduction
at the Trobriand trough (Johnson & Molnar 1972; Johnson
&Jaques1980;Abers&Roecker1991) in agreement with
the slab map present here. A small tear in the slab is
interpreted below the eastern margin of the Huon Penin-
sula; this fundamentally separates the wester n slab
domain from the remaining Solomon slab in the east.
Adjacent to east New Britain and New Ireland, the
curvature of the trench and subducted slab, and associ-
ated subduction of an originally flat oceanic crustal
sheet have resulted in the development of a vertical tear
in the slab (Figure 4); in line with findings suggested by
Cooper & Taylor (1989). At present, the tip of the tear ter-
minates beneath southern New Ireland and exhibits a
western and an eastern flank propagating beneath the
Tabar–Lihir–Tanga–Feni arc; the western flank propa-
gated beneath Lihir (where the slab lies some 550 km
below owing to the steep dip). The tear here is significant
because it provides a window where the asthenosphere
can penetrate from the rear of the slab to the front. To
Figure 4 3-D model of the Solomon slab comprising the sub-
ducted Solomon Sea plate, and associated crust of the Wood-
lark Basin and Australian plate subducted at the New
Britain and San Cristobal trenches. Depth is in kilometres;
the top surface of the slab is contoured at 20 km intervals
from the Earth’s surface (black) to termination of slab-
related seismicity at approximately 550 km depth (light
brown). Red line indicates the locations of the Ramu-Mark-
ham Fault (RMF)–New Britain trench (NBT)–San Cristobal
trench (SCT); other major structures are removed for clar-
ity; NB, New Britain; NI, New Ireland; SI, Solomon Islands;
SS, Solomon Sea; TLTF, Tabar–Lihir–Tanga–Feni arc. See
text for details.
610 R. J. Holm and S. W. Richards
Downloaded by [JAMES COOK UNIVERSITY], [R. J. Holm] at 15:30 20 August 2013
TAJE_A_824505.3d (TAJE) 09-08-2013 11:44
the east, the Solomon slab is dipping beneath the Solo-
mon Islands and reaches a maximum interpreted depth
of 500 km. The subducted slab here exhibits less ‘struc-
ture’than the slab to the west; however, research focused
on mapping the subducted extent of the Woodlark Basin
rift is ongoing.
Gravity and seismicity correlation
At the relatively shallow sub-crustal depths (in the order
of 50–100 km), upper crustal features can mask seismic
tomography and earthquake distribution. We instead
utilise regional seafloor gravity free-air anomaly data
obtained during the 1970 Hamme Cruise (Figure 3;
Australian Bureau of Mineral Resources 1970) to help
interpret crustal boundaries. In particular, we focus on
the region adjacent to the Huon Peninsula and north
coast of Papua New Guinea (Figure 1). Figure 5 presents
the gravity data together with seismicity and the
interpreted slab model. A large gravity-low anomaly is
observed trending sub-parallel to the northern coast of
Papua New Guinea defined by negative gravity values
(Figure 3). The gravity low is particularly intense to the
north of the Huon Peninsula (Davies et al. 1987; Honza
et al. 1987). This anomalous gravity-low also corresponds
with the location of intense seismicity beneath the Huon
Peninsula at depths of between 0 and 100 km (Figure 5).
Gravity and seismicity anomalies of the two datasets cor-
relate extremely well suggesting a relationship between
the two, to a depth of up to 100 km. This level of seismic-
ity has been attributed to the presence of Australian lith-
osphere at up to 100 km depth (Pegler et al. 1995;
Woodhead et al. 2010); however, this has only been
explored in 2-D sections adjacent to the eastern Huon
Peninsula without consideration given to the 3-D extent
of the seismicity.
An additional component of the anomalous gravity
low is present to the southeast of our interpreted upper
Figure 5 Gravity and seismicity correlation for the southwest Bismarck–Huon Peninsula region. All earthquakes illustrated
are from the NEIC earthquake database (1990–2010) and projected on the Bismarck Sea seafloor gravity anomaly map and
topography (a, c); and associated earthquake density distribution maps (b, d). (a, b) Gravity and seismic correlation between
40 and 100 km depth; the top surface of the slab map above 100 km depth is shown as a white shaded area (a) and area outlined
in black (b). A large area of anomalously high earthquake density trending east-southeast–west-northwest and outlined by
the red dashed line does not show a relationship with the defined windows for slab-related seismicity (a, b); furthermore, this
anomalous seismic region correlates well with the negative (low; purple) gravity anomaly to the north of- and beneath the
Huon Peninsula. We make note of the two highest density earthquake clusters; the largest in extent of the two lies at the south-
east tip of the Huon Peninsula, this is a region of overlap between both the slab map and anomalous seismicity/gravity corre-
lation implying multiple earthquake sources superimposed; the second is located at the approximate centre of the Finisterre
Terrane and holds the greatest observed density. (c, d) Gravity and seismicity correlation between 100 and 300 km depth; the
top surface of the slab below 100 km to termination is shown according to the previous description; as this surface represents
the top of the slab, where the top surface is above 100 km depth, earthquakes occurring below this depth and within the slab
are not atypical. Seismicity in this depth range correlates well with the outlined slab map and show little relationship with
gravity anomaly trends. A zone of intense, high density seismicity is present north of the Huon Peninsula, this has previously
been referred to as the ‘Finisterre Nest’(Abers & Roecker et al. 1991).
Bismarck Sea collision tectonics 611
Downloaded by [JAMES COOK UNIVERSITY], [R. J. Holm] at 15:30 20 August 2013
TAJE_A_824505.3d (TAJE) 09-08-2013 11:45
mantle crustal anomaly. This is not as seismically active
as the remainder of the anomalous zone and falls within
the normal bounds of slab-related seismicity for the
region. This anomaly is also consistent with elevated
bathymetry at the surface, and can therefore be attrib-
uted to thickened crust between the Ramu-Markham
Fault–New Britain trench and the Wonga Thrust.
Previous interpretations of geochemistry
Woodhead et al. (2010) concluded that the West Bismarck
and New Britain arcs are both ‘typical’subduction-
related volcanic arcs, and although contiguous along
strike, exhibit very different geochemical characteris-
tics. The arc was divided into two parts, the West Bis-
marck arc and New Britain arc with the line separating
the two drawn between the western-most volcanoes of
New Britain (Cape Gloucester, Langila, Aimaga, Tangi
and Gloucester) and the remainder of New Britain in the
east (Woodhead et al. 2010). In the most general terms,
the distinction between New Britain and West Bismarck
arcs equates to a tholeiitic–calc-alkaline transition
(Jakes & Gill 1970; Woodhead et al. 1998,2010), and may
reflect underlying differences in the nature and composi-
tion of the mantle wedge or subducting plate, or the pro-
cesses of mass transfer between the two, or alternatively
is a consequence of collisional processes during
accretion of the Adelbert and Finisterre Terranes
(Woodhead et al. 2010).
Along-arc geochemical trends of elements, ratios, and
isotopic variation utilising a compilation of data from
both Woodhead et al. (1998,2010) are shown in Figure 6.
Further characteristics of this arc will be discussed
below. Woodhead et al. (2010) identified important differ-
ences between the geochemistry of the two arcs. Arc lavas
from the New Britain volcanic front are derived from a
mantle source highly depleted in many incompatible
trace elements (Woodhead et al. 1998), while the least
evolved West Bismarck arc lavas generally have higher
HFSE contents than the New Britain volcanic front.
Extreme element depletion in the New Britain lavas is
typically attributed to prior melt extraction in the back-
arc Manus Basin, however, the same process does not
operate to the same extent on the mantle source of the
West Bismarck lavas (Woodhead et al. 1993,2010). The
Sm/La ratios (Figure 6), which are higher in New
Britain, suggest a depleted mantle source when compared
with the West Bismarck arc lavas. Furthermore, the
decrease in the Sm/La ratio to the west together with a
decrease in the size of volcanic edifices and eruption rate
(Johnson 1977) suggests the degree of mantle melting falls
dramatically from east to west (Woodhead et al. 2010).
In addition, Woodhead et al. (2010) finds Th/La ratios
in the West Bismarck arc are lower than that of bulk
Figure 6 Along-arc geochemical variation in selected major and trace elements, trace element ratios, and isotopic ratios for
the West Bismark and New Britain arcs (data from Woodhead et al. 1998,2010). See text for further discussion.
612 R. J. Holm and S. W. Richards
Downloaded by [JAMES COOK UNIVERSITY], [R. J. Holm] at 15:30 20 August 2013
TAJE_A_824505.3d (TAJE) 09-08-2013 11:45
continental crust and the ‘average arc’(Plank 2005) sug-
gesting that a sedimentary component is apparent in the
West Bismarck arc lavas. Woodhead et al. (2010) noted
that prior to collision with the South Bismarck plate, the
Australian plate likely carried high Th/La sediments
derived from mainland Papua New Guinea. This is sup-
ported by the similar Th/La ratios to average sediments
from the Solomon Sea (Woodhead et al. 1998), which also
contain a substantial volcaniclastic input derived from
the Papua New Guinea Highlands (Crook 1987). More-
over, Pb-isotopic compositions of the West Bismarck
lavas, which contain relatively radiogenic Pb compared
with Manus Basin MORB (Figure 6), suggest a strong
‘crustal’signature is evident, again similar to Pb compo-
sitions found in the Solomon Sea sediments. Hafnium
and Nd isotope ratios show opposite but similar trends
exhibiting the highest ratio values in the New Britain
arc and decreasing in the West Bismarck arc; Woodhead
et al. (2010) interpreted this as a response to arc–conti-
nent collision where increased proximity to a crustal
source dramatically increased the proportion of conti-
nent-derived detritus delivered to the subducting slab.
These geochemical observations are important and dif-
ferentiate the New Britain arc magmas, predominantly
mantle derived but with a very small sedimentary com-
ponent, from the West Bismarck arc where the sedimen-
tary component is interpreted to be much greater.
DISCUSSION
Crustal architecture
The seismological activity or inactivity of major struc-
tures and plate boundaries should be apparent over a
regional-scale, even taking into account the relatively
short geological window of earthquake recording. The
principal cause of seismicity in the Bismarck Sea region
is subduction at the New Britain trench; this has long
been recognised and accepted as the origin for shallow
through to deep earthquakes. On the same regional-
scale, additional earthquake trends related to major
structure and plate boundary activity are those of the
Bismarck Sea seismic lineation and New Guinea trench
in the north and northwest, and the Papuan Fold and
Thrust Belt to the southwest (Figures 1,2). These major
tectonic structures have long been defined and are well
understood. However, shallow to intermediate depth
seismicity of the northeast Papua New Guinea mainland
is characterised by a somewhat chaotic distribution of
earthquakes (Figure 2). While much of the shallow seis-
micity has previously been correlated with upper crustal
structure (Cooper & Taylor 1987; Abers & McCaffrey
1994; Stevens et al. 1998), it is evident that a significant
proportion cannot be clearly related to any recognised
structural control (Figures 2,5).
As presented in this study, the anomalous seismicity
is focused beneath the Finisterre Terrane and immedi-
ate adjacent areas (Figures 2,5). This zone is defined by
an uncharacteristically high abundance of earthquakes
compared with ‘typical background’seismicity
(Figure 5), and defines a zone extending from the surface
to approximately 100 km depth beneath northeast Papua
New Guinea. Further more, this feature correlates with
a negative gravity anomaly that cannot be easily related
to any near surface geological phenomena (Figure 3).
Similar gravity lows observed adjacent to trenches are
commonly interpreted as subducted low-density crust
(Morales et al. 1999; Mishra et al. 2000), or alternatively,
as crustal thickening and stacking of low-density crust
during orogenesis (Stern 1995; Casas et al. 1997). Both
scenarios are typical of convergent margin settings
much like the recent history of the Bismarck Sea region.
We propose the anomalous gravity-low in combination
with anomalous seismicity is the expression of a previ-
ously undefined crustal block underthrust beneath the
Adelbert and Finisterre Terranes during collision with
the Papua New Guinea margin. In addition, this crustal
block is interpreted to be continental crust that is the
underthrust and subducted leading edge of Papua New
Guinea (Figure 7). This is significant in that the nature
of the continental margin has not previously been con-
sidered in the context of the northern Papua New
Guinea accreted terranes and holds implications for the
dynamics of terrane collision processes. The extent of
the underthrust margin is further supported by CMT sol-
utions. These are illustrated in Figure 7 for between 40
and 90 km depth and highlight a regional compressional
stress field orientated WNW–ESE, consistent with the
trend of the underthrust margin; this regional stress dis-
tribution has previously been recognised in Australian
lithosphere by Woodhead et al. (2010). At the eastern
boundary of the underthrust margin, we see evidence
for more complex deformation occurring through trans-
lational source mechanisms and an additional dilational
regime, rotated into a generally northeast–southwest
orientation (Figure 7).
If we infer the underthrust Papua New Guinea mar-
gin is similar in extent to the margin prior to terrane
accretion, we can begin to add detail to this crustal block
and place it in the context of the surrounding structure
and geology expressed at the surface. Within the seismic
signature of the underthrust block (Figures 2,5) there
are earthquake clusters typified by an increase in the
density of seismic activity that are confined to the north
side of the Ramu-Markham Fault–New Britain trench
plate boundary. The regional context of these clusters,
and likewise the source regions, has not previously been
investigated. It is clear from Figure 2 that these clusters
are proximal to major structures; the Bundi Fault Zone,
Aure Fold Belt, and Owen Stanley Fault, where the struc-
tures are currently being underthrust beneath the Fini-
sterre Terrane at the Ramu-Markham Fault. We
interpret this seismicity as possible reactivation of the
former structural suture zones during their passage
beneath the Finisterre Terrane. It is reasonable to
assume this structure will continue to depth in the
downgoing crustal block, and likewise can be defined by
a similar earthquake cluster at greater depths as
observed between approximately 20 and 60 km
(Figures 2,5). Furthermore, we propose that based on
this premise, and similarities in geological context, that
is, ophiolite belts of similar interpreted age adjacent to,
and exhibiting analogous structural and tectonic rela-
tionships to major suture zones, that the Bundi Fault
Zone and Owen Stanley Fault can be correlated as the
same structural discontinuity (Figure 7). This provides
Bismarck Sea collision tectonics 613
Downloaded by [JAMES COOK UNIVERSITY], [R. J. Holm] at 15:30 20 August 2013
TAJE_A_824505.3d (TAJE) 09-08-2013 11:45
a new context for major structures within Papua New
Guinea as regionally significant suture zones rather
than discrete and unrelated geological phenomena.
Along-arc geochemical variation
Given the constraint provided by the new crustal archi-
tecture model presented here, we can begin to address
the implications of these findings and re-evaluate cur-
rent geological models of northeastern Papua New
Guinea. Most significantly, we can reinterpret the geo-
chemical signatures of the currently active New Britain
and West Bismarck arcs within a robust tectonic frame-
work. It becomes clear from the Solomon slab map that
the New Britain arc is related to the Solomon slab adja-
cent to the island of New Britain (Figure 7). However, in
the western Bismarck Sea, the north-dipping limb of the
subducted slab is located to the south of the West Bis-
marck arc and beneath the accreted Finisterre and Adel-
bert Terranes and the Papua New Guinea mainland
(Figure 7), spatially removed from the active West Bis-
marck arc. This observation immediately brings into
question the relationship between the West Bismarck
arc and the subducted Solomon slab proposed in previ-
ous studies. Instead, we suggest that the source of fluids
and potential crustal melts is readily available in the
form of the underthrust edge of Papua New Guinea crust
(Figure 7).
The extent of the underthrust margin at depth is out-
lined by both gravity and seismic signatures, consistent
with the interpretations of Davies et al. (1987), Honza
et al. (1987) and Woodhead et al. (2010). At a depth of
approximately 100 km in the mantle, the underthrust
continental margin is likely to be undergoing dewater-
ing processes and contributing fluids to the mantle
wedge. Given these fluids are derived from a continental-
crustal source rather than oceanic crust, this accounts
for the high sediment-signature input into the magmas
and comparatively reduced slab influence addressed
above, in contrast to the New Britain arc in the east.
This concept is further supported by the frontal edge of
the underthrust margin correlating spatially with the
overlying arc (Figure 7). Therefore, we define the New
Britain and West Bismarck arc as two distinct entities
with different source regions and different geochemical
affinities. These two arcs are the expression of either
slab-derived fluids (New Britain arc) or continental
crust-derived fluids (West Bismarck arc). There is, how-
ever, the added complication that the two arcs form a sin-
gle, more or less morphologically continuous volcanic
arc. Therefore, a zone must be present where fluids
derived from both subducted slab and underthrust conti-
nental crust are mixing and both contribute to arc mag-
matism (Figure 7). This model is consistent with an
observed continuum in the geochemical signatures
between the two arcs, transitioning from a slab
Figure 7 Interpretation of present-
day tectonic plate configuration
and magmatic arc distribution in
northeastern Papua New Guinea.
Gravity anomaly map is provided
as a base map. Bold black outlines
illustrate the extent of the under-
thrust continental crust, formerly
the leading edge of the Papua New
Guinea mainland; and the associ-
ated correlation of the Bundi Fault
Zone and Owen Stanley Fault in the
under-thrust crust. CMT solutions
are shown for the under-thrust
margin between 40 and 90 km
depth. Below the under-thrust mar-
gin, the distribution of subducted
oceanic crust of the Solomon slab is
shown for comparison, and is con-
toured at 40 and 100 km until termi-
nation to the slab. The ‘Bismarck
arc’is divided into the West Bis-
marck arc, New Britain arc, and
mixing zone between the two; these
are derived from continental crust,
oceanic slab, and a combination of
the two respectively. Cross-sections
through the plate arrangement are
provided to illustrate the 3-D frame-
work of the new plate arrangement
and context of corresponding fluid
sources of the equivalent magmatic
arc. See text for discussion.
614 R. J. Holm and S. W. Richards
Downloaded by [JAMES COOK UNIVERSITY], [R. J. Holm] at 15:30 20 August 2013
TAJE_A_824505.3d (TAJE) 09-08-2013 11:45
signature-dominated melt in the east through to the
crustal signature-dominated melt in the west (Figure 6).
This influence of slab-derived fluids is apparent up to
approximately 146.5E at which point the slab becomes
removed to the south beneath the underthrust continen-
tal margin and effectively blocked from any further con-
tribution of fluids in the arc building process (Figure 7).
This explains the delay in the apparent peak crustal sig-
natures in arc geochemistry up to this point as observed
by Woodhead et al. (2010).
It is also worth noting that further to the northwest
along the trend of the West Bismarck arc and towards
the Bismarck Sea seismic lineation, we see a geochemi-
cal trend that is consistent with a return to more mantle-
like signatures. We suggest the Bismarck Sea seismic lin-
eation behaves as a leaky transform in line with findings
from Llanes et al. (2009) resulting in variable continen-
tal-derived fluid contribution to arc magmatism.
Geodynamic evolution of the Bismarck Sea
Given the regional significance of the findings outlined
in this study, we propose a new plate tectonic reconstruc-
tion for the Bismarck Sea region that builds on previous
reconstructions of the region (e.g. Abbott 1995; Hill &
Raza 1999; Weiler & Coe 2000; Hall 2002) and incorporates
the new tectonic elements introduced in this paper.
Reconstructions of the Australian plate include all previ-
ously accreted terranes. GPS measurements of current,
geologically instantaneous plate motions (Tregoning
et al. 1998,1999; Tregoning 2002; Wallace et al. 2004) form
the basis for this new Bismarck Sea reconstruction
while the published timing for events and sea floor mag-
netic anomalies from Taylor (1979), Goodliffe et al.
(1997), Taylor et al. (1999) and Gaina & M€
uller (2007) are
used to infer the direction and rate of seafloor spreading.
Paleomagnetic rotation data for the Finisterre Terrane
from Weiler & Coe (2000) is used to further constrain
reconstructions.
New reconstructions highlighting the significance of
the leading Australian continental margin are shown in
Figure 8. Between 3 and 4 Ma the Australian plate col-
lided with the Adelbert-Finisterre Terrane, closing the
western New Britain trench and forming the Ramu-
Markham Fault (Abbott et al. 1994; Abbott 1995). This is
coincident with decoupling and the initial formation of
the North Bismarck plate, South Bismarck plate and
associated Bismarck Sea seismic lineation (Martinez &
Figure 8 Tectonic reconstruction of the Bismarck Sea region. The inferred Papua New Guinea northeastern continental mar-
gin is shown in grey, and highlights the continuation of the Bundi–Owen Stanley suture and the seaward continental shelf; yel-
low dotted line represents the Finisterre Volcanics of the Adelbert and Finisterre Ranges. Magnetic isochrons and spreading
centres are included for the Woodlark Basin (Taylor et al. 1999), Solomon Sea (Gaina & M€
uller 2007) and Manus Basin (Taylor
1979). Filled triangles and open triangles indicate normal and slow or extinct subduction respectively. AFT, Adelbert-
Finisterre Terrane; SBP, South Bismarck Plate; NBP, North Bismarck Plate; AP, Adelbert microplate; MSC, Manus Spreading
Centre; RMFZ, Ramu-Markham Fault Zone; OJP, Ontong Java Plateau. Reconstructions are presented in a fixed hot spot refer-
ence frame.
Bismarck Sea collision tectonics 615
Downloaded by [JAMES COOK UNIVERSITY], [R. J. Holm] at 15:30 20 August 2013
TAJE_A_824505.3d (TAJE) 09-08-2013 11:45
Taylor 1996). Continued advance of the Australian plate
impinging on the western South Bismarck plate resulted
in the onset of clockwise rotation of the South Bismarck
plate (e.g. Weiler & Coe 2000). Furthermore, the South
Bismarck plate became decoupled from the North Bis-
marck plate and the subducting Solomon Sea plate,
allowing the South Bismarck plate to rotate freely about
a pole southeast of the Finisterre Terrane (Tregoning
et al. 1999). Rotation of the South Bismarck plate resulted
in retreat of the New Britain trench in the east and rift-
ing and sea-floor spreading at the Manus spreading cen-
tre from 3 Ma (Taylor 1979; Martinez & Taylor 1996).
Fragmentation of the South Bismarck plate in response
to collision with the thickened crust of the Bundi–Owen
Stanley suture zone formed the independent Adelbert
microplate. Right lateral offset of the Oligocene–Lower
Miocene Finisterre Volcanics in the Adelbert and Fini-
sterre Terranes is interpreted to have resulted from
crustal displacement along the Adelbert microplate–
South Bismarck plate boundary. Continued north-
northeast motion of the Australian plate combined with
clockwise rotation of the South Bismarck plate resulted
in ongoing subduction of the Solomon Sea plate and ini-
tiation of rifting adjacent to the Manus spreading centre
with opening the Manus microplate from ca 1Ma
(Martinez & Taylor 1996).
The concept of a separate Adelbert microplate,
although presented here for the first time in a regional
context, is not a new idea and initially arose from geolog-
ical observations of the volcanic island-arc terranes in
the Adelbert and Finisterre Ranges (e.g. Jaques & Robin-
son 1977; Abbott et al. 1994; Wallace et al. 2004). A physi-
cal boundary between the Adelbert and Finisterre
Ranges was interpreted in early mapping of cross fault-
ing in the Finisterre Ranges (Jaques & Robinson 1977;
Abbott et al. 1994), however, dextral translational faults
defined by Abers & McCaffrey (1994) (Figure 1) have
been adopted as the Adelbert-Finisterre boundary in
this study. The most apparent contrast between the Adel-
bert microplate and South Bismarck plate is a dramatic
difference in elevation (Figure 1; Abbott 1995). Abbott
(1995) interpreted that the Gowop Limestone cap both
the Adelbert and Finisterre Ranges and that the contrast
in elevation of the Adelbert Range is due to less total
uplift of the Adelbert block, and cannot be attributed to
erosion in the Adelbert Range. This contrast is likely to
be the result of differential plate motion with northward
motion and clockwise rotation of the Adelbert micro-
plate relative to the South Bismarck plate resulting in a
lower rate of convergence with the Australian plate.
Such differential motion also explains the observed off-
set of the Finisterre Volcanics common to both ranges
(Figure 8; Jaques & Robinson 1977; Abbott et al. 1994).
Moreover, the nature of the continental crust under-
thrust beneath the Adelbert and Finisterre Terranes
likely holds implications for the degree of uplift. Recon-
structions show the Finisterre Terrane was forced over,
and currently overlies the thickened crust of the Bundi–
Owen Stanley suture zone, in contrast to the marginal
continental crust beneath the Adelbert Terrane, thus
resulting in reduced total uplift compared with the Fini-
sterre Ranges. This is a new approach to explain differ-
ential uplift of the Adelbert and Finisterre Terranes
where the nature and geometry of the prior continental
margin, in combination with collision obliquity, controls
accretion dynamics.
It is also apparent from previous reconstructions and
those presented here that arc–continent collision in
northeast Papua New Guinea was a consequence of obli-
que collision between the larger Australian and Pacific
plates. This process ‘zipped’shut the intervening
Solomon Sea between the Papua New Guinea mainland
and the outboard Adelbert and Finisterre Terranes
(Figure 8;Abbott1995;Hill&Raza1999). However, the zip-
ping process continues to the present-day along this mar-
gin, therefore, collision is considered time transgressive
and will continue to migrate eastwards. Figure 9 presents
forward-looking tectonic reconstructions where contin-
ued subduction of the Solomon Sea plate at the New
Britain trench will ultimately consume the Solomon Sea
leading to arc–continent collision between the Papuan
Peninsula (Australian plate) and New Britain. We esti-
mate it will take approximately 5 m.y. to consume the
Solomon Sea at present plate motion rates. As in the pre-
vious instance of the Finisterre Terrane, we predict slab
pull forces acting on the subducted Solomon Sea slab cou-
pled to the Australian plate at the point of collision will
lead to drawdown and underthrusting of the leading edge
of continental crust. This process initiates a plate configu-
ration where New Britain is allowed to overthrust the
continental margin. Throughout this process we predict
the apparent continual outboard migration of the associ-
ated subduction-derived magmatic arc as the accreting
plate overrides the loci of slab dewatering (Figure 9).
The tectonic process of arc–continent accretion is not
unique to the tectonic evolution of Papua New Guinea.
In the geological record, there are many recognised arc–
continent collision episodes (e.g. Teng 1990; Rosenbaum
et al. 2002; Whattam 2009; Najman et al. 2010), and simi-
larly occurrences of continental crust entering a trench
and failing to subduct (e.g. New Caledonia; Aitchison
et al. 1995; Rawling & Lister 2002; Spandler et al. 2005).
These collision events are typically distinguished in
field studies by the presence of ultra-high-pressure meta-
morphic terranes and/or obduction of ophiolite sequen-
ces, however, these observed rock types are not present
en masse in northeast Papua New Guinea, and it seems
apparent that the processes of terrane accretion and par-
tial subduction of continental crust remain active. In the
distant future, we can reasonably expect some exhuma-
tion of subducted continental crust will occur in line
with findings in ancient arc–continent collision events.
Recognising this process in the recent geological past,
and in the near future for New Britain, provides valu-
able insight into the dynamics and tectonic settings that
give rise to such geological phenomena. Furthermore,
the role of subducted continental crust as a potential
source of fluids in magmatic arc generation, in this case
the West Bismarck arc, is a relatively new avenue of geo-
logical research and will only occur under exceptional
tectonic circumstances. Nevertheless, recognition of the
expression of continental crust-derived arcs at the sur-
face and an understanding of the tectonic events leading
up to the generation of such apparent geological anoma-
lies will be invaluable in unravelling complex tectonic
regions globally.
616 R. J. Holm and S. W. Richards
Downloaded by [JAMES COOK UNIVERSITY], [R. J. Holm] at 15:30 20 August 2013
TAJE_A_824505.3d (TAJE) 09-08-2013 11:45
CONCLUSIONS
We re-evaluate the tectonics of northeastern Papua New
Guinea and the South Bismarck Sea region, and present
a new and robust model for the 3-D architecture of the
crust and upper mantle in this region. The new tectonic
plate model accounts for all the observed geological phe-
nomena including geology, geophysics, seismicity, arc
geochemistry, topography, bathymetry, landscape mor-
phology, and a newly developed slab model of crust sub-
ducted at the New Britain trench, while also drawing
from studies and concepts introduced by numerous
workers over decades of research. Compilation and re-
evaluation of these data reveal a previously unrecog-
nised crustal block at depth beneath the Finisterre Ter-
rane on the northern Papua New Guinea coast, defined
by anomalous seismicity and a correlative negative grav-
ity anomaly. We interpret this as the leading edge of Aus-
tralian continental crust, which has become partially
subducted beneath the Adelbert and Finisterre Terranes
during arc–continent collision and terrane accretion.
This continental crust extends to a depth of approxi-
mately 100 km in the mantle where it is currently under-
going dewatering reactions and contributing fluids to
the upper plate resulting in formation of the West
Bismarck arc adjacent to the northeastern Papua New
Guinea coastline. The West Bismarck arc is a distinct
arc, separate from the slab-derived New Britain arc to
the east, however, a mixing zone exists between the two
where fluids derived from both normal subduction pro-
cesses and also partially subducted continental crust are
contributing to arc magmatism. In light of these findings
we present a new tectonic reconstruction for the recent
development of the Bismarck Sea in conjunction with a
forward-looking reconstruction to highlight the role of
marginal continental crust in the dynamics of arc–conti-
nent collision and accretion. Furthermore, we stress the
importance of the recognition of continental crust and
marginal provinces in collisional tectonic settings and
their potential to contribute in arc building processes.
ACKNOWLEDGEMENTS
We would like to thank Nautilus Minerals for providing
data from the Hamme Cruise. We also thank Carl Spandler,
Thorsten Becker, Patrice Rey, Gideon Rosenbaum and
Robert Hall and an anonymous reviewer for helpful
criticisms and suggestions.
REFERENCES
ABBOTT L. D. 1995. Neogene tectonic reconstruction of the Adelbert-
Finisterre–New Britain collision, norther n Papua New Guinea.
Journal of Southeast Asian Earth Sciences 11,33–51.
ABBOTT L. D., SILVER E. A. & GALEWSKY J. 1994. Structural evolution of a
modern arc–continent collision in Papua New Guinea. Tectonics
13, 1007–1034.
ABERS G. A. & MCCAFFREY R. 1994. Active arc–continent collision:
Earthquakes, gravity anomalies, and fault kinematics in the
Huon–Finisterre collision zone, Papua New Guinea. Tectonics 13,
227–245.
ABERS G. A. & ROECKER S. W. 1991. Deep structure of an arc–continent
collision: Earthquake relocation and inversion for upper mantle
P and S wave velocities beneath Papua New Guinea. Journal of
Geophysical Research 96, 6379–6401.
AITCHISON J. C. , CLARKE G. L., MEFFRE S. & CLUZEL D. 1995. Eocene arc–
continent collision in New Caledonia and implications for
regional southwest Pacific tectonic evolution. Geology 23,161–
164.
AMANTE C. & EAKINS B. W. 2009. ETOPO1 1 Arc-Minute Global Relief
Model: Procedures, Data Sources and Analysis. NOAA Technical
Memorandum NESDIS National Geophysical Data Center,
Colorado, USA.
AUSTRALIAN BUREAU OF MINERAL RESOURCES 1970. Geophysical survey in
Gulf of Papua. Bismarck Sea, cruise id: HAMME70, BMR survey 5,
Australian Bureau of Mineral Resources, Canberra.
CASAS A., KEAREY P. , R IVERO L. & ADAM C. R. 1997 Gravity anomaly map
of the Pyrenean region and a comparison of the deep geological
Figure 9 Forward tectonic reconstruction of progressive arc
collision and accretion of New Britain to the Papua New
Guinea margin. (a) Schematic forward reconstruction of
New Britain relative to Papua New Guinea assuming contin-
ued northward motion of the Australian plate and clockwise
rotation of the South Bismarck plate. (b) Cross-sections illus-
trate a conceptual interpretation of collision between New
Britain and Papua New Guinea.
Bismarck Sea collision tectonics 617
Downloaded by [JAMES COOK UNIVERSITY], [R. J. Holm] at 15:30 20 August 2013
TAJE_A_824505.3d (TAJE) 09-08-2013 11:45
structure of the western and eastern Pyrenees. Earth and Plane-
tary Sciences Letters 150,65–78.
COOPER P. & T AY L O R B. 1987. Seismotectonics of New Guinea: a model for
arc reversal following arc–continent collision. Tectonics 6,53–67.
COOPER P. & T AYLOR B. 1989. Seismicity and focal mechanisms at the
New Britain trench related to deformation of the lithosphere.
Tectonophysics 164,25–40.
CROOK K. A. 1987. Petrology and mineral chemistry of sedimentary
rocks from the western Solomon Sea. Geo-Marine Letters 6,203–209.
DAVIES H. L. 2012. The geology of New Guinea—the cordilleran mar-
gin of the Australian continent. Episodes 35,87–102.
DAVIES H. L., LOCK J., T IFFIN D. L. , HONZ A E., OKUDA Y. , MURAKAMI F. &
KISIMOTO K. 1987. Convergent tectonics in the Huon Peninsula
region, Papua New Guinea. Geo-Marine Letters 7, 143–152.
DENHAM D. 1969. Distribution of Earthquakes in the New Guinea–Sol-
omon Islands region. Journal of Geophysical Research 74, 4290–
4299.
ENGDAHL E. R. 2006. Application of an improved algorithm to high pre-
cision relocation of ISC test events. Physics of the Earth and Plan-
etary Interiors 158,14–18.
ENGDAHL E. R., VAN DER HILST R. & BULAND R. 1998. Global teleseismic
earthquake relocation with improved travel times and proce-
dures for depth determination. Bulletin of the Seismological Soci-
ety of America 88, 722.
GAINA C. & M€
ULLER D. 2007. Cenozoic tectonic and depth/age evolution
of the Indonesian gateway and associated back-arc basins. Earth
Science Reviews 83, 177–203.
GOODLIFFE A. M., TAYLOR B., MARTINEZ F. , H EY R., MAEDA K. & OHNO K.
1997. Synchronous reorientation of the Woodlark Basin Spread-
ing Center. Earth and Planetary Science Letters 146, 233–242.
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 Indo-
nesia–Tonga region: insights from tomography. Earth and Plane-
tary Science Letters 201, 321–336.
HENDRIX M. S., DUMITRU T. A . & G RAHAM S. A. 1994. Late Oligocene–
early Miocene unroofing in the Chinese Tian Shan: An early
effect of the India–Asia collision. Geology 22, 487–490.
HILL K. C. & RAZA A. 1999. Arc–continent collision in Papua Guinea:
Constraints from fission track thermochronology. Tectonics 18,
950–966.
HONZA E., DAVIES H. L., KEENE J. B. & T IFFIN D. L. 1987. Plate boundaries
and evolution of the Solomon Sea region. Geo-Marine Letters 7,
161–168.
ISACKS B., OLIVER J. & SYKES L. R. 1968. Seismology and the new global
tectonics. Journal of Geophysical Research 73, 5855–5899.
JAKES P. & G ILL J. B. 1970. Rare earth elements and the island arc tho-
leiite series. Earth and Planetary Science Letters 9,17–28.
JAQUES A. L. & ROBINSON G. P. 1977. The continent–island arc collision
in northern Papua New Guinea. BMR Journal of Australian Geol-
ogy and Geophysics 2, 289–303.
JOHNSON R. W. 1977. Distribution and major element chemistry of late
Cenozoic volcanoes at the southern margin of the Bismarck Sea,
Papua New Guinea. Australian Bureau of Mineral Resources
Report, vol. 188.
JOHNSON R. W. & CHAPPELL B. W. 1979. Chemical analyses of rocks from
the late Cenozoic volcanoes of north-central New Britain and the
Witu Islands, Papua New Guinea. Australian Bureau of Mineral
Resources Report, vol. 209.
JOHNSON R. W. & JAQUES A. L. 1980. Continent–arc collision and rever-
sal of arc polarity: new interpretations from a critical area. Tecto-
nophysics 63,111–124.
JOHNSON T. & M OLNAR P. 1972. Focal mechanisms and plate tectonics of
the southwest Pacific. Journal of Geophysical Research 77, 5000–
5032.
JOHNSON R. W., MUTTER J. C . & ARCULUS R. J. 1979. Origin of the Willau-
mez-Manus Rise, Papua New Guinea. Earth and Planetary Sci-
ence Letters 44,247–260.
KLEY J. & VOIGT T. 2008. Late Cretaceous intraplate thrusting in cen-
tral Europe: Effect of Africa–Iberia–Europe convergence, not
Alpine collision. Geology 36, 839–842.
LLANES P. , S ILVER E., DAY S. & HOFFMAN G. 2009. Interactions between a
transform fault and arc volcanism in the Bismarck Sea, Papua
New Guinea. Geochemistry, Geophysics, Geosystems 10, Q06013.
MADSEN J. A. & L INDLEY I. D. 1994. Large-scale structures on Gazelle Pen-
insula, New Britain: Implications for the evolution of the New Brit-
ain Arc. Australian Journal of Earth Sciences 41,561–569.
MANN P. & T AIRA A. 2004. Global tectonic significance of the Solomon
Islands and Ontong Java Plateau convergent zone. Tectonophys ics
389, 137–190.
MARTINEZ F. & T AYLOR B. 1996. Backarc spreading, rifting, and micro-
plate rotation, between transform faults in the Man us Basin.
Marine Geophysical Research 18,203–224.
MILLER M. S., GORBATOVA. & KENNETT B. L. N. 2006. Three-dimensional
visualization of a near vertical slab tear beneath the southern
Mariana Arc. Geochemistry, Geophysics, Geosystems 7, Q06012.
MISHRA D. C. , SINGH B., TIWARI V. M . , G UPTA S. B. & RAO M. B. S. V. 2000.
Two cases of continental collisions and related tectonics during
the Proterozoic period in India—insights from gravity modeling
constrained by seismic and magnetotelluric studies. Precam-
brian Research 99, 149–169.
MORALES J., S ERRANO I., JABALOY A., GALINDO-ZALD
IVA R J., Z HAO D., T ORCAL
F. , V IDAL F. & G ONZ
ALEZ LODEIRO F. 1999. Active continental subduc-
tion beneath the Betic Cordillera and the Albor
an Sea. Geology
27, 735–738.
NAJMAN Y. , A PPEL E., BOUDAGHER-FADEL M., BOWN P. , C ARTER A., GARZANTI
E., GODIN L., HAN J., L IEBKE U. , OLIVE R G., PARRISH R. & VEZZOLI G.
2010. Timing of India–Asia collision: Geological, biostratigraphic,
and palaeomagnetic constraints. Journal of Geophysical Research
115, B12416.
O’KANE T. 2008. 3-D structure and tectonic evolution of the Papua New
Guinea and Solomon Islands region and its relationship to Cu–Au
mineralization. Unpublished Honours thesis, Research School of
Earth Sciecnes, The Australian National University.
PEGLER G., DAS S. & WOODHOUSE J. H. 1995. A seismolo gical study of the
eastern New Guinea and the western Solomon Sea re gions and
its tectonic implications. Geophysical Journal International 122,
961–981.
PETTERSON M. G., BABBS T. , NEAL C. R., MAHONEY J. J. , S AUNDERS A. D.,
DUNCAN R. A., TOLIA D., M AGU R., QOPOTO C., MAHOA H. & NATOGGA
D. 1999. Geological-tectonic framework of Solomon Islands, SW
Pacific: crustal accretion and growth with an intra-oceanic set-
ting. Tectonophysics 301,35–60.
PLANK T. 2005. Constraints from thorium/lanthanum on sediment
recycling at sub- duction zones and the evolution of the conti-
nents. Journal of Petrology 46,921–944.
RAWLING T. J. & L ISTER G. S. 2002. Large-scale structure of the eclogite-
blueschist belt of New Caledonia. Journal of Structural Geology
24, 1239–1258.
RICHARDS S., HOLM R. & BARBER G. 2011. When slab collide: A tectonic
assessment of deep earthquakes in the Tonga–Vanuatu region.
Geology 39, 787–790.
RICHARDS S., LISTER G. & KENNETT B. 2007. A slab in depth: Three-
dimensional geometry and evolution of the Indo-Australian plate.
Geochemistry Geophysics Geosystems 8, Q12003.
RIPPER I. D. 1982. Seismicity of the Indo-Australian/Solomon Sea plate
boundary in the southeast Papua region. Tectonophysics 87,355–
369.
ROSENBAUM G., LISTER G. S. & DUBOZ C. 2002. Relative motions of Africa,
Iberia and Europe during Alpine orogeny. Tectonopysics 359,117–
129.
SOBEL E. R. & DUMITRU T. A. 1997. Thrusting and exhumation around
the margins of the western Tarim basin during the India–Asia
collision. Journal of Geophysical Research 102, B3, 5043.
SPANDLER C., RUBATTO D. & HERMANN J. 2005. Late Cretaceous–Tertiary
tectonics of the southwest Pacific: Insights from U–Pb sensitive,
high resolution ion microprobe (SHRIMP) dating of eclogite
facies rocks from New Caledonia. Tectonics 24, TC3003.
STERN T. A. 1995. Gravity anomalies and crustal loading at and adja-
cent to the Alpine Fault, New Zealand. New Zealand Journal of
Geology and Geophysics 38, 593–600.
STEVENS C., MCCAFFREY R., SILVER E. A., SOMBO Z., ENGLISH P. & VAN DER
KEVIE J. 1998. Mid-crustal detachment and ramp faulting in the
Markham Valley, Papua New Guinea. Geology 26,847–850.
TAYLOR B. 1979. Bismarck Sea: Evolution of a back-arc basin. Geology
7,171–174.
TAYLOR B., GOODLIFFE A. M. & MARTINEZ F. 1999. How continents break
up: Insights from Papua New Guinea. Journal of Geophysical
Research 104, 7497–7512.
618 R. J. Holm and S. W. Richards
Downloaded by [JAMES COOK UNIVERSITY], [R. J. Holm] at 15:30 20 August 2013
TAJE_A_824505.3d (TAJE) 09-08-2013 11:45
TENG L. S. 1990. Geotectonic evolution of Late Cenozoic arc–continent
collision in Taiwan. Tectonophysics 183,57–76.
TREGONING P. 2002. Plate kinematics in the western Pacific derived
from geodetic observations. Journal of Geophysical Research 107,
B1 2020.
TREGONING P. , J ACKSON R. J., MCQUEEN H., LAMBECK K., STEVENS C., LITTLE
R. P., CURLEY R. & ROSA R. 1999. Motion of the South Bismarck
Plate, Papua New Guinea. Geophysical Research Letters 26, 3517–
3520.
TREGONING P. , L AMBECK K., STOLTZ A., MORGAN P. , M CCLUSKY S. C., VAN DER
BEEK P. , M CQUEEN H., JACKSON R. J., LITTLE R. P., LAING A. & MURPHY
B. 1998. Estimation of current plate motions in Papua New
Guinea from Global Positioning System observations. Journal of
Geophysical Research B: Solid Earth 103, 12181–12203.
WALLACE L. M., STEVENS C., SILVER E., MCCAFFREY R., LORATUNG W. , H ASI-
ATA S., STANAWAY R., CURLEY R., ROSA R. & TAUGALOIDI J. 2004. GPS
and seismological constraints on active tectonics and arc–conti-
nent collision in Papua New Guinea: Implications for mechanics
of microplate rotations in a plate boundary zone. Journal of Geo-
physical Research 19, B05404.
WEILER P. D . & C OE R. S. 2000. Rotations in the actively colliding
Finisterre Arc Terrane: paleomagnetic constraints on
Plio–Pleistocene evolution of the South Bismarck microplate,
northeastern Papua New Guinea. Tectonophysics 316,297–325.
WHATTAM S. A. 2009. Arc–continent collisional orogenesis in the SW
Pacific and the nature, source and correlation of emplaced ophio-
litic nappe components. Lithos 113,88–114.
WOODHEAD J. D. & JOHNSON R. W. 1993. Isotopic and trace element pro-
files across the New Britain island arc, Papua New Guinea. Con-
tributions to Mineralogy and Petrology 113,479–491.
WOODHEAD J. D., E GGINS S. M. & GAMBLE J. G. 1993. High field strength
and transition element systematics in coupled arc–back arc set-
tings: evidence for multi-phase melt extraction and a depleted
mantle wedge. Earth and Planetary Science Letters 114,491–504.
WOODHEAD J. D., E GGINS S. M. & JOHNSON R. W. 1998. Magma genesis in
the New Britain island arc: Further insights into melting and
mass transfer processes. Journal of Petrology 39, 1641–1668.
WOODHEAD J., H ERGT J., S ANDIFORD M. & JOHNSON W. 2010. The big
crunch: Physical and chemical expressions of arc/continent colli-
sion in the Western Bismarck arc. Journal of Volcanology and
Geothermal Research 190,11–24.
Received 7 June 2012; accepted 8 July 2013
SUPPLEMENTARY PAPER
Earthquake density distribution maps for depth bins of
0–20, 20–40, 40–60, 60–80, 80–100, 100–120 and 120–140 km.
Earthquake densities are contoured from high density
(white) to low density (blue).
Bismarck Sea collision tectonics 619
Downloaded by [JAMES COOK UNIVERSITY], [R. J. Holm] at 15:30 20 August 2013
