~ectono~h~s~c~,144 (1987) 337-352 337

EIsevier Science Publishers B.Y., Amsterd~ - Printed in The Netberiands

Plate boundary evolution in the Halmahera region, Indonesia

ROBERT HALL

(Received December 1.1986; revised version accepted March 10,1987)

Abstract

Hall, R., 1987. Plate boundaryevolution in the Halmahera region, Indonesia. ~e~~~no~~.y~jcs,144: 337-352,

H&mahera is situated in eastern Indonesia at the southwest comer of the Philippine Sea PIate. Active arc-arc
collision is in process in the Molucca Sea to the west of Halmahera. New stratigraphic observations from Halmahera
link this island and the east P~Iippin~ and record the history of subduction of the Molucca Sea lithosphere. The
HaImahera Basement Complex and the basement of east Mindanao were part of an arc and forearc of Late
Cretaceous-Early Tertiary age and have formed part of a single plate since the Late Eocene-Early Oligocene. There is
no evidence that HaImabera formed part of an Oligo-Miocene arc but arc volcanism, associated with eastwards
subduction of the Molucca Sea beneath Halmahera, began in the Pliocene and the Pliocene arc is built on a basement
of the early Tertiary arc. Arc volcanism ceased briefIy during the Pleistocene and the arc shifted westwards after an
episode of deformation. The present active arc is built upon deformed rocks of the Ptiocene arc. The combination of
new strati~ap~c info~ation from the genera islands and models of the present-day tectonic structure of the
region deduced from seismic and other geophysicat studies is used to constrain the tectonic evolution of the region
since the Miocene. Diacbronous collision at the western edge of the Philippine Sea PIate which began in Mmdanao in
the Late Miocene impeded the movement of the Philippine Sea Plate and further motion has been achieved by a
combination of strike-&p motion aiong the P~~ppine Fault, subduction at the Philippine Trench and subduction of
the Molucca Sea lithosphere beneath Halmahera.

Introduction seismic and other geophysical studies. The combi-

nation of these two approaches can be used to
H~rn~era is situated in the centre of a mozaic trace the evolution of plate boundaries in the
of microplates at the boundary between Austral- region.
asia, Eurasia and Pacifica in one of the most
seismically active regions of the Earth (Fig. 1). Its Present tectonic setting
K-shape resembles that of Sulawesi on the oppo-
site side of the Molucca Sea and in general terms The Molucca Sea separates the opposed Sangihe
Halmahera is a smaller scale version of Sulawesi: and Halmahera volcanic arcs (Fig. 1) interpreted
the western arms of the K form a volcanic arc as the expression of the subduction of the Molucca
while the eastern arms include ophiolites and sedi- Sea Plate to both east and west (Hatherton and
ments. The Halm~era region is remarkable in Dickinson, 1969; Hamilton, 1979). The Molucca
being the only example of active arc-arc collision Sea Plate has been almost completely eliminated
and in preserving the processes of subduction in by subduction (Cardwell et al., 1980; McCaffrey
both a stratigraphic record which can be de- et al., 1980), forcing it into an asymmetrical in-
termined by field-based geological studies and a verted U-shaped ~nfigura~on beneath the collid-
lithospheric record which can be interpreted from ing fore-arcs (Fig, 2). Collision is creating a high

~-~951/87/~3.50 Q 1987 Elsevier Science Publishers B.V.

338

20”
0

PHILIPPINE
KM.

PLATE

AUSTRALIAN PLATE
/sl

Fig. 1. Principal tectonic features of the Halmahera region after Hamilton (1979) and Silver (1981). Solid triangles are active
volcanoes of the Halmahera and Sangihe arcs. Slip rates along the Philippine Trench from Ranken et al. (1984).

central ridge to the Molucca Sea which is being gravity low is the Molucca Sea “melange wedge”
thrust onto the two colliding fore-arcs. This central or “collision complex” (Silver and Moore, 1978;
zone, marked by intense shallow seismicity and a Hamilton, 1979; Moore et al., 1981) which is
 339

Seismicity associated with the Philippine Trench

ceases south of about 2.5* N, arguing against
Hamilton’s interpretation, and Cardwell et al.
(1980) proposed a Halmahera-Palau transform
fault to connect the Halmahera subduction zone
with the Ayu Trough. In contrast, McCaffrey
(1982) suggested that the transform fault is located
further north and stops at the Philippine Trench.
In both of these models Halmahera is part of the
Philippine Sea Plate. These differences in interpre-
tation reflect the complex pattern of boundaries at
the junction of the Philippine, Australian and
Southeast Asian plates and their shifting position
resulting from collision in the Molucca Sea region.
~bLucc~ SEA PLATE
The present tectonic setting is an important
constraint on models of the short-term geological
evolution, but information from studies of land
geology is equally vital for reconstructions of the
Fig. 2. Present configuration of the Molucca Sea Plate in the development of the region and to distinguish be-
region between Halmahera and the Sangihe Arc after Cardwell tween interpretations of present-day tectonics,
et al. (1980). especially where-as in the Halmahera region-
the geophysical evidence is ambiguous. The length
of subducted lithosphere on seismic profiles across
the region (Cardwell et al., 1980; Fig. 2) indicates
exposed on the island of Talaud. This zone con- that Halmahera must have been situated at least
tinues northwards into Mindanao where collision 1000 km to the east of north Sulawesi before
occurred in the Miocene and the Philippine Fault subduction of the Molucca Sea began. However,
follows the suture zone (Moore and Silver, 1983). the timing of initiation of subduction beneath
The southern boundary of the region is the Sorong Halmahera is unknown. The geological link be-
Fault zone along which H~m~era is currently tween the P~~ppines, Talaud, Halmahera and
moving west with respect to Australasia (Ham- surrounding regions remains unclear and this is
ilton, 1979). East of Halmahera plate boundaries reflected in the number of different regional re-
are uncertain. Northeast of Halmahera the Philip- constructions proposed (see discussion in Moore
pine Trench is known to be very young (Hamilton, and Silver, 1983). Halmahera, situated in the centre
1979; Cardwell et al., 1980), with less than 150 km of this complex zone, is clearly a source of critical
of subducted lithosphere, but it does not extend information but its geology is not well known.
south of about 2ON. The seismic&y of the region, Van Bemmelen (1970) su~~sed the stratigra-
or its local absence, has been interpreted to phy of the islands in a single table based on
terminate the Philippine Trench, and thus bound reconnaissance by Verbeek (1908) Wanner (1913),
the Philippine Sea Plate at its southern end, in Brouwer (1923a, 1923b) and Kuenen (1935). Bes-
different ways. Hamilton (1979) extended the sho (1944) made a wartime study of the geology
Philippine Sea Plate boundary from the Philippine and later Verstappen (1960, 1964) made an aerial
Trench southwards as a thrust or strike-slip fault photographic study of the volcanoes. Hamilton
to the northern New Guinea margins and Sorong (1979) shows western Halmahera as a magmatic
Fault system; in his model Halmahera and the arc and the eastern part simply as melange. Re-
Philippine Sea are thus separate plates although cent mapping reviewed by Sukamto et al. (1981),
the northern extension of the Halmahera subduc- Soeria Atmadja (1981) and Silitonga et al. (1981)
tion zone towards Mindanao remains a problem. indicates that the Halmahera islands include an
 PLIOCENE

145 I
JURASSIC
MO I------
J
Fig. 3. Summary of the stratigraphy of Halmahera with timing and interpretation of events in the region. SW text for discussion.

active volcanic arc in the west, a fragment of and igneous and metamorphic rocks representing
probable continental basement on Bacan, and the deeper parts of the fore-am, were imbricated
ophiolites and blueschists (Burgath et al., 1983) together during the Late Eocene. The Halmahera
associated with deep-water Mesozoic and Tertiary Basement Complex can be traced into eastern
sediments in the east. A new field investigation of Mindanao (Ramreft et al., 1960; Moore and Silver,
Halmahera has provided important new informa- 1983) and probably further northwards in the
tion on the geological evolution of the region (Hall eastern Philippines (&rig, 1983). In contrast, the
et al., 1987a, 1987b,3 which is summarised in Fig. southern part of the island of Bacan at the south-
3. The account of the geology of Halmahera below western end of the Halmahera group has a base-
is based on this work and is followed by a discus- ment of high-grade cantinental metamorphic rocks
sion of the evolution of the region in the light of associated with a deformed and met~o~ho~
the new information. ophiolitic complex quite different to the basement
of eastern H&m&era. The metamorphic rocks are
interpreted to be part of the north Australian
continental margin basement which is separated
Eastern Halmahera has a basement of dismem- from the H&m&era Basement Complex by a splay
bered ophiolitic rocks with slices of Mesozoic and of the Sarong Fault system, and the deformed
Eocene sediments overlain .~~nfo~ably by ophiolite complex of Bacan is sifted to repre-
Middle-Late Oligocene and younger sedimentary sent magmatism in the fault zone.
and volcanic rocks. The Mesozoic and Eocene The basement of western H&m&era is largely
rocks reveal notable stratigraphical and petrologi- covered by Neogene-Recent sedimentary and
eal similarities to the Marianas fore-arc and the volcauic rocks and remains poorly known. The
eastern Halmahera Baement Complex is interpre- oldest rocks are volcanic elastics which are typi-
ted as a pre-Oligocene fore-arc. The sediments, cally unfossiliferous and include pyroclastics, lava
 341

breccias and sub-aerial conglomerates, locally hy- terrain forming the Halmahera basement was
drothermally altered and deeply weathered and strongly deformed and uplifted. The Oligocene
consequently extremely difficult to date. A was a period of uplift and deep erosion of the
volcanic conglomerate from the western coast of Basement Complex forming deep valleys contain-
central Halmahera contains possible rudist frag- ing fluviatile ophiolitic conglomerates now being
ments and about 20 km further east mudstones re-excavated by the present-day rivers. Slow subsi-
and marls interbedded with volcaniclastics contain dence began in eastern Halmahera in mid-late
planktonic forams of early Middle Eocene age. Oligocene, leading to deposition of marls, and reef
This evidence suggests that the younger volcanic carbonate deposition began further to the west in
arcs which built the western arms of Halmahera the Early Miocene. For the remainder of Miocene
cover the eroded Late Cretaceous-Early Tertiary times the whole region was the site of shallow
volcanic arc. The 1: 250,000 geological maps marine carbonate deposition. In eastern and
(Apandi and Sudana, 1980; Supriatna, 1980; central Halmahera there was a change from
Yasin, 1980) assign these basement rocks to a carbonate to marl deposition in the early Pliocene,
Bacan Formation, tentatively dated as Late Oligo- followed by a sharp increase in siliciclastic debris
cene-Early Miocene. However, this “formation” deposited as turbidites in a submarine fan setting.
includes a number of unrelated units such as Late Calc-alkaline volcanic debris appeared in the se-
Cretaceous breccias which are imbricated in the quence in the mid-Pliocene and the amount of
Basement Complex in eastern Halmahera, unde- volcanic material increased with time, initially as
formed late Paleogene volcanic breccias on Bacan tuffs and volcaniclastic turbidites and later as
(Yasin, 1980; Silitonga et al., 1981) as well as the lavas. This rapid transition is interpreted as the
Eocene rocks noted above. Hamilton (1979) sug- result of the initiation of subduction of the
gested that between the Oligocene and Early Molucca Sea lithosphere to the west of Halmahera
Miocene Halmahera was an east-facing island arc and the formation of a Pliocene volcanic arc in the
and a subduction polarity flip led to the present western province. Rupture of the lithosphere was
tectonic configuration. However, in the area of preceded by sudden downwarping of the crust
detailed investigation in the northeast arm and beneath eastern Halmahera, immediately behind
central Halmahera, there is no evidence for an the arc, resulting in rapid subsidence of the
Oligo-Miocene volcanic arc. There are no volcanic Miocene reef limestones. The sedimentary basin
rocks in the sequence and talc-alkaline volcanic formed was filled by an overall coarsening-up-
debris is notably absent from the late Paleogene wards sequence with an increasing volcaniclastic
and Neogene carbonates and the siliciclastic de- component marking shallowing of the basin and
bris which they contain indicates derivation from increasing arc activity with lavas and sub-aerial
the underlying Halmahera ophiolitic Basement volcanic breccias and conglomerates at the highest
Complex. Oligocene-Early Miocene volcanism is levels. The Pliocene arc was built on the eroded
reported from Waigeo (Van der Wegen, 1963) and basement of the early Tertiary arc and the posi-
Bacan (Yasin, 1980; Silitonga et al., 1981). Both of tion in which the lithosphere fractured, leading to
these islands are close to the Sorong Fault system subduction of the Molucca Sea Plate, was prob-
which is a transform fault zone with a history of ably determined by the thickened crust beneath
volcanic activity (Morris et al., 1983; Dow and the older arc.
Sukamto, 1984) and Bacan is situated on a splay A major deformation event affected the arc in
of the Sorong Fault which is marked by recent the mid-Pleistocene. This caused tilting of major
volcanic activity. It is more probable that Oligo- fault-bounded blocks, intense folding with local
Miocene volcanic activity was related to transform overthrusting at the junction of the western arc
fault motion to the south of Halmahera at the and eastern back-arc regions, uplift and erosion.
Pacific-Australian plate boundary rather than to Arc volcanism ceased briefly and the present ac-
a volcanic arc on Halmahera. tive arc was built unconformably on the deformed
At the end of the Eocene the arc and fore-arc older rocks after a shift of position of the volcanic
 342

arc westwards by about 30 km in central Pacific (Hayes and Lewis, 1984) and may be re-
Halmahera. There is no gap in the seismicity lated to a major change in motion direction of the
beneath Halmahera (Cardwell et al., 1980) to sug- Pacific Plate at 40 Ma (Uyeda and Ben-Avraham,
gest a cessation of subduction in the Late Pliocene 1972). A second completely new result of the
or Pleistocene and this deformation event prob- Halmahera field-work is the discovery that the
ably marks a tectonic event at one of the existing widely-quoted east-facing Oligocene-Early Mio-
or developing plate margins in the region. The cene Halmahera island arc, which is shown on all
nature of this event is considered below. pre-5 Ma reconstructions and attributed to Ham-
ilton (1979) has no evidence to support it. Volcanic
The evolution of the region rocks of Oligo-Miocene age are lacking on Halma-
hera, although they are present on Waigeo and
Previous reconstructions of the geological Bacan indicating volcanism along the transform
evolution of the region have been based princi- fault system separating the Pacific from Australa-
pally on recent seismicity and marine geophysics sia.
(Hamilton, 1979; Cardwell et al., 1980; Moore These two discoveries mean that the East
and Silver, 1983). This reconstruction is based on Mindanao-Halmahera ridge (EMH Ridge) must
the new data from Halmahera and information have been a continuous feature within a single
from the literature on surrounding regions. A plate, the Philippine Sea Plate, since the Early
fundamental constraint is the model of a Molucca Oligocene. As the Molucca Sea has been sub-
Sea slab dipping both east and west, first pro- ducted, the EMH Ridge has moved westwards
posed by Hatherton and Dickinson (1969) and relative to West Mindanao. The western margin of
refined by Cardwell et al. (1980). The only signifi- the Molucca Sea, the West Mindanao-Sangihe-
cant difference between the present-day plate north Sulawesi arc, has been an active margin
boundaries adopted here and those of Cardwell et since the Miocene; the age of this arc is discussed
al. (1980) is in the omission of their Halmahera- below. Collision between West and East Min-
Pulau transform fault which is discussed further danao is thought to have occurred in the Late
below. Information on the pre-Miocene geology of Miocene (Cardwell et al., 1980; Moore and Silver,
the region is still very limited and it is not yet 1983). The 10 Ma reconstruction (Fig. 4) shows
possible to attempt a reconstruction for the period the position of the EMH ridge and adjacent areas
before the Miocene. just before collision occurred in Mindanao. It
differs principally from earlier reconstructions
Initial configuration: approximately IO Ma (Cardwell et al., 1980; Moore and Silver, 1983) in
joining Halmahera and East Mindanao and in the
An important new result is the demonstration absence of subduction under Halmahera; both
of a link between the Basement Complex of east- features are required by the field evidence. The
ern Halmahera and basement rocks of eastern EMH Ridge separated the Philippine Sea from the
Mindanao. East Mindanao and Halmahera are Molucca Sea, but all of these were moving at this
part of a region of thickened crust representing a time as part of a single plate with the western
Late Cretaceous-Early Tertiary arc terrain which margin at ABC. AB represents the Philippine
originated within the Pacific region. It was de- Fault. Before the Late Miocene this was an active
formed and uplifted during the major plate re- margin, with subduction westwards beneath West
organization event in the Late Eocene-Early Mindanao, representing a continuation of the
Oligocene, possibly by collision with the north Sangihe Trench. BC represents the trench associ-
edge of New Guinea and represents the continua- ated with subduction beneath the Sangihe-North
tion of the Papuan arc terrain (Kroenke, 1983) Sulawesi volcanic arc (the Sangihe Trench). DE
from western New Guinea. This deformation event represents the western margin of East Mindanao.
is recognizable over a widespread region between EF represents a former offset of the Molucca Sea
New Guinea (Kroenke, 1983) and the western margin. FG represents the site of the trench asso-
 343

reconstructions but does mean that the rate of

subduction will have changed and this is discussed
further below.
(2) Volcanism on Halmahera began in the
Pliocene and a date of 3 Ma for the age of initial
volcanism has been taken for ease of calculation.
Volcanism is assumed to have begun when the
slab reached 100 km. All the volcanoes of the
CELEBES
present Halmahera and Sangihe active arcs are
situated more than 100 km above the Benioff
zone. The present length of the slab beneath
Halmahera is the entire length of subducted litho-
sphere and is approximately 250 km. Assuming a
constant subduction rate this means subduction
10 Ma began at 5 Ma (Early Pliocene). These numbers
/ Active subduction zone x Strike-slip fault
,&” can be adjusted but are consistent with the stratig-
/ Inactive subduction zone Active volcanic arc
,.aye
raphy of Halmahera. The rate of subduction is
Fig. 4. Reconstmction of Hahnahera region at 10 Ma. Western discussed further below.
boundary of Philippine Sea Plate is subduction zone between (3) Subduction at the Philippine Trench (in the
West Mindanao and North Sulawesi. See text for discussion.
sector between 8” N and 2” N) is considered to
have been initiated at the same time as subduction
at the Halmahera Trench following collision in
ciated with subduction beneath the Hahnahera arc Mindanao. The justification for this is that a
(the Halmahera Trench). The southern edge of the collisional event is most likely to initiate new
region is the Sorong Fault. The north arm of subduction zones, the timing of initiation of sub-
Sulawesi is shown about 200 km further south of duction beneath Halmahera is now reasonably
West Mindanao than at present; this takes account well known and on this time scale the rate of
of subduction at the North Sulawesi Trench. North subduction required at the Philippine Trench is
Sulawesi is not shown on the reconstructions be- reasonable. The precise date of initiation of sub-
cause the date of initiation of subduction and the duction at the Philippine Trench is unknown; a
rate of subduction are unknown. young age for the southern Philippine Trench is
discussed by Cardwell et al. (1980) and is con-
Assumptions in the reconstructions sistent with the lack of associated volcanism.
(4) All the slabs are assumed to be dipping at
(1) The relative direction of motion of the approximately 45“ perpendicular to the trench
Philippine Sea Plate is assumed to have remained axis. Cardwell et al. (1980) show that the dip on
constant for the last 6 Ma (Seno and Maruyama, the Philippine Sea slab in the Talaud region at
1984). The present vectors for the Philippine Sea about 7O N is steeper than for other island arcs,
Plate are given by Ranken et al. (1984). Two consistent with a young age for the subduction
features are particularly important (Fig. 1): (a) the zone in this sector. Further north the dip is less
high degree of obliquity with which the Philippine steep and Quatemary volcanoes are situated above
Sea Plate approaches the Philippine Trench, and the 100 km Benioff zone contour indicating that
(b) the increase in relative rates of slip moving the northern sector is older and the Philippine
southwards along the Philippine Trench. Naka- Trench is propagating southwards.
mura et al. (1984) suggest that the relative direc- (5) West Mindanao is held in a fixed position
tion of convergence between the Philippine Sea on all the diagrams as a point of reference.
Plate and the Eurasian Plate changed at about 1 (6) No subduction is shown under the Sangihe
Ma; such a change will not alter the form of the Trench after 6 Ma. It is proposed that the rate of
344

subduction slowed significantly after the collision problems of defining the age and movement
of East and West Mindanao, possibly to zero, amounts on the Philippine Fault. The effect of this
until quite recently. There is a large gap in the collision was to stop subduction beneath West
seismicity below West Mindanao between 100 and Mindanao ( AB. Fig. 4) and to severely hamper
600 km consistent with collision terminating sub- continued subduction of the Molucca Sea at the
duction in the North Sulawesi-Sangihe-West Sangihe Trench (BC, Fig. 4). Therefore, continued
Mindanao arc. The effect of including subduction movement of the Philippine Sea Plate was accom-
at the Sangihe Trench after 6 Ma is to widen the modated by (a) strike-slip motion on Philippine
Molucca Sea and move East Mindanao further Fault, (b) subduction of Philippine Sea westwards
south. at the Philippine Trench, (c) subduction of the
Molucca Sea eastwards at the Halmahera Trench.
Model Because of the angle between the relative motion
direction of the Philippine Sea Plate and the
(1) Consumption of the Molucca Sea began at Philippine Trench, subduction of a given length of
the Sangihe Trench. The exact timing is not known Molucca Sea lithosphere was equivalent to signifi-
although the age of volcanic rocks from North cantly less subduction at the Philippine Trench
Sulawesi indicates 12 Ma or earlier. (Fig. 6). The equivalent lengths of the subducted
(2) Collision occurred between West Mindanao slabs depend on (a) the orientations of the Philip-
and East Mindanao (Fig. 5) in the Late Miocene pine Trench and the Halmahera Trench. (b) the
(9-10 Ma). For simplicity the margins of West motion vector for the Philippine Sea Plate, and (c)
and East Mindanao at the trench (AI?, Fig. 4) and the dip on each slab. Assuming the only variable
the EMH Ridge (DE, Fig. 4) are assumed to have to be the relative orientation of the trenches, ap-
been parallel. This collision may have been di- proximately 100 km of subduction at the
achronous if the EMH ridge was oriented ob- Halmahera Trench is unexplained (equivalent to
liquely to the trench and may explain the apparent about 140 km of horizontal lithosphere before
subduction). This can be accounted for by strike-
slip motion on the Philippine Fault. Strike-slip
motion on the Philippine Fault can also account
for any Molucca Sea lithosphere which was sub-
ducted in the same period at the Sangihe Trench.
There are ways of satisfying the apparent dif-
ference of 100 km other than, or in addition to,
strike-slip motion on the Philippine Fault. Small
EURASIAN changes in the relative convergence direction of
the Philippine Sea Plate and the Southeast Asian
PLATE
Plate, or changes in the dip on the subduction
zones could reduce the figure. If the convergence
rate was greater in the region of Halmahera than
further north, the given length of Molucca Sea
lithosphere subducted would be equivalent to even
less Philippine Sea lithosphere subducted; because
of the position of the pole of rotation (Ranken et
al., 1984) this may account for up to half of the
100 km. However, if the Philippine slab dipped
5 Ma more steeply than the Halmahera slab, the oppo-
site will be true, although the effect would be
Fig. 5. Reconstruction of Halmahera region at 5 Ma. Late
Miocene collision in Mindanao results in initiation of subduc- smaller. Because of their uncertainty these effects
tion at Halmahera Trench. have been omitted in making the reconstructions.
 345

120 Depth of slob km. 1

Eurarian

A)_

Plate

equivalent to_fOkm

Fig. 6. Cartoon of plate boundaries between Mindauao and HaImahera during process of eastward shift of boundary of Philippine
Sea Plate. Convergence of Philippine Sea Plate and Eurasian Plate was accommodated by (A) strike-slip motion on Philippine Fault
and subduction at Philippine Trench, (B) subduction at Halmahera Trench and Philippine Trench, and (C) subduction at Halmahera
Trench.

(3) Collision of the EMH ridge with West northwestern edge, new subduction zones devel-
Mindanao in the Late Miocene was responsible oped wirhinthe Philippine Sea Plate (Fig. 6). In
for blocking further significant subduction at the the north this was most easily achieved by the
Sangihe Trench. For the period from about 5-6 propagation of the existing Philippine Trench
Ma to the late Pleistocene, there was very little southwards (the sector of the trench south of 8” N
subduction at the Sangihe Trench. The renewal of is considered to be much younger than the north-
active volcanism was associated with the final ern sector). At the southern edge of the P~lippine
phase of subduction of the Molucca Sea as Sea Plate subduction was initiated in the opposite
Molucca Sea lithosphere was forced beneath the direction, on the opposite side of the EMH ridge.
colliding plates above. Although active volcanism Here, initiation of a new trench was facilitated by
continued along the Sangihe arc, it is noteworthy the higher rate of convergence (currently > 10
that volcanism is dying out along the arc from N cm/yr according to Ranken et al., 1984) because
to S (Morrice et al., 1983). This is consistent with the Philippine Sea Plate is effectively rotating
a gradual change of the plate boundary between relative to lithosphere to the west of it.
Mindanao from subduction to strike-slip (Fig. 6, (5) A progressive development of trenches com-
B). Minor subduction continuing at the Sangihe bined with strike-slip faulting at the Philippine
Trench would be equivalent to some of the strike- Fault allowed a continuity of motion of the Philip-
slip motion on the Philippine Fault. pine Sea Plate relative to the Molucca Sea Plate
(4) Because of the continued motion of the and Eurasian-Southeast Asian Plate. Fitch (1972)
Philippine Sea Plate, which was impeded at its suggested that the oblique convergence at the
346

Philippine Trench can be decoupled into strike-slip

motion on the Philippine Fault and normal under-
thrusting at the Philippine Trench. In contrast,
Seno (1977) suggested that the Philippines behave
as a single lithospheric plate with the Philippine
Fault as an active intra-plate fault with a low slip
rate. The suggestion of Fitch (1972) is accepted
here, which is supported by determination of re-
cent seismic slip rates on the Philippine Fault
(Acharya, 1980) while recognizing that the Philip- Z
a
pine Islands probably represent a zone of complex
deformation which is not behaving as a simple
plate system (Cardwell et al., 1980; Hamburger et
al., 1983; Hayes and Lewis, 1984; C. Rangin, pers.
commun., 1986).
The most complex part of the Philippine Sea
c begins on ____I
Holinahero
Plate is that between Halmahera and East
Mindanao, approximately the area BEF (Fig. 4). 3 Ma
Transform faulting may be required to allow the Fig. 7. Reconstruction
southern and northern sections of the plate to
move relative to one another. An obvious feature
of the reconstruction is the way in which the
Halmahera Trench migrates westwards relative to
the Philippine Trench. This migration could be explanation of why the Snellius Ridge has the
accounted for by propagation of the Philippine characteristics of a subsided volcanic arc, is con-
Trench southwards with time; however, it may be tinuous with the Halmahera Arc, but volcanism
necessary to postulate the existence of a transform has ceased. Above the Philippine Trench there was
fault between the two subduction systems. Such a no volcanism because (a) oblique subduction
feature has been suggested by McCaffrey (1982). meant the slab was not very deep and (b) some of
Note that if the initiation of the Philippine Trench the convergence was being taken up by strike
is unrelated to that of the Halmahera Trench, a slip-motion on the Philippine Fault.
transform of even greater significance must be (7) At about 1 Ma the junction zone between
postulated. The reconstruction of Cardwell et al. the arc and the back-arc was deformed, arc
(1980) shows such a transform between Palau and volcanism ceased briefly and the active arc shifted
Halmahera. If no transform fault is postulated it westwards. There are several possible explanations
is necessary to accommodate between 0 and 100 of this event. Internal deformation of the Philip-
km of lithosphere shortening in the area BEF pine Sea Plate is required, as explained above, if
(Fig. 4). This could be achieved by internal defor- there is no transform fault linking the Halmahera
mation such as folding, strike-slip faulting and subduction zone and the Philippine Trench.
thrusting, and this may be the cause of the 1 Ma Nakamura et al. (1984) suggest a change in mo-
deformation event on Halmahera which clearly tion direction of the Philippine Sea Plate at 1 Ma
postdates the initiation of subduction. which may also be responsible. Other possibilities
(6) At about 3 Ma the slab beneath Halmahera include interaction between the east-dipping
reached 100 km and volcanism began (Fig. 7). To Molucca Sea lithosphere with adjacent plates. The
the north of Halmahera (beneath the Snellius Sorong Fault system forms the southern boundary
Ridge), because the trench was oblique to the to the Molucca Sea region, and north of the splay
convergence direction, the slab was less deep at 3 passing through Bacan the Quatemary arc (Fig. 1)
Ma and no volcanism resulted. This offers an follows the 100 km Benioff zone contour, whereas
 34-l

on Bacan the direction of the line of Quaternary

volcanoes trend almost at right angles to this 5 ::.
y
:::::::::::::::::::::
-:::::::::
jf;::
w :: q
yyzx::::::::::::::::::::::::I- ..
contour (Cardwell et al., 1980; Morris et al., 1983). _ \II;i ::::‘: :::::: +:::r
............::::::::::::::::::::::::::::
m :::w
The shift in position of active volcanicity could
therefore mark westwards motion of the continen- I z
’ <
tal fragment south of this splay dragging the east-
dipping slab of the Molucca Sea lithosphere west- a
0:
-I
wards. Such motion would steepen the east-dip- ul

ping limb of the Molucca Sea Plate (Cardwell et

Acfive volcanism renewed ‘iq
al., 1980) and shift the active arc westwards about in Songihe Arc 1;;1
.:---.
a pivot in north Halmahera (Fig. 2). An altema- during final stages of /?d period of volcomsm
q.Kz:::: becomesinactive
tive location is beneath northern Halmahera where
the east-dipping Molucca Sea Plate and the west-
dipping Philippine Sea Plate are in collision at a
depth of 100-150 km. The relative lengths of the
subducted slabs suggest that the Philippine Sea
PLATE collide beneath Holmohero
Plate collided with the Molucca Sea Plate beneath
northern Halmahera at about 1 Ma and the likely
effect of such a collision would be to force the
0.5 Ma
east-dipping Molucca Sea Plate westwards, thus Fig. 8. Reconstruction of Halmahera region at OSMa. Colli-
shifting the axis of volcanicity in the same direc- sion between Philippine Sea Plate and Molucca Sea Plate
beneath Halmahera results in westwards shift in position of
tion (Fig. 2).
Halmahera volcanic arc.
(8) By 0.5 Ma the whole of the zone between
Mindanao and Talaud was a zone of strike-slip
movement (Fig. 8). As the Molucca Sea closed the
Philippine Fault propagated southward as a 1 EURASIAN ::
\n J.
strike-slip fault and the Philippine Trench propa-
gated southward as a new subduction zone. This
accounts for the very young age of subduction at
the Philippine Trench in the Talaud section, as
suggested by the steep dip on the Philippine Sea
slab. It also explains the lack of significant defor-
mation of Neogene sediments shown on seismic fo N.E. ond S.

reflection profiles from the Davao Gulf, between

the Talaud and Sangihe Ridges.
(9) At present the Molucca sea is effectively
completely closed (Fig. 9). This is resulting in
arc-directed shallow thrusting in the Molucca Sea
(Silver and Moore, 1978). The field-work suggests
that blocking of subduction in the Halmahera
Trench may also be resulting in motion on strike-
slip faults in the northeast arm of Halmahera in
Present
an attempt to accommodate continued motion of
Fig. 9. Present tectonic setting in Halmahera region. Collision
the Philippine Sea Plate. This suggestion is sup-
between Halmahera and Sangihe forearcs results in thrusting
ported by seismicity in the northeast arm which of Molucca Sea collision complex onto the opposed forearcs.
also indicates strike-slip motion (I. Effendi and The Philippine Trench may now be linked to Sorong Fault
S.A.F. Murrell, pers. commun., 1986). zone by strike-slip faults through Halmahera.
Discussion volcanism after 5 Ma: exactly as predicted by the
model, which suggests that subduction effectively
Age of volcanism in the North Sulawesi-Sangihe- ceased at the Sangihe Trench from about 5--6 Ma.
West Mindanao arc On Mindanao Ranneft et al. (1960) record changes
in sedimentation from the late Miocene, but these
The model assumes that arc volcanism began may be related to collision rather than subduction
before 10 Ma. In north Sulawesi the oldest volcanic further south.
rocks reported (Effendi, 1976; Apandi, 1977) are
Early to Middle Miocene in age and Dow (1976)
Rate of subduction
records Late Miocene quartz-diorites intruding
older Miocene intermediate and volcanic rocks.
Further north less is known; Morrice et al. (1983) Using an average rate of subduction may be
assume that the Sangihe arc formed part of a misleading. The rate of subduction calculated at
single arc which stretched from the west arm of the Halmahera subduction zone is 7 cm/yr. Even
Sulawesi to Mindanao and quote Hamilton (1979) assuming about 100 km of subduction at the
as noting that volcanism was “...especially volu- Sangihe Trench for the same time period results in
minous from 5 to 14 Ma...“. These dates include a convergence rate of about 9 cm/yr. The present
those from the south arm, west-central Sulawesi rate of convergence is about lo-11 cm/yr at the
and the south end of the north arm, and range Philippine Trench (Ranken et al., 1984). There is
between 1.6 and 31 Ma, and although Hamilton some evidence to suggest a variation in the rate of
does note that volcanism was especially subduction. Nakamura et al. (1984) have sug-
voluminous between 14 and 5 Ma, it is not clear gested a change in the direction of relative motion
where in this large region this was so. The con- between the Eurasian and Philippine Sea Plates at
tinuation of this arc into West Mindanao is uncer- about 1 Ma. If the entire region west of the
tain; Ranneft et al. (1960) report lavas and volcanic Philippine Fault and Molucca Sea was part of a
breccias of probable Early Miocene age. single Eurasian Plate, which is uncertain, the rate
of subduction would have increased since before 1
Timing of subduction Ma the Philippine Sea Plate was moving in a
direction almost parallel to the Philippine Trench.
The model suggests that subduction beneath The Pleistocene unconformity below the present
Halmahera began at about 5 Ma. This is based on active arc volcanic rocks on Halmahera may be an
an estimate of the initiation of volcanism begin- expression of such a change in the subduction rate
ning at 3 Ma from the stratigraphic evidence, an (see above for other explanations). On North
assumption of a continuous and steady rate of Sulawesi the currently active volcanoes are being
subduction, and an assumption that volcanism built on a Miocene substratum, consistent both
began when the subducted slab reached 100 km. with discontinuous subduction and the suggestion
This interpretation is supported by stratigraphic that subduction virtually ceased at the Sangihe
evidence from Halmahera, where there was an end Trench between 5-6 Ma until very recently. Two
to the deposition of reef carbonates and a marked intervals of higher rates of subduction (say about
change in the rate of subsidence and a change to 5-2 Ma and 1-O Ma) would be consistent with
siliciclastic sedimentation. Equally, from the little the present rate of convergence and length of
that is known from adjacent regions a 5-6 Ma subducted lithosphere. An earlier interval of rapid
event seems plausible. In the Ayu Trough, im- subduction would explain how a velocity required
mediately east of Halmahera, there is a marked to break the Molucca Sea Plate and initiate sub-
decrease or cessation of spreading at 5-6 Ma duction at the Halmahera subduction zone could
(Weissel and Anderson, 1978). Volcanism in the be achieved (N.J. Price, pers. commun., 1986). It
Sulawesi arc was voluminous between 14-5 Ma should be possible to refine this aspect of the
(see above). There is thus an implication of fess model by precise dating of volcanic rocks.
 349

The Philippine Fault arcs (Sukamto, 1979), and part of the fore-arc of
the west-facing Halmahera arc (Moore and Silver,
The model predicts left-lateral motion on the 1983). The reason for these disparate interpre-
Philippine Fault with an average rate of 0.5 cm/yr tations is the evidence of collision in the southern
(the range could be O-l.0 cm/yr). The Philippine Molucca Sea with evidence for its absence north
Fault is still a region of considerable controversy. of Talaud. On Mindanao Moore and Silver (1983)
Hamilton (1977, 1979), apparently on the basis of show that the Agusan-Davao trough separates the
aerial photograph examination, states that the East and West Mindanao arcs and contains up to
Philippine Fault does not reach Mindanao “...as 6 km of Eocene to Recent sediments with a period
an active strike-slip feature, and is not now a of intense folding at the end of the Middle
major plate boundary”. In contrast, Karig (1983) Miocene. They trace the Philippine Fault system
states that the northern section is and has been southwards as far as the Talaud islands where
active “ . ..for much of the past 15 m.y., despite thrusts in the southern Molucca Sea are
arguments to the contrary...” by Rutland (1968) terminated. North of Talaud there are little-de-
and Hamilton (1979). Interestingly, in Luzon de- formed Neogene sediments in the Davao Gulf
tailed mapping by Rutland indicated three im- (Cardwell et al., 1980) not consistent with the very
portant phases of movement: Late Miocene, Plio- recent convergence in this area and west of Talaud
Pleistocene and Recent. Cardwell et al. (1980) also sediments lap onto the Talaud ridge indicating the
dispute Hamilton’s assertion that the fault does fault block is not a young feature (Moore and
not extend into Mindanao on the basis of geologi- Silver, 1983). On Talaud Moore et al. (1981) de-
cal mapping (Ranneft et al., 1960; Philippine scribe a post-Middle Miocene sedimentary se-
Bureau of Mines, 1963), airphoto investigations by quence resting unconformably on ophiolite slabs
Allen (1962), and the presence of large earth- which are Middle Eocene or older. The Middle
quakes on or near the fault in 1893 and 1911 Miocene to Pleistocene rocks are tuffaceous sand-
(Rowlett and Kelleher, 1976). Acharya (1980) re- stones, siltstones and shales with intercalations of
cords recent seismic slip rates of about 6 cm/yr. limestone, marl and conglomerate in which the
Moore and Silver (1983) report field observations sediments are dominated by volcanic debris. They
from southern Mindanao consistent with left- were deposited in deep water (> 2000 m) by
lateral strike-slip motion on the Philippine Fault. turbidity currents. Volcanic rocks occur only in
In the north Philippines Karig (1983) estimates the lower part of the Talaud sequence where they
the average rate of movement on the fault to have are of probable mid-Miocene age. The Basement
been near 1.5 cm/yr from the mid-Miocene to Complex of Halmahera is similar to that of Talaud
Recent. Considering the uncertainties, particularly but the Neogene sequence is quite different.
the inadequacies of data from Mindanao, this is Volcanic activity in the Halmahera arc did not
not inconsistent with the model. begin the Pliocene and volcanic rocks in-
until
crease in abundance as the sequence becomes
Talaud and its significance younger. This rules out an origin for Talaud in the
Halmahera arc.
The Talaud ridge at the centre of the Molucca These apparently conflicting observations can
Sea has previously been interpreted on the basis of be reconciled if Talaud formed part of the Miocene
marine geophysics as part of a “melange wedge” Sangihe arc which became inactive in the Late
(Silver and Moore, 1978; McCaffrey et al., 1980) Miocene according to the model. After the
including slices of the Molucca Sea lithosphere. Miocene arc became inactive, it shed debris into
Proposals for the origin of Talaud include part of the Molucca Sea to areas which subsided to
a west-facing East Mindanao-Talaud arc separate depths of several kilometres. This important ero-
from the Halmahera arc (Cardwell et al., 1980; sional break is recorded in the north arm of
Moore et al., 1981), an east-facing Talaud-Tifore Sulawesi where currently active volcanoes are built
arc independent of the Sangihe and Halmahera upon Miocene volcanics and plutonic rocks. The
northern Molucca Sea therefore included an inae- has the interesting feature that it corresponds to
tive arc on its western side, and along the Talaud the known age of spreading of many marginal
ridge strike-slip faulting on the Philippine Fault basins in the region, including the Caroline Sea,
system continued southwards into the Halmahera the South China Sea. the Iapan Sea, the Parece
subduction zone. Activity in the Sangihe arc has Veia Basin and the Shikoko Basin (Hayes and
been renewed only very recently (possibly west of Lewis, 1984). It also corresponds to a period
its Miocene position) and Talaud has been up- without arc volcanicity consistent with models of
lifted by the closure of the Molucca Sea. arc evolution (K&g, 1974; Scott and Kroenke,
1981). Alternatively the Molucca Sea may repre-
sent Eocene ocean crust like the Celebes Sea
(Weissel, 1983) and the West Philippine Sea.
It is interesting to note that all of the “melange”
reported by Moore et al. (1981) is of pre-mid Terrain accretion
Miocene age; none appears to be related to the
present collision. Indeed they note (p. 472) that Many erogenic belts incfude ~l~hthonous ter-
the Neogene “... strata are moderately to strongly rains although the mechanisms by which these
deformed, but unlike the melange zones, bedding terrains are assembled are still not understood.
is well preserved and becomes disrupted only lo- Karig et al. (1986) have suggested that strike-slip
cally adjacent to melange zones”. The similarity in motion has played an important role in the trans-
character and structure of the ophiolitic rocks to port and emplacement of terrains in the northern
those forming the Basement Complex on Halma- Phihppines. The development recorded by Halma-
hera suggests that much of the “melange” may hera and the interpretation suggested here is that
have been formed in the Late Eocene deforma- the EMH ridge has been in the process of trans-
tional event. If the present model is correct in ferring from the Philippine Sea Plate to the
suggesting the gradual change from thrusting to strike-&p complex of the Philippines at the junc-
strike-slip motion the juxtaposition of ophiulitic tion of the Eurasian and Philippine Sea plates.
basement and Neogene rocks may in part be a This process has involved a complex sequence of
consequence of the strike-slip deformation. It is events including strike-slip faulting and subduc-
noticeable that the maps and descriptions of tion in several stages and will be completed by
Moore et al. (1981) show steeply-dipping fault- either strike-slip faulting through the northeast
bounded slices of material in the “melange”, as on arm of Halmahera (Fig. 9) or by continued south-
Halmahera. It is likely that the Molucca Sea ward propagation of the Philippine Trench. when
“melange wedge” is composed largely of sedimen- this is complete the EMH ridge will be another of
tary material dominated by volcanic debris de- the allochthonous terrains of the Philippines. The
rived by erosion of the Miocene Sangihe arc, the complexity of events recorded during the last 10
Pliocene Halmahera arc, and material shed from Ma in the Hahnahera region confirms the warning
the present active arcs, which is now becoming of Karig et al. (1986) that certain geological keys
locally tectonically intercalated with its pre- should be used more circumspectly and supports
Miocene ophiolitic basement. their contention that current views of terrain
evolution need to be broadened from simple two-
dimensional models to mure complex models.

The age of the Molucca Sea is unknown. In Acknowledgements

view of the similar character of the Talaud and
Halmahera basements, it is possible that the Financial support for the field-work in In-
Molucca Sea formed by rifting of a formerly ex- donesia was provided by the Royal Society, the
tensive and continuous terrain in a short intervd Central Research Fund of London University,
of spreading in the Oligocene-Early Miocene. This Amoco International, British Petroleum and the
 351

University of London Consortium for Geological of Hahnahera, eastern Indonesia: initiation of a volcanic

Research in Southeast Asia. GRDC Bandung pro- island arc. J. Geol. Sot. London, in press.
Hamburger, M.W., Cardwell, R.K. and Isacks, B.L., 1983.
vided aerial photographs and invaluable practical
Seismotectonics of the Northern Philippine island arc. In:
assistance in Indonesia. I thank Mike Audley- D.E. Hayes (Editor), The Tectonic and Geologic Evolution
Charles for assistance and discussion during and of South-East Asian Seas and Islands, part 2. Am. Geo-
after field-work, Syarif Hidayat and Sahat L. Tob- phys. Union, Monogr., 27: l-22.
ing for their assistance with field-work, and Paul Hamilton, W.H., 1977. Subduction in the Indonesian region.
In: M. Talwani and W.C. Pitman (Editors), Island Arcs,
Ballantyne and F.T. Banner for their work on the
Deep Sea Trenches and Back Arc Basins. Am. Geophys.
material collected in the field.
Union, Maurice Ewing Ser., 1: 15-31.
Hamilton, W.H., 1979. Tectonics of the Indonesian region.
References U.S. Geol. Surv., Prof. Pap., 1078: 345 pp.
Hatherton, T. and Dickinson, W.R., 1969. The relationship
Acharya, H.K., 1980. Seismic slip on the Philippines Fault and between andesitic volcanism and seismicity in Indonesia,
its tectonic implications. Geology, 8: 40-42. the Lesser Antilles, and other island arcs. J. Geophys. Res.,
Allen, C.R., 1962. Circum-Pacific faulting in the Phihppines- 74: 5301-5310.
Taiwan region. J. Geophys. Res., 67: 4795-4812. Hayes, D.E. and Lewis, S.D., 1984. A geophysical study of the
Apandi, T., 1977. Geologic map of the Kotamobagu quadran- Manila trench, Luzon, Philippines. 1. Crustal structure,
gle, North Sulawesi. Indones. Geol. Res. Develop. Centre. gravity and regional tectonic evolution. J. Geophys. Res.,
Apandi, T. and Sudana, D., 1980. Geologic map of the Temate 89: 9171-9195.
quadrangle, North Maluku. Indones. Geol. Res. Develop. Karig, D.E., 1974. Evolution of arc systems in the western
Centre. Pacific. Annu. Rev. Earth Planet. Sci., 2: 51-75.
Bessho, B., 1944. Geology of the Halmahera islands. J. Geogr., Karig, D.E., 1983. Accreted terrains in the northern part of the
(LVI) 664: 145-203 (in Japanese). Philippine archipelago. Tectonics, 2: 21 l-236.
Brouwer, H.A., 1923a. Bijdrage tot de geologic van het eiland Karig, D.E., Sarewitz, D.R. and Haeck, G.D., 1986. Role of
Batjan. Jaarb. Mijnbouw Ned, Ind., 1921, Verh. II: 73-105. strike-slip faulting in the evolution of allochthonous ter-
Brouwer, H.A. 1923b. Geologische onderzoekingen op het ei- ranes in the Philippines. Geology, 14: 852-855.
land Halmaheira. Jaarb. Mijnbouw Ned. Ind., 1921, Verh. Kroenke, L.W., 198’3. Cenozoic development of the southwest
II: 3-72. Pacific. United Nations Economic and Social Commission
Burgath, K., Mohr, M. and Simandjuntak, W., 1983. New for Asia and the Pacific, Committee for Co-ordination of
discoveries of blueschist metamorphism and mineral occur- Mineral Prospecting for Mineral Resources in South Pacific
rences in the Halmahera-Gag ophiolite belt, East In- offshore areas (CCCOP/SOPAC) Tech. Bull., 6: 122 pp.
donesia. Bull. Dir. Min. Res. Indonesia, 13: 1-19. Kuenen, P.H., 1935. Contributions to the geology of the East
Cardwell, R.K., Isacks, B.L. and Karig, D.E., 1980. The spatial Indies from the Snellius expedition. Part I. Volcanoes.
distribution of earthquakes, focal mechanism solutions, and Leidsche Geol. Meded., 7(2): 273-331.
subducted lithosphere in the Philippine and northeastern McCaffrey, R., 1982. Lithospheric deformation within the
Indonesian islands. In: D.E. Hayes (Editor), The Tectonic Molucca Sea arc-arc collision: evidence from shallow and
and Geologic Evolution of South-East Asian Seas and intermediate earthquake activity. J. Geophys. Res., 87:
Islands. Am. Geophys. Union, Monogr., 23: l-35. 3663-3678.
Dow, D.B. and Sukamto, R., 1984. Western Irian Jaya: the McCaffrey, R., Silver, E.A. and Raitt, R.W., 1980. Crustal
end-product of oblique plate convergence in the late Ter- structure of the Molucca Sea collision zone, Indonesia. In:
tiary. Tectonophysics, 106: 109-139. D.E. Hayes (Editor), The Tectonic and Geologic Evolution
Dow, J., 1976. Porphyry copper exploration in Block II, of South-East Asian Seas and Islands. Am. Geophys.
Sulawesi. Indones. Geol. Res. Develop. Centre Newslett., Union, Monogr., 23: 161-177.
8(33): 6. Moore, G.F. and Silver, E.A., 1983. Collision processes in the
Effendi, A.C., 1976. Geologic map of the Manado quadrangle, Northern Molucca Sea. In: D.E. Hayes (Editor), The
North Sulawesi. Indones. Geol. Res. Develop. Centre. Tectonic and Geologic Evolution of Southeast Asian Seas
Fitch, T.J., 1972. Plate convergence, transcurrent faults and and Islands, Part 2. Am. Geophys. Union, Monogr., 27:
internal deformation adjacent to South-east Asia and the 360-372.
Western Pacific. J. Geophys. Res., 77: 4432-4460. Moore, G.F., Kadarisman, D., Evans, C.A. and Hawkins, J.W.,
Hall, R., Audley-Charles, M.G., Banner, F.T., Hidayat, S. and 1981. Geology of the Talaud Islands, Molucca Sea collision
Tobing, S.L., 1987a. Basement rocks of the Halmahera zone, northeast Indonesa. J. Struct. Geol., 3: 467-475.
region, eastern Indonesia: a Late Cretaceous-early Tertiary Morrice, M.G., Jezek, P.A., Gill. J.B., Whitford, D.J. and
fore-arc. J. Geol. Sot. London, in press. Monoarfa, M., 1983. An introduction to the Sangihe Arc:
Hall, R., Audley-Charles, M.G., Banner, F.T., Hidayat, S. and volcanism accompanying arc-arc collision in the Molucca
Tobing, S.L., 1987b. Late Paleogene-Quatemary geology Sea, Indonesia. J. Volcanol. Geotherm. Res., 19: 135-165.
 3.52

Morris, J.D., Jezek, P.A., Hart, S.R. and Gill, J.B.. 1983. The Silver, E.A. and Moore, J.C.. 1978. The Molucca Sea coilisron
Halmahera island arc, Molucca Sea collision zone, in- zone, Indonesia. J. Geophys. Res., 83: 1681-1691.
done% a geochemical survey. In: D.E. Hayes (Editor). Soerra Atmadja, R., 1981. Ophiolites in the Halmahera paired
The tectonic and geologic evolution of South-East Asian belts, East Indonesia. in: A.J. Barber and S. Wiryosuyonn
Seas and Islands, Part 2. Am. Geophys. Union, Monogr.. (Editors). The Geology and Tectonics of Eastern Indonesia.
23: 313-381. Spec. Publ. Indones. Geoi. Res. Develop. i‘entre. 2:
Nakamura, K.. Shimazaki, K. and Yonekura. N., I984. Sub- 363-372.
duction, bending and eduction. Present and Quatemary Sukamto, R., 1979. Tectonic significance of melange on the
tectonics of the northern border of the Philippine Sea plate. Talaud Islands northeastern Indonesia. Bull. Indones. Geol.
Bull. G&o]. Sec. Fr., (7), 26: 221-243. Res. Develop. Centre. 2: 7-19.
Philippine Bureau of Mines, 1963. Geological map of the Sukamto, R., Apandi. T., Supriatna. S. and Yasin, A.. 1981.
Philippines. 1 : l,~,~. 9 sheets. Manila. The geology and tectonics of Haimahera island and sur-
Ranken, B., Cardwell, R.K. and Karig, D.E.. 1984. Kinematics rounding areas. in: A.J. Barber and S. Wiryosuyono (Edi-
of the Philippine Sea plate. Tectonics. 3: 555-575. tors), The Geology and Tectonics of Eastern Indonesia.
Ranneft, T.S.M., Hopkins, R.W.. Froehch, A.J. and Gwinn, Spec. Publ. indones. Geol. Res. Develop. C’entre, 2:
J.W., 1960. Reconnaissance geology and oil possibilities of 349-362.
Mindanao. Am. Assoc. Petrol. Geol. Bull., 44: 529-568. Supriatna, S., 1980. Geologic map of the Morotai quadrangle.
Rowlett, H. and Keiieher, J., 1976. Evolving seismic and North Maiuku. Indones. Geoi. Res. Develop. Centre.
tectonic patterns along the western margin of the Philip- LJyeda, S. and Ben-Avraham, Z., 1972. Origin and deveiop-
pine Sea plate. J. Geophys. Res., 81: 3518-3524. ment of the Philippine Sea. Nature (London). Phys. Sci..
Rutland, R.W.R., 1968. A tectonic study of part of the Philip- 240: 176-178.
pine Fault Zone. Q. J. Geol. Sot. London, 123: 293- 325. Van Bemmelen. R.W., 1970. The Geology of Indonesia. Nijhoff.
Scott, R.B. and Kroenke, L., 1981. Periodicity of remnant arcs The Hague, 2nd ed.
and back-arc basins of the south Philippine Sea. Ocean. Van der Wegen, G., 1963. Geology van Waigeo <W. New
Acta Suppl. Prcc. Int. Geol. Congr., 26th. Paris, pp. Gumea). Geoi. Mijnbouw, 42: 3-12.
193-202. Verbeek, R.D.M., 1908. Moiukken-Versiag. Jaarb. Mijnwezen
Seno, T., 1977. The instantaneous rotation vector of the Phiiip- Ned. Oost-lndie 1908, Verh., 37: 836 pp.
pine Sea plate relative to the Eurasian plate. Tectonophysics, Verstappen, H.Th.. 1960. Geomorphological observations on
42: 209-226. the North Moiuccan-North Vogeikop island arcs. Nova
Seno, T. and Maruyama, S., 1984. Paieogeo~aphic reconstruc- Guinea Geoi. (Leiden). 3: 13-37.
tion and origin of the Phiiippine Sea. Tectonophysics. 102: Verstappen, H.Th.. 1964. Some volcanoes of Haimahera
53-84. (Moiuccas) and their geomorphoiogicai setting. Tydschr. K.
Silitonga, PH., Pudjowalujo, H. and Moliat, H., 1981. Geo- Ned. Aadrijkskd. Genoot., Sl(3): 297-316.
logical reconnaissance and mineral prospecting on Bacan Wanner, J., 1913. Zur Geologie der Inseln Obimajora und
Island (Moiuccas, Indonesia). in: A.J. Barber and S. Halmahera in den Molukken. Neues Jahrb. Mineral. Geol.
Wi~osuyono (Editors). The Geology and Tectonics of Paiaeontoi., Suppl., 36: 560-585.
Eastern Indonesia. Spec. Publ. Indones. Geoi. Res. De- Weissei, J.K., 1983. Evidence for Eocene oceanic crust in the
velop. Centre, 2: 373-381. Celebes Basin. In: D.E. Hayes (Editor), The Tectonic and
Silver. E.A.. 1981. A new tectonic map of the Molucca Sea and Geologic Evolution of South-East Asian Seas and Islands.
East Suiawesi, Indonesia with implications for hydrocarbon Am. Geophys. Union, Monogr.. 23: 37-4X.
potential and metallogenesis. In: A.J. Barber and S. Weissei. J.K. and Anderson, R.N., 1978. Is there a Caroline
Wiryosuyono (Editors), The Geology and Tectonics of plate? Earth Planet. Sci. Lett.. 41: 143-158.
Eastern Indonesia. Spec. Publ. Indones. Geoi. Res. De- Yasin, A., 1980. Geologic map of the Bacan quadrangle, North
velop. Centre, 2: 363-372. Maluku. Indones. Geol. Res. Develop. Centre.