Tim Charlton, Banda Arc Geologist


Regional tectonics 

The following is a manuscript (slightly modified) recently rejected for publication. But I still like it, so you can judge for yourself if it has any merit... 

The Bird's Head-Halmahera microplate: an unrecognised plate simplifies present-day SE Asia tectonics


A new microplate, the Bird's Head-Halmahera (BHH) microplate, is defined within the SE Asian zone of three-megaplate interaction (Eurasia, Indo-Australia and the Pacific/Philippine Sea plates). This microplate is structurally compound, but has moved as a coherent body since about the beginning of the Pliocene (since ~5-6Ma). At the present day it rotates anticlockwise relative to Australia at a rate of 4.7°/m.y. about a pole to the south of the Aru islands (9.4°S, 134.3°E); and 4.3°/m.y. clockwise relative to the Pacific plate about a pole east of northern Halmahera (2.1°N, 129.0°E). The microplate is effectively attached to the Pacific plate at the northern pole, and to the Australian plate at the southern pole. Anticlockwise rotation relative to Australia is accommodated by a zone of extension behind the microplate (the currently extending Aru Trough, and formerly by extension in Cenderawasih Bay), while compression in front of the microplate is absorbed across the Banda Arc. An apparently older triangular extension zone in the Ayu Trough to the NE of the BHH microplate may record an earlier phase of clockwise rotation of the Philippine Sea plate relative to the Pacific in what may be a virtual mirror-image of the BHH microplate rotation relative to Australia.


The present-day tectonic complexity of SE Asia is focused particularly in eastern Indonesia where the Alpine-Himalayan belt and the Pacific 'Ring of Fire' meet in a T-junction interaction of three major plates (Eurasia, Indo-Australia and the Pacific: Figure 20). A fourth large plate, the Philippine Sea plate, also extends into this region, although its motion is very similar to that of the Pacific plate in this area. The eastern Indonesia region is characterised by high seismicity and by numerous active terrane boundaries including ocean-floor spreading centres and intra-continental extension zones, strike-slip fault systems and zones of underthrusting/subduction. Although a number of interpretations have been proposed for the evolution of the eastern Indonesia collison complex (e.g. Hamilton, 1979; Hall, 1996, 2002, 2010; Rangin et al., 1999; Charlton, 2000), no simple plate model has yet been proposed that accounts adequately for the present-day tectonics of the region. This article attempts to synthesise a new account of the present-day tectonics by introducing a new microplate, the Bird's Head-Halmahera (BHH) microplate (Figure 21).

 bhh location map

Figure 20: SE Asia zone of three megaplate interaction (Pacific PA, Australia AU and Eurasia EU), and the location of the Bird's Head-Halmahera (BHH) microplate. M.S.C.Z.: Molucca Sea Collision Zone. S.F.Z.: Sorong Fault Zone


bhh microplate

Figure 21: Bird's Head-Halmahera microplate and seismicity.

Present-day crustal motions relative to Australia

 It is widely recognised that GPS data broadly define a Bird's Head microcontinent that is rotating anticlockwise relative to Australia (Puntodewo et al., 1994; Stevens et al., 2002), for which Bock et al. (2003) defined a rotation pole at 14.2°S, 136.6°E (angular velocity 2.92±0.45°/m.y.). This pole is based on GPS vectors from six measurement stations (SORO = Sorong, FAKF = Fakfak, KAIM = Kaimana, MANO = Manokwari, BIAK = Biak, and OBIX = Obi island). The match between predicted motion and measured GPS motion is not particularly strong, however, especially with regard to the MANO and BIAK sites on the northern margin of Cenderawasih Bay (for reasons that will be discussed later). Alternatively, Charlton (2010) proposed a new rotation pole located at 9.4°S, 134.3°E (rate of rotation 4.7°/m.y.). This pole was derived by excluding the BIAK and MANO sites, but instead including the TERN (= Ternate) GPS station in the definition of the microplate (Figure 21). The fit between predicted motion and GPS-determined motion is much improved with this new pole, particularly for motion azimuths which agree within 3° (mean <2°) for all five stations in this Bird's Head-Halmahera microplate, whereas the disparity for the six stations defining the Bock et al. (2003) Bird's Head microplate is up to 19° (mean ~7°: Charlton, 2010; see Table below).

charlton table 1

The boundaries of the BHH microplate are shown in Figure 21. Internally the microplate has fairly low seismicity within the context of the regionally high seismicity of eastern Indonesia. Only the northern extremity of the microplate has strong seismicity, but this is deep-seated, associated with subduction at the southern end of the Philippine Trench. Minor seismicity also extends latitudinally across the middle of the microplate, broadly associated with the Sorong Fault Zone, but the seismicity is insufficient to suggest that this major left-lateral fault zone is still active to any significant extent. Rather, the Sorong Fault system probably became essentially inactive when the BHH microplate came into existence at about 5-6Ma.

Compensational deformation

 If the BHH microplate is rotating counterclockwise relative to Australia, this rotation must be accommodated by overall extension within the Australian continent behind the microplate, and compression in front of it.

Figure 22 shows a map of the Aru Trough, the main active extension zone accommodating rotation of the BHH microplate. Seismicity is largely restricted to a triangular zone at the axis of the Aru Trough. The seismicity is primarily extensional in character, and the mean extension direction derived from the seismicity (Pubellier & Ego, 2002) is indicated by the double-headed arrow at the axis of the trough. The margins of the Aru Trough are characterised structurally by normal faults downthrowing primarily towards the trough in both the Aru islands to the east and in Kai Besar island to the west. The trend of the normal faulting is approximately SW-NE in the Aru islands, but closer to N-S in Kai Besar. In Kai Besar subhorizontal Eocene limestones of the Elat Formation commonly have a pronounced extensional fracture fabric that also suggests nearly east-west extension, and the second double-headed arrow shown over Kai Besar indicates the mean extension direction inferred from 37 measurements of the extensional fracture fabric on that island. Together these factors are indicative of active extension in the Aru Trough with a rotational component about a pole located near to the southern end of the trough. The triangular Aru Trough extension zone (a sphenochasm: Carey, 1958) is compensating the counterclockwise rotation of the Bird's Head-Halmahera microplate relative to Australia.

aru trough map 2

Figure 22: The Aru Trough, an active extension zone accommodating anticlockwise rotation of the Bird's Head-Halmahera microplate relative to Australia. Bathymetry at 1000m interval. Double headed arrows represent extension direction interpreted from earthquakes in the Aru Trough (Pubellier & Ego, 2002); and from extensional fabrics in outcrop, Kai Besar island. Red arrows are GPS vectors relative to Australia, as in Figure 21. 


Cenderawasih Bay to the north of the Aru Trough (Figure 21) is another interpreted triangular extension zone accommodating anticlockwise rotation of the BHH microplate relative to Australia, although this sphenochasm was active primarily during the Pliocene. Evidence for Pliocene extension in Cenderawasih Bay includes multi-kilometre thicknesses of Pliocene sediments in the Waipoga Basin on the eastern margin of the Bay (at least 3km of Pliocene section based on drilling results, and up to 8km based on seismic data: Visser and Hermes, 1962; Dow et al., 1986; Sapiie et al., 2010); and in Bintuni Basin to the west of Cenderawasih Bay where the Pliocene-Recent succession is up to 5km thick (Chevallier & Bordenave, 1986). These two areas of massive Pliocene subsidence are separated by a metamorphic core complex on the Wandamen peninsula which exposes amphibolites with peak metamorphism dated radiometrically at about 5Ma, with an exhumation age of 1.5Ma (Pigram et al., 1982, Pigram & Davies, 1987; see Charlton, 2010). These rocks have been exhumed from depths of more than 15km since the peak of metamorphism in the Late Miocene (Robinson et al., 1990).

At the present day the western margin of Cenderawasih Bay is characterised by predominantly right-lateral strike-slip faulting, and the eastern margin by left-lateral faulting. The crustal fragment lying immediately north of Cenderawasih Bay, including the MANO, BIAK and YAPE (= Yapen island) GPS stations (see Figure 23b), appears to be moving southward into the expanding triangular extension zone to the east of the BHH microplate.

The compressional deformation in front of the counterclockwise-rotating BHH microplate (relative to Australia) is absorbed across the Banda Arc subduction-collision complex (Figure 21). The unusual 180° oroclinal bend of the Banda Arc can be simply accounted for as an accommodation of the westward rotational displacement of the BHH microplate (Figure 23).

bhh 2 pole rotation

Figure 23:  GPS motions of the Bird's Head-Halmahera microplate (defined by the TERN, OBIX, SORO, FAKF and KAIM GPS sites) and adjacent GPS sites (a) relative to Australia; (b) relative to the Pacific plate. Circles are small circles around the respective rotation poles. GPS sites in the Banda Arc broadly move in the direction indicated by the Bird's Head-Halmahera rotation, but at a slower rate, suggesting compensatory shortening in front of the anticlockwise-rotating microplate (relative to Australia). Note that relative to the Pacific the MANO, BIAK and YAPE GPS site vectors are rotated clockwise relative to the predicted BHH-Pacific rotation as the crustal fragment north of Cenderawasih Bay moves southward into the expanding triangular pull-apart basin.

Age and rotation history of the Bird's Head-Halmahera microplate

The BHH microplate is geologically heterogeneous, consisting of a continental platform with Australian stratigraphic affinity in the Bird's Head (Pieters et al., 1983), and a non-continental island arc terrane in Halmahera (Hall et al., 1988a & b). Clearly this microplate is not a geologically long-term structural entity, but is a compound crustal block that is rotating as a coherent body at the present day, and for a relatively short time back into the geological past. Evidence from both the extensional history of Cenderawasih Bay (outlined briefly above, and in more detail in Charlton, 2010) and the cessation of movement on the Sorong Fault system (Charlton, 2000) suggest that the BHH microplate has been rotating anticlockwise relative to Australia since about the beginning of the Pliocene (about 5-6Ma). The current rate of rotation is estimated at 4.7°/m.y. relative to Australia, which suggests a minimum rotation of about 24° during that time. However, it appears that the pole of rotation may be migrating southward through time, with the zone of extension opening more in the manner of a zipper rather than about a single fixed rotation pole. As the driver of the rotation is presumably the westward motion of the Pacific plate relative to Australia, and as the motion of the Pacific relative to Australia is likely to have been fairly constant through the Pliocene-Recent period, the rate of rotation of the BHH microplate relative to Australia may have decreased through time. The rotation of the Bird's Head relative to Australia may, therefore, be somewhat greater than the 24° minimum; perhaps 30-40° of total rotation since the Early Pliocene.

Relationship to the Pacific and Philippine Sea plates

Defining the rotation pole for the BHH microplate relative to Australia also defines poles of rotation for BHH relative to the Pacific and Philippine Sea plates. Based on the NUVEL-1A Pacific-Australia rotation pole (de Mets et al., 1990, 1994), the BHH-Australia rotation pole (9.4°S, 134.3°E, 4.7°/m.y.) implies a Pacific to BHH pole at 2.1°N, 129.0°E (4.3°/m.y.). Relative to the Philippine Sea plate (Seno et al., 1993) the pole of rotation is located at 2.3°N, 127.5°E (3.4°/m.y.). These two poles are located very close together (Figure 21), reflecting the similarity in present-day plate motion between the Pacific and Philippine Sea plates in this area; and close to the geologically and seismically defined northern end of the BHH microplate. The Pacific and Philippine Sea plates are separated in the area immediately to the east of Halmahera by the northward-tapering triangular Ayu Trough spreading centre (Figure 20). Although the history of the Ayu Trough is rather poorly constrained, it is most widely considered that spreading occurred primarily before about 6Ma, with only very slow spreading since (Weissel & Anderson, 1978). It is notable that there appears to be a broad mirror-image symmetry between the extension in the Aru Trough and extension in Cenderawasih Bay-Aru Trough, and the Ayu Trough may, therefore, record an earlier phase of microcontinental rotation within the Pacific-Australia-Eurasia megaplate triple junction.


The recognition of a single Bird's Head-Halmahera microplate greatly simplifies the tectonic picture of SE Asia at the present day, with a single crustal block rotating anticlockwise relative to Australia and clockwise relative to the Pacific. This rotation (relative to Australia) is compensated by a zone of extension behind the microplate (the Aru Trough and formerly the Cenderawasih Bay triangular extension zones), and by a zone of compression in front of the microplate (the Banda Arc). This general plate configuration has existed since about the beginning of the Pliocene, and the complex regional geological evolution during the Pliocene and Quaternary can be largely explained in terms of this essentially simple plate configuration. Among other features, this tectonic configuration helps to explain:

• The present-day kinematics and the 'aberrant' 180° curvature of the Banda Arc, where the northern limb of the arc has been rotated anticlockwise in front of the clockwise rotating BHH microplate;

• The exclusively Pliocene and younger age of the eastern Banda volcanic arc which developed above a distinct subduction zone as an accommodation response to the BHH rotation;

• The contemporaneous (~6Ma) cessation of seafloor spreading in the Ayu Trough (Weissel & Anderson, 1978) and the the onset of spreading in the South Banda Basin (Honthaas et al, 1998);

• Extension in Cenderawasih Bay and the Aru Trough since ~5-6Ma.



Backarc spreading in the Banda Sea

Two zones of backarc spreading are recognised in the Banda Sea. The North Banda Basin is bounded by the Banggai-Sula continental fragment and the Sorong Fault Zone to the north, and by the Banda Ridges to the south (Figure 24). The South Banda Basin is located between the Banda Ridges to the north and the Plio-Quaternary Banda volcanic arc to the south. The Banda Ridges crustal domain separating the two oceanic basins is composed of both continental and non-continental fragments, in the south bounded by remnants of a former volcanic arc in the Pisang, Lucipara and NEC Ridges. This remnant arc was active between about 6-8Ma (Late Miocene).

banda  spreading

Figure 24: Neogene backarc spreading in the Banda Sea

North Banda Basin

The North Banda Basin developed between about 12.5Ma and 7.15Ma (Hinschberger et al., 2000). Rudyawan & Hall (2012) have shown that the Sorong Fault does not continue west of Sulabesi island, and does not link to the Matano Fault in SE Sulawesi. These authors concluded that that the Banggai-Sula block was not transported to the west by the Sorong Fault Zone.

It is suggested here that the Sorong Fault strike-slip system and the spreading in the North Banda Basin were cogenetic, with the left-lateral motion on the Sorong Fault linking westwards into a ridge/transform-like spreading system in the North Banda Basin. The Matano and Sorong fault strands probably formed a continuous strike-slip fault before the spreading, and the separation between the ends of the two faults suggests that the spreading in the North Banda Basin absorbed approximately 350km of strike-slip motion on the Sorong Fault system further east. The Banggai-Sula block was translated at least this far westward in the Sorong Fault system during the Middle-Late Miocene.

This timing for movement on the Sorong Fault Zone is earlier than I suggested previously (Charlton, 1996). It suggests that the Kais Formation pinnacle reefs and the basinal Klasafet Formation, respectively the reservoir and likely main source rock successions in the Salawati Basin oil fields of western Papua, developed during the main period of motion on the Sorong Fault.

South Banda Basin

The South Banda Basin developed between about 6Ma and 3.5Ma (Honthaas et al., 1998). The locus of rifting was apparently along the southern edge of the Pisang-Lucipara-NEC volcanic arc, which remained as a remnant arc as the active arc migrated away southwards. Several previous authors (e.g. Hamilton, 1981; Hall, 2001; Harris, 2006) have interpreted the spreading in the South Banda Basin in terms of trench rollback – i.e., advance of  the overriding plate relative to the subducting plate in an absolute plate motion framework. However, as Sundaland appears to be close to static in an absolute tectonic framework, it seems unlikely that the South Banda Basin spreading was the result of trench rollback because the full spreading rate in the South Banda Basin was approximately 60mm/year (Hinschberger et al., 2000), which is less than the regional convergence between Australia and Eurasia in this region (~75mm/year).

An alternative explanation of spreading in the South Banda Basin is based on a fixed-slot subduction zone geometry – i.e., that the Sunda-Banda subduction trench has remained essentially fixed in an absolute plate motion framework, not advancing towards the subducting Australian plate as would be the case with trench rollback. The Sunda-Banda arc at this time had an essentially linear, approximately east-west trend, but the eastern termination of the arc abutted against the Australian continental margin (somewhere in the region of present-day Buru-Seram). This eastern termination of the arc was driven northward with Australia relative to the Sunda Arc further west, and the South Banda Basin developed as a triangular backarc extension zone in order to accommodate the relative motion between Australia and Sundaland and maintain an essentially linear east-west subduction front. At the western tip of the South Banda extension zone near Bali the amount of backarc spreading was close to zero, but the rate of extension increased progressively eastwards to 60mm/year northeast of Wetar island (Figure 24), and further east still, near to the eastern limit of the South Banda Basin extension zone, the full relative motion between Australia and Sundaland (~75mm/year) was accommodated.

The jump in backarc spreading from the North to the South Banda Basin at about 6Ma coincided with the inactivation of the Sorong Fault, and with a regional reorganisation of plate boundaries that was associated with the initiation of the Bird's Head – Halmahera microplate shown in Figure 21. The termination of spreading in the South Banda Basin at ~3.5Ma was probably the consequence of hard collision between the southern Banda Arc and the Australian continental margin in the Timor area at this time. Since then the southern Banda Arc has been driven northward relative to the Sunda Arc, and now moves essentially as an element of the Australian plate. The pronounced present-day curvature of the formerly linear east-west Banda Arc and its associated curved Benioff zone is a consequence of the anticlockwise rotation of the Bird's Head - Halmahera plate since 6Ma.


© Tim Charlton
Wesite design in Guildford by: Sitec IT Ltd