milsom_etal_1996 Gravity fields Halmahera.pdf

TECTONICS, VOL. 15, NO. 1, PAGES, 84-93, FEBRUARY 1996

Gravity fields in eastern Halmahera and the Bonin Implications for ophiolite origin and emplacement

AI'C ø

John Milsom

Research School of Geological and Geophysical Sciences, Birkbeck College and University College London, London, England

Robert Hall

SE Asia Research Group, Department of Geology, Royal Holloway College, Egham, England

Tatang Padmawidjaja Geological Research and Development Centre, Bandung, Indonesia

Abstract. Classic ophiolites, as exemplified by the Troodos Massif in Cyprus and the Papuan Ultramafic Belt in eastern New Guinea, are large overthrust masses which are generally associated with large positive gravity anomalies. However, similar rocks occurring in extensive fragmented terranes which have also been described as ophiolitic do not produce large gravity effects. The eastern part of the island of Halmahera, in northeastern Indonesia, is an ophiolite of this latter type. On the two eastern arms of the island, a Mesozoic ophiolitic basement is overlain by, and imbricated with, Upper Cretaceous and Paleogene arc volcanic and sedimentary rocks. Bouguer gravity values are generally in the range +50 to +150 mGal and are characterised by steep local gradients indicative of shallow sources. The Bouguer gravity average suggests that the crust is at least 20 km thick, and it must be even thicker if a significant part of the anomalous gravity field is due to the presence of a cold and therefore dense, lithospheric slab within the asthenosphere, associated with the present-day subduction beneath Halmahera. The absence of any exposures of continental basement rocks or of quartzose sediments in eastern Halmahera suggests that these ophiolites have not been overthrust onto continental crust and that the thickening occurred in an intraoceanic island arc. The Paleogene arc was evidently characterised by volcanism occurring over an unusually wide area. In this it resembles the Izu-Bonin volcanic arc, which, like Halmahera, has been situated

at the margin of the Philippine Sea Plate throughout its history. The gravity field of the Halmahera ophiolite is comparable with that of the Bonin volcanic arc, but there is no Halmahera parallel to the very high gravity fields recorded over the Bonin Islands forearc ridge. The equivalents of this part of the Paleogene arc may be represented by the ophiolitic complexes now distributed along the northern margin of the orogenic belt in New Guinea.

Introduction

Ophiolites are associations of rocks of oceanic affinity [peridotires, gabbros, basalts, and cherts] exposed above sea

Copyright 1996 by the American Geophysical Union.

Paper number 95TC02353. 0278-7407/96/95TC-02353510.00

level. Many of the classic examples, such as those in New Caledonia [Crenn, 1953], Cyprus [Gass and Masson-Smith, 1963], Papua [Milsom, 1973] and Oman [Manghnani and Coleman, 1981], are associated with broad, intense Bouguer gravity anomalies with peak to trough amplitudes of more than 100 mGal. However, there are also extensive ophiolite terranes which are characterised by complex gravity variations of generally low (<50mGal) amplitude. The East Halmahera ophiolite, in northeast Indonesia at the margin of the Philippine Sea Plate, is of this second type. Its possible former links with high-gravity ophiolites in nearby New Guinea and comparisons with the Bonin Arc, also at the margin of the Philippine Sea Plate, suggest a related origin for the two types of ophiolites.

Halmahera is situated in the extreme west of the western

Pacific (Figure 1) and is associated with an active, east facing volcanic arc. Seismological data indicate subduction of at least 200 km of oceanic crust [Cardwell et al., 1980], but the process must soon be terminated by collision with the west-facing Sangihe Arc which is linked to northern Sulawesi. The region has attracted considerable attention as containing the only present- day example of arc-arc collision, and the bulk of the published geophysical data relates to studies offshore in the Molucca Sea, between the two arcs [cf. Silver and Moore, 1978; McCaffrey et al., 1980, McCaffrey, 1982; Moore and Silver, 1983]. Plate boundaries are poorly defined; the Philippine Trench is propagating southward but does not extend south of the latitude of northern Halmahera, and convergence between the Philippine Sea Plate and Eurasia must be transferred by dextral strike slip across Halmahera into the Molucca Sea [McCaffrey, 1982; Nichols et al., 1990]. Halmahera is thus part of the Philippine Sea Plate, the southern limit of which is the Sorong Fault Zone, a major sinistral strike-slip system which separates Halmahera and the Molucca Sea from the Banda Sea and the Australian margin in Irian Jaya [Hamilton, 1979; Hall et al., 1991 ].

Geology of the Ophiolite Province

Halmahera island (Figure 2) is over 300 km from north to south and 125 km from east to west. Its distinctive shape leads naturally to a division into northwest, northeast, southeast, and southwest arms. The western arms have been constructed largely by arc volcanism during the Cenozoic [Hall et al., 1988b], but

84

MILSOM ET AL.: GRAVITY FIELD OF EAST HALMAHERA OPHIOLITE 85

North New Guinea Palaeogene Collided Arc

Palaeogene volcanic ridges remnant arcs

Finisterre - New Britain Arc

East Halmahera - Waigeo Ophiolite Terrane

Hew Guinea Ophiolite Belt

Extensional Zones Trenches

.30*N

• 150øE

Chichijima

Hahajima

.20•N

-%

V.o•l

The WesternPacific \- 0

I Ultrarustic Beli

Figure 1. The western Pacific and the Philippine Sea Plate, showing the locations of Halmahera and the Bonin Islands, and ophiolites and collided arc fragments in New Guinea. The line AA' across the Bonin Arc shows the location of the gravity profile of Figure 6.

the basement of the two eastern arms consists of Mesozoic

ophiolitic rocks and Late Cretaceous and Paleogene arc-related igneous and sedimentary rocks [Ballantyne, 1991b; Hall et al., 1991 ]. The East Halmahera-Waigeo Ophiolite province or terrane of Sukamto et al. [1981] and Ballantyne and Hall [1990] includes these two arms, the islands of Gag, Gebe, and Waigeo which lie between Halmahera and the Kepala Burung (Bird's Head) region of Irian Jaya, and the intervening marine areas (Figure 2). Paleogene arc fragments and ophiolites are scattered throughout northern New Guinea [Pigram and Davies, 1987], and a case could be made for including these in the province; the possible relationships between these fragments and the Halmahera ophiolites are discussed in the concluding section of the present paper.

The Halmahera ophiolites are dismembered and volcanic rocks are relatively uncommon but there are components present of each level of an intact ophiolite, with the possible exception of sheeted dykes. There are also some metamorphic rocks, including foliated amphibolites and rare blueschists. Peridotites on East Halmahera include abundant serpentinised harzburgites which record a high degree of mantle partial melting and are similar to those of oceanic forearcs, together with rare, less depleted, therzolites compatible with a mantle residue after extraction of mid-ocean ridge basalt [Hall et al., 1988a, 1991; Ballantyne, 1990, 1991a b]. Cumulates are common and include dunites, olivine clinopyroxenites, wehrlites, and olivine gabbro-norites. There is chemical and mineralogical evidence for a genetic link between the harzburgites and the cumulates [Ballantyne, 1991 a].

86 MILSOM ET AL.: GRAVITY FIELD OF EAST HALMAHERA OPHIOLITE

' Philippine Sea

Molucca Sea

HALMAHERA

MOROTAI ß

EAST HALMAHERA- WAIGEO OPHI OLITE TERRANE

Weda

Bay

4-000

BACAN

GAG•

BURUNG ...

132':

Figure 2. The East Halmahera - Waigeo Ophiolite terrane. The subsea boundaries of the terrane are uncertain, and it is not clear whether Batanta should be included or excluded.

Diabase dykes associated with pillow lavas have been reported from central Halmahera [Burgath et al., 1983], but, although microgabbros are abundant, no sheeted complex has been found elsewhere on the island. Pieters et al. (unpublished manuscript, 1979) have suggested that dykes on Gag may have formed part of a sheeted complex. Volcanics in the ophiolite complex include boninitic rocks interpreted as cogenetic with the cumulates [Ballantyne, 199 lb], arc tholeiites, and amygdaloidal calcalkaline basalts chemically similar to ocean island and seamount volcanic rocks. Isotopic ages (Ballantyne [ 1990] and unpublished results) and fossils from the oldest overlying sedimentary rocks indicate a probable Jurassic age for the ophiolites.

The ophiolitic rocks are associated in all areas with Cretaceous-middle Eocene predominantly volcanic and volcaniclastic rocks and limestones deposited in an island arc setting. In some areas these rocks are faulted and sliced in with the ophiolites, but in many areas there are coherent sequences of volcanic and sedimentary rocks. The ophiolitic rocks are themselves intruded by arc-related plutonic rocks which chemical evidence suggests are comagmatic with the volcanic rocks [Ballantyne, 1991a]. The Cretaceous to middle Eocene formations include basaltic and andesitic volcanic rocks, coarse

to fine andesitic breccias and conglomerates, volcaniclastic sandstones, and siltstones containing similar volcanic debris, red mudstones, and pelagic and redeposited limestones. Well-dated

formations in East Halmahera are Upper Cretaceous and lower- middle Eocene; on Waigeo there are similar Lower Cretaceous and lower Eocene formations. The Cretaceous-middle Eocene

sediments are interpreted as the fill of basins situated near to a single long-lived or intermittently active arc. They could be back arc or forearc deposits, although a forearc setting is favoured for the Upper Cretaceous volcaniclastic sequences on Halmahera because of similarities to modem forearcs and the close spatial association with metamorphic rocks, including rare blueschists, of similar ages [Hall et al., 1988a; Ballantyne, 1990].

The pre-Tertiary and lower Paleogene ophiolitic and arc volcanic and sedimentary rocks are interpreted as having formed part of a single complex terrane [Hall et al., 1988a] assembled prior to the establishment of a major regional unconformity in the late middle Eocene. Late Cretaceous to middle Eocene

volcaniclastic rocks are folded and faulted and locally dip steeply. In contrast, younger rocks are not strongly deformed except in areas of localised late Neogene deformation. In SE Halmahera, upper Eocene and younger sediments include conglomerates, sandstones, siltstones, mudstones, coals, and shallow water limestones which accumulated in a marginal to shallow marine environment, but in the Northeast Arm them are

thick upper Eocene and Oligocene deep-water basinal sequences, dominated by arc basaltic pillow lavas, thick volcaniclastic turbidite sandstones, siltstones, and mudstones, locally with


pelagic limestones. Volcaniclastic turbidites of Oligocene age are well exposed on Waigeo, where they are deformed into large- scale open folds [CharItoh et al., 1991]. A second, Lower Miocene, regional unconformity separates the Paleogene and older rocks from a shelf carbonate sequence which is typically about 1 km thick.

Ophiolite Province Gravity Data

Gravity surveys in an ophiolite province might be expected to define large gravity anomalies associated with deep root zones. However, no such anomalies were recorded on Waigeo or Gag island, where gravity readings were taken in the course of the Indonesian-Australian Irian Jaya mapping project [Dow et al., 1986]. Results from these islands, together with those from northern Kepala Burung of Irian Jaya itself, have been incorporated in the regional Bouguer gravity map, the limits of which were selected to show the boundary between the ophiolite province and oceanic crust (Figure 3).

Subsequently, as part of an integrated geological-geophysical study of the Sorong Fault Zone mounted jointly by the Geological Research and Development Centre, Bandung, Indonesia, and the University of London, gravity readings were taken along all the accessible coasts in the Halmahera archipelago. The low drift and high reliability of the LaCoste-Romberg geodetic gravity meters allowed drift to be

adequately controlled on loops which in some cases lasted for several days. As on Waigeo, large anomalies which could be attributed to deep root zones were not found. Inland traverses would have provided important additional information, but the lack of roads, the rugged topography, and the thick rainforest presented obstacles which were insuperable in the time available for survey work.

Values of absolute gravity for the Halmahera surveys were obtained by reference to the accepted value of 978,164.80 mGal for the Indonesian National Base Station at Pattimura airport on Ambon [Adkins et al., 1978]. The data were reduced to Bouguer gravities using the 1967 International Gravity Formula applied to latitudes estimated from 1:250,000 topographic maps, elevations obtained by direct reference to sea level, and a Bouguer density of 2.67 Mg m -3. Overall errors in reduced values should not exceed 1 mGal, which is trivial compared with the more than 100-mGal total anomaly range recorded and the 10-mGal contour interval dictated by the steep gravity gradients.

Marine gravity data obtained east of Halmahera by U.S. and other research vessels control the offshore contours in Figure 3. The coverage is very uneven, and the contours shown have in places been interpolated across considerable areas in which data are lacking. The two most important data sources were the systematic survey of the Ayu Basin by Weissel and Anderson [ 1978] and the 1988 Eastern Indonesia cruise of the RRS Charles

Darwin. The latter has provided almost the only public domain data for the region between Halmahera and the Ayu Basin.

4øN

2øS 128øE

Kepala Burung

130øE 132øE 134øE

Bouguer gravity

Reduction density 2.67 Mg.m •

5O km

crust

..-• + 175 regal

Figure 3. Regional Bouguer gravity map of the western equatorial Pacific, based on marine data filed with the National Geophysical Data Center, Boulder, Colorado and land data from Dow et al. [ 1986] and from surveys described in the text. Light shading indicates areas underlain by oceanic crust in the Philippine Sea, Ayu Basin, and Pacific Ocean. The + 175-mGal contour is shown as indicating the approximate boundary between oceanic crust and the ophiolite province.


The Gravity Field of the East Halmahera - Waigeo Ophiolite

Bouguer gravity values in the area covered by the regional map (Figure 3) exceed +300 mGal in the Pacific and Philippine Sea oceanic basins and sporadically in the Ayu Basin, which contains the youngest oceanic crust in the area [Weissel and Anderson, 1978], and are close to zero in the centre of the Kepala Burung region of New Guinea. The average values in these provinces are consistent with oceanic and continental crustal thicknesses (6 km and 30 km), respectively. The steepest regional gradients occur between the + 150 and +250 mGal contours, and the + 175 mGal contour marks the approximate boundary between the East Halmahera- Waigeo Ophiolite province and oceanic crust.

Gravity contours on Waigeo, as shown by Dow et al. [1986] and briefly discussed by Milsom [1991 ], lack strong linear trends. Bouguer anomalies of more than + 150 mGal were recorded only

in the extreme east of the island, while the lowest values, of less

than +80 mGal, were at points on the western part of the south coast. There are minor outcrops of ophiolitic rocks in both the low and high gravity areas. Still lower values, of little more than +50 mGal, characterise the small islands a few kilometres farther

south and might suggest that these islands lie outside the ophiolite province. However, still farther south a closed high defined by steep gradients is centred on Batanta Island, which is composed mainly of Paleogene arc volcanic rocks [Dow et al., 1986].

The gravity pattern over eastern Halmahera (Figure 4) is less well defined than on Waigeo because the central parts of the two arms are remote from the gravity stations along the coasts. However, the E-W trend to the contours drawn across the

northern part of the NE Arm is well controlled, and there is field and aerial photographic evidence for a major E-W fault onshore in this area which downthrows the basement to the north by at least 2 or 3 km.

Farther south, Bouguer gravity varies rapidly along the coasts, the coast-to-coast distances are large, and the ambiguities are consequently greater. In the preferred solution (Figure 4) the inland areas have generally been contoured as gravity lows, but there could be a far greater linkage between the highs on opposite coasts, as is shown in the inset. This alternative requires less extreme gradients in some areas but involves abrupt changes in trend at the margins of Kau Bay and contour patterns which are discordant to most of the surface topographic and geological trends. The uncertainties are obviously considerable, but the two solutions, considered together, do demonstrate the improbability of Bouguer gravity values exceeding +150 mGal over more than a small portion of the area of the NE and SE Arms. Figure 4 also shows the western part of Halmahera, which was excluded from Figure 3 as being outside the limits of the ophiolite province. The strong local gradients and absolute Bouguer gravity values of the order of +200mGal observed in western Halmahera are

associated with the modem volcanic arc and collision tectonics

and not with the ophiolite. The gravity highs around Kau Bay, and also those near the

southeastern tip of the Southeast Arm, might conceivably owe their existence to the presence of unmapped unserpentinised ultrabasic rocks, but there is a notable correlation, similar to that

on Batanta, between relatively high gravity fields and the presence of arc volcanic rocks. The eastern end of the Southeast

1•27OE 1 • 28øE

0 100 km i

-2ON

I 129øE

HALMAHERA ONSHORE BOUGUER

ANOMALIES

Reduction density 2.67 Mg. rn"

Contour interval 10 regal

N

!

I //980d•100 120

40

Figure 4. Bouguer gravity map of Halmahera. Reduction density is 2.67 Mg m -3. Because all readings were taken very close to sea level, changes in reduction density would have virtually no effect on the contour pattern. The inset shows an alternative contouring of the data from eastern Halmahera in which highs or lows on opposite sides of the NE Arm have been linked wherever possible.

Arm where ophiolitic rocks are most common and where the highest gravity fields might be expected is not characterised by unusually high gravity fields. None of the highs recorded approach, in either amplitude or areal extent, those associated with classic ophiolites such as Cyprus [Gass and Masson-Smith, 1963] and the Papuan Ultramafic Belt [Milsore, 1973], but they do resemble features observed on island arcs with long and complex subduction histories. Ophiolitic rock types are often exposed in such arcs, but the relatively small gravity anomalies suggest that they lack direct "root zones" connecting them to the mantle.

Crustal Thicknesses in the Ophiolite Province

Initial Estimates

There have been no deep seismic refraction studies in any part of the ophiolite province, but the gravity data can be used to set reasonable limits to the range of possible crustal thicknesses.


Continental crust at sea level, 30 km thick and with a density contrast of 0.4 Mg m -3 across the Moho, is commonly taken in gravity modeling as the standard which produces zero values of Bouguer gravity at the surface. A 1-km change in elevation of the Moho in such a model produces a gravity change of about 16 mGal, and the Bouguer gravity values of between +50 and +150mGal within the ophiolite province therefore suggest crustal thicknesses of between 21 and 27 km. This far exceeds the

average thickness of oceanic crust. Patterns of variation in oceanic crustal thickness have been summarised by Mutter and Mutter [1993], who noted that crustal thickening of oceanic plateaus such as the Shatsky Rise and the Ontong Java Plateau usually takes place by thickening within Layer3 but that thickening in oceanic islands, whether island arc or hot-spot related, is achieved by increase in the proportion of extrusive and high-level intrusive rocks, i.e., in the thickness of Layer 2. The thick crust of East Halmahera would be expected to be of this latter type; it may therefore be slightly denser than standard continental crust and may make some excess positive contribution to the gravity field, in which case an even smaller elevation of the Moho above the 30-km level would be required to produce the observed Bouguer gravity levels.

The validity of using Bouguer gravity to estimate crustal thickness may be questioned because of the incomplete gravity coverage. However, nowhere do the fields, either recorded or extrapolated as discussed above, suggest thicknesses of much less than 20 km. In eastern Halmahera the crust may be even thicker than crudely estimated from the Bouguer anomalies because of the possibility of a positive contribution to the gravity field from the lithospheric slab subducted beneath Halmahera from the west.

Digression: Gravitational Effects of a Subducted Slab

If, as is now widely believed, the most important plate-driving force is the gravitational instability of cold oceanic lithosphere [cf. Mueller and Phillips, 1991], studies of the gravity field of active volcanic arcs must take into account the possible gravitational effects of the subducted slabs. Quantifying these effects is difficult, since neither the thicknesses of the regions of mass excess nor the density contrasts are well constrained, but McCaffrey et al. [1980] did incorporate a subducted slab into

their interpretation of a free-air gravity profile recorded along a transect west from northern Halmahera across the Molucca Sea.

Their analysis was based on the model presented by Grow and Bowin [ 1975] in which the slab-mantle density contrast was set at +0.05 Mg m -3. There are additional uncertainties in extending this modeling over the wider Halmahera region since the Benioff Zone beneath the arc is a very variable feature which, although reasonably well defined on the chosen transect, can be traced only as far south as IøN, i.e., to about 50 km north of Makian, the most southerly active volcano.

Figure 5 shows contours of the estimated slab contribution to the gravity field for the cases of minimum and maximum likely southward extent of the slab. Calculations have been based on

simple west to east subduction, strike-limited two-dimensional bodies, and the Grow and Bowin [1975] density model. The contours in Figure 5a have been derived by supposing a subducted slab to exist only to the north of Makian, where a Benioff Zone can actually be identified. In this region, the slab has been assumed to be a coherent body, although the actual

distribution of hypocentres suggests that it is fragmented and discontinuous. Figure 5b shows the contours derived using a continuous slab extending south as far as the Sorong Fault Zone. Subducted Philippine Sea lithosphere is included in the northern part of both models and contributes to the pronounced NE-SW contour trend. The calculated gravity effects increase westward from the west coast of Halmahera because, in the models used, the largest mass excesses are associated with the depression of cold lithosphere beneath the Molucca Sea and not with the lithosphere which dips beneath the arc.

There is little evidence in eastern Halmahera for any southward decrease in Bouguer gravity similar to that shown in Figure 5a. It therefore seems likely that despite the absence of seismic evidence for its presence, high-density lithosphere dips beneath the whole of the active volcanic chain and extends also

beneath the inactive volcanic centres as far south as Bacan. The

active volcanism on Makian and presence of volcanic cones on Bacan which have suffered little erosion, suggesting that volcanism there ceased only within the last hundred thousand years, support this conclusion. In this case, average Bouguer anomalies in the north and south of Halmahera would be

expected to be approximately equal, as in fact they are. The moderately high field levels in the ophiolite province cannot be taken as necessarily indicating large departures from crustal isostatic equilibrium, since any contribution from the subducted slab would be from a mass excess decoupled from the overiding plate.

Thickness of the Crust Beneath Halmahera

The moderate Bouguer gravity levels encountered in eastern Halmahera indicate that the terrane, although formed in an exclusively oceanic environment, now has a thickness and bulk de,.nsity approaching that of continental rather than oceanic crust. This conclusion might require modification if it could be shown that the region is underlain by low density mantle, but such an effect, if it exists, may merely serve to compensate for the positive effect of high-density subducted lithosphere.

The Gravity Field of the Bonin Arc

The Paleogene rocks of East Halmahera, which formed in an oceanic arc remote from continental influences, have been described by Ballantyne [1991a] as strikingly similar to rocks of similar age which now lie at the eastern margin of the Philippine Sea Plate in the Bonin and Mariana Arcs and forearcs [cf. Bloomer and Hawkins, 1983]. Moreover, recent palaeomagnetic studies [Hall et al., 1995] indicate that eastern Halmahera, like the Bonin Arc, formed part of the active margin of the Philippine Sea Plate during the Paleogene. If the early development of eastern Halmahera was indeed similar to that of the Bonin Arc, some parallels might be expected between their present-day gravity fields, even though the Neogene histories have been very different.

The most prominent gravity feature in the Bonin-Mariana region is the extraordinarily strong anomaly over the forearc Bonin Ridge. Free-air anomalies are close to +350mGal on Chichijima [Tanaka and Hosono, 1974] and approach or exceed +400mGal on and around Mukojima and Nakodojima at the northern end of the island chain [Honza and Tamaki, 1985].


_1

,70 60

Makian

Contour interval

_ 1øS 10mGal _

127øE 128øE 129øE I I I

b

IøN

ß .

..................

................. 70

5O

40

30

Contour interval

]o s 10 mGal

127øE I 129øE I I

Figure 5. Gravitational effects of subducted slabs in the Halmahera region. (a) The slab from the Molucca Sea is assumed to terminate abruptly at 0ø30'N, underlying only the region, between Makian and the northern tip of Halmahera, where there is present-day volcanic activity. (b) The slab from the Molucca Sea is assumed to terminate close to the equator. Both models include a slab subducted at the Philippine Trench terminating at about 2øN. All calculations are based on the model of Grow and Bowin [ 1975] in which the slab between 70 and 300 km depth has a positive density contrast of 0.05 Mg m -3 with its surroundings. Areas of shading show locations assumed, in each case, for the most recently subducted parts of the downgoing Halmahera and Philippine Sea Plates; arrows show directions of movement of these plates relative to Halmahera.

These latter values, which are probably the highest observable at sea level anywhere in the world, have been shown on numerous widely circulated maps [cf. Watts et al., 1976], but there has been little published quantitative analysis.

Free-air anomalies across the active Bonin volcanic arc are

more subdued but still exceed + 100 mGal in any areas where the seafloor approaches sea level. On one of the volcanic islands [Nishinoshima], sea-level values are slightly less than + 190 mGal [Ohkawa and Yokoyama, 1977]. Only within the deep forearc basin (Bonin Trough) is free-air gravity negative, and in this area the negative effect of the water column would amount to well over a hundred milligals.

The significance of the Bonin Arc for the interpretation of the gravity field of the East Halmahera ophiolite lies in the density differences that modeling indicates must exist between the Bonin forearc and the volcanic arc. Taylor [ 1992] has suggested that the entire Bonin Arc complex was initiated in the Eocene by a phase of boninitic magmatism which extended over a region 300 km wide, an event which has no present-day parallel. Surface exposures, the results of deep sea drilling, and analogies with older, better exposed arc terranes combine to suggest that the crust beneath the volcanic arc must now consist of an original oceanic basement extensively modified first by the addition of the boninites and later by tholeiitic and calc-alkaline magmatism [cf.

Taylor, 1992]. The end result of these processes probably differs little in bulk density or thickness from normal continental crust. Calc-alkaline magmatism would, however, have been largely absent from the forearc ridge, which might be expected to resemble "supra-subduction zone" ophiolites such as the Troodos Massif and the Papuan Ultramafic Belt.

Figure 6 shows a gravity profile and simple geological model for a transect (position shown in Figure 1) across the Bonin Arc. The line selected crosses the arc from west to east in the vicinity of the most dramatic free-air gravity variations, beginning in the west in the back arc Shikoku Basin and from them passing in succession across the volcanic arc ([although not in the vicinity of any of the known volcanic islands or seamounts), the Bonin Trough forearc basin, the Bonin Ridge, and the Izu-Bonin Trench, ending in a region of old Pacific Basin oceanic crust. The free-air profile is dominated by the high values over the Bonin Ridge; lesser, but still significant, features include free-air lows associated with the forearc basin and the subduction trench where

the minima, on this profile, are of the order of-60 mGal and -160 mGal, respectively. The difference in water depth of about 2 km, together with a contribution from the low-density sediment infill, is sufficient to produce the difference between the free-air anomalies over the trough and the volcanic arc, but the very high values over the Bonin Ridge demand a difference in density of at


400

200

-200

Bonin . . -- - Observed Arc lated

mGals

-0

10

20

30

Volcanic Bonin Bonin Bonin Shikoku Basin Arc Trough Ridge Trench Pacific Ocean

Sea Water •_c• ,l• Sea Water

+o.4_ // - V.E• • 5:1 _ _+ 4 Densiw •ff 4 Mg.m ntras• in I I I I I I I I I I I I I I

-600 Km -500 -400 -300 -200 -1 O0 0 1 O0 200

Figure 6. Gravity profile and density model for the Bonin Arc along Line AA' of Figure l, near 28øN. For convenience in modeling, density values are referenced to a standard continental crust 30 km thick and to a density contrast of +0.4 Mg m -3 across the Moho. This crust is in equilibrium with oceanic crust which is 6.5 km thick and 0.1 Mg m -3 denser than continental crust and which lies beneath 5 km of sea water. The model incorporates the effect of a subducted slab with an density excess of 0.05 Mg m -3 over its surroundings between 70 and 300 km depth. Free-air gravity profile based on the free-air gravity map of Honza and Tamaki [ 1985].

least 0.15 Mg m '3 between the crust of the volcanic ar, and the crust of the Bonin Ridge.

Although the arc-forearc density difference must extend to depths in excess of 10 km in order to retain geologically reasonable density contrasts, it must also extend virtually up to the seafloor to explain the steepness of the free-air gravity gradient between the forearc basin and the Bonin Ridge. This gradient, taken in conjunction with the bathymetry, also provides control on absolute density values. A density contrast of +0.23 Mg m '3 is dictated by the shape of the rock/water interface, and this translates into a density of 2.9 Mg m -3 for the uppermost parts of the ridge. If the crust were thin beneath the forearc basin, as beneath a conventional extensional basin, it would be still more difficult to model the gradient and would require still higher density at depth within the Bonin Ridge.

Support for the existence of a significant difference in density between the Bonin volcanic arc and forearc comes from a study by Horine et al. [ 1990] of the forearc farther north. Working in an area in which neither the forearc ridge nor the forearc basin is a prominent feature and concerning themselves primarily with the effects of serpentinite diapirism, these authors used densities of 2.69 Mg m -3 for the upper part of the forearc where no diapirism is present and 2.46 Mg m -3 for the re•ion beneath a known diapir. The absolute values are not well controlled, but, significantly, in the light of the results presented here, the model failed to explain a rapid decrease in free-air gravity in the direction of the volcanic arc. It would therefore seem that in this northern area also, the crust of the volcanic arc is notably less dense than the crust of the forearc.

The position of the Moho beneath the volcanic arc is not well controlled by seismic refraction data [Honza and Tamaki, 1985], but, given the parameters of the survey, the failure to record Moho refractions implies a depth of at least 20 km. To achieve agreement between observed and calculated fields with crust of

this thickness, a small positive density contrast (+0.05 Mg m -3) with standard continental crust has been assumed. The calculated

effect of the seafloor topography along the volcanic line is then too large, but no attempt has been made to achieve a perfect fit because the two-dimensional assumption is invalid where the transect crosses the volcanic chain. The chosen density does produce excellent agreement with observation at the western margin of the Bonin Trough forearc basin.

Below the levels illustrated, the model of Figure 6 incorporates a subducted slab which between 70 and 300 km depth is 0.05 Mg m -3 more dense than its surroundings. Position and depth extent are seismologically well controlled [Katsumata and Sykes, 1969; van der Hilst and Seno, 1993]. Because the main gravitational effect of this slab is concentrated east of the volcanic arc, its presence has relatively little influence on the estimate of crustal thickness.

Modeling of the Bonin Arc gravity field has been undertaken purely as an adjunct to investigations of specific aspects of the gravity results from Halmahera, and no claim is made that interpretation of this single profile constitutes a complete analysis. The model shown is only one of an infinite range of possible solutions but serves to illustrate two essential features of the real geology, these being the relatively thick crust of the volcanic arc and the high density of the Bonin Ridge.

Discussion

The possibilities of positive contributions from the subducted slabs, which would be expected to be larger in the Bonin Arc where the Benioff Zone is more continuous and extends to

greater depths, and of possible negative contributions from low- density mantle, more likely in the Bonin region because of the back arc extension in the Shikoku Basin, cannot be ignored, but average gravity fields in the vicinity of the Bonin volcanic arc


and in eastern Halmahera are compatible with similar crustal structures and thicknesses in the two regions. This similarity does not, however, extend to the Bonin forearc region, where gravity fields are up to 200 mGal higher than the highest levels recorded on Halmahera. This observation prompts speculation as to the location of any Halmahera equivalent of the Bonin Ridge.

There are three obvious possibilities, one being that there are not, and never were, any such equivalents. Another is that the former forearc, as well as the arc, is represented within the present East Halmahera ophiolite province but has been subjected to tectonic thickening, which would lower the Moho, and serpentinization, which would reduce the density. The third solution, which is the most interesting and is the one favoured here, is that the frontal segment of the forearc to the Eocene Halmahera volcanic arc is now distributed as slivers of ophiolite along the northern flanks of the central ranges in New Guinea. Recent plate reconstructions suggest that northern New Guinea was a largely strike-slip boundary through much of the Neogene and that during this period Halmahera tracked along the margin from east to west [Hall et al., 1995]. Under such a regime, fragments of the Philippine Sea margin could well have been transferred across the plate boundary. Of the New Guinea ophiolites, the Papuan Ultramafic Belt has a peak Bouguer gravity of +200mGal, despite being overthrust on continental crust, is associated with Paleogene boninites on Cape Vogel [Dallwitz et al., 1966] and has a chemistry consistent with supra- subduction zone origin; it is a tempting analogue of the Bonin

Ridge. Other ophiolites along the northern slopes of the central ranges are less well known but appear to be chemically similar to the Papuan belt [cf. Jaques, 1981 ] and are associated with strong positive gravity anomalies [Milsom, 1991 ]. Arc fragments in the north coast ranges of New Guinea [cf. Dow, 1977; Pigram and Davies, 1987] might, on the other hand, be direct equivalents of the East Halmahera ophiolite, as parts of the same subduction- related Eocene volcanic belt. If this hypothesis is correct, a general implication may be that nonaccreting forearc ridges and volcanic arc fragments can be relatively easily detached from their plate under transpression, simplifying their emplacement as supra-subduction zone ophiolites.

Acknowledgments. The Halmahera onshore gravity data were obtained in the course of the Sorong Fault Zone Project funded by UK Natural Environment Research Council grant GR3/7149. Additional financial support was received from the Royal Society and the University of London Consortium for Geological Research in Southeast Asia and Central Research Fund. T. Ishihara of the

Geological Survey of Japan supplied raw gravity data from the Bonin Islands and copies of relevant Japanese publications. We thank the present and former directors of the Geological Research and Development Centre, M. Untung, R. Sukamto and I. Bahar, for their support: Daniel Lelitoly, Didi Pandu Tasno, and Agus Haryono for their contributions to the gravity surveys; Simon Baker for artwork and Tony Barber and John Dewey for incisive reviews which prompted major revision and improvement.

References

Adkins, J., $. Sukardi, H. Said, and M. Untung, A Regional Gravity Base Station Network Jbr Indonesia, Publ. Tek. Geol. Surv. Indones. Ser. Geofis. 6, 75 pp., 1978.

Ballantyne, P.D., The ophiolitic basement rocks of the Halmahera region, eastern Indonesia: Petrology, mineralogy, geochemistry and igneous and metamorphic history, Ph.D. thesis, Univ. of London, London, 1990.

Ballantyne, P.D., Petrology and geochemistry of the plutonic rocks of the Halmahera ophiolite, eastern Indonesia, an analogue of modem oceanic forearcs, Geol. Soc. London Spec. Publ. 60, 179-202, 1991a.

Ballantyne, P.D., Petrological constraints upon the provenance and genesis of the East Halmahera ophiolite, J. SE Asian Earth Sci., 6, 259-269, 1991b.

Ballantyne, P.D., and R. Hall, The petrology of the Halmahera ophiolite, Indonesia: An early Tertiary forearc, in Ophiolites: Oceanic Crustal Analogues, edited by J. Malpas, E.M. Moores, A. Panayiotou, and C. Xenophontos, pp. 461-475, Geol. Survey Dept., Ministry of Agriculture and Natural Resources, Nicosia, Cypru s, 1990.

Bloomer, S.H., and J.W. Hawkins, Gabbroic and ultramafic rocks from the Mariana Trench: An

island arc ophiolite, in The Tectonic and Geologic Evolution of Southeast Asian Seas and Islands, Part 2, Geophys. Monogr. Ser., vol. 27, edited by D.E.Hayes, pp.294-317, AGU, Washington, D.C., 1983.

Burgath, K., M. Mohr, and W. Simandjuntak, New discoveries of blueschist metamorphism and mineral occurrences in the Halmahera-Gag ophiolite belt, East Indonesia, Dir. Min. Resour. Indones. Bull., 13, 1-19, 1983.

Cardwell, R. K., B. L. Isacks, and D. E. Karig, The spatial distribution of earthquakes, focal mechanism solutions and subducted

lithosphere in the Philippine and northeastern Indonesian islands, in The Tectonic and Geologic Evolution of Southeast Asian Seas and Islands, Geophys. Monogr. Ser. Vol. 23, edited by D.E.Hayes, pp. 1-36, Washington, D.C., 1980.

Charlton, T. R., R. Hall, and E. Partoyo, Geology and tectonic evolution of Waigeo Island, NE Indonesia. J. SE Asian Earth Sci., 6, 289-297, 1991.

Crenn, Y., Anomalies gravimetriques li6es aux roches basiques de Nouvelle Caledonie. Ann. Gdophys., 9, 291-299, 1953.

Dallwitz, W. B., D. H. Green, and J. E. Thompson, Clinoenstatite in a volcanic rock from the Cape Vogel area, Papua, J. Petrol., 7, 375-403, 1966.

Dow. D.B., A geological synthesis of Papua New Guinea, Bull. Bur. Miner. Resour. Geol. Geophys. Aust. 201, 41pp., 1977.

Dow, D.B., G.P. Robinson, U. Hartono, and N. Ratman, Geologic map of Irian Jaya, Indonesia (1:1,000,000 scale), Geol. Res. Dev. Cent., Bandung, Indonesia, 1986.

Gass, I.G., and D. Masson-Smith, Geology and gravity anomalies of the Troodos massif, Cyprus, Philos. Trans. R. Soc London A, 255, 417-467, 1963.

Grow, J.A., and C.O. Bowin, Evidence for high- density crust and mantle beneath the Chile Trench due to descending lithosphere, J. Geophys. Res., 80, 1449-1458, 1975.

Hall, R., M. G. Audley-Charles, F. T. Banner, S. Hidayat, and S. L. Tobing, The basement rocks of the Halmahera region, east Indonesia: a Late Cretaceous-Early Tertiary forearc, J. Geol. Soc. London, 145, 65-84, 1988a.

Hall, R., M.G. Audley-Charles, F. T. Banner, S. Hidayat, and S. L. Tobing, Late Paleogene- Quaternary Geology of Halmahera, Eastern

Indonesia: initiation of a volcanic island arc, J. Geol. Soc. London, 145, 577-590, 1988b.

Hall, R., G. J. Nichols, P. D. Ballantyne, T. Charlton, and J. Ali, The character and significance of basement rocks of the southern Molucca Sea region. J. SE Asian Earth Sci., 6, 249-258, 1991.

Hall, R., M. Fuller, J. R. Ali, and C. D. Anderson, The Philippine Sea Plate: Magnetism and reconstructions, in Active Margins and Marginal Basins of the Western Pacific, Geophys. Monogr. Ser., Vol. 88, edited by B. Taylor and J.H. Nutland, pp. 371-404, AGU, Washington, D.C., 1995.

Hamilton, W., Tectonics of the Indonesian region, U.S. Geol. Surv. ProJ; Pap. 1078, 1979.

Honza, E., and K. Tamaki, The Bonin Arc, in The Ocean Basins and Margins, vol. 7A, The Pacific Ocean, edited by A.E.M. Naim, F.G. Stehli, and S. Uyeda, pp. 459-502, Plenum Press, New York, 1985.

Horine, R. L., G. F. Moore, and B. Taylor, Structure of the outer Izu-Bonin forearc from

seismic reflection profiling and gravity modeling, edited by P. Fryer et al., Proc. Ocean Drilling Program, Initial Rep. 125, 81- 94, 1990.

Jaques, A.L., Petrology and petrogenesis of cumulate peridotites and gabbros from the Marum ophiolite complex, northern Papua New Guinea, J. Petrol., 22, 1-40, 1981.

Katsumata, M., and L.R. Sykes, Seismicity and tectonics of the western Pacific: Izu-Mariana-

Caroline and Ryukyu-Taiwan region, J. Geophys. Res. 74, 5923-5948, 1969.

Manghnani, M.H., and R.G. Coleman, Gravity profiles across the Samail ophiolite, Oman, J. Geophys. Res. 86, 2509-2525, 1981.

McCaffrey, R., Lithospheric deformation within the Molucca Sea arc-arc collision: evidence


from shallow and intermediate earthquake activity, J. Geophys. Res. 87, 3663-3678, 1982.

McCaffrey, R., E. A. Silver, and R W. Raitt, Crustal structure of the Molucca Sea collision

zone, Indonesia, in The TeCtonic and Geologic Evolution of Southeast Asian Seas and Islands. Geophys. Monogr. Ser., vol. 23, edited by D.E.Hayes, pp.161-177, AGU, Washington, D.C., 1980.

Milsom, J., Papuan Ultramafic Belt: Gravity anomalies and the emplacement of ophiolites, Geol. Soc. Am. Bull., 84, 2243-2258, 1973.

Milsom, J., Gravity measurements and terrane tectonics in the New Guinea region, J. SE Asian Earth Sci., 6, 319-328, 1991.

Moore, G. F., and E. A. Silver, Collision processes in the northern Molucca Sea, in The Tectonic and Geologic Evolution of Southeast Asian Seas and Islands, Part 2, Geophys. Monogr. Ser., Vol. 27, edited by D.E.Hayes, pp.360- 372, AGU, Washington, D.C. 1983.

Mueller, S., and R.J. Phillips, On the initiation of subduction, J. Geophys. Res 96, 651-665, 1991.

Mutter, C.Z., and J.C. Mutter, Variations in thickness of Layer 3 dominate oceanic crustal structure, Earth. Planet. Sci. Lett., 117, 295- 317, 1993.

Nichols, G., R. Hall, J. Milsom, D. Masson, L. Parson, N. Sikumbang, B. Dwiyanto, and H.

Kallagher, The southern termination of the Philippine Trench, Tectonophysics 183, 289- 303, 1990.

Ohkawa, S., and I. Yokoyama, Gravity anomalies on Nishinoshima volcano, Geophys. Bull. Hokkaido Univ. 36, 83-95, 1977.

Pigram, C. J., and H.L. Davies, Terranes and the accretion history of the New Guinea orogen, BMR J. Aust. Geol. Geophys., 10, 193-211, 1987.

Silver, E.A., and J.C. Moore, Molucca Sea collision zone, Indonesia, J. Geophys. Res., 83, 1681-1691, 1978.

Sukamto, R., T. Apandi, S. Supriatna, and A. Yasin, The geology and tectonics of Halmahera and the surrounding islands, in The Geology and Tectonics of Eastern Indonesia, edited by A. J. Barber and S. Wiryosujono, Spec. Publ. Geol. Res. Dev. Cent., Bandung,. 2,349-362, 1981.

Tanaka, M., and T. Hosono, Geomagnetic and gravity surveys in the Bonin Island (Chichijima), the Ogasawara Archipelago (in Japanese), J. Geod. Soc. Jpn, 20, 193-208, 1974.

Taylor, B., Rifting and volcano-tectonic evolution of the Izu-Bonin-Mariana Arc, edited by B. Taylor et al., Proc. Ocean Drill. Program, Sci. Results, 126, 627-651, 1992.

Van der Hilst, R., and T. Seno, Effects of relative plate motion on the deep structure and

penetration depth of slabs below the Izu-Bonin and Mariana island arcs, Earth Planet. Sci. Lett. 120, 395-407, 1993.

Watts, A.B., M. Talwani, and J.R. Cochrane, Gravity field of the northwest Pacific Ocean Basin and its margin, in The Geophysics of the Pacific Ocean Basin and its Margin, Geophys. Monogr. Ser., vol. 19, edited by G.H. Sutton, M. H. Manghnani, and R. Moberly, pp. 17-34, AGU, Washington, D.C., 1976.

Weissel, J.K., and Anderson, R.N.. Is there a Caroline plate ? Earth Planet. Sci. Lett., 41, 143-158,1978

John Milsom, Research School of Geological and Geophysical Sciences, Birkbeck College and University College London, Gower Street, London, WC 1E 6BT, England. (e-mail: [email protected])

Robert Hall, SE Asia Research Group, Department of Geology, Royal Holloway College, Egham, TW20 OEX, England (e-mail: [email protected])

Tatang Padmawidjaja, Geological Research and Development Centre, Jalan Diponegoro 57, Bandung, Indonesia.

(Received July 4, 1994; revised May 24, 1995; accepted June 15, 1995.)

Documents

milsom_etal_1996 Gravity fields Halmahera.pdf