smyth, h.r., hamilton, p.j., hall, r., and kinny, p.d., 2007, the

tters 258 (2007) 269–282www.elsevier.com/locate/epsl

Earth and Planetary Science Le

The deep crust beneath island arcs: Inherited zircons reveal aGondwana continental fragment beneath East Java, Indonesia

H.R. Smyth a,⁎, P.J. Hamilton b, R. Hall a, P.D. Kinny b

a SE Asia Research Group, Royal Holloway University of London, Egham, Surrey TW200EX, UKb Department of Applied Geology, Curtin University of Technology, Perth 6845, Australia

Received 21 November 2006; received in revised form 21 March 2007; accepted 21 March 2007

Available on

Editor: R.W. Carlson

line 1 April 2007

Abstract

Inherited zircons in Cenozoic sedimentary and igneous rocks of East Java range in age from Archean to Cenozoic. Thedistribution of zircons reveals two different basement types at depth. The igneous rocks of the Early Cenozoic arc, found along thesoutheast coast, contain only Archean to Cambrian zircons. In contrast, clastic rocks of north and west of East Java containCretaceous zircons, which are not found in the arc rocks to the south. The presence of Cretaceous zircons supports previousinterpretations that much of East Java is underlain by arc and ophiolitic rocks, accreted to the Southeast Asian margin duringCretaceous subduction. However, such accreted material cannot account for the older zircons. The age populations of Archean toCambrian zircons in the arc rocks are similar to Gondwana crust. We interpret the East Java Early Cenozoic arc to be underlain by acontinental fragment of Gondwana origin and not Cretaceous material as previously suggested. Melts rising through the crust,feeding the Early Cenozoic arc, picked up the ancient zircons through assimilation or partial melting. We suggest a WesternAustralian origin for the fragment, which rifted from Australia during the Mesozoic and collided with Southeast Asia, resulting inthe termination of Cretaceous subduction. Continental crust was therefore present at depth beneath the arc in south Java whenCenozoic subduction began in the Eocene.© 2007 Elsevier B.V. All rights reserved.

Keywords: East Java, Indonesia; Gondwana basement; zircon U–Pb SHRIMP

1. Introduction

The character of the deep crust beneath many vol-canic arcs in Southeast Asia is not known. There havebeen few seismic refraction studies, and the deep crustalrocks are often covered by thick sequences of young

⁎ Corresponding author. Presently of CASP, Department of EarthSciences, University of Cambridge, West Building, 181a HuntingdonRoad, Cambridge, CB3 0DH, UK. Tel.: +44 1223 277586; fax: +441223 276604.

E-mail address: [email protected] (H.R. Smyth).

0012-821X/$ - see front matter © 2007 Elsevier B.V. All rights reserved.doi:10.1016/j.epsl.2007.03.044

volcanic and sedimentary rocks. Xenoliths studies haveproven to be extremely useful in other volcanic arcs [1]where they have shown crustal character and magma–crustal interaction but there have been none in Java. InEast Java, Indonesia, the basement has long been con-sidered Cretaceous in age and arc and ophiolitic incharacter. No continental rocks are exposed or reportedon East Java. The assumed character of the crust is basedon small and isolated exposures of basement rocks and ahandful of basement-penetrating exploration wells. U–Pbdating of zircons in this study provides compelling ev-idence for the presence of a continental fragment of

mailto:[email protected]

http://dx.doi.org/10.1016/j.epsl.2007.03.044

270 H.R. Smyth et al. / Earth and Planetary Science Letters 258 (2007) 269–282

Gondwana affinity within the deep crust of East Java.This fragment was sampled by melts feeding Early Ce-nozoic arc volcanism.

2. Geological background

2.1. Regional setting

The island of Java is located within the Indonesianarchipelago between the landmasses of Eurasia andAustralia, on the southeast margin of the Eurasian Plate.The southeastern part of the Eurasian Plate is known asSundaland (Fig. 1), which is the Mesozoic continentalcore of Southeast Asia [2]. Sundaland is not a singlehomogenous block of continental crust or a stable con-tinental block; it was formed by the Early Mesozoicthrough the amalgamation of continental blocks [3,4].During the late Cretaceous terranes of arc and ophioliticmaterial of Cretaceous age were accreted to the southernmargin of Sundaland along a northeast–southwesttrending subduction zone [2]. Subduction moved to itspresent-day position and east–west orientation, south ofJava, along the Java Trench in the Early Paleogene [4].Cenozoic sequences exposed today on East Java restunconformably above terranes accreted in the Cretaceous.

2.2. Character of the basement

The interpreted NE–SW orientation of the Creta-ceous subduction zone is somewhat speculative, due to

Fig. 1. Current plate tectonic setting showing the extent of Sundaland [2] andSunda Arc are shown in black triangles.

limited exposure. The orientation was determined bylinking the exposures of subduction complexes inCentral Java to those in the Meratus Mountains in SEKalimantan (Fig. 2) [2,5]. There is no evidence of a relicsubduction complex cross-cutting the Java Sea, but thismay be due to thick Cenozoic sedimentary cover andlack of publicly available data from deep penetratingseismic surveys.

The character of basement beneathmost of East Java isunknown and surmised from spatially limited data.Exposures of basement rocks in Java are rare, and arefound only at Ciletuh in West Java, Karangsambung andthe Jiwo Hills in the western part of East Java (Fig. 2).These exposures account for less than 1%of the land area.Lithologies exposed are Cretaceous in age and typical ofarc and ophiolitic terranes, including basaltic pillow lavas,cherts, limestone, schists, and metasedimentary rocks[5–8]. In addition to those at Karangsambung there areoccurrences of high pressure metamorphic rocks such asjadeite–quartz–glaucophane bearing rocks and eclogiteinterpreted to be diagnostic of subduction zone metamor-phism [9]. Cretaceous ages have been determined for thebasement rocks atKarangsambung using radiolariawithinthe cherts [7] and K–Ar dates on muscovite from quartz–mica schist ranging from 110 to 124 Ma [5,9].

2.3. Volcanic activity

Java is essentially a volcanic island and two Cenozoicvolcanic arcs are preserved on East Java: the modern arc

the location of the study area of East Java. The modern volcanoes of the

Fig. 2. Reported basement configuration of the southern margin of Sundaland adapted from Hamilton [2], Parkinson et al. [5] and Miyazaki et al. [9].The ages quoted are K–Ar dates for metamorphic material from offshore bore holes [2] and outcrop [5,9].

271H.R. Smyth et al. / Earth and Planetary Science Letters 258 (2007) 269–282

which has been active from the Late Miocene and an arcof Early Cenozoic volcanoes known as the SouthernMountains Arc. The two arcs are separated by a distanceof 50 km, the modern arc occupying the central spine ofthe island and the eroded remnants of the SouthernMountains Arc exposed along the southern coast (Fig. 3).The older arc had been thought to be of Oligo-Mioceneage [10], but recent studies [6] indicate that the arc was

Fig. 3. Locations of samples containing populations of ancient zircon

active from the Middle Eocene until the Early Miocene.The erupted products range from andesite to rhyolite andform lava flows and domes, volcanic breccias, andextensive pyroclastic deposits including flow, fall, andsurge deposits. The arc rocks are intruded by numeroushigh-level igneous bodies that range from Eocene toPliocene in age and basalt to granodiorite in composition.Some of these intrusions are contemporaneous with arc

s. Early Cenozoic volcanic centres are marked as white circles.


activity and others post-date it and are associated with thevolcanism in the modern arc.

In addition to the volcanic and volcaniclastic rocksexposed in the southern part of East Java, there are thickCenozoic sedimentary sequences to the north of the arc.The source of these clastic sediments has long beenassumed to be Sundaland to the north and west[2,11,12].

As the erupted products of the Southern Moun-tains Arc are poorly dated a study was designed toutilise SIMS U–Pb dating of zircons to provide ageconstraints on the development of the arc. In addition itwas hoped that the zircons would provide provenanceinformation on the sedimentary rocks exposed withinand to the north of the arc, which are commonly assumedto be derived from the continental rocks of Sundaland[2,11,12]. This paper reports the unexpected results ofthis study.

2.4. Predicted age range

Assuming that the basement of East Java is, asHamilton[2] suggested, Cretaceous arc and ophiolitic fragments,and knowing the history of Cenozoic arc volcanism onthe island, the predicted zircon age range would beCretaceous to present-day. It was a surprise to discoverthat in addition to this predicted range there are significantnumbers of Cambrian and older zircon grains in many ofthe samples analysed. Below we document the range anddistribution of the old zircon grains identified in theigneous and clastic rocks fromEast Java, and discuss theirpotential origin.

3. Methods

3.1. Samples analysed

We report the results of zircon U–Pb SIMS dating of17 Cenozoic samples of igneous (6 intrusive bodies and4 extrusive rocks), volcaniclastic (n=3) and sedimen-tary (n=4) rocks. The samples are listed in Table 1 andtheir locations shown in Fig. 3. In the samples reportedhere more than 350 spot dates were measured and ofthese 290 were Cretaceous and older.

3.2. Analytical method

Zircon grains were analysed for Pb isotope compo-sition and U, Th and Pb concentrations using theSHRIMP facility at Curtin University of Technology.The results are summarised in Table 1 and further detailsof the samples and analytical methods are given in the

Supplementary material. Dates greater than 1250 Ma arecalculated from 207Pb/206Pb ratios, and from this groupanalyses with greater than 15% discordance were ex-cluded from any age plots. Dates less than 1250 Ma arecalculated from 206Pb/238U ratios as these yield morereliable results because of uncertainties in common Pbcorrections. All grains older than 100 Ma have beencorrected for common Pb using 204Pb, and those youngerthan 100 Ma have been corrected using 208Pb. All errorsquoted in the text are 2σm.

3.3. Probability of missing a population of ages

As this was the first study of its type in East Java,there was some uncertainty about the expected range ofdates. The strategy adopted was to analyse a relativelysmall number of grains (∼15–50) from a large numberof samples, rather than the more conventional approachof analysis of a large number of grains from a smallnumber of samples. As a large range of zircon ages wasobtained, consideration must be given as to whether thepopulation of zircons has been fully sampled. Thisproblem is discussed in detail in Dodson et al. [13] andCawood et al. [14] who show that for detrital zircons 60analyses are necessary to sample the entire populationwith greater than 95% confidence. For any population ofdates that is significant, say greater than 20% of the totalpopulation, there is less than a 3.5% chance that it willbe missed in a sample set comprising only the smallestnumber of the grains analysed in this study (n=15grains). We are thus confident that with 15–35 grainsanalysed the probability of sampling any populationcomprising more than 20% is high.

4. Results

4.1. Age range observed

When examining the entire dataset several distinctpeaks (Fig. 4A) can be identified in the entire popu-lation: a Cenozoic peak (n=55), a Cretaceous peak(n=77), three Cambrian to Proterozoic peaks (n=166)of 500–750 Ma, 900–1250 Ma, 1600–1850 Ma, and anArchean peak of 2500–2800 Ma (n=38). In additionthere are 9 zircon ages ranging from Devonian toJurassic but these do not form a significant peak. Thispaper details the Cretaceous and older zircon results.

4.2. Lithology and zircon ages

The distribution of zircons of different ages is shownon Fig. 4 and discussed in detail below.

Table 1Samples and ages of inherited zircons in East Java

Position Sample Region Lithology Cretaceous Ordovician–Jurassic

Proterozoic–Cambrian

Archean

Igneous Central Java Jhs2KaliS3 Nanggulan, WestProgo Mountains

Basaltic sill 0 1 12 8

Central Java Jhs2AND1 Nanggulan, WestProgo Mountains

Andesiteintrusion

1 0 12 3

SouthernMountains Arc

Jhs2AN7 Pendul Hill, Bayat,Klaten

Diorite 0 0 11 2


Jhs2PON4 Ponorogo Diorite 0 0 9 5


Jhs2Treng6 Trenggelek Diorite 0 0 23 1


Jhs2Jember10 Sukamede, Jember Granodiorite 0 0 14 2


Jhs3Prambanan Prambanan,Yogyakarta

Dacitic ash 0 0 9 1


Jhs2PAC7 Pacitan Dacitic ash 0 0 15 6


Jhs2TUN4 Tulunagung Dacitic pumicebreccia

0 0 19 8


Jhs2Turen14 Turen Andesitic breccia 0 1 11 0

Volcaniclastic Central Java Jhs2KK13 Karangsambung,Central Java

Reworked andesiticbreccia

27 0 1 0

Central Java Jhs2Langse1 Karangsambung,Central Java Coresample provided byCoparex Indonesia

Volcaniclasticsandstone

13 0 0 0


Jhs2Santren1 Batu Agung,Yogyakarta

Crystal-richtuffaceous sandstone

3 0 0 0

Sedimentary Shelf Jhs1-012 Lodan Quarries,Rembang

Quartz-rich sandstone 15 1 0 0

Basin Jhs2Lutut6 Kali Lutut,Semarang


SouthernMountainsArc

Jhs2SERMO6 SermoReservoir, WestProgo Mountains



Jhs2NW12 Nanggulan, WestProgo Mountains Quartz-rich sandstone

8 3 7 1

Total zircons (291) 77 9 166 38No. of samples (17) 8 6 14 11


4.2.1. Igneous rocksThe 6 intrusive and 4 extrusive igneous rocks contain

a range of Cambrian and older zircons (n=155). Thesamples lack Cretaceous zircons (Fig. 4), with the ex-ception of a sample from Central Java (Jhs2AND1)which yielded a single Cretaceous zircon. One samplefrom a basaltic sill in Central Java (Jhs2KaliS3) yieldeda single Carboniferous age. Of the igneous samples,only one extrusive rock from East Java (Jhs2Turen14)failed to yield Archean grains.

4.2.2. Clastic rocksThe 3 volcaniclastic rocks yielded a significant

number of Cretaceous zircons (n=42) but only the

westernmost sample (Jhs2KK13) yielded older zircons,a single grain of Mesoproterozoic age (1079±36 Ma).All 4 of the sedimentary rocks analysed were quartz-richsandstones. These sandstones yielded varied zirconages. All of the sandstones contain Cretaceous grains(n=33) (Fig. 4B). The northernmost sandstone samplefrom the edge of the Sunda Shelf (Jhs1-012) contains asingle Jurassic–Triassic grain, but no older grains. Theother sandstones from the south contain a range ofCambrian and older grains (n=22).

4.2.3. Spatial distribution of zircon datesThe samples containing different zircon dates do

have a clear geographical pattern. The Archean grains

Fig. 4. A. Age range of zircons (bins on the histogram represent 25 Ma and the values are plotted with 2σ errors) B. Samples with Cretaceous zircons.C. Samples with Ordovician to Jurassic zircons. D. Samples with Proterozoic to Cambrian zircons. E. Samples with Archean zircons.


are found only in rocks of the Southern Mountains Arc(Fig. 4E) and the igneous rocks of this area containno Cretaceous zircons. Cretaceous zircons are presentonly in rocks located in the west and north (Fig. 4B).Cretaceous zircons have been identified only in sedi-mentary and volcaniclastic rocks and one igneous rockfrom the western part of the study area. This distributionis interpreted to reveal the character of the crust beneathEast Java.

5. Discussion

5.1. Origin and incorporation of the zircons into therocks of East Java

The age range identified in the samples from EastJava is much greater than would be predicted whetherderived from a Sundaland source or local basementsource of Cretaceous arc and ophiolitic rocks.


5.1.1. Possible Sundaland SourcesSundaland, the continental core of Southeast Asia, is

the closest and most obvious source of very old zircons(Fig. 5). Rocks which could have provided abundantzircons are granites of southwest Borneo [2], the MalayPeninsula [15], the Thai–Malay Tin Belt [16] andSumatra [17,18]. These regions could have been asource of Mesozoic and Paleozoic material, but noArchean rocks are known, nor is there any indicationthat they are underlain by Archean crust.

There are abundant granites of Cretaceous agelocated in the Schwaner Mountains and the Thai–Malay Tin Belt which are known to have contributedmaterial to the Paleogene sedimentary sequences of

Fig. 5. Potential source areas for Cambr

Northern Borneo [19]. If these source regions alsoprovided material to East Java, the zircon populationswould be similar. The zircon populations of theseregions are dominated by Cretaceous zircons with agepeaks different from those in East Java, there are abun-dant Permian–Triassic zircons (Fig. 6C), which are rarein East Java, and Precambrian zircons are mainlyPaleoproterozoic with very rare Archean zircons [19].

In addition to the differences in zircon populations,there are paleogeographic factors to be considered. Theisland of Java is separated from Borneo by the shallowwaters of the Java Sea. Beneath the flat-lying siliciclas-tic and carbonate Pleistocene deposits of the SundaShelf there are numerous broad NE–SW trending ridges

ian and older zircons in East Java.

Fig. 6. Comparison of the Cambrian–Archean data set from East Java with sediments of the Perth Basin Australia and Northern Borneo. A. East Javadata set. B. Young placer deposits of the Perth Basin, known to be the product of erosion of the Proterozoic and Archean rocks of Western Australia[34]. C. Range of detrital zircon ages for Paleogene sandstones from Northern Borneo [19]. Bins on the histogram represent 25 Ma and the values areplotted with 2σ errors.



flanked by narrow deep Cenozoic basins. During theearlier part of the Cenozoic the ridges and basins wouldhave acted as barriers or traps preventing sedimentpassing to the south and east from Sundaland into EastJava. Indeed the ridges themselves may have acted assediment source areas. One such ridge on the edge ofSundaland, the Karimunjawa Arch (Fig. 5), is reportedto have been elevated throughout much of the LateEocene to Late Miocene [20,21]. The elevation of thearch has been interpreted to be responsible for thedifferent sedimentary environments recorded in the Eastand West Java Seas. To the east of the arch, the LowerMiocene and Oligocene sediments are marine, but to thewest most of the Lower Miocene and all of theOligocene sediments are non-marine [20,21]. TheKarimunjawa Arch is reported to have been a sourceof sediment to both the East and West Java Sea Basinsand potentially provided sediment to the area of present-day Java between the Late Eocene and Late Miocene.There are exposures of metasedimentary rocks on smallislands on the arch but none are dated as they areterrestrial. They are interpreted to be pre-Cenozoic inage and part of the continental basement of Sundaland.This basement could be the source of the Cretaceouszircon grains in East Java's clastic sequences.

5.1.2. Possible Gondwana SourcesThe Archean–Cambrian peaks identified coincide

with periods of major crustal growth represented by thePan African and Grenville age orogens and the Archeancratons [22,23]. These peaks are characteristic of muchof the Gondwana margin as recorded in the rocks ofsouthern and east India [24], eastern and northernAntarctica [23,25,26], eastern Africa [22,23] andwestern Australia [14,27–29]. The rocks of Sundalandare not known to share this history. Plate tectonicreconstructions of Gondwanaland prior to and followingbreak-up indicate that it is unlikely that either Antarcticor African material could have contributed to Sundaland[30–32]. The largest and closest area of Archean conti-nental crust is Australia, which formed part of Gond-wana until the Cretaceous.

Basement rocks in Western Australia are of Protero-zoic and Archean age (Fig. 6B). As these ancient rockshave been the focus of detailed zircon dating studies,many of the basement blocks have age fingerprints[28,29,33,34]. The rocks at the surface in WesternAustralia are representative of much of the Gondwanamargin. A large published dataset [34] from WesternAustralia was chosen to compare with that obtainedfrom East Java. The dataset [34] focuses on the youngplacer sediments of the Perth Basin, which are the

eroded products of Precambrian basement rocks whichare currently exposed in Western Australia. The detritalzircons dated from these placer deposits have threemajor age peaks and these can be correlated with knownsource areas [28,29,34]. The Cambrian to Neoproter-ozoic zircons (500–800 Ma) were produced by theerosion of the Pan African Pinjarra Orogenic Belt, theMesoproterozoic zircons (1000–1300 Ma) were derivedfrom the Grenvillian Albany Fraser Orogenic Belt, andthe Archean zircons (2500–3200 Ma) were eroded fromthe Yilgarn and Pilbara Cratons. There are also a limitednumber of dates which correspond to the formation ofthe Capricorn Orogenic Belt (1600–2000 Ma).

A range of Archean and Proterozoic peaks has alsobeen reported from the Higher Himalayan Gneisses inIndia [35]. Yoshida and Upreti [35] suggest that thesepeaks represent terranes derived from a Circum-EastAntarctic origin, which could have included part ofGreater India. However, the peaks identified do notmatch those reported in this study.

The Western Australia peaks are remarkably similarto those identified in the Southern Mountains Arc ofEast Java (Fig. 6). Given this correlation it is plausiblethat Western Australia or a region with a similar geo-logical history, such as other parts of the Gondwanamargin, was the source of material.

5.2. Incorporation of the zircons

Zircon is a highly refractory mineral. Zircon is knownto survive numerous sedimentary cycles, metamorphismand diagenesis [36]. The zircon grains identified in thesedimentary rocks of East Java could have beenincorporated through erosion and transport of sedimentfrom exposed source areas, such as the local basement,the Karimunjawa Arch or from volcanic sources by air-fall and subsequent reworking. The incorporation ofzircons into the igneous rocks requires more explanation.As well as being resistant to weathering, zircon isknown to survive episodes of melting [37]. Material ofGondwana origin must have been introduced to the meltsfeeding the Southern Mountains Arc. There are severalmechanisms whereby this introduction could occur.

5.2.1. Subducted sedimentSediment eroded from western Australia deposited

on the Indo-Australian plate could have been carriednorth to the Java Trench as Australia moved north.However, even today a fan similar in size to the modernBengal Fan would be required to bring material fromAustralia to the site of subduction at the Java Trench(Fig. 5). During the Early Cenozoic Australia lay very

Fig. 7. Plate tectonic reconstruction of SE Asia and Australia at 45 Ma [4]. The light grey shaded area around Australia shows the approximate extentof continental crust based on present-day bathymetry and mapping of the continental–ocean boundary. The ornamented area west of westernAustralia indicates the surmised area of continental crust that rifted from Australia in the Jurassic and early Cretaceous. We suggest that the East Javafragment probably originated in the southern part of this area.


far south of its current position (Fig. 7), and at 45 Ma thedistance between Australia and East Java would haveexceeded 1900 km [4]. If the fan model were correct, thesedimentary sequences on the Indian–Australian platecurrently entering the Java Trench would correspond toa relatively proximal position on the Cenozoic fan.However, today there is little sedimentary cover on theIndian–Australian plate south of Java [38]. In addi-tion the modern Bengal Fan is fed by erosion of theHimalayas, and there is no evidence of a similar moun-tain belt in NWAustralia during the Cenozoic.

Sediment of Gondwana origin could have beenintroduced into the Java Trench by axial transport fromthe Bird's Head microcontinent, New Guinea or fromthe Himalayas via the Bengal Fan, although both areunlikely. The basement rocks of the Bird's Head micro-continent were not available for erosion as the Bird'sHead was the site of active carbonate sedimentationduring most of the Cenozoic and so can be excluded as a

source [39]. If sediment was produced by the erosion ofthe Himalaya it would require transport via the BengalFan to the Java Trench and total transport distanceswould exceed 4000 km. In addition to this significantdistance, there were topographic barriers at the Inves-tigator Fracture and the Ninety East Ridge which wouldhave hampered the flow of sediment. The lack of BengalFan material today in the Java Trench offshore East Java[38] adds to the argument that this is not a plausiblemechanism. It is equally unlikely that sediment fromthe Himalayan region could travel a distance of over3500 km from Indochina across Sundaland, by-passingits deep basins, structural highs and ridges [40].

5.2.2. Subducted continental fragmentThe subduction of a fragment of continental crust

also seems unlikely. Buoyancy forces make subductionof continental crust difficult and even if these wereovercome, the fragment is required to have supplied


melts from the Eocene to the Miocene, a period ofaround 20 Ma. The fragment either remained stationarybeneath the Southern Mountains or was 1500 km inlength based on the present subduction rate of 75 mm/yr[41] and Cenozoic plate tectonic reconstructions [4].Neither of these scenarios is likely.

5.2.3. Collision/accretion of a continental fragmentAn alternative, simpler and more plausible explana-

tion is that there is a fragment of continental crust ofGondwana character beneath the Southern MountainsArc. This fragment must have been in place beforesubduction began in its current position along the JavaTrench. We suggest that the fragment collided with themargin of Sundaland terminating the Late Cretaceousphase of subduction. The origin of this fragment isdiscussed in detail below. Magmas rising through, orpooling in, the continental crust would lead to partialmelting or assimilation of material.

The presence of continental crust is supported by theevolved character of the products of the SouthernMountains Arc and the rocks which intrude them. Thevolcanic rocks range fromandesite to rhyolite (60–77wt.%SiO2) with an estimated average SiO2 content of 67 wt.%[6], and a significant proportion of the high-level intrusivebodies are granodiorites. These rocks are considerablymore evolved than the basic-intermediate eruptive productsof the present-day volcanoes. The chemistry of the modernarc volcanic rocks in East Java [42–44] indicates they arethe products of relatively primitive subduction meltshaving no significant interaction with underlying crust.The difference in chemistry of the products of the SouthernMountains Arc and the Sunda Arc could be explained in anumber of ways: fractional crystallization of basic magma,or contamination by melting of felsic continental basementcrust (C. Macpherson, 2006, pers. comm.). The evidencefrom zircon dating suggests that partial melting of con-tinental basement is the most likely explanation.

The distribution of pre-Cenozoic zircons (Fig. 4) andthe character of the arc rocks indicate that the conti-nental fragment is present only beneath the SouthernMountains whereas north of the Southern Mountainsthere is no continental basement. The deep crust to thenorth of the arc is probably Cretaceous arc or ophioliticterranes, as suggested by Hamilton [2], representingmaterial accreted during Cretaceous subduction orunderplated during collision.

Sribudiyani et al. [45] also inferred the presence of acontinental fragment beneath East Java. Their interpre-tation is based on the distribution of quartz-rich sand-stones assumed to have a continental origin. A detailedstudy of the provenance of these sandstones raised

doubts about the continental source as many of thesandstones contain considerable acidic volcanic mate-rial, probably derived from the Southern Mountains Arc[6]. Thus, although we agree with the interpretation ofSribudiyani et al. [45] of a continental fragment beneathEast Java, we dispute the evidence on which theirinterpretation is based, and consequently differ on thetiming of arrival of the continental fragment.

6. Character and origin of the crust beneath EastJava

6.1. Character of the crust

The distribution of pre-Cenozoic zircons in samplesfrom East Java provides new information about thecharacter of the deep crust about which little was pre-viously known. We suggest that the most likely expla-nation for the zircon age range sampled in East Java is thatthere is a fragment of continental crust of Archean age,Gondwana affinity and potentially of Western Australianorigin at depth beneath the Southern Mountains.

The size of the East Java fragment can be estimatedby utilising the spatial distribution of the different zirconage populations and other supporting evidence. Cur-rently none of the data suggests that the continentalfragment extends north beneath the modern Sunda Arc.Therefore, the northern boundary must be located northof the Southern Mountains but south of the modernSunda Arc (Fig. 8). The character of the crust beneaththe modern arc is still unclear, but there is no indicationof a continental contribution to the magmas feeding thearc [e.g. [42]]. The deep crust is likely to be similar incomposition to the Cretaceous arc and ophiolitic frag-ments known from exposures in Central Java.

The lack of Cretaceous material within the igneousrocks of the Southern Mountains Arc to the west of theYogyakarta area suggests that a terrane boundary exists inthis region between Cretaceous crust and the East Javacontinental fragment. This boundary is interpreted to co-incide with a major northeast–southwest trending struc-ture identified at approximately 110.5°E [6] (Fig. 8B).This lineament links several major features onshore andoffshore, a structural high in the Java Forearc, and thealigned centres of three Oligo-Miocene volcanoes [10] inthe West Progo Mountains, Mount Merapi and the un-usual K-rich back-arc volcano of Mount Muria. In theregion close to the boundary between the Cretaceous crustand the older Gondwana fragment mixing of zircons ofdifferent ages could easily occur through assimilation orpartial melting of the crust during magma ascent, or upliftand erosion of the basement rocks.

Fig. 8. Character of the crust on the southeastern margin of Sundaland.A. Extent of the continental fragment onshore East Java. B. Mapshowing the proposed extent of the East Java fragment beneathSulawesi. C. Sketch cross-section.


Terrane boundaries are rarely straight or uniformfeatures, but are embayed and irregular, and their shapeis dictated by original rift morphology and collisiontectonics. Therefore the margins of the East Java frag-ment could be structurally complex.

6.2. Origin and extent of the fragment

Several fragments rifted off the Australian marginduring theMesozoic in a phase of break-up preceding theseparation of India from Gondwana [3,46,47]. We sug-gest that one such fragment or fragments collided withthe margin of Sundaland during the Late Cretaceousforming the basement to the Southern Mountains Arc.Prior to the collision, arc and ophiolitic terranes wereaccreted and formed a belt on to the margin of Sundaland(Fig. 8). The ophiolitic blocks may be remnants of IndianOcean crust which preceded the continental fragment onthe northward-moving plate.

The timing of collision of the fragment pre-dates theMiddle Eocene as continental crust was already presentbeneath the arc when subduction began at that time. Wesuggest that the collision of this fragment resulted in thetermination of Cretaceous subduction. Following therenewal of subduction during the Middle Eocene, theCenozoic melts feeding the Southern Mountains Arcwere contaminated by interaction with the continentalfragment.

Manur and Barraclough [48] indicate the presence oftwo fragments (Fig. 8B), in the areas of the BaweanArch and Paternoster Platform, of Pre-Cenozoic conti-nental crust (the shape, orientation and age of which arevery poorly constrained) in the Java Sea. These occurbetween the belt of ophiolitic and arc blocks and theEast Java fragment. It is possible that these fragmentsare a continuation of the East Java fragment, but as thereis no evidence of continental crust to the north of theSouthern Mountains, beneath the Sunda Arc, it is dif-ficult to envisage how these fragments could be linked.

The eastern extent of the East Java fragment is notknown. However, in Sulawesi continental rocks of similarage to those identified in the East Java fragment have beenreported [49]. Geochemical and zircon dating evidencefrom the Malino Metamorphic complex of northwestSulawesi reveals the presence of Archean material withages up to 3500Ma [49,50].We suggest that the East Javafragment extends to the northwest, beneath westernSulawesi, forming a large arcuate block on the southeastmargin of Sundaland (Fig. 8B). Geochemical analysis ofthe modern volcanic products on Java [42–44] suggeststhat the magmas are not sampling continental crust.However, the rocks analysed have been basic, and we


suggest that a study of the more evolved volcanic rockswould be profitable. Basic rocks are not suitable for astudy of the kind reported here as they contain only rare orvery small zircons. The study of more evolved volcanicrocks from volcanoes between East Java and northwestSulawesi could identify the extent of the fragment.

7. Conclusions

This is the first firm evidence of a Gondwana con-tinental fragment in this part of Southeast Asia. Prior tothis study little was known about the character of thecrust beneath much of Java and in particular nothing wasknown about the crust beneath the Paleogene arc. Thecontinental fragment beneath the Southern Mountains ofEast Java was identified by the distribution of ancientzircons in the Cenozoic igneous and sedimentary rocks.U–Pb dating of zircons from East Java has alloweddistinct basement types to be mapped. In the north andwest there is Cretaceous accreted arc and ophioliticmaterial and in the south along the Southern MountainsArc there is compelling evidence for a fragment ofGondwana continental crust. The rising Cenozoic meltsfeeding the Southern Mountains Arc mixed with thecontinental crust and picked up ancient zircons throughassimilation or partial melting. This fragment could belarge and may extend beneath Sulawesi.

The zircons dated in southeast Java are very similarin age to those of the Gondwana terranes of India andAustralia. Mesozoic reconstructions of the region showa history of Mesozoic rifting of Gondwana fragmentsfrom northwest Australia and this is thought to be theultimate source of the zircons in the igneous and sedi-mentary rocks of East Java.

Acknowledgements

This project was funded by the SE Asia ResearchGroup at Royal Holloway, supported by an oil companyconsortium. Financial assistance for SHRIMP datingwas provided by CSIRO, Australia. We are grateful toGRDC, ITB, LIPI, Eko Budi Lelono, Peter Lunt, ColinMacpherson and Heather Handley. Jhs2Langse1 samplewas provided by Lundin Banyumas. We thank Jason Aliand an anonymous reviewer for helpful comments onthe manuscript.

Appendix A. Supplementary data

Supplementary data associated with this articlecan be found, in the online version, at doi:10.1016/j.epsl.2007.03.044.

References

[1] R. Hickey-Vargas, M.J. Abdollahi, M.A. Parada, L. López-Escobar, F.A. Frey, Crustal xenoliths from Calbuco Volcano,Andean Southern Zone: implications for crustal composition andmagma–crust interaction, Cont. Min. Pet. 119 (1995) 311–344.

[2] W. Hamilton, Tectonics of the Indonesian Region, USGSProfessional Paper, vol. 1078, 1979 345 pp.

[3] I. Metcalfe, Gondwanaland dispersion, Asian accretion andevolution of eastern Tethys, in: Z.X. Li, I. Metcalfe, C.M. Powell(Eds.), Breakup of Rodinia and Gondwanaland and Assembly ofAsia, vol. 43, 1996, pp. 605–624.

[4] R. Hall, Cenozoic geological and plate tectonic evolution of SEAsia and the SW Pacific: computer-based reconstructions, modeland animations, J. Asian Earth Sci. 20 (2002) 353–434.

[5] C.D. Parkinson, K. Miyazaki, K. Wakita, A.J. Barber, D.A.Carswell, An overview and tectonic synthesis of the pre-Tertiaryvery-high-pressure metamorphic and associated rocks of Java,Sulawesi andKalimantan, Indonesia, IslandArc 7 (1998) 184–200.

[6] H. Smyth, Eocene to Miocene basin history and volcanic activityin East Java, Indonesia, PhD-Thesis University of London, 2005,476 pp.

[7] K. Wakita, B.W. Munasri, Cretaceous radiolarians from the Luk-Ulo Melange Complex in the Karangsambung area, Central Java,Ind. J. Southeast Asian Earth Sci. 9 (1994) 29–43.

[8] K. Wakita, Cretaceous accretionary — collision complexes incentral Indonesia, J. Asian Earth Sci. 18 (2000) 739–749.

[9] K. Miyazaki, J. Sopaheluwakan, I. Zulkarnain, K. Wakita, Ajadeite–quartz–glaucophane rock from Karangsambung, centralJava, Indonesia, Island Arc 7 (1998) 223–230.

[10] R.W. van Bemmelen, The Geology of Indonesia, Govt. PrintingOffice, Nijhoff, The Hague, 1949, 730 pp.

[11] E. Sharaf, J.A. Simo, A.R. Carrol, M. Shields, Stratigraphicevolution of Oligocene–Miocene carbonates and siliciclastics,East Java basin, Indonesia, AAPG Bull. 89 (2005) 799–819.

[12] W. Ardhana, A depositional model for the early Middle MioceneNgrayong Formation and implications for exploration in the EastJava basin, 22nd Annual Indonesian Petroleum AssociationConvention, Jakarta, 1993, pp. 395–443.

[13] M.H. Dodson, W. Compston, I.S. Williams, J.F. Wilson, A searchfor ancient detrital zircons in Zimbabwean sediments, J. Geol.Soc. (Lond.) 145 (1988) 977–983.

[14] P.A. Cawood, A.A. Nemchin, A. Leverenz, A. Saeed, P.F.Ballance, U/Pb dating of detrital zircons: implications for theprovenance record of Gondwana margin terranes, Geol. Soc.Amer. Bull. 111 (1999) 1107–1119.

[15] T.C. Liew, R.W. Page, U–Pb zircon dating of granitoid plutonsfrom the West Coast of Peninsular Malaysia, J. Geol. Soc.(Lond.) 142 (1985) 515–526.

[16] E.J. Cobbing, D.I.J. Mallick, P.E.J. Pitfield, L.H. Teoh, Thegranites of the Southeast Asian tin belt, J. Geol. Soc. (Lond.) 143(1986) 537–550.

[17] W.J. McCourt, M.J. Crow, E.J. Cobbing, T.C. Amin, Mesozoicand Cenozoic Plutonic Evolution of SE Asia: evidence fromSumatra, Indonesia, in: R. Hall, D.J. Blundell (Eds.), TectonicEvolution of SE Asia, Geol. Soc. London Spec. Publ., vol. 106,1996, pp. 321–335.

[18] Imtihanah, Isotopic Dating of the Sumatran Fault System, MPhil-Thesis University of London, 150 pp.

[19] M.W.A. van Hattum, R. Hall, A.J. Pickard, G.J. Nichols, SoutheastAsian sediments not from Asia: provenance and geochronology ofnorth Borneo sandstones, Geology 34 (2006) 589–592.

http://dx.doi.org/doi:10.1016/j.epsl.2007.03.044

http://dx.doi.org/doi:10.1016/j.epsl.2007.03.044


[20] W.P. Bishop, Structure, stratigraphy and hydrocarbons offshoresouthern Kalimantan, Indonesia, AAPG Bull. 64 (1980) 37–58.

[21] M.C. Cater, Stratigraphy of the offshore area South of Kalimantan,Indonesia, 10th Annual Indonesian Petroleum Convention Jakarta,1981, pp. 269–284.

[22] J.J. Veevers, Pan-African is Pan-Gondwanaland: oblique con-vergence drives rotation during 650–500 Ma assembly, Geology31 (2003) 501–504.

[23] J. Jacobs, C.M. Fanning, F. Henjes-Kunst, M. Olesch, H.J. Paech,Continuation of the Mozambique Belt into East Antarctica:Grenville-age metamorphism and polyphase Pan-African high-grade events in central DronningMaud Land, J. Geol. 106 (1998)385–406.

[24] M. Santosh, T. Morimoto, Geochronology of the khondalite beltof Trivandrum Block, Southern India: electron probe ages andimplications for Gondwana tectonics, Gondwana Res. 9 (2006)261–278.

[25] I. Fitzsimons, A review of tectonic events in the East Antarcticshield and their implications for Gondwana and earlier super-continents, J. Asian Earth Sci. 31 (2000) 3–23.

[26] W.A. Gose, M.A. Helper, J.N. Connelly, F.E. Hutson, I.W.D.Dalziel, Paleomagnetic data and U–Pb isotopic age determina-tions from Coats Land, Antarctica: implications for lateProterozoic plate reconstructions, J. Geophys. Res. 102 (1997)7887–7902.

[27] J.J. Veevers (Ed.), Billion year earth history of Australia andneighbours in Gondwanaland, GEMOC Press, Sydney, 2000,400 pp.

[28] S.D. Pell, I.S. Williams, A.R. Chivas, The use of protolith zircon-age fingerprints in determining the protosource areas for someAustralian dune sands, Sediment. Geol. 109 (1997) 233–260.

[29] P.A. Cawood, A.A. Nemchin, Provenance record of a rift basin:U/Pb ages of detrital zircons from the Perth Basin, WesternAustralia, Sediment. Geol. 134 (2000) 209–234.

[30] S.K. Acharyya, Break-up of the Greater Indo-Australiancontinent and accretion of blocks framing south and east Asia,J. Geod. 26 (1998) 149–170.

[31] S.E. Bryan, A.E. Constantine, A. Ewart, R.W. Schon, J. Parianos,Early Cretaceous volcano-sedimentary successions along theeastern Australian continental margin: implications for the break-up of eastern Gondwana, Earth Planet. Sci. Lett. 153 (1997)85–102.

[32] C. Scotese, A continental drift flipbook, J. Geol. 112 (2004)729–741.

[33] P.A. Cawood, A.A. Nemchin, M. Freeman, K.N. Sircombe,Linking source and sedimentary basin: detrital zircon record ofsediment flux along a modern river system and implications forprovenance studies, Earth Planet. Sci. Lett. 210 (2003) 259–268.

[34] K.N. Sircombe, M.J. Freeman, Provenance of detrital zircons onthe Western Australian coastline — implications for the geo-logical history of the Perth basin and denudation of the Yilgarncraton, Geology 27 (1999) 879–882.

[35] M. Yoshida, B.N. Upreti, Neoproterozoic India within EastGondwana: constraints from recent geochronologic data fromHimalaua, Gondwana Res. 10 (2006) 349–356.

[36] M.A. Mange, H.F.W. Maurer, Heavy Minerals in Colour,Chapman and Hall, London, 1992 147 pp.

[37] J.M. Hanchar, P.W.O. Hoskin (Eds.), Zircon, MineralogicalSociety of America and Geochemical Society, Washington, 2003,500 pp.

[38] D.G. Masson, L.M. Parson, J. Milsom, G. Nichols, N. Sikumbang,B. Dwiyanto, H. Kallagher, Subduction of seamounts at the JavaTrench: a viewwith long-range sidescan sonar, Tectonophysics 185(1990) 51–65.

[39] G.J. Carman, New paleogeographic maps for the Coral Sea andDarai megasequences, Papua New Guinea, AAPG Bull. 76(1992) 1094.

[40] R. Hall, C.K. Morley, Sundaland Basins, in: P. Clift, P. Wang, W.Kuhnt, D.E. Hayes (Eds.), Continent–Ocean Interactions withinthe East Asian Marginal Seas, AGU Geophysical Monograph,vol. 149, American Geophysical Union, 2004, pp. 55–85.

[41] R. McCaffrey, Slip partitioning at convergent plate boundaries ofSEAsia, in: R. Hall, D.J. Blundell (Eds.), Tectonic Evolution of SEAsia, Geol. Soc. London Spec. Publ., vol. 106, 1996, pp. 3–18.

[42] H. Handley, Geochemical and Sr–Nd–Hf–O isotopic constraintson volcanic petrogenesis at the Sunda arc, Indonesia, PhD-Thesis, University of Durham, 2006, 206 pp.

[43] I.A. Nicholls, D.J. Whitford, K.L. Harris, B. Taylor, Variation inthe geochemistry of mantle sources for tholeiitic and calc–alkaline mafic magmas, western Sunda volcanic arc, Indonesia,Chem. Geol. 30 (1980) 177–199.

[44] G.E. Wheeler, R. Varne, J.D. Foden, M.J. Abbot, Geochemistry ofQuaternary volcanism in the Sunda–Banda arc, Indonesia, andthree-component genesis of island-arc basalt magmas, J. Volcanol.Geotherm. Res. 32 (1987) 137–160.

[45] N. Sribudiyani, R. Muchsin, T. Ryacudu, P. Kunto, I. Astono, B.Prasetya, S. Sapiie, A.H. Asikin, I. Harsolumakso, Yulianto, TheCollison of the East Java Microplate and its implications forhydrocarbon occurrences in the East Java Basin, 29th Annu-al Indonesian Petroleum Association Convention Jakarta, 2003,pp. 335–346.

[46] R.D. Muller, C. Gaina, S. Clarke, Seafloor spreading aroundAustralia, in: J.J. Veevers (Ed.), Billion-year earth history ofAustralia and neighbours in Gondwanaland, GEMOC Press,Sydney, 2000, pp. 18–28.

[47] C. Heine, R.D. Müller, C. Gaina, Reconstructing the lost EasternTethysOceanBasin: constraints for the convergence history of theSE Asian margin and marine gateways, in: P. Clift, P. Wang, w.Kuhnt, D.E. Hayes (Eds.), Continent–Ocean Interactions withinthe East Asian Marginal Seas, AGU Geophysical Monograph,vol. 149, American Geophysical Union, 2004, pp. 37–54.

[48] H. Manur, R. Barraclough, Structural control on hydrocarbonhabitat in the Bawean area, East Java Sea, 23rd Annual IndonesianPetroleum Association, Convention, Jakarta, 1994, pp. 129–144.

[49] T.M. van Leeuwen, Muhardjo, Stratigraphy and tectonic settingof the Cretaceous and Paleogene volcanic–sedimentary succes-sions in northwest Sulawesi, Indonesia: implications for theCenozoic evolution of Western and Northern Sulawesi, J. AsianEarth Sci. 25 (2005) 481–511.

[50] T.M. VanLeeuwen, C.M. Allen, A. Kadarusman, M. Elburg, M.J.Palin, Muhardjo, Suwijanto, Petrologic, isotopic, and radiometricage constraints on the origin and tectonic history of the MalinoMetamorphic complex, NW Sulawesi, Indonesia, J. Asian EarthSci. 29 (2007) 751–777.

smyth, h.r., hamilton, p.j., hall, r., and kinny, p.d., 2007, the

Documents