sign in sign up
<<
Home , Accretion (geology), Australian Plate, Halmahera Arc, Pacific Plate, Philippine Mobile Belt, Philippine Sea Plate

, VOL. 14, NO. 5, PAGES, 1117-1132, OCTOBER 1995

Cenozoicmotion of the Plate: Palaeomagnetic evidence from eastern

Robert Hall SE AsiaResearch Group, Department of ,Royal Holloway, London University, Egham, England United Kingdom

Jason R. Ali Departmentof Oceanography,Southampton University, Southampton, England, United Kingdom

Charles D. Anderson Departmentof GeologicalSciences, University of ,Santa Barbara

Abstract. The history of motion of the is understandingits evolution (Deep Sea Drilling Project poorly known becauseit is isolated from the oceanic ridge (DSDP)/OceanDrilling Program(ODP) Legs 58, 59, 60, 125, system.Interpretation of palaeomagneticresults from the plate and 126). Despitethis, it remainsa platewhose Tertiary motion has been controversial because declination data have been is poorlyknown. The principalobstacle to deducingthe history obtainedonly from the easternmargin where -related of motion is the isolationof the PhilippineSea Plate from the tectonicprocesses may have causedlocal rather than plate-wide global plate circuit since it is surroundedby subductionzones. rotations.New palaeomagneticdata relevantto the problemhave For this reasonits origin and movementhistory have remaineda been obtainedfrom 34 sites north of the SorongFault and 29 sourceof controversy.It has been suggested[Uyeda and Ben- sites within the Sorong system.These sites record south- Avraham, 1972] that the plate was formed when subductionof ward movement during the and northward movement the was initiated at -45 Ma along transformfaults duringthe .Sites within the SorongFault systemrecord within old floor, trapping the former spreading both counterclockwiseand clockwiserotations interpreted as the centrein the West PhilippineBasin, and this view has continued resultof Neogeneblock movementsat the southernboundary of to be cited in interpretationsof the plate's history[e.g., Stem and the Philippine Sea Plate. North of the , all sites Bloomer, 1992]. record clockwise declinations. Neogene rocks have small Palaeomagnetismmay help providea solutionto the problem. deflections consistent with rotation about the present-day Palaeomagneticdata indicate northwardmovement of the plate Eurasia-Philippine Sea Plate pole. -middle Eocene although the amount of rotation is less certain. The rotation rocks show consistent clockwise declination deflections of-40 ø. historyof the plate has a bearingon its origin, on the origin of Declinations of lower Eocene rocks indicate -90 ø of clockwise subduction-relatedmagmatism, and on its movement history, rotation.We proposethat the entire area north of the Sorong because there are hemisphere ambiguities in palaeolatitudes Faultin eastIndonesia has always been part of the PhilippineSea obtained from inclination data especiallyfor low-latitude sites. Plate and that the whole plate has rotated clockwise in a Most ocean drilling data lack declinationinformation, and land discontinuousmanner by approximately90 ø since the early palaeomagneticdata from the plate remain scarce[Haston and Eocene.The new data from north of the SorongFault providea Fuller, 1991]. Although many land studies have reported basis for determining rotation poles which satisfy all the declinationshifts implying clockwiserotations, the interpretation palaeomagneticdata from the PhilippineSea Plate and permit its of theseresults has been disputed.All samplingsites have been at reconstruction. the eastern edge of the plate where local subduction-related tectonic processesprovide a potential explanationfor observed Introduction rotations [Larson et al., 1975; McCabe and Uyeda, 1983] becauseseveral episodes of back arc spreadinghave extendedthe The Philippine Sea Plate (Figure 1) has been the sourceof plate's eastern margin since the Eocene. Distinguishinglocal numerousmodels associated with subduction-relatedprocesses of tectonic from plate-wide movementshas proved controversial. .Different partsof the plate have beenused as Haston and Fuller [ 1991] arguedthat since the middle Eocene examples for tectonic processesincluding initiation of the Philippine Sea Plate has behavedas a rigid unit with little subduction,development of intraoceanicsubduction zones, arc local tectonic deformation along its eastern margin, rotated rifting, back arc spreading,and forearc extension.There have continuouslyclockwise through 90 ø- 100 ø, and moved northward been two major phases of ocean drilling devoted to about 15ø-20ø. Koyama et al. [1992] reviewed tectonic models which could accountfor their own and older data. They accepted northwardmotion of the plate but concludedthat more data were Copyright1995 by theAmerican Geophysical Union. required to distinguish between models which accounted for Papernumber 95TC01694. declinationdeflections by rotation of the whole plate and those 0278-7407/95/95TC-01694510.00 which attributedpart of the deflectionsto local deformation.

1117 1118 HALL ET AL.: CENOZOICMOTION OF THE PHILIPPINESEA PLATE

EURASIAN

30'N

Shikoku PLATE PACIFIC

Dakin Donin I•iand•

Plateau, PLATE

Plateau

Plateau 20'N

• • Pamce • • Vela

Dakin II # PhilippineBasin

10øN

Sulu Caroline ,Sea O Pala Ridge

CAROLINE

Sea PLATE Eauripik Ridge

reo

...... •...... Di.•rnarck DandaSea AUSTRALIAN sea . PLATE

Figure 1. Principaltectonic features of the PhilippineSea Plateand surroundingareas. Double lines represent activeor formerspreading centres within the plate. The presentboundary between the PhilippineSea Plate and the AustralianPlate is a complexstrike-slip plate boundaryzone, with principallysinistral motion, which has been muchsimplified on the map.

In this study we presentnew palaeomagneticdata from the tectonicsyntheses of the regionalthough the Indonesianislands southernpart of the Philippine Sea Plate in easternIndonesia between Halmahera and Waigeo together make up the largest which are relevantto the historyof motionof the plate duringthe land area within the plate: -5x 104km2 spreadover past 50 m.y. This part of the plate has largely been ignoredin -25 x 104km 2. Our new data support large clockwise rotations of HALL ET AL.: CENOZOIC MOTION OF THE PHILIPPINE SEA PLATE 1119

the whole plate but indicate a more complex history of rotation Molucca Sea collision complex(Figure 2) where the Halmahera and latitudinalmovement than previouslysuggested. Using our and Sangibe Arcs are actively converging.The Molucca Sea new data, earlier palaeomagneticdata, and geologicalarguments, Plate has an inverted U-shapedconfiguration [Hatherton and it is possibleto estimate the position of rotation poles which Dickinson, 1969; Katili, 1975; Hamilton, 1979; Cardwell et al., permit reconstructionsof the Philippine Sea Plate and adjacent 1980; McCafire.v,1982] and dips eastunder Halmahera and west regions[Hall et al., 1995b]. The resultsof the modellingprovide under the Sangibe Arc. Regional seismicity[McCafire.v, 1982] a basis for estimating how much of the clockwise rotation suggeststhat the east dipping slab extends200-300 km beneath observedat the eastern plate margin is due to local tectonic Halmahera.The west dipping slab can be identifiedto 600 km effects. beneaththe [Cardwell et al., 1980]. PresentTectonic Setting of Eastern Indonesia The Philippine Sea Plate is currentlyrotating clockwise with respectto Eurasiaabout an Euler pole closeto its northernedge, Geologicalknowledge of easternIndonesia has increased basedon geological,geophysical, and seismologicalobservations considerablysince Hamilton's [1979] review of Indonesian along the plate margins [Ranken et al., 1984; Huchon, 1986; tectonics.Interpretations of present-daytectonics are basedon Send et al., 1987; Send et al., 1993]. The rate of convergence marine[e.g., Silver and Moore, 1978; Moore and Silver, 1983; increasessouthward, resulting in westward subductionat the Silver et al., 1983] and regionalseismic [Cardwell et al., 1980; PhilippineTrench, which terminatesjust north of Halmaheraat McCaffre.v, 1982] investigations.The IndonesianGeological 2ø50'N. The formed later than 5 Ma and is Research and Development Centre (GRDC) has carried out associatedwith less than 150 km of subductedlithosphere mappingwith internationalcollaboration, and knowledgeof the [Cardwell et al., 1980; Karig, 1975] implying that before the geologicaldevelopment of easternIndonesia is based on this Pliocene, parts of the east formed part of the work [e.g., Dow and Sukamto,1984; Pigram and Davies, 1987; Philippine Sea Plate. Hamilton [1979] joined the Philippine Hall, 1987; Hall and Nichols, 1990] andour unpublishedresults. Trenchto the New GuineaTrench with a strike-slipfault eastof EasternIndonesia includes the junctionbetween three major Halmaheraand placedHalmahera on a microplateseparate from plates(Figure 1). Relativeto Eurasia,treated as a singleplate, the the Philippine Sea Plate. However, recentstudies of seismicity AustralianPlate is moving NNE at about75 mm/yr [DeMets et and [Cardwell et al., 1980; McCaffre.v, 1982; al., 1990], and the Philippine Sea Plate movesWNW at about Moore and Silver, 1983; Nichols et al., 1990] indicate that the 100 mm/yr [Send et al., 1993]. The marginsof Eurasiain SE Halmahera-Waigeoislands form part of the PhilippineSea Plate Asia mustbe consideredas involvingnumerous small platesand [Hall and Nichols, 1990]. The southernboundary of both the plateboundaries that haveshifted and evolved on timescalesof a Molucca Sea and the Philippine Sea Plate is the Sorong Fault few million years. In the Philippines,active arcs are system(Figure 2), which is the westernend of a zone of left- separatedfrom the Eurasiancontinent by smallbasins floored by lateral strike-slipfaults extendingfrom northernNew Guinea . The Philippine arcs end in the south in the [Carey, 1958], resulting from the oblique convergence of

• ...... 132'E ...... 124'E/ • • ======Celebes Sea: e ...... ß...... i=,hilippin ...... ///////// Sea F'la ...... Lili ......

:Wayatoll...... •owonli ...... Molucca i i!iiiiii!i WAIGEO Sea Collision a^• :•iii '•"•"• ' ' ..... :::...... Complex I•A'rANTA ......

.... • ,,,a,,.•,•-.• • .' •..'...•.OLUC.CA•01•N• F^ULT' ' ...... • • • • • •

BANGGAI :

' ' ,' ': ::::::::::::::::::::::::::::::::::::::: ,• x x x x x x x x x x x x x ,• x ,• x x x x x x x x x x x x x x x x x x x x x x x x x x x ,• x x x

Figure2. Principaltectonic features of theSorong Fault Zone. Australian (indicated by crosses) is found only southof the SorongFault strandpassing through Bacan. North of this fault is crust of arc or ophioliticorigin whichis part of the PhilippineSea Plate.Solid trianglesindicate active volcanoes. Solid circles showthe locationof sitesfrom which palaeomagnetic samples discussed in thisstudy were collected. 1120 HALL ET AL.: CENOZOIC MOTION OF THE PHILIPPINE SEA PLATE

Australiaand the Pacific.The SorongFault hasits type locality [Hall et al., 1991]. The ophiolitesare remnantsof an early in the northernBird's Head region of [Visserand Mesozoicintraoceanic arc [Hall et al., 1988a, Ballantyne,1991a, Hermes, 1962; Dow and Sukamto, 1984]. We use the term b] and are overlain by, and imbricatedwith, upper Cretaceous- Sorong Fault Zone for an area including the region from the Eocene arc volcanic and sedimentaryrocks; arc plutonic rocks Bird's Head to east Sulawesi. intrude the ophiolites. These are interpretedas equivalentsof Paleogenearc rocksof northernNew Guinea [Dow and Sukamto, Geologyof the SorongFault Zone 1984; Pigram and Davies, 1987], the eastern Philippines [Rangin, 1991], and ridges of the north Philippine Sea This geologicalsummary is basedon fieldworkcarried out by [Tokuyama, 1985; Hall and Nichols, 1990]. A middle Eocene LondonUniversity and GRDC geologists[Hall et al., 1988a,b; unconformity marks plate reorganisation,perhaps linked to Nichols and Hall, 1991; Charlton et al., 1991; Hall et al., 1991; changein Pacific Plate motion at -42 Ma [Clague and Jarrard, Hall et al., 1995a, R. Hall et al., unpublishedresults, 1995]. The 1973]. Older rocks are overlain by shallowwater upper Eocene SorongFault systemjuxtaposes continental, arc, and ophiolitic limestonesand Oligocenebasaltic pillow lavasand volcaniclastic rocks(Figures 2 and 3). Australiancontinental crust is presenton turbidites. The lavas and volcaniclastic rocks are arc-related the islands of Obi and Bacan within the Sorong Fault system productswhich also have equivalentsin New Guinea [Dow and [Hamilton, 1979; Ali and Hall, 1995]. High-grademetamorphic Sukamto,1984; Pigram and Davies, 1987] and the Philippines rocksof probablePalaeozoic or greaterage are exposedon Bacan [Rangin, 1991]. Arc activityceased by the early Miocene. and Obi, and lower and middle sedimentary rocks, Fragmentsof continental,arc, and ophiolitic origin have a formingthe cover to the metamorphicbasement, are presentin common history from the early Miocene, and there was no southern Obi. significantNeogene volcanic activity along the SorongFault. The regionnorth of, andmuch of the areabetween the strands Collision betweena Philippine Sea Plate boundingarc and the of, the SorongFault have a basementof ophioliticand arc rocks Australian continental margin occurred at -25 Ma and led to

Qua•rnary Present Halmahe• Arc Volcanism -1.6 Kayasa Formation u MOVEMENT ON SORONG Pliocene - FAULT STRANDS v:v'v'v'v'v'v:v:4 L (E Halmahem)5aolat Formation _ -5.2 (WHalmahem) We•la Group NeogeneHalmahe• Arc Volcanism v :v.m v:4 u Volcanic Member v.v.v.v.v.v.v.v• MOLUCCA SEA SUBDUCTION BEGINS Miocene M .V.V .V•

L SORONG FAULT INITIATED iil111 L!m•?•,ng?ii1 I II I I I • I • • • -23.3 -PHILIPPINE SEA PLATE I I I I I I I I I U (Waigeo)F, umai Formation ARC COLLISION Olioocene L (Halmahem)Tawall Formation PhilippineSea Plate -35.4 JojokMember Arc Volcanism U

PACIFIC PLATE REORGANISATION Eocene M Sagea Formation Arc Volcanism L WayamliFormation -56.5 Paleocene -65.0 (NE Halmahera)Gau Lst Formation v v v v v v v,.l•:l:;:l• Arc Volcanism (fie Halmahem)Gowonli Formation v v v v v v v v v Cretaceous Arc Volcanism

Sup•-subductionZone Ju•ssic IgneousActivity : !•lack 5halee -- -- Micac•ou• 5•t• --

Triassic F,ock•• • Palaeozoic Age Ma

?HILIF?INE PLATE AU,STL!AN ORIGIN ORI,IN

Figure3. Simprifledstratigraphy of PhilippineSea Plate rocks north of theSorong Fault and within fault strands andAustralian-origin rocks now within the fault system. The ages of principalevents in thegeological history of the regionare basedon the timescaleof Harland et al. [ 1990]. HALL ET AL.: CENOZOIC MOTION OF THE PHILIPPINE SEA PLATE 1121

Table la. Agesand Locations of NeogeneSites

StratigraphicUnit Age Ma DatingMethod Latitude,Longitude

Kahatola, NW Halmahera KayasaFormation Pliocene 2.2 _.+0.7 K-At 1ø42.49'N, 127ø31.25'E TanjungLili, NE Halmahera Saolat Formation Mio-Pliocene 5 _+1 Forams 1ø15.55'N, 128ø43.62'E Kahatola, NW Halmahera Weda Group UpperMiocene 7 + 2 K-At 1ø41.08'N, 127ø31.90'E

creationof the left-lateralSorong Fault Zone [Ali and Hall, convertedto numericalages using the schemeof HaHand et al. 1995]. Northwardmovement of Australiaduring the Neogene [ 1990]. At eachsite, 25-mm coreswere obtained using a gasoline occurredwithout subduction,and the southernplate boundary poweredrock drill andoriented to _+2 ø usinga magneticcompass hasbeen a strike-slipplate boundary zone. Neogene convergence inclinometer. The magnetic stability of each specimen was between Eurasia and the Philippine Sea Plate resulted in assessedusing either stepwisealternating field (AF) demagneti- subductionof the Molucca Sea producingthe HalmaheraArc. sationor continuousthermal demagnetisation[Dunn and Fuller, Volcanicactivity continues at presentin the northernpart of the 1984]. Demagnetisationdata were analysed using equal area and . orthogonal[Zijderveld, 1967] plots from which characteristic The Halmahera-Waigeoislands today form part of the componentsof remanencefor each specimenwere determined. PhilippineSea Plate.They, togetherwith mostof the islands Site meandirections were computedusing the statisticsof Fisher withinthe fault system,include Tertiary rocks which have a long [1953]. Site dataare presentedin Tables1 to 3. volcanicarc history,and the oldestrocks known are ophiolites which also formed in an arc-related setting. The significant PalaeomagneticResults similaritiesto Tertiary rockselsewhere on the PhilippineSea Plate [Hall and Nichols, 1990], and to areassuch as the east The magnetisationsof the 254 sitessampled during this study Philippines,support the suggestionthat all theseareas formed fall into four groups: partof thePhilippine Sea Plate throughout most of theCenozoic. 1. The first is Mesozoic and Tertiary rocks of Australian We here test this hypothesis,which is basedon interpretationof origin.These rocks are typicallyweakly magnetised. Only a few present-day tectonics and geological evidence, using sites yielded directionalinformation and resultsare discussed palaeomagneticmethods. elsewhere[Ali and Hall, 1995]. 2. The secondgroup is Neogenelimestones. Despite attempts at measurementof remanenceusing 2G cryogenicmagnetometers PalaeomagneticMethods at Southamptonand University of California, Santa Barbara Palaeomagneticresults should provide a key test of tectonic (UCSB) and a CryogenicConsultants Limited magnetometerat modelsfor the region. With our new stratigraphicbase we have Southampton,the rocks were too weakly magneticto obtain been able to systematicallysample rocks of virtually all ages reliable data. betweenJurassic and Neogene. Palaeomagneticinvestigations 3. The third group is Neogene volcanogenicsedimentary were supportedby a dating programme,and the ages of the rocks.These rocks are stronglymagnetic and formedover 25% of formationsreported here, with one exception,were determined the samplescollected. However, they carry a large viscous during this study using K-Ar isotopic, nannoplankton,and remanence. After initial failure to obtain reliable data from these foraminiferalstudies. Biostratigraphicdeterminations, which in rocksin the laboratoryat Southampton,we attemptedto measure manycases allow resolutionto a single fossil zone, have been specimensfrom all of thesesites in the shieldedlaboratory at

Table lb. PalaeomagneticResults from Neogene Sites

In Situ Corrected Correct•½l Site RockType N A/T Inc Dec tz9s K Bedding Inc Dec RegionalTilt Inc Dec Polarity KayasaFormation HC6 andesite 7 6,1 42.7 191.5 1.9 1063.8 088/36 S 7.2 187.9 090/16 S 26.9 189.8 R HC7 andesite 6 4,2 42.9 175.6 7.4 83.6 322/25 NE 51.7 149.5 090/16 S 27.0 176.7 R Sa01atFormation HC22 calcisiltite 5 5,0 -4.2 5.5 4.2 332.6 180/22 W -2.5 7.1 N HC24 calcisiltite 9 0,9 -10.2 9.1 7.3 51.2 146/16 SW 0.5 10.1 N Weda Group HC 1 andesite 8 6,2 54.1 195.7 9.0 238.7 330/05 NE 52.4 191.5 090/16 S 38.4 192.2 R HC2 andesite 9 7,2 51.4 195.3 3.2 256.2 330/05 NE 55.1 192.0 090/16 S 35.7 192.2 R

AbbreviationsareN, numberof samplesat site;Inc, inclination in degrees;Dec, declination in degrees; %5, circle of 95% confidence aboutthe site mean; In Situ, sitemean direction before application of tectoniccorrection; Corrected, site mean direction after tectonic correction;A/T, demagnetisationmethod: A, alternatingfield; T, thermal;K, precisionparameter; Polarity: N, normal,R, reversed. 1 122 HALL ET AL.: CENOZOIC MOTION OF THE PHILIPPINE SEA PLATE

Table 2a. Ages andLocations of Oligoceneto Middle EoceneSites

Formation Age Ma DatingMethod Latitude,Longitude Doi, NW Halmahera Tawali Oligocene 27 _.+2 K-Ar 2ø15.50'N, 128ø46.00'E AkelamoRiver, NE Halmahera Tawali Oligocene 29 _+6 Nannofossils 1ø24.50'N, 128ø38.60'E Loleojaya,Kasiruta Tawali Lower Oligocene 32.3 _.+3 Forams,Nannofossils 0ø21.40'S, 127ø15.00'E Dikoh, Kasiruta Tawali Lower Oligocene 32.3 _.+3 Nannofossils 0ø23.60'S, 127ø6.40'E Jojok, Kasiruta Tawali Lower Oligocene 33.3 _.+0.3 Forams,Nannofossils 0 ø19.20'S, 127ø6.50'E TanjungMomfafa, Waigeo Rumai Oligocene 29 _+6 Literature 1ø24.50'N, 128ø38.60'E SageaRiver, SE Halmahera Sagea MiddleEocene 40 _.+2 Metamorphicage 0ø32.94'N, 128ø2.64'E

UCSB. However, when thermaldemagnetisation was carried out, of the low temperaturecomponent identified in this specimen erraticdemagnetisation trends were accompaniedby magneto- wasnot recordedin the other specimensfrom this site. mineralogicaltransformations at temperaturesabove 300øC. 4. The fourth group is rocks, mainly volcanic, from which reliable data were obtained. Approximately25% of the total Neogene samplesyielded reliable data and include the resultsreported below. At all sitesthe magnetisationis simple(Figure 4). In all NW Halmahera, Kahatola, Kayasa Formation but one of the examplesthe remanencecomprises a single The Kayasa Formation comprisesa sequenceof Plio- principal componentusually with a weak present-dayviscous Pleistoceneandesitic lavas and interbeddedpyroclastic rocks remanentmagnetization (VRM). The exceptionis the specimen restingunconformably on sandstonesof the OligoceneTawali from the Rumai Formation (Figure 4) in which the stable Formation.Flow-banding surfaces' parallel to flattenedpumice component was isolated between-350 and 580øC; the direction fragmentshave dips between25 ø and 35ø. Two sitesdrilled in a

Table 2b. PalaeomagneticResults from Oligoceneto Middle EoceneSites

In $itu Corrected Site Rock Type N A/T Inc Dec 595 K Bedding Inc Dec Polarity Tawali Formation, Doi, NW Halmahera HJ40 pillow 5 3,2 19.7 218.9 4.8 253.3 260/12 N 27.5 222.3 R Tawali Formation, Akelamo River, NE Halmahera HB61 siltstone 9 1,8 -1.7 46.7 4.1 159.6 340/35 E -33.1 41.6 Tawali Formation.Loleojaya. Kasiruta KJ2 basalt 8 8,0 -0.8 241.7 3.3 283.6 330/12 NE 11.2 241.7 KJ5 basalt 6 6,0 15.5 217.3 8.1 69.0 330/12 NE 26.5 215.5 KJ6 basalt 6 6,0 17.6 226.9 11.7 33.6 330/12 NE 29.3 225.7 KJ8 basalt 8 8,0 16.8 214.4 13.0 19.1 330/12 NE 27.5 212.2 KJ21 basalt 5 5,0 18.7 236.7 12.6 38.3 330/12 NE 30.7 236.4 KJ23 basalt 7 7,0 15.7 201.9 6.6 97.4 330/12 NE 25.0 199.1 KJ24 basalt 6 6,0 4.4 203.7 14.0 24.2 330/12 NE 14.0 202.5 Location Mean 7 13.0 220.3 12.8 23.2 24.2 218.9 Tawali Formation, Dikoh, Kasiruta KJ26 basalt 6 6,0 11.2 220.4 5.3 162.3 295/08 N 18.9 221.1 Tawali Formation.Jojok. Kasiruta KJ 10 siltstone 4 0,4 16.0 218.3 6.1 236.8 315/08 NE 23.9 217.9 R KJ 11 basalt 5 0,5 19.3 229.9 5.0 88.0 314/14 NE 33.1 231.2 R Rumai Formation

WC7 silt/sandstone 8 0,8 30.0 258.0 13.9 16.9 073/58 S 18.1 229.9 R WC8 silt/sandstone 8 0,8 -32.2 219.2 7.2 59.6 274/50 N 11.2 214.2 R Saeea Formation

HC37 siltstone 6 2,4 -36.7 55.4 19.6 12.7 172/40 W -0.3 61.4 N HC38 siltstone 10 0,10 - 16.4 23.6 7.6 47.0 146/36 W 13.8 24.0 N HC39 siltstone 7 0,7 - 10.6 38.9 8.5 50.9 172/26 W 8.2 39.3 N HC40 siltstone 12 0,12 -31.1 41.4 13.1 12.0 168/40 W -8.2 52.2 N Location Mean 4 -24.1 39.1 19.5 23.1 -2.1 41.6

Abbreviationsareas in Tablelb. Forthe Sagea Formation the location mean corrected values have %5, 29.8; K, 10.5. HALL ET AL.: CENOZOIC MOTION OF THE PHILIPPINE SEA PLATE 1 123

Table3a. Ages and Locations of LowerEocene and Cretaceous Sites

Formation Age Ma DatingMethod Latitude,Longitude

Wayamli, SE Halmahera Wayamli LowerEocene 47 +_3 Forams 1ø15.55'N, 128ø43.62'E Gowonli River, SE Halmahera Gowonli UpperCretaceous 74.5 +_4 Forams 0ø26.26'N 128ø38.10'E Tutuli River, NE Halmahera GauLimestone UpperCretaceous 78.5_+ 4.5 Forams,Nannofossils 1 o 12.27'N 128 ø 15.92'N sequenceof glassyandesites on the islandof Kahatola,NW I =-2.5 ø, and D =10.1ø, I = 0.5ø (Figure5), reducingthe angular Halmahera(Figure 2) havein situmean directions of D = 191.5ø, separationof the two vectors. These data suggesta small I=42.7 ø, {x95=1.9 ø, K=1064, and D=175.6 ø, I=42.9 ø, clockwiserotation of the sites, althoughthe declinationoffset is {x95= 7.4ø, K= 84 (Figure5). The in situ directionsbecome within limits generallyaccepted for secularvariation effects. widelyseparated after tilt correctionbased on the flow banding (Figure5), suggestingthat the dip is primary.The in situ NW Halmahera, Kahatola, Weda Group Volcanic Member directionsbecome less steep (Figure 5) if an alternativetilt Two sitesin upperMiocene andesite flows of the Weda Group correctionbased on dip of Oligocenesandstones beneath the Volcanic Member on southKahatola (Table 1a, Figure 2) yielded unconformityis used.This correction assumes that the Oligocene in situ directions of D = 195.7ø, I = 54.1 o, o•95= 9.0 ø, K = 239, sandstoneswere close to horizontalat the time of lava eruption and D= 195.3ø, I= 51.4ø, o•95= 3.2ø, K= 256 (Figure 5). As and that the lavas had a primary dip of about 20ø . This with the KayasaFormation (see above)a regionaltilt correction interpretationis consistent with the low dipsof Oligocenerocks wasapplied, which reduced the inclinationfor both sitesto -37 ø in NW Halmaheraand with evidencethat large-scaleblock tilting (Table 1b), suggestingthat the lavashad a primarydip, but which postdates2 Ma. The correctionyields site directionsof is steepfor upperMiocene rocks today located 1.7øN. It is likely D = 189.8ø, I = 26.9ø and D = 176.7ø, I = 27.0ø suggestinglittle that the effectsof secularvariation have not been fully averaged rotationof the site duringthe Pleistocene. in these sites and we do not use these data in tectonic interpretations. NE Halmahera, Tanjung Lili, SaolatFormation Five sites were drilled in shallow marine limestones and marls Oligocene-Middle Eocene of the upper Miocene-lowerPliocene Saolat Formationat Tanjung Lili, NE Halmahera (Table la, Figure 2). NW Halmahera, Doi, Tawali Formation Demagnetisationrevealed a singlesignificant component of remanence(Figure 4); thermal demagnetisationindicated Pillow basaltsoverlain by sandstoneswith a low dip, bothpart unblockingtemperatures in the range 530-560øC, suggesting that of the Tawali Formation, were drilled at one site on the island of magnetiteis the principalremanence carrier. Two siteshave in Doi, NW Halmahera(Table 2a). The pillow basaltshave upper situ directions of D= 5.5ø, I=-4.2 ø, o•95=4.2 ø, K= 333, and Oligocene K-Ar ages, and the sequence is unconformably D=9.1ø, I =-10.2 ø, o•0•=7.3ø , K=51 (Table lb). The overlain by lowermost Miocene rocks. The single component applicationof tilt correctionsproduces directions of D = 7.1ø, remanencehas magnetiteunblocking temperatures.The in situ

Table 3b. PalaeomagneticResults from LowerEocene and Cretaceous Sites

In Situ Corrected Site RockType N AFF Inc Dec c•95 K Bedding Inc Dec Polarity Wayamli Formation HC25 calcarenite 7 0,7 41.5 265.6 7.6 64.2 175/45 W -16.1 265.0 R HC26 calcarenite 9 0,9 38.3 261.1 8.7 36.3 175/50 W -19.2 262.7 R HC27 calcarenite 10 0,10 41.2 279.5 4.3 130.1 170/62 W - 16.5 277.9 R HB90 calcarenite 8 0,8 12.1 80.9 3.4 264.6 347/34 E -21.9 81.2 N HB93 calcarenite 6 0,6 -32.2 262.8 4.8 192.9 324/29 E -6.7 258.9 R Location Mean 5 16.4 265.6 36.4 5.4 -3.7 264.3 G0wonli Formation HC32 mudstone 9 8,1 -29.1 352.7 5.0 106.8 142/15 SW -20.7 358.4 N HC33 mudstone 8 4,4 -33.5 352.4 5.4 104.4 123/23 SW -15.5 359.0 N HB 101 mudstone 7 0,7 -28.3 354.8 9.0 45.8 133/31 W -12.9 1.4 N HB102 mudstone 7 0,7 -41.1 338.4 9.6 40.5 125/42 SW -12.8 355.4 N Location Mean 4 -33.1 348.4 9.1 103.2 -15.5 358.6 Tutuli Formation HC32 mudstone 9 8,1 -29.1 352.7 5.0 106.8 142/15 SW -20.7 358.4 N

Abbreviationsare as in Tablelb. Forthe Wyamli Formation the location mean corrected values have c•05, 15.9; K, 24.0.For the Gowonli Formationthe location mean corrected values have %•, 5.0;K, 338.3. 1124 HALL ET AL.: CENOZOIC MOTION OF THE PHILIPPINE SEA PLATE

(A) (B) (C) -4 w up NRM: 8.875 X I0 -3 W UP wup NRM: 2.375 X I0

-5 NRM: 4.243 X I0

90 1O0 •-•e

30• ß 30 10.•0

E DN )N E DN

(D) (E) (F) WUP wup wup

! / / -5 NRM: 3.973 x 10

NRM: ,,• / NRM:6.507 X 10-5 293 .• ß i oo.:-•. 350 / "t 0 ./•.•-• ..... S --456- •a;•e'ee '"e N

E DN E DN E DN

(G) (H) (I) w up W UP W UP

-5 -4 NRM::2.158 X 10-7 NRM: !.159 X I0 NRM: 1.113 X 10

225 20 412. e.,,..,e_ ..... e-o 526 23

531 $L/38'9• : : '"" -' N 213

2OO

600

E DN E DN E DN

Figure4. Representativedemagnetisation vector end pointplots (tilt-corrected). The solidcircles represent the remanencevector on the horizontalplane; open circles represent the vectoron a verticalplane. On theseplots, valuesin the range0-80/100 representapplied alternating fields (milliTeslas), whereas values in the range20-650 representthe thermal treatment(degrees celsius). Natural remanentmagnetization (NRM) values are in milliamperesper meter.The plotscorrespond to the followingformations: (a) KayasaFormation, Kahatola, (b) SaolatFormation, Tanjung Lili, (c) Weda Group VolcanicMember, Kahatola, (d) Tawall Formation,Doi, (e) RumaiFormation, Tanjung Momfafa, (f) SageaFormation, Sagea River, (g) WayamliFormation, Wayamli, (h) Gowonli Formation, Gowonli River, (i) Gau LimestoneFormation, Tutuli River. HALL ET AL.: CENOZOIC MOTION OF THE PHILIPPINE SEA PLATE 1125

N N N Jojok, Kasiruta On Jojok, just west of Kasiruta, is a sequenceof lavas and sedimentaryrocks about 30m in thickness which dips 8ø northeast.Two of five sites (Table 2a) collected yielded tilt correcteddirections of D = 217.9ø, I = 23.9ø, (x9s= 6.0 ø, K = 237 and D =231.2 ø, I = 33.1ø, (•s = 5.0ø, K = 88 (Table 2b, Figure 6). u? PLIOCENE'•• Dikoh, Kasiruta N N N Near Dikoh there is a 5-m sequenceof pillow lavas overlain by thinly bedded subhorizontalsandstones of lower Oligocene age(Table 2a). The lavashave an in situ directionof D = 220.4ø, I= 11.2ø, c•95= 5.3ø, K= 162, and a tilt-correcteddirection of D = 221. l o, I = 18.9ø (Table 2b, Figure6). The directionsfrom all localitieson Kasirutaare interpretedas

u reversepolarity magnetisations, clockwise deflected through 35- 40 ø with formation latitudes of 10-15øS. The similarity in PLION..•,.••//CE MIOCEN E• directionsfrom three separate areas indicates that the islandhas behavedas a singleblock since the Oligocene. Figure 5. Equal-area stereographicprojection of the mean directionsfor Neogenesites. Solid circlesrepresent downward Waigeo,Tanjung Momfafa, Rumai Formation dipping vectors, and the open circle representsan upward The volcaniclasticRumai Formation was drilled at Tanjung dipping vector. Site vectorsare shown with 95% confidence circles. These sites record small clockwise declinations consistent Momfafa,east Waigeo (Table 2a). Mediumbedded turbiditic sandstonesand siltstoneswith subordinateconglomerates are with the present-day motion of the Philippine Sea Plate. exposedin closeto tightsymmetrical folds with verticalaxial Inclinationsof volcanicrocks suggest that the effectsof secular planesand negligible plunges. Two siteson oppositelimbs of a variationhave not beenfully removed. smallsyncline have in situdirections of D = 258.0ø, I = 30.0ø, (z95=13.9 ø , K=17, and D=219.2 ø, I=-32.2 ø, (•95=7.2ø, K = 60 (Table 2b, Figure7). Applicationof the tilt corrections mean direction is D=218.9 ø, I=19.7 ø, N=5, (x95=4.8, producesdirections of D = 229.9ø I = 18.1o, and D = 214.2ø, K= 253, and the tilt-correctedmean direction is D= 222.3ø, i = 11.2ø (Table2b), significantlyreducing the angular separation I = 27.5ø (Table2b, Figure6). This is interpretedas a reverse of the two vectors(Figure 7) suggestingthat the remanence polaritymagnetisation indicating about 40 ø of clockwiserotation predatesfolding. These data suggest over 40 ø clockwiserotation and a formation latitude of ~ 15øS. and a formation latitude of about 8øS.

NE Halmahera,Akelamo River, Tawali Formation N N Oligocenevolcaniclastic siltstones of the Tawali Formation UP'FEEr,,,///••• OLIG weresampled in theAkelamo River, NE Halmahera(Table 2a). / .. DOI Onesite with a stablemagnetisation yielded an in situdirection of D = 46.7ø, I = - 1.7ø, N = 9, (x95= 4.1o, K = 160 and a tilt- TAWALl / OLIGOCENEFOF..MATION•\ . FOF,M^TION correcteddirection of D = 41.6ø, I=-33.1 ø (Table2b, Figure6). •. ^KELAMOI•IVEI• Thisis interpretedas a normalpolarity magnetisation indicating NEHAL• over40 ø of clockwiserotation and a formationlatitude of ~ 18øS. N N N

Kasiruta, Tawali Formation OLI•OCENE • OLI•OCENE The Tawali Formationis well exposedon the islandof Kasiruta(Figure 2) andconsists of basalticpillow lavas with interbeddedsedimentary rocks which provide excellent structural control.Specimens from all sitesyield essentiallysingle componentremanences with magnetite unblocking temperatures (530-560øC). Figure6. Equal-areastereographic projection of the tilt- corrected directions for Oligocene sites from the Tawall Loleojaya,Kasiruta Formation at Pulau Doi, NW Halmahera; Akelamo River, NE A 5-m sequenceof siltstonesdipping northeast at ~12ø Halmahera; Loleojaya, Kasiruta; Jojok, Kasiruta; Dikoh, overlainby 70m of basalticpillow lavasis exposednear Kasiruta.Solid circlesrepresent downward dipping vectors, and Loleojaya,east Kasiruta (Table 2a). Sevensites in the lavas theopen circle represents an upwarddipping vector. Site vectors yieldedan in situ directionof D= 220.3ø, I= 13.0ø, N= 7, are shown with 95% confidence circles. The sites include those (x95= 12.8ø, K = 23 anda tilt-correcteddirection of D = 218.9ø, with normal and reversed polarities, but all show similar I = 24.2ø (Table 2b, Figure6). clockwise deflected declinations of about 40 ø. 1126 HALL ET AL.: CENOZOIC MOTION OF THE PHILIPPINE SEA PLATE

SE Halmahera, Sagea River, SageaFormation N N Middle Eocene volcaniclasticrocks of the Sagea Formation were drilled in the SageaRiver, SE Halmahera(Table 2a). The LOWER rocks have a weak cleavage and were metamorphosedto subgreenschistfacies (T < 400øC) in the middle Eocene. Four sitesyielded an in situ component,defined between100øC and 450-500øC, with a direction of D= 39.1 ø, I= -24.1 ø, (x95= 19.5ø, K= 23 (Table 2b, Figure 7). Application of individual site tilt corrections increases the dispersion of the sites (D= 41.6ø, I = -2.10, (x95= 29.8ø, K = 11) suggestingthat the magnetisation '•, NEH^LIVI^HEIKA '•, N•_AL._M^HE_IKA postdatestilting. The formation is unconformablyoverlain by Miocene limestones with no appreciable dip suggestingno regional tilt correctionis required.The in situ directionis clearly not a recent low-temperaturecomponent. We interpret it as a N N middle Eocene normal polarity magnetisationassociated with metamorphismindicating-40 ø of clockwise rotation and a formation latitude of about 13øS. U?? CRETACEOUS-I-j \ N N I GOWONLiFORMA•T•3NI GOWONLiF•ORMATiON OLin/OC •. ,SEH^LM^HEF.• •, ,SE_H__ALM_AHEF.•--- F,.U W^I•EO / ,.•W•E,.•.O,., \ N N

UFF CIRE

'• NE"ALMAHERA '• CORRECTED /

.•,;;•,:.•... Figure 8. Equal-areastereographic projection of the site mean + - directions(before and after tilt correction)for the lower Eocene Wayamli Formation,NE Halmahera,the CretaceousGowonli 5AGEA FOF,M^TION 5AGEA FOF,M^TION Formation, SE Halmahera, and Gau Limestone Formation, NE •, 5E HALMAHEF-,A / 5E HALMAHEF-,A/ Halmahera. Solid circles representdownward dipping vectors, and open circles representupward dipping vectors.The site vectorsare shownwith 95% confidencecircles. For the Wayamli Formation, in situ and tilt-correcteddirections for site HC90 have been inverted, and the declinationsare interpretedto indicate Figure 7. Equal-areastereographic projections of the site mean -90 ø rotationsince the early Eocene.Inclinations of Cretaceous directions (before and after tilt correction) for the Oligocene rocksindicate subequatorial formation latitudes and suggest180 ø Rumai Formation, Waigeo, and the middle Eocene Sagea rotation since the late Cretaceous. Formation, SE Halmahera. Solid circles representdownward dipping vectors, and open circles representupward dipping Lower Eocene vectors. Site vectors are shown with 95% confidence circles. All these sites record similar declinations to those of the Tawali The lower-middle Eocene Wayamli Formation consistsof Formation despite being distributedthroughout a large area, redepositedslope limestonesand interbeddedred and green indicating a consistentpost-Oligocene rotation for the entire carbonatemudstones. At Wayamli the sequencedips at 30-60ø to region. both east and west. The strike of the beds remains constant over a HALL ET AL.: CENOZOIC MOTION OF THE PHILIPPINE SEA PLATE 1127

distance of a few kilometres indicating folds with negligible NE Halmahera, Tutuli River, Gau Limestone Formation plungealthough fold axesare not exposed. Folding is probablyof The upper CretaceousGau LimestoneFormation was sampled middle Eoceneage. Five sites(Table 3a) have an in situ mean in the Tutuli River, NE Halmahera (Table 3a) where there is a direction of D = 265.6 ø, I = 16.4ø, et95= 36.4 ø, K = 5 (Table 3b). continuous 60-m sequenceof thinly bedded limestones and In Figure8 the in situand tilt-corrected direction for siteHC90 is volcanogenic turbidires dipping NW at 50-70ø. Deformation inverted. After tilt correction,D = 264.3 ø, I =-3.7 ø, et95= 15.9ø, occurred during the middle Eocene. One site has an in situ K = 24. The localityincludes both normaland reversedpolarity direction of D= 9.3 ø, I= 23.9 ø, •z9s= 7.7 ø, K = 63 (Table 3b, sites, which are excellent evidence to suggest that the Figure 8 with a tilt-corrected direction of D= 1.4ø, I=-4.0 ø. magnetisationis primary.The tilt-correcteddata passboth the Demagnetisationisolated a single component(Figure 4) with McElhinny [1964] and the McFadden [1990] testsat the 95% unblockingtemperatures of 550-580øC,suggesting that magnetite confidence level indicating a prefolding magnetisation.We is the remanencecarder in these rocks. Assuming a primary interpretthis to indicate-85 ø of clockwiserotation with a remanence,the direction indicateseither a normal polarity with formation latitude of-2øN. negligible rotation or a reversepolarity with -180 ø rotation.The site must have been closeto the equatorat formation. Cretaceous Declination and Inclination Record

SE Halmahera, Gowonli River, Gowonli Formation In the Sorong Fault Zone, palaeomagneticdata collected Limestones and interbedded volcaniclastic mudstones of the during this project from rocks underlain by crust of arc or SenonianGowonli Formation were drilled in the Gowonli River, ophiolitic origin (Figure 2) define two principal areas(Figure 9) SE Halmahera(Table 3a). The sequencetypically dips to the SW with different tectonichistories: (1) the area north of the Sorong at about 15-30ø althoughlocally it is deformedby mesoscopic Fault and (2) the area within strandsof the SorongFault. openfolds. Folding is middleEocene in age. Four sitesyielded an in situ directionof D= 348.4ø, 1 = -33.1 ø, et95= 9.1ø, K= 103 Rotation North of the Sorong Fault (Table 3b, Figure8). Applicationof sitetilt correctionsproduces In the area north of the SorongFault we interpretdeclination a mean direction of D= 358.6ø, I=-15.5 ø, et9s=-5.0 ø, K= 338 shiftsto indicate long-term,but discontinuous,clockwise rotation (Figure8). This suggeststhat the remanencepredates the middle (Figure 10). Eocene folding. This directioncan be interpretedeither as a Neogene. The sites in upper Neogene volcanic and normal polarity remanenceindicating negligible net rotationand sedimentaryrocks record small clockwisedeclination deflections a formationlatitude of-8øS or as a reversepolarity remanence consistent with those expected from angular velocities and acquiredat -8øN with 180ø net rotation. rotation poles calculated for the present-day motion of the

Dr.Glina•on Change an&Lat;iOudinalApparent •hifO Northward

15• Oli•ocene- MiddleEocene

filEUTA

I 124'E •

Figure 9. Summaryof palaeomagneticdata obtainedfrom the SorongFault Zone. The observeddeclination deflection(with confidence limit) in degreesfor eachlocality is shownby theorientation of the solidarc segments. The apparentlatitude change in degreescalculated from the inclinationis indicatedby the numberon eacharc segment. 1128 HALL ET AL.: CENOZOIC MOTION OF THE PHILIPPINE SEA PLATE

40 ß ß More than 3 sites in locality mean 30 ß ß 3 or fewer sites in locality mean

20

::3 10 1 I , , I I T 10115 20 25 35 45 50 55 60 65 70 75 80 85 90 -10 /•e Ma lavae 'r -2O --• -0- -30

lOO

8o

60 ß [] Nort, h of ._o 40 ,5orong Fault, Zone .• 20 • o ß o Wit, hint, he 5 10 15 20 25 30 35 40 45 50 55 60 5orong Fault, Zone -20 Ma -40

-60

-80

Figure 10. Palaeolatitudeand declination versus age for sitesnorth of, andwithin, the Sorong Fault Zone.

Philippine Sea Plate [Rankenet al., 1984; Huchon, 1986; Senoet indicatethat (1) the region has behavedas a single rigid block al., 1987; Seno et al., 1993]. Becausethe rocksare all <7 Ma, since-40 Ma, (2) no significantrotation took place between-25 only the oldest volcanic rock sites show clear differencesfrom and -40 Ma, (3) approximately 40ø of clockwise rotation the present field direction. However, their inclination values occurred after -25 Ma, and (4) this rotation occurred at an suggestthat the effectsof secularvariation have not beenfully approximatelyconstant rate. removed.Despite drilling morethan 50 sitesin lower and middle Lower Eocene. The five sitesdrilled in steeplydipping lower Miocene rocks, we have obtained no interpretable results. Eocene limestones of the Wayamli Formation result have Shallow marine lower-middleMiocene limestoneswere very particular significance.The locality includes sites with normal weakly magnetic or showed unstable magnetic behaviour. and reversepolarities, and the resultspass a fold test. We assume Volcaniclastic rocks of this age carried a massive viscous rigid block behavioursince the beginningof the Eocene.The remanence,and althoughsome samples yielded reliable data, site declinationsof-90 ø are thereforeinterpreted to indicate -90 ø statisticswere inadequate;however, in most cases individual total rotation with -45 ø of clockwise rotation between -50 and samplesindicated clockwise declination shifts. -40 Ma. O!igocene.middleEocene. Rocksof agesbetween -25 Ma Cretaceous. Primary magnetisationswere obtainedfrom two and -40 Ma and with primary remanencesshow clockwise upper Cretaceousformations. These show northward declinations declination deflections of-40 ø. They comprise a variety of and low inclinations (indicating subequatorial tbrmation lithologiesincluding basalticand andesiticlavas, volcanogenic latitudes). We tentatively interpret these results to indicate sediments,and impure limestones sampled over a very largearea. rotationsbetween the late Cretaceousand the early Eoceneof The Tawali Formation includes sites with normal and reversed -90 ø but do not includethem in our tectonicmodelling. polarities but with the same clockwise deflected declinations (Figure 6). Resultsfrom sitesat one locality (Rumai Formation, Waigeo) pass a fold test. The 10 Tawali Formation sites on RotationsWithin the SorongFault Zone Kasiruta,from three localities spread over an area of-200 km2, The island of Obi, situatedwithin the SorongFault Zone, indicate that post-Oligocenefaulting has not significantly includes Australian continentalrocks together with arc and affectedthe declinationdeflections. The SageaFormation, with a ophiolitic rocks. The arc and ophiolitic rocks and their remanencereset during middle Eocene metamorphism,shows sedimentarycover record latitudinal shifts similar to rocks of 40ø clockwise rotation. All these results are interpreted to comparableage and characternorth of the Sorong Fault Zone HALL ET AL.: CENOZOIC MOTION OF THE PHILIPPINE SEA PLATE 1129

..'. .•:.: RotationPole , oOo-o-,••.. ß ======...... ::,•::..,.... 6)-5 Poeelble Polee :::::::::::::::::::::::::::.•:.. ß

' k Rotationof ...... PHILIPPINE,SEA / PLATE Figure 11. Positionof rotationpoles (relativeto presentnorth) calculatedfor the PhilippineSea Plate from palaeomagneticdata [Hall et al., 1995b].The polefor the interval0-5 Ma is fromSeno et al. [1993]. The shaded areashows the path within which rotation poles must lie to satisfythe palaeomagnetic data alone.

(Figure11). However,different parts of the islandhave different low southerly latitudes (10-15øS). Shallow inclinations (and rotationhistories (Figures 9 and 10). Sitesin north Obi indicate northerly declinations)are recordedby the upper Cretaceous rapidNeogene counterclockwise rotations of ~60ø;the southern Gowonli and Gau LimestoneFormations. Interpretation of their part of the islandhas rotatedin the oppositesense at a similar formation latitude depends upon the rotation history inferred rate [Ali and Hall, 1995]. The similarityof the ophioliticrocks of between the late Cretaceous and early Eocene. Clockwise Obi to thoseof east Halmahera,and the distributionof Neogene rotationof 90ø duringthis intervalimplies a northernhemisphere volcanicrocks, suggest a maximumpossible displacement of Obi, origin and reversedpolarity for the Cretaceousmagnetisation. relative to Halmahera, of ~200 km westward since the middle Alternatively, a counterclockwiserotation of 90ø implies a Eocene.We interpretthe rotationsto representblock movements southern hemisphere origin and normal polarity. Low within the left-lateralSorong Fault systemwhich occurred during palaeolatitudesare consistentwith the lithologicalcharacter of the Neogene.However, althoughthe arc and ophioliticbasement the Cretaceoussequences [Hall et al., 1988a]. andtheir coverrocks may havebeen fragmented and dispersed by processeswithin the $orongFault Zone, it appearsthat any such dispersalhas been essentially latitude-parallel, i.e., parallelto the Implications for the Philippine Sea Plate orientationof the $orongFault system. The area north of the SorongFault is currentlypart of the Philippine Sea Plate. We interpret geological evidence as Latitudinal Movements indicatingthat the PhilippineSea Plate has been a coherententity All Cretaceous-Neogenerocks underlain by arc and ophiolitic since the early Tertiary. Our new inclination data supportthis basementnorth of, and within, the Sorong Fault Zone region idea for the Neogene.Our estimatesof latitudeshifts since the have low inclinationsand were formed at low latitudes(Figure Oligoceneare similar to thoseof earlier palaeomagneticstudies 11). Our data indicate northwardmovement during the Neogene. of other parts of the Philippine Sea Plate [e.g., Larson et al.. Threegroups of sitesdrilled in Neogenelavas on Halmaheraand 1975; Louden, 1977; Koyama et al., 1992]. Previousstudies of Obi [Ali and Hall, 1995] yield anomalouslysteep inclinations inclination data from the Philippine Sea Plate [Keating and suggestingthat the effects of secularvariation have not been Herrero, 1980; Kinoshita, 19'80; Bleil, 1981; Kodama et al., averagedout. However, all three localities record northward 1983; Hirooka et al., 1985; Haston and Fuller, 1991; Haston et movements. All other localities indicate 10-15 ø northward al., 1992; Koyama et al. 1992] have been used to argue for motion sincethe early Miocene. The inclinationdata from older continuousnorthward movement since the Eocene.Using similar formationssuggest little or no latitudinalmovement between the rates of northward motion to those demonstrated in earlier studies late Eoceneand the early Miocene(Figure 11), with formationat would imply palaeolatitudesof at least 25øS for lower Eocene 1130 HALL ET AL.: CENOZOIC MOTION OF THE PHILIPPINE SEA PLATE

50 Ma

lOON 150OE Rot, at, ion Pole 40-50 Ma

_

Figure 12. Reconstructionsof the Philippine Sea Plate at 5, 25, and 50 Ma using the rotation polescalculated from the resultsof this studyby Hall et al. [1995b]. The shadedarea showsthe approximatearea of the plate inferred to have been subducted.

SorongFault Zone sites,whereas the Wayamli Formationrecords The SorongFault Zone resultsare a significantaddition to the subequatorialformation latitudes. However, a critical pointis that data set for the Philippine Sea Plate. The clockwiserotation the simple history inferred from DSDP data lacks declination historyrecorded in the area northof the SorongFault in east control and there is an equatorial ambiguity for pre-Miocene Indonesiais entirely consistentwith tectonicand geological sites.Louden [ 1977] explicitly pointedthis out, and althoughhe argumentsthat this region is part of the plate. The area has inferred a northwardshift of the plate with decreasingage, he behaved as a single block since at least the late Eocene and recognisedthat each mean palaeolatitudecould have originated probablysince the early Eoceneand includesrocks predating on either sideof the equator.A knowledgeof the rotationhistory openingof the West PhilippineSea Central Basin. Thus it should of the plate is thereforevital in order to correctlycalculate and containthe mostcomplete palaeomagnetic record of Philippine interpretpalaeolatitudes. Sea Plate motionand is morelikely than otherregions so far Palaeomagneticstudies of the platehave indicated clockwise studiedto record the past motion of the plate withoutlocal declinationshifts [Larson et al., 1975;Koyama et al., 1992],but tectoniccomplications. Most importantly,it is unaffectedby the amount of rotation which can be attributed to motion of the openingof basinsat the easternplate margin. The areawithin entire plate has remained controversial. Studies of marine SorongFault system, close to the present-daysouthern bound;,:.y magneticanomaly skewness in the PhilippineSea Plate also of the Philippine Sea Plate, includes sites with local rotation suggestclockwise rotations. Louden [1976] and Shih [1980] historiesdifferent from thosenorth of the fault. We interpret reportedpole positionswhich imply about 60ø of clockwise thesedata [Ali and Hall, 1995]to reflectdeformation at thep.'_,.ate rotationfor the West PhilippineBasin since -44 Ma. Bowinet al. edge and suggestthat in easternIndonesia, unlike areas at the [1978] reporteda pole implyingabout 50 ø of counterclockwise presenteastern edge of the plate, the distinctionbetween local rotation, but the anomaliesanalysed were poorly defined and plate-widerotations is clear. [Jarrardand Sasajima,1980]. In contrast,anomaly skewness Our new data, indicating little or no latitudinal shift for datafrom the ShikokuBasin [Chamot-Rooke et al., 1989] have Sorong Fault Zone sites between 40 and 25 Ma, as well as beeninterpreted to indicateno significantrotation since ~20 Ma. southwardmotion between50 and 40 Ma, are incompatiblewith An early studyof seamountmagnetisation in the ShikokuBasin modelsfor the Philippine Sea Plate that predictonly northward [Vacquierand Uyeda,1967] hasalso been used to arguethat motion with little or no rotation of the plate [e.g., Uyeda and therehas been no significantrotation of thePhilippine Sea Plate Ben-Avraharn,1972; $eno and Maruyarna, 1984]. Neither can since17 Ma [e.g.,Seno and Maruyama, 1984]. The agesof the the palaeomagneticdata be accountedfor by rotationabout any are unknown, but two of the three seamounts studied of the Eurasia-PhilippineSea Plate (EU-PH) poles for the are now known to be situatedon young seafloorclose to the present-day[Ranken et al., 1984; Huchon, 1986; Seno et al., spreadingaxis [Chamot-Rooke et al., 1987]and so can only have 1987; Seno et al., 1993]. This is unsurprisingsince the present recordedthe historyof the last 15 m.y. Giventhe likely errors motion of the plate was establishedbetween 3 and 5 Ma and assumptionsin thesestudies the ShikokuBasin resultscould [Huchon,1986; Senoand Maruyama, 1984;Jolivet et al., 1989]. permitup to 20ø rotationsince ~20 Ma [Hall et al., 1995b]. However, by assumingthat the new data from the area north of Koyama et al. [1992] concluded that there has been clockwise the SorongFault systemrecord the Tertiarymotion history of the rotationduring the Neogene,but there were insufficientdata to oldestpart of the PhilippineSea Plate, it is possibleto calculate separatethe componentdue to rotationof the entireplate from rotationpoles for the period before 5 Ma [Hall et al., 1995a, b]. that due to local tectonic deformation: Rotation amountsand poles (with respectto magneticnorth) HALL ET AL.: CENOZOIC MOTION OF THE PHILIPPINE SEA PLATE 1131

which best fit the declinationand inclination data for the period Indonesiahas beenpart of a singleplate sinceas long ago as the before 5 Ma are 35 ø clockwise about 15øN, 160øE between 5 and Eocene,(2) that this plate was the Philippine Sea Plate, and (3) 25 Ma; no rotation between 25 and 40 Ma; and 50 ø clockwise that the whole plate has rotated clockwise in an episodic way about 10øN, 150øE between40 and 50 Ma (Figure 12). Using through approximately90 ø since the early Tertiary. Additional thesepoles it is possibleto reconstructthe plate aftercalculating data from rocksof early Neogeneand Eoceneage would improve the rotational effects of opening the marginal basins. this model, particularlyby defining more preciselythe intervals Palaeolatitudes calculated from inclination data and observed of rapid rotation, but such data will not be easy to acquire, declinationsfrom all Philippine Sea Plate sites can then be because almost the entire area of the plate is underwater. comparedto predictionsbased on thismodel [Hall et al., 1995b]. Potentialtargets are partsof the West PhilippineSea Basin and The new modelprovides a closefit to all palaeomagneticdata, far areaswhich until recently formed part of the plate such as the betterthan discrepanciesin modelsassuming no rotationof the easternPhilippines. whole plate [cf. Seno and Maruyama, 1984], and also satisfies constraintson rotationinferred from magneticanomaly skewness Acknowledgments. This work was supportedby NERC award and magnetisationstudies. Our model therefore GR3/7149 and grants from the Royal Society, the University of resemblesthat of Haston and Fuller [ 1991] in which declinations London Central ResearchFund, and the University of London SE from sitesin the easternplate marginwere consideredto reflect Asia Geological Research Group. We thank S. J. Baker, F. T. motionof the entireplate. Only a few sitesrecord rotations which Banner, T. R. Charlton, E. M. Finch, M.D. Fuller, E. A. Hailwood, are differentfrom thosepredicted by our model,and we attribute H. Y. Ling, J. A. F. Malaihollo, G. J. Nichols, and S. J. Robertsfor these to local deformation [Hall et al., 1995a, b]. Furthermore, discussionand their contributionsto the Sorong Fault Project; S. Cisowski and J. R. Dunn for advice and assistance with regionalimplications of the plate reconstructionswhich follow palaeomagneticwork; and C. Rundle and D. Rex for help with from the palaeomagneticdata, such as the characterof plate isotopic dating. Logistical assistancewas provided by GRDC, boundaries and timing of changes in their character, are Bandungand the Director, R. Sukamto,with excellentfield support supportedindependently by other lines of geologicalevidence by D. A. Agustiyanto,S. Atmawinata, A. Haryono, Kusnama,T. [Hall et al., 1995a, b; Ali and Hall, 1995]. Therefore we Padmawijadja,and S. Pandjaitan.Reconstructions were made using conclude(1) that the area north of the Sorong Fault in eastern the CambridgePaleomap ATLAS program.

References

Ali, J. R., and R. Hall, Evolutionof the boundary Charlton, T. R., R. Hall, and E. Partoyo, The Baker, Origin and motion history of the between the Philippine Sea Plate and geology and tectonic evolution of Waigeo Philippine Sea Plate: Evidence from eastern Australia: palaeomagnetic evidence from Island, NE Indonesia, J. SE Asian Sci., Indonesia,Tectonophysics, in press,1995a. eastern Indonesia, Tectonophysics,in press, 6, 289-297, 1991. Hall, R., M. Fuller, J. R. Ali, and C. D. Anderson, 1995. Clague,D. A., and R. D. Jarrard,Tertiary Pacific The Philippine Sea Plate: Magnetism and Ballantyne, P. D., Petrologicalconstraints upon plate motion deducedfrom the Hawaiian- reconstructions, in Active Margins and the provenance and genesis of the East Emperor Chain, Geol. Soc. Am. Bull., 84, Marginal Basins,Geophys. Monogr. Ser. vol. Halmaheraophiolite, J. SE AsianEarth Sci., 6, 1135-1154, 1973. 88. edited by Taylor and J. H. Natland, pp. 259-269, 1991 a. DeMets, C., R. G. Gordon, D. F. Argus, and S. 371-404, AGU WashingtonD.C., 1995b. Ballantyne,P. D., Petrologyand geochemistryof Stein, Current plate motions,Geophys. J. Int., Hamilton, W.H., Tectonics of the Indonesian the plutonicrocks of the Halmaheraophiolite, 101,425-478, 1990. region,U.S. Geol.Surv. Prof Pap., 1078,345 eastern Indonesia, an analogue of modern Dow, D. B., and R. Sukamto,Western Irian Jaya: pp., 1979. oceanic forearcs, Geol. Soc. Spec. Publ. The end-productof obliqueplate convergence Harland,W. B., R. L. Armstrong,A. V. Cox, L. E. London, 60, 179-202, 1991b. in the late Tertiary, Tectonophysics,106, Craig, A. G. Smith, and D. G. Smith, A Bleil, U., Paleomagnetismof DSDP Leg 60 109-139, 1984. Geologic Time Scale 1989, 263 pp., sedimentsand igneousrocks from the Mariana Dunn, J. R., and M. Fuller, Thermal CambridgeUniv. Press,New York, 1990. region,Initial Rep. Deep Sea Drill. Proj., 60, demagnetizationwith measurementsat high Haston,R. B., and M. Fuller, Paleomagneticdata 855-876, 1981. temperatureusing a SQUID magnetometer from the PhilippineSea Plate and their tectonic Bowin, C., R. S. Lu, C. Lee, and H. Schouten, (abstract),Eos Trans. AGU, 65, 863, 1984. Plate convergence and in the significance,J. Geophys.Res., 96, 6073-6098, Fisher, R. A., Dispersionon a sphere,Proc. R. 1991. -Luzonregion, AAPG Bull., 62, 1645- Soc. London. A, 217, 295-305, 1953. Haston, R. B., L. B. Stokking, and J. R. Ali, 1672, 1978. Hall, R., Plate boundary evolution in the Cardwell, R. K., B. L. Isacks,and D. E. Karig, The Palaeomagneticdata from holes 782A, 784A Halmaheraregion, Indonesia, Tectonophysics, and 786A, Leg 125, Proc. Ocean Drill. spatial distribution of , focal 144, 337-352, 1987. ProgramSci. Res., 125, 535-545, 1992. mechanism solutions, and subducted Hall, R., and G. J. Nichols, amalgamation Hatherton,T., andW. Dickinson,The relationship lithospherein the Philippine and northeastern in the PhilippineSea margin,Tectonophysics, betweenandesitic volcanism and seismicityin Indonesian islands, in The Tectonic and 181,207-222, 1990. Indonesia,the LesserAntilles, and otherisland Geologic Evolution of SoutheastAsian Seas Hall, R., M. G. Audley-Charles,F. T. Banner,S. and Islands, Geophys.Monogr. Ser., vol. 23, Hidayat,and S. L. Tobing,The basementrocks arcs,J. Geophys.Res., 74, 5301-5310,1969. edited by D. E. Hayes, pp. 1-35, AGU, of the Halmaheraregion, east Indonesia:A Hirooka, K., T. Takahasi, H. Sakai, and T. WashingtonD.C., 1980. Late Cretaceous-EarlyTertiary forearc, J. Nakajima, Paleomagneticevidence of the Carey, S. W., A tectonicapproach to continental Geol. $oc. London, 145, 65-84, 1988a. northwarddrift of the Izu peninsula,central drift, paper presented at the Symposium Hall, R., M. G. Audley-Charles,F. T. Banner, S. Japan,in Formationof ActiveOcean Margins, Continental Drift, University of , Hidayat, and S. L. Tobing, Late - editedby N. Nasu, pp. 775-787,Terra Sci., Hobart, Tasmania, Australia, 1958. Quatemary geology of Halmahera, eastern Tokyo, 1985. Chamot-Rooke, N., V. Renard, and X. Le Pichon, Indonesia: Initiation of a volcanic , J. Huchon, P., Comment on "Kinematics of the Magnetic anomaliesin the ShikokuBasin: A Geol. Soc. London, 145, 577-590, 1988b. Philippine Sea Plate", Tectonics, 5, new interpretation,Earth Planet.Sci. Lett., 83, Hall, R., G. J. Nichols, P. D. Ballantyne,T. 1355-1383, 1986. 214-228, 1987. Charlton, and J. R. Ali, The characterand Jarrard, R. D.,and S. Sasajima, Paleomagnetic Chamot-Rooke,N., K. Tamaki, and K. Kobayashi, significanceof basementrocks of the southern synthesisfor SoutheastAsia: Constraintson Deskewed magnetic profiles of the Shikoku MoluccaSea region, J. SEAsian Earth Sci., 6, platemotions, in The Tectonicand Geologic Basinand the pastkinematics of the Philippine 249-258, 1991. Evolution of Southeast Asian Seas and Sea Plate (abstract),Eos, 70, 1365-1366, 1989. Hall, R., J. R. Ali, C. D. Anderson,and S. J. Islands, Geophys. Monogr. Ser., vol. 23, 1132 HALL ET AL.: CENOZOIC MOTION OF THE PHILIPPINE SEA PLATE

edited by D. E. Hayes, pp. 293-316, AGU, from shallow and intermediate 23, editedby D. E. Hayes,pp. 49-75, AGU, Washington,D.C., 1980. activity, J. Geophys. Res., 87, 3663-3678, WashingtonD.C., 1980. Jolivet, L., P. Huchon, and C. Rangin, Tectonic 1982. Silver, E. A., R. McCaffrey, and R. B. Smith, settingof the westernPacific marginalbasins, McElhinny, M. W., Statisticalsignificance of the Collision, rotation, and the initiation of Tectonophysics,160, 23-47, 1989. fold test in palaeomagnetism,Geophys. J. R. subduction in the evolution of Sulawesi, Karig, D. E., Basingenesis in the PhilippineSea, Astron. Soc., 8, 338-340, 1964. Indonesia,J. Geophys. Res., 88, 9407-9418, editedby D. E. Karig et al., Initial Rep. Deep McFadden, P. L., A new fold test for 1983. Sea Drill. Proj., 31,857-879, 1975. palcomagneticstudies, Geophys. J. lnt., 103, Silver, E. A., and J. C. Moore, The Molucca Sea Katili, J. A., Volcanismand plate tectonicsin the 163-169, 1990. collisionzone, Indonesia, J. Geophys.Res., 83, Indonesian island arcs, Tectonophysics,26, Moore, G. F., and E. A. Silver, Collisionprocesses 1681-1691, 1978. 165-188, 1975. in the Northern Molucca Sea, in The Tectonic Stern, R. J., and S. H. Bloomer, Subductionzone Keating, B. H., and E. Herrero, Palcomagnetic and Geologic Evolution of Southeast Asian infancy:Examples from the EoceneIzu-Bonin- studies of and andesites from DSDP Seasand Islands, Part 2, vol. 27, editedby D. Mariana and Jurassic California arcs, Geol. Leg 59, Initial Rep. Deep Sea Drill. Proj., 59, E. Hayes, Geophys. Monogr. Ser., pp. 360- Soc. Am. Bull., 104, 1621-1636, 1992. 533-544, 1980. 372, AGU, WashingtonD.C., 1983. Tokuyama,H., Dredgedigneous rocks from the Kinoshita,H., Palcomagnetismof sedimentcores Nichols, G. J., and R. Hall, Basin formation and Amami Plateau, in Geology of the Northern from DSDP Leg 58, Philippine Sea, Initial Neogenesedimentation in a backarcsetting, PhilippineSea, edited by T. Shiki,pp. 50-56, Rep.Deep SeaDrill. Proj., 58, 765-768, 1980. Halmahera, eastern Indonesia, Mar. Pet. Geol., Tokai Univ., Shinjuku,Japan, 1985. Kodama,K., B. H. Keating, and C. E. Helsley, 8, 50-61, 1991. Uyeda, S., and Z. Ben-Avraham,Origin and Palcomagnetismof the Bonin Islandsand its Nichols, G. J., R. Hall, J. S. Milsore, D. Masson, developmentof the PhilippineSea, Nature tectonicsignificance, Tectonophysics, 95, 25- L. Parson,N. Sikumbang,B. Dwiyanto,and H. Phys.Sci., 240, 176-178, 1972. 42, 1983. Kallagher, The southern termination of the Vacquier,V., and S. Uyeda,Palaeomagnetism of Koyama, M., S. M. Cisowski, and P. Pezard, Philippine Trench, Tectonophysics, 183, nine seamounts in the western Pacific and of Palcomagneticevidence for northwarddrift 289-303, 1990. three volcanoesin Japan, Bull. Earthquake and clockwise rotation of the Izu-Bonin forearc Pigram,C. J., and H. L. Davies, and the Res.Inst., Univ. Tokyo,45, 815-848, 1967. sincethe early Oligocene,Proc. OceanDrill. accretionhistory of the New Guinea orogen, Visser,W. A., andJ. J. Hermes,Geological results ProgramSci. Res., 126, 353-370, 1992. BMR J. Aust. Geol. Geophys., 10, 193-211, of the explorationfor oil in NetherlandsNew Larson,E. E., et al., Palcomagnetismof Miocene 1987. Guinea, K. Ned. Mijnbouwk. Genoot. Verh. volcanic rocks of Guam and the curvature of Rangin, C., The : A Geol. Ser., 20, 265 pp., 1962. the southern Mariana island arc, Geol. Soc. complex plate boundary,J. SE Asian Earth Zijderveld, J. D. A., A.C. demagnetizationof Am. Bull., 86, 346-350, 1975. Sci., 6, 307-318, 1991. rocks: analysis of results, in: Methods in Louden,K., Magneticanomalies in the West Ranken, B., R. K. Cardwell, and D. E. Karig, Paleomagnetism,edited by D. W Collinson,K. PhilippineBasin, in The Geophysicsof the Kinematics of the Philippine Sea Plate, M. Creer, and S. K. Runcorn,pp. 254-286,. PacificOcean Basin and Its Margin: a volume Tectonics,3, 555-575, 1984. Elsevier, New York. 1967. in honorof GeorgeP. Woolard,Geophys. Seno, T., and S. Maruyama, Palcogeographic Monogr.Ser., vol. 19, edited by G. H. Sutton, reconstructionand origin of the PhilippineSea, M. H. Manghnani,and R. Moberly,eds., pp. Tectonophysics,102, 53-84, 1984. J. R. Ali, Department of Oceanography, 253-267,AGU WashingtonD.C., 1976. Seno,T., T. Mofiyama S. A., Stein, D. F. Woods, SouthamptonUniversity, Southampton, SO9 5NH, Louden, K. E., Palcomagnetismof DSDP and C. R. Demets, Redetermination of U.K. sediments,phase shifting of magnetic Philippine Sea Plate motion, Seismol. Soc. C. D. Anderson,Department of Geological anomaliesand rotationsof the West Philippine JapanProgram Abstr., 1, 86, 1987. Sciences,University of California,Santa Barbara, Basin,J. Geophys.Res., 82, 2989-3002,1977. Seno,T., S. A. Stein and A. E. Gfipp, A modelfor CA 93106. McCabe,R., andS. Uyeda,Hypothetical model the motion of the Philippine Sea Plate R. Hall, SE Asia ResearchGroup, Department for the bendingof the Marianaarc, in: The consistentwith NUVEL-I and geologicaldata, of Geology,Royal Holloway,London University, Tectonicand GeologicEvolution of Southeast J. Geophys.Res., 98, 17,941-17,948, 1993. Egham,Surrey, TW20 0EX, U.K. Asian Seas and Islands, part 2 Geophys. Shih, T-C., Marine magneticanomalies from the (e-mail: robert.hall@ gl.rhbnc.ac.uk) Monogr.Ser., vol. 23, editedby D. E. Hayes, western Philippine Sea: Implicationsfor the pp.281-293, AGU, Washington D.C., 1983. evolutionof marginalbasins, in: The Tecwnic (ReceivedFebruary 15, 1995; McCaffrey,R., Lithosphericdeformation within and Geologic Evolution of Southeast Asian revisedJune 1, 1995; the Molucca Sea arc-arc collision: Evidence Seasand Islands,Geophys. Monogr. Ser., vol. acceptedJune 5, 1995.)