38. Geochemistry and Origin of Igneous Rocks from the Outer Tonga ...
Transcript of 38. Geochemistry and Origin of Igneous Rocks from the Outer Tonga ...
Hawkins, J., Parson, L., Allan, J., et al., 1994Proceedings of the Ocean Drilling Program, Scientific Results, Vol. 135
38. GEOCHEMISTRY AND ORIGIN OF IGNEOUS ROCKSFROM THE OUTER TONGA FOREARC (SITE 841)1
Sherman H. Bloomer,2 Anthony Ewart,3 Janet M. Hergt,4 and Wilfred B. Bryan5
ABSTRACT
Two igneous rock units were recovered at Site 841. More than 200 m of island-arc rhyolites, rhyolitic tuffs, lapilli tuffs, andpumice breccias, divided into five units, compose the basement at the site. These rhyolitic volcanics are late middle Eocene orolder and formed part of a subaerial rhyolitic volcano. These low-K rhyolites were produced by fractional crystallization of amore mafic arc-tholeiitic lava or by dehydration melting of lower crustal arc tholeiites. The Site 841 basement rocks are similarin composition to high-SiO2 lavas in the Eocene basement on 'Eua and crystallized from depleted island-arc-tholeiitic basalts likethose exposed on 'Eua. No evidence is present in the rhyolites, or in the clasts enclosed within them, for boninite series magmasat Site 841. The Site 841 rhyolitic complex bears no resemblance to Cretaceous rhyolites from the Lord Howe Rise, which areenriched in K and incompatible elements. The volcanic rocks at Site 841 are part of a widely distributed Eocene volcanic episodethat marked the earliest phases of subduction in the Tonga region; they are not part of an older crustal fragment.
The second igneous sequence is a series of basaltic dikes and sills that intruded Miocene sediments. These basalts have traceelement abundances and ratios identical to upper Miocene lavas from the Lau Ridge. The Site 841 basalts do not have anygeochemical characteristics that suggest they were generated by unusual thermal conditions in the shallow sub-forearc mantle.They are most reasonably interpreted as intrusions fed by basement dikes propagated from the associated active arc.
No evidence for local serpentinite exposures, like those that are common in the Mariana forearc, was found at Site 841. Theresults from Site 841 provide strong support for hypotheses of forearc evolution that have been advanced for the Izu-Bonin-Mariana system.
INTRODUCTION
Forearcs are an integral part of intraoceanic convergent margins.Their basement contains a record of the early evolution of the sub-duction system (e.g., Hussong and Uyeda, 1982; Bloomer, 1983;Taylor, 1992) and the volcaniclastic sediments covering them docu-ment the tectonic and geochemical evolution of the arc.
Marine geologic studies and Deep Sea Drilling Project (DSDP)and Ocean Drilling Program (ODP) transects of the intraoceanic Izu-Bonin-Mariana (IBM) forearcs have generated four important newideas about the origin and evolution of these terranes. The first, andmost striking, is the recognition that the initial phases of volcanism inthese subduction zones developed nearly synchronously in the middleto late Eocene over a zone up to 300 km wide (Taylor, 1992) andthousands of kilometers long (Stern and Bloomer, 1992; Taylor, 1992).This early, or "infant," arc volcanism was characterized by the eruptionof very depleted boninitic and arc-tholeiitic lava compositions (e.g.,Meijer, 1980; Ishii, 1985; Bloomer, 1987; Murton et al., 1992) anddeveloped in an extensional setting (as evidenced by dikes on Chichi-jima in the Bonin Islands; Umino, 1985), in places possibly to the pointof true seafloor spreading (Stern and Bloomer, 1992). This mode ofarc volcanism is unlike that developed during the "normal" or maturephases of arc activity and is a plausible mechanism for developingsupra-subduction zone ophiolites (what Pearce et al. [1984] termed"pre-arc" ophiolites; Robertson, 1990; Stern and Bloomer, 1992).
A second new observation is that the exposure of Eocene arcbasement immediately adjacent to the axis of the Mariana-Bonin
Hawkins, J., Parson, L., Allan, J., et al., 1994. Proc. ODP, Sci. Results, 135: CollegeStation, TX (Ocean Drilling Program).
2 Department of Earth Sciences, Boston University, Boston, MA 02215, U.S.A.' Department of Geology and Mineralogy, University of Queensland, St. Lucia,
Brisbane, Queensland. 4067, Australia.4 Department of Earth Sciences. The Open University, Walton Hall, Milton Keynes,
MK7 6AA, United Kingdom (present address: School of Earth Sciences, The Universityof Melbourne, Parkville, Victoria, 3052, Australia).
• Department of Geology and Geophysics, Woods Hole Oceanographic Institution,Woods Hole, MA 02543, U.S.A.
Trench requires some subduction erosion since the Eocene (Hussongand Uyeda, 1982; Bloomer, 1983). However, if the Eocene basementdid form by unusual, voluminous volcanism in an extensional envi-ronment, the required amount of erosion may only be 20-50 km(Johnson et al., 1991; Stern and Bloomer, 1992; Taylor, 1992). Clearevidence is also present of episodic, post-Eocene accretion of PacificPlate sediments and crustal fragments to the outer forearc (Karig andRanken, 1983; Bloomer, 1983; Johnson et al., 1991).
The large volume of serpentinite in the outer forearc has been asurprising discovery. The serpentinites occur as diapirs, with associ-ated serpentine mud flows (Fryer et al., 1985; Bloomer, 1983). Pore-water geochemical studies suggest that the fluids responsible for theserpentinization are derived in part by dehydration of the downgoingslab (Haggerty, 1987; Mottl et al., 1989).
Finally, the drilling transects documented episodes of magmaticrejuvenation in supposedly "cold" forearcs. Sills and dikes cuttingMiocene and younger sediments were found in two holes drilled inthe outer Mariana-Bonin forearc (Shipboard Scientific Party, 1990a,1990b). These intrusions were emplaced in the forearc at a time whenthe active arc was well to the west, and indicate either a reheating ofthe forearc or intrusion of long dikes and sills from the active arc. Leg126 also found evidence for rifting and volcanism within the Izu-Bonin forearc during the Oligocene, at a time when the main activearc farther to the west (Taylor, 1992).
These observations in the Izu-Mariana-Bonin system have painteda far different picture of intraoceanic forearcs than existed 11 yr agowhen the results of DSDP Leg 60 documented in-situ Eocene bonin-itic basement in the outer Mariana forearc (Hussong and Uyeda,1982). These forearcs are geologically complex, and are clearly directproducts of subduction, not just trapped pieces of oceanic crust. Theyhave geologic and geochemical similarities to supra-subduction zoneophiolites like those in Troodos and Oman and represent a previouslyunappreciated type of crustal construction.
Before Leg 135, it was not known whether the hypotheses offorearc evolution derived from the IBM studies were generally appli-cable to other intraoceanic forearcs. Seismic and dredging studies inthe Tonga forearc had suggested that arc basement did occur in the
S.H. BLOOMER. A. EWART, J.M. HERGT, W.B. BRYAN
outer forearc (cf. Fisher and Engel, 1969; Bloomer and Fisher, 1987),but there were not good constraints on the age and composition of thisbasement. Site 841 was drilled just west of the outer trench-slopebreak to determine the age and composition of the forearc basementand to examine the evolution of the Tonga forearc as a test of thehypotheses proposed for the Izu-Marian-Bonin system. A secondobjective at the site was to recover the volcaniclastic sediments over-lying the basement to constrain the variation of arc volcanic composi-tions through time and the uplift and subsidence history of the forearc.
Two igneous sequences were recovered from Hole 84IB. Theuppermost is a series of basaltic dikes and sills intruded into Miocenesediments and the lower one is a sequence of low-K rhyolitic tuffs,rhyolites, and pumice breccias that constitute the forearc basement atthis location. This silicic complex is older than late middle Eocene,as there are late middle Eocene calcareous volcanic sandstones, con-taining rhyolitic clasts, normally faulted against the volcanics.
The purpose of this paper is to present detailed geochemical analy-ses of the two major igneous groups from Hole 84IB and to examinethe tectonic implications of their compositions. We will also draw somecomparisons with the evolution of the Izu-Mariana-Bonin forearc.
BACKGROUND
Regional Geology
The Tonga forearc is a part of a complex array of arcs and backarcsthat have developed as a result of the subduction of the Pacific Platebelow the Indo-Australia Plate since at least middle Eocene time (e.g.,Herzer and Exon, 1985). This subduction is presently near perpen-dicular to the trench at a rate of about 16 cm/yr (including an 8 cm/yropening rate in the Lau Basin; Pelletier and Louat, 1989). There aretwo chains of islands parallel to the trench (Fig. 1), separated fromeach other by the deep, sedimented Tofua Trough (Raitt et al., 1955).The eastern chain, the Tonga platform, includes extinct, coralline-capped Eocene to Miocene volcanic islands and the western chain,the Tofua Arc, comprises active submarine and subaerial volcanoeserupting basalt, basaltic andesite, and minor dacite (e.g., Ewart et al.,1977). West of the Tofua Arc lies the actively spreading Lau Basin,which is bounded on the west by the inactive Lau Ridge, a remnantarc (Cole et al., 1990).
The Tonga forearc and platform have existed since Eocene time,when they formed as part of the ancient Melanesian arc, which includedfragments of what are now the Fiji, Lau, Tonga, and New Hebridesarcs. The development of the subduction zones associated with thesearcs was preceded by rifting of the eastern Australian margin in LateCretaceous time; the Lord Howe Rise and the Norfolk Ridge are blocksof this margin that were rafted to the east (Burns and Andrews, 1973).These ridges were separated from the margin by seafloor spreading inthe New Caledonia Basin in late Paleocene time and in the TasmanBasin and Coral Sea through early Eocene time (Bentz, 1974).
The west-dipping subduction that built the Melanesian arc wasinitiated sometime after the early Eocene (Herzer and Exon, 1985).The early growth of the arc was accompanied by spreading in theSouth Fiji Basin, which continued until about 26 Ma (Malahoff et al.,1982). The New Hebrides arc separated from the Lau-Fiji-Tonga arcduring the late Miocene, after the initiation of an east-dipping sub-duction zone beneath the New Hebrides arc (Herzer and Exon, 1985).Volcanism in the Tonga system continued through the Miocene, alongthe Lau Ridge. The Tonga platform accumulated a thick clastic sedi-mentary cover during the Miocene, as it was the forearc east of theactive Lau Ridge (Herzer and Exon, 1985). This forearc sedimentaryfill is cut locally by Miocene dikes and sills (21.3 and 13.9 Ma;Cunningham and Anscombe, 1985). The main phase of this period ofarc volcanic activity is represented by the Lau Group volcanics on theLau Ridge (Cole et al., 1985). A regional unconformity developedover the Tonga platform in the late Miocene or early Pliocene, as aconsequence of uplift associated with the initiation of spreading inthe Lau Basin (Herzer and Exon, 1985;Packham, 1985). This opening
15° S
2O°S
25"S
180°W 17O°W
Figure 1. Location of Site 841 (star). Filled circles show places where volcanicor plutonic rocks of island arc affinities have been dredged from the forearcand trench slope (Fisher and Engel, 1969; Bloomer and Fisher, 1987; Valuer,O'Connor,etal., 1985;Gnibidenkoetal., 1985; Sharaskinetal., 1983;Falloonet al., 1987). Bathymetry in meters.
of the Lau Basin about 6 Ma ago (Parson, Hawkins, Allan, et al., 1992)split the Lau Ridge from the Tonga Platform and led to the develop-ment of the presently active arc to the west of the Tonga platform (Fig.2). Volcanism continued on the Lau Ridge till about 2.4 Ma, as theLau Basin opened. This volcanism produced the Korobasaga Group,a large-ion-lithophile-enriched, slightly alkalic, arc series (Cole et al.,1985). The opening of the Lau Basin has rafted the Tonga platformto the east, and rotated the entire forearc clockwise (see Sager andMacLeod, this volume).
Forearc Geology
The forearc is flanked on the west by the Tonga platform, aninactive frontal arc now covered by a thick sedimentary sequence(Fig. 1). The basement of the Tonga platform is exposed only on theisland of 'Eua (Fig. 1). That basement comprises depleted island-arc -tholeiitic basalts, andesites, and dacites (Ewart and Bryan, 1972;Hawkins and Falvey, 1985; Cunningham and Anscombe, 1985). Thebasement rocks yield radiometric ages of 46 to 40 Ma (Ewart et al.,1977; Duncan et al., 1985) and, locally, are unconformably overlainby upper middle Eocene calcareous conglomerates and breccias (Cun-ningham and Anscombe, 1985). Oligocene lava flows (33-31 Ma)and Miocene dikes (19-17 Ma) indicate later episodes of arc activity(Duncan et al., 1985) that predate the opening of the Lau Basin whenthe Tonga platform was still part of the Lau Ridge.
Seismic reflection profiles suggest that the Eocene basement on'Eua can be traced across the forearc and crops out on the landward
-
GEOCHEMISTRY AND ORIGIN OF IGNEOUS ROCKS
Site 841
Arc crust
Oceanic crust
Vertical exaggeration = 1.5
Figure 2. Schematic cross section of the Tonga forearc showing the relativepositions of 'Ata, the active arc, 'Eua, on the eastern Tonga Ridge, and Site 841on the trench-slope break. Crustal thicknesses are roughly after Raitt et al. (1955);these are not well constrained. Depth to Benioff zone from Billington (1980).
trench slope, just outboard of the trench-slope break (Greene andWong, 1983; Gnibidenko et al, 1985). Fragments of arc-tholeiiticbasalts and andesites, boninites, highly depleted serpentinites, andgabbros have been dredged from a number of places on the landwardslope (Fig. 1) supporting this suggestion that the entire forearc isfloored by arc crust. The age of the outer forearc basement is notconstrained by these dredge samples. Boninite glasses from the north-ern trench have yielded ages of < 10 Ma (A. Crawford, pers. comm.,1992), but gabbroic rocks from the same suite have a 50 Ma age(Acland et al., 1992). Boninitic glasses commonly yield anomalouslyyoung ages (Tsunakawa, 1983); the boninites at the northern trenchslope are in a complicated region at the terminus of the active arc andLau Basin, along a trench-trench transform. They may indeed be young,but their tectonic significance is difficult to assess at present. Thisambiguity concerning the age of the forearc basement made one of theprincipal objectives of Site 841 to determine the basement age in aportion of the forearc that has undergone only normal convergence andis well-removed from the loci of Oligocene and younger arc volcanism.
The exposure of arc volcanic rocks in the mid- and upper-land-ward trench slopes has been taken as evidence for subduction erosionof the forearc (Bloomer and Fisher, 1987; Valuer, O'Connor, et al,1985) since such exposures are unlikely to have been generated insitu. Lonsdale (1986) has postulated that subduction erosion has beenaccelerated by the passage of the Louisville Ridge southward alongthe forearc. There has been, however, minor accretion of seamountfragments and offshore pelagic sediments to the lower landward slopes(Bloomer and Fisher, 1987; Valuer, O'Connor, et al., 1985).
The forearc has been affected by a number of tectonic events.Drilling at Site 840 on the Tonga platform documented two periodsof rapid subsidence on the crest of the ridge, one after 9 Ma, coincidingwith the breakup of the Melanesian Arc, and one dating from 5.25Ma, related to arc rifting during the opening of the Lau Basin (Clift,this volume). The passage of the Louisville Seamount Chain south-ward down the forearc does not appear to have produced any docu-mentable uplift or subsidence in the forearc (Clift, this volume),though it does appear to have produced some faulting and fracturingof the forearc (Gnibidenko et al., 1985; Bloomer et al., 1988).
Site 841
Site 841 was located just west of the trench-slope break (Figs.1-2), 48 km inboard of the axis of the Tonga Trench, 140 km outboardof the active arc island of 'Ata, and about 70 km outboard of theeastern edge of the Tonga platform (the equivalent position of 'Eua).The major objective for the site was to determine the age and compo-sition of the forearc basement, to develop a model for the origin of
2 0 0 ^
6 0 0 -
I Plθist.-Pliocθnθclays, sands
II upper Miocene
clays, silts, sands
II upper Miocene
siltstoneconglomerate
IV lower mid. Miocene
volcanicsandstone
V. Iwr. Olig.-u. Eoc.calc. sandstones.
rhyolitesand
rhyolitictuffs
6 0 0 -
w§o&
<ri
8 0 0 -
2A breccias
2B rhyolitβ
2C pumice breccia
3Atuff3B iapillituff3Ciapilli tuff
3D rhyolite
breccias
4 Iapilli tuff
5 sheared breccias
6 welded
Iapilli
tuffs
8 0 0 -
Figure 3. Schematic stratigraphy of the principal sedimentary and igneous unitsfound at Site 841 (from Shipboard Scientific Party, 1992).
the forearc, and to compare the evolution of the Tonga forearc tohypotheses of forearc evolution derived in the IBM region. A second-ary objective was to recover the overlying sedimentary section, whichcould be used to understand the tectonic history of the forearc and thegeochemical evolution of the adjacent island arc.
Hole 841B penetrated 834.2 mbsf in a water depth of 4810 m. The210.55 m of drilled igneous basement (with about 14% recovery) atthe site is composed of low-K rhyolites, rhyolitic tuffs, breccias,welded tuffs, and Iapilli tuffs (Fig. 3). The welded fragments, theabsence of interbedded sediments and marine fauna, and the lack ofdensity sorting within the pumice breccias indicate a subaerial originfor the rhyolitic complex. Two possibilities were considered for theorigin of this complex (Parson, Hawkins, Allan, et al., 1992). First,this could be an Eocene construct equivalent to the rocks that areexposed on 'Eua. Alternatively, the Site 841 basement could be afragment of a rifted block like the Cretaceous (93.7 ± 1.2 Ma; van derLingen, 1973) Lord Howe Rise, which also comprises high-silicaIapilli tuffs and rhyolites. Preliminary K-Ar dating of one of therhyolites yielded ages of 44 and 45 Ma, and a suspect date of 67 Ma(see McDougall, this volume). Illites from the rhyolitic tuffs yield
SH. BLOOMER. A. EWART. J.M. HERGT, W.B. BRYAN
K-Ar ages of 52.8,51.4, and 31.2 Ma (D. Schöps, pers. comm., 1993).These ages cluster in the Eocene, supporting the former hypothesis(see McDougall, this volume).
Fifty-five m of Eocene and lower Oligocene calcareous volcanicsandstones, with basal ages of about 41.3 Ma (foraminifer Zone PI4;Parson, Hawkins, Allan, et al, 1992), are faulted against the silicicbasement. This fault strikes north to northeast and dips east at about10°-30° (see MacLeod, this volume). Parasitic structures are consis-tent with normal motion on this fault. The fauna in the Eocene sedi-ments indicate very shallow-water (photic zone) depositional condi-tions (see Chaproniere, this volume). In combination with the evi-dence for subaerial eruption of the rhyolitic basement complex, it isclear that the basement at Site 841 has subsided over 5000 m sincethe late Eocene.
The Eocene-Oligocene sediments are overlain by 91 m of lowermiddle Miocene turbiditic sandstones and siltstones; these are uncon-formably overlain by 125 m of upper Miocene siltstones, sandstones,and volcanic conglomerates, some with reworked Eocene fauna, and277 m of late Miocene vitric silts and sandstones. This Miocenesection is intruded by several orthopyroxene-clinopyroxene-plagio-clase-phyric basaltic dikes and sills. Fifty-six m of Pleistocene toPliocene clays, sands, and ashes cap the section. The entire sedimen-tary sequence is cut by microfaults, most of which are normal faults,although minor reverse faults also occur (see MacLeod, this volume).Two major faults cut the sections at 449-458 and 605 mbsf. The upperfaults dips about 60°, has a normal displacement, and may haveaccommodated over 1 km of motion (see MacLeod, this volume). Thelower fault separates the Eocene calcareous sediments from the rhy-olitic basement.
The two igneous groups at the site—the intrusives into the Mio-cene sediments and the pre-late Eocene rhyolitic basement—recordtwo different magmatic pulses in the history of the Tonga forearc.We will examine the geochemistry and significance of each igneousgroup separately.
METHODS
Major and trace element data reported here for igneous rocks fromSite 841 are from the JOIDES Resolution shipboard analyses, BostonUniversity, the Woods Hole Oceanographic Institution, the Universityof Durham, the University of Queensland, Monash University, andthe Australian National University. Shipboard analyses were by X-ray fluorescence (XRF); methods are discussed in Parson, Hawkins,Allan, et al. (1992). Samples were ignited at 1100°C before beinganalyzed. A discussion of shipboard precision and sample contami-nation during grinding are presented in Hergt and Sims (this volume).
Analyses at the Woods Hole Oceanographic Institution are ofwhole-rock powders by XRF. Techniques are discussed in Bryan etal. (this volume). Samples were ignited before being analyzed. Analy-ses from the University of Queensland for major elements and for Rb,Zr, Zn, Y, Ni, Cr, V, Se, Co, Cu, and Zn were also by XRF. Be and Naon the same samples were determined by atomic absorption spec-trometry; some of these samples were analyzed for rare earths byinstrumental neutron activation analysis (INAA) at the AustralianNational University (B.W. Chappell, analyst) and for rare-earth ele-ments (REEs), Cs, Ba, Nb, U, Th, Pb, Ta, Hf, W, TI, and Ga byinductively coupled plasma source mass spectrometry (ICP-MS) atMonash University (W.W. Ahlers, analyst). Samples were ignitedbefore being weighed and analyzed. All analyses were of bulk-rockpowders, except for the analysis of Sample 135-841B-50R-1,5-9 cm,which was of glass separated from a plagioclase-quartz phyric pitch-stone. Details of the analytical methods and errors are included inEwart et al. (this volume).
Analyses from the University of Durham were of glass separatesand were done by ICP-MS. The methodology is given in Hergt andNilsson (this volume). Samples were not ignited before being ana-lyzed and major element analyses were normalized to 100%.
Analyses from Boston University were by inductively coupledplasma emission (ICP) spectrometry. Samples were dried at 110°Cbefore being weighed, but they were not ignited. Loss on ignition(LOI) was not determined for the samples because of the small sizeof many of them. Analyses of samples from Sections 135-841B-57R-I and -59R-1 were of whole-rock samples crushed in diamonitemortars. Glass was separated from a pitchstone from Section 135-841B-47R-2; analyses from Sections 135-841B-23R-4, -25R-4, and-26R-2 are of hand-picked separates from crushed and sieved bulkrock. This separation was done to try to remove the ubiquitous sili-ceous and carbonate veining in the dike and sill samples. The cleanedsamples were powdered in a diamonite mortar, dried at 110°C, andmixed 1:3 with a lithium metaborate flux. The mixtures were fusedat 1000°C for 20 min and the glasses then dissolved in 50 mL of 10%HNO3. These solutions were used for trace element determinations;major elements were determined on 1-mL aliquots from the traceelement solutions, which were diluted with 25 mL of 10% HNO3 andspiked with Li2CO3 as a matrix modifier. Standard curves were cal-culated using aqueous standards and USGS reference rock powders.
A number of small volcanic rock inclusions from the sedimentarysection and from the igneous basement were analyzed at Boston Uni-versity. These inclusions range from large clasts (over 1 g) to small,individual grains picked from crushed sediment or basement samples.The latter were as small as 1 mg. These samples were crushed in adiamonite mortar, or if less than 0.1 g, simply mixed directly withlithium metaborate flux, after being washed and dried. A flux:rockratio of 3:1 was used for samples larger than 0.02 g; these mixtureswere fused and dissolved at 50:1 in dilute nitric acid. Smaller sampleswere mixed with 0.05 g flux and dissolved in 13 mL of solution. Someof the solutions were too dilute to obtain reliable minor and traceelement analyses. Weighing errors for the very small samples pro-duced sums ranging from 90% to 117%. Sums outside this range werediscarded, and sums less than 95% and greater than 105% were thennormalized to 100%.
Analyses of interlaboratory reference powders, prepared aboardship in an agate mill, are listed in Table 1. The analyses from thedifferent laboratories and different techniques generally agree verywell. Also listed are analyses of USGS reference rocks W-2 andBHVO, and an estimate of precision for the Boston University lab,based on replicate analyses of the interlaboratory reference powders.
The Pb-, Nd-, and Sr-isotopic analyses were performed at theOpen University. Techniques are described in Hergt and Hawkes-worth (this volume).
Analyses of rocks from Site 841 are plotted normalized to normalmid-ocean-ridge basalt (N-MORB) compositions in several figures.The normalizing values used are from Sun and McDonough (1989).
POST-LATE MIOCENE DIKES AND SILLS (UNIT 1)
Description
Nine intervals of basalt, ranging in thickness from 7 cm to 18 m,were drilled within Miocene sediments between 324.76 and 497.68mbsf. These basalts are petrographically similar and were denoted asSubunits 1A through IL The basalt-sediment contacts, when recov-ered, are near horizontal to steeply dipping. Contacts include glassymargins on basalts, brecciated hyaloclastite zones, and baked sedi-ments, suggesting that the units are dikes and sills rather than flows.Subunits 1A-1G are intruded into upper Miocene sediments. Subunit1H occurs as broken fragments at the base of a major fault that juxta-poses upper Miocene and lower middle Miocene sediments. SubunitII is within lower mid-Miocene sediments.
Formation MicroScanner (FMS) records show that the distribu-tion of the basaltic sills and dikes is more complex than shown by therecovered core. For example, Subunit 1D was defined as a 17-m-thicksill between 383 and 400 mbsf. The FMS images show that the unitactually occurs between 380 and 396 mbsf as 11 individual intrusionswith thicknesses from 5 to 134 cm. The contacts of the dikes strike
GEOCHEMISTRY AND ORIGIN OF IGNEOUS ROCKS
1 0 0 -
fflerO
I1 0 -
1 -
0.1
A
+O×Δπ•
•—i 1 1 1 1 1 —
Eua, Eocene
Lau, Miocene
Korobasaga, Pliocene
Tofua Arc, Recent
Ata Island, Recent
Hole 841B, Unit 1
×
Rb Ba Nb La Ce Sr Zr TiO
Figure 4. Comparison of compositions of average Unit 1 basalt with arcvolcanic units in the Tonga region with 52%-56% SiO2, normalized to N-MORB (values from Sun and McDonough, 1989). Analyses from 'Eua fromCunningham and Anscombe (1985) and Hawkins and Falvey (1985); Lau andKorobasaga analyses from Cole et al. (1985); Tofua Arc (active arc) averagecompositions from Ewart and Hawkesworth (1987), and average for 'AtaIsland (active arc) from Valuer, Stevenson, Scholl et al. (1985).
211° to 31° and, with one exception, dip toward the trench at 2° to59° (see MacLeod, this volume; dips are calculated after rotation ofsedimentary layering to horizontal).
The sills and dikes are orthopyroxene-clinopyroxene-plagioclasephyric basalts. Plagioclase is the dominant phenocryst, constituting4-12 modal %. Clinopyroxene ranges from trace amounts to 5 modal%and up to 2 modal% orthopyroxene was found, though it is absent insome samples. The basalts have 3%-5% vesicles, which tend to be con-centrated toward the contacts of the intrusions. A detailed petrographyof each subunit can be found in Parson, Hawkins, Allan, et al. (1992).
Most of the samples have been affected by a low-temperaturealteration. Siliceous, carbonate, K-feldspar, and thaumasite veins arecommon. Minor amounts of sulfides, principally pyrite, are devel-oped in some of the veins. The groundmass is typically 5%-50%replaced by clays, chlorite, and zeolites. The basalt in Subunit II isunusual in its lack of alteration.
Geochemistry
Eleven different sill or dike samples were analyzed; all are fromthe intrusions in upper Miocene sediments. SiO2 ranges from 50% to54%; the shipboard analyses appear to be high in comparison withthe Queensland and Boston University analyses, even with correc-tions for LOI (Table 2). The average composition for the intrusions is51.7% SiO2, including the shipboard analyses, and the samples shouldbe referred to as basalts (Le Bas et al., 1986).
All of the samples have essentially identical compositions, withininterlaboratory errors (Tables 2-3). The samples are relatively enrichedin Ba, K, and Sr; they are depleted in Nb and Ta; and they have slightlight-rare-earth-element (LREE) depletions (Fig. 4 and Table 2). Theirvery similar compositions support the shipboard conclusion, based onpetrography, that the individual basalts are all part of a single intrusiveevent. There are no analyses from Subunits IC, 1G, 1H, or II, but theirigneous mineralogy and textures are identical to the analyzed basalts;we interpret them to be part of the same intrusive episode.
Three of the basalt samples have Ba concentrations greater than250 ppm, and Ba in general varies more than other elements (Table 2).The three high-Ba samples are all noted as being extensively veined;
38.9
Q_
sPJ
21
37.7
• Site 841-dikes* Site 841-rhyolitθπ Ataθ'Eua
Tafahi andNiuatoputapu
18.1 18.9
15.65
Q_
15.45
18.1 206Pb/^Pb18.9
0.51315
T32
T3»
14
3f
0.51285
-\Sitθ 834
\Λ \ éPacific^V/
' MORB X ,Lau Group
--
Modern Lau Basin
\<^ o
IndianMORB^ • ^ ^ ,
1
. I t Korobasaga Group
‰r~^ Tofua \ ~y V Arc X
Tafahi and " " / V öNiuatoputapu
i i • i i
0.7025 0.7029 _7 . . 0.70378 7 Sr / 8 6 Sr
0.7041
Figure 5. Isotopic compositions of Site 841 igneous rocks compared to fieldsfor Pacific and Indian MORB, modern Lau Basin Basalts, the southern TongaArc (Tofua Arc), Tafahi Island (northern Tonga Arc), and Pacific and TongaTrench sediments. For a discussion of the analytical methods and sources ofreference fields, see Hergt and Nilsson (this volume). Dark-shaded Sr-Nd field= Lau Group volcanics, and light-shaded field = Korobasaga Group (both fromCole et al., 1990).
the high Ba concentrations are likely a consequence of alteration,perhaps from the development of authigenic alkali feldspar veins.
The compositions of the basalt intrusives are typical for arc vol-canics of all ages from the Tonga system (Fig. 4 and Table 3). Theyhave the same element enrichments and depletions and fall well withinthe range of arc compositions of similar SiO2. Their isotopic ratiosare within the range of modern Tonga Arc and Lau Basin basalts(Table 4 and Fig. 5) and do not show a pronounced sediment contami-nation signature (see Hergt and Nilsson, this volume). Their elementaland isotopic compositions indicate that the Unit 1 basalts, like mostarc lavas in the Tonga region, are fractionates of mantle-derived melts.
The Unit 1 basalts are not as depleted in large-ion-lithophile ele-ments (LILEs) and high-field-strength elements (HFSEs) as are theEocene lavas on 'Eua (Table 3). Neither are the basalts as depleted as
629
S.H. BLOOMER, A. EWART, J.M. HERGT, W.B. BRYAN
Table 1. Comparison of major and trace element analyses of shipboard powders (ground in agate) used for interlaboratory comparison.
Sample
LaboratoryMethod
BUICP
Major elements (wt%):SiO2TiO2A12O3Fe2O3yFeOMnOMgOCaONa2O
P 2 O 5SumLOI
52.841.04
15.4910.54*
0.1626.15
10.602.260.630.15
99.85—
Trace elements (ppm):ZnNiCrVCuZrSeYSrBaRbNbCsUThPbCoTaHfWGaTILaCePrNdSmEuGdTbDyHoErTinYbLu
876907
26S1111003623
204172NDND
—
—————
11
—
———
—
W-2
Reported
52.811.06
15.4910.86*
0.1676.39
10.892.210.630.14
807092
2601051003623
190175NDND
10.5
BHVO
BUICP
49.162.90
13.4712.29*
0.1677.21
11.312.250.510.29
99.30—
Ill1282983151381753127
398121NDND
15
—
Reported
49.902.69
13.8512.23*
0.1707.31
11.332.290.540.28
1051203003201401803127
420135NDND
17
Precision ofreplicates
BUICP
0.320.010.060.03
0.0010.070.090.040.010.003
—
5.50.11.11.40.90.70.30.11.60.9
NDND
0.3
135-834B-33R-2, 105-110 cm
QueenslandXRF, AA
49.321.27
17.773.804.530.138.17
11.803.190.070.10
100.15f3.16
6488
31318756953123.7
156100.41.6
—
34
—
—
—
—
WHOIXRF
49.561.23
18.018.47*
0.148.08
12.272.770.080.13
99.97t—
57100264208
6597
—26
156142
46
—
BUICP
48.641.24
17.708.57*
0.1317.78
11.642.940.0750.13
98.85—
7089
326210
74923426
16813
———
—————4
——
——
—
DurhamICP-MS
—1.18——
0.142—————
—
65100301211
738135.924.8
18140.41.30.010.060.070.5
39.60.121.740.02
170.043.149.821.618.992.811.063.710.653.840.862.570.372.10.43
135-836B-5R-2, 65-74 cm
QueenslandXRF,AA
49.450.83
15.953.585.280.159.59
13.512.170.050.04
100.60t3.21
6995
25424810839.53016.6
1361910.95
—
0.9441—————
———
————
—
WHOI BUXRF ICP
49.30 48.210.78 0.83
15.82 15.289.14* 9.49*
0.14 0.1409.87 9.44
14.06 13.251.82 2.100.06 0.0450.06 0.06
100.22t 98.33— —
62 70107 98213 314239 266110 10841 37
47.519 19
137 1362 17.52 —— —— —
———
49 —__ —— —— —— —— —— 2
—_ —— —— —
——
— _— —— —— —
——
— —
DurhamICP-MS
—0.81
——
0.139—————
—
64106317271104344919
14370.60.390.010.050.050.65
45.10.050.850.02
140.081.314.260.724.181.70.632.230.452.80.591.850.271.660.37
the lavas of the presently active arc (Table 3). The Site 841 intrusivesbear the closest resemblance to lavas of the Lau Group, a sequenceof 14- to 5.4-Ma basalts to dacites from the Lau Ridge remnant arc(Cole et al., 1985; Table 3). The least altered basalts have Y/Zr andBa/Zr (the most diagnostic difference between arc units of differentages in the Tonga region) very close to those of the Lau Group lavas(Fig. 6). The Ba/La ratios in the least altered samples range from 24to 43; the Lau Group has Ba/La ratios of 10 to 37, whereas the activeTofua Arc has Ba/La ratios of 36 to 160 (Cole et al., 1990). Unit 1Sr/Nd values are 21 to 30, overlapping the range of the Lau Group(23 to 25), but they are much lower than lavas of the active arc (34 to116; Cole et al., 1990). The La/Zr of Unit 1 lavas are like those of theLau Group, although their La and Zr abundances are at the low endof the range for those lavas (Fig. 6). The Unit 1 basalts have Sr andNd isotopic compositions within the range for lavas of the Koro-basaga Group, a lava series that developed late in the evolution of theLau Ridge (4.4-2.4 Ma) as the Lau Basin opened (Cole et al., 1985;Fig. 5). However, the Unit 1 lavas are not anywhere near as enrichedin LILEs as the Korobasaga Group (Fig. 3 and Table 3).
The element abundances and ratios show a striking similaritybetween the Unit 1 basalt intrusives and the late Miocene Lau Group.
This similarity is consistent with the apparent age of the intrusives. Ifall of the dikes and sills are indeed part of a single intrusive event,that event must be late Miocene (foraminifer Zone NI6 to Subzone17a; 12-6 Ma; Parson, Hawkins, Allan, et al., 1992) or younger.
Discussion
The discovery of upper Miocene or younger sills and dikes in theouter forearc is surprising. Site 841 is now over 140 km from thepresently active Tofua arc, and at least 75 km from what may havebeen the axis of late Miocene volcanism (near 'Eua; Fig. 2). The outerparts of the forearc are generally assumed to be relatively cold, andfar removed from arc volcanic activity since the late Eocene.
However, this is not the first report of such intrusives in forearcbasin sediments. A diabase sill of basaltic andesite composition,intruded into mid-Miocene sediments, was cored at Site 793 in theBonin forearc, 75 km east of the presently active arc (ShipboardScientific Party, 1990b). The diabase has baked sediment along itsmargin, clearly showing that it was intruded into the sediments, andtherefore is younger than mid-Miocene. A basalt interval was recov-ered within late Pliocene sediments at Site 781 in the Mariana forearc;
GEOCHEMISTRY AND ORIGIN OF IGNEOUS ROCKS
Table 1 (continued).
Sample:
Laboratory:Method:
135-835B-7R-2, 75-84 cm
QueenslancXRF, AA
Major elements (wt%):SiO2TiO2A12O3Fe2O3
FeOMnOMgOCaONa2OK2OP2O5SumLOI
50.801.04
15.524.155.750.138.19
11.432.590.120.07
99.79t3.00
Trace elements (ppm):ZnNiCrVCuZrSeYSrBaRbNbCsUThPbCoTaHfWGaTILaCePrNdSmEuGdTbDyHoErTmYbLu
9758
108289
85483321.9
12346
1.90.7
———
0.9836———————————————————
[ WHOIXRF
51.770.85
16.379.12*—
0.138.31
12.862.030.130.08
100.82t—
7679
1212467843
—24
13146
3—————
49———————————————————
BUICP
50.620.86
16.209.26*—0.1217.98
12.712.200.080.10
100.13—
10674
20425275383821
12834—
———————
——1.5——
————
———
DurhamICP-MS
—0.89——
0.125—————
—
9376
193276
7233.839.220.6
13228.3
1.90.450.010.130.120.79
39.80.041.020.06
160.041.574.560.754.671.790.772.70.533.280.762.140.311.910.39
135-836A-3H-3, 33^3 cm
QueenslandXRF, AA
55.511.35
14.952.949.450.203.988.453.390.230.13
100.58f1.68
1131110
431109724030
136652.71.6
31
—
WHOXRF
55.291.34
14.8413.23*
0.214.108.632.890.260.14
99.73t—
972110
42211073
34147586
49
—
1 BUICP
54.341.32
14.6013.14*
0.203.998.423.000.240.16
99.42—
82188
399117683632
15054
2.5
—
DurhamICP-MS
—1.35——
0.20—————
—
122154.9
4121117339.236
156544.10.990.140.090.161.23
34.20.082.020.03
180.102.728.321.368.513.091.11
48.70.855.41.243.770.523.360.73
135-836A-3H-4
QueenslandXRF, AA
57.031.21
14.702.788.930.263.658.123.260.280.13
100.35t2.06
112134
38912076.94032.5
134794.42.1
—
—1.4
34——
————
———
——
WHOIXRF
56.671.19
14.6311.35*
—0.263.828.333.010.290.15
99.70t—
96228
36712477
—36
147857
————
46——
————
—
——
———
, 88-100 cm
BUICP
55.601.16
14.4012.46*
—.249
3.647.963.110.280.17
99.02—
114187
354132723634
14865
—
———————
—3——
————
————
DurhamICP-MS
—1.21
———0.254—————
—
107184.2
37712774.738.936.7
15460.34.11.190.120.130.231.36
34.50.122.190.10
180.372.938.751.479.053.421.184.800.885.671.234.000.603.600.78
no contacts were recovered, and the sample could be a sill or a flow(Shipboard Scientific Party, 1990a). Site 781 is about 100 km east ofthe Mariana active arc.
The recovery of mafic intrusions or flows within forearc basinsediments in three different forearcs indicates that magmatic activitymay be far more common in forearcs than has been recognized. Thereare two possible explanations for the origin of these intrusives. First,they may represent melts generated in the mantle immediately be-neath the forearc. This has been suggested for the Site 781 flow(Marlow et al., 1992). Alternatively, these mafic magmas may beinjected into the forearc from the associated active arc along dikes orsills. Two arguments need to be examined in determining whichexplanation is the more likely. The first concerns the chemical com-position of the intrusives, the second the physical constraints on dikeinjection and melt generation.
The Site 781 intrusives, which occur within late Miocene sedi-ments, are strikingly similar to lavas that were erupting from the activearc in the late Miocene. The intrusives, in all regards, are typicalisland-arc lavas. The same is true of sills and flows from Sites 781and 793 (Fig. 7). Each of the mafic units has element abundances andratios typical of "main" arc eruptives and, in particular, are similar to
lavas erupted from the arc active when the units were likely intruded(Miocene for Site 793; late Pliocene-Pleistocene for Site 781).
If the Lau Ridge is restored against the Tonga platform, where itwould have been in the Miocene, it appears that the crest of the activearc had to be over 100 km from the trench. There are Miocene dikeson 'Eua, but no substantial Miocene volcanic section, which suggeststhat 'Eua was near the eastern side of the active arc province. Giventhat geometry, Site 841 must have occupied a position in the Miocenemuch like that in which it presently occurs. If it was less than 60 kmfrom the trench the surface of the subducting slab was at most 40 kmbelow the site (taking a slab geometry similar to that which nowexists). If the Unit 1 intrusives were generated in the mantle, they hadto be made at very shallow depths, presumably at unusually lowtemperatures, between the top of the slab and the base of the crust.This mantle might be expected to be extremely depleted because ofolder episodes of melt generation. Basaltic melts from that thin cornerwedge should carry some distinctive chemical signatures: a greatersilica saturation because of the low pressure of melting, very highvolatile contents because of the high water concentrations needed tomelt the cold mantle, signs of low degrees of melting because of thecool thermal regime, and very depleted incompatible element com-
631
SH. BLOOMER. A. EWART, J.M. HERGT, W.B. BRYAN
Table 1 (continued).
Sample:
Laboratory:Method:
135-834B-13R-1,
QueenslandXRF, AA
Major elements (wt%):SiO 2
TiO 2
A1 2O 3
F e 2 O 3
FeOMnOMgOCaONa 2 OK 2OP2O5SumLO I
49.971.18
16.843.794.910.158.82
11.552.890.130.10
100.33t4.90
Trace elements (ppm):ZnNiCrVCuZrSeYSrBaRb1MbCsUThPbCoTaHfWGaTILaCePT
\TdSmEuGdTbDyHoErTmYbLu
7105
202210
68S13021.9
16437
1.21.2
———
1.8135
———
———
——————
—
WHOIXRF
50.121.13
16.908.88*
0.148.74
12.002.660.140.12
100.02t—
02106189225
7586
2616146
2
———
48
———
——
—
——
—
130-131 cm
BUICP
48.921.14
16.489.02*
0.1338.44
11.322.800.120.12
98.48,—
7490
243237
847S
3723
17537—
———
—
—
3
—
—
—
—
DurhamICP-MS
—1.15
——
0.136—
———
—
7993
241252
8173.736.823.2
17828.1
1.10.960.030.050.101.31
39.10.091.70.03
160.162.638.321.267.682.671.033.450.613.710.82.30.351.970.42
Notes: Values marked with an asterisk (*) represent total iron asFe2θ3; a dash (—) indicates values that were not determined;a dagger (t) represents major element analyses performed ona volatile-free basis. BU = Boston University, WHOI = WoodsHole Oceanographic Institution, ICP = inductively coupledplasma source spectrometry, ICP-MS = ICP mass spectrome-try, AA = atomic absorption, and XRF = X-ray fluorescence.ND = no data.
positions because of the nature of the mantle. These are, in fact, allcharacteristic of the boninitic volcanic rocks that constitute an impor-tant phase of early arc volcanism in the Izu-Bonin-Marianas (Arculuset al., 1992) and that are generated at low pressures (<6 kb; van derLaan et al., 1989). Similar geometric arguments can be made formelting in the mantle beneath Sites 781 and 793.
None of the forearc basin intrusives in the Tonga, Mariana, orBonin forearcs have such a geochemical signature. They are all, infact, exactly like the lavas erupted from their associated active arcs.An interesting contrast is provided by the sill at Site 793 in theIzu-Bonin forearc and the Oligocene basement there. The sill hasTi/Zr =110, Ba/Ce = 8 and K/Ba = 61, all values typical of the activearc. The Oligocene basement, however, has Ti/Zr = 63, Ba/Ce = 4.6,
and K/Ba = 132 to 600. These values are characteristic of the depletedboninitic and tholeiitic lavas that form the Eocene forearc basement,rocks formed by shallow melting of a depleted, hydrous mantle wedge(Murton et al., 1992; Taylor et al., 1992). The Oligocene basement atSite 793, in fact, formed by rifting and magmatism within the forearc,so it is not surprising that basement has the same geochemical sig-nature as the Eocene volcanics (Taylor, 1992). The sill within the Mio-cene sediments, however, has a chemical signature much more likethe Miocene and recent active arc.
The chemical evidence, then, is most consistent with the derivationof the forearc basin dikes and sills in both the Tonga and Mariana-Boninforearcs from the active arc. Do the physical constraints on melt gener-ation and dike injection give the same answer? Melt generation in adepleted mantle wedge, even under water-saturated conditions, willrequire temperatures in excess of 1000°C (van der Laan et al., 1989).The sites of the Miocene and Pliocene or younger dike injections are45 to 100 km from the trench. There was a well-established island arcto the west of the forearc when each of the dikes or flows was formed;the implication is that there existed a "normal," steady-state subductionzone. In such a case, temperatures at 50 km depth, 40-60 km behindthe trench, 20 Ma after the beginning of subduction, should be on theorder of 500°C (Fryer and Fryer, 1985). This requires a substantialthermal anomaly to generate melts in situ in the corner of the mantlewedge. Ridge subduction could provide a mechanism for such ananomaly, but there is no other evidence suggesting the subduction ofa ridge in any of the three forearcs during the appropriate time period.
If, on the other hand, the dikes and flows are derived from theassociated active arc, they have been injected very large distances intothe forearc. Site 841 was likely at least 70 km, and perhaps over 100km, from the Miocene arc. Site 781 is 100 km from the active arc;Site 793 is 75 km from the active arc and probably occupied aboutthe same position relative to the Miocene arc (Taylor, 1992). Theseare not unreasonable distance for large basaltic dikes and sills. ThePalisades sill is 300 m thick and 80 km long, the Whin Sill in England75 m thick and 125 km long, and Icelandic dikes are not uncommonlytens of meters wide and over 100 km long (Philpotts, 1990). TheHigganum dike in New England runs nearly 250 km through thebasement (Philpotts and Martello, 1986). The emplacement of dikesis likely to be facilitated by fracturing of the forearc. All of theseintraoceanic forearcs are in extension (Mrozowski and Hayes, 1980;HussongandUyeda, 1982; Fryer and Fryer, 1985;Taylor, 1992). Boththe Tonga and the Mariana forearcs are cut by both trench-paralleland trench-perpendicular faults and fractures (Fryer and Fryer, 1985;Herzer and Exon, 1985; Bloomer et al., 1988; MacLeod, this volume).These fractures would facilitate the intrusion of low-viscosity basalticmelts outward from the arc.
It is unlikely that the sills, dikes, and flows sampled have beenemplaced all the way from the arc through the forearc basin sediments.Typical width/length ratios for shallow dikes emplaced by hydraulicfracture in consolidated sediments are I0"2 to I0"3 (Fedotov, 1978).The 18-m-thick dike sequence at Site 841 could only have traveled 18km in such a case. Also, injection through wet sediments would haveproduced very high cooling rates and shorter propagation lengths.However, in the deeper crust and in higher density rocks, basaltic dikestypically have length-to-width ratios of I0"3 to lO~4, or for the same18-m-thick unit distances of 18-180 km (Fedotov, 1978). The preex-isting fractures in the forearc would only increase the distance suchdikes could be injected.
We suggest that the dikes and sills found in Site 841, and in similarpositions in the Mariana-Bonin forearc, are derived from their asso-ciated active arc. They are fed by large dikes propagated through theforearc basement; the sampled intervals are intrusive bodies that havebeen fed from that deeper seated large dike. The intersection ofmultiple sets of faults, like those described on the Tonga platform(Herzer and Exon, 1985), could provide pathways for vertical move-ment of melt from a central feeder dike or sill. A very similar geometryhas been described for en-echelon dike sets in New England, where
632
GEOCHEMISTRY AND ORIGIN OF IGNEOUS ROCKS
i
Korobasagaandθsitβs
Korobasagabasalts
-óO
Tofua
Lau basalts
'Eua
10
3
0.3 0.5Y/Zr
o Averages, arcvolcanic units
Hole 841B,Unit 1
Korobasagabasalts
Korobasagaandesites
0.7 50 100Zr (ppm)
Figure 6. Comparison of Unit 1 basalts and other arc units in the Tonga region. The high-Ba Unit 1 samples are all noted as extensively veined. Thereference fields are shown as averages (dots) and standard deviations (bars) for published analyses of the particular groups. Sources as in Figure 3. 'Euavolcanic rocks are Eocene in age, Lau Group basalts are Miocene, Korobasaga Group rocks are Pliocene, and rocks from 'Ata and the Tofua Arc arePleistocene to Recent.
individual 4 to 10 km near-surface dikes are fed by the deeper andregionally extensive Higganum dike (Philpotts and Martello, 1986).This interpretation is mechanically reasonable and is more consistentwith the chemistry of the intrusives than is an origin by in-situ meltingof the mantle wedge beneath the forearc sites.
PRE-LATE EOCENE SILICIC BASEMENT(UNITS 2-6)
Description
At Site 841,209.55 m of volcanic basement (605.05 to 815.6 mbsf)were drilled; these rocks were divided into five units (Fig. 3). Unit 2,at the top of the rhyolitic basement, is 41.9 m of rhyolites and pumicebreccias, which are divided into three subunits. Subunit 2A is a 7.25-m-thick fault zone, developed within rhyolites, that separates the base-ment from the overlying calcareous sandstones. Subunit 2B (19.51 mthick) is a quartz-plagioclase phyric rhyolite flow or dome; it is theleast altered unit within the basement complex (Figs. 8A-8B). Therhyolite averages 14% modal Plagioclase (An48^0), 7% quartz, 0.6%clinopyroxene (augite, En35Wθ40Fs25 to En32Wθ39Fs29), 0.6% orthopy-roxene (hypersthene-ferrohypersthene En5|Wo3Fs46 to En^WfyFs^),trace amounts of hornblende, and 0.5% opaques (magnetite and il-menite), with 8% vesicles in a hydrated, perlitic groundmass. Repre-sentative analyses of the pyroxene and opaque phases are in Table 5.Locally within the unit, there are transitions to a poorly welded rhy-olitic pumice. A similar poorly welded pumice breccia constitutes the22.4-m-thick Subunit 2C.
Unit 3 (25.6 m) includes rhyolitic lapilli tuffs (Subunit 3A), weldedrhyolitic lapilli tuffs (Subunit 3B), laminated crystal to lapilli tuffs
(Subunit 3C), and rhyolite breccias (Subunit 3D). Subunit 3B isbounded by clearly erosive contacts. The tuffs in Subunit 3C haveconvoluted and steeply dipping laminations that are cut by clasticdikes of sand to silt sized tuffaceous sediment. The tuffs coarsen inthe lower 50 cm of the unit, and there is a concentration of large (2.5cm) lithic clasts at the base. The laminated unit could be a surgedeposit or a reworked ash flow. Subunit 3D includes angular frag-ments of rhyolite, some of which are welded, with pumice andsiltstone clasts. The rhyolites have 3%-5% quartz phenocrysts, 1%-5% Plagioclase, and <l % magnetite in a granular matrix of quartz andfeldspar. Symplectic plagioclase-quartz overgrowths on the pheno-crysts are common in the rhyolite clasts.
Unit 4 is a 10.2-m-thick welded rhyolitic lapilli tuff with clasts ofpumice, rhyolite, and basaltic andesite. The flattened pumice clastsare more concentrated toward the top of the unit, whereas the lithicclasts become more abundant, and larger, toward the base. The baseof the unit is sheared and is gradational into Unit 5, an 18.8-m-thicksequence of sheared pumice breccias. Unit 6, at the base of the hole,is at least 39 m thick. Recovery in the unit was less than 3% and it isdifficult to characterize its lithology. Recovered fragments includewelded tuffs, pumices, and mafic clasts. There are also mafic inclu-sions in some of the tuffs.
Alteration is pervasive throughout the basement cores, with theexception of the rhyolites in Subunit 2B. Illite, kaolinite, chlorite,pyrite, calcite, quartz and albite all are variously developed as replace-ments of the groundmass and as overgrowths and alterations of phe-nocrysts (see Schöps and Herzig, this volume). More detailed petro-graphic descriptions of each igneous unit are in Parson, Hawkins,Allan, et al. (1992).
633
S.H. BLOOMER, A. EWART, J.M. HERGT, W.B. BRYAN
Table 2. Analyses of samples from Hole 841B Unit 1.
Core:Interval (cm):Subunit:Laboratory:
18R-1137-140
1ADurham
18R-228-32
1AShip
23R-417-21
IBBU
23R-48-12
IBShip
25R-1105-109
IDShip
25R-421-22
IDDurham
25R-438^6
IDBU
25R-442-46
IDQueensland
26R-288-91
IEShip
26R-281-86
IEBU
26R-284-86
IFShip
Major elements (wt%):SiO2TiO2A12°3Fe2O3*MnOMgOCaONa2OK2OP 2 O 5
TotalLOI
Trace elements (ppm):ZnNiCrVCuZrSeYSrBaRbNbCsUThPbCoTaHfWTIBeLaCePrNdSmHuGdTbDyHoErTmYbLu
——————————
—
——
—
57
25.6241
76.12.510.630.010.160.250.97
0.041.640.06
—3.139.171.478.692.811.013.970.673.940.872.580.42.40.51
54.131.01
16.0412.530.224.559.522.780.160.15
101.09t1.29
1232014
34512271
23216
9021
—
————
—
50.660.98
15.6811.960.244.549.323.230.150.16
96.93**—
1482026
358114643925
24879
—
3
————
—
52.591.06
16.8313.060.274.76
10.262.840.20.17
102.04t1.01
1732217
37611456
24229256
51
————
—
52.591.05
16.6213.130.244.8
10.182.90.090.15
101.74t0.72
1322112
38712874
24230125
11
—
————
—
—
—
64.3
26.1256123.6
1.40.67
BDL0.180.251.67
0.121.680.07
3.249.081.548.422.840.973.950.643.980.912.690.412.480.48
49.661.04
16.8012.510.234.649.543.600.170.18
98.36**—
1282228
380133664026
250124
——
——
—
3.1
———————————
50.25 52.581.21 1.08
17.36 16.4613.43 13.080.26 0.254.98 4.939.69 9.613.26 3.280.17 0.450.15 0.17
100.76t 101.16t3.39 1.55
186 16618 2026 (24) 17
416 382115 11769 7643 (38) —23.5 25
225 227145 310
2 40.77 10.14 —0.180.15 —4.58 —
37 —0.06 —2.90 —0.09 —1.44 —0.18 —4.17(3.4) —
11.7(9.3) —2.05 —
10.8 (7.8) —3.46 (2.6) —1.16 (.99) —4.08 (3.6) —0.72 (0.6) —4.36 —0.93(0.8) —2.63 —0.38 —2.52 (2.6) —0.41 (0.4) —
51.031.05
16.1512.780.234.809.233.100.920.18
99.47**—
1462227
386140.56740.526.5
241294.5
—————————
——3.4—————————————
51.851.03
15.6412.420.24.369.332.940.260.16
100.98t0.87
901812
35113274
—25
213140
42
————————————————————————
Notes: Single asterisk (*) represents total iron as Fe2O3; double asterisk (**) indicates that the sum does not include volatiles; dagger (t) indicates major elementanalyses perfoimed on a volatile-free basis; dashes (—) indicate values that were not determined. Values in parentheses were determined by instrumentalneutron activation from Australian National University, B.W. Chappell, analyst; other analyses as in text. BDL = below detection level.
Subaerial vs. Subaqueous Origin
The environment of eruption of the Site 841 rhyolites has animportant bearing on the tectonic history of the forearc. Clear evi-dence of welding exists in some of the tuffs (Fig. 8C), indicating asubaerial or very shallow submarine origin. No interbedded sedi-ments are present in any of the units, and all of the units are completelydevoid of microfossils. There is a very large volume of pumice withinthe section, which indicates eruption subaerially or in very shallowwater, or which requires very high volatile contents. However, thephenocryst assemblage in the rhyolites is anhydrous, implying lowconcentrations of water or temperatures too high for the crystal-lization of biotite or amphibole. More important, many of the pumicebreccias are poorly sorted; they include abundant low-density pumiceclasts intimately mixed with higher density rhyolite and andesiticclasts. The pumice clasts would normally be separated from the lithicclasts by flotation in a submarine eruption. Only Unit 4 shows evi-dence of density sorting but is still poorly sorted.
The volcanic facies at Site 841 are similar to facies associated withsubaerial eruptions in places like the Taupo volcanic zone (Ewart,1968). Unit 2 is similar to late-stage ash flow deposits associated withplagioclase-quartz rhyolite domes (Ewart, 1968). Poorly consolidatedash-flow deposits (producing pumice breccias), ignimbrites, and pla-gioclase-quartz rhyolites are a common association throughout theTaupo zone (Ewart, 1968). We consider the textures of the volcanicunits at Site 841 to be indicative of subaerial eruption. This is con-sistent with the interpretation that the environment of deposition forthe calcareous sandstones above the rhyolites was in the photic zone(see Chaproniere, this volume). These sandstones contain fragmentsof rhyolitic debris like those found in the basement rocks.
Geochemistry
Most of the rocks recovered from Site 841 are tuffaceous orpumiceous and are extensively altered. The rhyolites of Subunit 2B,and rhyolite clasts within Subunits 3C-3D and Unit 4, provide the
634
GEOCHEMISTRY AND ORIGIN OF IGNEOUS ROCKS
Table 3. Average compositions of Site 841 Unit 1 lavas compared with compositions of arc lavas of different ages fromthe Tonga region.
Location:
Source:
Major elements (wt%):SiO2TiO2
A12O3
Fe 2 O 3 *MnOMgOCaON a ^K2OP 2 θ 5Sum
Trace elements (ppm):ZnNiCrVCuZrSeYSrBaRbNbPbLaCe
Ratios:Ba/ZrTi/ZrY/ZrBa/La
Hole 84IBUnitl
Table 1Average
51.521.05
16.3312.720.244.679.593.090.280.16
99.65
1442020
376124674025
234157
2.7412.413.499.81
2.394
0.3745
'Eua
Cunninghamand
Anscombe(1985)
53.950.77
16.359.660.184.526.153.330.400.06
95.37
758
31280
22—4021
14072—
1.24—
58
'Eua
Hawkins andFalvey(1985)
54.870.71
18.3010.320.215.368.072.580.200.08
100.70
49
181
42
2611537
1
—
0.9101
0.62—
Lau Group
Cole et al.(1985)
55.161.14
18.089.370.173.628.543.300.840.25
100.47
9014
45973035
331138
11.2226
19
1.470
0.3623
KorobasagaGroup
Cole et al.(1985)
55.760.73
18.587.440.154.418.812.991.240.29
100.40
801627
17947662120
533309
3724.39
20
4.766
0.3034
Tonga Arc
Ewart andHawkesworth
(1987)
53.210.51
17.0910.920.185.37
11.451.570.340.07
100.71
842352
305—18—12
196884.40.561.81.53.7
4.9170
0.6759
'Ata
Valuer,Stevenson,
And Scholl,et al. (1985)
52.280.74
17.1310.980.195.48
11.132.030.520.14
100.62
578651
482—46—19
251159
11—
93.58.9
3.596
0.4145
Notes: Dashes (—) indicate values that were not determined; single asterisk (*) indicates total iron as Fe2O3.
6 0 0 -
Sitθ 793 Zr(ppm) 150
Mariana-Bonin:
I I Eocene
Oligocene
Miocene
Active arc
Forearcintrusivθs
Table 4. Isotopic analyses of volcanic rocks from Hole 841B.
Figure 7. Compositions of forearc basin dikes and sills from Site 781 (inPliocene-Pleistocene sediments; Shipboard Scientific party, 1990a) and Site793 (in Miocene sediments; Shipboard Scientific Party, 1990b) compared witharc volcanic units of different ages in the Mariana system. Eocene analyses forforearc lavas from Bloomer (1987) and Reagan and Meijer (1984); Oligoceneanalyses from Reagan and Meijer (1984) and Wood et al. (1980); Mioceneanalyses from Wood et al. (1980); active arc analyses for selected volcanoesfrom Bloomer et al. (1989).
Core, section:Interval (cm):Subunit:
206pb/204pb
207pb/204pb
208pb/204pb
87 S r / 86 S r
143 N d / 144 N d
18R-1137-140
1A
18.70715.53638.2960.703380
0.513096
25R-3118-119
ID
18.716
15.53738.324
0.513085
25R-421-22
ID
18.709
15.54738.3100.7034660.513085
50R-116-18
2B
18.750
15.53938.341
0.513116
Note: Analytical methods detailed in Hergt and Nilsson (this volume).
only materials suitable for analysis. All of the analyzed samples arehigh-silica, low-K rhyolites (Table 6). The freshest samples, theperlitic glasses in Subunit 2B, average 77.6% SiO2 and 5.6% totalalkalis (Table 6), making them rhyolites sensu strictu, by the criteriaof Le B as et al. (1986). We will refer to these rocks as low-K rhyolites,to emphasize their low K20/Na20 (0.12-0.4; Ewart, 1979). Some ofthe rhyolitic clasts show very high SiO2 (>80%; Table 6); this resultsfrom secondary silica developed in post-crystallization graphic inter-growths with feldspar (Shipboard Scientific Party, 1992). All of therhyolites have low contents of Zr, Y, Ba, and REEs, considering theirhigh SiO2 concentrations. All of the analyzed rocks have similar
635
A cm9 -
11 -
1 3 -
15
17
1 9 -
21 - 1
C cm
24-i
2 5 -
26
2 7 -
28 - 1
Figure 8. A. Plagioclase-quartz phyric rhyolite pitchstone from Subunit 2B (Interval 135-841B-49R-1, 9-21 cm). B. Photomicrograph of Subunit 2B rhyolite; field of view is 6 mm from topto bottom (Sample 135-841B-50R-1, 14-17 cm). C. Welded tuff from Unit 6 (Sample 135-841B-66R-CC, 24-28 cm).
GEOCHEMISTRY AND ORIGIN OF IGNEOUS ROCKS
Table 5. Representative analyses of phenocryst phases in Hole841B Subunit 2B.
SiO2
TiO2
A12O3
FeO*MnOMgOCaONa2OZnOSum
Fe2O3
FeOSum
T(°C)
!og f02
Magnetite
0.1414.052.11
76.570.650.810.09
BDL0.29
94.71
38.4941.9498.57
Low-Mgilmenite
BDL47.22
0.1449.56
0.801.470.09BDLBDL
99.28
11.8338.92
100.47
906-11.87
High-Mgilmenite
0.0248.02
0.1748.11
0.632.860.03BDLBDL
99.82
11.8237.48
101.01
907-11.85
Average
clinopyroxene(TV = 1 0 )
51.770.210.76
15.380.68
11.4418.970.25BDL
99.46
15.3899.46
Average
orthopyroxene
(TV = 9)
51.210.130.36
30.711.22
15.771.330.05
BDL100.78
0.7130.07
100.85
Note: Data determined on Section 135-841B-50R-1,5-9 cm. Single asterisk(*) indicates total iron as FeO. Fe•203 and FeO partitioning based onStormer (1983). Temperatures and oxygen fugacities from Buddingtonand Lindsley (1964) and Ghiorso and Carmichael (1981). Analyses byelectron microprobe at the University of Queensland. BDL = belowdetection limit.
chemical compositions, indicating that the units at Site 841 are partof a single magma series. However, the rhyolites of Subunit 2B arethe freshest samples and the only ones that are clearly part of a flowor dome. The subsequent discussion of the chemistry of the basementunits will refer to the average composition of the Subunit 2B rhyolites(Table 7).
The two hypotheses advanced for the origin of the basement atSite 841 (Parson, Hawkins, Allan, et al., 1992) were that it was partof an Eocene sequence correlative to rocks on 'Eua, or that it was afragment derived from rifted Cretaceous crust, like that exposed onthe Lord Howe Rise, that was dispersed by the original breakup ofthe Southwest Pacific (Burns and Andrews, 1973). Analyses of LordHowe Rise basement samples from DSDP Leg 21, Site 207 (Table 6)show that they are high-K2O rhyolites similar to other continental riftrhyolitic sequences (Fig. 9). Although they have Nb, Ta, and Tidepletions like the Site 841 rocks, the Lord Howe Rise rhyolites arehighly enriched in incompatible elements (Fig. 10) and have pro-nounced LREE enrichments (Fig. 11). The Site 841 rhyolites clearlyare not materials derived from a crustal fragment similar to the Creta-ceous rhyolites of the Lord Howe Rise.
The other hypothesis is that the Site 841 rhyolites are part of anEocene complex developed during the earliest subduction of thePacific Plate beneath the Indo-Australian Plate. The rhyolites have a
Table 6. Analyses of rhyolitic volcanic rocks from Hole 841B Units 2-6.
Hole:Core, section:Interval (cm):Subunit:Laboratory:
841B49R-110-13
2BShip
841B47R-2
147-1492BBU
841B50R-1
5-92B
Queensland
841B50R-116-18
2BDurham
841B54R-2
142-1463C
WHOI
841B56R-CC
2-33D
Ship
841B57R-110-13
3DBU
841B59R-214-18
3DBU
84 IB62R-CC
11-154
Ship
207A34R-2
110-113Lord Howe Rise
Queensland
207A35R-3
108-111Lord Howe Rise
Queensland
207A40R-198-101
Lord Howe RiseQueensland
207A42R-243-^6
Lord Howe RiseQueensland
Major elements (wt%)SiO2TO;A12O3FeöOj*MnOMgOCaONa2OKiOP2O5SumLOI
Trace elements (ppm):ZnNiCrVCuZrSeYSrBaRbW ,CsUTilPbCoTaHfWTIBeLaCePrNdSmEuGdTbDy
ErTmYbLu
78.460.25
12.352.370.070.251.834.521.460.05
101.61+5.42
35203
10175
—6249
198142———
—
—
—
20
——————————
75.640.24
11.462.150.060.331.714.551.240.04
97.42**—
43<5<5<521
116150
190—
———
—
—
5—
———
———
——
74.520.37
12.873.290.070.82.934.031.230.04
100.20+4.66
571
19(15)34
1134
16(13)4463
19110.4
1.130.25 (0.40)0.280.393.065
4.340.080.79
4.33 (3.5)13.7(11.2)2.57
13.9(11)4.92 (3.6)1.15(0.96)6.42 (5.5)1.24(0.8)7.901.75(1.6)5.200.805.35 (5.4)0.88 (0.79)
78.630.16
12.312.060.090.211.594.310.820
100.00+—
—
—
161—
70179
111.20.370.250.424.94
0.124.760.14
4.2312.952.15
12.614.961.127.51.129.011.985.851.016.231.12
76.310.34
12.433.720.031.811.323.850.680.06
100.19+—
77
3325
148—5.372
7
——
13
—
—
—
———
—
80.50.28
11.562.480.020.951.803.000.380.04
101.01 +2.59
4440
159
184—6476
11682
——
19
—
—
—
—
80.840.24
11.440.920.011.081.433.250.220.05
99.45**—
207
14310534786—
4—
—
77.250.26
11.792.210.010.511.653.450.670.00
97.79**—
3314
22145
94461
104
—
5—
—
—
75.940.23
10.834.350.060.744.973.181.390.06
101.75+5.17
0109
21137—50
1378942
——
20
—
—
—
—
75.970.18
13.522.380.0120.790.302.754.090.02
100.01 +9.08
131
22
3341
5232
19319830.410.73.65
16.125
31.968.31.649.925.1
47.6101
12.344.910.20.48
10.31.579.01.845.020.754.790.77
76.890.20
12.892.010.0140.270.463.234.430.03
100.42+5.24
165
4——
3682
7663
54718837.510.23.96
14.527.9
21.816.631.79
10.75.3
7514219.670.315.41.39
14.82.22
12.52.546.981.056.91.03
77.500.17
1.870.0240.340.55
4.300.02
100.33+9.18
128—53—
25533
2811.33.66
14.520.4
—1.767.451.74
17.85.2
55.3115
14.251.911.60.54
11.61.759.791.97
0.805.170.79
75.770.26
13.152.580 010,161.093.144.140.03
100.36+5.34
174—53—
4926
1451260
17430.210.33.68
13.342.0
—1.63
1.6912.46.5
59.5124
15.255.312.42.23
12.21.81
10.02.045.530.825.250.80
Notes: Single asterisk (*) indicates total iron as Fe2θ3; double asterisk (**) indicates sum that does not include volatiles; dagger (+) indicates that major element analyses were performed on a volatile-free basis;dashes (197 indicate values that were not determined; values in parentheses by instrumental neutron activation from Australian National University, B.W. Chappell, analyst; other analyses as in text. BU = BostonUniversity, WHOI = Woods Hole Oceanographic Institution. BDL = below detection level.
637
S.H. BLOOMER, A. EWART, J.M. HERGT, W.B. BRYAN
Table 7. Average composition of Hole 841B Subunit 2B rhyolites compared with other western Pacific dacites andrhyolites.
Location:
Source:
Major elements (wt%):SiO2
TiO2
A12O3
Fe2O3*MnOMgOCaONa2OK2OP 2 O 5
SumLOI
Trace elements (ppm):ZnNiCrVCuZrSeYSrBaRbNbCsUThPbCoTaHfLaCePrNdStnEuGdTbDyHoErTinYbLu
Site 841Subunit 2B
AverageTable 5
77.580.22
12.042.190.070.261.714.461.170.03
99.685.42
39364
16173126156
19012
1.40.310.270.414.0040.104.554.28
13.322.36
13.264.941.146.961.188.461.875.530.915.791.00
Lord Howe Rise
AverageTable 5
76.530.20
13.042.210.030.390.603.034.240.03
100.207.21
150
43
3733
6168
5381923210.63.74
14.628.82.51.797.81
59.4120
15.355.612.4
1.1612.2
1.8410.32.105.710.865.530.85
SD
0.700.030.350.280.010.240.300.180.130.01
1.92
20
10
7429
46444
11.93.60.40.1318.10.50.120.84
10152.79.31.90.71.60.241.30.270.750.120.810.11
Saipan
BoniniticrhyoliteMeijer(1983)
78.200.17
10.581.470.040.221.283.541.690.12
97.253.00
14
125
9217515
——
0.851.6
5.509.96
7.712.480.543.31
3.42
2.13
2.40.43
Site 786
61R-4,96-102 cm
Pearce etal. (1992)
70.880.32
13.384.630.070.563.524.841.720.08
100.002.93
5748
3857811215
12317528
1.390.360.430.832.35
0.122.94.53
10.91.427.341.880.632.430.372.500.551.420.241.430.23
Site 786
67R-1,3 3 ^ 1 cmPearce etal. (1992)
76.030.23
12.512.600.030.401.194.252.710.05
100.000.69
1053
112238771412
10243024
0.85——0.522.46
0.062.223.066.160.814.721.350.511.420.191.530.290.920.181.160.13
Marianaforearc
Boniniticdacite
Bloomer(1983)
67.450.19
13.205.570.093.553.973.781.370.06
98.851.4
—249885—
76—7
427—————————7.02—3.961.000.300.98—0.86—0.48———
Marianaforearc
Tholeiiticdacite
Bloomer(1983)
67.540.56
13.487.140.130.814.003.341.200.16
98.362.24
—8
2021
—99
—34
1183110—————————
11.76—8.152.620.813.98—4.52—3.18—3.62—
Notes: Single asterisk (*) indicates total iron as Fe2O3; dashes (—) indicate values that were not determined. SD = standard deviation.
clear subduction signature, with enrichments in LILEs, low compat-ible-element concentrations, depletions in Nb and Ta, and pronouncedPb enrichments (Fig. 10). They have low abundances of REEs as wellas LREE depletions (Fig. 11). Temperatures calculated from coexist-ing magnetite and ilmenite in the rhyolite give temperatures of 906°Cand oxygen fugacitiesof 1 0 " 82 (using the curves of Buddington andLindsley [1964] with the recalculation procedure of Stormer [1983]and the fitting procedure of Ghiorso and Carmichael [ 1981 ]; Table 5).These data, as argued for similar Eocene rhyolites from Saipan, aremost consistent with derivation of the rhyolites by fractional crystal-lization of a more mafic parent (Meijer, 1983). Rhyolites are not theproduct of hydrous melting of the mantle (Green, 1973) and thecompatible element concentrations in the rhyolites are far too low forthem to be mantle-derived melts. Melts of the downgoing slab shouldproduce dacites (Ringwood, 1974) with pronounced heavy-rare-earth(HREE) depletions, because of equilibration with garnet eclogite, andLILE enrichments.
Melting of the altered lower oceanic crust beneath the Eoceneforearc under hydrous conditions could produce rhyolites at 680°-720°C (Helz, 1976). These melts, if the crust were in amphibolite facies
conditions, should have higher A12O3 (Helz, 1976), more pronouncedEu depletions, greater LREE enrichments, and possibly higher HREEelement "tails" than the Site 841 rhyolites. The crystallization tempera-tures for the rhyolites are also much higher than those for anatectitesfrom a basaltic amphibolite (680°-720°C; Helz, 1976; Meijer, 1983).
However, Beard and Lofgren (1991) have recently shown thatdehydration melting, rather than water-saturated melting, of green-stones at low pressures does produce melts very similar to low-K2Oarc rhyolites at 850°-900°C. The restite from this melting lacks am-phibole and includes Plagioclase (>50 modal%), orthopyroxene andclinopyroxene. The experimental rhyolites with 78% SiO2 have 12%A12O3, 2% CaO, 0.5% TiO2, and 0.5% MgO, values very similar tothose in the Site 841 Subunit 2B rhyolites. The Eu anomalies in theSubunit 2B rhyolites are unusually small if they are indeed derivedfrom partial melts that were in equilibrium with a plagioclase-richresidue. Also, production of the rhyolites by partial melting of thecrust might be expected to produce a bimodal rock suite. However,the few analyses of arc clasts contained within the Site 841 rhyolites(see below) show a diverse range of SiO2 from 50% to 70%, suggest-ing that the basement is not bimodal. However, the evidence for the
638
GEOCHEMISTRY AND ORIGIN OF IGNEOUS ROCKS
Site 841 RhyoliteLord Howe Rise Rhyolites
CDSDP Site 207)O Mt Somers Rhyolites
WhitsundayRhyolite - dacite field
Ab OrFigure 9. Q-Or-Ab normative projection for rhyolite compositions from the Southwest Pacific. LordHowe Rise rhyolites (this paper), Whitsunday rhyolites in Queensland Province, Australia (Ewart etal., in press) and Mt. Somers rhyolites are all examples of Cretaceous rhyolitic fields associated withthe fragmentation of the Southwest Pacific continental margins. Note the distinct compositions of theSite 841 rhyolite. Curve I is the water saturated curve at 100 MPa for compositions of An3, after Jamesand Hamilton (1969); curves 2 and 3 are the calculated 4-phase surfaces in the system Ab-An-Or-Q-H2O for aw = 1.0 and 0.2, respectively, at 200 MPa (after Nekvasil, 1988).
104
103
GOccO 1Q2
Q. 10
(O
1
0.1
a Lord Howe Rise rhyolite
O Hole 786B, boninitic rhyolitβ
• Hole 841B Unit 2
I I I I I I I I I I I I I I I I I I T I I I I I I I I ICs Ba Th Ta La Pb Sr Sm Hf Ti Tb Y Er Yb
Rb U Nb K CΘ Pr Nd Zr Eu Gd Dy Ho Tm Lu
Figure 10. N-MORB-normalized compositions of Site 841 Subuit2B analyses,Lord Howe Rise rhyolites, and Site 786 boninitic dacite (Murton et al., 1992).Normalizing values from Sun and McDonough (1989).
basement composition is fragmentary and the possibility that the Site841 rhyolites represent dehydration melting of lower arc crust cannotbe ruled out.
The Site 841 rhyolites are, therefore, most reasonably interpretedas derivatives of a more mafic magma, requiring a substantial volumeof mafic cumulates at depth, or as dehydration melts of lower arc crust,leaving a substantial granulite facies mafic residue. In either case, the
parent material must have been very depleted in incompatible ele-ments, given the low incompatible element abundances in the rhy-olites. There are two mafic-silicic associations described from Eocenesequences in the Izu-Bonin-Mariana and Tonga forearcs. Boniniticseries lavas have been described from a number of places in the IBMforearc (e.g., Hussong and Uyeda, 1982; Bloomer and Hawkins, 1987;Meijer, 1980; Reagan and Meijer, 1984; Ishii, 1985; Arculus et al.,1992; Pearce et al., 1992). These boninitic associations include daciticand rhyolitic lavas (as on Saipan, first described by Schmidt [1957])that are interpreted to have developed by fractional crystallization of aboninitic magma (Table 7; Meijer, 1983; Bloomer, 1987; Murton et al.,1992). Associated with these boninitic series, in time and space, aredepleted island-arc-tholeiitic series lavas. These also include dacitic torhyolitic fractionates (e.g., Bloomer, 1987; Murton et al, 1992) thatare not as depleted in incompatible elements as their boninite seriescounterparts. Similar tholeiitic series basalts, andesites, dacites, andrhyolites comprise the Eocene basement on 'Eua (Ewart and Bryan,1972; Hawkins and Falvey, 1985; Cunningham and Anscombe, 1985).
The rhyolites from Site 841 Subunit 2B show a closer similarityto tholeiite series lavas than to the boninite series. They are not asdepleted in HREEs and HFSEs as are high-silica rocks derived fromboninitic parents, such as those on Saipan and in Site 786 (Figs. 10and 12). They are, however, very similar in element abundances andratios to tholeiite series dacites from 'Eua (Fig. 12). In fact, they havetrace element abundances nearly identical to the 73% SiO2 volcanicson 'Eua (Table 7 and Fig. 12).
Stern et al. (1991) noticed that boninitic-series lavas in the Mari-ana forearc were characterized by high K/Ba and low Ba/La, in
SH. BLOOMER, A. EWART, J.M. HERGT, W.B. BRYAN
100
T3co
-CO
Q.ECO
CO 10
Site 841 Rhyolite
La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
Figure 11. Chondrite-normalized, rare-earth-element diagrams for Site 841 rhyolite, Lord Howe Rise rhyolites, andrhyolites from the East Queensland Cretaceous Province (striped field, from Ewart et al., in press).
100
CD
DC 10 +
D Saipan boninitic rhyolitβ
Q Mariana forearctholθiitic dacitβ
• Hole 841B, Unit 2
0-1 <—\—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—h-Rb U Nb K2O Ce Nd Zr TiO2 Dy Er Lu
Ba Th Ta La Sr Sm Eu Gd Yb
CO
uu
10
1
0 1
-π\ /r\\ fi
3- H 1 1 1
g
\ ÅD
O
•
I 1 1 1 1 1 1 1 H
'Eua, 62% SiO2
'Eua, 73% SiO2
Hole 841B, Unit 2,78% SiO2
—I 1 1 1 1 H-Ba Ta La Pb Nd Hf TiO2 Dy Yb
Th K2O Ce Sr Sm Eu Gd Y Lu
Figure 12. Hole 84IB rhyolite compared to other siliceous Eocene forearclavas. Normalizing values from Sun and McDonough (1989). Saipan analysisfrom Meijer (1983), Mariana forearc dacite from Bloomer (1987), and 'Euaanalyses from Cunningham and Anscombe (1985).
400-
* _
100"
mm
H
N-MORB•
Mariana forearc/boninites
HSite 786 HCB
"@Sitθ786lCB
TEua
Site 786 L C B W
Site 841
o• Ata
rhyolitθ11
•Tofua
10 Ba/La 60
Figure 13. Trace element ratios for Site 841 lavas compared with boninite andlava on an island-arc-tholeiitic series. Average compositions for Ata and TofuaArc from Ewart and Hawkesworth (1987); average and standard deviation forMariana boninites from Stern et al. (1991), averages and standard deviationsfor 'Eua from Cunningham and Anscombe (1985), averages for Site 786boninites from Murton et al. (1992; LCB = low-Ca boninite, ICB = intermedi-ate-Ca boninites, and HCB = high-Ca boninite); N-MORB average from Sunand McDonough (1989).
contrast to the normal arc signature of high Ba/La and low K/Ba (Fig.13). They suggested that this signature was unique to the boniniticeruptives that characterized the early phases of subduction. Indeed,the boninitic series at Site 786 show a low Ba/La signature, althoughK/Ba varies between high-Ca and low-Ca boninites (Fig. 13). TheSite 841 rhyolites have high Ba/La and low K/Ba, suggesting anaffinity to an arc-tholeiitic, not boninitic, magma stem.
Volcanic Clast Compositions
The extreme degree of fractionation of rhyolitic rocks makes itdifficult to be specific about the nature of the parental lavas in theEocene forearc. Furthermore, no information is available about thebasement on which the silicic edifice was built. However, a varietyof mafic clasts are present within the section at Site 841 that mayprovide a clue about the nature of both the parental lava for therhyolites and about the basement beneath them.
We have examined three populations of clasts to search for infor-mation about the Eocene history of the forearc (Table 8). The first
GEOCHEMISTRY AND ORIGIN OF IGNEOUS ROCKS
Table 8. Analyses of clasts from Hole 841B basement and sediments.
Core, section:Interval (cm):Sample wt (g):Type:
48R-129-310.0059
Arc
Major elements (wt%):SiO2TiO2A12O3Fe2O3*MnOMgOCaONa2OK2OP2O5Sum
Normalizedfrom:
50.810.62
15.7413.130.32
10.282.565.401.13—
lOO.OOt
90.10
Trace elements (ppm):ZnNiCrVCuZrSeYSrBa
489—————
—27—
48R-129-310.0059
Arc
51.441.13
13.4117.500.437.743.673.371.31—
lOO.OOt
94.90
433——
296——52—
152—
51R-324-280.0008
Arc
54.320.81
20.507.410.082.788.644.860.611.43
lOO.OOt
117.19
——————————
51R-324-280.0012Arc?
53.681.11
18.988.830.102.319.903.851.23—
lOO.OOt
104.93
———
————
52R-166-690.2427
Arc
75.680.38
11.253.900.0270.902.184.990.330.09
99.64
109
102142
147—
24210372
Inclusions from rhyolitic basement complex
52R-1126-1290.2514ORB
49.141.51
10.8611.590.267
11.9711.32
1.910.100.20
98.67
95137218251
1603743
12223
52R-1126-1290.0553ORB
46.461.48
16.5015.660.224
12.025.921.460.150.11
99.87
29292
233273
764429
16458
62R-153-570.1051
Arc
82.610.548.514.320.0331.002.842.081.540.046
103.47
77
1587
4845
119
63R-191-940.0439ORB
48.872.67
17.4916.820.0985.743.723.252.180.28
100.84
542
132467197
155
63R-191-940.2204ORB
46.912.21
17.6114.630.0975.762.573.391.990.42
95.167
890
—295
1274797
10696
63R-1108-1100.2663ORB?c
54.211.55
19.059.790.0813.394.395.091.070.35
98.62
394<IO<IO177<151043170
16794
65R-CC25-270.0009
Ar
70.961.08
10.838.940.0493.151.582.780.64
lOO.OOt
109.70
—
————
65R-CC25-270.0023
Arc
62.10.77
17.678.750.0575.290.883.690.78
lOO.OOt
105.83
———
—————
65R-CC25-270.2019
Arc
84.770.339.192.820.0210.570.633.060.430.028
101.82
471712<l17
1258
438451
67R-CC0-5
0.24477
61.941.45
15.589.390.052.140.527.100.120.16
98.44
7441273215389411513642
68R-CC12-150.2529
Arc
53.401.05
15.4810.940.117.144.473.040.160.09
95.88
7815263863052422210450
Table 8 (continued).
Core, section:Interval (cm):Sample wt (g)Type:
Inclusions fromEocene sediments
45R-138^0
: 0.0045Arc?
Major elements (wt%):SiO2TiO2
A12O3Fe2O3*MnOMgOCaONa2OK2OP 2 O 5
Sum
Normalizedfrom:
51.501.07
14.3113.060.1364.885.823.321.53—
95.63
Trace elements (ppm):ZnNiCrVCuZrSeYSrBa
252——
322——
53—
208278
46R-CC
0.0012Arc
48.280.87
13.7115.420.1227.375.09———
90.86
— —— —— —— —— —— —— —— —— —— —
Inclusions from Oli;
41R-CC15-200.0047
Arc
50.460.88
16.1311.450.2747.286.463.201.11—
97.24
263———272———207318
sediments
42R-235^00.0025
Arc
65.600.99
16.996.750.530.714.272.461.830.66
lOO.OOt
109.69
———————
——
jocene
43R-1116-1190.0057Arc?
54.311.14
14.3911.640.1775.329.182.821.020.46
lOO.OOt
105.00
232
355695
65
189269
22R-281-850.0056
7
49.231.04
18.3910.910.2236.499.764.030.19
100.26
180229180
148258
Inclusions from Miocene volcaniclastics with reworked Eocene fauna
22R-281-850.0022
9
51.241.10
15.6214.870.2416.774.175.000.990.67
lOO.OOt
107.50
—
22R-281-850.0115ORB?
49.221.81
16.5214.220.1633.83
10.423.130.770.09
lOO.OOt
105.12
8218339
392
40
264261
22R-281-850.1257
Arc
64.000.48
16.484.590.0521.292.677.370.420.07
97.35
3177
8419381415
93
22R-281-85
0.0022Arc
57.250.75
14.4211.640.134.094.804.182.738.01
lOO.OOt
108.70
—
———
—
——
22R-281-850.0829
Arc
64.660.58
12.0810.660.133.383.182.223.120.19
lOO.OOt
94.64
901835
1872735223879
697
22R-2125-1290.2482
Arc
50.400.74
15.2012.000.2529.311.715.820.220.25
95.65
190115413181304532349841
22R-2125-1290.2451ORB?
45.601.29
18.6414.370.22
15.452.102.090.230.15
lOO.OOt
91.95
99153470323
59045216342
22R-2125-1290.1033
Arc
54.320.71
18.7111.370.163.192.608.700.250.06
lOO.OOt
93.15
772013
206434230115663
group comprises clasts contained within the rhyolites. These includeaphyric greenstones, mafic grains and pebbles, and silicic pebbles.Similar clasts occur within the Oligocene and Eocene sediments imme-diately overlying the rhyolites. A third group of clasts comes from aMiocene breccia/conglomerate sequence in which there is a reworkedEocene fauna, indicating that Eocene basement was locally exposed
during sedimentation. Many of these clasts are extensively altered,and their mobile element concentrations are suspect; also many weretoo small to allow analysis of trace elements.
The compositions of the clasts are shown in Figure 14, comparedto fields for their most likely sources. The field for 'Eua representsbasaltic to rhyolitic Eocene arc compositions, whereas fields for E-
641
S.H. BLOOMER, A. EWART, J.M. HERGT, W.B. BRYAN
Table 8 (continued)
Core, section:Interval (cm):Sample wt (g):Type:
22R-2125-1290.0234ORB?
Major elements (wt%):SiO2
TiO2
A12O3
Fe2O3*MnOMgOCaONa2OK2OP 2 O 5Sum
Normalizedfrom:
48.351.03
14.6115.670.318.255.904.110.700.25
98.93
Trace elements (ppm):ZnNiCrVCuZrSeYSrBa
150110360165
99404781
187
Inclusions from Miocene Volcaniclastics with reworked Eocene fauna
23R-112-140.2149
Arc
51.380.62
17.0216.150.296.471.925.410.740.03
lOO.OOt
92.06
153347
251
49241552
133
23R-112-140.0058
Arc
53.160.56
20.789.550.153.194.157.421.04
—lOO.OOt
94.56
152
—
74400
23R-2113-1150.0056
7
53.781.03
14.9012.320.193.716.364.580.840.21
97.71
171
331
49179380
24R-CC13-150.2604ORB?
50.131.13
16.5112.930.295.16
10.273.480.110.26
lOO.OOt
90.75
1222931
436109804330
25787
24R-CC13-150.1141Arc?
54.271.01
14.8512.760.264.128.383.870.200.07
99.72
972428
229
633724
213112
29R-152-540.0366ORB?
52.541.40
15.9915.140.223.773.645.911.380.09
100.001
94.23
168132
51713623530
11798
29R-152-540.0147
9
50.881.17
14.6815.550.245.844.626.000.650.42
99.63
15612021
238——72—5388
29R-171-730.1367
Arc
53.240.74
13.7715.880.236.953.924.091.160.16
lOO.OOt
91.47
1955271
30315252439
2232930
29R-171-730.0079ORB
48.812.19
15.3115.630.209.394.553.820.041.00
lOO.OOt
108.00
161147181228——30—21—
Notes: Single asterisk (*) represents total iron as Fe2O3; dashes (M) indicate values that were not determined. Some smallsamples yielded sums between 90% and 120% because of weighing errors or because of very high volatile contentscaused by alteration. These analyses were normalized to 100% (noted by t); the original sum is noted below eachanalysis. Volumes of small samples were insufficient to allow complete trace element analysis. Type designations arebased on empirical comparison to the fields in Figure 13 and discriminant analysis using Ti, V, CaO, Fe2O3, MgO,SiO2, and Sr. ORB = ocean ridge basalts, Arc = compositions similar to Eocene volcanics on 'Eua, and ? = clasts thatcould not be assigned a definitive origin.
MORB and N-MORB from the subducting Pacific Plate east of theTonga Trench (Bloomer and Fisher, 1987) represent reasonable com-positions for the basement on which the Eocene arc might have beenbuilt. The origins assigned to the clasts in Table 8 are based oncomparison of the clast compositions to these fields (empirically andby discriminant analysis).
Two groups of clast types can be identified (Table 8 and Fig. 14);some of the clasts cannot be confidently assigned to either group. Theprincipal population includes basalts, andesites, and dacites with lowTiO2 and low Ti/V that are clearly of arc origin (Table 8). None of thispopulation has the low Ti, Ca, and Al signatures typical of boniniticseries lavas, supporting the conclusion that the Site 84 ̂ volcaniccomplex is part of an arc-tholeiitic magma series.
The second population is small, but appears to comprise clasts ofocean ridge basalt composition. These clasts have TiO2 greater than1.4% (some greater than 2%), high Ni and Cr concentrations, and highTi/V. Some of the highest Ti samples may represent enriched MORBor intraplate compositions, which do exist in the offshore plate (Bloomerand Fisher, 1987). These ocean-ridge-like clasts occur in both thereworked Miocene sediments and within the rhyolitic basement. Thelatter group is very significant, as it means that mid-ocean-ridge crustexisted beneath the rhyolitic arc edifice that built the Site 841 basement.
Discussion
The rhyolitic complex found at Site 841 was one of the mostsurprising discoveries of Leg 135. There are a number of similaritiesbetween this arc sequence and Eocene arc lavas that have been drilledin the forearcs of the IBM systems. The Site 841 rhyolites appear tobe mid-Eocene in age, on the basis of the age of the overlying
sediments and preliminary K-Ar ages. The IBM basement is all partof a middle to late Eocene construct (Taylor, 1992). The rhyoliticmagma at Site 841 was derived by fractional crystallization of a moremafic, arc-tholeiitic lava, very similar to volcanic rocks in the Eocenebasement on 'Eua or by dehydration melting of similar Eocene vol-canic rocks in the arc basement. The IBM forearc is a composite oflow-Ca to high-Ca boninites and depleted arc-tholeiite-series lavas(NatlandandTarney, 1981;Bloomer, 1983,1987; Arculusetal., 1992;Murton et al., 1992; Taylor, 1992). Boninitic lavas have not beenfound with any of the igneous rocks on 'Eua or at Site 841; they havebeen dredged from the northern termination of the Tonga arc, butthe measured age of those boninites is much younger than Eocene.Boninitic rocks have not been found in most of the Tonga forearc,though the volume of dredge and drill sampling in most of the forearcis still so sparse that their existence cannot be ruled out.
The basement at Site 841 has subsided over 5000 m since the lateEocene. This remarkably large subsidence is a consequence both ofthermal subsidence and of subduction erosion. Substantial subsidencehas been documented for forearc sites in the IBM system (Lagabrielleet al., 1992; Taylor, 1992). The boninitic volcano drilled at Site 786in the Bonin forearc is also inferred to have been above sea level inthe Eocene (Lagabrielle et al., 1992).
The results from Site 841 also support hypotheses for forearcevolution proposed in the IBM system. There the initiation of sub-duction was accompanied by the development of a very wide (up to300 km) and extensive (up to 3000 km long) zone of arc volcanismin the middle to late Eocene (Natland and Tarney, 1981; Taylor, 1992;Stern and Bloomer, 1992). This style of volcanism is required by thedistribution of middle to late Eocene arc volcanics in the forearcs andfrontal arcs. This Eocene volcanism developed in an extensional en-
GEOCHEMISTRY AND ORIGIN OF IGNEOUS ROCKS
0.8 -
CO
OCM
1ü
0.4 -
LÜ
i
1
1TTTrmEED
•
IE
]
2 4Fθ2θ3* / MgO
400
Q.
a.
200
E-MORB offshore
Eua
N-MORB offshore
S1
1.0TiO2
T2.0
100 -
Figure 14. Compositions of clasts in Site 841 sections. Solid circles = clasts within the rhyolitic basement (Units 2-6); open circles = clastsfrom Eocene-Oligocene calcareous sandstones; and dotted squares = clasts within Miocene sediments that contain reworked Eocene fauna.Fields for Eocene arc lavas on 'Eua from Cunningham and Anscombe (1985); fields for N-MORB and E-MORB, which form the Pacific Platein the Tonga Trench, from Bloomer and Fisher (1987).
vironment, and in places may have occurred as true seafloor spreading(Stern and Bloomer, 1992).
A similar model can be applied to the Tonga forearc. Site 841 onthe trench-slope break, and the basement on 'Eua at the edge of theTonga platform, bracket a zone at least 65 km wide, on both sides ofwhich identical lavas were erupted at 46-40 Ma. Arc lavas and plu-tonics are the principal materials dredged from the landward trenchslope from the Louisville Seamount Chain intersection all the way tothe north end of the trench (Fig. 1), a distance of over 1000 km. Theoccurrence of ocean-ridge basalt clasts within the rhyolites at Site 841is significant, because it requires that the older oceanic crust on whichthe Eocene arc was building had been rifted, but that crust was notcompletely replaced by juvenile arc volcanism. The age of the base-ment in the forearc is only well constrained at Site 841 and on 'Eua.Additional sampling is essential if we are to confirm that the Tongaforearc is indeed constructed in the same manner as the IBM forearcsystems. However, the available evidence is completely consistentwith the model for forearc evolution that has been proposed by Sternand Bloomer (1992) (Fig. 15).
This model supposes that subduction zones begin by sinking ofold, unstable lithosphere, perhaps initiated at a fracture zone. Theinitial "subduction" is entirely a vertical movement (Fig. 15). Aeon-sequence of that sinking is a strong trench "suction" that extends theupper plate and begins to rift apart the preexisting oceanic crust.
Dehydration of the sinking slab initiates melting in the extendinglithosphere, augmented by an influx of asthenosphere driven by thesinking lithosphere. This melting generates boninitic and arc-tholei-itic sequences. The dominant magma product depends upon the initialthermal structure of the mantle and the degree of depletion of thepreexisting lithosphere. Highly depleted lithosphere produces bonin-itic-series melts; less depleted, perhaps deeper lithosphere and theinflowing asthenosphere produce arc-tholeiitic melts. These arc erup-tives build atop the rifted oceanic crust. Zones of particularly highmagma supply may yield subaerial volcanic edifices like those foundat Site 841 and on Saipan (Meijer, 1980). If the degree of extensionis sufficiently high, the arc volcanism may proceed in a seafloorspreading mode (Stern and Bloomer, 1992). That stage was appar-ently not reached beneath Site 841.
The results from Site 841 provide strong support for the modelsof forearc evolution that have been proposed for the IBM system. Theearliest phases of subduction in the Tonga region appear to have beenaccompanied by broadly distributed island arc-tholeiitic volcanism.
SUMMARY
The basement at Site 841 comprises over 200 m of rhyolites,pumice breccias, and tuffs that are part of a middle to late Eocenesubaerial island-arc volcano. The rhyolites were derived by fractional
643
S.H. BLOOMER, A. EWART. J.M. HERGT. W.B. BRYAN
kmVerticalexaggeration = 1 o 50 100
see detail below.
•*tt> ExtensionT TrenchM Mantle flow
I U Old ocean crust
E23 Old lithosphere
| Infant arc crust
[ 3 Zone of intensemelting
: : • • : : •
Figure 15. Model for the initial phases of subduction (after Stern and Bloomer,1992). Subduction is initiated by sinking of old lithosphere, producing exten-sion in the overlying oceanic plate and melting in the mantle wedge (lithosphereand inflowing asthenosphere) above the sinking plate. Arc lavas erupt on, andcover, extended ocean crust. Sites of large magma supply may develop silicic,subaerial edifices.
crystallization of a more mafic arc-tholeiitic parent (implying a sig-nificant volume of mafic cumulates at depth) or by dehydrationmelting of lower arc crust (implying a substantial granulite faciesmafic residue at depth). The Site 841 rhyolites are nearly identical tohigh-SiO2 lavas in the Eocene basement on 'Eua and are not the sameas the Cretaceous rhyolites on the Lord Howe Rise, and are not partof a pre-Eocene crustal fragment.
The age and composition of the Site 841 rhyolites, and their posi-tion 65 km outboard of same-aged rocks on 'Eua supports models forforearc development that require an early episode of broadly distrib-uted arc volcanism. Clasts of ocean-ridge basalt within the rhyolitessuggests that the Eocene volcano was built on rifted oceanic crust.
Unlike the Mariana forearc, no evidence was found at Site 841 forextensive exposures of serpentine in the forearc. Such diapiric bodiesmay not exist here, or they may be too far away to have shed debrisinto the sedimentary section at Site 841.
Basaltic dikes and sills intrude upper lower Miocene and mid-Miocene sediments at Site 841. These basalts appear to be part of oneintrusive episode; they are identical to upper Miocene lavas eruptedon the Lau Ridge and show no geochemical signature to suggest thatthey were generated by anomalous melting in the sub-forearc mantle.We consider it most likely that these intrusions were fed from base-ment dikes propagating out from the Miocene active arc. A similarorigin is suggested for forearc basin intrusions in the Mariana andBonin forearcs.
ACKNOWLEDGMENTS
We thank the technicians, crew, and drillers aboard Leg 135 fortheir enthusiasm and skill during the drilling at Site 841. Our ideasabout forearc evolution have benefited from discussions with BobStern, Julian Pearce, Brian Taylor, Kristen Nilsson, Jim Hawkins,Patty Fryer, and Tony Crawford. Shawn Dolan and Jennifer Chaseprovided invaluable assistance with the analyses of the Site 841samples. We would also like to thank members of the Western PacificPanel; David Scholl, Tracy Vallier, and Andy Stevenson of the U.S.Geological Survey; and Jim Hawkins for their efforts in including Site841 in the Leg 135 prospectus. Careful reviews by Robert Stern,Richard Arculus, and James Hawkins improved the final manuscript.This work was supported by a grant from JOI-USSAC.
REFERENCES
Acland, A.S., Vysotsky, S.V., Pearce, J.A., and Parson, L.M., 1992. Geochem-istry of a boninitic crustal section from the northern Tonga Trench [pre-sented at Back-Arc Development and Crustal Accretion in the WesternPacific, Cambridge, England, May 1990]. (Abstract)
Arculus, R.J., Pearce, J.A., Murton, B.J., and van der Laan, S.R., 1992. Igneousstratigraphy and major-element geochemistry of Holes 786A and 786B. InFryer, P., Pearce, J.A., Stokking, L.B., et al., Proc. ODP, Sci. Results, 125:College Station, TX (Ocean Drilling Program), 143-169.
Beard, J.S., and Lofgren, G.G., 1991. Dehydration melting and water-saturatedmelting of basaltic and andesitic greenstones and amphibolites at 1, 3 and6.9 kb. J. Petrol., 32:365-401.
Bentz, F.P., 1974. Marine geology of the southern Lord Howe Rise, southwestPacific. In Burk, CA., and Drake, C.L. (Eds.), The Geology of ContinentalMargins: New York (Springer-Verlag), 537-547.
Billington, S., 1980. The morphology and tectonics of the subducted litho-sphere in the Tonga-Kermadec-Fiji region from seismicity and focalmechanism solutions [Ph.D. dissert.]. Cornell Univ., Ithaca, NY.
Bloomer, S.H., 1983. Distribution and origin of igneous rocks from thelandward slopes of the Mariana Trench: implications for its structure andevolution. J. Geophys. Res., 88:7411-7428.
, 1987. Geochemical characteristics of boninite- and tholeiite-seriesvolcanic rocks from the Mariana forearc and the role of an incompatibleelement-enriched fluid in arc petrogenesis. Spec. Pap.—Geol. Soc. Am.,215:151-164.
Bloomer, S.H., and Fisher, R.L., 1987. Petrology and geochemistry of igneousrocks from the Tonga Trench—a nonaccreting plate boundary. J. Geol.,95:469-495.
Bloomer, S.H., and Hawkins, J.W., 1987. Petrology and geochemistry ofboninite series volcanic rocks from the Mariana trench. Contrib. Mineral.Petrol., 97:361-377.
Bloomer, S.H., Parsons, L., Huggett, Q., and Pearce, J., 1988. GLORIAsurveys of the Tonga and Kermadec Trenches. Eos, 69:1442.
Bloomer, S.H., Stern, R.J., Fisk, E., and Geschwind, CH., 1989. Shoshoniticvolcanism in the northern Mariana Arc, 1. Mineralogy and major and traceelement characteristics. /. Geophys. Res., 94:4469-4496.
Buddington, A.F., and Lindsley, D.H., 1964. Iron-titanium oxides minerals andsynthetic equivalents. /. Petrol., 5:310-357.
Burns, R.E., and Andrews, J.E., 1973. Regional aspects of deep sea drilling inthe southwest Pacific. In Burns, R.E., Andrews, J.E., et al., Init. Repts.DSDP, 21: Washington (U.S. Govt. Printing Office), 897-906.
Cole, J.W., Gill, J.B., and Woodhall, D., 1985. Petrologic history of the LauRidge, Fiji. In Scholl, D.W., and Vallier, T.L. (Eds.), Geology and OffshoreResources of Pacific Island Arcs—Tonga Region. Circum-Pac. Counc.Energy Miner. Resour., Earth Sci. Ser., 2:379-414.
Cole, J.W., Graham, I.J., and Gibson, I.L., 1990. Magmatic evolution of LateCenozoic volcanic rocks of the Lau Ridge, Fiji. Contrib. Mineral. Petrol.,104:540-554.
Cunningham, J.K., and Anscombe, K.J., 1985. Geology of 'Eua and otherislands, Kingdom of Tonga. In Scholl, D.W, and Vallier, T.L. (Eds.),Geology and Offshore Resources of Pacific Island Arcs—Tonga Region.Circum-Pac. Counc. Energy Miner. Resour., Earth Sci. Sen, 2:221-258.
Abbreviations for names of organizations and publication titles in ODP reference listsfollow the style given in Chemical Abstracts Service Source Index (published byAmerican Chemical Society).
644
GEOCHEMISTRY AND ORIGIN OF IGNEOUS ROCKS
Duncan, R.A., Valuer, T., and Falvey, D.A., 1985. Volcanic episodes at 'EuaTonga. In Scholl, D.W., and Valuer, T.L. (Eds.), Geology and OffshoreResources of Pacific Island Arcs—Tonga Region. Circum-Pac. Counc.Energy Miner. Resour., Earth Sci. Ser., 2:281-290.
Ewart, A., 1968. The petrography of the central north island rhyolitic lavas,part 2. Regional petrography including notes on associated ash-flowpumice deposits. N.Z. J. Geol. Geophys., 11:478-545.
, 1979. A review of the mineralogy and chemistry of Tertiary-Recentdacitic, latitic, rhyolitic and related salic volcanic rocks. In Barker, F. (Ed.),Trondhjemites, Dacites, and Related Rocks: Amsterdam (Elsevier), 13-121.
Ewart, A., Brothers, R.N., and Mateen, A., 1977. An outline of the geology andgeochemistry, and the possible petrogenetic evolution of the volcanic rocksof the Tonga-Kermadec-New Zealand island arc. J. Volcanol. Geotherm.Res., 2:205-250.
Ewart, A., and Bryan, W., 1972. Petrography and geochemistry of the igneousrocks from 'Eua, Tonga Islands. Geol. Soc. Am. Bull., 83:3281-3298.
Ewart, A., and Hawkesworth, C.J., 1987. The Pleistocene-Recent Tonga-Ker-madec arc lavas: interpretation of new isotope and rare earth data in termsof a depleted mantle source model. J. Petrol., 28:495-530.
Ewart, A., Schon, R.W., and Chappell, B.W., in press. The Cretaceous vol-canic-plutonic province of the central Queensland (Australia) coast—a riftrelated "calc-alkaline" province. Trans. R. Soc. Edinburgh: Earth Sci.
Falloon, T.J., Green, D.H., and Crawford, A.J., 1987. Dredged igneous rocksfrom the northern termination of the Tofua magmatic arc, Tonga andadjacent Lau Basin. Aust. J. Earth Sci., 34:487-506.
Fedotov, S.A., 1978. Ascent of basic magmas in the crust and the mechanismof basaltic fissure eruptions. Int. Geol. Rev., 20:33-47.
Fisher, R.L., and Engel, CG., 1969. Ultramafic and basaltic rocks dredgedfrom the nearshore flank of the Tonga Trench. Geol. Soc. Am. Bull.,80:1373-1378.
Fryer, P., Ambos, E.L., and Hussong, D.M., 1985. Origin and emplacement ofMariana forearc seamounts. Geology, 13:774-777.
Fryer, P., and Fryer, G.J., 1987. Origins of nonvolcanic seamounts in a forearcenvironment. In Keating, B.H., Fryer, P., Batiza, R., and Boehlert, G.W.(Eds.), Seamounts, Islands, and Atolls. Am. Geophys. Union, Monogr.Ser., 43:61-72.
Ghiorso, M.S., and Carmichael, I.S.E., 1981. A Fortran IV computer programfor evaluating temperatures and oxygen fugacities from the compositionsof coexisting iron-titanium oxides. Comput. Geosci., 7:123-129.
Gnibidenko, H.S., Anosov, G.I., Argentov, V.V., and Pushchin, I.K., 1985.Tectonics of the Tonga-Kermadec Trench and Ozbourn Seamount junctionarea. Tectonophysics, 112:357-383.
Green, D.H., 1973. Experimental melting studies on a model upper mantlecomposition at high pressure under water-saturated and water-undersatu-rated conditions. Earth Planet. Sci. Lett., 19:37-53.
Greene, H.G., and Wong, FL., 1983. Hydrocarbon resource studies of thesouthwestern Pacific, 1982. Open File Rep.—U.S. Geol. Surv., 82-293.
Haggerty, J.A., 1987. Petrology and geochemistry of Neogene sedimentaryrocks from Mariana forearc seamounts: implications for emplacement ofthe seamounts. In Keating, B., Fryer, P., Batiza, R., and Boehlert, G.W.(Eds.), Seamounts, Islands, and Atolls. Am. Geophys. Union., Geophys.Monogr. Ser., 43:175-185.
Hawkins, J.W., and Falvey, D.A., 1985. Petrology of andesitic dikes and flowsfrom 'Eua, Tonga. In Scholl, D.W., and Valuer, T.L. (Eds.), Geology andOffshore Resources of Pacific Island Arcs—Tonga Region. Circum-Pac.Counc. Energy Miner. Resour., Earth Sci. Ser., 2:269-280.
Helz, R.T., 1976. Phase relations of basalts in their melting ranges at pH->O =5kb. Part II. Melt compositions. J. Petrol., 17:139-193.
Herzer, R.H., and Exon, N.F., 1985. Structure and basin analysis of thesouthern Tonga forearc. In Scholl, D.W., and Vallier, T.L. (Eds.), Geologyand Offshore Resources of Pacific Island Arcs—Tonga Region. Circum-Pac. Counc. Energy Miner. Resour., Earth Sci. Ser., 2:55-73.
Hussong, D.M., and Uyeda, S., 1982. Tectonic processes and the history of theMariana Arc: a synthesis of the results of Deep Sea Drilling Project Leg60. In Hussong, D.M., Uyeda, S., et al., Init. Repts. DSDP, 60: Washington(U.S. Govt. Printing Office), 909-929.
Ishii, T, 1985. Dredged samples from the Ogasawara fore-arc seamount or"Ogasawara paleoland"-"fore-arc ophiolite." In Nasu, N., Kobayashi, K.,Kushiro, I., Kagami, H. (Eds.), Formation of Active Ocean Margins:Tokyo (Terra Sci. Publ.), 307-342.
James, R.S., and Hamilton, D.L., 1969. Phase relations in the system NaAlSi3O8-CaAl->Og-SiO^ at 1 kilobar water vapour pressure. Contrib. Mineral. Petrol.,21:111-141.
Johnson, L.E., Fryer, P., Taylor, B., Silk, M., Jones, D.L., Sliter, W.V., Itaya,T., and Ishii, T, 1991. New evidence for crustal accretion in the outerMariana forearc: Cretaceous radiolarian cherts and MORB-like lavas.Geology, 19:811-814.
Karig, D.E., and Ranken, D., 1983. Marine geology of the forearc region,southern Mariana island arc. In Hayes, D.E. (Ed.), The Tectonic andGeologic Evolution of Southeast Asian Seas and Islands (Pt. 2). Am.Geophys. Union, Monogr. Ser., 27:266-280.
Lagabrielle, Y., Sizun, J.-P., and Arculus, R.J., 1992. The constructional anddeformational history of the igneous basement penetrated at Site 786. InFryer, P., Pearce, J.A., Stokking, L.B., et al, Proc. ODP, Sci. Results, 125:College Station, TX (Ocean Drilling Program), 263-276.
Le Bas, MJ., Le Maitre, R.W., Streckeisen, A., and Zanettin, B., 1986. Achemical classification of volcanic rocks based on the total alkali-silicadiagram. J. Petrol., 27:745-750.
Lonsdale, P., 1986. Amultibeam reconnaissance of the Tonga Trench axis and itsintersection with Louisville Guyot Chain. Mar. Geophys. Res., 8:295-327.
Malahoff, A., Feden, R., II, and Fleming, H.S., 1982. Magnetic anomalies andtectonic fabric of marginal basins north of New Zealand. J. Geophys. Res.,87:4109-4125.
Marlow, M.S., Johnson, L.E., Pearce, J.A., Fryer, PB., Pickthorn, L.G., andMurton, B.J., 1992. Pleistocene volcanic rocks in the Mariana forearcrevealed by drilling at Site 781. In Fryer, P., Pearce, J.A., Stokking, L.B.,et al, Proc. ODP, Sci. Results, 125: College Station, TX (Ocean DrillingProgram), 293-310.
Meijer, A., 1980. Primitive island arc volcanism and a boninite series: exam-ples from the western Pacific island arc system. In Hayes, D.E. (Ed.),Tectonic and Geologic Evolution of Southwest Asian Seas and Islands.Am. Geophys. Union, Geophys. Monogr. Ser., 23:269-282.
, 1983. The origin of low-K rhyolites from the Mariana frontal arc.Contrib. Mineral. Petrol., 83:45-51.
Mottl, MJ., Haggerty, J.A., and Scientific Party of Ocean Drilling ProgramLeg 125, 1989. Upwelling of Cl-poor, S-, and C-rich waters through aserpentinite seamount, Mariana Forearc: ODP Leg 125. Eos, 70:1382.
Mrozowski, C.L., and Hayes, D.E., 1980. A seismic reflection study of faultingin the Mariana fore arc. In Hayes, D.E. (Ed.), The Tectonic and GeologicEvolution of Southeast Asian Seas and Islands. Am. Geophys. Union,Geophys. Monogr. Ser., 23:223-234.
Murton, B.J., Peate, D.W., Arculus, R.J., Pearce, J.A., and van der Laan, S., 1992.Trace-element geochemistry of volcanic rocks from Site 786: the Izu-Boninforearc. In Fryer, P., Pearce, J.A., Stokking, L.B., et al., Proc. ODP, Sci.Results, 125: College Station, TX (Ocean Drilling Program), 211-235.
Natland, J.H., and Tarney, J., 1982. Petrologic evolution of the Mariana arcand back-arc basin system—a synthesis of drilling results in the SouthPhilippine Sea. In Hussong, D.M., Uyeda, S., et al., Init. Repts. DSDP, 60:Washington (U.S. Govt. Printing Office), 877-908.
Nekvasil, H., 1988. Calculated effect of anorthite component on the crystal-lization paths of H7O-undersaturated haplogranitic melts. Am. Mineral.,73:966-982.
Packham, G.H., 1985. Vertical tectonics on the Tonga Ridge from the Ton-gatapu oil exploration wells. In Scholl, D.W., and Vallier, T.L. (Eds.),Geology and Offshore Resources of Pacific Island Arcs—Tonga Region.Circum-Pac. Counc. Energy Miner. Resour., Earth Sci. Ser., 2:291-300.
Parson, L., Hawkins, J., Allan, J., et al., 1992. Proc. ODP, Init. Repts., 135:College Station, TX (Ocean Drilling Program).
Pearce, J.A., Lippard, SJ., and Roberts, S., 1984. Characteristics and tectonicsignificance of supra-subduction zone ophiolites. In Kokelaar, B.P., andHowells, M.F. (Eds.), Marginal Basin Geology. Geol. Soc. Spec. Publ.London, 16:74-94.
Pearce, J.A., Thirlwall, M.F, Ingram, G., Murton, B.J., Arculus, R.J., and vander Laan, S.R., 1992. Isotopic evidence for the origin of boninites andrelated rocks drilled in the Izu-Bonin (Ogasawara) forearc, Leg 125. InFryer, P., Pearce, J.A., Stokking, L.B., et al., Proc. ODP, Sci. Results, 125:College Station, TX (Ocean Drilling Program), 237-261.
Pelletier, B., and Louat, R., 1989. Seismotectonics and present day relativeplate motions in the Tonga-Lau and Kermadec-Havre region. Tectono-physics, 165:237-250.
Philpotts, A.R., 1990. Principles of Igneous and Metamorphic Petrology:Englewood Cliffs, NJ (Prentice Hall).
Philpotts, A.R., and Martello, A., 1986. Diabase feeder dikes for the Mesozoicbasalts in southern New England. Am. J. Sci., 286:105-126.
Raitt, R.W., Fisher. R.L., and Mason, R.G., 1955. Tonga Trench. In Poldervaart,A. (Ed.), Crust of the Earth. Spec. Pap.—Geol. Soc. Am., 62:237-254.
S.H. BLOOMER, A. EWART, J.M. HERGT, W.B. BRYAN
Reagan. M.K., and Meijer, A., 1984. Geology and geochemistry of early arcvolcanic rocks from Guam. Geol. Soc. Am. Bull., 95:701-713.
Ringwood, A.E.. 1974. The petrological evolution of island arc systems. J.Geol. Soc. London, 130:183-204.
Robertson, A.H.F., 1990. Tectonic evolution of Cyprus. In Malpas, J., Moores,E.M., Panayiotou, A., and Xenophontos, C. (Eds.), Ophiolites: OceanicCrustal Analogues, Proc. Symp. "TROODOS 1987. " Cyprus Geol. Surv.Dep., 235-249.
Schmidt, R.G., 1957. Petrology of the volcanic rocks. In Cloud, P.E., et al.,Geology of Saipan, Mariana Islands (Pt. 2, Chapt. B). Geol. Surv. Prof.Pap. U.S., 280:127-172.
Sharaskin, A.Y., Pustchin, I.K., Zlobin, S.K., and Kolesov, G.M., 1983. Twoophiolitic sequences from the basement of the northern Tonga Arc. Ofioliti,8:411-430.
Shipboard Scientific Party, 1990a. Site 781. In Fryer, P., Pearce, J.A., Stokking,L.B., et al., Proc. ODP, Init. Repts., 125: College Station, TX (OceanDrilling Program), 179-195.
, 1990b. Site 793. //; Taylor, B., Fujioka, K., et al., Proc. ODP, Init.Repts., 126: College Station, TX (Ocean Drilling Program), 315-403.
. 1992. Site 841. In Parson, L., Hawkins, J., Allan, J., et al., Proc.ODP, Init. Repts., 135: College Station, TX (Ocean Drilling Program),571-677.
Stern, R.J.. and Bloomer, S.H., 1992. Subduction zone infancy: examples fromthe Eocene Izu-Bonin-Mariana and Jurassic California Arcs. Geol. Soc.Am. Bull, 104:1621-1636.
Stern, R.J., Morris, J., Bloomer, S.H., and Hawkins, J.W., 1991. The source ofthe subduction component in convergent margin magmas: trace elementand radiogenic isotope evidence from Eocene boninites, Mariana forearc.Geochim. Cosmochim. Acta, 55:1467-1481.
Starmer, J.C., Jr., 1983. The effects of recalculation on estimates of tempera-ture and oxygen fugacity from analyses of multicomponent iron-titaniumoxides. Am. Mineral., 68:586-594.
Sun, S.-S., and McDonough, W.F.. 1989. Chemical and isotopic systematicsof oceanic basalts: implications for mantle composition and processes. //;Saunders, A.D., and Norry, M.J. (Eds.). Magmatism in the Ocean Basins.Geol. Soc. Spec. Publ. London, 42:313-345.
Taylor, B., 1992. Rifting and the volcanic-tectonic evolution of the Izu-Bonin-Mariana Arc. In Taylor, B., Fujioka, K., et al., Proc. ODP, Sci. Results,126: College Station, TX (Ocean Drilling Program), 627-651.
Taylor, R.N., Lapierre, H., Vidal, P., Nesbitt, R.W., and Croudace, I.W., 1992.Igneous geochemistry and petrogenesis of the Izu-Bonin forearc basin. //;Taylor, B., Fujioka, K., et al., Proc. ODP, Sci. Results, 126: College Station,TX (Ocean Drilling Program), 405-430.
Tsunakawa, H., 1983. K-Ar dating on volcanic rocks in the Bonin Islands andits tectonic implication. Tectonophysics, 95:221-232.
Umino, S., 1985. Volcanic geology of Chichijima, the Bonin Islands (OgasawaraIslands). Chishitsugaku Zasshi [J. Geol. Soc. Jpn.], 91:505-523.
Valuer, T.L., O'Connor, R.M., Scholl, D.W., Stevenson, A.J., and Quinterno,PJ., 1985. Petrology of rocks dredged from the landward slope of theTonga Trench: implications for middle Miocene volcanism and subsidenceof the Tonga Ridge. //; Scholl, D., and Valuer, T.L. (Eds.), Geology andOffshore Resources of Pacific Island Arcs—Tonga Region. Circum-Pac.Counc. Energy Miner. Resour., Earth Sci. Ser., 2:109-120.
Valuer, T.L., Stevenson, A.J., and Scholl, D.W., 1985. Petrology of igneousrocks from 'Ata Island, Kingdom of Tonga. In Scholl, D.W., and Vallier,T.L. (Eds.), Geology and Offshore Resources of Pacific Island Arcs—Tonga Region. Circum-Pac. Counc. Energy Miner. Resour., Earth Sci. Sen,2:301-316.
van der Laan, S.R., Flower, M.F.J., and Koster van Groos, A.F., 1989. Experi-mental evidence for the origin of boninites: near-liquidus phase relationsto 7.5 kbar. In Crawford, A.J. (Ed.), Boninites and Related Rocks: London(Unwin Hyman), 112-147.
van der Lingen, G., 1973. The Lord Howe Rise rhyolites. In Burns, R.E.,Andrews, J.E., et al., Init. Repts. DSDP, 21: Washington (U.S. Govt.Printing Office), 523-540.
Wood, D. A., Mattey, D.P., Joron, J.L., Marsh, N.G., Tarney, J., and Treuil, M.,1981. A geochemical study of 17 selected samples from basement coresrecovered at Sites 447, 448, 449, 450, and 451, Deep Sea Drilling ProjectLeg 59. In Kroenke, L., Scott, R., et al., Init. Repts. DSDP, 59: Washington(U.S. Govt. Printing Office), 743-752.
Date of initial receipt: 22 June 1992Date of acceptance: 7 June 1993Ms 135SR-129
646