Journal of Petrology Pages 1641-1668 © 1998 Oxford University Press

Magma Genesis in the New Britain Island Arc: Further Insights into Melting and Mass Transfer Processes
Introduction
   Geological setting and samples studied
Analytical Methods
   Trace elements
   Isotopic analyses
Results
Mantle Domains
Prior Depletions
   Models for HFSE depletion
   Relationships to back-arc spreading
   Niobium and tantalum
   Summary
Palaeo-enrichments
Fluids And Melts: The 'Subduction Signature' In Depleted Volcanic Front Suites
   Subduction components
   Europium and cerium anomalies and redox conditions
   Subducted sediments vs slab basalt involvement
Across-arc Trends-The Geochemical Architecture Of A Subduction Zone
Conclusions
Acknowledgements
References

Magma Genesis in the New Britain Island Arc: Further Insights into Melting and Mass Transfer Processes

J. D. WOODHEAD1*, S. M. EGGINS2 AND R. W. JOHNSON3

1SCHOOL OF EARTH SCIENCES, THE UNIVERSITY OF MELBOURNE, PARKVILLE, VIC. 3052, AUSTRALIA
2DEPARTMENT OF GEOLOGY, AUSTRALIAN NATIONAL UNIVERSITY, CANBERRA, A.C.T. 0200, AUSTRALIA
3AUSTRALIAN GEOLOGICAL SURVEY ORGANISATION, GPO BOX 378, CANBERRA, A.C.T. 2601, AUSTRALIA

RECEIVED JULY 22, 1997; REVISED TYPESCRIPT ACCEPTED MARCH 20, 1998

Quaternary volcanic rocks from the New Britain island arc display a wide range in chemical compositions. The source of the lavas shares isotopic characteristics with Indian Ocean type mid-ocean ridge basalt (MORB). In contrast, the high field strength elements (HFSE) are extremely depleted in the volcanic front rocks compared with MORB. We propose that this results from a previous melt-extraction event-hypotheses invoking residual phases in either the mantle wedge or subducting slab cannot account for the depletion relative to MORB. In addition, elements other than the HFSE are also affected. Chemical signatures in parts of the New Britain arc and Manus Basin may relate to a previous subduction episode along the now inactive Vitiaz-West Melanesian trench. Ultradepleted volcanic front basalts invariably have strong 'fluid'-related trace element signatures, including high Sr/Nd and U/Th (and 238U disequilibrium), together with positive Eu anomalies that can be related to the mobility of Eu2+ in the slab-derived flux. Negative Ce anomalies are attributed to a minor sedimentary component. Across-arc geochemical profiles record a decrease in the degree of partial melting and diminishing influence of a slab-derived fluid with depth, superimposed upon the depleted mantle composition beneath the volcanic front. Element partitioning into (and not necessarily the source of) the fluid is considered to exert strong control on the chemistry of volcanic front magmas, a feature that may go some way to explaining the contradictory estimates of the slab flux derived from isotope vs trace element data in many subduction suites.

Keywords: arcs; depletion; europium; fluid; subduction

INTRODUCTION

The New Britain region of Papua New Guinea represents an outstanding opportunity to reach an understanding of the processes of magma genesis in an oceanic island arc. The Quaternary volcanoes found there define the eastern part of the Bismarck volcanic arc, and have formed in response to northward subduction of the small Solomon plate beneath the Bismarck plate (Fig. 1). The New Britain arc is outstanding for two main reasons:

(1) Arc volcanism in the central sector of the island has taken place over an exceptionally wide range of depths to the Wadati-Benioff zone-from ~100 km deep at the 'volcanic front' (closest to the submarine trench) down to ~600 km in the northwest beneath the Witu Islands. The reasons for the correspondingly large width of the volcanic arc are still unknown, but isotopic and elemental differences between the rocks are systematic as depths to the Wadati-Benioff zone increase. We believe that these differences provide unparalleled insights into the geochemical architecture of subduction-zone systems.

(2) Rocks of the New Britain volcanic front are low in potassium, range from basalt to rhyolite, and have (as illustrated below) exceptional depletions in high field strength elements (HFSE). Indeed, they may represent the most HFSE-depleted arc rocks known. These rocks therefore should provide valuable constraints on the controls of HFSE abundances in island-arc magmas. The depletions also mean that the nature of any slab-derived components should be faithfully recorded in critically selected trace-element ratios.


Figure 1. Volcanic zones E to Hn of the New Britain island arc shown in relation to the schematic plate boundaries of part of the Papua New Guinea region. Numbers enclosed by circles represent sample locations on the subducting Solomon plate and in the vicinity of the Manus spreading centre. Dashed line indicates 2000 m water depth. PP, Pacific plate.


Rocks of the New Britain volcanic front were cited by Jakes & Gill, (1970) as typical of those of the so-called Island Arc Tholeiite Series. Magma genesis in New Britain has been considered in several previous studies (e.g. Lowder & Carmichael, 1970; Blake & Ewart, 1974; Johnson, 1977; DePaolo & Johnson, 1979; Basaltic Volcanism Study Project, 1981). Woodhead & Johnson, (1993) more recently provided a detailed assessment of isotopic and trace-element profiles across the arc, and proposed a model for the across-arc differences. Additional U-Th and Be-B constraints were provided by Gill et al., (1993).

The study reported here represents an extension of the work undertaken by Woodhead & Johnson, (1993). We report for the first time ICP-MS (inductively coupled plasma mass spectrometry) trace element and Sr/Nd/Pb-isotope analyses of rocks from the full New Britain and Manus Basin arc-trench system, thus complementing and extending substantially the previously reported data sets:

(1) Samples of sediments and basalts from the subducting Solomon plate were analysed to identify possible chemical components from the downgoing slab.

(2) Samples from the Manus back-arc basin were analysed also, to determine the possible end-member composition of mantle-wedge material that may have been unaffected by a slab contribution (normal MORB has been dredged from the basin).

(3) A suite of samples from the Sulu Range volcanic cluster in the New Britain volcanic front was analysed using the same methods as in (1) and (2). Sulu Range was chosen because its least fractionated basalts were known from X-ray fluorescence analyses (Johnson & Chappell, 1979) to be exceptionally depleted in HFSE.

We use these new data to consider further the origin of the New Britain magmas in relation to the results of other recent work on island arc magmas. Several contemporary issues are addressed, but discussion is concentrated on elemental depletion in the mantle wedge (particularly in HFSE) and the nature of the 'subduction signature' as seen most clearly in the depleted volcanic front rocks. We conclude with some commentary on more general aspects of arc geochemistry, including the across-arc differences in rock composition.

Geological setting and samples studied

New Britain lies at the margin of the Bismarck and Solomon plates (Fig. 1). The morphology, geology and structure of the western Solomon Sea were investigated by the R.V. Natsushima in 1983 and 1984. The age of the oceanic basement is constrained by magnetic lineation (Joshima et al., 1987) and heat flow (Joshima & Honza, 1987) data to between 28 and 34 Ma, and estimates of the rate of convergence with the overriding Bismarck plate are in the range 9-12 cm/yr (Basaltic Volcanism Study Project, 1981). Samples from the Solomon Sea were collected during the Natsushima cruise using dredges and a free-fall grab only, and hence it is uncertain to what extent the recovered material is representative of that present at depth. The only large collection of volcanic rocks was recovered from Station 29, a scarp of unknown origin in the centre of the basin. Most of the basalts and sediments used in this study come from this site, with additional sediments from Site 33, slightly to the west. Chemically the Station 29 basalts are ferrobasalts comparable with evolved MORB from the East Pacific Rise (Davies & Price, 1987). Sediments are predominantly poorly indurated mudrocks which are unlikely to be older than Pliocene, but which Crook, (1987) considered to be representative of the hemipelagic deposits covering the oceanic basement.

The Manus Basin is a fast-opening back-arc basin behind the New Britain volcanic front (Taylor, 1979). It straddles part of the Pacific-Bismarck plate boundary where plate motion takes place along three major left-lateral transform faults and a range of different extensional segments (Martinez & Taylor, 1996). Two linked spreading axes in the basin are (1) an extensional transform zone in the west and (2) the Manus spreading centre (MSC) in the east. The MSC is a wedge of new seafloor crust that is about 72 km wide i