Issuu

Geological Society of Australia

ABSTRACTS Number

SGGMP - Port Macquarie 2005

Specialist Group in Geochemistry, Mineralogy and Petrology First biennial conference Port Macquarie, New South Wales -

July, 2005

Geological Society of Australia ABSTRACTS Number 76

SGGMP - Port Macquarie 2005

Editor: Ian T Graham

Specialist Group in Geochemistry, Mineralogy and Petrology First biennial conference Port Macquarie, New South Wales - is'** July, 2005

Produced By: Ian Graham Geoscience The Australian Museum 6 College Street, Sydney NSW 2010, Australia

June-July 2005

ISSN 0729 01IX

Purchase requests should be addressed to: Business Manager Geological Society of Australia Suite 706, 301 George Street Sydney, NSW 2000 © Geological Society of Austraha Specialist Group in Geochemistry, Mineralogy and Petrology (SGGMP)

TABLE OF CONTENTS Page INVITED SPEAKERS Jon Davidson Progress and problems in understanding arc magmatism.

Peter A. Williams Secondary minerals: scientific curiosities or geochemical treasure troves?

OTHER SPEAKERS Michael J. Baker, Anthony J. Crawford and Ron F. Berry Geochemistry of mafic igneous rocks of the central Georgetown Inlier, North Queensland.

Jane B. Barron, Lawrence M. Barron, Terry P. Memagh and Greg Duncan Decompression exsolution microstructures in gamet from Bingara, northeast NSW: first evidence for incompletely exhumed ultrahigh-pressure slab material (UHP terrane) at depth.

Lawrence M. Barron, Jane B. Barron and Terrence P. Memagh Relation between remnant pressure in / and strain birefringence around inclusions in diamond.

Vickie C. Bennett and A.P. Nutman Evidence for evolving mantle regimes from 143-neodymium and 176-hafriium isotopic records.

Ralph S. Bottrill From chromite to crocoite: chromium mineralogy and geochemistry at Dundas, Western Tasmania.

Paul Carr, Brian Jones and Bruce Selleck The father of mineralogy, his son, and frozen prawns - the cool mineral connection.

Stephanie A. Carroll and Nathan R. Daczko Plutonism and metamorphism at the root of a Cretaceous magmatic arc.

Benjamin E. Cohen, Kurt M. Knesel and Paulo M. Vasconcelos HFSE fractionations in a continental, intraplate volcanic system. Southeast Queensland, Australia.

Anthony Crawford, Leonid Danyushevsky, Roman Leslie, Sofia Tetroeva, Trevor Falloon and Alicia Verbeeten. The Hunter Ridge, SW Pacific: hot subduction slab edge effects and the boninite-adakite- high-Nb basalt association.

Nathan R. Daczko and Geoffrey L. Clarke High-pressure mafic migmatites, Fiordland, New Zealand: does migmatisation promote recrystallisation to garnet granulite?

Joel Fitzherbert and Geoff Clarke 58 A sequence of corona formation in two-pyroxene gabbro, Fiordland New Zealand: intrusion, rapid post-magmatic cooling and transformation of gabbro to garnet granulite. Rhiannon George, S. Turner, R. Price, C. Cook and B. Finney Tracking crustal differentiation and assimilation processes at arc volcanoes: a Uranium series isotope perspective.

Daniel Halas, Ian Nicholls and Roland Maas Geochemistry of Mesozoic rocks of Victoria: melting of source mantle remote from sites of Gondwana break-up.

Janet Hergt, Jon Woodhead and Stephen Anderson Determination of U-Th and REE in apatite for (U-Th-Sm)/He dating.

Zara Heyworth, Rhiannon M. George, Bruce F. Schaefer and Simon P. Turner Insights into magma generation and evolution at White Island, New Zealand.

Terrence P. Memagh and Frank P. Bierlein Geochemical modeling of mineralization scenarios for turbidite-hosted gold deposits in the Lachlan Fold Belt.

Luke A. Milan, Nathan R. Daczko, Ian Tumbull and Andrew Allibone Thermobarometry of an Earl Cretaceous high-pressure contact metamorphic aureole near Resolution Island, Fiordland, New Zealand.

Sunne G. Nielsen, Mark Rehkamper, Marc Norman and Alex Halliday Thallium isotopic evidence for ferromanganese sediments in the mantle source of Hawaiian basalts.

Marc D. Norman, J.M. Rhodes and M.O. Garcia Compositions of Hawaiian basalts preclude eclogite mantle plumes.

David Och, Evan Leitch and Graziella Caprarelli The geology of the Port Macquarie - Tacking Point coastal tract.

Robin Offler Metamorphism in the southern New England Fold Belt - an overview.

100

Robin Offler, Horst Zwingmann, Lin Sutherland and Ian Graham 105 Significance of orientation, age and geochemical composition of dykes emplaced before and during the opening of the Tasman Sea. Nick Petford, David Wertheim, Dougal Jerram and Jon Davidson Slurry flow and structures formation in a magma mush: the Basement Sill, McMurdo Dry Valleys, Antarctica.

108

Alice C. Pliophs and Kelsie A. Dadd Facies analysis, geochemistry and tectonic setting of the Frampton Volcanics, Southeastern New South Wales.

111

Elizabeth Price Limestone petrography of the Riversleigh World Heritage Area, Northwestern Queensland.

118

Flavian C. Schroter, Geoff L. Clarke, R.W. White and N.J. Pearson 122 Systematics in two phase REE and Y partitioning coefficients in mafic granulites. Alanna L. Simpson, Kurt M. Knesel and Michael A. Dungan Cumulate recycling in arc magmas.

125

Ian E. Smith and Richard C. Price The significance of silicic melts in subduction related volcanic suites.

130

Lin Sutherland, Ian Graham, Gayle Webb, Ross Pogson, Gaston Giuliani and Anthony Fallick New ruby-sapphire sources, Yarrowitch basaltic field, eastem NSW.

133

Jean-Yves Talbot and Nathan R. Daczko Tectonic significance of low-grade mineralization of seafloor spreading-related faults, Macquarie Island.

137

Simon Turner Time scales of magmatic processes: a review of recent U-series results.

143

Ron H. Vernon Where do high-level S-type granite magmas come from?

144

Emmanuel W. Wembenyui, K.D. Collerson and J.X. Zhao Geochemical and isotopic signatures of Mount Cameroon lavas (West Africa).

145

Helen M. Williams, A.N. Halliday, C.A. McCammon, A.H. Peslier, N. Teusch, 149 S. Levasseur and J.-P. Burg Iron isotopes as a potential new tool in igneous geochemistry and cosmochemistry. Megan L. WilUams and Paul F. Carr Origin of carbonates and prehnite in the Prospect Intrusion, NSW.

153

Jon Woodhead, Janet Hergt and Roger Kemp Spatial resolution and the analysis of complex geometries in LA-MC-ICPMS.

158

Progress and problems in understanding arc magmatism

Jon Davidson (Invited Speaker) Department of Earth Sciences, University of Durham, Durham, DH13LE, UK i .p.Davidson@durham.ac.uk

There are two principle factors which influence nearly all aspects of arc magmatism, from chemical compositions to petrology, textures, distribution and eruption/ emplacement behaviour. These are: Sources (which of the components in the subduction zone system - slab, mantle and lithosphere of the overriding plate - have contributed to the magmas), and Processes (the effects of melting, crystallization, mixing, contamination and interaction with the surroundings during storage and delivery) Technological advances, particularly in geochemistry, have resulted in a strong bias towards isotopic and trace element data in trying to constrain the sources and processes that determine the characteristics of arc magmas. Far from converging on a "universal theory" the progressive accumulation of geochemical approaches has resulted in a current lack of consensus on what sources or processes are important. Indeed there has even been idealogical backtracking on the important issue of what is the contribution from the subducted slab. With the advent of plate tectonics and the elucidation of subduction zone geometry it was assumed that the subducted plate simply melted as it was thrust into the hot mantle, producing the andesites and granites which were known to be common along convergent plate margins. By the time Jim Gill's 1981 book was published, petrological and geochemical research, especially trace element studies, leaned heavily towards an interpretation that arc magmatism reflected melting of the mantle wedge, the solidus of which was lowered due to fluxing by fluid from the subducted slab. In this interpretation primary magmas are basaltic and andesites/granites are differentiates. In 1990 the proverbial spanner in the works was Defant and Drummond's suggestion that a special type of arc magma, adakites, were melts of the subducted slab. In principle, adakites are the exception that proves the rule - their compositions (dacitic, high Sr, low HREE) are distinct from "typical arc magmas" and their association with very young, and therefore hot, subducted slabs provided a reasonable mechanism for slab melting in exceptional circumstances. In practice, slab melting was back on the table. Recent geochemical studies have advocated multiple contributions from sediment and basaltic slab, transferred in both fluids and melts. As usual, "the picture is more complicated than we had originally assumed!" This contribution will summarise what I think we know (or at least what I think we think we know) about arc magmas in 2005. It will inevitably be incomplete, and many of the issues which I see as problems will likely have solutions which others have already thought of It will also draw shamelessly on my own research and prejudices. It would be nice not to have to conclude that every arc is different, and that arc magmas are the result of variable

contributions from a multitude of sources acted on by a diverse range of processes at different times and places in their evolution, but

Some zeroth order observations

It may be profitable to stand back from the morass of geochemical data and ask what we really confidently know (as oppose to infer) about arc rnagmatism. I submit that: 1. There is a geometric relationship between the slab and the arc. In the simplest sense magmatic arcs clearly lie above a dipping slab, although the notion that the arc-slab depth is fixed due to a pressure-sensitive dehydration control on melting has been recently challenged (England et al 2004). The superposition of arc and slab indicates that there is a dynamic spatial control on magma generation - something about the architecture of subduction zones causes melting, and may, in tum, influence the composition of the melts. 2. Despite the wide variety of arc magma compositions, there are surely some general characteristics which distinguish magmas at subduction zones from those of other tectonic settings. In particular, arc magmas are distinct in: a. Being differentiated (Davidson, 1996); high Mg basalts are rare, and the abundance of intermediate rocks such as andesites was an important factor in early claims that these were primary magmas. b. Having a characteristic trace element pattern with relatively high abundances of LILE, and relatively low abundances of HFSE. c. Having high H2O contents. This property is difficult to measure, especially in subaerial rocks where original H2O has been degassed, but recent studies of melt inclusions and submarine glasses lend confidence to this generality (Wallace, 2005). It is worth noting that the combination of a and c - i.e. high water contents in relatively silicic, and therefore viscous, magmas - is why arc volcanism is typically explosive.

What w^e can infer about arc magmas

It is reasonable to suppose that primary arc magmas represent variable contributions from slab and mantle, but it is clear that primary magmas, like mantle xenoliths are "as rare as hen's teeth" (to borrow Dick Arculus' phrase). In most studies the slab contribution dominates discussion and modelling, probably because, despite typically minor volumetric contribution, it has a major effect on incompatible trace element and isotopic compositions. The wedge is volumetrically the dominant contributor in most arc magma systems, and it is important to constrain the composition of the mantle wedge if we are to accurately quantify mass fluxes at arcs.

The mantle wedge

To a first approximation many authors have assumed that the mantle wedge is equivalent to MORB source. Is this reasonable? The geometry of subduction zones restricts the mantle source to that of the wedge-shaped region above the slab, which effectively isolates the wedge from the mantle below. Even

allowing for convection of the wedge, it is unlikely to involve mantle deeper than --670 km. Furthermore geographic provinciality among MORB sources (Indian, Pacific and Atlantic Ocean MORBs have distinct geochemical characteristics) are replicated among arc systems, as documented especially clearly in the SW Pacific (Hergt and Hawkesworth, 1994). Thus if MORBs sample the upper mantle, then arcs sample that same mantle interval. If arc magmas share a common source with MORB then the geochemical difference between primary arc magmas and primary MORB must be due to; 1. An additional contribution from the subducted slab at arcs (more of this below), and/or 2. Differences in the degree of melting/ fertility of the MORB source in arcs vs. mid ocean ridges At arcs, melting may be controlled by the amount of fluid added from the subducted slab (Stolper and Newman, 1994) - the greater the fluid contribution the larger the degree of melting and therefore the lower the relative abundances of incompatible elements. In such a scenario there is a complex interplay between the element budget contributed in the fluid from the slab, versus the change in element distribution consequent on varying the degree of partial melting. HFSE distributions in arc magmas are typically assumed to reflect source characteristics uncompromised by subsequent fluid addition (Pearce et al, 1999), since these elements are not thought to be fluid-mobile. Even this belief has been recently called into question by Hf isotope provinciality, which appears to reflect that of the slab rather than the underlying wedge (Woodhead et al., 2001). Regardless, arc sources are evidently at least as depleted as MORB source and undergo probably higher degrees of melting than MORB. There may also be a component of decompression melting - Plank and Langmuir (1988) suggest that the degree of melting reflects the vertical extent of the pathway over which decompression accompanies ascent. Thus, the greater the vertical ascent the greater the degree of partial melting. They equate the ascent path with the depth from the base of the crust to the top of the slab. For a roughly constant depth of Benioff zone, this means that the thinner the crust, the longer the decompression melting pathway and the greater the degree of melting. Correlations between geochemistry and thickness of the overriding plate are consistent with this model - but can also be explained by interaction with, and depth of fractionation control by, the upper plate (e.g. Leeman, 1983; Coulon and Thorpe, 1981). The Slab (which I define as the subducted oceanic lithosphere along with its complement of superjacent sediment) The slab contribution may consist of some combination of; 1. Fluid, derived by dehydration and metamorphism of the oceanic crust and/or sediments. The properties of this enigmatic fluid (supercritical? H2O or H2O-CO2 mixture?) and its partitioning behaviour are poorly known. Experimental studies have largely confirmed predicted behaviour with elements such as LILE fluid-soluble whereas HFSE are immobile (e.g. Tatsumi et al, 1986; Brenan et al, 1995; Keppler, 1996). 2. Sediment or sediment melt. Global correlations between the trace element inventory of the downgoing sediment column and the trace element distributions of associated magmatic

rocks at a given arc lend credence to the suggestion that sediments are recycled at subduction zones (Plank and Langmuir, 1993). The identification of live cosmogenic in arc magmas is difficult to explain by any other mechanism than recycling of subducted seafloor sediment on timescales <10 myrs (Tera et al, 1986). For simple geochemical modelling bulk sediment addition is commonly assumed. However it is difficult to visualise how sediment can be mechanically mixed into the superadjacent wedge, to provide a suitable contaminated peridotite source. Melting may be a more likely mechanism for sediment addition. Indeed in some cases sediment melting is demanded to satisfy geochemical constraints (e.g. Elliot et al, 1997). 3. Melts of the oceanic crust. In the extreme, melting of the basaltic slab at high pressure can produce adakites, which are essentially dacitic partial melts in equilibrium with an eclogite source. Eruption of pure adakite at the surface is rare, and it has been suggested that many slab melts are contaminated by the hot peridotite wedge through which they ascend, producing high Mg andesites (Kelemen and Ghiorso, 1986). More difficult to identify though would be small adakitic contributions to otherwise typical arc magmas.

What informs our constraints on arc magmatism? A box-model type approach enables us to identify potential sources which might contribute to arc magmatism, and the processes which control these contributions. However, we need to now determine what sources actually do contribute. Two different perspectives can be adopted to address this problem: 1) modelling the thermal structure of the subduction zone and using this to predict melting and dehydration processes and the partitioning of elements as a consequence. This is a forward modelling approach which makes predictions about the compositions of magmas generated, which in turn can be compared with; 2) analysis of magma geochemistry and inverse modelling to constrain the sources and processes by which these compositions are generated. Thermal structure The thermal structure of subduction zones has been the subject of extensive modelling (e.g. Davies and Stevenson, 1992) and experimental study (Kincaid and Griffiths, 2004). It depends on a large number of factors including slab age, time since subduction initiation, subduction rate, subduction angle and the convection dynamics of the mantle wedge. As a result we can expect each subduction zone to have a unique thermal structure, and at any given arc that thermal structure will typically evolve with time. Intrinsic in modelling are several feedback factors. For instance, the wedge convection, which controls temperature advection, is dependent on the mantle viscosity, which is sensitive to H2O content. The H2O content of the mantle wedge depends on how and where it is added from the dehydrating slab - which, in turn, is controlled by the temperature imposed by the convecting wedge. And so on. The temperature structure relative to depth (pressure) determines what phases dehydrate where, and whether melting (of sediments or basaltic slab) can occur. Current published thermal structures are very accommodating for geochemists showing that the slab/ sediments may - or may not - melt (van Keken et al,, 2002).

The observation that magmatism is not universal at subduction zones and is absent where slab dips are very shallow, seems to indicate at least that convecting mantle wedge is required to be superposed over the dehydrating slab. In the case of shallow subduction, fluids will be released into cool (and relatively refractory) lithospheric mantle and will not trigger melting. Geochemical interpretation The largest body of research into subduction zone magmatism, perhaps inevitably, lies in the petrology and geochemistry of the arc magmatic rocks themselves. As alluded to above, there is, in general, strong geochemical evidence for a fluid component (e.g. high LILE/HFSE) and a sediment component (e.g. elevated ^^Be, trace element correlations between input and output) in most subduction-related magmas. Progress beyond this generality to determine how much sediment/ fluid, how it is added and when in the subduction process, has proved difficult (Figure 1). Trace element data can appear contradictory. For instance B/Be data presented by Morris et al (1990) seem to indicate that each arc defines a unique mixing trajectory between mantle and a specific sedimentifluid mixture. On the other hand, Hawkesworth et aL, (1997) using Ba/Th vs Th (and analogues) showed that global arc data define an array between putative "sediment" and "fluid" end members rather than mixtures between mantle and sediment-fluid mixtures. In terms of the mechanism of adding in slab components, distinct fluid and sediment additions have been proposed at a number of arcs. In the Marianas Elliot et al (1997) suggested that a sediment component was added to the wedge (to satisfy Th isotope constraints) before the fluid, which generated relative U enrichment and presumably triggered melting. Likewise at the Tonga arc. Turner and Hawkesworth (1997) use U-Th disequilibria to indicate fluid addition shortly before eruption, with ^^Be added in a slab melt significantly earlier. In both these cases it is difficult to reconcile the geochemical conclusions with the implication that sediment melts are generated at lower temperatures and pressures than fluids, or that dehydration is even possible from a slab which has had melt removed. Geochemical inferences can be compared with the predictions of thermal models - although there is some danger in using them to inform thermal structure. It is unwise, for instance to conclude that geochemistry indicates the involvement of sediment melts and then use as a constraint for thermal modelling that "we know slab surface temperatures must exceed the solidus of sediments".

Pause for thought - what do the geochemical data really tell us? At this stage a note of caution should be added. The uniformly differentiated nature of arc magmas means that we are always extrapolating - often over large compositional gaps - to assumed primary magma compositions. Can the effects of differentiation on isotopic and trace element ratios be accounted for or ignored?

POSSIBLE INTERPRETATIONS

COMPONENTS 800 ^ 600

Each arc mixes wedge with a specific (sediment+fluid) slab component (txjt arc arrays ought tothen" to then be perpendicular to global array,' f, which they tend not to be

400

200 ^

rnmmmKJ 10

Pelagic sediments

T-

2-stage pnocess; sediment added to wedge already metasomatised by fluid ( or fluid added to wedge already contaminated by sediment ( — • )

Fluids

100-

50Aleutians crust . and mantl^

i/ 100

B/Be

Th ppm

Arc data an-ays

200

300

High Ba/Th, high (^^su/^^OTh) )f arc data arrays Fluids (added jus< prior to - possiWy causing - melting)

Th ppm

400

B/Be

^ Low Ba/Th. high Th C e n d s of arc data arrays

B/Be

....Implies that Be is fluid mobile

Degree of melting (m) depends on anrxxint offluidadoea (f); less fluid, less melting, higher Ih. But this does not account for isotopic variations which correlate with Th ppm.

Overprinting of source to which a slab component has been added (1), by subsequent crustal contamination ((2); consistent with increasing 5I8O, 87Sr/86Sr.La/Yb in many individual arrays towards NcustO)

Th ppm

Figure 1. Illustration of potential ambiguity in interpreting trace element data - components and data fields/ arrays from Hawkesworth et al. (1997) and Morris et al (1990).

In principle it is actually very difficult to distinguish the effects of crustal contamination from those of sediment addition to the mantle source. Incompatible trace element or isotopic mixing trajectories are broadly similar with differences only in the amount added - a 1% addition of sediment to the mantle wedge will have a similar effect to adding 10% sediment during contamination of a mantle-derived magma. In the case of island arcs built on oceanic lithosphere, crustal contamination is commonly ignored. However, island arcs must be built on a sediment layer equivalent to - or greater if offscraping and underplating are significant than that which is subducted. Island arcs such as the Lesser Antilles have been shown to exhibit the effects of intra-crustal contamination (Davidson, 1987). Furthermore, there is no reason why altered basaltic crust cannot act as a contaminant, it simply shows up less well in

7 geochemical data due to its general compositional similarity to the ascending basaltic magma and its young age. Even the ubiquitous "fluid" signature, which apparently defines arc magmas, cannot be a priori wholly ascribed to addition of fluid from the slab. The trace element signature of arc magmas is shared with that of the continental crust, so crustal contamination of any mantle-derived basalt will tend to impart an "arc" signature. Recent work has concentrated on picking rocks apart geochemically, focusing on crystals and melt inclusions. A disturbing consensus is materialising from these studies; 1) most arc rocks are mechanical mixtures of melt and crystals which have not co-evolved (Davidson et al, 2005). 2) the crystals provide records of magmatic evolution (Davidson et al, 1998) - they commonly indicate multiple origins with significant recycling of earlier crystal debris (Dungan and Davidson, 2004). 3) melts (groundmass) are commonly andesitic and rhyolitic, and melt inclusions may be quite contaminated rather than primitive, representing compositions produced at cooling boundary layers (Danyushevsky et al, 2002).

Moho

1: Basalt from the mantle wedge crosses Moho and stalls at density filter 2: Basalt sills cool and crystallise, heating and melting crust 3. Cumulates transferred below Moho 4. Shallow level differentiation of magmas which have been shaped by deep crustal processes

Figure 2: Simple model of magma supply from mantle (left) is being replaced by more realistic model of deep processing overprinted in part by shallow level processes of magma mixing and crystallisation (right) The textbook "balloon and straw" model of a magma system is being replaced by a model involving a sub-volcanic plexus of dykes and sills representing the subvolcanic storage and delivery system (Figure 2). There is also growing evidence that basaltic addition to the crust stalls and differentiates in the lower crust (Annen et al, 2002; Petford and Gallagher, 2001), fractionating and mixing with crustal melts generated by the addition of heat. A consequence

8 of this scenario is that melts may evolve at deep levels and entrain crystals at any level, so that the geochemistry may be influenced by phases such as amphibole and garnet which are not seen in the rocks, while those crystals that are found in the rocks have little relationship with the melt in which they are now hosted. Careful petrographers of arc rocks have of course always been aware of this from the common disequilibrium textures observed. Another potential control offered by this perspective is the generation of calcalkaline and arc tholeiitic suites respectively. In regions of thicker crust (oceanic or continental) there will be an increasing tendency to process magmas at greater depths, allowing for a greater role of amphibole and/or garnet and for lesser control by plagioclase, all of which may contribute to calc alkaline differentiation trends (Grove and Kinzler, 1986). Given all of the above it is perhaps a surprise that: 1. Suites of rocks from arc plutons/ volcanoes do typically define fairly simple trends, and 2. Source compositions can be teased from arc rocks using care

Implications for the origin and evolution of the continental crust The broad geochemical similarity between arcs and the continental crust has led to the supposition that the continental crust has been generated through subduction zone processes. The composition of primary/ primitive magmas at subduction zones becomes critical here. If it is andesitic then it would presumably be high Mg andesite which can be shown to be generated from hydrous mantle, but these are rare in arcs today and are quite distinct from the broadly andesitic continental crust. It is more likely to be basaltic, accounting for the rare primitive basalts that are erupted in arcs and implying that andesites are simply differentiates. The abundance of differentiated material then indicates that arcs are efficient sites of magma processing. If we accept that the main flux from the mantle at subduction zones is primitive basalt then either; • Arcs are not the principle factories for generating continental crust (but we know that differentiated crust-like material is produced at arcs, and there are no more likely alternative sites), or • The continental crust is growing more mafic with time (this would be consistent with the conclusion that the present-day flux is basaltic, but there is no evidence for a secular change in composition through time), or • There is an additional process which returns mafic/ ultramafic material to the mantle such that the net flux to the crust across the Moho is andesitic. In our current understanding of arcs as magma factories, they are ideal sites for processing and cannibalising materials, particularly in the deep crust where complementary cumulates can be transferred into the convecting mantle.

References

Annen, C., and Sparks, R.S.J., 2002. Effects of repetitive emplacement of basaltic intrusions on thermal evolution and melt generation in the crust. Earth and Planetary Science Letters, 203: 937-955. Brenan, J.M., Shaw, H.F., Ryerson, F.J. and Phinney, D.L., 1995. Mineral-aqueous fluid partitioning of trace elements at 900°C and 2.0 GPa: Constraints on the trace element chemistry of mantle and deep cmstal fluids. Geochimica et Cosmochimica Acta, 59: 3331-3350. Coulon, C., and Thorpe, R.S., 1981. Role of continental crust in petrogenesis of orogenic volcanic associations. Tectonophysics, 77: 79-93. Danyushevsky L.V., Sokolov, S., and Falloon, T.J., 2002. Melt inclusions in olivine phenocrysts: using diffusive re-equilibration to determine the cooling history of a crystal, with implications for the origin of olivine-phyric volcanic rocks. Journal of Petrology, 43: 1651-1671. Davidson, J.P., 1987. Cmstal contamination versus subduction zone enrichment: Examples from the Lesser Antilles and implications for mantle source compositions of island arc volcanics. Geochimica et Cosmochimica Acta, 51:2185-2198. Davidson, J.P., 1996. Deciphering Mantle and Cmstal Signatures in Subduction Zone Magmatism. In : Bebout, G.E., Scholl, D.W., Kirby, S.H. and Piatt, J.P. (eds). Subduction: Top to Bottom. American Geophysical Union Monograph 96: 251-262. Davidson, J.P., Tepley, F.J.III, and Knesel, K.M., 1998. Crystal isotope stratigraphy; A method for Constraining Magma Differentiation Pathways. Eos,19 {IS): 185, 189, 193. Davidson, J.P., Hora, J.M., Garrison, J.M. and Dungan, M.A., 2005. Cmstal Forensics in Arc Magmas. In: Leeman, W.P. and Davidson, J.P. (eds). Energy and Mass Fluxes in Volcanic Arcs. Special SOTA Issue. Journal ofVolcanologyand Geothermal Research, 140: 157-170. Davies, J.H and Stevenson, D.J., 1992. Physical model of source region of subduction zone volcanics. Journal of Geophysical Research, 97: 2037-2070. Defant, M.J., and Dmmmond, M.S., 1990. Derivation of some modem arc magmas by melting of young subducted lithosphere. Nature, 347: 662-665. Dungan, M.A. and Davidson, J.P., 2004. Partial assimilative recycling of the mafic plutonic roots of arc volcanoes: An example from the Chilean Andes. Geology, 32: 773-776. Elliot, T., Plank, T., Zindler, A., White, W., and Bourdon, B., 1997. Element transport from slab to volcanic front at the Mariana arc. Journal of Geophysical Research, 102: 14,991-15,019. England, P., Engdahl, R. and Thatcher, W., 2004. Systematic variation in the depths of slabs beneath arc volcanoes. Geophysics Journal International, 156: 277-408. Gill, J.B., 1981. Orogenic andesites and plate tectonics. Springer-Verlag, New York. Hawkesworth, C.J , Turner, S.P., McDermott, F., Peate, D.W. and van Calsteren, P., 1997. U-Th isotopes in arc magmas: implications for element transfer from the subducted cmst. Science, 276: 551-555. Hergt, J.M. and Hawkesworth, C.J., 1994. Pb, Sr, and Nd isotopic evolution of the Lau Basin implications for mantle dynamics during back-arc opening. Proceedings of ODP Leg 135: 505-517. Kelemen, P.B. and Ghiorso, M.S., 1986. Assimilation of peridotite in zoned calc-alkaline plutonic complexes: evidence from the Big Jim complex, Washington Cascades. Contributions to Mineralogy and Petrology, 94: 1228.

10 Keppler, H., 1996. Constraints from partitioning experiments on the composition of subduction-zone fluids. Nature, 380: 237-240. Kincaid, C. and Griffiths, R.W., 2004. Variability in flow and temperatures within mantle subduction zones. Geochemistry Geophysics Geosystems, 5, Q06002, doi:10.1029/2003GC000666. Grove, T.L. and Kinzler, R.J., 1986. Petrogenesis of andesites. Annual Reviews of Earth and Planetary Science, 14:417-454. Leeman, W.P., 1983. The influence of crustal structure on compositions of subduction-related magmas. In: Aramaki, S. and Kushiro, I., (eds). Arc Volcanism. Journal of Volcanology and Geothermal Research, 18: 561588. Morris, J.D., W.P. Leeman, and F. Tera, 1990. The subducted component in island arc lavas: constraints from Be isotopes and B-Be systematics. Nature, 344: 31-36. Pearce, J. A., P.D. Kempton, P.D., No well, G.M. and Noble, S.R., 1999. Hf-Nd element and isotope perspective on the nature and provenance of mantle and subduction components in western Pacific arc-basin systems. Journal of Petrology, 40: 1579-1611. Petford, N., and Gallagher, K., 2001. Partial melting of mafic (amphibolitic) lower crust by periodic influx of basaltic magma. Earth and Planetary Science Letters, 193: 483^99. Plank, T. and C.H. Langmuir, 1988. An evaluation of the global variations in the major element chemistry of arc basalts. Earth and Planetary Science Letters, 90: 349-370. Plank, T., and Langmuir, C.H., 1993. Tracing trace elements from sediment input to volcanic output at subduction zones. Nature, 362: 739-743. Stolper, E., and Newman, S., 1994. The role of water in the petrogenesis of Mariana Trough magmas. Earth and Planetary Science Letters, 121: 293-326. Tatsumi, Y., Hamilton, D.L., and Nesbitt, R.W., 1986. Chemical characteristics of fluid phase released from a subducted lithosphere and the origin of arc magmas: evidence from high-pressure experiments and natural rocks. Journal of Volcanology and Geothermal Research, 29: 293-309. Tera, F., Brown, L., Morris, J., Selwyn Sacks, I., Klein, J. and Middleton, R., 1986. Sediment incorporation in islands arc magma: inferences from ^^Be. Geochimica et Cosmochimica Acta, 50: 535-550. Turner, S., and Hawkesworth, C., 1997. Constraints on flux rates and mantle dynamics beneath island arcs from Tonga-Kermadec lava geochemistry. Nature, 389: 568-573. van Keken, P.E., Kiefer, B. and Peacock, S.M., 2002. High-resolution models of subduction zones: Implications for mineral dehydration reactions and the transport of water into the deep mantle. Geochemistry Geophysics, Geosystems, 3, 1056, doi: 10.1029/2001GC000256. Wallace, P.J., 2005. Volatiles in Subduction zone magmas: concentrations and fluxes based on melt inclusions and volcanic gas data. In: Leeman, W.P. and Davidson, J.P. (eds). Energy and Mass Fluxes in Volcanic Arcs. Special SOTA Issue. Journal of Volcanology and Geothermal Research, 140: 217-240. Woodhead, J.D., Hergt, J.M., Davidson, J.P. and Eggins, S.M., 2001. Hafriium isotope evidence for "conservative" element mobility during subduction zone processes. Earth and Planetary Science Letters, 192: 331-346.

Secondary minerals: scientific curiosities or geochemical treasure troves? Peter A. Williams (Invited Speaker) School of Science, Food and Horticulture, University of Western Sydney, Locked Bag 1797, Penrith South DC NSW 1797; p.williams@uws.edu.au

Primary sulfides and sulfosalts, usually formed deep in the Earth at elevated temperatures and under reducing conditions, are chemically unstable under oxidizing conditions above the water table. Oxidation leads to a suite of minerals with elements attaining, for the most part, their highest oxidation states. For example, pyrite reacts with oxygen and water to form goethite, protons, and sulfate ions, as shown below. FeS2(s) + 15/402(g) + 5/2H20(l)

FeOOH(s) + 4ll\aq)

+ 2S04^"(aq)

Similarly, As-bearing sulfosalts are oxidized to give, inter alia, arsenate(III) and arsenate(V) species. In general terms, sulfides and sulfosalts that are quite insoluble in water are transformed to sulfates and their oxidized equivalents, and these are much more soluble in aqueous solution. Other anions contributed from ground and vadose waters, such as phosphate, carbonate and chloride, can react with metal ions to form secondary minerals of great complexity and beauty (Williams, 1990). The chemical environment in which these processes occur is highly variable and can be quite hostile, chemically speaking. Oxidation potentials reach the limits of the thermodynamic stability of water, exemplified by the formation of such species as Cr04^', IO3", IO4" and CIO4" . The range of pH spans solutions that are about 5M in H2SO4, generated by reactions akin to that above followed by dehydration, to values around 12 achieved by the hydrolysis of Fedeficient silicate phases. About one half of the 4000 or so known minerals occur in the oxidized zones of base metal ore bodies. Many are thermodynamically stable with respect to prevailing solution conditions, but others are metastable, being formed as a result of kinetic influences. The latter are more or less ephemeral phases and ultimately react to give more stable assemblages if conditions permit. A typically stable oxidized Pb-Cu-Zn assemblage formed at neutral pH in the presence of C02(g) would comprise malachite (Cu2C03(0H)2), azurite (Cu3(C03)2(OH)2), cerussite (PbC03) and smithsonite (ZnC03). These, together with a small number of sulfates, chlorides, arsenates, silicates and phosphates characterize the bulk of the secondary base metal mineral assemblages found in a host of settings (Williams, 1990). The products obtained by reacting a Cu(II) chloride solution with OH- ion (Pollard et aL, 1989) provide an example of an evolving metastable assemblage. In this case, the first precipitated phase is claringbullite (Cu4Cl(OH)7), followed sequentially by botallackite, atacamite and clinoatacamite (all polymorphs of composition Cu2Cl(OH)3), or paratacamite (also Cu2C1(OH)3), if suitable amounts of Zn^^(aq) or Ni^^(aq) are available in solution. Dehydration of the reaction medium will serve to preserve any of these phases, all of which form a complex Ostwald Step Rule cascade of progressively more stable species (Morse and Casey, 1988).

The characterization of new secondary minerals of the base metals and their congeners has been referred to disparagingly as mere "stamp collecting." However, nothing could be further from the truth. Secondary minerals are nothing more that the inorganic chemicals that reflect the chemical conditions that prevailed during their formation. If we can leam to "read the evidence," we can unlock the door to the geochemical past and apply the findings in a host of ways. Some examples of how one might go about this are given below. The origins and expressions of geochemical anomalies are complex and involve many complementary pathways. Secondary minerals, however, represent the metal ion buffer between insoluble species and the dispersion in elements in the regolith. In the oxidized zone of the New Cobar deposit near Cobar, New South Wales separate suites of oxidized base metal carbonate and arsenate minerals are present (Leverett et aL, 2003). The arsenate suite includes bayldonite (Cu3Pb(As04)2(OH)2), chenevixite (Cu2Fe2(As04)2(0H)4 H20), duftite (CuPbAs040H), gartrellite (PbCuFe(As04)2(0H) H20), mimetite (Pb5(As04)3Cl), olivenite (CU2ASO4OH), philipsbomite (PbAl3(As04)2(0H)5-H20) and segnitite (PbFe3(As04)2(0H)5-H20), with minor amounts of agardite (Cu6(REE)(As04)3(0H)6-3H20) occurring as late overgrowths. The arsenate minerals have a very limited spatial distribution. Superimposed on this is a later copper suite dominated by azurite and malachite, these being widely dispersed in the enclosing host rocks. It is possible to reconstruct the solution conditions responsible for the generation of the separate suites and thus to explain the differential geochemical dispersion of Cu and Pb±As from the oxidizing deposit. A phase diagram for the important arsenate species is shown in Figure 1. Field associations of the arsenate minerals indicate the general conditions under which the suite crystallized. Olivenitebayldonite-duftite intergrowths at a Pb^^ activity of 10"^ (higher values serve to obliterate the olivenite field except at unrealistically high Cu^"^ activities) indicate a solution pH that is acidic, in line with observations of members of the jarosite super group and the absence of comwallite (Cu5(As04)2(0H)4) and clinoclase (CuAs04(0H)3) in the deposit. An average Cu^^ activity of 10""^ is reasonable (about the centre of the bayldonite field in Figure 1) and the presence of mimetite is significant; its stability field is contoured on Figure 1 for varying chloride activities. A value for chloride activity of 10'^ during the mineralizing event is appropriate, as higher values would serve to obliterate the duftite and bayldonite fields. A similar sulfate activity (10"^) is justified by reference to related groundwater compositions in other deposits (Williams, 1990) and the absence of anglesite for the given lead ion activity. A good proxy, therefore, for the bulk composition of the mineralizing solutions is 10'^ M NaCl plus 10'^ M Na2S04. This was used by Leverett et al (2004) to model total dissolved Cu, Pb and As concentrations giving rise to the assemblage, these being 4.90 x 10"^, 7.94 x 10'^ and 1.45 X 10"^ M, respectively. For the separate carbonate event, a similar approach gave a total dissolved Cu concentration of 1.22 x 10"^ mol dm"^, a value almost two orders of magnitude higher than in the case of the arsenate mineralization. These calculations show both how and why the elements Cu, Pb and As are differentially dispersed in this setting and point to an exploration geochemistry model for deposits in the Cobar Basin (Leverett et al, 2004).

-1

-2' -3-

^+ -4OA O

Philipsbornite

—-5-6-7-8Figure 1. Stability field diagram for selected Cu and Pb arsenates at 298 K, with application to the New Cobar deposit; Pb^"^ activity is 10"^ and CI" activities equal 10^ 10'\ 10"^ (bold), and 10'^ (as indicated).

A similar approach illustrates the reason for the more extensive dispersion of Cu associated with the E26 deposit at Northparkes, New South Wales, versus the nearby E22 and E27 ore bodies. Upper oxidized zones in all three are dominated by the secondary Cu(II) phosphates libethenite (CU2PO4OH) and pseudomalachite (Cu5(P04)2(OH)4) and, uniquely to E26, sampleite (NaCaCu5(P04)4Cl-5H20) (Crane et al, 1998; McLean et al, 2004). At depth, zones dominated by malachite and azurite (and, notably, atacamite in E26) occur. A controlling factor in the development of the very different oxidized zone in E26 is the requirement of both high phosphate and chloride activities for sampleite to form (Clissold et al., 2005) as compared to libethenite, pseudomalachite and cometite (Cu3P04(0H)3) (Crane et al, 2001). This is reflected in the compositions of present groundwaters (McLean et aL, 2004); those associated with E26 are much more saline (NaCl). Aside from the above application of concepts involving equilibrium to geochemical phenomena, similar considerations apply to the development of hydrometallurgical methods of extraction of metals from oxidised ores. Leaching of copper, for example, represents nothing more complicated than a reversal of the solution conditions responsible for their formation. Fine-tuning of leach circuits and consideration of undesirable build-up of dissolved species such as phosphate and arsenate can be dealt with on the basis of a knowledge of the oxidised assemblage. Finally, the formation of secondary base metal minerals in polluted environments cannot be overlooked. Pb can be immobilised in a variety of phosphates and

arsenates in such settings and this has considerable potential for the remediation of certain contaminated sites (Morin et al, 2001; Ruby et al, 1994). These are but a few examples of the fact that, despite the complexity of the chemistry of the supergene environment, comparatively simple solution models can be constructed to yield quite subtle geochemical information. This can be applied in turn to the solution of real problems of scientific and commercial significance. However, much remains to be accomplished before it could be argued that a comprehensive knowledge of chemical processes in the supergene environment is to hand. There is a pressing need for thermodynamic data for many secondary minerals, together with related compounds known only as synthetic phases. Because of the magnitude of the task, there is a requirement for better predictive models for thermodynamic parameters with an accuracy at 298.2 K of about 5 kJ mol ^ Apart from this, very little is known of the solution characteristics responsible for the promotion or inhibition of nucleation of solid phases. The example above conceming basic Cu(II) chlorides is a case in point. Excess chloride ion in solution almost completely inhibits the nucleation of the most stable product at 25''C (Pollard et aL, 1989), and atacamite can persist in Nature for millions of years. Solution speciation must exert a kinetic control on the system, but just how this happens remains a mystery. A further example concems the addition of base to Cu(II) in the presence of sulfate. Depending upon temperature, precipitates of metastable wroewolfeite (Cu4S04(0H)6-2H20) or posnjakite (Cu4S04(0H)6-H20) change with time, in contact with the reaction solution, to the thermodynamically stable phase, brochantite, (Cu4S04(0H)6). Titration in the reverse sense gives a precipitate of spertiniite, C U ( 0 H ) 2 , which is also metastable with respect to brochantite or tenorite (CuO), depending upon the relative amounts of reagents added (Williams, 1990). Conditions leading to langite (CU4S04(0H)6-2H20) are enigmatic. Finally, it should be recognized that aqueous inorganic systems in the natural environment are rarely disconnected from biological processes. Aside from the importance of organic ligands in metal speciation, certain inorganic species may have a biological origin, at least in part. Carbonate is an obvious example, and transformations of sulfur species are catalysed by micro-organisms (see, for example, Krumbein, 1983). Other more bizarre cases are known. A member of the exotic connellitebuttgenbachite series from the Great Australia mine, Cloncurry, Qld, Australia, contains considerable nitrate substituting for sulfate in the lattice (Hibbs et al, 2003). Here it is associated with gerhardtite (Cu2N03(0H)3) and likasite (Cu3N03(0H)5-2H20) and measurements of stable nitrogen isotope ratios show that much of the nitrate is derived from nitrogen fixation to ammonia and oxidation to nitrate in and around termite mounds. A multidisciplinary effort is imperative to solve these several problems and others related to supergene processes of importance in the exploration, processing and pollution contexts (WilHams, 2005). Acknowledgemeiits I wish to acknowledge the work of my colleagues and students, particularly Professor Peter Leverett, Dr Dai Hibbs, Meagan Clissold, Adam McKinnon and Jim Sharpe, for their contributions to the particular examples chosen to illustrate some of the themes developed.

15 References Clissold, M.E., Leverett, P., and Williams, P.A., 2005. (unpublished results). Crane, M.J., Sharpe, J.L., and Williams, P.A., 1998. The mineralogy of the oxidized zones of the E22 and E27 ore bodies at Northparkes, New South Wales. Australian Journal of Mineralogy, 4: 1-8. Crane, M.J., Sharpe, J.L., and Williams, P.A., 2001. Formation of chrysocolla and secondary copper phosphate in the highly weathered supergene zones of some Australian deposits. Records of the Australian Museum, 53: 49-56. Hibbs, D.E., Leverett, P.,and Williams, P.A., 2003. Connellite-buttgenbachite from the Great Australia mine, Cloncurry: a crystal structural formula. Australian Journal of Mineralogy, 9: 39-42. Krumbein, W.E., (ed), 1983. Microbial Geochemistry. Blackwell, Oxford. Leverett, P., McKinnon, A.R., and Williams, P.A., 2003. Mineralogy of the oxidised zone of the New Cobar orebody. In: Roach, LC. (ed). Advances in Regolith. CRC LEME, Canberra, pp. 267-270. Leverett, P., McKinnon, A.R., and Williams, P.A., 2004. A supergene exploration model for Cobar style deposits. In: McQueen, K.G., and Scott, K.R. (eds). Exploration Field Workshop Cobar Region 2004 Proceedings. CRC LEME, Perth, pp. 46-50. McLean, A.L., Chapman, J., Sharpe, J.L., and Williams, P.A., 2004. The mineralogy of the oxidized zone of the E26 ore body at Northparkes, New South Wales. Australian Journal of Mineralogy, 10: 53-58. Morin, G., Juillot, F., Idlefonse, P., Calas, G., Samama, J.-C., Chevallier, P., and Brown, Jr, G.E., 2001. Mineralogy of lead in a soil developed on a Pb-mineralized sandstone (Largentiere, France). American Mineralogist, 86: 92-104. Morse, J.W., and Casey, W.H., 1988. Ostwald processes and mineral paragenesis in sediments. American Journal of Science, 288: 537-560. Pollard, A.M., Thomas, R.G., and Williams, P.A., 1989. Synthesis and stabilities of the basic copper(II) chlorides atacamite, paratacamite and botallackite. Mineralogical Magazine, 53: 557-563. Ruby, M.V., Davis, A., and Nicholson, A., 1994. In situ formation of lead phosphates in soils as a method to immobilize lead. Environmental Science and Technology, 28: 646-654. Williams, P.A., 1990. Oxide Zone Geochemistry. Ellis Horwood, Chichester. Williams, P.A., 2005. Solutions in the "big laboratory": towards a model for metals at the Earth's surface. Pure and Applied Chemistry, 77: 643-651.

16 Geochemistry of mafic igneous rocks of the central Georgetown Inlier, North Queensland. Michael J. Baker, Anthony J. Crawford and Ron F. Berry Centre for Ore Deposit Research and School of Earth Sciences, University of Tasmania, Australia

Proterozoic mafic igneous rocks are exposed over a wide area of the central Georgetown Inlier in north Queensland. They comprise the Dead Horse Metabasalt, a pile of massive and pillowed tholeiitic lavas, and the Cobbold Metadolerite that outcrops as a multitude of sills and lesser dykes. Both units appear to be comagmatic and both are metamorphosed to lower greenschist to amphibolite facies, and were emplaced during deposition of the lower Etheridge Group. Preliminary geochemical and geochronological studies indicate the mafic rocks crystallised between 1700-1600 Ma, possibly in two separate pulses. The mafic rocks are predominately low-K tholeiites. They are geochemically similar to Palaeoproterozoic mafic rocks from the eastem Mt Isa Inlier in northeastern Australia, and in particular show strong compositional similarities to the basalts of the Toole Creek Volcanics. The geochemical similarities between the mafic rocks of the Etheridge Group and those of the Mt Isa Inlier are of regional significance, as they may also point to a tectonic relationship between the two, and hence place the Georgetown Inlier within the Proterozoic makeup of northern Australia. Introduction The Georgetown Inlier, at approximately 30 000 km^, is the largest of four regions of contemporaneous Precambrian rocks in north-east Queensland (Fig. lb). Early to middle Proterozoic sedimentary and mafic rocks comprise the oldest parts of the inlier, the majority of which are assigned to the Etheridge Group (Withnall and McKenzie, 1980). The mafic rocks are common to the lower part of the Etheridge Group, and are assigned to two units, the extrusive Dead Horse Metabasalt, and the intrusive Cobbold Metadolerite (Withnall, 1985). Both the metabasalt and the metadolerite are stratigraphically bounded by metasedimentary rocks of the Robertson River Subgroup, a sequence of low-grade metasedimentary rocks that form part of the early to middle Proterozoic Etheridge Group (Fig. Ic). Significantly, no mafic rocks occur in the Etheridge Group above the Robertson River Subgroup. This study focuses on determining the geochemical characteristics of the mafic rocks, as well as any significant variations between the rocks, and similarities with other Proterozoic mafic rocks of northeastern Australia. This information is critical for detailed evaluation of tectonic models that claim the Georgetown Inlier may or may not have been part of a Proterozoic to Palaeozoic continental block that also contained the Mt Isa and Willyama Inliers (Giles et al, 2002; Boger and Hansen, 2004). Accompanying the geochemical study, preliminary LAICPMS multigrain zircon dating of the meta-igneous rocks is used in order to better constrain the depositional and emplacement ages of the lower Etheridge Group.

Geology, geochemistry and age of the mafic rocks: preliminary results In the low-grade area of the Robertson River Subgroup, the mafic rocks have been metamorphosed to mid-greenschist facies. The Dead Horse MetabasaU contains significant compositional layering, and accompanying textural variation. Layers are anywhere up to 1000 m thick, consisting of a base of massive, fine- to medium-grained structureless flows, which commonly resemble the metadolerite sills in texture and grainsize. Stratigraphically overlying the doleritic layers are predominantly fine-grained, locally massive basalt flows, interlayered with pillowed lavas. The metadolerite occurs as numerous coarse-grained sills and lesser dykes both above and below the metabasalt. Preliminary multigrain zircon dating of three samples containing felsic differentiates as pegmatoidal schlieren in several metadolerite sills have returned ^^^Pb/^^^Pb ages of 1651 ± 33 Ma, 1643 ± 19 Ma, 1642 ± 23 Ma. The ages indicate a minimum depositional age for the lower Etheridge Group of around 1600 Ma. Major, and some trace element data, have been obtained from a group of 22 metadolerite and 13 metabasalt samples from the low-grade area of the Etheridge Group (Fig. Id). Data to be presented will also include recent ICP-MS trace element analyses of 15 mafic samples. The geochemistry of the meta-igneous rocks of the lower Etheridge Group resembles that of modem, relatively evolved low-K continental tholeiites and tholeiites generated during continental rifling leading to break-up. Both the metabasalt and metadolerite are Fe-rich (Fe203* 14.9-17.5 wt% and 13.1-18.4 wt% respectively) and define a strong Fe-enrichment trend (Fig. 2a). Correlations between MgO and immobile incompatible elements (e.g. Ti02, Zr) in the mafic rocks, and very limited Zr/Nb and Ti/Zr (Fig. 2d) are consistent with the metabasalt and metadolerite having being derived from a chemically similar parent magma (Fig. 2b, c). Comparison of the lower Etheridge Group mafic rocks with a suite of metaigneous rocks from the Toole Creek Volcanics of the upper Soldiers Cap Group, eastem Mt Isa Inlier, indicate the mafic rocks of the lower Etheridge Group contain relatively higher amounts of Fe, but are otherwise compositionally very similar to the Mt Isa rocks (Fig. 2a). The geochemical similarities between the Toole Creek Volcanics and lower Etheridge Group mafic rocks, and differences from the other major mafic suites in the Mt Isa Inlier, are significant, and need to be assessed further in terms of regional correlations. However, further detailed investigation of the geochemical similarities between mafic rocks of the lower Etheridge Group and of other eastem Australian Palaeoproterozoic examples is necessary before any attempt at correlation can be made. References

Boger, S. D., and Hansen, D., 2004. Metamorphic evolution of the Georgetown Inlier, northeast Queensland, Australia; evidence for an accreted Palaeoproterozoic terrane? Journal of Metamorphic Geology, 22: 511-527. Giles, D., Betts, P. G., and Lister, G. S., 2002. Far-field continental backarc setting for the 1.80-1.67 Ga basins of northeastern Australia. Geology, 30: 823-826. OZCHEM 2000. Ozchem National Whole Rock Geochemistry GIS Database. Geoscience Australia, Canberra. http://wv^w.ga.gov.au/oracle/#geochem Withnall, I. W., 1985. Geochemistry and tectonic significance of Proterozoic mafic rocks from the Georgetown Inlier, north Queensland. BMR Journal of Australian Geology and Geophysics, 9: 339-351. Withnall, I. W., and Mackenzie, D. E., 1980. New and revised stratigraphic units in the Proterozoic Georgetown Inlier, north Queensland. Queensland Government Mining Journal, 81: 28-43.

Fig. 1 Location and simplified geology of the Georgetown Inlier. (a) Location of northeastern Australian Proterozoic Inliers. (b) Enlargement of (a), (c) Simplified geology of the Georgetown Inlier. 1 Etheridge Group; 2 Dead Horse Metabasalt; 3 Mesoproterozoic granites; 4 undifferentiated Phanerozoic intrusives; 5 undifferentiated Phanerozoic sedimentary and extrusive igneous rocks, (d) Area of study.

20.00

•

* 14.00

200

•

• a.

° D

§: 100

Ll_ 8.00 2.00 2.00

4.00

6.00

8.00

MgO

•

A^ -

•

• °

6.00

4.00

CtJ 8.00

10.00

MgO

250

2.00

1.00

200

° ° ° A J^rA A• • ^Ehl^jt^ pD • .• •A

g1.50

150

•

° nxCD a •

100 50

0.50 0.00 2.00

2.00

10.00

2.50

150

4.00

6.00

MgO

8.00

10.00

Zr/Nb

Fig. 2 Variation of major and trace elements in mafic rocks from the Dead Horse Metabasalt (open triangles) and Cobbold Metadolerite (open squares). In (a), Fe203* = Fe0+Fe203. The shaded fields are defined by 21 meta-igneous samples from the Toole Creek Volcanics, eastern Mt isa Inlier (OZCHEM, 2000).

Decompression exsolution microstructures in garnet from Bingara, northeastern NSW: first evidence for incompletely exhumed ultrahighpressure slab material (UHP terrane) at depth. Jane B. Barron^ Lawrence M. Barron^, Terry P. Mernagh^ and Greg Duncan^. ^Consulting Petrologist, 7 Fairview Ave., St Ives, NSW 2075, Australia ^Geological Survey, New South Wales Department of Primary Industries PO BOX 344, Hunter Region Mail Centre, NSW 2310, Australia ^Geoscience Australia GPO Box 378, Canberra ACT 2601, Australia "^Rimfire Pacific Mining NL, Level 13, 379 Queen Street Brisbane, Queensland 4000, Australia

About two million diamonds were mined from Tertiary alluvial deposits at Bingara-Copeton in the Phanerozoic New England Fold Belt, New South Wales, Australia, more than 1500 km distant from the nearest craton. An exploration program of the modem drainage system and soils in the Bingara district was undertaken by Rimfire Pacific Mining NL to search for a hard rock diamond source. Eight prospects were identified by regional magnetic anomalies, and 118 samples were processed for heavy minerals. Five diamonds were recovered; two macrodiamonds and three small diamonds that retain some delicate features. More importantly, abundant high pressure (HP) and ultrahigh-pressure (UHP) metamorphic minerals were also recovered, particularly gamet. We reported (Economic Geology, in press) compositions of 3467 Bingara gamets and classified them according to Schulze (2003). Based on the gamet compositions and interpreted protoliths, a conceptual model was also presented for subduction formation of Bingara-Copeton diamonds, in Carboniferous and Triassic slabs, and their two-stage delivery to the surface, firstly by partial exhumation, then by capture in local shallow-sourced basaltic magmas without deep-sourced kimberlitic or lamproitic volcanism. Our two-stage delivery mechanism explains how diamonds may be sourced within buried UHP terranes. Compelling evidence of post-subduction exhumation of slab material from great depth (>200km) is known from UHP terranes worldwide. Microstructures, produced by decompression exsolution in gamets from completely exhumed protoliths, are characteristic. Published examples of similar exsolution microstructures are rare, and indicate gamet compositions stable only in rocks subducted to great depths and then exhumed, such as those reported in eclogites from the Dabie Sulu UHP terrane, China (Ye et al, 2000; Zhang et al, 2003), Westem Gneiss Region of Norway (Van Roermund et al, 1998, 2000, 2001), and in gamets from metapelites from the Greek Rhodope (Mposkos and Kostopoulos, 2001). Here, we present a preliminary account of similar microstructures in analysed gamets from Bingara, confirming their derivation from partly exhumed slab material that is also our proposed diamond source. Chemically identified UHP-crustal gamets (with major element chemistry and strongly elevated Na20 indicating a cmstal-derived source) indicate a contribution from deeply subducted continental crustal material at most prospects. Additionally, UHP-crustal gamets with low Na values but identified by decompression exsolution microstmctures, comprise about 9% of the 3467 Bingara gamets analysed.

20 Some UHP-crustal garnets show decompression microstructures defined by abundant (exsolved) rods of mtile (confirmed by Raman spectroscopy) that are crystallographically oriented parallel to {111} in the host garnet (Fig. 1).

Figure 1. Photomicrograph showing rutile exsolution in garnet, crossed polarized light. Rutile rods are parallel to {111} of garnet (polished thin section RF IV S4, sample 119, Braemar Prospect). Other exsolved phases include rods of apatite, ilmenite and sulfide, and also present are inclusions of omphacite, orthopyroxene and kyanite. We conclude that the population of mantle-formed garnet varies strongly from prospect to prospect in the Bingara area, and since the implicit parent rocks are not exposed at the surface, those local sources must be volcanic. Each of eight volcanic centres (eight prospects) shows entrainment of different parcels of partly exhumed target rock at depth, and the geographic distribution of garnet types maps this variation (Fig. 2). Garnets from various eclogitic parageneses, including garnets from Group 1 (diamond) eclogite (oceanic crustal protolith) are more abundant (or more accessible) at depth in the southwest comer of the area, and UHP-cmstal gamets (continental crustal protolith) are increasingly important to the north. Two rounded eclogitic diamonds, found where Group 1 (diamond) eclogitic garnet is present (Back Creek Trevallyn and Tom and Jerry Prospects), are typical of the eclogitic BingaraCopeton alluvial diamonds. In contrast, three very small diamonds, with delicate shapes indicating a proximal source, were found where UHP-cmstal gamets represent the highest metamorphic grade (Five Mile Prospect).

2004 09 41.cdr

Figure 2 Relative distribution of diamonds recovered and chemically identified gamet types by frequency at the different prospects, with a pie chart for all types (large circle) and eclogitic types (small circle).

22 References Mposkos, E.D., and Kostopoulos, D.K., 2001. Diamond, former coesite and supersilicic garnet in metasedimentary rocks from the Greek Rhodope: a new ultrahigh-pressure metamorphic province established. Earth and Planetary Science Letters, 192: 497-506. Van Roermund, H.L.M., and Drury, M.R., 1998. Ultra-high pressure (P>6 GPa) garnet peridotites in Western Norway: exhumation of mantle rocks from >185 Km depth. Terra Nova, 10: 295-301. Van Roermund, H.L.M., Drury, M.R., Bamhoom, A.A., and De Ronde, A., 2000. Non-silicate inclusions in garnet from an ultra-deep orogenic peridotite. Geological Journal, 35: 209-229. Van Roermund, H.L.M., Drury, M.R., Bamhoom, A.A., and De Ronde, A., 2001. Relict majoritic gamet microstructures from ultra-deep orogenic peridotites in Westem Norway. Journal of Petrology, 42: 117-130. Ye, K., Cong, B., and Ye, D., 2000. The possible subduction of continental material to depths greater than 200 Km. Nature, 407: 734-736, Zhang, R-Y., Liou, J.G., and Zheng, J-p., 2003. Petrogenesis of eclogite in ultramaflc rock from the Sulu UHP terrane, Eastem China. Geological Society of America, Abstracts with Programs, 35: 636.

23 Relation between remnant pressure in/and strain birefringence around inclusions in diamond. Lawrence M. Barron^ Jane B. Barron^, and Terrence P. Mernagh^ ^Geological Survey of New South Wales, Australia, barronjandl@optusnet.com.au ^BJ Barron Petrologist, 7 Fairview Avenue, St. Ives, NSW, 2075 Australia ^Geoscience Australia, GPO Box 378, Canberra ACT 2601 Australia Laser Raman spectroscopy is used to identify mineral inclusions in sealed inclusion chambers in some Australian diamonds and to determine the remnant pressure within the inclusions: 30.3, 33.1 and 34.7 kbar for coesite; 4.2, 13.6 and 23 kbar for diopsidic omphacite; and 8 kbar for grossular garnet, each determination in a different stone. These remnant pressures cause various degrees of strain birefringence in the diamond adjacent to the inclusion. Pressure on diamond causes the refractive index to change in a linear way so the strain birefnngence represents the rate of change of the refractive index as the inclusion is approached, and the anisotropy of the pressure gradient around the inclusion. The Berek compensator is used to measure the maximum strain birefringence in the host diamond around each inclusion, yielding values between 0.0037-0.0244 retardation for the above diamonds. On a graph (Fig. 1) of remnant pressure versus strain birefringence, our data set is small with wide scatter and large potential error, so only a straight-line fit is justified, with the form Pr = K * 6, for remnant pressure (Pr) in kbar and the maximum strain birefnngence (5, represented by a difference in refractive index equivalents). Calibration values are K = 10000/6.0 for spherical inclusions (16 measurements, R2 = 0.94), K = 10000/5.8 for elongate inclusions (10 measurements, R2 = 0.75) and K=10000/5.9 for all inclusions (26 measurements, R2 = 0.91). There is no statistical justification for considering these calibrations as different, so only the last is recommended and is for all inclusions, and is independent of the inclusion shape and orientation. Being linear, this equation ignores potential non-elastic behaviour at high remnant pressures (>30 kbar). This work provides a simple means to investigate inclusions in diamonds, especially in combination with other techniques. For example, a planar sheaf of "furry" graphite, in an Argyle diamond, has a Raman spectrum that could be interpreted as shifted due to a remnant pressure of up to 28 kbar. However, the lack of strain birefringence around this graphite implies the Raman spectra may be indicating an unexpected form of carbon, such as single walled carbon nanotubes at 1 bar remnant pressure. Despite the simplicity of this new technique, it is practical and analytical. With standard petrological equipment and a diamond on magic tape, it is now possible to estabhsh the remnant pressure on any inclusion, or to critique published images of diamond in cross polarised light. The new technique combines with other techniques in a powerful way - for instance it gives much better control for identifying minerals with a complex variable spectrographic response due to solid solution or a variable structural state (e.g. graphite).

24 T3 OC

0.03

+ Elongate O Equant • PAG12 E 0.025 CO All data oo 0.02 - - - Linear (Equant) Linear (Elongate) co Linear (All data) D)

I 0.015 m

y = 0.00059X R^ = 0.89379

y = 0.00060X R^ = 0.93447

y = 0.00058X R^ = 0.68627

.1X 0.005 (0 10

Pr kbar

Figure 1. Maximum strain birefringence, in diamond adjacent to an inclusion, plotted against the remnant pressure determined by laser Raman spectroscopy. The square is for a graphite platy inclusion with a [111] orientation within diamond PAG 12 (Nasdala et al, 2005; pressure determined by XRD). The simple visual confirmation that an inclusion is under positive remnant pressure means that any subsequent isotopic age dating of the inclusion will tend to generate the age of crystallisation. Conversely, an inclusion with a negative remnant pressure (a vacuum) will tend to yield the age of emplacement. Lastly, Barron (2003, 2005) showed that for most minerals, a remnant pressure value will tend to reflect the conditions of diamond formation. Thus the new technique can: 1. within a suite of diamonds from an established provenance, aid provisional identification of included minerals, or conversely, 2. knowing the included mineral, provisionally identify the diamond formation conditions for a diamond out of context. Research is underway to extend the (Pr , 6) concept to inclusions in other minerals such as garnet and corundum.

25 Acknowledgements The following sources are thanked for loan of diamonds: D. Austen (Argyle, DA series), Cluff Resources N.L. (Copeton and Bingara, L and JB series), and B. and C. Ribeaux (Mount Airly). Published with the permission of the Director General of the New South Wales Department of Primary Industries (Mineral Resources) and the CEO of Geoscience AustraUa. References

Barron, L.M., 2003. A simple model for the pressure preservation index of inclusions in diamond. American Mineralogist, 88: 1615-1619. Barron, L.M., 2005. A Linear Model and Topology for the Host - Inclusion Mineral System Involving Diamond. Canadian Mineralogist, 42: 341-362. Nasdala, L., Hofmeister, Harris, J.W., and Glinnemann , 2005. Growth zoning and strain patterns inside diamond crystals as revealed by Raman tomography. American Mineralogist, 90: 745-748.

26 Evidence for evolving mantle regimes from 143-neodymium and 176-hafnium isotopic records VICKIE .C. BENNETT AND A.P. NUTMAN

Research School of Earth Sciences and Planetary Science Institute, The Australian National University, Canberra, ACT 0200 AUSTRALIA The ground truth of models of processes of mantle differentiation and crustal growth, particularly for the early Earth, relies on accurate determinations of original isotopic compositions. The ^'^^Sm-^'^^Nd isotopic system can potentially provide the most complete and accurate record owing to the well defined decay constant and bulk silicate Earth parameters for this system. However for both decay schemes careful analytical work must be combined with intensive field and geochronologic studies leading to judicious sample selection. Mantle isotopic evolution for ^ Nd is frequently depicted as linearly increasing from a chondritic primitive mantle composition at 4.56 Ga to modem MORB compositions. This straight line evolution implies a progressive depletion of the mantle and a continuity of rates and processes from ancient to modem Earth. The test of this model relies on clarifying and extending the isotopic record by finding the oldest, least altered mantle-derived rocks, providing them with age and geologic context and determining accurate initial compositions. Our continuing field and laboratory investigations have identified areas of southwest Greenland including the Isua supracrustal belt and the islands near Nuuk (Qilangarssuit and Innersuartuut), as having the most extensive record of early (3800-3850 Ma) Earth chemistry. New initial ^Nd (3850 Ma) from >30 mafic samples are in a narrow range of +2 to +4 (±0.5), providing a robust control on early mantle depletion. Using our new Nd isotopic results from >3800Ma mafic samples from SW Greenland, a minimum average degree of Sm/Nd fractionation of their mantle source can be calculated by assuming the fi-actionation event occurred close to Earth's formation at 4.56 Ga. This indicates that the pre-3850 Ma upper mantle is characterised by a minimum average a composition that is similar to the degree of depletion seen in the present-day MORB source mantle. This provides strong evidence for rapid formation of enriched and depleted mantle reserviors very early in Earth's history. The picture emerging from intensive studies of the 3600 Ma-3900 Ma Greenland mafic and felsic suites contrasts sharply with that inferred for the host rocks of the >4.0 Ga zircons from the Jack Hills, WA. The Jack Hills zircons are interpreted as being derived from evolved granites with crystallization temperatures of <750 C (Watson and Harrison, 2005). The oldest materials from Greenland are 3850 Ma "juvenile" tonalites with extimated crystallisation temperatures of >800 C. The Jack Hills zircons exhibit extreme isotopic heterogeneity with both highly positive and highly negative initial Guu attesting to the presence of both highly depleted mantle sources and highly evolved crustal souces (Harrison et aL, 2005). In contrast the Nd isotopic data from the early Archean samples show only a relatively narrow range of initial values. Significantly, from a large database (n>140) of both published and unpublished Nd isotopic data for >3760 Ma samples from southwest Greenland no negative

27 6Nd values have been determined. It is not clear if the change from a "Jack Hills mode" operative at >4000 Ma, with inferred minimum melt granitic crust and isotopically heterogenous reserviors and the "Greenland mode" with higher temperature juvenile TTG suites and a homogenous mantle source at ca. 3800 Ma reflects temporal evolution in tectonic styles, or alternatively, the existence of multiple tectonic settings in the early Earth. Published compilations of Nd isotopic data (e.g. McCulloch and Bennett, 1994; Bowring and Housch, 1995) defining mantle evolution curves have emphasized data from felsic rocks. Here, by combining the new Nd data from mafic suites in southwest Greenland with compilations of data from the literature for mafic/ultramafic rock suites through time, a more revised and possibly more accurate image of mantle isotopic evolution is obtained. The emerging global Nd (and to a lesser extent Hf) isotopic patterns reveal an increasingly detailed picture of rapid and extreme early planetary differentiation, likely associated with accretionary processes, followed by partial re-homogenisation of the mantle before 3.9 Ga. The 3850 Ma mantle was characterised by compositions of ca. +3 followed by near Httle change in 6Nd for greater than 1 billion years, from 3.85 to at least 2.7 Ga. The apparent large scale changing pattern of isotopic evolution must record mantle regimes characterised by different mixing times and/or variable (size, number and origin) enriched and depleted reservoirs. Early Archean rocks record rapid early depletion of some portions of the upper mantle. The extended period of near constant Cwd records balancing of the effects of crust extraction by either rapid crustal recycling or mixing with deep mantle enriched domains; post-Archean Nd isotopic compositions reflect the transition to a modem style tectonic regime. References

Bowring, S., and Housh, T., 1995. Science, 269: 1535-1540. Harrison, T.M., et al, 2005. Goldschmidt 2005 conference abstract, Geochimica et Cosmochimica Acta, 69 (lOS): 390. McCulloch, M.T., and Bennett, V.C., 1994. Geochimica et Cosmochimica Acta, 58: 4717-4738. Watson, E.G., and Harrison, T., M., 2005. Science, 308: 841-844.

143 Nd evolution- sampled mantle MORB ophiolites

Reflects large scale variations in mantle dynamic regimes.

Rapid Early differentiation

Nd (initial) 6

T Q

Well mixed (and buffered) for >1 billion years.

Linear evolution (balanced crustal growth/ recycling) Little evidence for enriched and depleted sources.

chondrites

-2

0.0

1.0

2.0

3.0

4.0

age (Ga)

Figure 1. Nd isotopic evolution of the mantle through time. Data are averages with standard deviations for mafic/ultramafic whole rock suites as compiled from literature sources and unpublished ANU data. The apparent inflections may represent large-scale changes in mantle dynamics.

29 From chromite to crocoite: chromium mineralogy and geochemistry at Dundas, western Tasmania Ralph S. Bottrill Mineral Resources Tasmania

Introduction The Dundas Ag-Pb mineral field is associated with Cambrian serpentinites and serpentinitederived greywackes and conglomerates; silver-lead lodes lie mostly along the brecciated contacts of these igneous bodies. During the Devonian Tabberabberan Orogeny, there were some pronounced deformation events and extensive granite emplacement and concomitant metasomatism and serpentinisation. Many galena-rich carbonate veins and other mineral deposits formed during this orogeny, particularly in brecciated contact zones between the ultramafics and sedimentary rocks. The breccia zones are mostly carbonate-rich (dolomite-ankerite and lesser siderite-rhodochrosite), with some quartz-rich zones and sulfide pods and veins, containing galena, sphalerite and minor pyrite, chalcopyrite, arsenopyrite and antimony sulphosalts. Fuchsitic mica (chromian muscovite) is variable to rich in these zones, and is derived mostly from mica in the sedimentary rocks and Cr from chromite, mostly from the ultramafics. The serpentinites and serpentine-derived sediments are commonly chromite-rich and locally rich in stichtite (Mg6(Cr,Al)2(C03)(0H)i6.4H20), and these minerals are the probable sources of the Cr in many minerals in the surrounding rocks. The timing of the stichtite is unknown. During the Cenozoic, deep weathering ensued due to high rainfall during tropical climates, continuing through to the present cool, wet temperate conditions. This weathering affected both the serpentinites and wallrocks; the carbonate and Pb-Zn-Sb-Ag-As rich lodes and listwanites adjacent to the serpentinites were particularly highly oxidised and leached. The presence of some Al-rich secondary minerals like gibbsite, philipsbomite and hinsdalite indicates a low pH (<4). The mobility of chromium is largely controlled by its mineralogy and oxidation state: Cr3+ is generally much less mobile than Cr^^ (Smith and Martell, 1976). Most Cr^"^ at Dundas probably resides in chromite, some of which was metasomatically altered to mametite, stichtite/barbertonite and fiichsite during metamorphism, during which some Cr ^ was probably also oxidised to Cr^^. Some of this chromium would have been released into solution during weathering, especially where the pH was relatively low, and fluids were oxidised, as most chromates are relatively soluble. The manganese oxides present in some lodes may also have oxidised some Cr^"^ to Cr^"^ in solution (Weaver and Hochella, 2003), enabling its precipitation as crocoite where lead is present in solution or host rocks. The Cr-rich serpentinites and serpentine-derived sediments are the probable sources of the Cr in many minerals in the lodes, including crocoite (PbCr04), grimaldiite (CrO(OH)), petterdite (PbCr2(C03)2(0H)4.H20) a Cr-Mn oxide, and Cr-rich mimetite (Pb5(As04)3Cl) and Cr-rich philipsbomite (PbAl3(As04)(As030H)(0H)6).

The world's best and most plentiful crocoite specimens come from the Adelaide and Red Lead silver-lead mines in the Dundas mineral field. The Dundas Extended, West Comet, Kosminsky, Platts Prospect, and the Kapi mine also contained crocoite. The mines originally all worked Ag- and Pb-rich sulfide and carbonate-rich primary ores, but are mostly deeply weathered, to goethite-rich vuggy gossans. Secondary minerals, besides those mentioned above, include cerussite, anglesite, dundasite, coronadite, pyrolusite, pyromorphite, chlorargyrite, goethite, gibbsite and chalcophanite. Genesis of stichtite Carbon and oxygen isotopic studies (Erik Melchiorre, pers. comm.) suggest that the stichtite at Dundas is a supergene mineral, derived from prolonged deep weathering of chromitemagnesiochromite under alkaline conditions. Other workers also suspect stichtite and other hydrotalcite-group minerals to be due to supergene weathering (Taylor et al, 1991), while others ascribe them to low temperature hydrothermal alteration by alkaline fluids. A hydrothermal origin is more in concord with the occurrence of stichtite in both outcrop and in massive, unweathered serpentinites, at the depths of up to 200 m or more in some diamond drillholes at Dundas. The timing of this hydrothermal alteration is unclear. It may post-date the major Devonian deformation, intrusion and alteration events in western Tasmania or it may relate to a later thermal event (e.g. related to the emplacement of the widespread Jurassic dolerites. Late Cretaceous thermal events or Cenozoic basalts in Tasmania). This would help to resolve the discrepancy with oxygen isotope values mentioned above. The evidence for stichtite replacing both chromite and serpentine indicate the mineral is formed by a relatively simple reaction: Fe(Cr,Al)204 + 2Mg3Si205(0H)4 + COs^" + I3H2O + 20H" + ^02 chromite serpentine ->

Mg6(Cr,Al)2(C03)(0H)i6.4H20 + 4H3Si04- (aq.) + i/2Fe203 stichtite

This reaction involves little introduction of material in or out of the system except for the volatiles H2O, CO2 and O2 and the strong leaching of silica. The iron in the chromite appears to be largely oxidised and incorporated as ferric iron in stichtite and/or the coexisting spinel. Silica is moderately soluble under the alkaline conditions likely in fluids buffered by these ultrabasic rocks, and appears to be removed from the rock; no silica-rich minerals coexist with the stichtite. The apparent lack of stichtite occurrences in many compositionally and tectonically similar chromite-rich serpentinites in Tasmania, suggests that some unique conditions are required for its formation, including: • abundant chromite • brecciation or shearing, creating permeability

31 • a high fluid flow to remove silica • C02-rich, oxidised and high pH hydrothermal fluids. Genesis of Crocoite and associated minerals The origin of the supergene minerals of the Dundas district is complex and poorly understood. The main factors are the superposition of some altered Cr-rich rocks with galena-rich veins, and deep weathering of these rocks by acid groundwaters. The rare chromium oxyhydroxide, grimaldiite appears in small pulverulent masses in some crocoite lodes (Red Lead mine). These may represent leached original stichtite masses, and some enclose small chromite grains (as stichtite typically does). The direct formation of grimaldiite by the leaching of chromite seems unlikely. Clayey patches in the same zone vary from bright green to white in colour and contain variable mixtures of gibbsite-, halloysite-, kaolinite- and illite-type clays, suggesting leaching of fuchsite-rich patches under low pH conditions. This indicates that the mobility of Cr is very variable in these altered ultramafic rocks, and probably relates to pH and Eh variability. Acid leaching of the A1 from these minerals probably directly contributes most of the A1 in the dundasite and philipsbomite. Lead is less soluble than most base metals, but the galena in the Dundas lodes oxidised to form a number of supergene minerals. Where Cr was lacking, the first supergene lead mineral formed was probably anglesite, usually as a fine-grained replacement of galena, due to relatively rapid oxidation. This may be locally overgrown or replaced by cerussite, but this is generally relatively uncommon in most of the Dundas area, suggesting the pH may have been a little too low in general. Where phosphate was available, pyromorphite formed, and some cerussite was replaced by it (Piatt's Prospect). Hinsdalite may locally replace pyromorphite, probably under more acid (low pH) conditions. Similarly, some mimetite formed where As (probably from sulphosalts or arsenopyrite) was locally high (although this mimetite contains significant Cr also). Philipsbomite may locally replace mimetite, probably again under more acid conditions (although direct pseudomorphs have not been noted). Hinsdalite and philipsbomite indicate low pH (probably <4 in general). Some Pb was probably mobilised further at low pH to react with Mn oxides to form coronadite. Petterdite contains Cr^"^, suggesting it may form in a locally more reducing environment than crocoite (containing Cr^"^). It forms close to galena, which may buffer the Eh of local fluids to a level that stabilises petterdite. Petterdite forms a limited solid solution with dundasite, the A1 analogue, which is locally abundant, often with cerussite, indicating higher pH. In general, however, the supergene lead mineralogy was dominated by crocoite. Crocoite is less soluble and more stable than most lead minerals at low pH (Liggins, 1982; Smith and Martell, 1976), which may be one of the main reasons for its abundance at Dundas, besides the great supply of both Pb and Cr, as outlined above. Crocoite may directly overgrow galena or occur in fuchsitic clays, suggesting both Pb and Cr can be mobile under different conditions at Dundas. It may also occur in wallrocks with no obvious primary Pb or Cr minerals, perhaps suggesting local formation by the mixing of Pb and Cr-rich fluids. Thus some Cr^"^ rich fluids are moving into weathered lead-rich veins, and some lead-bearing fluids

32 are moving from the lead lodes into the Cr-rich wallrocks. However, Pb and Cr may be ubiquitous in many of the wallrocks, and these metals may not have travelled very far in most situations. Summary The main factors in the genesis of the secondary chromium minerals of the Dundas district are the strong alteration of some chromite-rich rocks, including the emplacement of some galenarich veins, and the deep weathering of these rocks. The stichtite and fuchsite are considered to result from late stage, low temperature hydrothermal alteration. Prolonged weathering may have converted these minerals to grimaldiite in part, and remobilised some of the Cr as both Cr^"^ and Cr^"^. Lead was released by weathering of galena and mostly combined with Cr^"^ as crocoite, but some substituted in mimetite and Mn oxides. Where the Cr^"^ was dominant, petterdite and philipsbomite formed rather than crocoite, but these appear metastable in the present oxidised zone.

33 The father of mineralogy, his son, and frozen prawns - the cool mineral connection. Paul Carr^ Brian Jones^ and Bruce Selleck^ ^School of Earth and Environmental Sciences, University of Wollongong, Wollongong, NSW 2522 ^Department of Geology, Colgate University, Hamilton, NY, 13346-1398, USA Introduction Glendonite is but one of several names used for a group of pseudomorphs that have been recorded in a wide range of geographic locations and in strata ranging from Precambrian to Recent in age. These pseudomorphs are composed of granular carbonate (usually calcite) and have been described previously under a variety of names including barley corns, chrysanthemum stones, fundylite, gennoishi, glendonite, hedgehogs, jarrowite, opal pineapples, pseudogaylussite, thinolite, and White Sea homlets. Samples from the original locality at Glendon in the Hunter Valley of New South Wales were presented to the eminent American mineralogist, James Dana, during his visit to Australia in 1839-40 and were later described and figured in his report on the geology of New South Wales (Dana, 1849). The name "glendonite", however, was not proposed until more than fifty years later when David et al (1905) published a detailed description of the occurrence, morphology and significance of glendonites from Glendon and three other localities in the Sydney Basin. Precursor mineral The identity of the precursor mineral to glendonites has been the subject of considerable debate throughout the last century with the majority of studies relying on crystallographic data for mineral identification. Several minerals including anhydrite (CaS04), aragonite (CaCOs), celestite (SrS04), gaylussite (Na2Ca(C03)2®5H20), glauberite (Na2Ca(S04)2), gypsum (CaS04*2H20), sulphur (S) and thenardite (Na2S04) have been suggested as possible precursors, with glauberite and thenardite probably being the most favoured choices. The lack of agreement between the early investigators about the identify of the precursor mineral to glendonites and other similar pseudomorphs is not surprising due to the ambiguity of the crystallographic data and because the most likely forerunner was not discovered in nature until relatively recently. Synthetically produced calcium carbonate hexahydrate (CaC03®6H20) had been known to chemists before James Dana visited the Hunter Valley and described the original glendonites, but it was not found in nature until the mineral ikaite was discovered in the very cold waters of Ikka Fjord, southern Greenland (Pauly, 1963). This discovery provided the first indication of the identity of the elusive precursor mineral and also vindicated the suspicions held by at least some mineralogists more than a century ago. According to Edward Dana (1884), who investigated the thinolites from Lake Lahontan, Nevada, "the original mineral was one which does not appear thus far to have been observed in its natural condition, although, as will be shown later, it probably has occurred abundantly at numerous other localities." Following the discovery of ikaite it took nearly another two decades before Kaplan (1979) recognized that the composition, morphology and distribution

34 of this new mineral matched those required for the precursor to glendonites. Ikaite has now received widejspread acceptance as the precursor mineral for glendonites and related pseudomorphs (e.g. Suess et al, 1982; Carr et al, 1989; Swainson and Hammond, 2001). Significance of ikaite In the last two decades ikaite has been recognized in a wide range of facies ranging from lakes to estuaries, shallow-marine shelves and deep oceanic basins. All occurrences, however, are characterized by cold, alkaline conditions and, for at least some occurrences, elevated phosphate concentrations in precipitating waters. These elevated phosphate levels apparently suppress precipitation of other calcium carbonate polymorphs (Buchardt et ai, 2001). In addition to these natural occurrences, the white spots formed in the shells of frozen prawns have also been identified as ikaite (Mikkelsen et al, 1999). Ikaite crystals rapidly degrade to calcite plus water when warmed above 4°C (Bischoff et al, 1993) and may result in the generation of calcite aggregates whose external form pseudomorphs the ikaite crystal morphology. The limited stability of ikaite make it a robust indicator of coldwater conditions and thus the recognition of its pseudomorphs in the geological record is of great paleoclimatic significance. The restricted temperature stability and rapid conversion of ikaite to pseudomorphic calcite aggregates (i.e. glendonites) provides the potential for using the stable isotopic characteristics of glendonite calcite as a proxy for water isotope chemistry in the depositional or early burial setting. The current model for the stable isotopic evolution of Phanerozoic seawater is based on biogenic carbonate records and the availability of another proxy for seawater chemistry would be of considerable value in validating the long-term trends (Veizer et aL, 1999). Conversion of ikaite to calcite, however, involves a significant volume decrease resulting in development of high porosity and the potential for later precipitation of void-filling calcite under physical and chemical conditions that may be very different from the original depositional setting. This limits the reliability of stable isotopic results unless the primary calcite phase that was formed from ikaite conversion can be recognized and analyzed. Glendonites are widespread in the Sydney Basin and consist of brown to amber, blocky calcite crystals, 2-50 micron across, that form diffuse agglomerations and more regular dendritic arrays. This amber to brown calcite is interpreted as the calcite that formed during the early conversion of ikaite to calcite (cf Greinert and Derkachev, 2004). Pore space within the brown to amber calcite network is filled with clear, millimeter- to micron-scale, secondary calcite. Brachiopods associated with glendonites in the Sydney Basin are well-preserved, with apparently minimal recrystallization and loss of primary microfabric. Joints within the glendonite-bearing units are mineralized with calcite and rare quartz that are probably related to later tectonic or burial phenomena, but some may be related to hydrothermal fluid systems developed near basalt flows coeval with sediment deposition. Stable isotopic data Stable isotope analyses were carried out on hand-picked separates and bulk glendonite samples, together with samples of biogenic calcite (brachiopods), calcite spar from mineralized joints, and samples of coalified wood from the Sydney Basin. Samples of glendonites and associated biogenic carbonate from elsewhere were also analysed and the new data were integrated with published data to assess the reliability of glendonites as climate proxies and as records of stable isotopic chemistry of waters in the depositional setting.

35 The stable isotope signatures of modem (transformed) ikaite define distinctive trends. Most ikaite samples have a narrow range of S^^Opdb (+1 to +4.5%o) consistent with precipitation from normal marine waters at temperatures within the mineral's stability field. The broader range of 5^^Cpdb (-19 to -32%o) reflects mixing of relatively depleted carbonate from methane oxidation with relatively enriched seawater carbonate. Data for samples from the type locality in Ikka Fjord define a different array resulting in precipitation from waters formed by mixing of meteoric, submarine spring water with seawater (Buchardt et al, 2001). Stable isotope data for Sydney Basin brachiopods have a restricted range well within the range of Permian biogenic low-magnesium calcite reported in numerous studies as summarized by Veizer et al (1999). The modest spread of the Sydney Basin data may suggest some recrystallization or filling of brachiopod shell punctae by later diagenetic cement. Sydney Basin glendonite bulk samples and hand-picked separates form a broad array, particularly in terms of 5^^0pdb (-5.2 to - 21.4%o). The absence of any glendonite 5^^0pdb values more enriched than -5%o may indicate initial precipitation from relatively depleted waters from mixed meteoric-seawater sources. Alternatively, the relative depletion in 5^^0pdb of glendonites compared to the brachiopod samples, may reflect physical mixing of a more depleted phase analogous to the late spar cement in mineralized joints. The separated amber calcite, which is presumed to represent the primary calcite from the transformation of ikaite during early burial, shows no significant difference from the bulk glendonite samples or the white/clear spar. This suggests that the scale of intergrowth of primary calcite and later spar is much finer than the mm-scale fragments that were separated by hand-picking. References

Bischoff, J.L., Fitzpatrick, J.A., and Rosenbauer, R.J., 1993. The solubility and stabilisation of Ikaite (CaC03*6H20) from 0° to 25°: Environmental and palaeoclimatic implications for thinolite tufa. Journal of Geology, 101: 21-33. Buchardt, B., Israelson, C., Seaman, P., and Stockmann, G., 2001. Ikaite tufa towers in Ikka Fjord, southwest Greenland; their formation by mixing of seawater and alkaline springwater. Journal of Sedimentary Research, 71: 176-189. Carr, P.F., Jones, B.G., and Middleton, R.G., 1989. Precursor and formation of glendonites in the Sydney Basin. A ustralian Mineralogist, 4:3-12. Dana, E.S., 1884. A crystallographic study of the thinolite of Lake Lahontan. United States Geological Survey Bulletin, 12: 429-450. Dana, J.D., 1849. United States Exploring Expedition. During the years 1838, 1839, 1840, 1841, 1842, Under the command of Charles Wilkes, U.S.N. Vol. X. Geology, C. Sherman, Philadelphia. David, T.W.E., Taylor, T.G., Woohiough, W.G., and Foxall, H.G., 1905. Occurrence of the pseudomorph glendonites in New South Wales. Records of the Geological Survey of New South Wales, 8: 162-179. Greinert, J., and Derkachev, A., 2004. Glendonites and methane-derived Mg-calcites in the Sea of Okhotsk, eastern Siberia: Implications of a venting-related ikaite/glendonite formation. Marine Geology, 204: 129-144. Kaplan, M.E., 1979. Calcite pseudomorphs (pseudogaylussite, jarrowite, thinolite, glendonite, gennoishi. White Sea homlets) in sedimentary rocks. Plenum Publishing Corporation 1980. Translated from Lithologiya I Poleznye,5: 125-141.

36 Mikkelsen, A., Anderson, A.B., Engelsen, S.B., Hansen, H.C., Larsen, O., and Skibsted L.H., 1999. Presence and dehydration of ikaite, calcium carbonate hexahydrate, in frozen shrimp shell. Journal of Agriculture and Food Chemistry, 47: 911-917. Pauly, H., 1963. Ikaite, a new mineral from Greenland. Arctic, 16: 263-264. Suess, E., Balzer, W., Hesse, K-F., Muller,P.J., lingerer, C.A., and Wefer,G., 1982. Calcium carbonate hexahydrate from organic-rich sediments of the Arctic Shelf: Precursors of glendonites. Science, 216: 11281130. Swainson, LP., and Hammond, R.P., 2001. Ikaite, CaC03*6H20: Cold comfort for glendonites as palaeothermometers. American Mineralogist, 86: 1530-1533. Veizer et al, 1999. ^^Sr/^^Sr,

and

evolution of Phanerozoic seawater. Chemical Geology, 161: 59-88.

37 Plutonism and metamorphism at the root of a Cretaceous magmatic arc Stephanie A. Carroll and Nathan R. Daczko GEMOC Key centre, Department of Earth and Planetary Sciences, Macquarie University Sydney, NSW 2109, Australia In Fiordland, South Island, New Zealand, migmatites are spatially associated with the mafic plutons that are inferred to have led to or enhanced crustal melting and melt migration. The gabbro-diorite Western Fiordland Orthogneiss, that extends in a '--25 km wide belt for over 150 km, represents a well-exposed section of mafic lower crust. The patchy development of garnet reaction zones at Lake Grave, located on the west coast of Fiordland, south-east of Sutherland Sound, is described here for the first time. These patchily developed gamet reaction zones are argued to represent a rarely preserved early stage of the migration of trondhjemitic partial melts produced in high-P mafic migmatites to structurally higher levels. The trondhjemitic dykes and veins that contain orthopyroxene, clinopyroxene, plagioclase (with anti-perthite textures), and quartz form from partial melts sourced from mafic lower crust with the melts escaping the lower crust via fracture propagation and dyking. Bulk rock geochemical data presented, support models of formation of melts with adakitic signature by partial melting of mafic deep continental crust as initially suggested by Stevenson et al. (2005). The high NaaO contents and KsO/NasO-- 0.5, Ni (20-40 ppm) and Cr (30-50 ppm) in the original definition of Martin (1999) are not essential features of adakites produced by melting of a mafic lower crust under conditions that has gamet as a residual mineral in the lower crust. Regional Geology The South Island of New Zealand has been geologically divided into (i) the Eastem Province, (ii) the Westem Province and (iii) the Median Tectonic Zone (also referred to as the Median Batholith by Mortimer et al, 1999) separates these two provinces (Kimbrough et al, 1993). Late Triassic stitching plutons mark the amalgamation of the Eastem Province and the Median Tectonic Zone. Early Cretaceous plutons stitch the Median Tectonic Zone and Westem Provinces (Hollis et al, 2003). In addition to intrusive contacts of the stitching plutons, the boundaries between the three provinces are faults. Cenozoic motion along the Alpine Fault has disrupted and displaced the above-mentioned terranes (Figure 1). The gabbroic to dioritic Westem Fiordland Orthogneiss represents the main Cretaceous plutonism in Fiordland (Hollis et al, 2003), and was emplaced and metamorphosed to granulite facies between 126-119 Ma, shortly after the Arthur River Complex of northem Fiordland (Hollis et al, 2003). This unit represents a rare, young, extensive, well-preserved belt of high-P granulites that are exposed along the entire length of the Fiordland Block (Figure 1). The Westem Fiordland Orthogneiss is a composite arc batholith, >3000 km^ (Klepeis et al, 2003), emplaced into Palaeozoic orthogneiss and paragneiss and Mesozoic orthogneiss at depths of 20 -25 km. The degree of deformation and metamorphism are very variable in the Westem Fiordland Orthogneiss. The microstmctures range from igneous to completely granoblastic and gneissic,

38 while the low strain areas preserve relict igneous clinopyroxene-orthopyroxene-plagioclase assemblages, in the more deformed regions these igneous microstmctures are recrystallised to metamorphic aggregates of clinopyroxene-orthopyroxene-plagioclase with or without hornblende, biotite, clinozoisite, K-feldspar and quartz (Clarke et al, 2000). FIORDLAND WBSTLAKD

undiri

l-TOSJDLAND Figure 1: Fiordland geology & temperature estimations.

The development of garnet-bearing replacement textures in the Western Fiordland Orthogneiss confirm a high-P history and the transition of two pyroxene assemblages to gamet-clinopyroxene assemblages has been linked to burial from mid crustal levels (c. 25 km) to lower crustal levels (c. 45 km) (Clarke et ai, 2000; Daczko et al, 2001b). The Western Fiordland Orthogneiss has a distinctive chemistry, similar to that of the Separation Point Batholith that outcrops to the north-east and east, consistent with it being the metamorphosed root of the Cretaceous Separation Point Batholith. Data - Geochemistry The geochemistry of the Westem Fiordland Orthogneiss has been studied extensively in northern Fiordland by various authors (Mattinson et aL, 1986; Daczko et al., 2002; Hollis et aL, 2004). More recent studies have concentrated on the central localities in Fiordland, including Doubtful Sound. New geochemical data and the use of discrimination diagrams help to classify the main rock types in the Westem Fiordland Orthogneiss at Lake Grave, Doubtful Sound and other Mesozoic orthogneiss units located in and around Breaksea Sound. A secondary aim of geochemical analysis was to determine if these mafic intrusive rocks have adakitic affinities and if so, do they have a distinct chemistry which links them to melting of a mafic source at high pressure.

39 Lake Grave Samples from Lake Grave are classified as gabbroic diorites that plot in the alkaline field of the plutonic rock classification of Le Maitre et al (1986) (these rocks are now metamorphosed to gabbroic-dioritic gneiss). Si02 contents range from 53.7 - 54.8 wt%, AI2O3 from 17.3-18.1 wt% and total alkalis (Na20+K20= 5.03 - 5.64 wt%) are high. Gabbroic gneiss samples forming the Westem Fiordland Orthogneiss in this location are alkali-calcic, with a Peacock Index of 56.01 and metaluminous (ASI-0.8). Doubtful Sound Samples from Doubtful Sound include gabbroic gneisses, syeno-dioritic gneisses and dioritic gneiss based on the plutonic rock classification of Le Maitre et al (1986). Samples plot in the alkaline and calc-alkaline fields on this diagram. The rocks at this site are characterised by moderate Si02 contents, ranging from 48.55 - 56 wt%, high AI2O3 between 17.67 - 20.42 wt% and high total alkalis (Na20+K20= 4.01 - 7.29 wt%). The gabbroic and dioritic gneiss found in this location are also alkali-calcic, with a Peacock Index of 56.25 and are metaluminous (ASI-'0.8). Breaksea Sound Samples in Breaksea Sound also include gabbroic-dioritic, syeno-dioritic and dioritic gneisses that also plot in the alkaline and calc-alkaline field on the TAS diagram. Breaksea Sound lithologies are characterised by moderate Si02 content, ranging from 52.03 - 61.83 wt%, high AI2O3 between 17.97-19.35 wt% and high alkahs (Na20+K20= 5.05 - 6.7 wt%). The orthogneiss at this locality are calc-alkalic, with a Peacock Index of 57.35 and are also metaluminous (ASI~0.8). Previous studies (Daczko et al, 2002; Stevenson et al, 2005) have suggested that the Westem Fiordland Orthogneiss is the metamorphosed root of the Separation Point Batholith, primarily based on similarities in the geochemistry of both suites. Muir et al (1998) provided evidence of crystallisation of the Separation Point Batholith from an adakitic melt. To further test whether there is a geochemical relationship between the Westem Fiordland Orthogneiss and the Separation Point Batholith, discrimination diagrams are employed to ascertain whether the Westem Fiordland Orthogneiss has an adakitic signature, similar to that of the Separation Point Batholith. The results show that both groups share similar chemistry consistent with formation by partial melting of mafic lower continental crust producing these adakitic melts (Figure 2). KEY

2G0 C1C2

700

Lake Gra^B # Gabbroic whole rock iGNS • Breaks«a # Sound who^ rock

600

500 >

400

Figure: 2

300

200-

Island anc: andDsite-dacite-rihyolite field MORS field

100-'

Figure 2: Geochemical field for adakites.

40 Thermobarometry and Mineral Chemistry Mineral chemistry determined by electron microprobe analysis was used in thermobarometric calculations. Minerals defining early igneous assemblages at Lake Grave indicate temperatures of ~750°C (Figure 1). Thermobarometric results of gamet-clinopyroxene granulite facies assemblages in the garnet reaction zones indicate development in the deep crust (P=10-13.70 kbar) at elevated temperatures (T >780'C) across Fiordland. The results of this study show that peak pressures were similar across the entire Fiordland block. Very consistent pressure estimates of about 15 kbar have been obtained from the Breaksea Gneiss (Figure 3). The presence of omphacitic pyroxene in this same sample suggest these high pressure estimates are reliable and this unit may be the deepest exhumed rock yet found in Fiordland (Figure 3).

WESHANB

!Vf csozoic Arc:

^ UfjJjfT, Mim^Jvit^. MidMi

li&t

Piiloozciic BdscniL'nl

Figure 3: Pressure estimates, Fiordland.

Discussion Garnet Reaction Zones: This study is the first to describe the development of garnet reaction zones in diffuse patch-like domains adjacent to trondhjemitic veins which may represent early stages in the formation of garnet granulite across Fiordland. The lower abundance of homblende as the hydrous phase or an accessory phase such as clinozoisite, may have controlled the weaker patchy development of garnet reaction zones at Lake Grave. Most veins at the centre of garnet reaction zones are dominantly plagioclase feldspar, inconsistent with them representing a crystallised melt sourced from the mafic migmatites. The veins are better interpreted as a cumulate missing much of the original trondhjemitic liquid that has intruded to structurally higher levels, leaving behind the "cumulate" plagioclase feldspar. This study is the first to recognise orthopyroxene in true trondhjemitic veins, suggesting again that the

veins at Lake Grave may reflect the rare preservation of a short-lived, early stage in the development of the garnet reaction zones, where the liquid rock (sourced from the mafic migmatites) is preserved. Conclusions Bulk rock chemical data presented in this study confirms the suggestion of Stevenson et al (2005) that the Western Fiordland Orthogneiss have adakitic chemistry as a result of formation in a deep part of a mafic crust where garnet was part of the restite. The term adakitic is recommended to cover igneous rocks that have the main characteristic of adakites but fall outside the restrictive ranges of some trace elements in the original definition of Martin (1999) [e.g. Ni (20-40 ppm) and Cr (30-50 ppm)]. The mineral chemistry/end member calculations defined omphacite (high temperature and pressure pyroxene) for the first time in Fiordland, suggesting that the Breaksea Gneiss may represent the deepest exhumed level in Fiordland. Anti-perthite textures in plagioclase in veins at Lake Grave are consistent with a high-temperature feldspar forming from a rarely preserved trondhjemitic liquid. Trace element pattems of garnet from veins that cut the high-P mafic granulites are chemically distinct from hornblende in the host gneiss, suggesting that the leucosome veins invaded the gabbroic gneiss and therefore form an injection migmatite. Furthermore the trace element chemistry of the vein-hosted gamet is similar to that reported from northem Fiordland (Schroter et al.^ 2004) and suggests that a more dioritic composition may have been the source for the injecting leucosome veins. Thermobarometric P-T estimates constrain peak metamorphic pressures of up to 15.5 kbar consistent with the presence of high-pressure omphacite in equilibrium with gamet and the gamet granulite assemblages. This suggests that the Fiordland block was coherently exhumed as one large block of lower crust. The systematic approach taken in this study has led to a better understanding of the plutonic and metamorphic evolution of the high-P mafic migmatites in Fiordland. Plutonism in Fiordland is dominated by mafic rocks which underwent migmatisation and preserved partial melting textures in which water undersaturated partial melting of the lower continental crust generated more intermediate magmas. Petrographic, geochemical and field descriptions show that the migmatites are mainly dioritic in composition. Partial melts formed during migmatisation of the surrounding country rocks producing trondhjemitic veins and dykes that invaded country rock gabbroic gneiss inducing the limited recrystallisation to gamet granulite. References

Clarke G.L., Klepeis, K.A., and Clarke, G.L., 2000. Cretaceous high-P granulites at Milford Sound, New Zealand: metamorphic history and emplacement in a convergent margin setting. Journal of Metamorphic Geology, 18: 359-374 Daczko, N. R., Klepeis, K. A., and Clarke, G.L., 2001. Evidence of Early Cretaceous colHsional-style orogenesis in northem Fiordland, New Zealand and its effects on the evolution of the lower crust. Journal of Structural Geology, 23 (4): 693.

42 Daczko, N. R., Klepeis, K. A., and Clarke, G.L., 2002. Thermomechanical evolution of the crust during convergence and deep cmstal pluton emplacement in the Western Province of Fiordland, New Zealand. Tectonics, 21(4). Hollis, J. A., Clarke, G. L., Daczko, N. R., Klepeis, K. A., and Ireland, T. R., 2003. Geochronology and geochemistry of high-pressure granulites of the Arthur River Complex, Fiordland, New Zealand: Cretaceous magmatism and metamorphism on the palaeo-Pacific Margin. Journal ofMetamorphic Geology, 21(3): 299. Kimbrough, D. L., Tulloch, A.J., Coombs, D.S., Landis, C.A., Johnston, M.R., and Mattinson, J.M., 1994. Uranium-lead zircon ages from the Median Tectonic Zone, New Zealand. New Zealand Journal of Geology and Geophysics, 37: 393-419. Klepeis, K.A., Clarke, G.L., and Rushmer, T., 2003. Magma transport and coupling between deformation and magmatism in the continental lithosphere, GSA Today, Jan 2003: 4-11. Martin, H., 1999. Adakitic magmas: modem analogues of Archean granitoids. Lithos, 46: 411-429. Mattinson, J. L., Kimbrough, D.L., and Bradshaw, J.Y., 1986. Western Fiordland orthogneiss: Early Cretaceous arc magmatism and granulite facies metamorphism. New Zealand. Contributions to Mineral Petrology, 92: 383392. Muir, R.J., Ireland, T.R., Weaver, S.D., Bradshaw, J.D., Evans, J.A., Eby, G.N., and Shelley, D., 1998. Geochronology and geochemistry of a Mesozoic magmatic arc system, Fiordland, New Zealand. Journal of the Geological Society, London, 155: 1037-1053. Schroter, F.C., Stevenson, J.A., Daczko, N.R., Clarke, G.L., and Pearson, N.J., 2004. Trace element partitioning during high-P partial melting and melt-rock interaction; an example from northern Fiordland, New Zealand. Journal ofMetamorphic Geology, 22: 443-457. Stevenson, J.A., Daczko, N.R., Clarke, G.L., Pearson, N., and Klepeis, K.A., 2005. Direct observation of adakite melts generated in the lower continental crust, Fiordland, New Zealand. Terra Nova, 17: 73-79. Wandres, A. M., Weaver, S.D., Shelley, D., and Bradshaw, J.D., 1998. Change from calc-alkaline to adakitic magmatism recorded in the Early Cretaceous Darran Complex, Fiordland, New Zealand. New Zealand Journal of Geology and Geophysics, 41: 1-14.

43 HFSE fractionations in a continental, intraplate volcanic system, Southeast Queensland, Australia Benjamin E. Cohen*, Kurt M. Knesel, and Paulo M. Vasconcelos Department of Earth Sciences, University of Queensland, St Lucia 4072 *Email: b.cohen@uq.edu.au Introduction The high-field-strength-element (HFSE) pairs, Zr-Hf and Nb-Ta, represent powerful tools in the investigation of geochemical processes involved in the differentiation of the upper mantle and growth of the continental crust. It is generally thought that the identical charge and very similar ionic radii make it difficult to fractionate the ratios of these elements during typical differentiation processes (e.g., Jochum et al., 1986). However, with the improvement in analytical techniques over the last 10 years, high-precision measurements in both experimental and natural systems have revealed significantly different partition coefficients for Zr-Hf and Nb-Ta in various upper-mantle and crustal systems (e.g., Johnson, 1998; Klein et al, 2000) and striking examples of Zr-Hf and Nb-Ta fractionations among different crustal and mantle reservoirs (e.g., David et al, 2000; Weyer et al, 2003). In hght of these results, we investigate here Zr-Hf, Nb-Ta and Zr-Nb variations among a suite of well-characterized mafic to intermediate lavas from a classic, continental, intraplate volcanic setting to examine the effects of source composition and mixing, partial melting, and crystal fractionation on HFSE systematics. Cenozoic magmatism in southeast Queensland Oligocene volcanic rocks in southeast Queensland (Fig. 1) comprise voluminous mafic lavas along with lesser volumes of metaluminous trachytes, peralkaline rhyolites (comendites) and peraluminous rhyolites (Fig. 2). Recent "^^Ar/^^Ar analyses (Cohen et aL, 2005) indicate the mafic lavas erupted first (31.0 to 27.5 Ma), followed closely by the metaluminous trachytes and the comendites (27.0 to 26.3 Ma). These ages, plus others from Fraser Island and Flinders Peak, define a north-south progression along a 320 km NNE-lineament that is consistent with a hotspot origin for east Australian volcanism and yields a plate velocity, relative to a fixed hotspot reference, of 58 ±10 mm/yr. Petrographic, geochemical and isotopic observations indicate that the metaluminous trachytes and comendites were derived via extensive fractional crystallisation from parent magmas similar to the mafic lavas (Ewart et aL, 1988 and references therein). In contrast, the peraluminous rhyolites, which were erupted at about 28 Ma, appear to have been derived largely by partial melting of the crust. The mafic lavas studied here can be divided into two groups (Figs 2-4). In the Maleny area (Fig. 3), the stratigraphically lowest lavas (type 1) are characterized by lower concentrations of incompatible elements (with respect to peridotite melting), at a given MgO content, relative to lavas in the upper part of the section, as well as those from Mt Mee and Mt Pinbarren (type 2). No simple differentiation process, such as the degree of partial melting in the source, fractional crystallization or crustal contamination, can easily relate the more enriched, type-2 lavas to the type-1 lavas (Cohen, 2002). Nevertheless, within each suite, elements that are compatible during differentiation, such as Ni, Cr, and Sr, decrease and incompatible elements,

44 such as Rb, Zr, increase with decreasing MgO (Fig. 4). These trends, along with decreasing Ca0/Al203, are consistent with up to 20%fractionationof oHvine > pyroxene > plagioclase. An additional 40-60%fractionationof plagioclase, clinopyroxene, olivine, and minor apatite and magnetitefromparent magma(s) similar in composition to the type-2, enriched lavas is required to produce the major and trace element characteristics of the metaluminous trachytes. b N

^ " 153-E / / \.

jlometers

I Noosa J 26.5°S

55 60 65 70 SI02(wt% anhydrous)

Fig. 1: (a) Distribution of Cainozoic volcanic rocks in eastem Australia, with the location of the study area, (b) Detailed map of the Sunshine Coast hinterland, with the main study locations of the metaluminous trachytes and peralkaline rocks from the Glasshouse Mountains (GM) (solid black areas) and the Maleny mafic lavas (unfilled black outlines).

Fig. 2: Total-alkalis vs. Si02 for rocks from the study area, showing the bimodal nature of Oligocene magmatism in southeast Queensland.

Top, Maleny (Type 2)

La Ce Pr Nd

Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Element

MgO(wt%)

Fig. 3: Representative REE patterns for mafic lavas. Type-1 lavas are restricted to the lower and middle portions of the Maleny lava sequence. Type-2 lavas are more widespread, comprising the upper Maleny flows and all outlying mafic lavas.

Fig. 4: MgO vs. Sr for mafic lavas. Type2 lavas have a considerably higher concentration of Sr for the same MgO, and extend to more evolved compositions than the type-1 lavas. Filled symbols are ICPMS; open symbols are XRF.

45 HFSE Fractionations The well-characterized compositional spread of these lavas provides an excellent framework within which to assess the effects of differentiation processes on HFSE behaviour. Towards this end, we discuss below Zr-Hf, Nb-Ta and Zr-Nb variations, determined by high-precision ICP-MS, among the mafic to intermediate lavas from this suite. Zr-Hf Zr/Hf ratios in the mafic lavas are superchondritic, with type-1 lavas (7.8-7.6 wt% MgO) clustering between 39.5 and 40.5 and the more enriched, type-2 lavas (7.8-2.9 wt% MgO) ranging from about 41.5 to 48.5. The metaluminous trachytes overlap with those of the most fractionated type-2 lavas and extend to ratios as high as 52.4. A positive correlation of Zr/Hf with Rb and Th (not shown), and a negative correlation with Sc (Fig. 5) are consistent with differentiation of the trachytes via fractionation crystallization of type-2 mafic lavas, with the later controlled by precipitation of clinopyroxene (e.g., David et al, 2000). Nb-Ta The mafic lavas show near-chondritic Nb/Ta ratios of 16.0-17.0, similar to that of the presentday depleted mantle. In contrast, the metaluminous trachytes are characterized by Nb/Ta ratios ranging from 17.8 to 19.1. A moderately well-defined correlation (r^=0.82, not shown) between Nb concentrations and Nb/Ta ratios for type-2 mafic lavas and metaluminous trachytes indicates that Nb/Ta can also be fractionated during differentiation. Removal of about 2-3% titanomagnetite is required to produce the observed variations (Fig. 6) and is consistent with least-squares modelling of major-element compositions.

Sc (ppm)

Fig, 5: Sc vs. Zr/Hf ratio for mafic lavas and metaluminous trachytes. Sc, which is compatible in clinopyroxene, is well correlated with Zr/Hf M2 was excluded from the regression as it contains abundant megacrysts, including pyroxenes. Symbols are as shown in Fig. 2.

150 50 "^00 Fig. 6: Rb vs. Nb/Ta ratio for mafic lavas and metaluminous trachytes. Curves show evolution of residual melt following fractionation of plagioclase (50%), olivine (20%) and clinopyroxene (27%), with (circles) and without (crosses) magnetite (3%) from two different type-2 starting compositions. Symbols are as shown in Fig. 2. ou/ 1 Rb (ppm)

400 (a)

600

A-AM2 A 300-

a 200H

100-

A A

I J^

70 % crystallised X

200

^ oO

ICP-MS data (excluding M2) R2 = 0.99

MORB

^ 400

I 20

30 Nb (ppm)

1 20

/ <

(to

r/ °

46 /

X50% crystallised

o ^

Xo - OtFract nal crystbasal alistats'(iXRF) on herioSEQId' 80 40 60 Nb (ppm)

Fig. 7: (a) Nb vs. Zr for mafic lavas. Tie lines join XRF and ICP-MS analyses for the same sample. M2, which is highly evolved (2.92 wt % MgO), shows potential effects of crustmagma interaction, (b) Nb vs. Zr for other southeast Queensland mafic rocks, which also show a strong correlation between Zr and Nb over a wide range of concentrations, which cannot be accounted for by fractional crystallisation alone. Curves show the effects of fractionation of olivine (50%), clinopyroxene (30%), and plagioclase (20%). Pb-isotopic evidence (Ewart et al 1988) suggests that the two samples with high Zr (arrowed) have been contaminated by the crust. Zr-Nb Zr and Nb are also well correlated for both our ICP-MS and older XRF data for the mafic lavas (Fig. 7), with the exception of three samples showing petrographic and geochemical evidence for crustal contamination (the effect of which is to increase the Zr/Nb ratio). Similar to the Zr-Hf and Nb-Ta variations discussed above,fi-actionalcrystallisation can account for some of the range of the correlated Zr and Nb concentrations among the variably fi-actionated mafic lavas. However, at comparable MgO contents, the distinction between the type-1 and type-2 lavas remains. Given the differences in bulk distribution coefficients for Zr and Nb during melting in the upper mantle, the Zr-Nb relationships among the most primitive samples in the two suites cannot be explained simply by variable degrees of melting (e.g., Kamber and Collerson, 2000). We show that these variations require mixing between depleted and enriched end-members. These are best represented by a depleted upper mantle and an EMllike component, respectively. Type-2 lavas reflect a greater contribution fi'om the enriched end-member, a scenario that is consistent with the general trace element and isotopic characteristics of these lavas (Ewart et al, 1988 and references therein). References

Cohen, B. E., 2002. Geochronology and geochemistry of southeast Queensland Tertiary volcanism. B.Sc. (Hons.), University of Queensland, Brisbane (unpubL). Cohen, B. E., Vasconcelos, P. M., and Knesel, K. M., 2005. ^^Ai/^^Ai constraints on the timing of Oligocene intraplate volcanism in southeast Queensland. Australian Journal of Earth Sciences (submitted). David, K., Schiano, P., and Allegre, C. J., 2000. Assessment of the Zr/Hf fractionation in oceanic basalts and continental materials during petrogenetic processes. Earth and Planetary Science Letters, 178: 285-301. Ewart, A., Chappell, B. W., and Menzies, M. A., 1988. An overview of the geochemical and isotopic characteristics of Eastem Australian Cainozoic volcanic provinces. In: Menzies, M. A., and Cox, K. (eds).

47 Oceanic and Continental Lithosphere: Similarities and Differences. Journal of Petrology^ Special Publication: pp.255-274. Jochum, K. P., Seufert, H. M., Spettel, B., and Palme, H., 1986. The solar system abundances of Nb, Ta and Y, and the relative abundances of refractory lithophile elements in differentiated planetary bodies. Geochimica et Cosmochimica Acta, 50: 1173-1183. Johnson, T. M., 1998. Experimental determination of partition coefficients for rare earth and high-field-strength elements between clinopyroxene, garnet, and basaltic melt at high pressures. Contributions to Mineralogy and Petrology, 133: 60-68. Kamber, B. S., and Collerson, K. D., 2000. Zr/Nb Systematics of ocean island basalts re-assessed - the case for binary mixing. Journal of Petrology, 41: 1007-1021. Klein, M. H.-G. S., Seek, H. A., and Shimizu, N., 2000. Experimental partitioning of high field strength and rare earth elements between clinopyroxene and garnet in andesitic to tonalitic systems. Geochimica et Cosmochimica Acta, 64: 99-115. Weyer, S., Munker, C., and Mezger, K., 2003. Nb/Ta, Zr/Hf and REE in the depleted mantle: implications for the differentiation history of the crust-mantle system. Earth and Planetary Science Letters, 205: 309-324.

48 The Hunter Ridge, SW Pacific: hot subduction, slab edge effects and the boninite - adakite - high-Nb basalt association Anthony Crawford, Leonid Danyushevsky, Roman Leslie, Sofia Tetroeva, Trevor Falloon, and Alicia Verbeeten. Centre for Ore Deposit Research, University of Tasmania, GPO Box 252-79, Hobart, Tasmania, Australia 7001 The Hunter Ridge (aka the Hunter Fracture Zone), a well-defined curvilinear belt of ridge/trough topography extending from Fiji to the southern Vanuatu arc, has been generally considered to mark the trace of a transform fault system linking the opposite-facing Tongan and Vanuatu arc systems, although several authors have proposed that it may have accommodated some oblique subduction of the South Fiji Basin lithosphere. The absence of an inclined seismic zone and shallow thrust-type focal mechanisms suggest that currently the Hunter Ridge is not accommodating plate convergence. The regional plate kinematic model of Auzende et al (1996) shows initiation ~7 Ma of a major E-W orientated spreading ridge across the North Fiji Basin. Spreading on this new ridge system is considered to have triggered subduction along the Hunter Ridge, which continued until a major reorganisation of spreading geometry in the North Fiji Basin occurred Ma, broadly synchronous with the initial opening of the Lau Basin. The southern (Matthew and Hunter Iss) and northem (Kadavu Is. Group of Fiji) ends of the Hunter Ridge are dominated by rocks that have geochemical signatures indicating that slab melts were involved in their petrogenesis. However, modem high-Mg andesites from Matthew and Hunter Islands are considerably more primitive compositions than the typical andesitic to dacitic lavas (3-1 Ma) that form Kadavu, probably reflecting thickness of the underlying crust. Kadavu lavas have passed through the thickened arc (Eocene - Pliocene) crust, whereas the high-Mg andesites at the southern end of the ridge have erupted upon young backarc basin crust. Kadavu is dominated by high-K adakitic andesites and dacites, and an even younger unusual high-K, relatively high-Nb basaltic suite (Ngaloa Group). Major and trace element geochemistry of the adakitic andesites strongly support their derivation via partial melting of subducted South Fiji Basin oceanic crust and limited interaction of these melts with lithospheric mantle during ascent to eruption. Ngaloa Group basalts also show some important eclogitic slab-melt features (LanAT3n=18-27 and low Y (4-22ppm) contents, high Sr (2000-3000ppm) and SrA" (93-205) values similar to those of the adakites). However, the mafic nature of Ngaloa Group volcanics and their primitive phenocryst compositions (abundant olivine F089.91 phenocrysts) preclude their derivation solely from partial melting of subducted basaltic crust, and require significant involvement of mantle wedge peridotite. Some compositional features of these volumetrically minor Ngaloa Group basalts, particularly their high Na20 (3.7-4%), P2O5 (0.7-0.8%), K2O (1.7%), strikingly low FeO* (6.1-6.8%) contents, and their strong enrichment of incompatible elements, are typical features of low degree partial melts of peridotite. It is hypothesised that siliceous melts (andesites or dacites) broadly similar to those constituting the associated adakitic suites were produced by melting of eclogite in the subducted slab, but reacted with, and stalled in the overlying mantle wedge

49 peridotite. Subsequent shallow partial melting of this slab melt-metasomatised peridotite produced the Ngaloa parental magmas. Both the adakitic lavas and Ngaloa basalts have ^^Sr/^^Sr values from 0.70290-0.70335, and ^^^Nd/^^Nd values of 0.51301-0.51304, falling within or very close to the high ^^Sr/^^Sr - low ^'^^Nd/^'^Nd end of the Pacific MORB field. Pb isotopic compositions for the Kadavu adakites plot close to or on the NHRL and at the more radiogenic end of the Pacific MORB range. Involvement of old continental crust or pelagic sediments in the petrogenesis of these lavas can be confidently ruled out. Pb isotopic compositions of Ngaloa basalts are very uniform ^^¥^^Pb=15.52-15.53; H^>b=38.26-38.34), and plot very close to the NHRL, either just within or very slightly below (^^^Pb/^^^Pb) the field defined for Pacific MORB. Importantly, the Ngaloa basalts have very similar Pb isotope compositions to basalts from the subducting South Fiji Basin - Loyalty Basin, and the melts/fluids involved in both adakite and Ngaloa high-Nb basalt generation had not suffered significant reaction with seawater, given the MORB-like Sr isotopic values of these lavas. Despite the MORB-like Nd-Sr-Pb isotopic compositions of these adakitic and basaltic lavas, N-MORB normalised multi-element diagrams show a significant positive Pb anomaly relative to LREE. Clearly, dehydration fluids and/or siliceous melts of subducting oceanic, have the ability to mobilised and concentrate (relative to LREE and HFSE) Pb from the basaltic portion of subducted oceanic crust with little or no modification of Pb isotopic ratios. Furthermore, all Pb in the erupted adakite - high-Nb basalt lavas has been sourced from the subducted oceanic crust, as the upper plate mantle wedge in the N Fiji Basin carries a significant 'Indian Oceanic' realm isotopic fingerprint. No Indian-type Pb was scavenged by these magmas during their passage through the N Fiji Basin lithospheric mantle. Between Kadavu and the southern end of the Hunter Ridge exposed on Matthew and Hunter islands, dredged rocks include primitive typical- and low-Ti arc tholeiite basalts, high-Ca boninite lavas, and some primitive basahs with adakitic affinities not unlike the Ngaloa Group basalts. This unusual magmatic association reflects magma genesis in an abnormally hot environment where old, cold South Fiji Basin crust was subducted (3-7 Ma) immediately beneath the southward propagating major spreading centre in the upper North Fiji Basin plate. To the east of Kadavu, in the Koro Sea, basalts with subdued OEB affinities have been erupted during the last several million years. Isotopically, these lavas are significantly less radiogenic than those from the Samoan plume 300 km to the northeast, and overlap the field for North Fiji Basin lavas. We suggest that the so-called Fijian OIB are relatively low-degree melts of broadly MORB-type regional asthenospheric mantle, such as that ascending in ridge-type extensional zones in the N Fiji Basin to the west and the northem Lau Basin to the east. Rollback of the subducting slab beneath the Hunter Ridge occurred as a response to growth of the upper (N Fiji Basin) plate until about 3 Ma, providing 'room' at the northem termination of the slab beneath the Koro Sea for invasion by asthenospheric mantle providing BABB to the Lau and N Fiji Basins. Limited, and possibly slightly deeper melt generation from this mantle provided the 'OIB' basalts that ring the Koro Sea.

Figure 1: Location of the Hunter Ridge in the SW Pacific

High-pressure mafic migmatites, Fiordland, New Zealand: does migmatisation promote recrystallisation to garnet granulite? Nathan R. Daczko^and Geoffrey L. Clarke^ ^GEMOC Key Centre, Department of Earth and Planetary Sciences, Macquarie University, Sydney, NSW 2109, Australia ^School of Geosciences, University of Sydney, NSW 2006, Australia Mafic migmatitic structures occur in regular networks across a ---25 km-wide, >150 km-long belt in Fiordland, New Zealand. Commonly, narrow anorthositic dykes and less commonly trondhjemitic dykes cut two-pyroxene-homblende-bearing orthogneiss and form a type of injection migmatite. Patchy zones, <10 cm-wide, show spatially restricted recrystallisation and dehydration to garnet granulite adjacent to the dykes. The source for the leucosome in the injection migmatites is rarely exposed, but locally may be spatially linked to the injection migmatites. Whole-rock and in-situ mineral major and trace element data indicate the recrystallisation to garnet granulite was mostly isochemical and also indicate a geochemical link between the sites of partial melting and the injection migmatites. Mineral chemistry analyses and thermobarometric calculations indicate granulite facies conditions of T > 750°C and P = 12-14 kbar accompanied the migmatisation. This study provides direct observation of the processes associated with the generation and migration of felsic magma in the lower crust. Regional Geology The geology of the South Island of New Zealand is divided into three domains. A belt of rocks referred to as the Median Tectonic Zone or Median Batholith separates Eastern and Western Provinces. The Western Province contains extensive Lower Palaeozoic metasediments that are cut by Devonian and Carboniferous granitoids (Fig. 1). Rocks of the Median Tectonic Zone and the Western Province were intruded by plutons of the 126-105 Ma Western Fiordland Orthogneiss / Separation Point Suite. The Arthur River Complex is a belt of granulite facies orthogneiss that lies at the boundary between the Median Tectonic Zone and Westem Province rocks in northern Fiordland; it has been grouped with the Median Tectonic Zone (Hollis et al, 2003). The Arthur River Complex is intruded by the Westem Fiordland Orthogneiss at its southem-most extent.

52 New Zealand Fiordland

Fiordland WFB

KMiljbrd j^omid

EFB SWFB

Figure 1: Fiordland is located in the SW of the South Island of New Zealand, east of the Alpine Fault. The Fiordland block comprises the (i) western Fiordland belt (WFB), including the Palaeozoic Tuhua Sequence and Cretaceous Western Fiordland Orthogneiss; (ii) eastern Fiordland belt (EFB), including the Median Tectonic Zone and Arthur River Complex; and (ii) southwest Fiordland block, including low-grade Palaeozoic units. The Pembroke Granulite is located directly north of Milford Sound. Introduction Garnet granulite (Fig. 2) is commonly associated with anorthositic veins that cut twopyroxene-amphibole-bearing assemblages in the Arthur River Complex and Western Fiordland Orthogneiss (Blattner, 1976; Oliver, 1977; Bradshaw, 1989; Daczko et al, 2001). Data drawn from the field setting and petrography of garnet granulite facies assemblages that occur in unique circumstance throughout Fiordland are central to many of the interpretations of the tectonic evolution of the Fiordland region. Excellent exposures in recently glaciated rocks that experienced only partial recrystallisation have lead to a good understanding of metamorphic reactions involved in the change from two-pyroxene-amphibole-bearing orthogneiss to garnet granulite assemblages (Blattner, 1976; Oliver, 1977; Bradshaw, 1989). However, the tectonic and metamorphic processes that lead to the localised development of garnet granulite have yet to be resolved. In Fiordland, the garnet reaction zone structures occur between Milford and Breaksea Sounds, over an area of >150 by -25 km. One of the best-exposed examples occurs in the Pembroke Valley, NW of Milford Sound (Fig. 1), where vertical and planar garnet reaction zones cut granulite facies gabbroic to dioritic gneisses in spectacular rectilinear patterns. The various models for the development of the garnet granulites in Fiordland (Blattner, 1976; Oliver, 1977; Bradshaw, 1989; Daczko et al, 2001) conceptually build from an open or closed system approach to the study of migmatites, and related gamet reaction zones, which may or may not have involved the presence of a silicate melt. Quantitative assessments of migmatite petrogenesis, constrained by geochemical and isotopic analysis, converge on four main mechanisms: (i) injection of a sihcate Hquid (Sederholm, 1907, 1934; Buddington, 1948; Olsen, 1983); (ii) sub-solidus metamorphic differentiation (Eskola, 1932; Sawyer and Robin, 1986); (iii) metasomatism (Olsen, 1985); and (iv) partial melting (Holmquist, 1921; Mehnert, 1968, 1973; Brown, 1973; Sawyer, 1987). Each process may operate solely, or in combination with one or more other processes.

Figure 2: NE-trending, narrow, anorthositic dyke and associated garnet reaction zone cutting compositional layering in gabbroic gneiss in the Pembroke Granulite.

Major physical and chemical relationships In gabbroic gneiss of the Pembroke Valley, cm- to decimetre-wide anorthositic dykes are surrounded by fine-grained garnet granulite that variably recrystallised a host two-pyroxene hornblende granulite at conditions of T>750°C and P ^ 4 kbar (Clarke et al, 2000). The anorthositic dykes cut contacts between gabbroic gneiss, dioritic gneiss and ultramafic gneiss without host rock displacement (Fig. 2), and have Sr and Nd isotopic ratios indistinguishable from those of their hosts. Garnet reaction zones flanking the dykes terminate at gabbroicdioritic gneiss contacts (fig. 2d, Daczko et aL, 2001); at such locations, the anorthositic dyke is continuous with a planar felsic segregation commonly involving a septum of coarse-grained garnet surrounded by anorthositic leucosome. Klepeis and Clarke (2004) have summarised regional tectonic relationships of the evolving Cretaceous orogen, involving migmatite formation in the root of a mafic arc undergoing convergence with a continental margin. Peak conditions established for rocks of the Pembroke Valley are at temperatures well above their water-saturated solidi (e.g. Rushmer, 1991; Lopez and Castro, 2001) and the limited development of migmatitic structures require that there was no free water at peak conditions. Pervasive garnet and spatially related leucosome throughout dioritic gneiss have been used to infer that the migmatite was produced by closed-system partial melting (Daczko et al, 2001). In this interpretation, garnet was a peritectic product of the melting step, balancing Fe and Mg components supplied by reactants hornblende and clinozoisite. Textures within the leucosomes are consistent with the interpretation that the dioritic gneiss partially melted, using the microstructural criteria of Sawyer (1999). Limited retrogression of the peak assemblage suggests that partial melt must have escaped the dioritic gneiss (e.g. Waters and Whales, 1984; White and Powell, 2002). Sites of inferred melt escape are represented by the planar leucosome structures that continue in surrounding gabbroic gneiss. Work completed prior to Daczko et al (2001) had ignored these intermediate migmatites. Gabbroic gneiss, the common rock type in the Pembroke Valley, has 3 main components (Fig. 2). Anorthositic dykes that have sharp boundaries and large euhedral grains of gamet and, less commonly, diopside, hornblende or scapolite cut the host two-pyroxene-homblende gneiss. Small proportions of phengitic white mica, kyanite and clinozoisite may also be present. Gamet reaction zones mostly separate these two components; the width of the zones mostly

54 reflects hornblende mode of the host, and sometimes the gamet reaction zones lack an anorthositic core. Daczko et al (2001) observed that gamet in the anorthositic dykes is texturally distinct to gamet in the gamet reaction zones but similar to gamet in migmatitic stmctures in the dioritic gneiss. This observation led to the prediction that the anorthositic dyke v^as sourced from the partial melting of the dioritic gneiss. In a detailed study of the trace element content of the main silicates, Schroter et al (2004) confirmed this prediction: gamet in anorthositic dykes hosted by gabbroic gneiss has REE contents that distinguish it from gamet in adjacent gamet reaction zones, but it is identical to peritectic gamet in the dioritic gneiss that may be tens to hundreds of metres away. Gamet in anorthositic dykes (hosted by gabbroic gneiss) is also subtly zoned in pyrope and grossular components, and systematically zoned in REEs. The zoning profile is most pronounced in the heavy REEs, and is similar to spessartine growth zoning profiles in low-grade gamet from elsewhere (e.g. Tracy, 1982; Marmo et al, 2002). With the exception of the rectihnear networks, gabbroic gneiss lacks pervasive leucosome development and peak conditions are inferred to have been at lower T than the solidus for such less fertile compositions (Daczko et al, 2001). The chemical and textural distinctions between gamet in dyke and gamet reaction zone, the sharp boundaries of the dykes, and the lack of petrographic evidence for relics of the host rock in the dyke have been used to support the interpretation that the anorthositic dyke was injected into the gabbroic gneiss. The chemical and spatial links are consistent with the mobilisate having been sourced from the dioritic gneiss (fig. 5 Daczko et al, 2001). The anorthositic dykes in the gabbroic gneiss thus form an injection migmatite, albeit an unusual one with injected dykes forming the rectilinear network. Discussion Granulite facies assemblages that record T>750°C are a common feature in high-grade Precambrian terranes, and require that water activity at peak conditions was less than one to avoid pervasive partial melting (e.g. Lopez and Castro, 2001). Dehydration is commonly inferred to have occurred in response to a prograde reaction sequence or progressive melt removal (e.g. White and Powell, 2002). Small proportions of scapolite in mafic rocks can reflect the consequent enrichment of CO2 in any remaining fluid. In such circumstances, the degree of partial melting will be mostly controlled by the availability of hydrous minerals to react with quartz and plagioclase. Migmatitic stmctures preserved in dioritic gneiss from the Pembroke Valley are consistent with it having been previously richer in quartz and hydrous minerals, but the lack of a low-grade equivalent prevents any quantitative constraint on the proportion of melt removed. The lower alkali and silica content of the gabbroic gneiss would have made it less fertile at peak conditions. In this context, it is worthwhile evaluating what fluids might have induced metasomatism in the gabbroic gneiss to form the gamet reaction zones. A decision needs to be made as to whether the anorthositic dykes and gamet reaction zones were contemporaneous. Protoliths to the host rock were emplaced only a few million years before the partial melting, most probably into a mid or lower cmstal setting (Clarke et al, 2000; Mollis et al, 2003), but the period allows the processes to have been separate. Models proposed by Blattner (1976) and Bradshaw (1989) involving the ingress of carbonic fluids along a pre-existing vein network leading to dehydration of rocks adjacent to the fluid pathways suffer two main problems: (1) there is no confirmed source or sink for such fluids; and (2) the termination of the gamet

55 reaction zones at lithological boundaries of rocks that are mineralogically close to the gabbroic gneiss begs special cases of fluid flow. Continuity of the dykes across the gabbroicdioritic gneiss boundaries, at which point the garnet reaction zones terminate, qualitatively indicates a genetic link. Locations where garnet reaction zones lack a central anorthositic dyke may be terminations of a dyke in the third dimension, or sites of prior melt flux (Daczko etaL, 2001). Assuming a genetic link between dyke and gamet reaction zone, the causal fluid: (i) needs to have been capable of generating the sharp boundaries of the anorthositic dykes; (ii) had a radiogenic signature close to that of calk-alkaline hosts; (iii) been capable of inducing dehydration; and (iv) been capable of forming mineral assemblages that vary on a metre scale and include texturally-equilibrated gamet, diopside and homblende. Furthermore, the gametforming process in the anorthositic dykes needs to account for growth zoning preserved by REEs. The injection of a hydrous fluid along specific fractures would lead to the focussed partial melting of the gabbroic gneiss, as such fluid would immediately bring the rock to solidus conditions. A rapid, limited influx on a fracture network might form structures similar to those observed in the Pembroke Valley. However, geochemical data indicate that the anorthositic dykes are exotic to the gabbroic gneiss and little material other than water was removed from the gamet reaction zones. An altemative interpretation involves a specific fluid having formed both the anorthositic dykes and gamet reaction zone by metasomatism. This would require extensive alkali metasomatism of the host rock to form the dyke, and less intense metasomatism to form the gamet reaction zones. Such a fluid would need to be undersaturated in water, to avoid partially melting the host, and have had the unusual composition of being alkali-rich yet silica-poor, as the anorthositic dykes lack quartz. There are no petrographic relics of host rock in the dykes, nor any outgoing pathways of material demanded by the metasomatic origin. The observations best match the behaviour of one fluid: a silicate liquid formed by partially melting the dioritic gneiss (Daczko et al, 2001). However, the unusual metasomatic relationships in the gabbroic gneiss beg scepticism for the inferred genetic link. The common separation of silicate melts and hydrous fluids, appropriate to upper crustal conditions, may be inappropriate at deep crustal and upper mantle conditions (Shen and Keppler, 1997; Newton and Manning, 2003); with increasing pressure, the solubility of water in silicate melts, and the solubihty of silicate melt in hydrous fluids, increases. On-going work in this field will establish whether a silicate melt that behaved more like a hydrous fluid could have formed at conditions recorded by the Pembroke Granulite. Identical gamet reaction zones occur in Cretaceous rocks of the Kohistan arc, northern Pakistan (Yamamoto and Yoshino, 1998). Though distinguished by host rock composition and hence critical hydrous-mineral breakdown reactions, the dehydration structures in India and Sri Lanka are also steeply dipping and regionally persistent (Raith and Srikantappa, 1993). The persistence of the vertical network through westem Fiordland is consistent with comparatively rapid uplift of that terrane having not involved significant rotation of the preserved sites on a horizontal pole. The migration of silicate melt is a common feature at granulite facies conditions, though such rocks are commonly extensively deformed. On the basis of related links to structures in Pakistan, we prefer the interpretation that the Pembroke example is a well-exposed example of a common process, the evidence for which is usually destroyed by recrystallisation.

56 References Blattner, P., 1976. Replacement of hornblende by garnet in granulite facies assemblages near Milford Sound, New Zealand. Contributions to Mineralogy and Petrology, 55: 181-190. Bradshaw, J. Y., 1989. Early Cretaceous vein-related garnet granulite in Fiordland, Southwest New Zealand; a case for infiltration of mantle-derived CO2- rich fluids. Journal of Geology, 97: 697-717. Brown, M., 1973. The definition of metatexis, diatexis and migmatite. Proceedings of the Geological Association, 84: 371-382. Buddington, A. F., 1948. Origin of granitic rocks of the northwest Adirondacks. Geological Society of America Memoir, 28: 21-43. Clarke, G.L., Klepeis, K.A., and Daczko, N.R., 2000. Cretaceous granulites at Milford Sound, New Zealand; metamorphic history and emplacement in a convergent margin setting. Journal of Metamorphic Geology, 18: 359-374. Daczko, N.R., Clarke, G.L, and Klepeis, K.A., 2001. The transformation of two-pyroxene hornblende granulite to garnet granulite: simultaneous melting and fracturing of the lower crust, Fiordland, New Zealand. Journal of Metamorphic Geology, 19: 547-560. Eskola, P., 1932. On the principles of metamorphic differentiation. Bulletin Commission Geologique de Finlande, 97: 68-77. Mollis, J.A., Clarke, G.L., Klepeis, K.A., Daczko, N.R., and Ireland, T.R., 2003. Geochronology and geochemistry of high-pressure granulites of the Arthur River Complex, Fiordland, New Zealand: Cretaceous magmatism and metamorphism on the palaeo-Pacific Margin. Journal of Metamorphic Geology, 21: 299-313. Holmquist, P.J., 1921. Typen und Nomenklatur der Adergesteine. Geologiske Foreningens i Stockholm Forhandlingar, 43: 613-631. Klepeis, K.A., and Clarke, G. L., 2004. Evolution of an exposed lower crustal attachment zone in Fiordland, New Zealand. In: Grocott, J., McCaffrey, K., Taylor, G., and Tikoff, B. Vertical coupling and decoupling in the lithosphere. Geological Society Special Publication, (in press). Lopez, S., and Castro, A., 2001. Determination of the fluid-absent solidus and supersolidus phase relationships of MORB-derived amphibolites in the range 4-14 kbar. American Mineralogist, 86: 1396-1403. Marmo, B., Clarke, G.L., and Powell, R., 2002. Fractionation of bulk rock composition due to porphyroblast growth: effects on eclogite facies mineral equilibria, Pam Peninsula, New Caledonia. Journal of Metamorphic Geology, 20: 151-166. Mehnert, K. R., 1968. Migmatites and the origin of granitic rocks. Elsevier, Amsterdam: 393p. Mehnert, K.R., 1973. Initial melting at grain boundaries of quartz and feldspar in gneisses and granulites. Neues Jahrbuch fur Mineralaogie, 4: 165-183. Newton, R.C., and Manning, C.E., 2003. Activity coefficient and polymerisation of aqueous silica at 800°C, 12 kbar, from solubility measurements on Si02-buffering mineral assemblages. Contributions to Mineralogy and Petrology, 146: 135-143. Oliver, G.J.H., 1977. Feldspathic homblende and garnet granulites and associated anorthosite pegmatites from Doubtful Sound, Fiordland, New Zealand. Contributions to Mineralogy and Petrology, 65: 111-121. Olsen, S.N., 1983. A quantitative approach to local mass balance in migmatites. In: Atherton, M.P. and Gribble, C.D., (eds.). Migmatites, melting and metamorphism. Shiva Publishing, Nantwich, U.K: pp. 201-233.

57 Olsen, S.N., 1985. Mass Balance in Migmatites. In: Ashworth, J.R. (ed.). Migmatites, Glasgow: Blackie, pp. 145-179. Raith, M., and Srikantappa, C., 1993. Arrested chamockite formation at Kottavattam, southern India. Journal of Metamorphic Geology, 11: 815-832. Rushmer, T., 1991. Partial melting of two anq)hibolites: contrasting experimental results under fluid-absent conditions. Contributions to Mineralogy and Petrology, 107: 41-59. Sawyer, E.W., 1987. The role of partial melting and fractional crystallisation in determining discordant migmatite leucosome conpositions. Journal of Petrology, 28: 445-473. Sawyer, E. W., 1999. Criteria for the recognition of partial melting. Physics and Chemistry of the Earth. Part A: Solid Earth and Geodesy, 24: 269-279. Sawyer, E.W. and Robin, P.-Y.F., 1986. The subsolidus segregation of layer-parallel quartz-feldspar veins in greenschist to upper amphibolite facies metasediments. Journal of Metamorphic Geology, 4: 237-261. Sederholm, J. J., 1907. Om granit och gneis. English summary: On granite and gneiss. Bulletin Commision Geologique de Finlande, 23: 1-90. Sederholm, J. J., 1934. On migmatites and associated Precambrian rocks of southwestern Finland. Bulletin Commision Geologique de Finlande, 107: 1-68. Schroter, F.C., Stevenson, J.A., Daczko, N.R., Clarke, G.L., Pearson, N.J., and Klepeis, K.A, 2004. Trace element partitioning during high-P partial melting and melt-rock interaction: an example from northern Fiordland, New Zealand. Journal of Metamorphic Geology, 22: 443-457. Shen, A.H., and Keppler, H., 1997. Direct observation of complete miscibility in the albite-H20 system. Nature, 385: 710-712. Tracy R. J., 1982. Compositional zoning and inclusions in metamorphic minerals. Reviews in Mineralogy, 10, 355-394. Waters, D.J., and Whales, C.J., 1984. Dehydration melting and the granulite transition in metapelites from southern Namaqualand, S. Africa. Contributions to Mineralogy and Petrology, 88: 269-275. White, R.W., and Powell, R., 2002. Melt loss and the preservation of granulite facies mineral assemblages. Journal of Metamorphic Geology, 20: 621-632. Yamamoto, H., and Yoshino, T., 1998. Superposition of replacements in the mafic granulites of the Jijal complex of the Kohistan arc, northern Pakistan: dehydration and rehydration within deep arc crust. Lithos, 43: 219-234.

58 A sequence of corona formation in two-pyroxene gabbro, Fiordland New Zealand: intrusion, rapid post-magmatic cooling and transformation of gabbro to gamet granulite. JOEL FITZHERBERT AND GEOFF CLARKE School of Geosciences, The University of Sydney, NSW, Australia, 2006, Australia (email: joel@Reosci.usvd.edu.au)

Coronitic-gabbro intrusives of the Pembroke Valley, New Zealand preserve an unusual sequence of coronas developed between igneous Fe-Mg silicates, Fe-Ti oxides and plagioclase. The igneous assemblage does not contain gamet, constraining the depth of gabbro intrusion to less than c. 35 km at temperatures above 1000 Two distinct stages of corona crystallisation (Ci and C2) at lower crustal levels have been identified. Ci coronas involve symplectic intergrowths of pargasite-kyanite-clinopyroxene-orthoclase-plagioclaseclinozoisite-quartz+/-orthopyroxene, and formed during incipient hydration at the boundary between the two-pyroxene and gamet granulite facies at c. 25-35 km depth and T «800 (Fig. 2a,b). Both igneous and Ci corona mineral assemblages are cut by a series of ramifying anorthositic veins (Fig. 1).

•Fig. 2a,b -Fig. 2c,d •t '".-i

-Fig.2e,f

Figure 1. Gamet reaction zone from the Pembroke Valley. Notice the Si fabric defined by aligned pyroxene and pargasite. Ci coronas are exclusively preserved within the host lithology. C2 GRZ coronas are represented by the bleached zone. Notice the anorthositic vein in the middle of the GRZ.

59 Adjacent and parallel to these veins, igneous and Ci corona minerals are pseudomorphed by gamet-clinopyroxene-rutile-bearing C2 coronas; the zone of C2 corona formation has been termed the garnet reaction zone (GRZ) (Fig. 2c,d,e,f). C2 corona mineral assemblage and mineral composition vary consistently from the coronitic-gabbro/GRZ boundary through the GRZ/central vein boundary. This variation represents decreasing degree of recrystallisation of igneous and Ci mineral assemblages away from the central vein (Fig. 2c,d,e,f). C2 coronas formed via dehydration and recrystallisation of igneous and Ci mineral assemblages in the garnet granulite facies at P «1.1 GPa (c. 30-35 km depth) and T «750 ""C. The sequence of corona development at Pembroke represents the post-magmatic recrystallisation of gabbroic lithologies at constant or slightly increasing pressure and declining temperature. Importantly, isotopic dates indicate that the intrusive and garnet granulite forming events were separated by as little as 11-17 Ma. This implies that the process of cooling (isobaric or up pressure) to produce high-P granulites is not just restricted to decreasing heat flow in old continental root zones, but may also occur in a transient setting during relatively rapid (11-17 Ma) isobaric cooling of young mafic under plate in an intracontinental arc setting.

Figure 2. The sequence of corona formation at Pembroke from coronitic-gabbro (a,b) to gabbro/GRZ boundary and finally through to the GRZ/central vein boundary (e,f). a) Ci corona developed between igneous orthopyroxene and plagioclase. Dashed white line denotes the boundary between igneous orthopyroxene and Ci clinopyroxene. The dark zone directly adjacent to clinopyroxene is a zone of orthopyroxene-pargasite intergrowth (white arrow). Sym = symplectic intergrowths of kyanite-quartz-plagioclase. Clinozoisite forms intergowths with quartz and plagioclase at the boundary. Base of photomicrograph 4mm wide, b) Ci corona developed between a magnetite-ilmenite intergrowth. Finely intergrown kyaniteplagioclase-quartz-pargasite from sympletctic intergrowths adjacent to the oxide grain, while adjacent to igneous plagioclase kyanite-quartz-plagioclase are the symplectite phases. Black arrow points an igneous orthopyroxene grain. Base of photomicrograph 4mm. c) Combined C1/C2 corona from the GRZ/coronitic-gabbro boundary. Here C2 garnet occupies the position once occupied by pargasite in the Ci corona, Ci kyanite needles are continuous through C2 garnet. Base of photomicrograph 4mm. d) C2 corona developed around ilmenite from the coronitic-gabbro/GRZ boundary. Kyanite needles are continuous from the ilmenite core, though the C2 garnet and into plagioclase. Base of photomicrograph 4mm. e) Welldeveloped C2 garnet necklace around igneous orthopyroxene and Ci clinopyroxene at the GRZ/vein boundary. Note the abundant intergrown quartz adjacent the gamet-clinopyroxene interface. Note that the majority of the Ci corona has been pseudomorphed by C2 gamet-clinopyroxene-quartz intergrowths. The only similarity to Cj coronas preserved here is the twophase pyroxene in the core of the corona. Plagioclase here is commonly charged with fine kyanite inclusions. Base of photomicrograph 8mm.. f) Well-developed C2 garnet necklace around rutile at the GRZ/vein boundary. Here ilmenite is no longer present and the pure Ti oxide is enveloped by C2 garnet and quartz. Base of photomicrograph 4mm.

Tracking crustal differentiation and assimilation processes at arc volcanoes: a Uranium series isotope perspective Rhiannon George^ S. Turner^ R. Price^ C. Cook^, B. Finney^ ^GEMOC Key Centre, Department of Earth and Planetary Sciences, Macquarie University, Sydney NSW 2109 ^School of Science and Technology, University of Waikato, Private Bag 3105, Hamilton, New Zealand ^Department of Earth Sciences, University of Bristol, Queens Rd, Bristol, BS8 IRJ, UK Parent-daughter isotope systems that decay on time scales appropriate to the rates of processes themselves have proved powerful tools in tracking the histories of many arc-related processes from fluid release during plate subduction to dynamic melting of the mantle wedge. However, the rates inferred for mantle processes can only be considered real if the effects of continental crustal interaction are demonstrably minimal. Increasingly, it appears that short-lived isotope disequilibria are systematically different between oceanic and continental arc regimes and the extent to which this is controlled by crustal processes is an imperative that must be addressed if we are to gain reliable time information as well as derive realistic global mass flux models. A related issue is how we can reconcile growing evidence for small-scale, large amplitude isotope heterogeneity in magmas with evidence derived from whole-rocks. In order to try and deconvolve mantle and crustal processes in this context we consider two contrasting case studies. In the first, we examine U-Th-Ra disequilibria in a suite of historic eruptives from Ruapehu volcano at the southern end of the Taupo volcanic zone. New Zealand. The case for variable involvement of lower crustal and shallower level lithologies through time is clear-cut at this volcano (Price et al, 2005), as are micro-scale variations in bulk chemistry between groundmass glass and coexisting crystals. In the second, we examine U-Th disequilibria in an andesitic-dacitic eruptive suite from a typical oceanic arc volcano at Okmok, Umnak Island in the Aleutian arc. In this example the geochemical effects of crustal interactions on magmas are necessarily much more subtle as the contrasts between incoming mantle magmas and the crust itself will be small. Here, detailed MELTS modelling in combination with in-situ fractional crystallisation models show that the major and trace element variations observed over the last 10000 years can be explained by crystal fractionation and accumulation alone. However, we show new Sr and O isotope evidence which, in combination with ^^^U-^^^Th disequilibria, provide robust, yet tantalising glimpses into the behaviour of the U-series nuclides during shallow-level crustal interaction. Ruapehu volcano, New Zealand On an equiline diagram the Ruapehu samples form a subtle positive array that extends from Taupo rhyolites and is intermediate between them and Kermadec lavas. It is conceivable from this data that a simple model of mixing between mantle-derived melts and Taupo crustal melts may explain the Ruapehu disequilibria. However, other reservoirs may be involved. For example, the range of calculated disequilibria derived from elemental Th/U ratios in averaged metasedimentary xenoliths, Torlesse and Waipapa basement averages could also be involved.

62 Thus an alternative, although speculative, model is that Ruapehu samples reflect melts sourced from the lower to middle crust (as sampled in the xenolith suite), mixed with those generated at shallower levels (represented by Taupo-like magmas). Irrespective of the exact details of likely components, the key aspect of these systematics is that for the first time, there is sufficient information available to show that an inclined array on the equiline diagram for a single volcanic suite is unequivocally the product of open system processes: mixing between mantle or lower to mid- crustal melts and shallow upper crustal components. By implication, this dictates that the slope of the Ruapehu array has no time significance, and that if anything, the mixing processes took place over very short intervals with respect to the time scale of ^^^Th decay. Okmok volcano, Aleutian arc. At face value, ^^^U-^^^Th disequilibria in young volcanics from Okmok suggest time scales of 50 kyr or less for some combination of mantle-crustal processes to take place. Given good evidence in the arc as a whole for fluid transfer from the subducted plate to take less than 10 kyr to reach the arc volcano (George et al, 2003), this additional time could reasonably be supposed to relate to a crustal phenomenon. Extremely good negative correlations between Sr and O isotopes are consistent with crustal assimilation of low hydrothermally-altered wallrocks. While the correlation between these two parameters and (^^^Th/^^^Th) appears to be less coherent, it is striking that these 18 activity ratios broadly increase with increasing Sr isotope ratios and decreasing 5 O. The assimilant could therefore be hydrothermally-altered arc crust of relatively recent age of less than 100 kyr. The implications of our two case studies are considerable, and highlight two key, scaledependent aspects. At the plate level, careful consideration should be exercised when inferring time scales from U-Th disequilibria in along-arc datasets if subtle effects of shallow level crustal interaction are observable at the single-edifice scale. Moving from the macroscopic to the microscopic, it would appear that for the ^^^U-^^^Th system where the U and Th partitioning between crystals and melt are heavily balanced in favour of the melt, whole-rock measurements must remain relatively impervious to the small-scale, large amplitude geochemical fluctuations resolvable in some other systems.

63 Geochemistry of Mesozoic rocks of Victoria: melting of source mantle remote from sites of Gondwana break-up Daniel Halas^ Ian Nicholls^ and Roland Maas^ ^ School of Geosciences, PO Box 28E, Monash University, Clayton, VIC 3800 ^ School of Earth Sciences, University of Melbourne, Parkville, VIC 3010 During continental break-up, major mantle melting episodes may generate large-scale subaerial volcanic activity, giving rise to continental flood basalt (CFB) provinces, the rocks of which provide important geochemical evidence for the nature of underlying mantle reservoirs. In Tasmania, vast volumes of Early- to Mid-Jurassic (-175 Ma; Brauns et al, 2000) basaltic rocks are believed, with similar occurrences in Antarctica, to represent a CFB province known as the Ferrar Large Igneous Province (Ferrar LIP). In other parts of southeastern Australia, Mesozoic magmatism took place on a much smaller scale, with occurrences scattered from South Australia through Victoria into New South Wales (Sutherland, 1978). Rocks of this region most similar to those of the Ferrar LIP are the Jurassic low-Ti tholeiites of Kangaroo Island (Foden et al. 2002). Rocks partly of similar age (-^190-150 Ma?) but very different geochemistry occur amongst the widespread small occurrences of mainly alkaline basaltic to phonolitic rocks (Coleraine Volcanic Group Morand et al 2003) of the Dundas Tableland, far westem Victoria (Day, 1983; Hergt et al 1991). Similar alkaline basalts occur as --190 Ma dykes in central Gippsland (the Freestone dykes - Soesoo et al 1999; Elburg and Soesoo, 1999). In central and eastern Victoria, further known examples of Jurassic alkaline volcanic and shallow intrusive rocks include the "lamprophyre" dykes of the Bendigo Goldfield and the Gallows Hill phonolite. Both these occurrences have been dated at --160-150 Ma. In south-central Victoria are a number of occurrences of Cretaceous (--100-80 Ma) volcanic and shallow intrusive rocks, mostly basaltic. These include lava flows and minor intrusives (dykes and plugs) of the San Remo-Bass-Kilcunda-Wonthaggi coastal strip, and lavas further east within South Gippland. Similar basalts form the earliest lava-dominated sub-provinces (Ballan Graben, Poowong) of the extensive, dominantly Cenozoic Older Volcanics Province (Day, 1983; Price et al 2003). To date, geochemical studies of Victorian Mesozoic igneous rocks have been very limited, reflecting the small scale of occurrences and broad variety of types. This paper examines their geochemistry and assesses their probable mantle sources, with emphasis on the recognition of contributions from subcontinental lithospheric mantle (SCLM). Also assessed are reasons why the scale of Mesozoic magmatic activity in southeastern Australia during Gondwana breakup was much smaller than that in Tasmania and Antarctica. The simplest possibility is that the former took place distant from the main zone of plume activity, asthenospheric mantle upwelling and continental extension.

64 Victorian Mesozoic basaltic rocks represent a broad spectrum, from strongly alkaline lamprophyres and sodic alkaline basalts to tholeiitic basalts (Fig. 1). Their compositions suggest derivation by a wide range of degrees of mantle source melting, followed by crystal fractionation.

• Dundas Tableland • Central Victorian Goldfields • Ballan Graben + South Gtppsland

Si02 (wt%) Fig. 1: Alkalis vs silica diagram illustrating the wide range of tholeiitic to strongly alkaline compositions of Victorian Mesozoic basaltic rocks Rare earth element patterns for all types are consistently LREE-enriched (Fig. 2), indicating mantle sources enriched in these and other incompatible elements.

DTP average

1000

B — BGP average •A—South Gippsland nephelinite (D6x1) -h— South Gippsland tholeiites (avg) OIB average

•C 100 ^ "D

monchiquites

o o a E C3 V)

Fig. 2: Rare earth element patterns for representative samples or averages for Victorian Mesozoic basaltic groups Initial isotopic compositions (^^Sr/^^Sr = 0.70274 to 0.70559, Snci = -3.8 to + 7.3) show strong similarities to those of plume-related oceanic island basalts and continental basalts worldwide,

65 but they are also within the range of the sub-continental lithospheric mantle (SCLM) beneath southeastern Australia (Fig. 3). Most do not appear to have undergone significant crustal contamination. By comparison with hypothetical oceanic mantle "end member" source components these basaltic rocks show dominant "Prevalent Mantle" (PREMA) characteristics, in some cases together with the signatures of enriched mantle (EM) components. • Ballan Graben • Dundas Tableland A South Gippsland • monchiquites XNVP plains basalts -f NVP basalts

0.7020

0.7030

0.7040 87

0.7050

0.7060

0.7070

Sr/®®Sr (i)

Fig. 3: Initial Sr- and Nd-isotope data for Victorian Mesozoic basaltic groups, compared with those for basalts of the Late Cenozoic-Recent Newer Volcanics Province. Melting models based upon REE patterns suggest that the entire range of basaltic magmas represented could be generated from metasomatised spinel Iherzolite sources, with highly alkaline lamprophyres/basalts formed by -3-5 % melting, transitional basalts by ~9-13 % melting and tholeiitic basalts by -^15-23 % melting. Strong evidence for mildly enriched sources confirms that the SCLM was probably the dominant mantle source region. Comparison of trace element and isotopic compositions of the Victorian basalts with those of the tholeiitic basaltic rocks of the Ferrar LIP, including Kangaroo Island Jurassic basalts, indicates strongly contrasting mantle source regions. Volumes of highly enriched lithospheric mantle probably involved in the origin of Ferrar LIP magmas (Brauns et al 2000) and South Australian Jurassic kimberlitic magmas do not appear to have influenced the Victorian Mesozoic magmas, which were emplaced east and/or north of the likely sites of ancient SCLM enrichment (Foden et al 2002).

66 extension above plume

—I

Victoria

Ferrar/Karoo

small degree melting

plume

large degree melting

material underplating lithosphere asthenosphere -2000 km

Fig. 4:

A possible scenario for relationships between Jurassic basaltic magmatism in Victoria and the Ferrar Large Igneous Province.

Underplating of the SE Australian SCLM by plume mantle material remote from the main zone of extension and melting associated with plume activity during Gondwana break up may explain the small-volume nature and wide distribution of the Victorian occurrences (Fig. 4). References Brauns, C.M., Hergt, J.M., Woodhead, J.D., and Maas, R., 2000. Os isotopes and the Origin of the Tasmanian Dolerites. Journal of Petrology, 41: 905-918. Day, R.A., 1983. Petrology and Geochemistry of the Older Volcanics (Victoria)-Distribution, Characterization & Petrogenesis. Unpublished PhD. Thesis, Monash University. Elburg, M.A., and Soesoo, A., 1999. Jurassic alkali-rich volcanism in Victoria (Australia): lithospheric versus asthenoshperic source. Journal of African Earth Sciences, 269-280. Foden, J., Song, S.H., Turner, S., Elburg, M., Smith, P.B., Van der Steldt, B., and Van Penghs, D.V, 2002. Geochemical evolution of lithospheric mantle beneath S.E. South Australia. Chemical Geology, 182: 663-695. Hergt, J.M, Peate, D.W., and Hawkesworth, C.J, 1991. The Petrogenesis of Mesozoic Gondwana low-Ti flood basalts. Earth and Planetary Science Letters, 105: 134-148. Morand, V.J., Wohlt, K.E., Cayley, R.A., Taylor, D.H., Kemp, A.I.S., Simons, B.A., and Magart, A.P.M., 2003. Glenelg Special Map Area Geological Report. Geological Survey of Victoria, Report 123, Geological Survey of Victoria, 261 pp. Price, R.C., Nicholls, LA., and Gray, C.M., 2003. Cainozoic igneous activity. Chapter, 12 (pp. 362-375). In: Birch, W.D. {Q^,Geology of Victoria. Geological Society of Australia, Special Publication 23: 842 pp. Sutherland, F.L., 1978. Mesozoic-Cainozoic Volcanism of Australia. In: Scheibner, E. (ed). The Phanerozoic Structure of Australia and Variations in Tectonic Style. Tectonophysics, 48 (3-4): 413-427.

67 Determination of U-Th and REE in apatite for (U-Th-Sm)/He dating Janet Hergt\ Jon Woodhead\ Stephen Anderson^ ^School of Earth Sciences, The University of Melbourne, 3010, VIC, Austraha Varian Austraha Pty Ltd, 3170, VIC, Austraha Renewed interest in (U-Th)/He dating, particularly of apatite grains, demands a reliable, rapid and sensitive analytical approach to the acquisition of Th, U and Sm concentrations. We have developed protocols in our laboratory, at the University of Melbourne, capable of meeting these challenges without recourse to isotope dilution techniques or the requirement for desolvation systems. The sensitivity of the Varian quadrupole ICPMS employed in this work enables routine analysis at dilutions of several hundred thousand, representing concentrations in solution of 0.1 -0.001 ppb. Typically, a total of 0.05-0.1 ng of Th and U are available for analysis, and U contents as low as 0.1 ppm in the solid have been determined with sufficient reliability for geochronological purposes. Analytical methods Although a number of studies report the trace element abundances of apatite grains acquired via SIMS or LA-ICPMS (e.g., Belousova et al, 2001), successful (U-Th-Sm)/He dating requires analysis of these elements from the entire volume of the grain (or group of grains). Given the small sample sizes (typically 0.005-0.02 mg), even dissolution in small volumes of acid represent dilution factors of tens to hundreds of thousand. In order to conduct routine ICPMS measurements at such dilutions, particularly in the absence of a desolvation sample introduction system, an instrument with high sensitivity, low backgrounds and stable signals is essential. The advantage in avoiding the use of a desolvation system is that this removes the unwanted elemental fractionations that such additional processes introduce. The Varian quadrupole ICPMS has been developed to create high performance ion optics employing a 90° ion mirror. This reduces contamination, while increasing efficiency and sensitivity and improving long term signal stability. The introduction of curved fringe rods reduces the number of excited neutral ions prior to the quadrupole, resulting in low background counts and improved detection limits. For routine bulk-rock analysis, the Varian instrument in our laboratory is tuned to achieve sensitivities of million cps/ppm (Be), -300 million cps/ppm (In) and --^150 million cps/ppm (Th) with oxide and doubly charge ion ratios of around 1% and 3% respectively. In the case of apatite work, where the low mass range is less important, the instrument is tuned to optimise the high mass range, typically yielding ~300-400 million cps/ppm Th (at the expense of Be which is often reduced to around 3 million cps/ppm) at similar or lower oxide and doubly charged ion levels. These tune settings are both within 'normal' sensitivity mode, for which the continuum background is less than 5 cps. Results The sources of uncertainties in (U-Th)/He age determinations are complex (e.g., Farley et al, 1996; House et al., 1997; Farley, 2000). In terms of the Sm, Th, and U data, the errors in calculated ages will obviously be influenced by whether the uncertainties for each element are

offset to values that are too high, or low (i.e., in the same direction) or if one is high and another low such that these cancel out to some degree. Fortunately there are several mechanisms by which the data quality can be assessed. First, Standard apatites (variable in U, Th, Sm and He, but of a single and well-documented age) are run several times in each batch of unknowns. The Melbourne laboratory (and indeed most laboratories) employ the "Durango Apatite". We also use an in-house Mud Tank apatite standard. Thus, when the He data and U-Th-Sm data are combined, this approach provides a robust monitor of data quality. Second, given that the analytical routine is very similar to "normal" trace element runs, welldocumented standard reference materials (e.g., USGS rock standards) can be included to monitor accuracy of trace element abundances. Using this approach, our values for Sm, U and Th are generally within 1% (sometimes 2%) of recommended values. Finally, as little additional effort is involved, we obtain the range of rare earth and other elements in all samples. The smoothness of chondrite-normalised REE data provide a useful measure of the quality of the run and patterns are plotted for all samples as a quality control check. Examples of REE, Th and U data acquired using this approach are listed in Tables 1 and 2 and illustrated in Figure 1. Table 1 summarises Th and U data that correlate with the REE data in Table 2 and displayed in Figure 2. Note that the relative standard deviations are generally less than 3% and often less than 1%. As indicated by the data, this is not because the concentrations of U and Th are high in these solutions, nor does it simply mean that apatites contain high U and Th contents. Apatite 3 contains around 0.1 ppm in the solid; nevertheless, the uncertainties are only 2.3% for Th and 6.6% for U even at such low abundances. Similar data are listed for Sm determinations in Table 2. Apart from being useful in the (U-Th)/He studies, Sm provides a 'typical' measure of individual REE contents as it is seldom as highly enriched as La or depleted as Lu (or vice versa). Table 1. An example of typical apatite Th and U results. Three 'unknown' apatites are run in conjunction with the Durango and Mud Tank standards. Note the standards are heterogeneous in both Th and U, and that the total ng values determined are combined with the He data to assess the accuracy of the age information.

Table 2. Sm data are listed here for the same samples in order to provide examples of typical REE concentrations measured in solution, the counts per second such solutions generate, and the uncertainties obtained.

0.082 0.124 0.403 1.857 1.702 0.586

0.168 0.254 0.822 3.788 3.437 1.196

10599 13362 31334 127967 118442 44227

0.002 0.001 0.007 0.008 0.011 0.008

2.70 1.12 1.37 1.81 2.95 0.66

The full REE data for these six apatite runs are illustrated in Figure 1. These display a range in both abundance in solution (i.e., approximately 4 orders of magnitude) and REE pattern. The highest uncertainties occur in the Lu for Apatite 2 (12.5% rsd) as this had a count rate of only 484 c/s. Despite this, the REE pattern remains remarkably smooth even at these very low abundances.

Figure 1. Chondrite normalised REE patterns for the apatite samples listed in Tables 1 and 2. Note that the data here are not recalculated as ppm in the solid, but are the total ng of each element as determined during the analysis, normalised to chondrite. The top pattern (solid diamonds) is the Durango standard, having 47 ppb La and 0.12 ppb Lu in solution. The two smooth convex upward curves (open diamonds and crosses) are the patterns for the two Mud Tank standards. The final three patterns are Apatite 1 (solid triangles, 0.52 ppb La in

70 solution), Apatite 2 (open circles, 0.003 ppb Lu in solution) and Apatite 3 (solid squares). Note: Pm and Tm are interpolated values. Conclusions In order to conduct high-throughput (U-Th-Sm)/He dating, it is essential to have a means of measuring U, Th and Sm abundances rapidly, routinely and on very small sample volumes. Isotope dilution is unnecessary as long as careful independent monitoring protocols are established, and this permits a far broader range of elemental abundances to be determined (e.g., REE element suites). Our approach does not require a desolvation sample introduction system despite the high dilution factors routinely employed. This removes the significant and undesirable elemental fi-actionations usually experienced with such systems, thereby improving the quality and reliability of the results. This same approach has the potential to open up new opportunities for the trace element characterisation of apatite (and other accessory minerals) in a wide range of rock-types, providing new ways of constraining petrogenetic models. References

Belousova, E.A., Walters, S., Griffin, W.L., and O'Reilly, S.Y., 2001. Trace-element signatures of apatites in granitoidsfromthe Mt Isa Inlier, northwestern Queensland. Australian Journal ofEarth Sciences, 48: 603-619. Farley, K.A., Wolf, R.A., and Silver, L.T., 1996. The effects of long alpha-stopping distances on (U-Th)/He ages. Geochimica et Cosmochimica Acta, 60: 4223-4229. Farley, K.A., 2000. Helium diffusion from apatite: general behaviour as illustrated by Durango Fluorapatite. Journal Geophysical Research, 105: 2903-2914. House, M.A., Wernicke, B.P., Farley, K.A., and Dumitru, T.A., 1997. Cenozoic thermal evolution of the central Sierra Nevada, California, from (U-Th)/He thermochronometry. Earth and Planetary Science Letters, 151: 167179.

Insights into magma generation and evolution at White Island, New Zealand Zara Heyworth^ Rhiannon M. George^, Bruce F. Schaefer^ and Simon P. Turner^ ^School of Geosciences, Monash University ^GEMOC, Macquarie University Subduction zones are a key factory in the geochemical evolution of the continental crust and upper mantle. The conditions during partial melting and liquid lines of descent (LLD) can provide valuable insights into magma generation and evolution. White Island is an active arc volcano in the Taupo Volcanic Zone (TVZ), New Zealand and provides an excellent environment to assess models of melt generation and evolution within the broader context of arc magmatism. White Island is situated 50 km offshore in the Bay of Plenty and is composed of andesiticdacitic lavas and small-volume pyroclastics. Historically, the volcano has been characterised by continuous fumarolic and hydrothermal activity, punctuated by frequent but relatively minor phreatic, phreatomagmatic and strombolian eruptions. Lavas are calc-alkaline with compositional ranges from 55-63% Si02 (wt%) and 7.5-10% MgO (wt%). High-Mg andesites, erupted at White Island from 1977-2000, are relatively rare on a global scale but may have implications for the melting conditions and source of the primary magmas. Knowledge of the physical conditions during crystallisation can provide insights into subsequent magmatic evolution. Using the two-pyroxene thermometry of Lindley (1983), based on coexisting ortho- and clino-pyroxene compositions, the crystallisation temperature ranges from HOOT to 1200T with a separate group at around 700T (Figure 1). Ternaryfeldspar thermometry (Fuhrman and Lindley, 1988) also yields low equilibration temperatures of less than 750T. Crystallisation pressures were calculated using a thermobarometer that involves the jadeite (Jd; NaAlSi206)-diopside/hedenbergite (DiHd; Ca(Mg, Fe)Si206) exchange equilibrium between clinopyroxene and Hquid (Putirka et al, 1996). Lavas erupted between 1977 and 1992 have an average crystallisation pressure of ~ 4 kbars, equivalent to depths of about 12 km. In contrast, the crystallisation pressures for lavas erupted in 2000 are <1 kbar, indicative of very shallow crystallisation/equilibration depths (<3 km). However, such pressures are consistent with previous estimates at Mt Ruapehu (0.5-7 kbars; Graham and Hackett, 1987; Gamble et al., 1990) and a magma chamber depth of 2-7 km beneath White Island (Cole et aL, 2000).

72 Hd

Di X phenocryst •

prehistoric phenocryst

A prehistoric groundmass inclusion

4- HCA --- polythermal boundary of 'forbidden zone'

Figure 1 Crystallisation temperatures of White Island magmas with coexisting ortho- and clino-pyroxene compositions based on the thermometry of Lindley (1983). A typical phenocryst assemblage of White Island lavas contains 40% plagioclase, 50% clinopyroxene, 5% orthopyroxene and 5% olivine with a total phenocryst mode of 45%. For historic pyroxene phenocrysts, En content varies from 70-76 in orthopyroxene and 44-56 in clinopyroxene. Plagioclase composition ranges from An63 to An98 and olivine composition ranges from Fogo to F092. However, the liquid line of descent (LLD) indicates that these proportions do not reflect the true fractionating assemblage (Figure 2). All phenocryst assemblages lie above the LLD and thus contain more plagioclase than the true fractionating assemblage. A negative correlation between AI2O3 and MgO ftirther supports the notion that plagioclase did not dominate the fractionating assemblage.

• WR

»a XR

Figure 2 A plot of MgO against A1203 for whole rock (WR), olivine (Ol), pyroxene (Px) and plagioclase (PI) illustrating that the erupted assemblages do not represent the true fractionating assemblage. Solid squares represent phenocryst assemblages erupted between 1977-2000 at White Island and the dashed line represents the LLD.

Previous work indicates that high-Mg andesites can form from the reaction of partial melts of subducted, eclogite-facies basalt with the overlying mantle wedge (e.g. Pe-Piper, 1991; Shiraki et al, 1994) or by direct partial melting of extremely hydrous upper mantle at relatively low temperatures (e.g. experimental results of Hirose, 1997). The high-Mg andesites at White Island have SrA^ ratios of 7-8 and LaAT^ of 3.5-4 compared to SrA^ ratio values >40 and La/Yb ratio values of 48 in slab melts (e.g. Yogodzinski et aL„ 1995). These values argue against the formation of these melts in the presence of residual garnet in the melt source region. Therefore, it is unlikely that the high-Mg lavas were generated as an eclogitefacies slab melt and hydrous melting at low temperatures and pressures seems more likely. Useries disequilibria data are being collected to constrain the time scales of fluid addition and partial melting.

73 References

Cole, J. W., Thordarson, T., and Burt, R. M., 2000. Magma origin and evolution of white Island (Whakaari) volcano. Bay of Plenty, New Zealand. Journal of Petrology, 41; 867-895. Fuhrman, M., and Lindley, D. H., 1988. Ternary-feldspar modeling and thermometry. American Mineralogist, 73: 201-215. Gamble, J. A., Smith, L, Graham, I. J., Kokelaar, B. P., Cole, J. W., Houghton, B., and Wilson, C., 1990. The petrology, phase relations and tectonic setting of basalts from the Taupo Volcanic Zone, New Zealand and the Kermadec Island Arc- Havre Trough, SW Pacific. Journal of Volcanology & Geothermal Research, 43: 253270. Graham, I. J., and Hackett, W. R., 1987. Petrology of calc-alkaline lavas from Ruapehu volcano and related vents, Taupo Volcanic Zone, New Zealand. Journal of Petrology, 28: 531-567. Lindley, D. H., 1983. Pyroxene thermometry. American Mineralogist, 68: 417-493. Putirka, K., Johnson, M., Kinzler, R., Longhi, J., and Walker, D., 1996. Thermobarometry of mafic igneous rocks based on clinopyroxene-liquid equilibria, 0-30 kbar. Contributions to Mineralogy and Petrology, 123: 92108.

Yogodzinski, G. M., Kay, S. M., Kay, R. W., Volynets, O. N., and Koloskov, A. V., 1995. Magnesian andesite in the western Aleutian Komandorsky region: implications for slab melting and processes in the mantle wedge. Geological Society of America Bulletin, 107: 505-519.

74 Geochemical modeling of mineralisation scenarios for turbidite-hosted gold deposits in the Lachlan Fold Belt Terrence P. Mernagh^ and Frank P. Bierlein^ ^Geoscience Australia, GPO Box 378, Canberra, ACT 2601 ^Tectonics Special Research Centre, University of Western Australia, Crawley WA 6009 Traditionally, gold mineralisation in the Lachlan Fold Belt of southeastern Australia has been assumed to be exclusively associated with structurally controlled dilatational sites (Foster and Gray, 2000). Most of these deposits formed at intermediate crustal levels where gold precipitation is considered to be closely linked with fluctuations in fluid pressure during seismic events. Brittle - ductile deformation cycles corresponding to seismic episodes produce the laminated and sigmoidal or en-echelon veins that are characteristic of most turbidite-hosted gold deposits. Quantification of the factors controlling the distribution and size of these deposits has long been the focus of intensive efforts by researchers and exploration geologists alike, but there has been little rigorous chemical modeling to determine exactly which chemical parameters have the largest effect on gold deposition in this setting. Therefore, we have used the HCh modeling software (Shvarov and Bastrakov, 1999) to construct a suitable chemical model for gold mineralisation in the Lachlan Fold Belt and then we systematically varied some of the chemical parameters to investigate the effects of these changes on the processes of gold deposition and to determine which fluid compositions are geologically reasonable. The initial fluid composition was based on fluid inclusion data from the Fosterville gold deposit (Memagh, 2001), which is in good agreement with other fluid inclusion studies of lode-gold deposits (Ridley and Diamond, 2000). The Lachlan Fold Belt Model The model used for gold mineralisation in the Lachlan Fold Belt involves a C02-rich, low salinity, aqueous fluid flowing upwards through a fault in a sedimentary rock (composed of 90 % sandstone and 10% shale). 100% -1 90% • 80% 70% 2 o c i

60% -

^d)

50% -

E 3 O

40% -

30% 20% 10% 0% 0.00001

0.0001

0.001

0.01

0.1

Kg Rock Equilibrated with 1 kg Fluid

100

75 Figure 1 Predicted alteration assemblage for increasing rockiwater ratio (left to right) in turbidites infiltrated by a C02-bearing fluid at 300 T (Log /O2 = -29.0; pH = 5.0; [CO2] = 1.14m; [H2S] = .009m> The fluid was initially allowed to undergo phase separation (in the veins) as the pressure dropped from lithostatic to hydrostatic. The modeling was carried out at temperatures from 320 to 200 T and pressures from 100 to 1100 bar. Figure 1 shows the results obtained at 300 ®C and 185 bar. In this scenario, gold is precipitated in the veins over the complete temperature range. In agreement with field observations, the veins are predicted to be mostly quartz (> 93 vol.%) with minor amounts of pyrite, arsenopyrite and muscovite (sericite) precipitating above 230 ""C. The predicted alteration assemblage within the host rocks contains pyrite, arsenopyrite, calcite, muscovite (sericite), chlorite and feldspar. Non-boiling fluid To test the importance of boiling on gold mineralisation we deliberately prevented the fluid from boiling by raising the confining pressure to 1000 bar above the saturated vapour pressure. In contrast to the model above, gold is only precipitated in the vein at temperatures above 310 suggesting that boiling may be an important catalyst at lower temperatures. The vein mineralogy is similar to that of the boiling system but no arsenopyrite and less pyrite and muscovite are precipitated. The aheration assemblage in the host rocks contains siderite rather than chlorite. Less K-feldspar and albite are formed with excess sodium being taken up in paragonite (Fig. 2).

; muscovite S

paragonite

20) c a> E >o

0.00001

0.0001

0.01

0.1

Kg Rock Equilibrated with 1 kg Fluid

Figure 2 Predicted alteration assemblage for increasing rock:water ratio (left to right) in turbidites infiltrated by a non-boiling CO2 -bearing fluid at 300 ^C (Log /O2 = -28.3; pH = 4.8; [CO2] = 1.17m; [H2S] = .008m)

76 Low CO2 Fluid

To test the effects of CO2 on mineralisation all aqueous carbon species were removed from the initial fluid prior to reaction with the turbidites (sandstone and shale). In this case gold is only precipitated in the vein at temperatures above 300 This result is similar to that for the non-boiling fluid and suggests that C02-rich fluids are more susceptible to phase-separation and hence, more efficient for gold mineralisation, particularly at lower temperatures. However, the vein mineralogy differs from that above. At temperatures above 310 the veins are dominated by quartz and arsenopyrite while at lower temperatures they are dominated by quartz and pyrite or quartz-only at temperatures below 240 Alteration of the host rocks initially leads to the precipitation of quartz, muscovite and pyrite (at low rockiwater ratios) followed by the later precipitation of arsenopyrite and epidote and wairakite (Ca-zeolite). The thermodynamics suggests that wairakite is more stable than K-feldspar under the conditions that precipitate arsenopyrite but this may not necessarily be the case in nature (Fig. 3). 100% 90%

: muscovite

80%

0) S

0> 3

E >o

70% 60% 50% 40%

epidote

arsenopynte

30%

quartz

20% 10%

o%-

0.00001

calcite

0.0001

I 0.001

0.01

100

Kg Rock Equilibrated with 1 kg Fluid

Figure 3 Predicted alteration assemblage for increasing rockiwater ratio (left to right) in turbidites infiltrated by a low CO2 fluid at 300 T (Log f02 = -28.9; pH = 4.7; [CO2] = 0.33m; [H2S] = .01m) Low Total Sulfur Fluid

hi this scenario all H2S and HS" were removed from the initial fluid prior to phase separation in the veins and reaction with the host turbidites (sandstone and shale). Although sulfur species are still formed from the sulfides in the host rocks, the concentration of sulfide species in solution is over a factor of ten less than for reference fluid in the initial model. However,

gold still precipitates at all temperatures but the amount of gold precipitated is about a factor of ten less. At temperatures above 310 a quartz-chlorite assemblage dominates the veins. Below 310 quartz and small amounts of carbonate (siderite) are precipitated. The siderite is replaced by graphite at temps below 220 This type of vein mineralogy is not commonly observed in gold deposits in the Lachlan Fold Belt suggesting that the fluids are not usually depleted in sulfur species (Fig. 4). 100%

0.00001

0.0001

0.001

0.01

0.1

Kg Rock Equilibrated with 1 kg Fluid

Figure 4 Predicted alteration assemblage for increasing rockiwater ratio (left to right) in turbidites infiltrated by an initially low total sulfur fluid at 300 T (Log /O2 = -29.2; pH - 5.1; [CO2] = 1.14m; [H2S = .0004m) Summary Chemical modeling of gold mineralisation in the Lachlan Fold Belt shows that gold can be precipitated over a wide temperature range (from 320 to 200 in this study) from CO2bearing, low salinity, aqueous fluid flowing upwards through faults in turbiditic sequences. In agreement with field observations, the veins are predicted to be mostly quartz (> 93 vol.%) with minor amounts of pyrite, arsenopyrite and muscovite (sericite) precipitating above 230 ®C. The predicted alteration assemblage contains pyrite, arsenopyrite, calcite, muscovite (sericite), chlorite and feldspar. Varying some of the chemical characteristics of the initial fluid has resulted in the following changes to the model: Preventing the fluid from boiling stops gold precipitating below 310 ""C but has little effect on the vein mineralogy or the mineralogy of the surrounding alteration assemblage. Removing CO2fromthe fluid also prevents gold precipitation in the veins below 300 ""C. The modeling also generates an alteration assemblage with a number of Ca-rich silicate minerals as less

78 carbonate exists in this system. Removing sulfur species from the initial fluid decreases the amount of gold precipitated by more than a factor of ten, which is to be expected if sulfur ligands are the main species for gold transport. However, the vein assemblage and the lack of sulfide minerals in the surrounding alteration assemblage also suggest that sulfur species are important in this mineral system. The results are in agreement with the widely accepted premise that gold is transported as bisulfide complexes and that the ore-bearing fluid is typically a low-salinity, mixed aqueouscarbonic fluid with low-moderate CO2 contents (Ridley and Diamond, 2000). However, the modeling has shown that the absence of certain physico-chemical processes or fluid constituents, such as boiling or lack of CO2 may inhibit gold precipitation in some environments. Acknowledgements This study is supported by ARC Discovery Grant DP DP0342488. TPM publishes with permission from the CEO of Geoscience Australia. References Foster, D. A., and Gray, D. R., 2000. Evolution and structure of the Lachlan Fold Belt (Orogen) of eastern Australia. Annual Review of Earth and Planetary Sciences, 28: 47-80. Memagh, T.P., 2001. A fluid inclusion study of the Fosterville Mine: a turbidite-hosted gold field in the western Lachlan Fold Belt, Victoria, Australia. Chemical Geology, 173: 91-106. Ridley, J. R., and Diamond, L. W., 2000. Fluid chemistry of orogenic lode gold deposits and implications for genetic models. In: Hagemann, S. G., and Brown, P. E., (eds). Gold in 2000. Reviews in Economic Geology, Society of Economic Geologists, Inc., Littleton, U.S.A., 13: 141-162. Shvarov, Y., and Bastrakov, E., 1999. HCh: a software package for geochemical equilibrium modeling. User's QmdiQ.AGSORecord, 1999/25: 60pp.

79 Thermobarometry of an Early Cretaceous high-pressure contact metamorphic aureole near Resolution Island, Fiordland, New Zealand. Luke A. Milan^ Nathan R. Daczko^ Ian Turnbull^ and Andrew Allibone^ ^GEMOC Key centre. Department of Earth and Planetary Sciences, Macquarie University, Sydney, NSW 2109, Australia ^Institute of Geological and Nuclear Sciences, Dunedin, New Zealand Recent fieldwork in the Resolution Island / Breaksea Sound areas of southern Fiordland, New Zealand has delineated two distinct metamorphic paragneiss terrains in contact with the Early Cretaceous Western Fiordland Orthogneiss batholith. One terrain preserves rare examples of intrusive relationships between the batholith and host rocks, the Palaeozoic Tuhua Sequence and thermobarometric estimations and petrography show that the region adjacent to the intrusive contact records a metamorphic fingerprint (-11-15 kbar, -650-850°C) of the conditions at the time of emplacement or soon thereafter. High-pressure and temperature estimates are recorded up to several kilometres away fi-om the contact, suggesting a large lower crustal aureole. The second metamorphic terrain preserves moderate pressures (~7-8.5 kbar), synonymous with regional mefamorphism during the Palaeozoic and shows no evidence of a Cretaceous high-pressure overprint. This terrain is juxtaposed against the Cretaceous batholith across a previously un-recognised shear zone. Regional Geology New Zealand represents a fragment of the Gondwana Pacific margin that was rifted from the Australian and Antarctic segments during the Late Cretaceous and Early Cenozoic (Mortimer, 1995). South Island geology has been divided into the Eastem and Western provinces (see Figure 1), separated by a belt of rocks called the Median Batholith (Landis and Coombes, 1967; Kimbrough et al, 1994; Mortimer et al, 1999). The Eastem province formed in a convergent magmatic arc, outboard of the Pacific margin of Gondwana. This province consists of arc-derived volcanic rocks, related sedimentary sequences and accretionary complexes of Upper Palaeozoic to Mesozoic ages (MacKinnon, 1983; Bradshaw, 1989; Mortimer, 1995). The Median Batholith is comprised of arc-related plutons, volcanic rocks and sedimentary rocks of Triassic to Jurassic age (ICimbrough et al, 1994). The Westem province is comprised of deeply exhumed Palaeozoic paragneiss and orthogneiss of the Tuhua Sequence that are cut by Devonian and Carboniferous granitoids. The Westem province has been displaced by the Alpine Fault with -500 km of right lateral movement, where upper cmstal sections of the Westem province continue in the Westland-Nelson region (Muir et al, 1996; Wandres et al, 1998; Ireland and Gibson, 1998; Mortimer et al, 1999). The Westem province in Fiordland preserves a polyphase mid Palaeozoic history of a low pressure/high temperature metamorphism followed by moderate pressure (5-7kbar)/high temperature metamorphism at 330 Ma (Ireland and Gibson, 1998). The Fiordland region is dominated by the Westem Fiordland Orthogneiss (WFO), a large 126-116 Ma batholith (170 x 25 km) (Kimbrough et al, 1994; Mattinson et al, 1996; Muir et aL, 1998; Tulloch et al, 2000; HolHs et al, 2004). The WFO was emplaced into thickened cmst (-30 km) and attained peak pressures equivalent to 45 km during the early Cretaceous

80 (Daczko et al, 2001). The Palaeozoic Tuhua Sequence structurally overlies the WFO batholith, and has locally experienced high-P metamorphism (Mattinson et al, 1986; Bradshaw, 1989; Bradshaw and Kimbrough, 1989; Ireland and Gibson, 1998; Daczko et al, 2002). The WFO typically ranges from mafic to intermediate in composition and is geochemically distinct from older rocks, with high Sr and low Y concentrations (HiSY of Tulloch and Kimbrough, 2003). The WFO batholith is part of a paired belt of plutons, with the older outboard rocks (Median Bathohth) having low Sr to Y ratios (LoSY). The WFO comprises the deeply exhumed portions of the inboard belt (Tulloch and Kimbrough, 2003).

• B• • •

Ductile Fault

Figure 1: Geological map of field area in Resolution Island area: note medium-P in the west high-P to the east. Inset A: Locality of Fiordland, Inset B: Location of the Resolution Island with respect to Fiordland. The controversy Work on unravelling Fiordland's metamorphic and geochronological history has resulted in several conflicting interpretations. The granulite facies conditions associated with the emplacement of the WFO are well documented, yet there is uncertainty as to the extent to which the adjacent Tuhua Sequence was affected by the granulite facies event. Discord in the literature revolves around the differing interpretations of the tectonic setting in which the WFO was intruded and the processes that produced the high-P garnet granulite assemblages in the WFO. Three models have been put forward in the literature: 1. Gibson (1988) and Gibson and Ireland (1995), have suggested that the WFO intruded into the lower crust (P>12 kbar) in an extensional tectonic setting. Ductile normal faulting during the mid-Cretaceous is inferred to have exhumed the WFO soon after emplacement.

81 2. Oliver (1990) and Brown (1996) offer a magma loading model that involved intrusion of the WFO at mid crustal levels, and subsequent magma loading to deeper crustal levels led to high-P granulite facies metamorphism of the WFO. 3. Finally, a model involving emplacement of the WFO into mid crustal levels, followed by contraction and overthrusting, producing subsequent crustal loading and an up-pressure history was proposed by Mattinson et al (1986), Bradshaw (1989), and Bradshaw and Kimbrough (1989). This theory has been further ratified more recently by investigations of fold/thrust and crustal loading, as shown in Clarke et al (2000), Daczko et al (2001), Daczko et al (2002), and Klepeis et al (2003). Petrological Units The rocks in this field area were previously un-mapped, such that their relationship with rocks that are reasonably well understood elsewhere in Fiordland, is tentative. The WFO within the field area has been mapped as three distinct units based on field observations. Although texturally distinct, whole-rock geochemical data shows that these units are all gabbroic to dioritic and are compositionally quite similar. They each show HiSY characteristics, synonymous with the WFO (Tulloch and Kimbrough, 2003). The Tuhua Sequence exposed in the field area is possibly related to the Deep Cove Gneiss (Oliver, 1980) that outcrops to the north in Doubtful Sound. The unit commonly comprises heterogenous psammitic lithologies, with minor interlayered amphibolite. Palaeozoic sequences of Resolution Island and north to Lake Beattie contain minor marble and pelitic lithologies. The metasediments preserve a well-defined schistose to gneissic foliation and wellequilibrated amphibolite facies assemblages. Assemblages include plagioclase + hornblende + biotite + quartz ± garnet ± K-feldspar ± muscovite ± epidote ± chlorite ± meionite ± wollastonite ± clinopyroxene ± titanite ± ilmenite ± rutile. The Tuhua Sequence on Resolution Is. has been intruded by a Carboniferous/Cretaceous(?) granite informally known as the Indian Island Granite. No similar granites have been mapped in the Tuhua Sequence, east in Wet Jacket Arm and Breaksea Sound. Thermobarometry and Mineral Chemistry Mineral chemistry determined by electron microprobe analysis was used in thermobarometric calculations, (see examples in Table 1). The thermobarometry results (Table 2) show there are distinct differences in the metamorphic history in the areas investigated. The Palaeozoic metasediments of central Resolution Is. and those to the north at Lake Beattie, record moderate temperature and pressure conditions (~600-700°C, -6.7-8.5 kbar). The suite of rocks to the east, outcropping in the Wet Jacket Arm / Breaksea Sound areas, record high pressures (--11-15 kbar) and temperatures (-'650-850°C). An intrusive contact mapped east of the shear zone on Resolution Is. (73873) contains wollastonite, suggesting the WFO intruded into marble or calcareous metasediment (marble is observed elsewhere in the Tuhua Sequence). Meionite-rich scapolite occurs in the high-P terrain in a sample near Breaksea Sound (71410).

82 Discussion and Conclusion The Tuhua Sequence of metasediments studied to the east, in Wet Jacket Arm and Breaksea Sound, preserve a clear intrusive relationship with the WFO. The country rock metasediments exhibit contact metamorphism as a result with the intrusion of the WFO, with the presence of localised deformation, recrystallisation associated with partial melting and migmatites, rafting of the country rock and dykeing. Here the Palaeozoic metasediments preserve strong evidence of a Cretaceous thermal disturbance, consistent with a well-developed high-P metamorphic aureole. High-grade assemblages are found in all samples analysed in this region, with two of the samples located kilometres from the mapped intrusive contact with the WFO. The presence of meionite (71410) is indicative of a high metamorphic grade at contact aureoles (Deer etaL, 1992). In contrast, the Palaeozoic Tuhua Sequence on Resolution Is. and at Lake Beattie to the north only record moderate pressure assemblages, identical to those found further north by Ireland and Gibson (1998). This moderate-P event was synchronous with the emplacement of granites in Fiordland (Muir et al, 1994), and may have occurred during emplacement of the Indian Island Granite on Resolution Island (of possible Carboniferous/Cretaceous age). A mylonite zone has been mapped that juxtaposes the moderate-P Palaeozoic rocks with the WFO. The lack of a high-P overprint in the Palaeozoic rocks suggests significant displacement across the shear zone. An intrusive contact mapped east of the ductile fault on Resolution Is. records a high-P assemblage, resembling the intrusive contact further east. This study has mapped high-P contact aureoles for the first time in Fiordland, clearly associated with the intrusion of the WFO into the Tuhua Sequence. Intrusional relationships provide an excellent opportunity to study the metamorphic history of the WFO and its affect on the host rocks. There exist a number of possible explanations for the moderate-P metamorphic terrain. The history of these rocks could have involved loading to -25-30 km depth during or preceding the WFO emplacement. However, if they were situated outside the contact aureole and did not recrystallise during the short-lived Cretaceous collision event, they may have experienced but did not re-equilibrate to deeper crustal conditions. A second possibility is that these rocks never experienced high-P conditions in the lower crust, and the newly mapped mylonite zone has juxtaposed different crustal levels, consistent with significant displacement ('-^10 km). The lack of granites in the second high-P terrain to the east may suggest that this terrain is distinct to the first, and perhaps was at an even deeper crustal level when the granites were emplaced. A third possibility is that the second terrain may correlate with metasediments found in George Sound that also preserve intrusive relationships with the WFO. The George Sound metasediments record moderate to high-P conditions, contain no Carboniferous/Cretaceous granitoids, and are significantly younger with source components of a likely Permo-Triassic age (Daczko et aL, 2002; Hollis et al, 2004). Future detrital zircon studies may resolve this issue.

83 The new field, petrographic and thermobarometric data presented here are consistent with the emplacement of the WFO at high-P conditions consistent with arguments made by Gibson et al (1988). Future work will focus on the tectonic significance of the shear zone that juxtaposes the moderate- and high-P terrains to evaluate the role it may have played in the exhumation of the high-P terrain. References Bhattacharya, A., Mohanty, L., Maji, A., Sen, S.K., and Raith, M., 1992. Non-ideal mixing in the phlogopiteannite binary; constrains from experimental data on Mg-Fe partitioning and reformulation of the biotite-gamet geothermometer. Contributions to Mineralogy and Petrology, 111 (1): 87 - 93. Bradshaw, J.Y., 1989. Origin and metamorphic history of an Early Cretaceous polybaric granulite terrain, Fiordland, southwest New Zealand. Contributions to Mineralogy and Petrology, 103: 346-360. Bradshaw, J. D., and Kimbrough, D.L., 1989. Comment: Age constraints on metamorphism and the development of a metamorphic core complex in Fiordland, Southern New Zealand. Geology, 17: 380 - 381. Brown, E. H., 1996. High-pressure metamorphism caused by magma loading in Fiordland, New Zealand. Journal of Metamorphic Geology, 14: 441-452. Daczko, N. R., Klepeis, K. A., and Clarke, G.L., 2001. Evidence of Early Cretaceous collisional-style orogenesis in northern Fiordland, New Zealand and its effects on the evolution of the lower crust. Journal of Structural Geology, 23 (4): 693. Daczko, N. R., Klepeis, K. A., and Clarke, G.L., 2002. Thermomechanical evolution of the crust during convergence and deep crustal pluton emplacement in the Western Province of Fiordland, New Zealand. Tectonics, 21 (4). Deer, W. A., Howie, R.A., and Zussman, J., 1992. An Introduction to the rock forming minerals. Prentice-Hall. Eckert, J. O., Newton, R.C., and Kleppa, O.J., 1991. The AH of reaction and recalibration of gamet-pyroxeneplagioclase-quartz geobarometers in the CMAS system by solution caliometry. American Mineralogist, 76: 148 160. Ferry, J. M., and. Spear., F.S., 1978. Experimental calibration of the partitioning of Fe and Mg between biotite and garnet. Contributions to Mineralogy and Petrology, 66 (2): 113 - 117. Gibson, G. M., and Ireland, T.R., 1995. Granulite formation during continental extension in Fiordland. Nature, 375: 479-482. Gibson, G. M., McDougall, I., and Ireland, T.R., 1988. Age constraints on metamorphism and the development of a metamorphic core complex in Fiordland, southern New Zealand. Geology, 16: 405-408. Hollis, J. A., Clarke, G. L., Daczko, N. R., Klepeis, K. A., and Ireland, T. R., 2003. Geochronology and geochemistry of high-pressure granulites of the Arthur River Complex, Fiordland, New Zealand: Cretaceous magmatism and metamorphism on the palaeo-Pacific Margin. Journal of Metamorphic Geology, 21 (3): 299. Ireland, T. R., and Gibson., G.M., 1998. SHRIMP monazite and zircon geochronology of high-grade metamorphism in New Zealand. Journal of Metamorphic Geology, 16: 149-167. Kimbrough, D. L., Tulloch, A.J., Coombs, D.S., Landis, C.A., Johnston, M.R., and Mattinson, J.M., 1994. Uranium-lead zircon ages from the Median Tectonic Zone, New Zealand. New Zealand Journal of Geology and Geophysics, 37: 393-419. Klepeis, K. A., Clarke, G.L., and Rushmer, T., 2003. Magma transport and coupling between deformation and magmatism in the continental lithosphere. GSA Today, 13 (1): 4 - 1 1 .

84 Landis, C. A., and Coombs, D.S., 1967. Metamorphic belts and orogenesis in southern New Zealand. Tectonics, 4: 501-518. MacKinnon, T.. C., 1983. Origin of Torlesse terrane and coeval rocks. South Island, New Zealand. Tectonophysics, 94: 383 - 392. Mattinson, J. L., Kimbrough, D.L., and Bradshaw, J.Y., 1986. Western Fiordland orthogneiss: Early Cretaceous arc magmatism and granulite facies metamorphism. New Zealand. Contributions to Mineralogy and Petrology, 92: 383-392. Mortimer, N., 1995. Triassic to Early Cretaceous tectonic evolution of New Zealand terranes: A summary of recent data and an integrated model. Proceedings of the 1995 PACRIM Congress, Australasian Institute of Mining and Metallurgy, Carlton, Victoria, Australia. Mortimer, N., Tulloch, A.J., Spark, R.N., Walker, N.W., Ladley, E., Allibone, A., and Kimbrough, D.L., 1999. Overview of the Median Batholith, New Zealand: A new Interpretation of the geology of the Median Tectonic Zone and adjacent rocks. Journal of African Earth Sciences, 29: 257-268. Muir, R. J., Ireland T. R., Weaver, S.D., and Bradshaw, J.D., 1994. Ion Microprobe U-Pb Zircon Geochronology Of Granitic Magmatism In The Western Province Of The South-Island, New-Zealand. Chemical Geology, 113 (1-2): 171. Powell, R., 1985. Regression diagnostics and robust regression in geothermometer/geobarometer calibration; the gamet-clinopyroxene geothermometer revisited. Journal of Metamorphic Geology, 3: 231 - 243. Tulloch, A. J., and Kimbrough, D.L., 2003. Paired Plutonic Belts in convergent margins and the development of high SrfY magmatism: Peninsular Ranges batholith of Baja-Califomia and Median batholith of New Zealand. Geological Society of America Special Paper, 374: 275-295. Wandres, A. M., Weaver, S.D., Shelley, D., and Bradshaw, J.D., 1998. Change from calc-alkaline to adakitic magmatism recorded in the Early Cretaceous Darran Complex, Fiordland, New Zealand. New Zealand Journal of Geology and Geophysics, 41: 1-14.

Si02 Ti02 A1203 Cr203 Fe203 FeO MnO MgO CaO Na20 K20 Total

PI 60.54 0.03 24.61 0.02 0.00 0.06 0.01 0.00 6.59 7.94 0.16 99.97

37.58 0.21 20.24 0.00 0.00 28.72 1.77 4.10 7.45 0.04 0.00 100.10

Medium P 73811 Bi 35.83 3.03 15.65 0.00 0.00 20.98 0.20 10.08 0.01 0.11 9.43 95.34

Chi 26.85 0.07 18.66 0.00 0.00 25.35 0.35 16.24 0.21 0.03 0.02 87.76

amph 40.22 0.57 14.30 0.02 0.00 19.66 0.53 8.04 11.43 1.47 1.04 97.27

G 39.13 0.05 21.59 0.03 0.00 20.40 0.54 7.72 10.80 0.02 0.00 100.29

High pressure Intrusive contact 73873A Ru Bi CPX pi 0.03 58.67 37.05 52.60 0.00 0.23 99.12 4.27 0.00 25.53 15.42 3.91 0.11 0.03 0.07 0.02 0.00 0.00 0.00 0.00 5.24 0.26 0.00 11.61 0.04 0.02 0.00 0.03 14.08 0.00 0.00 16.78 0.11 0.02 22.98 7.71 0.00 0.04 0.93 7.12 0.00 0.01 0.33 9.07 99.41 94.34 100.06 99.63

38.84 0.07 21.68 0.05 0.00 21.68 0.12 5.86 12.50 0.01 0.01 100.83

High P contact a 71239 Bi PI 37.28 57.60 0.00 2.93 16.79 26.30 0.00 0.15 0.00 0.00 13.59 0.05 0.02 0.03 14.54 0.00 8.70 0.02 0.17 6.75 0.06 9.63 95.13 99.48

chl 29.55 0.44 18.62 0.20 0.00 18.81 0.13 18.52 0.08 0.02 1.50 87.87

Epi 38.76 0.20 29.25 0.00 4.97 0.00 0.00 0.09 24.03 0.00 0.00 96.81

Amph 41.89 0.88 16.28 0.13 0.00 12.63 0.06 10.88 12.07 1.14 1.09 97.05

Oxygens

Si Ti At Cr Fe3+ Fe2+ Mn Mg Ca Na K Total

2.988 0.013 1.898 0.000 0.000 1.910 0.119 0.486 0.634 0.006 0.000 8.054

2.697 0.001 1.292 0.001 0.000 0.002 0.000 0.000 0.315 0.686 0.009 5.003

5.521 0.351 2.843 0.000 0.000 2.704 0.026 2.314 0.002 0.034 1.854 15.651

5.663 0.011 4.639 0.000 0.000 4.471 0.063 5.104 0.047 0.011 0.005 20.014

6.174 0.065 2.589 0.002 0.000 2.524 0.069 1.840 1.880 0.437 0.203 15.784

2.988 0.003 1.944 0.002 0.000 1.303 0.035 0.878 0.883 0.003 0.000 8.038

2.638 0.000 1.353 0.001 0.000 0.000 0.001 0.000 0.371 0.621 0.019 5.005

5.492 0.476 2.694 0.003 0.000 1.439 0.005 3.707 0.004 0.010 1.715 15.546

1.931 0.006 0.169 0.002 0.000 0.161 0.001 0.770 0.904 0.066 0.000 4.011

0.000 0.997 0.000 0.001 0.000 0.003 0.000 0.000 0.002 0.000 0.000 1.003

2.977 0.004 1.959 0.003 0.000 1.389 0.008 0.669 1.027 0.002 0.001 8.039

2.593 0.000 1.396 0.000 0.000 0.002 0.001 0.000 0.420 0.589 0.003 5.005

5.533 0.327 2.939 0.017 0.000 1.687 0.002 3.216 0.003 0.050 1.824 15.598

6.020 0.067 4.473 0.033 0.000 3.206 0.023 5.624 0.017 0.008 0.389 19.859

6.017 0.024 5.353 0.000 0.580 0.000 0.000 0.020 3.997 0.000 0.001 15.993

6.191 0.098 2.837 0.016 0.000 1.561 0.008 2.396 1.912 0.326 0.205 15.549

T a b l e 11 Microprobe analyses from three representative samples, medium P, high P and an intrusive contact. Resolution Is

Resolution Is

Lake Beattie

Resolution Is

73879

73811

73868

73856

73815

644i, 6382

Resolution Is Medium P

Sample Method

Directly Calibrated T (PC)

640i, 6272

815i, 7372

620i, 6242

Directly Calibrated P (Kbar)

6.726

8.62 5

7.145

7.91 5

8.285

Thermocalc t>C

T=755 ± 30

T=595 ± 37.5

T=671 ± 51

T=561± 15

T=703 ± 45

Kbar

P=9.4 ± 1.37

P=8.5 ± 1.86

P=6.9 ± 1.53

P=7.1 ± .72

P=7.5 ± 1.58

Resolution Island Intrusive Contact 73873A 73873B

Sample Method

Directly Calibrated T 765i, 7432, 7593 Directly Calibrated P (Kbar)

931 3

Wet Jacket and Breaksea region 71239

High P 71399

71410

754i, 7472

824i, 7142

682i, 7042

12.61 4

10.83 s

Thermocalc PC

T=974 ± 94.5

T=772 ± 19.5

T=830 ± 34.5

T=674 ± 79.5

Kbar

P=16.6 ± 1.83

P=14.1 ± 0.6

P=15.4 ± 1.19

P=11.2 ± 4.11

T a b l e 2: Xhermobarometry of 10 samples from the Resolution Is./ Breaksea Sound region, Fiordland. Thermobarometric techniques: (G-Bi)l. Ferry and Spear (1978), 2. Bhattacharya et al. (1992) (hw), (G-CPX) 3. Powell (1985), (G-CPX-Pl-Q) 4. Eckert et al. (1991), (G-Hbl-Pl-Q) 5. Kohn and Spear (1990), 6. GASP. Thermocalc: version 3.25. Assemblages used in Thermocalc ( + H 2 O , q) 71239: g-bi-amph-plag-epi-ilm-sph. 71399: g-bi-plag-mu-epi-sph-mic. 71410: g-bi-amph-epi-sph-me. 73811 :g-bi-plag-amph-chl-ilm. 73815: gplag-amph-chl-hem-mic. 73856:g-bi-plag-amph-chl-epi-hem. 73868:g-bi-plag-amph-ilm-mic. 73873A:g-biplag-CPX-ru-sph. 73879:g-bi-plag-mu-ky-mic.

86 Thallium isotopic evidence for ferromanganese sediments in the mantle source of Hawaiian basalts Sune G. Nielsen^ ^ Mark Rehkamper^'^, Marc Norman"^ and Alex Halliday^'^ ^Department of Earth Sciences, ETH Zurich, Sonneggstrasse 5, 8092 Zurich, Switzerland ^GEMOC, Department of Earth and Planetary Sciences, Macquarie University, NSW Împerial College, London, U.K "^Research School of Earth Sciences,. Australian National University, Canberra, ACT ^Department of Earth Sciences, University of Oxford, Oxford, 0X1 3PR, U.K Ocean island basalts (OIB) are generally thought to be the surface expression of mantle plumes. As the chemical and isotopic signatures of OIB are different to those of the mantlederived basalts erupted at mid-ocean ridges (MORE) it is clear that the sources of these types of magmatism must also be different. One model infers that OIBs originate from a deep mantle source, which was previously contaminated with ocean crust that entered the mantle via subduction zones (Hofmann and White, 1982). Based on ^^Ôs/^^Ôs and ^^Ôs/^^Ôs isotopic evidence, it was recently proposed that the Earth's core may also contribute up to 1% by weight to some mantle plumes (Brandon et aL, 1999, 2003; Walker et al, 1995, 1997). These arguments have been supported by the elevated Fe/Mn ratios of Hawaiian picrites that also exhibit Os isotope anomalies (Humayun et al, 2004). However, the lack of tungsten isotope anomalies in some of the same samples (Schersten et al, 2004) appears to exclude a core source, though it has been suggested that tungsten, Os and Fe/Mn ratios could be decoupled during core-mantle reaction processes (Humayun et al, 2004). As an altemative to core-mantle interaction, it was noted by several authors that oceanic ferromanganese (Fe-Mn) sediments display very high Pt/Os, such that they would develop high ^^Ôs/^^Ôs over time (Schersten et al, 2004; Baker and Jensen, 2004; Raviza et al, 2001). Addition of ancient FeMn sediments into a mantle source, therefore, could also explain the elevated Os/ Os ratios observed for some mantle plumes. ThalHum is one of the heaviest elements in the periodic table (isotopes ^^^Tl and ^^^Tl) and has a large ionic radius (Shannon, 1976), which makes it incompatible in mantle minerals and leads to its enrichment in melts and the continental crust (Heinrichs et al, 1980; Nielsen et al^ 2004), as well as a strong affinity for subduction-related fluids (Noll et al., 1996). Moreover, T1 is strongly enriched in both sediments (Heinrichs et al, 1980) and altered basalts (Jochum and Verma, 1996) compared with the mantle (McDonough and Sun, 1995) such that even volumetrically small contributions to the mantle should contribute significantly to the overall T1 isotope composition of the mixture. The advent of the technique of multicollector inductively coupled plasma mass spectrometry (MC-ICPMS) recently allowed the first high-precision isotope composition measurements of T1 (Rehkamper and Halliday, 1999). Ratios of T1 stable isotopes are reported as the deviation of a sample from the NIST SRM 997 T1 standard in parts per 10000: 1 O ^

1QO

87 Although T1 is a very heavy element, large isotopic fractionations of up to 20s-units between seawater and seafloor ferromanganese authigenic (chemical precipitates from seawater) sediment deposits were observed (Rehkamper and Halliday, 1999; Rehkamper et al, 2002). Thallium becomes incorporated into authigenic phases by adsorption onto the Fe-Mn oxyhydroxide particles and compared with seawater is enriched by a factor of >10^ in these deposits. Interestingly, it was pointed out by Baker and Jensen (2004) that addition of such sediments into a mantle source should generate ^^Ôs isotope anomalies that are accompanied by significant thallium (Tl) isotope variations. Such a co-variation is expected because Fe-Mn sediments have high Tl concentrations and fractionated Tl isotope compositions with s^^^Tl values as high as +15 (Rehkamper et al, 2002, 2004). Moreover, recent studies have shown that the continental crust and upper mantle are characterised by relatively constant Tl isotope ratios of s^^^Tl = -2 ± 0.5 (Nielsen et al., 2005 a,b). This indicates that the Tl isotope system may be an excellent monitor of Fe-Mn sediment additions to OIB source regions. We have therefore measured the Tl abundances and isotope compositions of nine Hawaiian picrites that have previously been investigated for ^^Ôs-^^Ôs isotope systematics (Brandon et al, 1999) and two additional picritesfromMauna Kea and Kilauea (Norman and Garcia, 1999). The Tl isotope compositions of the Hawaiian picrites vary from s^^^Tl = -3.1 to +3.8, whereby the most positive values are about 6 s-units "heavier" compared to the depleted mantle as represented by MORB (Nielsen et al, 2005a). Additionally, the Tl isotope data display a negative correlation with Cs/Tl ratios (Fig. 1).

Figure 1

8205T1 Fig. 1: Thallium isotope compositions plotted against the Cs/Tl ratio of Hawaii picrite samples. The mantle (large open square) is represented by MORB. Tick-marks represent amounts by weight of low-temperature altered MORB (open circles) and Fe-Mn sediments (small open squares) added into a mantle source.

88 Thallium and Cs exhibit very similar incompatibilities in igneous processes (Heinrichs et al, 1980; Shaw, 1952) and are therefore not expected to fractionate during partial melting or magma differentiation processes. As a consequence, it appears likely that a component with positive s^^^Tl and low Cs/Tl contributed to the Hawaiian lavas. In addition to high T1 contents and positive s^^^Tl-values, Fe-Mn sediments are known to display low Cs/Tl ratios (Hein et al, 2000). The correlation of Fig. 1 is therefore most readily explained by the addition of Fe-Mn sediments into the mantle source region of Hawaiian magmatism. However, as T1 and ^^^Os do not correlate it appears that isotope anomalies for these two tracers have different sources. Thus it is unlikely that the ^^^Os isotope anomalies are generated by additions of Fe-Mn sediments into a mantle source. Hence, this leaves coremantle interaction as the only viable current explanation for the Os isotope variations of the Hawaiian picrites. References

Baker, J.A., and Jensen, K.K., 2004. Coupled ^^Ôs-^^Ôs enrichments in the Earth's mantle - core-mantle interaction or recycling of ferromanganese crusts and nodules? Earth and Planetary Science Letters, 220 (3-4): 277-286. Brandon, A.D., et al, 1999. ^^Ôs-^^Ôs systematics of Hawaiian picrites. Earth and Planetary Science Letters, 174 (1-2): 25-42. Brandon, A.D., et aL, 2003. ^^Ôs-^^Ôs systematics of Gorgona Island komatiites: implications for early growth of the inner core. Earth and Planetary Science Letters, 206 (3-4): 411-426. Hein, J.R., et al, 2000. Cobalt-rich ferromanganese crusts in the Pacific. In: Cronan, D.S. (ed). Handbook of Marine Mineral Deposits. CRC Press, Boca Raton: pp. 239-280. Heinrichs, H., Schulzdobrick, B., and Wedepohl, K.H., 1980. Terrestrial Geochemistry of Cd, Bi, Tl, Pb, Zn and Rb. Geochimica Et Cosmochimica Acta, 44 (10): 1519-1533. Hofmann, A.W., White, W.M., 1982. Mantle plumes from ancient oceanic crust. Earth and Planetary Science Letters, 57: 421-436. Humayun, M., Qin, L.P., and Norman, M.D., 2004. Geochemical evidence for excess iron in the mantle beneath Hawaii. Science, 306: 91-94. Jochum, K.P., and Verma, S.P., 1996. Extreme enrichment of Sb, Tl and other trace elements in altered MORB. Chemical Geology, 130: 289-299. McDonough, W.F., and Sun, S.-s., 1995. The composition of the Earth. Chemical Geology, 120: 223-253. Nielsen, S.G., et al, 2004. The thallium isotope composition of the upper continental crust and rivers - An investigation of the continental sources of dissolved marine thallium. Geochimica et Cosmichimica Acta, (in press). Nielsen, S.G., et al, 2005a. Thermal fluid fluxes calculated from the isotopic mass balance of thallium in the ocean crust. Science: (submitted). Nielsen, S.G., et al, 2005b. The thallium isotope composition of the upper continental crust and rivers - An investigation of the continental sources of dissolved marine thallium. Geochimica et Cosmochimica Acta, 69 (8): 2007-2019. Noll, P.D., et al, 1996. The role of hydrothermal fluids in the production of subduction zone magmas: Evidence from siderophile and chalcophile trace elements and boron. Geochimica et Cosmochimica Acta, 60 (4): 587-611.

89 Norman, M.D., and Garcia, M.O., 1999. Primitive magmas and source characteristics of the Hawaiian plume: petrology and geochemistry of shield picrites. Earth and Planetary Science Letters, 168 (1-2): 27-44. Ravizza, G., Blusztajn, J., and Prichard, H.M., 2001. Re-Os systematics and platinum-group element distribution in metalliferous sediments from the Troodos ophiolite. Earth and Planetary Science Letters, 188 (3-4): 369-381. Rehkamper, M., and Halliday, A.N., 1999. The precise measurement of T1 isotopic compositions by MCICPMS: Application to the analysis of geological materials and meteorites. Geochimica et Cosmochimica Acta, 63 (6): 935-944. Rehkamper, M., et al, 2002. Thallium isotope variations in seawater and hydrogenetic, diagenetic, and hydrothermal ferromanganese deposits. Earth and Planetary Science Letters, 197 (1-2): 65-81. Rehkamper, M., et al, 2004. Cenozoic marine geochemistry of thallium deduced from isotopic studies of ferromanganese crusts and pelagic sediments. Earth and Planetary Science Letters, 219 (1-2): 77-91. Schersten, A., et al, 2004. Tungsten isotope evidence that mantle plumes contain no contribution from the Earth's core. Nature, 427(6971): 234-237. Shannon, R.D., 1976. Revised effective ionic radii and systematic studies of interatomic distances in halides and chalcogenides. Acta Crystallographica, A32 (5): 751-767. Shaw, D.M., 1952. The geochemistry of thallium. Geochimica et Cosmochimica Acta, 2 (2): 118-154. Walker, R.J., et al, 1997. Applications of the isotope system to geochemistry and cosmochemistry. Geochimica et Cosmochimica Acta, 61(22): 4799-4807. Walker, R.J., Morgan, J.W., and Horan, M.F., 1995. ^^^Osmium enrichment in some plumes: evidence for coremantle interaction? Science, 269: 819-822.

Compositions of Hawaiian basalts preclude eclogite mantle plumes Marc D. Norman^ J.M. Rhodes^, and M.O. Garcia^ ^Research School of Earth Sciences, Australian National University, Canberra ACT 0200 AU ^Dept. of Geosciences, University of Massachusetts, Amherst MA 01003 USA ^Dept. of Geology and Geophysics, University of Hawai'i, Honolulu HI 96822 USA Ocean island volcanoes are often considered to be the surface expression of deep-seated mantle plumes. The extreme variability of radiogenic isotopic compositions in ocean island basalts (OIB) relative to mid-ocean ridge basalts illustrates the diversity of mantle sources contributing to OIB, and has lead to the widespread view that recycled oceanic and continental crust is a common constituent of mantle plumes (Zindler and Hart, 1986). Due to the long stirring times of mantle heterogeneities, subducted crust may persist in the mantle as physically discrete packets or filaments, and be incorporated into mantle plumes from their source regions (Davies, 1990; Fametani et al, 2002). On ascent of the plume, these packets of former oceanic crust would convert to eclogite in the upper mantle and, depending on their bulk composition, contribute to melt production within the plume (Yoder, 1976). The physical and chemical consequences of recycled eclogite within mantle plumes are not well understood, and several models have been proposed over the past decade attempting to explain the geochemistry of OIB in terms of variable contributions from eclogitic and peridotitic source components. One of the focal points for the current debate has been the Hawaiian plume. This is due in part to the geochemical diversity of Hawaiian tholeiites (Frey and Rhodes, 1993; Norman and Garcia, 1999; Bhchert-Toft et al, 2003; Rhodes and Vollinger, 2004), and also to the fact that the Hawaiian swell has the largest apparent buoyancy flux (Sleep, 1990) and therefore serves as a type example of the mantle plume paradigm. Particularly vexing is the combination of unusually high Si02 contents and enriched isotopic compositions of tholeiites from Mauna Loa and Ko'olau volcanoes compared to other Hawaiian volcanoes such as Kilauea and Mauna Kea (Roden et al, 1994; Garcia et al, 1995; Rhodes and Vollinger, 2004). Hauri (1996) proposed that the major element and isotopic variability of Hawaiian tholeiites results from mixing of siliceous melts produced from recycled eclogite, with basaltic liquids derived from adjacent peridotite. Norman and Garcia (1999) criticised this model on geochemical grounds, and Yaxley and Green (1998) pointed out that siliceous melts would perish quickly in the mantle due to reaction with peridotite. However, this reaction would produce hyper-fertile peridotite rich in pyroxene and with the potential for enhanced melt production, which may be a suitable source for OIB and continental flood basalts (Yaxley, 2000). Recently, Takahashi and Nakajima (2002) and Sobolev et al (2005) proposed more extreme variants of the eclogite melting model. Takahashi and Nakajima (2002) invoked virtually complete melting of eclogite pods within the Hawaiian plume to produce primary, silica-rich tholeiitic magmas with unusually low MgO contents (Si02 53-54 wt%, MgO 6-7 wt%) as one

endmember for the Hawaiian geochemical array. Sobolev et al (2005) emphasised the unusually high Ni contents of many olivine phenocrysts from Hawaiian tholeiites, and extended the hybridisation model of Yaxley and Green (1998, 2002) to infer an olivine-free, pyroxenite source, which contributed approximately half of all Hawaiian magmas erupted over the past 1 million years. Here we examine the trace element compositions of a suite of tholeiites from Mauna Loa volcano, emphasising the Sc, V, and Ni abundances of inferred parental magmas. These particular trace elements should be sensitive to large variations in the proportions of olivine and pyroxene in the source region during melting, and so should provide a test for the presence of olivine-free pyroxenite as a major source component in the Hawaiian plume. We show that there are no apparent differences in Sc, V, and Ni abundances of parental magmas from Mauna Loa and Kilauea volcanoes, and that a peridotite source provides a better explanation for the observed compositions of these tholeiites than an olivine-free pyroxenite source. Samples and Methods Dredge samples of basaltic lavas were collected from the western and southern flanks of Mauna Loa volcano during a 1999 voyage of the Moana Wave, an oceanographic research ship operated by the University of Hawaii. A map of sample locations and petrographic descriptions of representative samples are given by Davis et al (2003). The samples range from moderately evolved, aphyric tholeiites to picrites with over 30% olivine phenocrysts (Fig. 1). Figure 1. Olivine abundances in Mauna Loa dredge samples, 1999

20 - 7 - ^ MgO 18

1 CEL 14 1 a 0 12 j E 10 i 0 81

iractlon^lon 15-1 7% MgO parental magmas

olivine cumulates

6 1

i 21 0 -f 4

8 10 12 14 16 18 20 22 24 26 28 30 32 34

% Modal Olivine Phenocrysts

Whole rock powders were prepared by crushing in either tungsten carbide or alumina ring mills. Major and trace element compositions were determined by XRF at the University of Massachusetts. Trace element compositions were also determined by solution aspiration ICPMS at the University of Tasmania following dissolution of the sample in HF-HNO3.

Results MgO contents of the Mauna Loa tholeiites range from 6-24 wt% with higher Si02 contents at a given MgO compared to tholeiites from Kilauea volcano (Fig. 2). Major and trace element compositions are consistent with olivine fractionation or accumulation as the primary process controlling the compositions of these lavas. Mauna Loa and Kilauea tholeiites have similar Ni contents at a given MgO (Fig. 2). 52

1400-

1 1

•

• •

1200-

Mauna Loa Kilauea

lOOO;

800; 49 . SiOj wt%

Ni ppm

600-

n*.

48 400; 47

15 ' MgO wt%

200;

• 25

15 MgO wt%

Discussion Estimates of parental magma compositions can be made by examining log-log plots of elements that are compatible in olivine (Mg, Ni) vs. elements that are incompatible in olivine (Al, Zr, Y, REE). On these types of diagrams, olivine accumulation can be distinguished from fractional crystallisation by a break in slope, which corresponds to the parental magma composition (Fig. 3). 200 fractional crystallisation

fractional crystallisation 30

«bi parental magmas 15-17% MgO 700-800 ppm Ni

100 90

parenfa/

magmas

ppm

80 70 60 olivine accumulation

100

1000

Ni ppm

olivine accumulation 50

1000

100

Ni ppm

93 Using this type of approach allows identification of a suite of Mauna Loa lavas that have major and trace element compositions suitable as parental magmas. These parental magmas have 15-16% MgO, 600-700 ppm Ni, 850-1050 ppm Cr, 25-27 ppm Sc, and 215-230 ppm V based on their measured compositions. The estimated parental magma compositions can be compared with predictions jfrom melting models, assuming peridotite and pyroxenite sources. Such calculations show that olivine-free pyroxenite source regions (PYX) such as those proposed by Sobolev et al (2005) should produce melts with significantly lower V and Sc contents than are observed for any Mauna Loa tholeiite (Fig. 5). Because Sc, V, and Ni are moderately compatible elements during melting, the compositions of calculated melts do not change appreciably for the extent of melting inferred for Hawaiian tholeiites (5-15%). In contrast, a peridotite source (PER) adequately accounts for the compositions of Mauna Loa parental melts, and the similar Ni-ScV compositions of tholeiitesfromKilauea and Mauna Loa (Fig. 5). 1600 1400

•

250

300

1200 - P Y X

1000 800

600 .

PER

Ni ppm

400 200

300

100

150

200 Vppm

The melting model assumes primitive mantle relative abundances in both sources and the following source modal abundances: PYX (opx 40%, cpx 35%, gnt 15%, ol 10%), PER (ol 55%, opx 25%, cpx 15%, gnt 5%, sp 2%). References

Blichert-Toft, J., Weis, D., Maerschalk, C., Agranier, A., and Albarede, F., 2003. Geochemistry, Geophysics, Geosystems, 4:10.1029/2002GC000340. Davies, G.F., 1990. Earth and Planetary Science Letters, 99: 94-109. Davis, M.G., Garcia, M.O., and Wallace, P., 2003. Contributions to Mineralogy and Petrology, 144: 570-591. Fametani, C.G., Legras, B., and Tackley, P.J., 2002. Earth and Planetary Science Letters, 196: 1-15. Frey, F.A., and Rhodes, J.M., 1993. Philosophical Transactions of the Royal Society, London, A 342: 121-136. Garcia, M.O., Hulsebosch, T., and Rhodes, J.M., 1995. American Geophysical Union, Geophysics Monograph, 92:219-239.

94 Hauri, E.K., 1996. Nature, 382: 415-419. Norman, M.D., and Garcia, M.O., 1999. Earth and Planetary Science Letters, 168: 27-44. Rhodes, J.M., and Vollinger, M.J., 2004. Geochemistry, Geophysics, Geosystems, 5: 10.1029/2002GC000434. Roden, M.F., Trull, T., Hart, S.R., and Frey, F.A., 1994. Geochimica et Cosmochimica Acta, 58: 1431-1440. Sleep, N.H., 1990. Journal of Geophysical Research, 95: 6715-6736. Sobolev, A.v., Hofmann, A.W., Sobolev, S.V., and Nikogosian, 2005. Nature, 434: 590-597. Takahashi, E., and Nakajima, K., 2002. American Geophysical Union, Geophysics Monograph, 128: 403-418. Yaxley, G.M., and Green, D.H., 1998. Schweiz. Mineral Petrogr. Mitt., 78: 243-255. Yaxley, G.M., 2000. Contributions to Mineralogy and Petrology, 139: 326-338. Yoder, H.S., 1976. Generation of basaltic magma. National Academy of Science, Washington DC. Zindler, A., and Hart, S.R., 1986. Earth and Planetary Science Letters, 14: 493-571.

95 The geology of the Port Macquarie-Tacking Point coastal tract

David Och, Evan C Leitch and Graziella Caprarelli Dept of Environmental Sciences, University of Technology, Sydney The presence of a wide range of well-exposed rocks including high-pressure metamorphics, serpentinite, gabbro, a variety of mostly mafic dyke rocks, disrupted sedimentary rocks and commonly pillowed basalt has long attracted attention to the coastal outcrops between Port Macquarie and Tacking Point. These rocks, seemingly more varied and more highly deformed than those found in the poorly exposed inland to the southwest, have commonly been interpreted as a discreet NNW trending zone possibly defining a major fault (Brunker et ^z/., 1970; Barron et al, 1976; Leitch, 1980a; Gilligan et a/., 1987) or serpentinite melange zone (Aitchison et a/., 1994). However, our recent investigations indicate that the rocks are largely a continuation of those to the southwest that are probably equally varied and comparably deformed and we have abandoned use of the term Port Macquarie complex used by several earlier workers for the coastal rocks. Rather they are considered part of Port Macquarie Block and provide a more detailed geological history for the eastem part of the block than has been read from inland exposures (Leitch, 1980b). This note summarises our present understanding of the geological development of the coastal tract. A detailed geochemical study of the igneous rocks is presently being undertaken fi-om which only preliminary conclusions are here indicated. Units recognized are described below. Port Macquarie serpentinite melange Extensive exposures of chrysotile-lizardite serpentinite found between Town and Rocky beaches comprise the Port Macquarie serpentinite melange, a name also applied to smaller bodies further south. The melange consists mostly of massive serpentinite variably transected by zones of foliated serpentinite that range from discrete narrow veins to wide schistose bands, the latter enclosing phacoids of the massive rock. Masses of Watonga Formation rocks (broken formation composed of altered basalt, chert, siltstone and sandstone) exceeding 100m in longest exposed dimension are completely surrounded by schistose serpentinite whereas metre-scale phacoids of blueschist and variably altered mafic rock (rodingite) occur less commonly. Contacts between serpentinite melange bodies and surrounding rocks are sharp. The serpentinite adjacent to the contacts is sheared and locally talcose whereas the immediate country rocks are variously sheared, brecciated or show little localised deformation. Little deformed dolerite, 'lamprophyre', and gabbro bodies have intruded the serpentinite melange. The Port Macquarie serpentinite melange shares many characters with serpentinite bodies elsewhere in New England that are associated with plagiogranites yielding Early Cambrian ages (c.535 Ma) (Aitchison and Ireland, 1995).

96 Rocky Beach Metamorphic Melange Rounded phacoids of eclogite, blueschist, omphacitite and tremolite marble embedded in a chlorite-actinolite schist matrix characterize two lenses of metamorphic melange apparently totally surrounded by serpentinite melange at the north end of Rocky Beach. Collectively the rocks show a complex metamorphic history involving both prograde and retrograde blueschist facies metamorphism separated by an episode of static eclogitic metamorphism at a pressures in excess of 1.8 Gpa (>54 km burial) and a temperature of about 560®C (Och et a/.,2003). The origin of the melange structure is incompletely established but it probably formed during a late greenschist metamorphic episode with the matrix a product of complete retrogression of earlier high pressure rocks and the locus of most of the accompanying strain. K-Ar dating of white mica suggests that retrograde blueschist metamorphism probably occurred around 470 Ma (Middle Ordovician) (Fukui et al.,1995), Watonga Formation The Watonga Formation, mostly chert and slaty siltstone with less common basalt and uncommon lithic sandstone, is the most widespread unit in the coastal tract and forms most of the eastem part of the Port Macquarie Block. Many contacts between rock-types are shears and bedding is discontinuous. Boudinage of more competent rocks produced lenticular melange and rarely pseudoconglomerate, and the unit is mostly broken formation. Despite this, several chert-dominated zones have been mapped, and coherent basalt masses are locally prominent. Notable features of the chert are the very discontinuous nature of the thin intemal layering, mostly defined by the presence of chloritic films, and the widespread occurrence of polyclinal folds and numerous small faults of diverse orientation. Both pillowed and sheet basalt flows are widespread and accompanied by breccias, jasper pods, narrow dykes and rarely interpillow limestone. The geochemistry of many basalts, all of which are significantly altered, is MORB-like although Aitchison et al (1994) recorded an alkaline basah and the incorporation of chert blocks and clastic sedimentary rocks within basalt flows and breccia masses suggests that some are the product of off-axis activity. Microfossils from Watonga Formation chert indicate a middle Palaeozoic age. Similar rocks are widely known from the New England Fold Belt the melange character of which is widely accepted as a product of Carboniferous subduction-accretion. Mafic intrusive rocks A wide range of mostly mafic igneous rocks has intruded the Port Macquarie serpentinite melange and the Watonga Formation. The rocks are texturally diverse and form syn-, late-, and post-deformational bodies that vary considerably in their degree of alteration. On the basis of their geochemistry we have divided them into two groups: a group of broadly calcalkaline affinity, and a suite that shows transitional calcalkaline-MORB character. The former are more deformed and altered than the latter which are hence considered younger. Calcalkaline mafic rocks Serpentinite melange at Town Beach is intruded by small bodies of dark gabbro and later inhomogeneous pegmatitic gabbro collectively referred to as Town Beach gabbro. These rocks are extensively altered, contain talcose serpentinite xenoliths and are cut by narrow mylonite zones.

97 North from Tacking Point to Miners Beach dykes and small stocks of dolerite and gabbro emplaced within the Watonga Formation are collectively referred to as the Tacking Point igneous complex. The igneous rocks are texturally diverse but of very similar calcalkaline geochemical character. They have been considerably altered, probably under greenschist facies conditions, and are broken by shear zones and numerous, mostly small displacement, faults. Gabbros include both leucocratic and melanocratic variants with prominent magmatic layering, pyroxenite schlieren, pegmatitic segregations and narrow aplite segregations. The igneous rocks clearly intrude broken formation around Tacking Point and a prominent chertdominated zone on the headland at the south end of Miners Beach. Watonga Formation rocks have been incorporated within the igneous material and also form screens separating intrusions. A prominent mass of basaltic rock 100m across is surrounded by gabbro. This we

interpret as a roof pendant which, together with the numerous small mafic intrusions, the overall heterogeneous character of the complex, and the presence of numerous stoped blocks of country rocks, suggests the complex represents the top of a mafic body emplaced during ongoing brittle deformation. Considered co-magmatic with the Tacking Point igneous complex are a suite of disrupted highly altered metadolerite dykes most common in the Miners Beach region where they intrude earlier deformed Watonga Formation. The more deformed and metamorphosed of Leitch's (1980b) Karikeree Metadolerite are possibly members of this suite. Fine-grained, dark green - black dykes, termed lamprophyres by Barron et a/. (1976), are readily distinguished both in the field and geochemically, and here referred to as 'melanocratic dykes'. They form narrow (<lm) planar bodies with prominent chilled margins, commonly succeeded by a variably vesicular zone, and with a coarser green centre. The rocks, moderately altered with widespread development of chlorite, have a characteristically low silica content (<45 wt%) and high MgO (>16 wt%) and Ni and Cr contents (325ppm and 1056 ppm respectively). They were interpreted by Aitchison et al (1994) as of boninitic affinity. The dykes cut the Tacking Point igneous complex as well as the Watonga Formation and serpentinite melange. They are offset by late small-displacement faults. The age of the calcalkaline rocks at Tacking Point is not well established. Reports of Early Carboniferous radiometric ages for the gabbro and melanocratic dykes are referred to by Aitchison et al. (1994) for Tacking Point rocks but we are unaware of any satisfactory documentation of these ages. We favour a somewhat younger age (Late Carboniferous or Early Permian) based on their emplacement subsequent to the development of Watonga Formation melange. Dolerites of transitional calcalkaline-MORB affinity South from Shelly Beach mildly altered and little deformed dolerite dykes transect the Watonga Formation, serpentinite melange, and the Tacking Point igneous complex. The dykes have well-defined chilled margins, are commonly plagioclase-phyric, and range in width from about 0.2m to 15m. These dykes have higher Ti02 content and lower Ni and Cr contents than other mafic intrusive rocks at Port Macquarie, except for the higher Ti02 (>2 wt%) Karikeree Metadolerite samples reported by Leitch (1980b) that we are presently analyzing for trace elements. Although the latter include more highly altered and deformed rocks than those from the coastal exposures, similarly undeformed and little altered dolerites

98 are also present and the rocks are likely correlatives. The Karikeree Metadolerite intrudes probable Early Permian slates and if correlation between them and the 'transitional' dolerites is accepted then the 'transitional' rocks are no older than late Early Permian. Felsic dykes Uncommon felsic dykes that yield zircon SHRIMP dates as young as 240 Ma (Aitchison and Ireland, 1995) intrude the Town Beach gabbro and the Tacking Point igneous complex. The rocks are considered minor representatives of the New England Batholith larger masses of which were emplaced elsewhere in the eastem New England Fold Belt in the Middle-Late Triassic. Geological synthesis Our study of the coastal rocks reveal that the eastem Port Macquarie Block has had a protracted geological history. The protolith of the Rocky Beach metamorphic melange was

carried to mantle depths in a (?)Neoproterozoic - Ordovician subduction system. A slice of the resulting high-pressure metamorphic rocks was later incorporated into the mantle section of the upper plate from where it was subsequently mobilized, together with the surrounding ultramafic rocks, and emplaced within the Watonga Formation. Neither the timing of these movements nor the mechanisms whereby the rocks rose into the upper crust are clearly established. K-Ar dating indicates that by the Middle Ordovician the high-pressure rocks had undergone blueschist facies retrograde metamorphism and subsequently remained below the argon closure temperature for phengite (Fukui et tz/.,1995).The widespread presence of retrograde glaucophane-rich rocks of this age associated with serpentinite in the southern New England Fold Belt indicates that this was the product of a regional event, possibly related to disruption of the subduction system. Retrogression involved hydration of the rocks and was accompanied by the formation of irregular massive patches and veins of glaucophane and phengite in some phacoids, but highly foliated fabrics involving these phases in others. While it is tempting to relate this retrogression to the alteration producing the Port Macquarie serpentinite this is unlikely because, with the exception of a small lens of antigorite serpentinite within the Rocky Beach melange, chrysotile is the dominant serpentine mineral. Textures suggest that chrysotile is the primary hydration product and is not replacing antigorite that would have been the stable serpentine mineral under the blueschist metamorphism (see Och et aL, 2003). It is more likely that serpentinisation accompanied the retrogressive greenschist facies formation of the chlorite-actinolite matrix of the Rocky Beach melange and that both it and the serpentinite were multiply deformed and their melange structures developed during thrust emplacement within the Watonga Formation. This could have occurred during the latest Carboniferous, immediately before deposition of Early Permian sedimentary rocks that locally unconformably overly serpentinite elsewhere in New England. Calcalkaline igneous activity post-dated emplacement of the serpentine melange. Gabbro and metadolerite were probably emplaced during the waning stages of the same period of compressive tectonism that had earlier led to thrust emplacement of the Port Macquarie and Rocky Beach melanges. Melanocratic dykes were the last products of this activity. The nature of these igneous rocks suggests the presence of a contemporaneous subducting slab beneath this region, possibly a steepening remnant of the slab responsible for Devonian and Carboniferous volcanism in the westem New England Fold Belt, that fragmented at about this time (CaprareUi and Leitch, 2001).

99 The dykes of transitional calcalkaline-MORB geochemical character are of uncertain age. They post-date most deformation in the eastern part of the Port Macquarie Block, suggesting post-Permian emplacement, but they are geochemically similar to Early Permian volcanics described by Caprarelli and Leitch (2001) from further south in the New England Fold Belt and may share some geochemical characters with deformed mafic volcanic rocks from the Early Permian Nambucca Slate Belt to the north (Leitch and Asthana, 1984). It is possible that intense Permian deformation in eastem New England was focused in weak Early Permian clastic sequences and that already tectonised older rocks were little affected, accounting for the contrast in the structural state of the dolerites emplaced in different geological units. References Aitchison, J.C., Blake, M.C., Flood, P.G., and Jayko, A.S., 1994. Paleozoic ophiolitic assemblages within the southern New England orogen of eastem Australia: Implications for the growth of Gondwana. Tectonics, 13: 1135-1149. Aitchison, J.C., and Ireland, T.R., 1995. Age profile of ophiolitic rocks across the Late Palaeozoic New England Orogen, New South Wales: implications for tectonic models. Australian Journal of Earth Sciences, 42: 11-23. Barron, B.J., Scheibner, E., and Slansky, E., 1976. A dismembered ophiolite suite at Port Macquarie, New South Wales. Records of the Geological Survey of New South Wales, 18: 69-102. Brunker, R.L., Offenberg, A.C., and Cameron, R.G., 1970. Hastings 1:250 000 Geological Sheet. Geological Survey of New South Wales, Sydney. Caprarelli, G., and Leitch, E.C., 2001. Geochemical evidence from Lower Permian volcanic rocks of northeast New South Wales for asthenospheric up welling following slab breakoff Australian Journal of Earth Sciences, 48: 151-166. Fukui, S., Watanabe, T., Itaya, T., and Leitch, E.C., 1995. Middle Ordovician high PT metamorphic rocks in eastem Australia: Evidence from K-Ar ages. Tectonics, 14: 1014-1020. Gilligan, L.B., Brownlow, J.W., and Cameron, R.G., 1987. Tamworth-Hastings 1:250 000 Metallogenic Map SH/56-13, SH/56-14. Geological Survey of New South Wales, Sydney. Leitch, E.C., 1980a. The Great Serpentine Belt of New South Wales: diverse mafic-ultramafic complexes set in a Palaeozoic arc. In: Panayiotu, A. (ed). Ophiolites. International Ophiolite Symposium, Cyprus 1979, Ministry of Agriculture and Natural Resources, Cyprus, Proceedings:pp. 637-648. Leitch, E.C., 1980b. Rock units, structure and metamorphism of the Port Macquarie Block, eastem New England Fold Belt. Proceedings of the Linnean Society of New South Wales, 104: 273-292. Leitch, E.C., and Asthana, D., 1985. The geological development of the Thora district, northern margin of the Nambucca Slate Belt, eastem New England Fold Belt. Proceedings of the Linnean Society of New South Wales, 108: 119-140. Och, D.J., Leitch, E.C., Caprarelli, G., and Watanabe, T., 2003. Blueschist and eclogite in tectonic melange. Port Macquarie, New South Wales, Australia, Mineralogical Magazine, 67: 609-624.

100

Metamorphism in the southern New England Fold Belt - an overview Robin Offler Discipline of Earth Sciences, School of Environmental and Life Sciences, University of Newcastle, Callaghan, NSW 2308 The New England Fold Belt is a major north-trending Palaeozoic structure extending for over 1600 km along the eastern margin of Australia. The southern part is divided into three segments that from west to east consist of a volcanic arc, forearc basin (Tamworth Belt; TB) and an accretion-subduction complex (Tablelands Complex; TC). Dividing the TB and TC is the Peel-Manning Fault System (PMFS) a major structure ca 350 km in length separating the Late Silurian to Carboniferous arc and forearc basin sequences from Middle Silurian to Early Carboniferous accretionary-subduction sequences (Figure). The PMFS is made up of serpentinised and stongly deformed ophiolitic rocks of supra-subduction zone affinity containing exotic blocks of varying age, origin and composition (Sano et al, 2004) adjacent to which are fault bounded Early Palaeozoic sequences. Metamorphic pattern The rocks in the SNEFB record a complex tectonothermal history produced by changes in tectonic setting through time. As a result, a complicated metamorphic pattern has evolved that subsequently has been disassembled during the Late Permo-Triassic Hunter-Bowen Orogeny (HBO). Arc-forearc basin (Tamworth Belt). In the Tamworth Belt, petrographic, x-ray diffraction and phytoclast reflectance (Ro) studies reveal two main types of metamorphism, namely burial and contact. Burial metamorphism This is recorded in the Early Palaeozoic sequences adjacent to the PMFS and in DevonianCarboniferous sequences of the TB and Hastings Block (HB). It is manifested by mineral assemblages of the zeolite and prehnite-pumpellyite facies. Overall, zeolite facies assemblages are more common in the Carboniferous sequences and prehnite-pumpellyite facies assemblages in the Devonian sequences of the TB. Locally, the association laumontiteprehnite ± pumpellyite is present in a Late Devonian, intraoceanic arc sequence at Keepit Dam (Offler et al, 1997). Metamorphic zones defined by the presence of heulandite/clinoptilolite ± stilbite, laumontite, and prehnite-pumpellyite ±epidote have been recognized with increasing stratigraphic depth in the TB. Mixed-layer clays, chlorite-smectite and illite-smectite are common in the lowest grade rocks. Phytoclast reflectance and Kiibler Index (KI; Kiibler, 1967) studies support the burial metamorphic origin of the assemblages in the Devonian-Carboniferous sequences. Contact metamorphism Superimposed on the burial metamorphic pattem is contact metamorphism associated with the emplacement of Permian and Triassic granitoids. Chlorite-vermiculite, laumontite ± prehnite, prehnite ± pumpellyite, epidote-chlorite ± actinolite ± biotite and biotite ±homblende form in

101 rocks of appropriate composition with increasing grade within the aureole of the Barrington Granodiorite, Tamworth Belt; rare wairakite and hypersthene develop locally (Baker, 1983; R.Offler unpubl. data). Adjacent to the Triassic granitoids in the Hastings Block (HB), illite+chlorite-vermiculite, chlorite-muscovite, biotite-cordierite progressively develop in the aureoles (Offler et al, 1997). Throughout the forearc sequences, veins of extensional and shear origin are developed that contain assemblages that mirror those found in the host rocks suggesting that thickening during the HBO had not been sufficient to change the P-T conditions. Subduction-accretion Complex (Tablelands Complex). The crystallinity (KI) of K-white micas in pelites and assemblages in meta-basites and greywackes, reveal a complex metamorphic pattern in the TC. Juxtaposed fault bounded blocks show inter and intra variation in grade of metamorphism (Offler and Hand, 1988; Black, 1998; Morton, 1998). In many of the blocks, younging of the sequences towards the forearc basin and an increase in grade from the bottom to the top of the sequence, is apparent which is typical of accretionary terranes (Merriman and Frey, 1999). The studies also indicate that zeolite, prehnite-pumpellyite and greenschist facies assemblages are dominant and that locally pumpellyite-actinolite and blueschist facies are developed. Superimposed on this pattem in some areas, is a high T-low P imprint associated with the emplacement of the granites of the S-type Hillgrove Suite (-^-300 Ma; Collins et al^ 1993) in a back-arc basin setting produced during slab rollback (Jenkins et aL, 2002). Permian rift basins in the aceretion-subduction complex. Early Permian sequences show burial (zeolite to prehnite-pumpellyite facies. Manning Group; Jenkins and Offler, 1996) and/or orogenic metamorphism with the highest grade (greenschist facies; Leitch, 1976; Offler and Brime, 1994) reached in the Nambucca Block (NB). In the NB, Leitch (1976) delineated four zones of regional metamorphism, based on prehnite only in l^stilpnomelane-pumpellyite in 2 '=^>pumpellyite-actinolite and lack of prehnite in 3 •^^^disappearance of pumpellyite in 4. Elsewhere in the accretion complex, the grade of the Permian rocks is low (zeolite facies). P-T conditions Arc-forearc basin. Few studies involving the determination of P-T conditions have been carried on the rocks in this belt. For the Mid to Late Devonian sequences at Glenrock Station, Offler (1991) has suggested that P-T conditions of~l. 7 kb and ~300^C prevailed during burial which accords with the back arc setting existing at this time (Offler and Gamble, 2002). In other areas, the existence of prehnite-pumpellyite facies assemblages in many of the Devonian sequences limits T to <270®C and P<4 kb depending on Fe content of the coexisting phases (Powell et al, 1993). Exotic blocks and aceretion-subduction complex (TC) P-T conditions changed substantially throughout the history of the complex. In the exotic blocks containing blueschists of Ordovician age (Fukui et al, 1995), assemblages formed during uplift at T <380''C and P ~ 8 kb (Offler, 1999). Conditions changed for blueschists in sequences of younger age than the exotic blocks (Nowendoc, T = 280®C, P = 6.5 kb (Dl); T =

102 345''C; P = 5.7 kb (D2); M.Handpers. comm,; Bingara P = 5.5 kb, R.Offler unpubl. data) and substantially at the time of the emplacement of the -300 Ma granites of the Hillgrove Suite. Peak metamorphic conditions of T = 600-650®C and P = 2-3 kb were attained for the Tia Complex (Dirks et a/.,1992) and 660''C and 2.8-3.8 kb for the Wongwibinda Complex (Farrell, 1992). b cell data obtained from K-white micas support the P-T conditions determined for the blueschists in the TC (x = 9.038; geothermal gradient 16® C/km). They also indicate a rise in geothermal gradient during subduction of younger sequences (x = 9.015; geothermal gradient 23®C/km). This rise in geothermal gradient may be due to subduction of younger, warmer oceanic crust but other factors such as advective heat transfer from fluid flow may have also played a role. Rift basins. Data available is limited because few Early Permian sequences in the rift basins have been studied. Those in the Nambucca Block (NB) have undergone burial metamorphism at T and P-^ 1.5 kb suggesting a geothermal gradient of ~ 54''C/km (R.Offler unpubl. data). Subsequently, they were deformed at T = 360-370''C and P = 4-7 kb (Leitch, 1976); recent studies suggest temperatures and pressures were lower (T = 278^C, P = --3 kb; geothermal gradient ~25®C/km; R.Offler unpubl. data). Thickening caused by the southward push of the Coffs Harbour Block during oroclinal bending led to this moderately low geothermal gradient operating during deformation (Johnston et al, 2002). In the Manning Group reflectance studies indicate maximum burial metamorphic temperatures of 260®C have been attained and a geothermal gradient of 24^C/km (Jenkins and Offler, 1996) much lower than in the NB suggesting that crustal extension was less in this basin than in the NB. References

Baker, C.K., 1983. Phytoclasts in metamorphic rocks. PhD Thesis, University of Newcastle (unpubl). Black, K., 1998. Structural analysis of the rocks in the Limbri Area, NSW: implications for their tectonic setting and plate motion directions. BSc (Hons), University of Newcastle (unpubl). Collins, W.J, Offler, R., Farrell, T.R., and Landenberger, B., 1993. A revised Late Palaeozoic-Early Mesozoic tectonic history for the southern New England Fold Belt. In: Flood, P.G., and Atchison, J.C., (eds). New England Orogen, Eastern Australia. Department of Geology & Geophysics, University of New England, Armidale, NSW: 69-84. Dirks, P.H.G.M., Hand, M., Collins, W.J., and Offler, R., 1992. Structural-metamorphic evolution of the Tia Complex, New England Fold Belt: thermal overprint of an accretion-subduction complex in a compressional back arc setting. Journal of Structural Geology, 14: 669-668. Farrell, T.R., 1992. Deformation, metamorphism and migmatite genesis in the Wongwibinda Complex, eastern Australia. PhD Thesis University of Newcastle, Australia (unpubl). Fukui, S., Watanabe, T., Itaya, T., and Leitch, E.C., 1995. Middle Ordovician high PT metamorphic rocks in eastern Australia: Evidence from K-Ar ages. Tectonics, 14: 1014-1020. Greentree, M.R., 1998. The "Keepit Volcano:" A source for the Devonian Mostyn Vale and Baldwin Formations? BSc (Hons) thesis, Macquarie University (unpubl).

103 Jenkins, R.B., and Offler, R., 1996. Metamorphism and deformation of the Early Permian extensional basin sequence: the Manning Group, southern New England Orogen. Australian Journal of Earth Sciences, 43: 423436. Jenkins, R.B., Landenberger, B., and Collins, W.J., 2002. Late Palaeozoic retreating and advancing subduction boundary in the New England Fold Belt, New South Wales. Australian Journal of Earth Sciences, 49: 467-489. Johnston, A., Offler, R., and Liu, S., 2002. Structural fabric evidence for indentation tectonics in the Nambucca Block, southern New England Fold Belt, NSW. Australian Journal of Earth Sciences, 49: 407-421. Kubler, B., 1967. La cristallinite de I'illite et les zones tout a fait superieures du metamorphisme. In: Etages Tectoniques - Collogue de Neuchdtel Avril 1966, Universite Neuchatel , a la Baconniere, Neuchatel (Suisse), 105-121. Leitch, E.G., 1976. Zonation of Low Grade Regional Metmorphic Rocks, Nambucca Slate Belt, northeastern New South Wales. Journal of the Geological Society of Australia, 22: 413-422. Merriman, R.J., and Frey, M., 1999. Patterns of very low-grade metamorphism in metapelitic rocks. In: Frey, M., and Robinson, D. (eds). Low-grade metamorphism, Blackwell Science: 61-107. Morton, M., 1998. Temporal change in geothermal gradient, subduction-accretion complex, southern New England fold Belt. BSc (Hons) Thesis, University of Newcastie (unpubl). Offler, R., 1991. Significance of metabasite assemblages in Devonian forearc basin sequences. New England Fold belt, NSW. IGCP Project 294 Meeting, Auckland, N.Z. Abstract, 39. Offler R., 1999. Origin of blueschist "knockers", Glenrock Station, NSW. In: Flood, P.G. (ed). NEO'99 Conference. New England Orogen. Regional Geology, Tectonics and Metallogenesis. University of New England, Armidale: 35-44. Offler R., and Hand, M., 1988. Metamorphism in the forearc and subduction complex sequences of the southern New England Fold Belt. In: Kleeman, J.D. (ed). Proceedings of Symposium "New England Orogen, Tectonics and Metallogenesis. " University of New England, Armidale, NSW: 78-86. Offler, R., Roberts, J., Lennox, P., and Gibson, J., 1997. Metamorphism in Palaeozoic forearc basin sequences, southern New England Fold Belt, N.S.W., Australia. In: Qian, X.L., et al (eds). Proceedings of the 30th Geological Congress 17, Part II: 241-250. Offler, R., and Brime, 1994. Characterisation of the low grade metamorphism in the Nambucca Block (NSW, Australia). Revisita Geologica de Chile, 21: 285-293. Offler, R., and Gamble, J., 2002. Evolution of an intra-oceanic island arc during the late Silurian to Late Devonian, New England Fold Belt. Australian Journal of Earth Sciences, 49: 349-366. Powell, W.G., Carmichael, D.M., and Hodgson, C.J., 1993. Conditions and timing of metamorphism in the southern Atibiti greenstone belt, Quebec. Canadian Journal of Earth Sciences, 32: 787-805. Sano, S., Offler, R., Hyodo, H., and Watanabe, T., 2004. Geochemistry and chronology of tectonic blocks in serpentinite melange of the southern New England Fold Belt, NSW, Australia. Gondwana Research, 7: 817-831.

104

105 Significance of orientation, age and geochemical composition of dykes emplaced before and during the opening of the Tasman Sea. Robin Offler^ Horst Zwingmann^ Lin Sutherland^ and Ian Graham^ ^Discipline of Earth Sciences, School of Environmental & Life Sciences, University of Newcastle, Callaghan, NSW 2308 (robin.offler@newcastle.edu.au) ^CSIRO - Division of Petroleum Resources, PO Box 1130, Perth, 6102, AustraUa and Centre of Excellence in Mass Spectrometry - School of Applied Geology, Curtin University, WA 6102, Austraha (horst.zwingmann@csiro.au) ^Geoscience, The Australian Museum, 6 College St, Sydney, NSW 2010 (lins@austmus.gov.au) Studies of basaltic dykes emplaced in Carboniferous and Permian sequences, have been carried out to ascertain their age, orientation and composition. They are exposed as isolated intrusions or swarms on wave cut platforms, railway and road cuttings, north, northeast and south of Newcastle (Embleton et al 1981; Maxwell, 1990; Sutherland and Graham, 2003). KAr dating of plagioclase feldspars extracted from the dykes reveal ages varying from 266 to 59 Ma indicating they were emplaced prior to and during the opening of the Tasman Sea (84 53 Ma). During this period, dyke orientation changed from meridional to WNW/NW ESE/SE and then to NE - SW with time indicating a change in far field extension direction a3 related to the breakup of Gondwana and formation of the Tasman Sea. Between 76 and 72 Ma, dyke orientation changed from WNW/ESE to NE/SW suggesting a major reorientation in a3. This may be related to a major change in orientation of the ridge axis at this time (Gaina etal, 1998). Major, trace and REE analyses of the dykes show that low-Ti (Ti02 < 2.0 wt%; TiA^=400700) and high-Ti (Ti02 > 2 wt%; TiA^=300-500) basalts are present (Figs. 1 and 2). Zr/Ti02Nh/Y and TiA^-NbA^ ratios show that most of the former are tholeiitic, the latter are mildly alkaline and some are transitional between the two magma types. Younger basalts (72-59Ma) have a tholeiitic magmatic affinity. LREE enriched chondrite normalized patterns ([LaATbJN = 19.7-9.04) are typical of the mildly alkaline basalts and LREE, flat, depleted patterns ([La/Yb]N =1.72-1.41) of the tholeiitic basalts. Transitional varieties have [LaAfbjN ratios varying froml2.2 to 6.63. The high-Ti basalts have characteristic OIB-like, primitive mantle normalized pattems that show moderately strong enrichment in incompatible elements (LaN = 74-32, LuN = 6.22-2.16) contrasting with the transitional samples (LaN = 49-26, LuN = 3.51-3.38) and the low-Ti basalts (LaN = 11-4, LuN = 7.43-2.57). A few low-Ti basalts show pattems characteristic of continental tholeiitic flood basalts. The Nb-Zr-Y discrimination diagram of Meschedes (1986) reveals that the high-Ti basalts have a within plate alkali basaltic affinity and the low-Ti basalts N-MORB/volcanic arc characteristics. An intercontinental rift setting for most of the high-Ti basalts is suggested by the Y/15-La/10-Nb/8 discrimination diagram of Cabanis and Lecolle (1989), and the low-Ti basalts either have signatures of weakly enriched E-MORB, calc-alkaline volcanic basalts

106 (CAB) or back arc basin basalts (BAB). NbA'-ZrA^ and ThA^-NbAT) plots (Fitton et al, 1997; Pearce and Peate 1995) reveal that the basalts with the BAB signature have N-MORB affinities and those with the CAB signature show contamination by continental crust. Prior to 76Ma, mildly alkaline basalts were derived from an enriched mantle source and segregated at depths varying from ~60 to 90km according to their Nb/La and LaAT) ratios (Abdel-Rahman, 2002). Subsequently, tholeiitic magmas derived from a mixed lithosphericasthenospheric or lithospheric mantle source were emplaced and segregated at depths of 3545km or less, reflecting a change in crustal thickness related to the continual opening of the Tasman Sea. This study shows that the variation in orientation and composition of the basaltic dykes through time are a reflection of the change in extension direction and crustal thickness that occurred prior to and during the formation of the Tasman Sea. References Abdel-Rahman, A.M., 2002. Mesozoic volcanism in the Middle East: geochemical, isotopic and petrogenetic evolution of extension-related alkali basalts from Central Lebanon. Geological Magazine, 139: 621-640. Cabanis B., and Lecolle, M., 1989. Le diagramme La/10-Y/15-Nb/8:an outil pour la discrimination des series volcaniques et la mise en evidence des processus de melange et /ou de contamination crustale. Comptes Rendus de VAcademic des Sciences Series //, 309: 2023-2029. Embleton, B.J.J., Schmidt, P.W., Hamilton, L.H., and Riley, G.H., 1981. Dating volcanism in the Sydney Basin: evidence from K-Ar ages and palaeomagnetism. In: Sutherland, F.L., Franklin, B.J., and Waltho, A.E. (eds). Volcanism in eastern Australia: with case histories from New South Wales. Special Publication!, NSW Division, Geological Socviety of Australia: pp. 59-72. Fitton, J.G., Saunders, A.D., Norry, M.J., Hardason, B.S., and Taylor, R.N., 1997. Thermal and chemical structure of the Iceland plume. Earth and Planetary Science Letters, 153: 197-208. Gaina, C., Roest, W.R., Muller, R.D., and Symonds, P., 2004. The Opening of the Tasman Sea: A gravity anomaly animation. Earth Interactions, 200-4: 1-23. Maxwell, S., 1990. Geochemical characterization of dykes intruding the northeast Sydney Basin, Australia. In: Parker, A.J., Rickwood, P.C., and Tucker, D.H. (eds). Mafic dykes and emplacement mechanisms. Balkema, Rotterdam, pp. 415-419. Meschede, M., 1986. A method of discriminating between different types of mid-ocean ridge basalts and continental tholeiites with the Nb-Zr-Y diagram. Chemical Geology, 56: 207-218. Pearce, J.A., and Peate, D.W., 1995. Tectonic implications of the composition of volcanic arc magmas. Annual Review of Earth and Planetary Science Letters, 23: 251-285. Sutherland, F.L., and Graham, I.T., 2003. Geology of Barringt6on Tops Plateau: Its rocks, minerals and gemstones, New South Wales, Australia. The Australian Museum Society, Sydney. 56p. Winchester J.A., and Floyd, P.A., 1977. Geochemical discrimination of different magma series and their differentiation products using immobile elements. Chemical Geology, 20: 325-343.

107

Fig. 1 Nobbys Head alkaline basalt in thin-section (PPL) showing glomeroporphyritic texture and prominent Titaniferous-bearing augite, Fe-Ti oxides and brown amphibole.

Figure 2. Norah Head tholeiitic basalt in thin-section (PPL) showing microporphyritic texture with clinopyroxene, altered olivine and interstitial glass and chlorite.

108 Slurry flow and structures formation in a magma mush: the Basement Sill, McMurdo Dry Valleys, Antarctica Nick Petford\ David Wertheim^ Dougal Jerram^ and Jon Davidson^ ^ Centre for Earth and Environmental Science Research, Kingston University, UK ^ Department of Earth Sciences, University of Durham, UK

The McMurdo Dry Valleys magmatic system, Antarctica, forming part of the Ferrar dolerite Large Igneous Province, comprises a vertical stack of four interconnected sills linked to surface flows of the Kirkpatrick flood basalts\ The lowermost intrusion, the Basement Sill, offers unprecedented exposure through a magmatic slurry flow (congested melt-particle mixture), comprised of abundant Opx pheoncrysts (MgO up to 20%). The overall geometry of the axially confined slurry is tongue-like, with the sill margins relatively aphiric^ The 3D nature of the exposure (100%) permits petrographical and structural observations to be made in unprecedented detail. A wide range of structures related to transport phenomena and deformation of the slurry during flow, including grading, layering and melt segregation have been recorded and are being used as the basis for sophisticated physics-based modeling of the magma emplacement process^. Key properties of the Opx tongue that have bearing upon emplacement and solidification include microscale matrix properties (grain size, roughness, porosity and permeability, packing density and fluctuations, see Fig Ic and Table 1), and the assumed fluid properties (viscosity and density) of the intergranular melt phase. It is noteworthy that features long recognized in classical layered intrusions formed here on a timescale (short) governed by the local cooling rate of a c. 300 m thick sill. Analysis Our preliminary analysis of the structures in the Opx tongue is based on detailed examination of rock textures; theory is modified from relevant studies in chemical engineering dealing with multiphase flow. It has been shown recently that fluid flow through an immobile, heterogeneous porous bed of a non-linear, dilatant fluid exhibits structure formation^ (layers, clots etc) in the cross streamwise direction. To understand this phenomenon, lengthscales are distinguished. Domains in the Opx tongue consist of grains of more or less equal size, but the packing is non-regular. The smallest relevant lengthscale is obviously the grain size. The next scale is the 'cell' scale, consisting of a small assembly of bed particles (in practical terms, one particle and its immediate neighbours). The assembly is the smallest set on which a porosity and average fluid velocity can be defined. If fluctuations on the cell scale are considered, any irregularly packed granular bed is heterogeneous. Qualitatively the formation of structures (Fig.l) is understood as follows. A small patch in which the permeability is lower than its immediate surroundings will possess a relatively lower fluid velocity in the direction of the average flow. By continuity the flow in the regions in the cross streamwise direction around the lower permeability patch must be directed towards it, thus relatively increasing the magnitude of seepage flow in these regions. At the same time the fluid flow velocity dips in the streamwise direction. All this holds for Newtonian fluids. Non-Newtonian fluids are distinguished as either dilatant or pseudo-plastic, depending on whether the tangent viscosity decreases or increases with increasing shearing strain rate. The seepage flow of a dilatant non-Newtonian fluid through a porous medium (Opx tongue) is discussed and the findings are

109 applied to explain a possible mechanism for the genesis of the observed layering. Information on conditions under which the touching Opx framework may collapse to expel interstitial fluid is also under investigation"^. A distinction between structures formed during transport, where the shearing regime dominates, and post- emplacement features due to mechanical compaction of the mush, is required.

Fig 1. a) Rhythmic Layering in Basement Sill Opx tongue, b) Detail of plagioclase-rich segregation, c) Granular texture of Opx tongue enhanced to reveal porosity (green) in three regions of the image. The quality of the derived image analysis data, summarized in Table 1, is dependent on the threshold value, but shows that variations in Opx packing density on lengthscales < 10 mm can in principle be quantified across the section. Estimated error in porosity based on pixel conversion is < 10%.

110

Region 1 2 3

White pixels 39436 6164 13600

Non-wliite pixels 120022 68 61642

Threshold 150 150 150

Porosity(%) 25 99 19

Table 1. Estimated relict porosity from analysis of regions 1-3 shown in Fig. Ic.

References 1. Marsh, B., 2004. Eos, 85: 497-502. 2. Charrier, A., and Marsh, B., 2004. Eos Trans AGU^S, v24A-03. 3. Koenders, M.A., and Kilchherr, R., 2002. Journal of Applied Physic^ D 35: 1-8. 4. Jerram, D.A., Cheadle, M.J., and Philpotts, A.R., 2003. Journal of Petrology, 44: 2033-2051.

Ill

Facies analysis, geochemistry and tectonic setting of the Frampton Volcanics, southeastern New South Wales Alice C. Plioplis and Kelsie A. Dadd National Key Centre for GEMOC and Department of Earth and Planetary Sciences, Macquarie University, Sydney, 2109, Australia The Frampton Volcanics, in the Eastem Lachlan Fold Belt, southeastern New South Wales, form part of the fill of the Early Silurian Tumut basin. They comprise felsic volcanic rocks, rare felsic and mafic dykes, and sedimentary rocks. Mapping in this study has delineated three major volcanic facies in the Frampton Volcanics, including crystal-rich rhyolite, flow-foliated rhyolite and plagioclase-crystal breccia facies as well as sedimentary facies including mudstone and sandstone facies. The volcanic facies are interpreted to have formed in a subaqueous environment as lava flows and domes. They are rhyolitic in composition and can be likened to other S-type volcanic rocks of the Lachlan Fold Belt in proximity to the Tumut basin. A palaeogeographic reconstruction shows that the Frampton Volcanics were emplaced below storm wave base in the Tumut basin following the initial rifting stages. This contrasts with interpretations for the Frampton Volcanics up to this point. Several geochemical analogues worldwide were considered with the best fit being Taupo Volcanic Zone. This suggests the Frampton Volcanics and Tumut basin occupied a backarc environment during the Early Silurian. Introduction The Frampton Volcanics (Figure 1) lie on the western edge of the Eastem Lachlan Fold Belt and have a U-Pb zircon age of 428 ± 6 Ma (Stuart-Smith et al, 1992). Basden (1990) placed the Frampton Volcanics within the Tumut Trough sequence, a basin structure floored by oceanic crust represented by mafic and ultramafic units including the Honeysuckle Beds, the North Mooney Complex and the Coolac Serpentinite, and filled with volcanic and sedimentary rocks. Stuart-Smith et al (1992) rejected this view, suggesting that the Tumut Trough was not a valid structure. They suggested the structure was a pull-apart basin called the Tumut Basin, filled by the Blowering Formation, Honeysuckle beds, Brungle Creek Metabasalt and Wyangle Formation. Scheibner and Basden (1998) placed the Frampton Volcanics on the westem edge of the Tumut Trough, adopting the model of Basden (1990) for basin formation, with the Goobarragandra Volcanics on the eastem edge (Figure 1). This paper documents the volcanic and sedimentary facies in an area of the Frampton Volcanics to place constraints on the environment of deposition and palaeogeography of the Tumut basin. The geochemistry of the silicic volcanic rocks is used to constrain the tectonic setting and provide a modem analogue.

112

Fades

The most extensive facies in the Frampton Volcanics mapped in this study is the crystal-rich rhyolite. It crops out mostly as north-south trending ridges. Crude columnar jointing is present but rare. The facies consists of flow-foliated and massive outcrop with foliation up to 10 mm in thickness and defined by ahemating of dark and light layers. Crystal size and abundance is variable, but averages between 0.5 mm and 2.5 mm and 24-36%. The groundmass is devitrified with spherulitic and micropoikilitic textures and relict perlitic cracking indicative of a previously glassy groundmass. Moulds (1999, unpubl.) noted a matrix-supported breccia

LFB S-typ» Tfolcinics g g g Prattipton "Vfelcinics ggOoobairagandra -VbUsamcs ggiLaidlaw-Vblcanios )PH|K«Uys Pbin -Vbleuiics

Figure 1. Distribution of the Frampton Volcanics in relation to other S-type volcanic rocks in the LFB (modified after Chappell et al, 1991). with 1-15 cm clasts of crystal-rich rhyolite within this facies, interpreted here as an autobreccia. The flow-foliated rhyolite facies comprises 13-20% subhedral to anhedral-shaped plagioclase, K-feldspar and quartz crystals, 0.25-2.75 mm in size. The groundmass is devitrified with micropoikilitic and spherulitic textures and quartz stringers. The foliation is seen as light and dark layers defined by variation in the size of the devitrification microstructures. Moulds (1999, unpubl.) observed peperite at the base of a unit of the flow-foliated rhyolite facies, in contact with the underlying mudstone facies. Outcrop of breccia containing angular clasts of flow-foliated rhyolite 3-30 cm in size, in a matrix of smaller clasts was found within this facies. Clasts fit together in a 'jigsaw' manner and are matrix supported suggesting this is an autobreccia. The plagioclase-crystal breccia facies is interbedded with the crystal-rich rhyolite facies. The facies occurs as north-south trending beds mostly in the western part of the study area. It comprises angular to subangular plagioclase crystals up to 5 cm and quartz crystals and lithic clasts to 3 cm in size in a devitrified matrix with micropoikilitic and spherulitic textures. Most

113 of the facies is poorly sorted and massive but grading, with younging direction to the west, is occasionally visible at the top of beds. The plagioclase-crystal breccia grades into a finegrained rock with light and dark grey-coloured laminae. Sedimentary facies include the mudstone and sandstone facies. Millimetre-thick layers of sand-sized grains of quartz and feldspar define occasional bedding in the mudstone facies. The sandstone facies comprises angular to subangular quartz, feldspar and opaque grains up to 2 mm in size and is interbedded with the mudstone facies. Graded bedding occurs as sandstone grading into mudstone. A polymictic conglomerate occurs toward the base of the Frampton Volcanics and consists of clasts several millimetres to 20 cm in size of granite, granodiorite, diorite, metasandstone, slate, siltstone and limestone (Skilbeck et al, 1992; Moulds 1999, unpubl.). Moulds (1999, unpubl.) interpreted this unit as a series of debris flows based on the massive, poorly sorted nature of the conglomerate. The mudstone and sandstone facies are interpreted as being deposited by turbidity currents based on the presence of normally graded beds. These facies were most likely deposited as gravity-driven flows between periods of volcanism. The presence of peperite at the base of beds of the crystal-rich rhyolite facies indicates this facies was also deposited in a subaqueous environment below storm wave base. Geochemistry Major and trace element geochemistry is used to classify and characterise the Frampton Volcanics. Data for the Frampton Volcanics fi-om this study, Moulds (1999, unpubl.) and Duggan (OZCHEM database) is compared to three S-type volcanic suites from the Lachlan Fold Belt and to data from silicic volcanic sequences in various tectonic settings worldwide, in order to facilitate placing the Frampton Volcanics in an appropriate tectonic setting. The Frampton Volcanics are rhyolitic to dacitic in composition with Si02 from 68-78%. Major element geochemistry is characterised by negative trends for Ti02, MgO, CaO, AI2O3, Fe203 and P2O5 on Harker variation diagrams indicative of fractional crystallisation of ferromagnesian minerals, ilmenite, plagioclase and apatite. Trends for trace elements are not as distinct as those from major elements with a wide scatter as Si02 abundances increase (e.g. Figure 2). Compatible trace elements show similar trends to the major elements. The major and trace element geochemistry of the Frampton Volcanics is similar to the S-type Goobarragandra Volcanics, Laidlaw Volcanics and Kellys Plain Volcanics (Chappell et al, 1991; Wybom et al,\9%\) (Figure 2) suggesting a similar source and petrogenesis. Silicic volcanism occurs in a variety of tectonic settings. The Frampton Volcanics were compared to silicic rocks from several of these settings (continental rifl, subduction zone, melting of arc crust and hot spot) and are most similar in major and trace element abundances from Yellowstone Plateau Volcanic Field (hot spot) and Taupo Volcanic Centre (backarc basin) (Figure 3). Considering there is no evidence for hot spot related mafic magmatism, it is more likely that a Taupo-style setting is appropriate for the Frampton Volcanics even though the Taupo Volcanic Zone magmas are not S-type. This match implies the presence of a related

114 subduction zone providing the mafic melts necessary for crustal melting and places the Frampton Volcanics in a backarc extensional basin (cf., Cole, 1990).

—

0.8

0.4 03 0.2

Z 07 O O.I 0

ioo

250

I: 200

% 1.20O 0-80

CO '50

OX)

100 50

0.40 • 0.00

120

lOOO

100

800

QU 600

400

N 65

Si02 wt% • Frampton Volcanics O-'Laidlaw Volcanics (Wybom) Goobarragandra Volcanics ( W y b o m ) ( C ? K e l l y s Plain Volcanics ( W y b o m ) F i g u r e 2. M a j o r a n d trace e l e m e n t variation diagrams for LFB volcanic rocks and Frampton Volcanics.

200 0 65

m 70

• ^• 75

Si02wt% • Frampton Volcanics A Y e l l o w s t o n e (Hildreth et aL, 1 9 8 4 ) • T a u p o (Sutton et al., 1 9 9 5 ) F i g u r e 3. M a j o r and trace e l e m e n t variation diagrams for Frampton Volcanics, Taupo Volcanic Centre and Y e l l o w s t o n e Plateau V o l c a n i c Field.

The tectonic discrimination diagrams of Pearce et al (1984) were also used in this study, with analyses plotting mostly within the Volcanic Arc Granite field, and some in the Within-Plate Granite field. Palaeogeography Characteristics of the Frampton Volcanics in the study area, such as sedimentary rocks with graded bedding, and the dominance of flow-foliated lava over pyroclastic flow products, indicate that this sequence was deposited in a subaqueous, below storm wave base setting (Cas and Wright, 1987). Subaqueous silicic volcanism differs from that of subaerial settings as it is typically characterised by lava flows and domes. The silicic magma in deep subaqueous settings may erupt more fi'equently as lava due to the high hydrostatic pressure suppressing vesiculation and preventing the escape of volatiles (Cas and Wright, 1987; Goto and Tsuchiya, 2004). Flows and domes have a distinct physical geometry and internal structure (e.g. Fink and Griffiths, 1998; Bonnichsen and Kauffman, 1987). In general, silicic lava flows are laterally extensive bodies, are evenly porphyritic or aphanitic and massive or flow-foliated, with flowfoliation usually subparallel to the base and top of flow contacts (McPhie et al, 1993). A brecciated carapace usually overlies coherent lava flows (McPhie et al, 1993) but is not always preserved (cf Eucarro Rhyolite, Allen and McPhie, 2002). Subaqueous silicic lavas

115 usually do not extend more than a few kilometers, and domes up to a few hundred of meters from their source (McPhie et al, 1993). Physical features associated with subaqueous domes include feeder dykes, rims of hyaloclastite that are meter to several tens of meters in thickness, radial columnar joints, peperitic textures, minor reworked pyroclastic material and autobreccia, with flow-foliation usually parallel to the dome margins (e.g. DeRita et al, 2001; Goto and McPhie, 1998; Goto and Tsuchiya, 2004). Most of the features of silicic domes and flows have been observed in the Frampton Volcanics, including flow foliation, autobreccia, possible columnar jointing, peperite at boundaries of mudstone and crystal-rich rhyolite facies, and reworked pyroclastics in the form of the plagioclase-crystal breccia facies. Feeder dykes and hyaloclastite carapaces were not observed. Due to the steep regional bedding orientation, a cross-section view of the Frampton Volcanics allows reconstruction of the palaeogeography based on stratigraphy and mode of deposition for each facies. Activity in the region began with the deposition of the polymictic conglomerate facies as a series of debris flows from a submarine canyon below a pebble beach (Moulds, 1999, unpubl.). Subsequent volcanism involved the extrusion of domes and flows (crystal-rich and flow-foliated rhyolite facies) onto the seafloor concomitant with the background sedimentation of fine-grained medial to distal turbidity currents. Coarse-grained debris-flows in the form of plagioclase-crystal breccia were deposited off the flanks of larger domes or a volcano in the vicinity during the later build up of the volcanic pile. Conclusions The Frampton Volcanics, in the area of study, comprise three major facies, including crystalrich rhyolite, flow-foliated rhyolite and plagioclase-crystal breccia facies, as well as minor sedimentary facies. The crystal-rich rhyolite facies and the flow-foliated rhyolite facies are interpreted as a series of lava flows and/or domes erupted and emplaced in a subaqueous environment. The plagioclase crystal breccia is interpreted as the product of debris-flow from a proximal volcanic source. The reinterpretation of the major volcanic facies as deep subaqueous lava domes and flows constrains the timing of volcanism in the Frampton Volcanics to after initial rifting of the Tumut basin. This contrasts with previous interpretations (eg, Stuart-Smith et al, 1992; Basden, 1990). Based on a comparison with the geochemistry of other Lachlan Fold Belt S-type volcanic sequences, the Frampton Volcanics may be magmatically related to the Goobarragandra Volcanics, Laidlaw Volcanics and Kellys Plain Volcanics. A possible modem analogue for the Tumut basin, and the site of emplacement of the Frampton Volcanics, is the Taupo Volcanic Zone, a major silicic volcanic centre in a backarc setting.

116 References

Allen S.R., and McPhie, J., 2002. The Eucarro Rhyolite, Gawler Range Volcanics, South Australia: A >675 km^ compositionally zoned lava of Mesoproterozoic age. Geological Society of America Bulletin, 114: 1592-1609. Allen S.R., and McPhie, J., 2003. Phenocryst fragments in rhyolitic lavas and lava domes. Journal of Volcanology and Geothermal Research, 126: 263-283. Basden, H., 1990. Geology of the Tumut 1:100,000 Sheet 8S27. Geological Survey of New South Wales, Sydney. Bonnichsen, B., and Kauffman, D.F., 1987. Physical features of rhyolite lava flows in the Snake River Plain volcanic province, southwestern Idaho. Geological Society ofAmerica Special Paper 212: 119-145. Cole, J.W., 1990. Structural control and origin of volcanism in the Taupo volcanic zone. New Zealand. Bulletin of Volcanology, 52: 445-459. Cas, R.A.F., and Wright, J.V., 1987. Volcanic Successions, Modern and Ancient. Allen & Unwin, London. Chappell, B.W, English, P.M., Kmg, P.L., White, A.J.R., and Wybom, D., 1991. Granites and related rocks of the Lachlan Fold Belt (1:1,250,000 scale map), Bureau of Mineral Resources, Geology and Geophysics, Canberra, Australia. De Rita, D., Giordano, G., and Cecili, A., 2001. A model for submarine rhyolite dome growth: Ponza Island (central Italy). Journal of Volcanology and Geothermal Research, 107: 221-239. Fink, J.H., and Griffiths, R.W., 1998. Morphology, eruption rates, and rheology of lava domes: Insights from laboratory models. Journal of Geophysical Research, 103: 527-545. Goto, Y., and McPhie, J., 1998. Endogenous growth of a Miocene submarine dacite cryptodome, Rebun Island, Hokkaido, Japan. Journal of Volcanology and Geothermal Research, 84: 273-286. Goto, Y., and Tsuchiya, N., 2004. Morphology and growth style of a Miocene submarine dacite lava dome at Atsumi, northeast Japan. Journal of Volcanology and Geothermal Research, 134: 255-275. Hildreth, W., Christiansen, R.L., and CNeil, J.R., 1984. Catastrophic isotopic modification of rhyolitic magma at times of caldera subsidence, Yellowstone Plateau Volcanic Field. Journal of Geophysical Research, 89: 83398369. McPhie, J., Doyle, M., and Allen, R., 1993. Volcanic Textures. Centre for Ore Deposit and Exploration Studies, University of Tasmania, Hobart. Moulds, T., 1999. The Geology, Palaeogeography and Tectonic Setting of the Frampton Volcanics, southeastern New South Wales, B.Sc. Hons. Thesis, Macquarie University, Sydney (unpubL). Pearce, J.A., Harris, N.B.W., and Tindle, A.G., 1984. Trace Element Discrimination Diagrams for the Tectonic Interpretation of Granitic Rocks. Journal of Petrology, 25: 956-983. Scheibner, E., and Basden,' H. (eds.), 1998. Geology of New South Wales - Synthesis. Volume 2 Geological Evolution. Geological Survey of New South Wales, Memoir, Geology 13, 666 pp. Skilbeck, C.G., Frankel, E., Dadd, K.A., and Leitch, E.C., 1992. Polymictic conglomerates of the Frampton Volcanics: pre-Benambran volcanic arc derivatives or the products of post-Benambran uplift? Geological Society of Australia Abstracts, 32: 36. Stuart-Smith, P.G., Hill, R.L, Rickard, M.J., and Etheridge, M.A., 1992. The stratigraphy and deformation history of the Tumut region: Implications for the development of the Lachlan Fold Belt. Tectonophysics, 214: 211-237.

117 Sutton, A.N., Blake, S., and Wilson, CJ.N., 1995. An outline geochemistry of rhyolite emptives from Taupo volcanic centre, New Zealand. Journal ofVolcanology and Geothermal Research, 68: 153-175. Wybom, D., Chappell, B.W, and Johnston, R.M., 1981. Three S-type volcanic suites from the Lachlan Fold Belt, southeast Australia. Journal of Geophysical Research, 86: 10335-10348.

118

Limestone petrography of the Riversleigh World Heritage Area, Northwestern Queensland. Elizabeth Price Biological, Earth and Environmental Sciences University of New South Wales, Sydney, Australia

Introduction The Riversleigh World Heritage area in northwestern Queensland contains outcrops of freshwater Oligo-Miocene limestone that have been the focus of intense palaeontological research since the 1980's (Archer, Hand and Godthelp, 1994, 1995, 2000). The deposits at Riversleigh contain abundant and often exceptionally preserved vertebrate fossils. These fossils have had a significant impact on our understanding of the evolution of Australia's fauna. Despite the fact that new taxa are still being described (and found) the focus of research is shifting to understanding the context in which these fossils are found. Riversleigh's fossil-rich sediments occupy a large area, with the fossils being found in over two hundred individual sites. Most Riversleigh limestones formed in freshwater pools and rivers, cave deposits or open fissures. Palaeo-ecosystems as currently interpreted (Archer, Hand and Godthelp, 1995; Price, 2001) ranged from complex closed forests (i.e. rainforest) in the Early and Middle Miocene, to more open forests by the Late Miocene and Pliocene. The major component of the Riversleigh limestone is calcite, with minor dolomite, chalcedony, manganese oxides or phosphates in isolated localities. These minerals have been determined from XRD as well as through diagnostic features seen in thin section. The thin sections show the complex range of features present within the Riversleigh limestones. Many invertebrates are also found in the limestone. Carbonates There are two main types of limestone at Riversleigh: the Cenozoic limestone that hosts the fossils (this is tentatively named the Carl Creek Limestone), and the Cambrian Thomtonia Limestone, which is the carbonate source for the Cenozoic deposits. The Cambrian limestone is mainly composed of dolomite, was formed in shallow seas, and in some places trilobite fossils can be found. The two limestones are relatively difficult to tell apart in the field. From a distance, some structural features may be seen in the Cambrian limestone, or there may be a layer of chert nodules, but chert nodules are also found within the Cenozoic deposits. Without the aid of a petrographic microscope in the field, the best way to tell them apart is firstly to see if there are fossils present (trilobites have not been found close to the Riversleigh area) and then to hit the rock - the Cambrian limestone has hydrocarbons present and emits a sulfurous smell when hit. The petrology of the Cenozoic limestone varies across the sites, as does the faunal assemblages. Some of the sites have a more terrestrial component, some more aquatic and

119 some more indicative of a cave enviromnent. The sites that indicate deposition within a cave can have speleothems present such as flowstone, stalactites / stalagmites, cave pearls and straws. These are usually macroscopic features.

Fig. 1. (Left): Microbial growth on raft; (Right): Flowstone.

Using petrography and SEM, many of the microscopic features of the Riversleigh limestone have been observed. One of the major discoveries of this study has been that the limestone is a mixture of inorganic and organic calcite (cf Pentecost, 1991). Bacteria and algae contributed significantly to the formation of the Riversleigh limestone. Many of the features visible in thin section could be indicative of multiple environments. Small rafts of calcite are formed from calcite precipitating at the water surface in stagnant water that is supersaturated with calcium carbonate. This could occur within a cave environment or within a stagnant body of water beside a river or creek. Gastropods, ostracods and charophytes have been identified and would normally be found in a fluvial / lacustrine environment, but they could also have been washed into a cave during a flood event.

Fig. 2. (Left): Charophyte with gastropods and ostracods in micrite. (Right): Banding within oncoids.

Large groups of oncoids were originally thought to be cave pearls and so could be used to indicate that the deposit had formed in a cave, but oncoids can also form at the edge of lakes by bacterial or algal interaction. Only recently has the SEM been used to look at the internal structure of these 'cave pearls' and they indicate that they were formed with the aid of bacteria. This does not rule them out as cave pearls. Many studies have been conducted on speleothems and it is becoming increasingly apparent that bacteria play a major role in secondary carbonate deposition within caves, (e.g. Northup and Lavoie, 2001).

120

Flowstone appears to be the one feature that can reHably be used as an indicator of cave deposition. These deposits are usually composed of clear sparite and are easily seen in the field. In some cases there has been recrystallisation of the original sparite. Another good indicator of a cave environment is based on the presence of two things - an abundance of bat bones and phosphate. The phosphate is derivedfi-ombat guano and can alter the sediments under and around it.

Fig. 3. (Left): Flowstone, showing recrystallisation and original growth zones (XP). (Right): Matrix altered by phosphate (XP).

Matrix Sparite and micrite make-up the majority of the matrix. Clean sparite is associated with cave deposits, while micrite could befi-ommany different depositional environments. Clays within the micrite have been analysed, and while further study into them is needed, illite does not seem to be a major component. Many sites have undergone neomorphism and recrystallisation and therefore it is the clasts and structures that can become the more useful palaeoenvironmental indicators. Clasts Quartz is the most abundant clast present within Riversleigh limestone. The grains are usually sub-rounded to sub-angular and are not well sorted. A few sites have such an abundance of quartz that they could be termed calcarenites. Most of the other clasts are of invertebrate and vertebrate fossils, but carbonate clasts, such as rafts, peloids and biological encrustations are also common. Most sites have an assortment of these clasts, but as yet only the vertebrate fossils have been catalogued. Zircons, chert and quartzite clasts have also been found. Some of the chert has come from the Cambrian limestone, but other clasts seem to have been formed by the remobilization of silica into Cenozoic carbonate structures. The quartzite and majority of the zircons have come from Precambrian deposits. Studies to determine if any of the zircons are Cenozoic in age is currently being undertaken.

121

Fig. 4. (Left): Planorbid gastropod in goethite rich, sandy limestone. (Right): Limestone showing three depositional events. One has abundant peloids, another mainly micrite & one has abundant quartz & chert.

Conclusion Petrographic thin sections are used in the initial stages of interpretation. They show some of the diversity of microfossils atid a large array of petrographic features. These can give a good indication of the depositional environment, but the fact that limestone readily undergoes recrystallisation and neomorphism has to be taken into account. A number of the sites appear to be able to group together based on their petrology, whilst others contain unique characteristics. These characteristics can range from strange algal growths, or large black sections within a site (analysed to be a barium manganese oxide), to phosphate nodules that contain plant and insect material. The current understanding of Riversleigh is highly dependent on fossil evidence, whether this is for the purpose of dating the fossils (e.g. stage of evolution / biocorrelation) or providing palaeoenvironmental interpretations. This study aims to investigate how the geology can be used to provide independent evidence to test these hypotheses. Aspects that will be investigated include petrography, geochemical and isotope analysis and the potential dating of zircons. Dating methods using ICPMS are also being explored. References Archer, M., Hand, SJ., and Godthelp, H., 1994. Riversleigh. Reed,

edition): 264pp.

Archer, M., Hand, S.J. & Godthelp, H., 1995, Tertiary environmental and biotic change in Australia. In: Vreba, E.S., Denton, G.H., Partridge, T.C., and Buckle, L.H. (eds). Palaeoclimate and Evolution, with emphasis on human origins. Yale University Press: 77-90. Archer, M., Hand, S.J., and Godthelp, H., 2000. Australia's lost world: prehistoric animals of Riversleigh. Indiana University Press. 264pp. Northup D. E., and Lavoie K. H., 2001. Geomicrobiology of Caves: A Review. Geomicrobiology Journal, 18 (3): 199-222 (24) Pentecost, A., 1991. Calcification processes in algae and cyanobacteria. In: Riding, R. (ed). Classification of Microbial Carbonates. Springer Verlag, Berlin: 3-20. Price, E., 2001. The Geology of Riversleigh. BSc (Honours) Thesis, UNSW {unpubl).

122

Systematics in two phase REE and Y partitioning coefficients in mafic granulites

Flavian C. Schroter^ Geoff L. Clarke\ R.W. White^ and N. J. Pearson^ ^School of Geosciences, F05, University of Sydney, Sydney, NSW 2006, Australia (g:eoffc@geosci.usvd.edii.aii) ^School of Earth Sciences, University of Melbourne, Parkville, Victoria 3052, Australia ^GEMOC ARC National Key Centre, Department of Earth and Planetary Sciences, Macquarie University, Sydney, NSW 2109, AustraHa

Trace elements are believed to behave and distribute 'passively' into minerals onto distinct crystal-lattice-sites. The incorporation of available trace elements in minerals is dependent on the crystal structure and in particular the crystal lattice site of the mineral hosting the trace element in question. It is well established that the partition coefficient (D) of isovalent trace elements (/) between mineral and melt in equilibrium describes a parabola (Fig. 1), when log D/ is plotted versus the ionic radius (r) of the trace element set (r/), provided the set of isovalent trace elements populates a single mineral lattice site (e.g. Onuma et aL, 1968; Jensen, 1973). The shape and position of the parabola can be determent by the optimum radius of the crystal lattice site (ro), the strain-free partition coefficient (Do - which describes the partition coefficient of a fictitious cation with the exact radius of ro, hence determines the peak of the parabola) and E, the Young's Modulus. The Young's Modulus describes the elasticity of the mineral site and is expressed in GPa. This factor determines the envelope of the parabola (Brice, 1975; Blundy and Wood, 1994). For a single crystal lattice site, each set of isovalent cations (monovalent, divalent, trivalent etc.) describes their own parabola, with only ro being identical. Experimental studies in the last decade have shown that the REE patterns of minerals in equilibrium with melt or fluid is dependent on pressure, temperature, mineral chemistry (in particular that of the crystal-latticesite in question) and water content of the melt (e.g. Blundy and Wood, 1994; Klein et al, 1991 \ cation radius

\ o

Fig.L A schematic display of the partitioning model (Brice, 1975; Blundy & Wood, 1994) of trace elements for one crystal-lattice-site, with variations (/S) in the parameters (ro, Do and E). The size of the crystal-lattice-site and therefore the size of the strain-free (fictive) cation are given with ro. DQ determines the distribution coefficient (min/melt or min/fluid) of ro (peak), whereas the Young's Modulus (E) controls the envelop of the parabola. All parameters are dependent on P, T, and X (mineral composition), ro primarily on Xand Do dominantly on P and T (Blundy & Wood, 1994; Wood & Blundy, 2002).

123 van Westrenen et al, 1999; Green et al, 2000; Wood and Blundy, 2002). It has also been shown that these factors are valid for partition coefficients between minerals and aqueous fluid (e.g. Stalder et al, 1998). In natural metamorphic samples the equilibrium phase to the mineral assemblage can't be measured. Nevertheless, in-situ analyses of major and accessory minerals in subsolidus assemblages reveals systematic incorporation of REE and Y. These elements show different concentrations and patterns of distribution for each mineral within a given assemblage under equilibrium conditions. REE pattem in minerals has been found to be dependent on mode of coexisting minerals. Hence, the REE pattems of minerals in isolation are highly variable and useless by themselves. Despite the differences between two minerals in the REE pattems, it is found that mineral partition coefficient (D"^^"^"^") is similar under similar pressure and temperature conditions. Variations of D-values for a given mineral pair under varying conditions is then again likely to depend on the physio-chemical parameters (ro, Do and E) shown to control the mineral/melt distribution. The first example is fi'om a metamorphosed layered mafic sill, in which each layer displays the common assemblage of hornblende (hbl), clinopyroxene (cpx), orthopyroxene (opx) and plagioclase (plag). Four layers (A to D) have been distinguished and display variation in their mineral modes and grain size. Layer A contains additional gamet, whereas layer D contains additional biotite and apatite. Hornblende and clinopyroxene are the dominant REE and Y carriers within the common assemblage. The changes of REE and Y contents in homblende and clinopyroxene between the different layers are larger by magnitudes compared to the changes in bulk rock compositions, which are therefore negligible. Rather the modes of coexisting minerals competing for the same element (group) in question drive the changes in concentration of REE and Y. Furthermore, additional minerals in the assemblage might influence the shape of the REE and Y pattem of coexisting minerals. For example, additional gamet in layer A and its strong affinity for heavy REE (HREE) enforces a selective decline in this element group in homblende and clinopyroxene. Despite varying REE pattems in the homblende and clinopyroxene between the different layers, the pattem (REE and Y distribution coefficient of homblende/clinopyroxene) remains similar. Apatite in layer D prefers light REE (LREE) over HREE but has a lesser effect on the REE pattem, because of its low abundance and its less selective nature within the REE compared to gamet. In the second example, preliminary data of variable values under changing physiochemical conditions expressed in corona formation around homblende and gamet in a single mafic gneiss sample are presented. Corona formation took place under conditions of near isothermal decompression at -TSO^'C and a decrease in pressure of around 1 kbar within 8 to 5 kbar.

124 References

Blundy, J., and Wood, B., 1994. Prediction of crystal-melt partition coefficients from elastic moduli. Nature, 372: 452-454. Brice, J. C., 1975. Some thermodynamic aspects of the growth of strained crystals. Journal of Crystal Growth, 28 (2): 249-253. Green, T. H., Blundy, J. D., Adam, J., and Yaxley, G. M., 2000. SIMS determination of trace element partition coefficients between garnet, clinopyroxene and hydrous basaltic liquids at 2-7.5 GPa and 1080-1200°C. Lithos, 53 (3-4): 165-187. Jensen, B. B., 1973. Patterns of trace element partitioning. Geochimica et Cosmochimica Acta, 37 (10): 22272242. Klein, M., Stosch, H.-G., and Seek, H. A., 1997. Partitioning of high field-strength and rare-earth elements between amphibole and quartz-dioritic to tonalitic melts: an experimental study. Chemical Geology, 138 (3-4): 257-271. Onuma, N., Higuchi, H., Wakita, H., and Nagasawa, H., 1968. Trace element partion between two pyroxenes and the host lava. Earth and Planetary Science Letters, 5: 47-51. Stalder, R., Foley, S. F., Brey, G. P., and Horn, I., 1998. Mineral aqueous fluid partitioning of trace elements at 900-1200 ®C and 3.0-5.7 GPa: New experimental data for garnet, clinopyroxene, and rutile, and implications for mantle metasomatism. Geochimica et Cosmochimica Acta, 62 (10): 1781-1801. van Westrenen, W., Blundy, J., and Wood, B., 1999. Crystal-chemical controls on trace element partitioning between garnet and anhydrous silicate melt. American Mineralogist, 84 (5-6): 838-847. Wood, B. J., and Blundy, J. D., 2002. The effect of H2O on crystal-melt partitioning of trace elements. Geochimica et Cosmochimica Acta, 66 (20): 3647-3656.

125

Cumulate recycling in arc magmas Alanna L. Simpson^*, Kurt M. Knesel* and Michael A. Dungan^ ^University of Queensland, St Lucia, Brisbane, 4072, QLD, Australia ^Section des Sciences de la Terre, Universite de Geneve, 13, Rue des Maraichers, 1211 Geneve 3, Switzerland *email: a,simpson@earth.uq.edu.au Tatara San Pedro Volcano, Chile The Tatara San Pedro Centre (TSPC), a large Quaternary volcanic centre located in the Southern Volcanic Zone (SVZ) of Chile, has erupted '-^55 km^ of chemically diverse lUvas over the past million years. These lavas span much of the compositional range of the SVZ as a whole, yet long-term differentiation trends are rare at the TSPC. Instead petrographic, mineralogical and geochemical data are consistent with short-term modulation of magma compositions through diverse interactions between new arrivals of mafic to intermediate magmas from depth and variably evolved and contaminated magma and/or crystal mush resident in the mid to upper crust (Dungan, et al, 2001). Clues to unravelling the timing and dynamics of these processes are held in xenoliths and xenocrystic fragments contained in both mafic and silicic magmas erupted throughout the history of this arc volcano (Costa et al, 2002; Dungan and Davidson, 2004). The best-preserved suite of these cumulate materials consists of gabbroic xenoliths in a late Holocene eruption of dacite from Volcan San Pedro. The majority of these samples are clinopyroxene norites and leuconorites, along with minor homblende and olivine norites. Many of the xenoliths display subsolidus and deformation features and some contain amphibole and phlogopite oikocrysts. These hydrous phases also occur in microfractures cutting plagioclase crystals and are interpreted to have formed during reaction of late-stage, evolved fluids and melt that migrated through the refractory cumulate pile (Costa and Singer, 2002). Micro-xenolithic fragments of clinopyroxenitic and troctolitic cumulates, which conspicuously lack abundant homblende and phlogopite, are also present in flows of the 235 ka, crystal-rich Upper Placeta San Pedro (UPSP) basalts (Dungan et al, 2001). Here we describe similar xenolithic fragments brought to the surface in one of the eariiest basalts erupted at the TSPC (925 -826 ka; Singer et al, 1997), and we examine the role cumulate recycling may play in the evolution of both mafic and silicic magmas at arc volcanoes. Evidence for cumulate recycling in the earliest TSPC basalts The oldest preserved record of volcanism at the TSPC comprises a volumetrically minor series of mafic to intermediate lavas and laharic breccias, the Sin Nombre lavas, intercalated between two major silicic units, a basal dacite and a capping zoned rhyodacite-rhyolite unit. A group of these mafic lavas show low but variable abundances of compatible elements, characteristic of relatively evolved magma, coupled with low-incompatible-element abundances more indicative of a primitive signature. These basalts form linear arrays on bivariant element plots, with a high-A^Os, high-CaO, low-Zr-Nb-Ba-Rb endmember and a high-Fe203, low-MgO, low-Sr endmember (Fig. 1). They are also quite crystal-rich (45% modal crystals) and contain anorthite, clinopyroxene and olivine phenocrysts, along with xenocrystic crystal clots (<10 mm) set within an andesine, clinopyroxene, iron-oxide

126 groundmass. The clots display variable modal mineralogies and textures, including (1) abundant poikilitic plagioclase and minor olivine enclosed in clinopyroxene oikocrysts, (2) dihedral-linked olivine enclosed within plagioclase oikocrysts and (3) clinopyroxene megacrysts surrounding plagioclase microphenocrysts. Electron microprobe (EMP) analyses of phenocryst and clot phases reveal homogenous cores out of equilibrium with the groundmass composition. In contrast, crystal rims have similar chemistry to phases within the groundmass (Fig. 2). Olivine crystals have unzoned cores, averaging Fogi, with low MnO and CaO abundances surrounded by a transition zone in which forsterite decreases and MnO increases. The outer ~20 |im rims are on average F070, and display an increase in CaO and a drop in NiO. Clinopyroxene crystals show similar profiles to olivine, with a decrease in Mg#, wollastonite and alumina and increase in MnO between the core and rim. Unzoned plagioclase cores are highly calcic, average Ang?, with low FeO. Rims are distinguished by a sharp drop in anorthite to ~An4o and an increase in FeO. The outer S-lO^im of the rim is slightly reversed with an increase in MgO. Application of the single-pyroxene geothermometer of Lindsley (1983) and the plagioclase-magmatic water saturation calculation of Sisson and Grove (1993) indicates the phenocrysts and clots crystallised under watersaturated (~6 wt%) conditions at temperatures ranging from about 1100 to 1150°C. In contrast, groundmass crystals appear to have grown under relatively hot, dry F\g l.Biwiantdernatplots.S>ri±)oIs: Sq - Sin Nombre Basalt (open-gm); magmatic conditions, ~1240°C and ~2wt%, respectively. dianTonds-MEMbosattrian^&d^ grey area - Holocene gabbro xenoliths; It grey area are the TSPC volcanics.

"typical

TO.85

Mg#/100

Wo/100

0.75

0.65

10.55

Fig2a)EMPtm\€r3e of olivine phenocryst. b) EMP traverse of clinopyroxene. c)NDIC image of plagiodase with EMP traverse.

niQtJ^nr^i fnm\ 100

200

127 The compositional and textural similarities of phenocrysts and clot phases, compared to the gm contrasting composition of groundmass phases, indicate the phenocrysts are disaggregated 0.70414 from crystal clots. The distinction between the phenocrysts/crystal clots and the host magma is 0.70410 furthered by ^^Sr/^^Sr ratios of mineral and groundmass separates. Clinopyroxene and plagioclase have indistinguishable ratios of 0.70406 calculated bulk cumulate 0.70407 and 0.70408, close to average TSPC composition values, whereas the groundmass is relatively 0.70402 radiogenic at 0.70415 (Fig. 3). These resuhs 20 10 Rb 15 are consistent with mixing of low-Rb, lowFig. 3. Rb versus ^^Sr/^^Sr, showing a mixing curve to ^^Sr/^^Sr and high-Rb, high-^^Sr/^^Sr endproduce whole-rock ratios. Bulk cumulate composition (mcfepw^mpl^^ 0.94 ppm, 0.70408; Groundmass [Sr] 557 ppm, members (see Fig. 3 caption). 0.70418

87sr/86sr

[Rb] 21.0 ppm, ^^Sr/S^sr 0.70415. The Sr isotope data >iddsa51+177]VIaisochrai(MSWD5,pnDbe^ 0.002), an age inconsistent with a stratigraphic position between 925 and 826 ka silicic units.

A tale of two magmas: recycling of H A B cumulates by tholeiitic basaltic andesite

Cumulates, as represented by the phenocrysts and crystal clots, are interpreted to be mobilised from a relatively cool, water-rich, subsolidus/ solidus mafic intrusion that was compositionally heterogeneous and possibly layered. The largely unzoned nature of the crystals implies that open-system percolation of melts/fluids occurred during crystallisation. The high AI2O3 and CaO and low MgO and Ni contents indicate the magma had undergone significant olivine ± clinopyroxene fractionation, but limited or no plagioclase fractionation, to form a low-MgO, high alumina basaltic (HAB) magma.

The host magma, roughly represented by the groundmass, plots within the high-Fe (Arculus, 2003) or tholeiitic (Miyashiro, 1974) field on a Fe/Mg versus silica plot. Low Si02, MgO, AI2O3 and CaO contents are consistent with considerable olivine + plagioclase ± clinopyroxene fractional crystallisation, but high Fe203 and low Mg# (-45) indicate magnetite was not an early crystallising phase. The delayed onset of magnetite crystallisation may be symptomatic of low oxygen frigacity and/or insufficient available water to stabilise magnetite (Sisson and Grove, 1993; Bemdt et al, 2005). Our best estimate for the bulk composition of cumulate end-member lies within the field of gabbroic xenoliths in Holocene San Pedro eruptives (Costa et al, 2002). However, as note above, neither the Sin Nombre nor the UPSP cumulates contain amphibole or phlogopite. This conspicuous absence is consistent with a model in which partial melting along grain boundaries, mainly hornblende, phlogopite and plagioclase, facilitated xenolith disaggregation. Assimilation of such grain-boundary melts enriched in incompatible trace elements along with variable retention of refractory, xenocrystic olivine + clinopyroxene ± plagioclase provides an attractive mechanism for the decoupling among incompatible and compatible elements in UPSP lavas (Dungan and Davidson, 2004). We argue that such a process also explains the juxtaposition of a primitive incompatible element signature with an evolved compatible element trait in the Sin Nombre basalt.

128 a) 90-

\mi\) SO^ffj

Timing of recycling event

Olivine

TO-

^overgrowth Fo

tfr

8 7060.

W/ o// 0

100

200

Clinopyroxene

The timescale for recycling can be constrained using Fe-Mg diffusion rates in olivine (Chakraborty, 1997) and clinopyroxene (Dimanov and Sautter, 2000; Klugel, 2001), and the 1-D diffusion model of Klugel (2001) and microprobe profiles of crystal edges, which record interaction of Mg-rich xenocrysts and Mg-poor melt. Assuming an initial, step-function compositional profile between the core and rim of the crystals (which appears valid because some crystals display relatively constant rim compositions, representing about 5-20]Lim of crystal growth in equilibrium with the host magma), we estimate timescales of 1 month to 1 year between entrainment of cumulates and eruption (Fig. 4).

A cumulate source for silicic magmatism?

Silicic (>63wt% Si02) magmatism has occurred sporadically throughout the 1 Ma history of the TSPC and can be divided into groups with low and high HREE-Y abundances (Fig. 5). The high-HREE-Y group can be explained by relatively simple fractionation of mantle-derived magmas (Dungan et 65al, 2001), similar to silicic magmas in the southern segment of the SVZ (e.g., Gerlach et al, 1988). In contrast, the lowHREE-Y silicic magmas appear to be the products of fluid fluxing and melting of gabbroic crust, which crystallised from primitive mantle-derived magma that interacted with garnet-bearing lower crust (Feeley et al., 1998; Costa and 100 200 Distance (microns) Singer, 2002). The oldest silicic magmas at the TSPC (925Fig. 4. Results of Fe-Mg diffusive 826 ka) fall within this crustally derived group and contain modelling in (a) olivine and (b) xenocrysts of orthopyroxene, amphibole, plagioclase and dinopyroxene Timecuvesaemoddled using IDdif fosionthroi^asemMnite biotite, similar in composition to late-crystallising phases in crystal with a planar interface. Values of the San Pedro gabbroic xenoliths. We propose these phases Datl200X:are5xl0 ^^m^solivine aid 1 X lO'^m^/s clinopyroxene represent a restite assemblage, where amphibole and orthopyroxene are stable during melting. Plagioclase is embayed with considerable regions of melt and relict biotite is mantled by amphibole, signifying the bulk of the silicic magma was formed during incongruent melting of biotite and plagioclase in the presence of refractory orthopyroxene and amphibole. Moreover, REE patterns in these silicic magmas display a concave pattern consistent with retention of amphibole in the source region. 75-

129

Fig.5. Rb versus Y plot displaying the Sin >fciTtieB^(9quares)2ndM^ dacite-rhyolite (circles). Other data displayed atefelPSPbasalls(diarrmds>TalaraDacite (slars)areifiddsofd«H(A»enearrulafies,aIl TSPC volcanic analyses and the sub-volcanic plutonics.

References Arculus, RJ., 2003. Journal of Petrology, 44: 929-935. Bemdt, J., Koepke, J., and Holtz, F., 2005. Journal of Petrology, 45: 135-167. Chakraborty, S., \991, Journal of Geophysical Research, 102: 12317-12331. Costa, P., Dungan, M.A., and Singer, B.S., 2002. Journal of Petrology, 43: 219-241. Costa, F., and Singer, B., 2002. Journal of Petrology, 43: 1571-1593. Dimanov, A., and Sautter, V., 2000. European Journal of Mineralogy, 12: 749-760. Dungan, M.A., Wulff, A., and Thompson, R., 2001. Journal of Petrology, 42: 555-626. Dungan, M.A., and Davidson, J., 2004. Geology, 32: 773-776. Feeley, T.C., Dungan, M.A., and Frey, F.A., 1998. Contributions to Mineralogy and Petrology, 131: 393-411. Gerlach, D.C., Frey, F.A., and Moreno-Roa, H., 1988. Journal of Petrology, 29: 333-382. Klugel, A., 2001. Contributions to Mineralogy and Petrology, 141: 1-14. Lindsley, D.H., 1983. American Mineralogist, 68: 477-493. Miyashiro, A., 1974. American Journal of Science, 274: 321-355. Singer, B.S., Thompson, R.A., Dungan, M.A., Freeley, T.C., Nelson, S.T., Pickens, J.C.m, Brown, L.L., Wulff, A.W., Davidson, J.P., and Metzger, J., 1997. Geological Society of America Bulletin, 109: 127-142. Sisson, T.W., and Grove, T.L., 1993. Contributions to Mineralogy and Petrology, 113: 143-160.

130

The significance of silicic melts in subduction related volcanic suites Ian E M Smith^ and Richard C Price^ ^Department of Geology, University of Auckland, PB92019, Auckland, New Zealand ^School of Science and Technology, University of Waikato, PBS 105, Hamilton, New Zealand

The petrological characteristics of subduction-related magmatic systems are determined by the thermal flux produced by the movement of magmas through the crust. The effect of this process through time is an upward movement of isotherms which results in development of the system toward more geochemically evolved magma compositions. At one extreme this can result in a system dominated by rhyolite. Recent work in andesitic volcanic systems of the North Island of New Zealand shows that the composition of erupted magmas varies systematically with time. We interpret this change as indicating the increasing involvement of a crustal component as the isotherm rises in response to magmatic flux. Andesites are generated through the interaction of mantle-derived magmas with lower crustal melts and restites and are complex mixtures of crustal melt, strongly fractionated mantle melt, restitic crystals, lithic lower crustal fragments and phenocrysts derived from mantle melts. Rhyolitic magmatic systems in continental, subduction linked settings represent the end member state of this evolutionary spectrum of petrological processes where the crustal component has come to dominate (Price et al, 2005; Fig.l). Early stage- preconditioning of cnist by andesitic niagrnatism

Fig.l Stages in the evolution of an andesitic magmatic system to a rhyolite dominated system. (From Price et al, 2005)

Crustal underplating

M a n t l e derived basaltic m e l t s

131 Later stage- Extensive cnistal recycling Formation of extensive rhyolitic magma acciimidations develop over ~100-40 ka

High heat flow extensive recycling of andesitic underplate

A widely accepted paradigm is that continental crast is generated in continental subduction systems where an important process is the generation of large volume silicic magmas by crastal anatexis. By contrast, oceanic subduction systems are considered to produce mainly mafic magma compositions in the compositional range basalt to basaltic andesite. Detailed examination of oceanic arcs suggests that this general conception is not valid and one of the clearest examples of this is the Tonga-Kermadec Arc which is the northward continuation of the subduction system that generated the andesites and rhyolite volcanoes of North Island New Zealand. The arc is a chain of volcanoes nearly 3000 km long on a convergent boundary between the Australian and Pacific plates that is almost entirely oceanic. The arc is represented by the subaerial volcanoes of Tonga and the Kermadec Islands and by the abundance of very large submarine volcanoes, many of them only known from recent submarine mapping. A surprising feature of the subaerial volcanoes and also of dredge sampling of submarine portions of the arc is the abundance of felsic compositions most of which are the products of eruptions in recent times (Fig.2). These are commonly associated with caldera structures formed during eruptions that are comparable in size to many continental silicic eruptions

Hg.2. Histogram of the distribution of Si02 (wt.%) in samples from the Kermadec Arc. The horizontal scale is from 45 to 75 wt.% Si02^

Oceanic arcs can be seen as evolving through four stages as follows. Generation of subduction-related basaltic magmas initiates a volcanic system on oceanic crust. Underplating magmas cool and crystallize and hydrothermal convection develops through the lower crust. At the top of the system relatively fractionated magmas begin to build a volcanic edifice. During arc infancy (0.5-1.0 my) heat is transferred by convection to the lower arc crust (Fig.3). Hydration of crust through reaction of pyroxene + olivine -bearing lithologies with hydrous fluids produces amphibole. Surficial eruption continues with eruption of basaltandesite magmas.

132 A stage of arc adolescence (1-2 Ma; Fig.3) commences as the temperature of a significant volume of lower crust approaches the amphibole-saturated solidus at 850-950®C. Initiation of dehydration melting fluxes the crust and melt volumes of a few km^ to tens of km^ are rapidly generated. A 20-30% melt fraction segregates from a granulitic residue and ascends to upper levels in the system. Triggering mechanisms include episodic transfer of extensional strain into the crust or a pulse of magma associated with a major recharge event. Felsic magmatic eruptive activity may be interspersed with continuing basaltic-andesitic activity. Arc maturity (>3 Ma) sees a continuation of basaltic to andesitic activity (Fig.3). The lower crust having undergone dehydration melting is now anhydrous granulite significantly below its solidus temperature and it acts as a thermal insulator preventing convection of hydrothermal fluids. Further anatexis can only occur if appropriate source materials remain to participate in the hydration-dehydration cycle.

Adolescent { arc

Infant arc

Mantle

,im ««frc MrcO lWOO' C"

Mature arc

Mantle

Fig.3 Stages in the evolution of an oceanic arc (from Smith et al 2003) In the thermal evolution of an oceanic arc system the processes of underplating, together with the continuous magmatic (and thermal) flux, can generate a cmstal thickness in which dehydration melting of under-plated arc material generates felsic magmas. Further, this condition can represent a unique 'adolescent' stage in a developing oceanic arc, because once the felsic melts are extracted the lower crust becomes an infertile residue (Fig.3). A consistent conclusion is that the scale of silicic magmatism in the oceanic portion of the TVZ-Kermadec-Tonga Arc is comparable to that of the continental portion and we suggest that the processes driving these systems are also comparable and that the processes operating in oceanic arcs are not fundamentally differentfromthose in continental systems.

References

Price, R. C., Gamble, J. A., Smith, I.E.M., Eggins, S.,. and Wright, I.C., 2005. An integrated model for the temporal evolution of andesites and rhyolites and crustal development in New Zealand's North Island. Journal of Volcanology and Geothermal Research, 140: 1-24. Smith, I.E.M., Worthington, T.J., Stewart, R.B., Price, R.C., and Gamble, J.A., 2003. Felsic volcanism in the Kermadec Arc, southwest Pacific: crustal recycling in an oceanic setting. In: Larter, R.D., and Leat, P.T. (eds). Intra-Oceanic Subduction Systems: Tectonic and Magmatic Processes. Geological Society of London Special Publication 219: 99-118

133

New ruby-sapphire sources, Yarrowitch basaltic field, eastern NSW. Lin Sutherland^'^, Ian Graham^ Gayle Webb\ Ross Pogson\ Gaston Giuliani^ and Anthony Fallick"* ^ Geoscience, Australian Museum, 6 College St Sydney, NSW 2010 Email: 1 ins@austmus. gov. au ^ School of Science, University of Western Sydney, North Parramatta, NSW 2197 ^ IRD and C.R.P.G / C.N.R.S., BP20, 54501 Vandouevre - Les Nancy, France ^ Isotope Geoscience Unit, SUERC, East Kilbride, Rankine avenue, Glasgow G75 OQF, Scotland Introduction Gem corundum prospects abound in eastern Australian Cenozoic basaltic terrains, and some deposits support major mining and treatment ventures (Coldham, 2003). Sapphire deposits dominate (e.g. New England, NSW), but ruby-bearing deposits are receiving increasing investigation. The main ruby deposit on the Barrington Plateau (Sutherland and Graham, 2003; Roberts et al, 2004) is now in production. The research reported here is focused on the Yarrowitch basalt field some 75 km north of Barrington, on the highland divide, some 85-100 km WNW of Port Macquarie. The gem-bearing basalt field here (Figure 1) forms a southeastern extension of the Walcha Volcanic Province, which has had only limited study (Duggan, 1989). The Yarrowitch gemfleld Limited exploration to date has revealed that the gemfield covers an area of at least 100 km^ and is centred around the headwaters of Fenwicks Creek, near the hamlet of Yarrowitch. The gem corundums (ruby, sapphire) have been found in present-day and palaeo alluvial deposits, and in rare cases, embedded within pyroclastics. Each specific site has different concentrations of corundum, zircon and spinel. Some sites are dominated by zircon, whilst at others zircon is uncommon. Among the corundums, sapphire is dominant at most sites, but ruby is moderately common and has been found at 8 discrete sites. Basalt characteristics Several pyroclastic vents and dykes were identified along a north-south trending zone in the basalt field (Figure 1). The basalts are largely little-evolved basalts and basanites, with many containing mantle peridotite xenoliths. Limited K-Ar dating on the basalts gave 58 ± 2 Ma at Yarrowitch; 48 =t 1 Ma from 11 km east of Yarrowitch; and 45.9 ± 1 Ma from 5 km SE of Yarrowitch at the top of the basalts overlying the ruby source in Fenwicks Creek (Jones, 1987; author's unpublished data).

134

(O km

Fig. 1. Geological sketch map, Yarrowitch basaltic gemfield. Basalt exposures and soils are enclosed by thick lines, with Yarrowitch field (Y) in central area and part of the remaining Walcha basalt province on left. The outer blank areas represent Palaeozoic folded basement. The Oxley Highway (double line) traverses the Yarrowitch field along with connecting secondary roads (dashed lines); drainage (thin lines); pyroclastic pipes, deposits and dykes related to volcanic centres (black areas); basalt age-dating sites (encircled cross); and gem corundum - zircon sites (asterisks). Based on the Hastings 1: 250 000 Geological Map Sheet, with updated road systems. Subordinate New South Wales map (NSW) indicates relative positions of the Yarrowitch (Y) and Barrington (B) basaltic gemfields to Sydney (S) and Port Macquarie (PM). Most of the gem corundum sites lie at approximately 1100m asl. For the Yarrowitch field, the highest basalt elevation is at 1292m, whilst the lowest basalt is at ~ 950m asl.

135 Ruby-sapphire characteristics The ruby forms part of a zircon, corundum, spinel, ilmenite (zircospilic) association, in which abundant zircon provides some age control on both the eruptive surface cooling (reset fission track ages) and the pre-transport source formation (U-Pb SHRIMP) ages of the gem minerals. These xenocrysts derived from the basalts typically show magmatic corrosion, but need not themselves have a magmatic origin. The rubies reach up to 8mm in size and grade from red and purple-red into pink, mauve and other colours, as the trace Cr chromophore contents decrease relative to trace Fe and Ti. Sapphirine and Cr-bearing spinel inclusions identified in the ruby here suggest a metamorphic origin, similar to that of the Harrington ruby suite. Some sapphires in these deposits show similar characteristics to those of the blue-green-yellow suite of magmatic origin. Small (< 2mm) pale blue sapphires derived from feldspathic tuff-like deposits form a distinctive minor 'magmatic' component here, so that Yarrowitch represents a polygenetic corundum association (Graham et al, 2004). Oxygen isotope data on the corundums give distinctly different values for the metamorphic ruby suite (3.6 - 3.9 7oo) compared to the magmatic blue-green sapphire suite (5.3 - 5.4 7oo). Zircon characteristics The zircons occur in pyroclastic breccias near the base of the basalt sequence, in deep lead deposits below later basahs, and in present-day alluvial sediments. They exhibit a range of crystal forms, mostly with magmatic growth features, and colours include white, yellow, orange, brown, pink, red, and some unusual bronze and blue-green types. Analyses show that Hf is the main substitution in the zircon (Hf02 0.55-1.15 wt%). The unusual colours show no obvious trace element links and may result from later heating events. Fission track dating of the zircons (U 49-444 ppm) records several eruptive events at 62, 55 and 40 ± 3 Ma, with one young event at 3 ± 2 Ma on the outskirts of the field at Tobins Camp (Sutherland, 1993). The U-Pb SHRIMP dating of the zircons (U 53-535ppm, Th 25-761ppm, U/Th 0.7-2), suggests early formation close to inception of basalt activity (60-66 ± 2 Ma). This includes the lateerupted Tobins zircons. Thus, magmatic gem formation appears to be early Palaeogene, with repeated later eruptions bringing up both magmatic (blue-green-yellow sapphire suite) and metamorphic (ruby suite) material. Yarrowitch gemfield comparisons In the presence of both metamorphic and magmatic ruby-sapphire suites, abundant zircons of several eruptive ages and mostly alkaline basalt hosts, the Yarrowitch gemfield closely resembles the Barrington gemfield to the south (Sutherland and Fanning, 2001). Both have extended activity (over 55-60 million years) of zircon formation and basaltic eruptions, but starting slightly earlier at Yarrowitch. The Yarrowitch Field however, largely represents a single early zircon-forming event compared to two zircon-forming events at Barrington. Yarrowitch also yields additional clues to ruby formation. Some of the ruby-bearing volcanic and volcaniclastic deposits include fragments of altered ultramafic and mafic lithologies related to ultramafic complexes in nearby Palaeozoic basement rocks. Interactions of thermal alkaline fluids and melts at contacts with Cr-rich spinels in the underlying ultramafic bodies provides a plausible mechanism for introducing Cr into a corundum-crystallising metamorphic process. Yarrowitch rubies, however, are distinct in O isotope values (3.6 - 3.9 7oo) compared to Barrington rubies (5.1 - 6.2 7oo; Giuliani et al, 2005), so may differ in some underlying genesis. The full potential of the Yarrowitch ruby field remains for future assessment.

136 Acknowledgments Ray Andrews of Gingers Creek and Joe Terp of Port Macquarie introduced and guided the authors in their studies of the Yarrowitch field. The Australian Museum provided funding for the research. Peter English, Armidale, provided some samples and petrologic reports. Zircon fission track dating was facilitated by Dr Paul Green of Geotrack International, Melbourne and the U-Pb dating represents a collaborative study with Dr Richard Armstrong from The Research School of Earth Sciences, ANU. K-Ar dating was done by Dr Horst Zwingmann of CSIRO Petroleum, Perth. Electron microprobe analyses were aided by Curt Stocksiek of the School of Science, University of Western Sydney. Oxygen isotope analyses of the corundums were made through IRD and C.R.P.G./C.N.R.S.and SUERC institutions, Vandoeuvre-lesNancy, France and Glasgow, Scotland. References

Coldham, T., 2003. The history and importance of heat treatment of Australian sapphires. The Australian Gemmologist, 21: 50-62. Duggan, M.B., 1989. Walcha. IN: Johnson, R.W. (compiler), Intraplate Volcanism in Eastern Australia and New Zealand. Cambridge University Press, Cambridge, p. 123. Giuliani, G., Fallick, A.E., Gamier, V., France-Lanord, Ch., Ohenstetter, D., and Schwarz, D., 2005. Oxygen isotope composition as a tracer for the origin of rubies and sapphires. Geology, 33: 249-252. Graham, I.T., Sutherland, F.L., Webb, G.B., and Fanning, C.M., 2004. Polygenetic corundums from New South Wales gemfields. IN: Khanchuk, A.I., Gonevchuk, G.A., Mitrokhin, A.N., Simanenko, I.F., Cook, N.J., and Seltmann, R. (editors), Metallogeny of the Pacific Northwest: Tectonics, magmatism and metallogeny of active continental margins, pp. 336-339. Jones, D.G., 1987. K-Ar isotopic dates. New South Wales. Geological Survey of New South Wales, Department of Mineral Resources, Report No. GS1986/237. Roberts, D.L., Sutherland, F.L., Mollis, J.D., Kennewell, P., and Graham, I.T., 2004. Gemstone characteristics, North-East Barrington Plateau, NSW. Journal and Proceedings of the Royal Society of New South Wales, 137: 99-122. Sutherland, F.L., 1993. Late thermal events based on zircon fission track ages in northeastern New South Wales and southeastern Queensland: Links to Sydney Basin seismicity?. Australian Journal of Earth Sciences, 40: 461470. Sutherland, F.L., and Fanning, C.M., 2001. Gem-bearing basaltic volcanism, Barrington, New South Wales: Cenozoic evolution based on basalt K-Ar ages and zircon fission track and U-Pb isotope dating. Australian Journal of Earth Sciences, 48: 221-237. Sutherland, F.L., and Graham, LT., 2003. Geology of the Barrington Tops Plateau. Its Rocks, Minerals and Gemstones, New South Wales, Australia. The Australian Museum Society, Sydney, 56pp.

137

Tectonic significance of low-grade mineralisation of seafloor spreadingrelated faults, Macquarie Island Jean-Yves Talbot and Nathan R. Daczko GEMOC Key Centre, Department of Earth and Planetary Sciences, Macquarie University, Sydney, NSW 2109, Australia The Macquarie Island "ophiolite" preserves a range of seafloor spreading-related faults that are heavily mineralised. Mineral assemblages include epidote, prehnite, calcite, chlorite, amphibole, and zeolites, with or without oxides and sulfides. The mineral assemblages and electron microprobe compositions are used to infer relative temperatures of the mineralisation. These data are combined with structural measurements and field observations to constrain the tectonic evolution of the island. The data best fit a model involving formation of Macquarie Island crust at the inside comer of a ridge-transform intersection. Regional Geology Macquarie Island is located approximately 1200 km southwest of New Zealand. It is the only subaerial exposure of non-plume-related oceanic crust that still lies within the basin in which it formed. The crust of Macquarie Island formed about 10 million years ago at a slowspreading ridge segment along the Australian-Pacific plate boundary (Duncan et aL, 1988; Vame et aL, 2000). Along this boundary, the oceanic spreading, which had started in Eocene time, became increasingly oblique until extension become nearly parallel to the present-day plate boundary. Then, the boundary evolved into the present transform plate boundary (Lamarche et al, 1997; Massel et al, 2000). Therefore, the oceanic crust exposed on Macquarie Island formed during the last stages of seafloor spreading in this area. The subsequent right-lateral transpression along the Australian-Pacific plate boundary has resulted in uplift along the Macquarie Ridge Complex. Macquarie Island, which is the only subaerial exposure of this uplifted area, lies about 4.5 km east of the major active plate boundary. Macquarie Island exposes oceanic crust and uppermost mantle rocks ranging from extrusive lavas (pillow basalts, massive basalt flows and hyaloclastites) and minor sedimentary rocks (sedimentary breccias, sandstones, mudstones, limestones and cherts) to sheeted dykes (basalts and dolerites), gabbros (typical gabbros, olivine gabbros and troctolites) and partly serpentinized peridotites (harzburgites, wehrlites and dunites) (Figure 1). Extrusive rocks outcrop mainly in the southem two-thirds of the island with deeper rocks being well-exposed in the northem part. The controversy Recently, several models were proposed to explain the formation of the island. Wertz et al (2003) proposed that Macquarie Island formed at an inside comer of a spreading ridgetransform fault intersection. Rivizzigno and Karson (2004) consider that many structural features of the island result from an oblique seafloor spreading setting. Finally, Dijkstra and Cawood (2004) argue that the growth of the Macquarie Island crust involved multiple magmatic events.

138 A detailed study of faults and associated mineralisation has been conducted to constrain the origins of this unusual ophiolite. 158°50'

158°55'j

54°30' 0

2 4km

Finch-Langdon

Brothers Point

Green Gorge

Waterfall Bay

54°45' Carol ine^^ Cove Hiird Point

Key E ] Volcanic m Dykes m Gabbro • Peridotite S Fault ^ Location of samples

Figure 1 - Geological map of Macquarie Island showing both the recent faults and faults that formed during seafloor spreading. Faults The numerous faults that cut Macquarie Island can be grouped into two major categories (Figure 1): 1) seafloor spreading-related faults and 2) recent faults that formed in the subsequent transform or transpressional setting following the cessation of volcanism. The two groups differ markedly in faulting style and orientation. In addition, hydrothermal fluids have deposited silicates, oxides, carbonates and sulfides along some faults. The recent faults generally trend NE and NNE. In the southem part of the island, the strike changes to NNW and NW. A dominantly normal sense of motion is observed. These faults have sharp, fresh fault scarps that affect the topography and have no or very limited mineralisation. The few mineralised recent faults formed at very low temperatures (clay, chlorite, muscovite; Daczko et al, 2003). Conversely, seafloor spreading-related faults have low relief traces and show abundant hydrothermal mineralisation (mainly epidote and prehnite) of higher temperatures. Mineralisation is observed in discrete faults, fractures, veins, or brecciated fault zones. This seafloor spreading-related faulting and associated mineralisation have been investigated in several sites on the island in this study (Figure 1).

139 The seafloor spreading-related faults crosscutting the extrusive and sedimentary rocks display variable strike and dip. Strikes are quite scattered and range from NE-SW (Hurd Point) to NW-SE (Green Gorge) through E-W and N-S and can also be highly variable within a site. Two subsets are generally observed within a site, one with shallow to moderate dips and another with steeper dips, as for instance at Hurd Point (Figure 2a). Lineations are rare but show mainly a shallow plunge suggesting a dominantly strike-slip motion. Slicks on fault planes and offsets indicate either sinistral or dextral kinematics. In the sheeted dykes of Waterfall Bay, two types of fault are observed: NW-SE-trending discrete mineralised faults with a shallow to moderate dip (Figure 2b) and NW-SE striking brecciated fault zones with a steeper dip (Figure 2c). On the brecciated fault zones, kinematics is commonly normal. The relative chronology of those two types of faults is not clear. However, in other sheeted dykes, north of Sandy Bay, the discrete mineralised faults are earlier than the brecciated fault zones. Petrology and Mineral Chemistry The petrology and mineral chemistry of mineralisation and recrystallised host rocks of seafloor spreading-related faults have been analysed. Most of the studied samples come from the extrusive rocks or the sheeted dykes. A few sedimentary breccias and one gabbro sample have also been investigated. CHnopyroxene is commonly altered and maybe totally replaced by a secondary amphibole. Clinopyroxene compositions range from augite to diopside. Plagioclase compositions range from albite to bytownite (An72) (Figure 3a). Rare orthoclase grains are observed (Figure 3a). The most albitic plagioclase and the orthoclase are probably secondary hydrothermal minerals. Amphibole is observed in sheeted dykes and gabbros and with only very few exceptions is absent from extrusive and sedimentary rocks. Most of the amphiboles replace clinopyroxene grains but some also occur in veinlets. Amphibole grains are mainly greenish and fibrous. By using the classification of Leake et al (1997), the amphibole compositions range from actinolite to low alumina Mg-homblende. Chlorite occurs as irregular clusters or isolated grains in variable amount and also filling vesicles. Most of the chlorite grains are greenish but colourless grains also occur. Two types can be distinguished for the composition point of view. Chlorites from Caroline Cove near a massive sulfide deposit have low Si (from 27.4 to 30.9 wt%), high Al (from 13.8 to 18.3 wt%) contents, high #Fe (average of Fe^"'/(Fe^''+Mg)*100 = 33.0) and low Ca+Na+K. Conversely, chlorites from other extrusive and sedimentary rocks have high Si (from 28.3 to 43.6 wt%), low Al (from 8.2 to 14.7 wt%) contents, low #Fe (average 20.8) and high Ca+Na+K. In the classification of Hey (1954), chlorites from Caroline Cove range between pycnochlorite and diabantite, whereas the other sites show compositions ranging from diabantite and penninite to talc-chlorite (Figure 3b). The high Ca+Na+K content of the second type of chlorites may indicate the presence of interlayered di-octahedral clays, such as illite and smectite, and therefore lower temperatures (Deer et al, 1992).

140

n = 61 Kurd Point

n = 69

Waterfall Bay Discrete mineralised faults N

n = 27

Waterfall Bay Brecciated fault zones Figure 2 - Stereonets (equal area projection, lower hemisphere) of mineralised seafloor spreading-related faults, (a) Hurd point, (b) Waterfall Bay, discrete mineralised faults, (c) Waterfall Bay, brecciated faults zones.

141 orthoclase

•

Hurd Point

•

Caroline Cove

Chlorite 1.0

A Green Gorge

Brunsvigite

X Brothers Point

0.8

Finch-Langdon (extrusives and sediments) o Finch-Langdon (gabbro) • Waterfall Bay (sheeted dykes)

Pseudothuringite

0.6

Ripidolite Dialiiantite

'ycnochlorite

ti. w 0.4 ri

•

Conindophilite

0.2 Sheridanitc Clinochlore

•—^A^x

Pennintte

Talc-chlorite

0.0 5.0

6.0

8.0

anorthite

Epidote

Prehnite

1.2

2.5

1.0

2.0

0.8

£ 1.0

0.6 0.4

0.5 0.0 3.0

7.0

0.2 0.0 3.5

4.0

4.5

5.0

5.5

6.0

2.9

3.1

3.3

3.5

3.7

3.9

Figure 3 - Mineral composition of (a) feldspar, (b) chlorite, (c) epidote and (d) prehnite of mineralised samples. Cross and full symbols are for extrusive and sedimentary rocks, open symbols are for sheeted dykes and gabbro. Epidote, prehnite and calcite are the most abundant hydrothermal minerals. Epidote occurs as irregular patches, clusters or in veinlets and is mainly present in sheeted dykes, gabbros and near the sulfide deposit of Caroline Cove. Epidotes have nearly constant Si02 and CaO content but vary significantly in AI2O3 and Fe203. Pistacite content, i.e. ranges from 10.6 to 41.1, that is from clinozoisite to pistacite composition. Except in epidote from a sheeted dyke, silica fills the tetrahedral position and thus nearly all aluminium is in sixfold-coordination. The strong negative correlation between Fe and A1 supports the presence of Fe in sixfold-coordination (Figure 3c) and thus its ferric state. Samples from sheeted dykes and gabbro display high A1 and low Fe content, consistent with crystallisation under higher temperature and/or lower oxygen fugacity (Liou et al, 1983). Prehnite is mainly observed in veinlets in all types of rocks. Prehnite has a constant Si02 and CaO content but displays significant variations in AI2O3 and especially Fe203. A strong negative correlation is observed between Fe and A1 but unlike the epidotes, composition varies within samples (Figure 3d). Calcite is mainly observed in veinlets but occurs also as isolated grains. Analcite is observed in veinlets in a few samples, mainly in extrusive and sedimentary rocks. Zeolites are found in a few samples of extrusive and sedimentary rocks and occur in veinlets or vesicles. Opaque minerals such as oxides (ilmenite, haematite) and sulfides are also present in variable amount along some faults. Discussion and Conclusion Seafloor spreading-related faults bear low-grade mineralisation ranging from zeolite facies to greenschist facies. Most of the samples display a mineralogy characteristic of the prehnite facies. The highest temperature mineralisation is found only in the deepest rocks (sheeted

142 dykes, gabbro) with more amphiboles and higher temperature epidotes. In these, minerahsed faults show strike-slip and normal motions. Extrusive and sedimentary rocks show a lower temperature range for mineralisation, from zeolite facies to prehnite facies. Among the extrusive samples, the highest temperature mineralisation is found in the southem part of the island (Hurd Point, Caroline Cove) or close to the Finch-Langdon fault (Unity Point). In the extrusive and sedimentary rocks, only strike-slip faults are observed. The grade of mineralisation seems to be above all related to the structural level. The dominant shallow plunge of mineral slickenlines in many mineralised structures is consistent with dominantly strike-slip motion. In addition, the faults show highly variable strike at many sites and between sites across the island. These observations best fit the interpretation of Wertz et al (2003), consistent with complex structures produced by the interaction of the seafloor spreading ridge and transform structure identified km offshore. References Daczko, N.R., Wertz, K.L., Mosher, S., Coffin, M.F., and Meckel, T.A., 2003. Extension along the AustralianPacific transpressional transform plate boundary near Macquarie Island. Geochemistry, Geophysics, Geosystems, 4. Deer, W.A, Howie, R.A., and Zussman, J., 1992. An introduction to the rock-forming minerals. Longman Scientific and Technical, Harlow, 2nd edition, 696 p. Dijkstra, A.H., and Cawood, P.A., 2004. Base-up growth of ocean crust by multiple phases of magmatism: field evidence from Macquarie Island. Journal of the Geological Society, London, 161: 739-742. Duncan, R.A., Vame, R., Banks, M.R., and Smith, S.J., 1988. The age and distribution of the igneous rocks of Macquarie Island. Proceedings Macquarie Island Symposium, 122: 45-50. Hey, M.H., 1954. A new review of the chlorites. Mineralogical Magazine, 30: 277-292. Lamarche, G., Collot, J.-Y., Wood, R.A., Sosson, M., Sutherland, R., and Delteil, J., 1997. The OligoceneMiocene Pacific-Australian plate boundary, south of New Zealand: evolution from oceanic spreading to strikeslip faulting. Earth and Planetary Science Letters, 148: 129-139. Leake, B.E., et al, 1997. Nomenclature of amphiboles: Report of the subcommittee on amphiboles of the International Mineralogical Association, Commission on new minerals and mineral names. The Canadian Mineralogist, 35: 219-246. Liou, J.G., Kim, H.S., and Mamyama, S., 1983. Prehnite-epidote equilibria and their petrologic applications. Journal of Petrology, 24: 321-342. Massel, C., Coffm, M.F., Mann, P., Mosher, S., Frohlich, C., Schuur, C.L., Kamer, G.D., Ramsay, D., and Lebrun, J.-F., 2000. Neotectonics of the Macquarie Ridge Complex, Australia-Pacific plate boundary. Journal of Geophysical Research, 105: 13457-13480. Rivizzigno, P.A., and Karson, J.A., 2004. Structural expression of oblique seafloor spreading in the Macquarie Island ophiolite, Southem Ocean. Geology, 32: 125-128. Vame, R., Brown, A.V., and Falloon, T., 2000. Macquarie Island; its geology, structural history, and the timing and tectonic setting of its N-MORB to E-MORB magmatism. In: Dilek, Y., Moores, E.M., Elthon, D., and Nicolas, A., (eds). Ophiolites and oceanic crust; new insights from field studies and the Ocean Drilling Program. Geological Society of America Special Paper 349: 301-320. Wertz, K.L., Mosher, S., Daczko, N.R., and Coffm, M.F., 2003. Macquarie Island's Finch-Langdon fault: A ridge-transform inside-comer stmcture. Geology, 31: 661-664.

143

Time scales of magmatic processes: a review of recent U-series results Simon Turner GEMOC, Department of Earth and Planetary Sciences, Macquarie University, Sydney NSW 2109

U-series isotope measurements have revolutionised the Earth sciences by offering the only quantitative constraints on time scales applicable to the physical processes that take place on the Earth. Over the past decade numerous studies have been undertaken on magmatic systems from a variety of tectonic settings providing new information about the mechanisms of melt formation, transport, differentiation and degassing. There are now several detailed U-series traverses across ocean islands which show a symmetrical pattern of disequilibria. This is most consistent with axisymmetric upwelling driven by thermal buoyancy. Inversion of U-Th and U-Pa disequilibria indicate that the rates of this upwelling are on the order of 5-10's cm per year and the magnitude of disequilibria correlate with independent estimates of buoyancy flux. Using these models, RaTh disequilibria are consistent with, but do not demand, rapid channelled melt ascent. Beneath island arcs, fluid addition of U appears to occur over lO's kyr with the last increments adding Ra only a few 100-1000 years prior to eruption. This supports models in which melting is controlled by the thermal structure of the mantle wedge and provides the most compelling case for channelled melt ascent. U-Pa data require some form of dynamic melting involving a matrix flow rate which is similar to the local rate of convergence. Magma differentiation appears to occur over lO's kyr consistent with the time scales for conductive cooling at mid to lower crustal depths. Degassing post-dates differentiation and occurs over decades prior to eruption. At midocean ridges, U-Th-Pa disequilibria require dynamic melting models in which the peridotite matrix is upwelling through the melting region at a rate of a few cm per year. The rate of meh ascent, and whether this occurs via percolative or channelled flow, remains controversial but recent ^^^Pb data may require melt ascent in decades.

144

Where do high-level S-type granite magmas come from? Ron H. Vernon Department of Earth & Planetary Sciences and National Key Centre for GEMOC, Macquarie University, Sydney, NSW 2109, Australia. rv^emon@els.mq.edu.au

Experimental evidence indicates that high-level (contact aureole, batholithic) S-type granite magmas are formed at granuHte facies conditions (850-950°C), whereas metasedimentary enclaves in these granites have prograde amphibolite facies assemblages, and hence are xenoliths, not restite. Amphibolite facies migmatite complexes are suitable sources for small S-type diatexites ("dirty granites"), such as the Cooma Granodiorite. However, migmatites are probably unsuitable for the production of larger high-level S-type granites, except for leucogranites. Microgranitoid (typically microtonalite) enclaves in S-type granites of the Lachlan Fold Belt are former globules of hybrid magma formed outside the pluton by mixing of peraluminous felsic magma with more mafic magma of ultimate mantle derivation. Mantle-derived mafic magma is probably necessary as a heat source for cmstal melting at relatively low pressure, and may mix with anatectic magmas in a "MASH zone." A suitable model for the upper parts of such a zone may be the Hidaka Complex, northem Hokkaido, Japan, in which hi^-temperature (> 900°C) peraluminous orthopyroxene-gamet tonalite sheets and abundant mafic rocks have intruded granulite facies metasediments at the deepest exposed levels, and cordierite tonalite sheets have intruded amphibolite facies metasediments at higher levels. Though some of the orthopyroxenegamet tonalite contains small orthopyroxene-gamet aggregates that may be restite, and some contains partly melted granulite facies xenoliths (therefore not restite), much of the orthopyroxene-gamet tonahte has cleared itself of restite. Because it is intrusive, the tonalite magma must have been formed at deeper levels (possibly several km deeper than the deepest rocks exposed) in a possible "MASH zone" characterized by abundant mafic and mixed magmas, as well as granulite facies restite. Such a zone may be a suitable source for high-level S-type granite magmas, as well as the hybrid magmas represented by the microtonalite enclaves. The higher-level amphibolite facies metasediments may be a suitable source for the metasedimentary xenoliths.

145

Geochemical and isotopic signatures of Mount Cameroon lavas (West Africa) Emmanuel W. Wembenyui*, K.D. Collerson and J.X. Zhao Advanced Centre for Queensland University Isotope Research Excellence (ACQUIRE), The University of Queensland, Qld 4072 Australia. *E-mail: s4Q40985@student.uq.edu.au and wembenyui@yahoo.co.uk Mount Cameroon is a plume-generated construct at the passive continental margin of the African West coast. It falls within the ocean-continent boundary of the Cameroon Volcanic Line (CVL), which is a volcanic lineament situated between the West African and Congo Cratons. The lineament is an ~ 1600 km long chain of Tertiary to Recent alkaline volcanic centres that are located on both continental and oceanic crust and are partly superimposed upon the Central African Shear Zone (Ubangoh et al, 1988). Mount Cameroon is underlain by the Pan African basement, which comprises schists, gneisses and alkaline leucogranites (Nnange et al, 2001). There has been volcanic activity since prehistoric times as indicated by the remnants of many undated lavas around its flanks. Recent documented events occurred in 1909, 1922, 1959, 1982, 1999 and 2000, producing lavas that may be used to unravel petrogenetic processes that account for the chemical variability observed in them with the goal of establishing a model of the plumbing system. Understanding of magma transport leads to a better knowledge of eruption timing, with implications for hazard mitigation and civil defence (Gamble et al, 1999). Geochemical data from representative lava samples of 1959, 1982, 1999 and 2000 eruptions, as well as from two prehistoric events at Ideneau and Buea, around the flanks of Mount Cameroon indicate that all the rocks are basic alkaline and silica undersaturated with Si02 content varying from 43.76-47.59 wt%, MgO content in the range 4.57-9.89 wt% and Mg number (Mg#=100Mg/(Mg + Fe^^) ranging between 45.02 and 58.63. The Ideneau suite of rocks is the most primitive of the studied rocks with MgO > 9 wt% and Si02 < 44 wt%. Total Alkalis versus Silica (TAS) diagram shows that all samples fall within the tephrite/basanite, trachybasalt and basalt fields (Fig. 1).

SiO^ (wt%)

Fig. 1: Classification of Mount Cameroon lavas on the Total Alkali vs. Silica, (TAS) diagram. The diagram shows that the lavas are alkahne.

146

Lavas produced under low degrees of mantle melting are generally alkaline, silica under saturated and characterized by high concentrations of Ti02, Na20, K2O, low Mg# as well as low ratios of Ca0/Al203 (Turner and Hawkesworth, 1995). These petrochemical characteristics are consistent with Mount Cameroon lavas (Si02 in the range 43-48 wt.%), thus are presumed to have been formed by low degrees of partial melting. The most primitive samples are the Ideneau lavas. They are distinct from all other lavas by exhibiting the lowest Si02 < 45 wt% and AI2O3 < 1 3 wt% content and by the highest MgO concentrations, generally > 9 wt%. This is consistent with formation under higher pressures (i.e greater depths) than the majority of the lavas. High-pressure magmas are generally Al-poor (Green and Ringwood, 1967) and when melting is initiated at great depths, the resulting melts are marked by low concentrations of Si02, AI2O3, Na20 and high MgO, total Fe203, CaO and Ca0/Al203 (Klein and Langmuir, 1987). On Harker plots, lavas from 1959, 1982, 1999 and Buea define a broad cluster marked by internal groupings, which clearly distinguish between the various eruption episodes whereas the Ideneau lavas plot as a distinct unit (Fig. 2).

Fig. 2: Harker diagrams of Mount Cameroon lavas showing major elements plotted against Si02 (wt.%). The Ideneau lavas tend to show a persistent offset from the bulk of rock suites.

Lava compositional changes, in addition to showing the important role of fractional crystallization of different magma batches, are interpreted to be the hallmark of multiple volcanological and petrologic processes such as mafic recharge and magma mixing (Hobden et al,, 1999), and mingling between fresh magma inputs and stagnant melts arrested as dykes and sills in the near surface (Gamble et al, 1999). In addition to fractional crystallization and partial melting, the geochemical heterogeneity observed in the studied rocks can be explained by mixing between remnant magmas of earlier flows and new magma inputs. Mixing of shallow stored magma with new mafic injections suggests that after initial eruption, the volume of the magma is sufficiently large that it can remain molten and not crystallise significantly over years to decades (Teasdale et a/., 2005). Chondrite normalized trace element patterns indicate that the lavas are enriched in Light REE but depleted in the Middle and Heavy REE, with no Eu anomaly, which is consistent with the subdued role of plagioclase as a fractionating phase. The evolution of the lavas appears driven by fractional crystallization of mainly olivine and clinopyroxene, which is confirmed by a strong

147

positive correlation between compatible elements and Mg# where Sc, Ni and Cr are seen to decrease with decreasing Mg#. On the other hand, primitive mantle normalized plots display distinctive ODB-like Rb deficiencies and negative K-anomalies, which are in general agreement with the arrays generated from volcanic rocks in the oceanic sector of the Cameroon Line (Lee et a/., 1994), (Figs. 3a and b).

Fig. 3: (a) Chondrite normalized REE plots and (b) Spider diagram of Mount Cameroon lavas showing trace elements normalized to the composition of the primitive mantle. MORB, OIB and average continental cmst are included for comparison.

Low Rb and K abundances be explained by the presence of an alkali bearing phase such as phlogopite or K-richterite in magma source regions (Stein and Hofmann, 1992). The Ideneau lavas are also distinct in that they display the highest compatible element abundances of all the Mount Cameroon rocks for example Ni > 145 ppm and Cr > 400 ppm whereas samples from all other localities display remarkably lower abundances but more scatter Ni (29 - 87 ppm) and Cr (28 - 144 ppm). This is consistent with the conclusion that the Ideneau suite of rocks are the most primitive and the least modified by fractional crystallization processes, whereas the others have evolved to varying degrees through fractionation. The results also indicate that the temporal variations in major and trace elements do not mirror variability in isotopic data. Pb-Sr-Nd data, ^^W^Vb=20.2-20.4, ^^Sr/^^Sr=0.70326-0.70336 and ^^^Nd/^^Nd=0.512755-0.512855 or 8Nd(0)=+2.3 - +3.4, indicate that the lavas are isotopically homogeneous and characterised by a high or 'HIMU' signature, indicative of high timeintegrated Pb enrichment resulting from U+Th/Pb fractionation. Temporal changes in major element chemistry, which are not accompanied by systematic variations in isotope composition have been interpreted to represent changes in the rate of melt supply to magma reservoirs (Regelous et al, 1999). Therefore magma supply beneath Mount Cameroon is not continuous, but rather temporally varies as a fimction of the prevailing geodynamic processes. The homogeneity of isotopic data for both the Prehistoric and Neogene lavas allows the following conclusions to be made: (a) All lavas from the different events are co-magmatic because isotopes are more diagnostic petrogenetic indicators relative to major and trace elements;

148 (b) The magmas are essentially free of crustal contamination as isotopes are highly sensitive to assimilation of isotopically different crust but insensitive to other petrogenetic processes such as fractional crystallization and partial melting, etc. Mount Cameroon is located within a passive continental margin, hence the heterogeneity observed in the major and trace element data cannot be explained by incorporation of ancient subducted material, but simply reflect modification by metasomatic and/or petrogenetic processes such as fractional crystallization, variable degrees of mixing and partial melting that acted on the lavas during ascent; (c) They are generated from a single plume that has remained almost unchanged through its eruptive history (ca. 9 Ma).

References

Gamble, J. A., Wood, C. P., Price, R. C., Smith, 1. E. M., Stewart, R. B., and Waight, T., 1999. A fifty year perspective of magmatic evolution on Ruapehu Volcano, New Zealand: verification of open system behaviour in an arc volcano. Earth and Planetary Science Letters, 170 (3): 301-314. Green, D. H., and Ringwood, A. E., 1967. The genesis of basaltic magmas. Contributions to Mineralogy and Petrology, 15: 103-190. Hobden, B. J., Houghton, B. F., Davidson, J. P., and Weaver, S. D., 1999. Small and short-lived magma batches at composite volcanoes: time windows at Tongariro volcano. New Zealand. Journal of the Geological Society, 156: 865-868. Klein, E. M., and Langmuir, C. H., 1987. Global correlations of ocean ridge basalt chemistry with axial depth and crustal thickness. Journal of Geophysical Research-Solid Earth and Planets, 92 (B8): 8089-8115. Lee, D. C., Halliday, A. N., Fitton, J. G., and Poli, G., 1994. Isotopic variations with distance and time in the volcanic islands of the Cameroon Line - evidence for a mantle plume origin. Earth and Planetary Science Letters, 123 (1-4): 119-138. Nnange, J. M., Poudjom Djomani, Y. H., Fairhead, J. D., and Ebinger, C. J., 2001. Determination of the isostatic compensation mechanism of the region of the Adamawa dome, West Central Africa using the admittance technique of gravity data. Journal of Science and Technology (AJST) Science and Engineering Series, 1 (4): 29-35. Regelous, M., Niu, Y., Wendt, J. L, Batiza, R., Greig, A., and Collerson, K. D., 1999. Variations in the geochemistry of magmatism on the East Pacific Rise at 10 degrees 30 ' N since 800 ka. Earth and Planetary Science Letters, 168 (1-2): 45-63. Stein, M., and Hofmann, A. W., 1992. Fossil plume head beneath the Arabian lithosphere. Earth and Planetary Science Letters, 114 (1): 193-209. Teasdale, R., Geist, D., Kurz, M., and Harpp, K., 2005. 1998 eruption at Volcan Cerro Azul, Galapagos Islands: 1. Syn-emptive petrogenesis. Bulletin ofVolcanology, 67 (2): 170-185. Turner, S., and Hawkesworth, C., 1995. The nature of the sub-continental mantle - constraints from the majorelement composition of continental flood basalts. Chemical Geology, 120 (3-4): 295-314. Ubangoh, R. U., Pacca, L G., and Nyobe, J. B., 1988. Palaeomagnetism of the continental sector of the Cameroon Line, West Africa. Geophysical Journal International, 135 (2): 362-374.

149

Iron isotopes as a potential new tool in igneous geochemistry and cosmochemistry Helen M. Williams^*, A.N. Halliday^, C. A. McCammon^ A.H. Peslie/, N. Teutsch\ S. Levasseur^ J.-P. Burg^ ^Department of Earth Sciences, ETH-Zurich, Sonneggstrasse 5, CH-8092 Switzerland •Current address: GEMOC National Key Centre, Department of Earth and Planetary Sciences, Macquarie University, NSW 2109, Australia. ^Department of Earth Sciences, University of Oxford, Parks Road, Oxford, 0X1 3PR, United Kingdom ^Bayerisches Geoinstitut, Universitat Bayreuth, D-95440 Bayreuth, Germany "^Texas Center for Superconductivity and Advanced Materials, University of Houston, HSC Building 202, Houston, TX 77204, USA Recent advances in analytical techniques, such as high-precision multi-collector inductivelycoupled plasma mass spectrometry (MC-ICP-MS) have accelerated the development of many new and powerful isotopic tracers that can be applied to a wide range of problems in the geosciences. The stable isotopes of the transition metal elements, e.g. Fe, Cu and Zn, have attracted considerable interest over the last five years as they can be applied to fields as diverse as economic geology, mantle geochemistry, bigeochemistry, paleoceanography, and cosmochemistry. Iron is the one of the most abundant elements on Earth and its mobility and chemical behaviour are strongly affected by redox processes, occurring both at the Earth's surface and within its deep interior. Theoretical Mossbauer studies of stable iron isotopic fractionation predict that significant iron isotope fractionation will be controlled by the coordination and oxidation state of iron in different minerals (Polyakov and Mineev, 2 0 0 0 ) . However, variations in the iron isotope compositions of mantle rocks and meteorites have only become recognised recently. The reason for this is that these isotopic variations are so small that they can only be detected using extremely high-precision MC-ICP-MS analytical techniques, which have been in a state of development over the last five years. The most recent iron isotope studies have shown that significant iron isotope variations exist in igneous bulk-rocks, minerals and meteorites. Iron isotope fractionations of ~ 0 . 2 % o have been found between ohvine and clinopyroxene in mantle peridotites, and between olivine and metal in pallasites (Zhu et al, 2 0 0 2 ) . Mantle spinels show considerable iron isotope variation, which may reflect mantle redox processes (Williams et al, 2 0 0 4 ) . A significant difference may also exist between the iron isotope compositions of bulk igneous rocks from the Earth and those from the Moon, which has been interpreted in terms of isotopic fractionation induced by the evaporation and condensation of iron following the moon-forming giant impact (Poitrasson et al, 2 0 0 4 ) . However, the processes governing iron isotope fractionation in igneous rocks remain poorly understood, limiting their use as a tracer. Here we show that there are significant variations in the iron isotope compositions (expressed as which denotes the deviation of the ^^Fe/^'^Fe ratio from that of the pure iron standard IRMM-14 in parts per 1 , 0 0 0 ) of mantle rocks ( 0 . 9 % o ) and minerals (olivines 0 . 6 % o ,

150 clinopyroxenes 0.9%o and orthopyroxenes 0.8%o), although spinels still show the greatest total variation of L7%o (data taken from Williams et al, 2 0 0 4 , 2 0 0 5 ) . Positive linear correlations with slopes that are, within error, equal to unity are found between the S^^'^^'^Fe values of coexisting orthopyroxene, clinopyroxene and olivine (Fig. 1), strongly suggesting that the S^^^^'^Fe values of these minerals reflect intra-sample mineral-mineral isotopic equilibrium. Positive correlations between the ^^^'^^^Fe values of silicate minerals and spinels also exist (Fig 1), although they are more scattered, which could be caused by late disturbance of mineral-spinel isotopic equilibrium. Bulk-rock, clinopyroxene and spinel ^"^Fe values correlate with chemical indices of both melt extraction and oxidation (Fig. 2). These variations cannot be explained by any kind of fluid or melt mixing or contamination scenario, as simple mass-balance calculations demonstrate that iron is simply too abundant in the mantle to be sensitive to such processes. Melt extraction, in combination with significant changes in mantle oxidation state, Figure 1 may be the best explanation for Fe isotopic variations in mantle Slopo=1-11±0.53 (2o) lntercept=-0.03±0.13{2a) r peridotites. Iron isotopes R'=0.B4 A may therefore provide a powerful new proxy for monitoring the redox evolution of the mantle, which is critical to understanding mantle melting, core-mantle exchange, volatile speciation and the evolution of the mantle and atmosphere. .1.6 Fe values of clinopyroxene (cpx) and spinel vs. S^^^^'^Fe values of orthopyroxene (opx) and olivine (oliv) from the same peridotite. Error are 2 SD.

Figure 2

2 values of clinopyroxene (cpx), spinel and whole-rock plotted against chemical indicators of depletion (Cr#, MgOcpx) and oxidation (Fe^VDFespmei, Alogf02). Errors for S^^^^^Fe values and spinel Fe^VEFe are 2 SD. Errors on AlogfD2 are the maximum range in Alogf02 values that could be produced from different combinations of pressure, temperature and mineral end-member abundances. They are ±0.7 log-bars for all samples except P6 and PI2, for which they are ± 1 . 0 log-bars. The values, expressed relative to I R M M - 1 4 , for M O R B (average 0.18%o) and O I B (average

151 0.12%o) are taken from (Beard et al, 2003), Alogf02 values are taken from Kiline et al (1983) and Christie et al

(1986).

Figure 3

Iron isotopes can also be used to place constraints on the formation and -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 evolution of planetary cores. The magmatic iron meteorites (e.g. groups IIAB, IIIAB and IVAB) are considered to be remnants of the cores of differentiated asteroids and may be used to place preliminary constraints on the behavior of iron isotopes during planetary core formation and crystallization. Theory (Polyakov and Mineev, 2000) predicts that there will be significant iron isotope fractionation between native iron, silicates (e.g. olivine) and iron sulphide (troilite, FeS), with isotopically heavy iron being preferentially incorporated in the metal relative to silicates and FeS. We present iron isotope data for metal and sulfide fractions extracted from iron meteorites that provide a general confirmation of these predictions (Fig. 3), as all of the sulphides extracted from the iron meteorites are isotopically lighter than the corresponding metal fractions by 0.3 to 0.7%o. The iron isotope compositions of the metal fractions separated from iron meteorites show very little variation, although metals from the IIAB group of magmatic iron meteorites appear to be slightly heavier than those of the other magmatic and non-magmatic groups. The heavier iron isotope compositions of the IIAB metals relative to the IIIAB and IVAB metals may be related to the large amounts of sulphur (ca. 17 wt%; Chabot, 2004) inferred to have been in the initial core of the IIAB parent body asteroid. Such a large amount of sulphur inferred for the IIAB core means a large FeS sink for isotopically light iron, leaving the metallic part of the IIAB core relatively depleted in light iron, or, to put it another way, enriched in isotopically heavy iron. In comparison, the IVAB iron meteorites are believed to originated from an parent body with a core that had no more than 1% sulphur, which provides little or no sink for isotopically light iron. This is consistent with our data, where the IVAB metals are, in general, isotopically lighter than the IIAB metals (Fig. 3). If this isotopic fractionation between metal and troilite can be regarded as representative of the high-pressure, high-temperature sulfide-melt fractionation during core crystallization in the major terrestrial liA sulphides

Murphy

, >

IIIA sulphide

illB sulphides |

|—

lilA metal

Mount Edith , Grant . Chupaderos 1 Bear Creek

. i

iVA sulphide

»B metals

' ' Sao Juilao de Moreira I — I t - H Sikhote Aiin 1 Mount Joy =11

Merceditas

^ ^

liA metals

1 Coahuila (x2) H Bennett County H Negrillos North Chile 1 Gressk {x2) II ^ J —I Guadtupe y Calvo

I Navajo

f+O-

_ H

- I Duchesne —I Bushman Land h - I Duel Hill h - ^ — I Yanhuitlan

- I Tawallah Valley

IllB metals

IVA metals

IVB metals

- I Warburton Range - f | Tiacotepec (x2)

1 Arispe

h E ^ Clark County Nelson County

HE silicate (x2)

| — H ^ ^ — I watson

IIICD sulphides

lAB sulphides i I—

—

-I--

" •

IC metal

|||F m e t a l s

HE metal

IIICD metals

1 Seelasgen — I — I Mundrabilla (x2)

I — C a n y o n Diablo lAB metals I — • — I Odessa t l i l H Toluca (x2) F # H Caddo County

152 planetary bodies (i.e. the Earth, Moon, Mars and the asteroid 4 Vesta) then we can potentially use the fractionation between iron metal and sulphide to infer the initial amounts of sulphur in the cores of these planets and to refine constraints on the relative core sizes of the terrestrial planets and the nature of core crystallization and core-mantle equilibration processes. References Beard, B.L., Johnson, C.M., Von Damm, K.L., and Poulson, R.L., 2003. Iron isotope constraints on Fe cycling and mass balance in oxygenated Earth oceans. Geology, 31 (7): 629-632. Chabot, N.L., 2004. Sulfur contents of the parental metallic cores of magmatic iron meteorites. Geochimica et Cosmochimica Acta, 68 (17): 3607-3618 Christie, D.M., Carmichael, I.S.E., and Langmuir, C.H., 1986. Oxidation-states of midocean ridge basalt glasses. Earth and Planetary Science Letters, 79 (3-4): 397-411. Kiline, A., Carmichael, I.S.E., Rivers, M.L., and Sack, R.O., 1983. The ferric-ferrous ratio of natural silicate liquids equilibrated in air. Contributions to Mineralogy and Petrology, 83: 136-140. Poitrasson, F., Halliday, A.N., Lee, D.-C., Levasseur, S., and Teutsch, N., 2004. Iron isotope differences between Earth, Moon, Mars and Vesta as possible records of contrasted accretion mechanisms. Earth and Planetary Science Letters, 223 (3-4): 253-266. Polyakov, V.B., and Mineev, S.D., 2000. The use of Mossbauer spectroscopy in stable isotope geochemistry. Geochimica et Cosmochimica Acta, 64 (5): 849-865. Williams, H.M., McCammon, C., Peslier, A.H., Halliday, A.N., Teutsch, N., Levasseur, S., and Burg, J.-P., 2004. Iron isotope fractionation and the oxygen fiigacity of the mantle. Science, 304: 1656-1659. Williams, H.M., Peslier, A.H., McCammon, C., Halliday, A.N., Levasseur, S., Teutsch, N., and Burg, J.-P., 2005. Systematic iron isotope variations in mantle rocks and minerals: The effects of partial melting and oxygen fiigacity. Earth and Planetary Science Letters, 235 (1-2): 435-452. Zhu, X.K., Guo, Y., Williams, R.J.P., Williams, CNions, R.K., Matthews, A., Belshaw, N.S., Canters, G.W., de Waal, E.C., Weser, U., Burgess, B.K., and Salvato, B., 2002. Mass fractionation processes of transition metal isotopes. Earth and Planetary Science Letters, 200 (1-2): 47-62.

153

Origin of carbonates and prehnite in the Prospect Intrusion, NSW Megan L. Williams and Paul F. Carr School of Earth and Environmental Sciences, University of Wollongong, Wollongong, NSW, 2522

Introduction The Prospect Intrusion is a small alkaline dolerite intrusion that is well-known both as a classic differentiated alkaline intrusion and also for its wide range of secondary minerals. The intrusion, located approximately 30 km west of Sydney, has been quarried as a source of aggregate since at least the 1830s and during its early history was visited by several famous scientists including Charles Darwin and James Dana. The first extensive study of the Prospect Intrusion was undertaken by Wilshire (1967) who used a combination of field, petrographic and major element data to identify the major rock types and their distribution. He proposed that emplacement resulted fi-om a single injection of magma followed by in-situ differentiation via fractional crystallisation and the diffusion of volatiles. Wilshire (1967) argued that these volatiles were entirely magmatic in origin and became trapped in the upper half of the intrusion to produce a wide variety of secondary minerals. These secondary minerals were described by England (1994) who also concluded that they were formed from residual hydrothermal fluids. The aim of the present study is to use the mineral assemblage together with isotopic data for carbonates and prehnite from the Prospect Intrusion to determine the characteristics of the parental hydrothermal fluids and to constrain the temperatures of these fluids. Geology The Prospect Intrusion is one of several Mesozoic igneous masses in the Sydney Basin. These masses form part of a large group of coeval intrusions extending across Antarctica, South Afiica, and the southern part of Australia that are believed to be associated with the early stages of the breakup of Gondwana (McBimey, 1984; Allen, 1999). The Prospect Intrusion was emplaced into the Sydney Basin sequence at the junction of the Triassic, fluvial Hawkesbury Sandstone and overlying shallow-marine Ashfield Shale implying a maximum cover of only -200 m (Wilshire, 1967; Herbert, 1997). K-Ar isotopic data for biotite from Prospect reported in Evemden and Richards (1962) provide an age of 172 Ma that, although probably a minimum for the emplacement age, is consistent with the stratigraphic relationships. The intrusion is dish-shaped in cross-section and oval in plan with maximum lateral dimensions of approximately 2700 x 1200 m, and a maximum thickness of 135 m. Seven major rock types (comprising basalt, picrite, alkaline olivine dolerite, alkaline dolerite, pegmatite, aplite and syenite) have been defined on the basis of field observations, petrographic and chemical data (Wilshire 1967). Three picrite zones form the lower half of the intrusion, and the upper half consists of interlayered alkaline olivine dolerite and alkaline dolerite with pods and veins of pegmatite, aplite and syenite. The development of a basaltic chilled margin provided

154 both an ideal insulating carapace conducive to slow cooling and attendant differentiation, and also a barrier that prevented escape of volatiles. Secondary minerals Most secondary minerals in the Prospect Intrusion are found in the upper half where they occur as interstitial and joint fillings, amygdales, and pseudomorphs after primary minerals. The major secondary minerals comprise analcime, prehnite, calcite and zeolites, particularly laumontite, natrolite and heulandite. In addition, minor amounts of apophyllite, aragonite, barite, chabazite, chlorite, marcasite, montmorillonite, opal, pectolite, phillipsite, pyrite, quartz and siderite have also been recognised. Studies by Wilshire (1967) and England (1994), supported by our observations, indicate that the order of formation of the major secondary species from oldest to youngest is: analcime

carbonates

laumontite -> prehnite-> heulandite -> carbonates

Wilshire (1967) considered that the volatiles responsible for the formation of the secondary minerals originated from within the intrusion on the basis of the following observations and deductions: 1. Chilled margin basalts show extensive carbonation, whereas shales near the contact lack carbonate. 2. Pegmatite pods contain abundant amygdales and the highest concentrations of alteration products, implying early concentration of volatiles. 3. The most extensive alteration occurs in the differentiated rocks in the upper half of the intrusion where the volatiles were concentrated. P-T conditions The secondary mineral phases at Prospect are typical of assemblages developed during hydrothermal alteration of mafic igneous rocks at low pressures. The presence of zeolites and prehnite but absence of epidote constrains the temperatures of formation. The minimum temperature for formation of epidote is ~200°C (Bird et al, 1984), thereby providing an upper limit for the temperature of the fluids. Laumontite has an upper stability limit of 230°C (Liou et ai, 1987) and a lower stability limit of 165-180°C at low pressures (Cho et al, 1987). These stability data imply that laumontite and the earliest carbonates originated from fluids with temperatures in the range of 165-200°C. The intrusion must have acted as the heat source for these fluids, as the high level of emplacement (<200 m cover) precludes elevated temperatures being related to the regional geothermal gradient. Prehnite is stable at these temperatures and pressures (Carr et al, 1999), and in the Prospect Intrusion it formed almost simultaneously with laumontite. In some cases it replaces laumontite, indicating lower temperatures (<165°C). Heulandite formed prior to the second episode of carbonate deposition and after laumontite, suggesting even cooler fluids were responsible for its formation. The upper stability limit of heulandite at low pressure is ~100°C (Noh and Boles, 1993), and it has also been found where the maximum temperature is known to be a few tens of degrees (Carr et al., 1999). Thus the second

155 generation of carbonates probably originated from fluids with temperatures between several tens of degrees and lOO^C. Carbonates Eleven carbonate samples were analysed for 6^^0smow, 5^^Cpdb and ^^Sr/^^Sr. These isotopic data show a wide range of values that are essentially continuous for both 5^^Cpdb and ^^Sr/^^Sr between -10.5%o and +3.4%o, and 0.70394 and 0.70713 respectively. The 6^^0smow data, however, are bimodal with Group 1 samples (3 calcites) having values between +4.5%o and +8.3%o, and Group 2 samples (7 calcites and 1 aragonite) with values between +19.4%o and +21.8%o.

The C and O isotopic signatures of the Group 1 calcites are consistent with derivation from a magmatic source, and the trend of decreasing C with increasing O isotopic values may reflect falling temperatures. Similarly, the ^^Sr/^^Sr values for these three calcites are consistent with derivation from a magmatic source. The much higher 5^^0smow values ( ~ + 2 0 % o ) for the Group 2 carbonates, combined with their S^^Cpdb values, imply formation at low temperatures from mixed meteoric and pore fluids. The measured ^^Sr/^^Sr values for these carbonates range from - 0 . 7 0 4 implying a magmatic origin, to - 0 . 7 0 7 implying substantial input from marine water and/or evolved sediments. Assuming equilibrium between the carbonate and source water, and a value o f - l % o for 5^^0smow of CO2 in equilibrium with water, the data can be used to calculate the temperature of carbonate formation. These data suggest that calcites in Group 1 formed at 160-195°G, the Group 2 calcites formed at 58-73°C while the aragonite formed at 51°C. These calculated temperatures are remarkably consistent with the temperatures suggested by the secondary mineral assemblages. Prehnite Six prehnite samples were analysed for 6^^0smow and 5Dsmow. All samples have very similar isotopic ratios with the ranges for 6^^0smow and SDsmow being - 0 . 5 % o to + 1 . 7 % o , and - 1 6 7 . 1 % o to - 1 5 8 . 3 % o respectively. This close similarity in isotopic ratios indicates that the parental fluids for all analysed prehnite samples had the same origin and isotopic composition. Values for 6Dsmow for the associated fluid cannot be calculated because a fractionation equation for prehnite-H2 has not been derived. The fractionation equation for prehnite-02 of Zheng (1993), however, can be used to calculate for the fluid from the measured isotopic data and an assumed temperature of formation. The calculated values for fluids in equilibrium with the prehnite are more negative than the measured 6^^0smow values for prehnite and become progressively more negative as the temperature of formation decreases. The fluid is interpreted to have originated from meteoric water alone, or a mixture of meteoric and magmatic water. Irrespective of the value of 5Dsmow for the associated fluid, this fluid must plot to the left of the fields for both primary magmatic and metamorphic waters, and progressively closer to the meteoric water line as the temperature of formation decreases.

156 Conclusions The clear subdivision of carbonates into two groups on the basis of their 6^^0smow values indicates derivation from fluids with different temperatures and/or compositions. Some of the secondary minerals of the Prospect Intrusion formed from magmatic fluids as suggested by Wilshire (1967), but others formed from hydrothermal fluids as the intrusion cooled. Calculated temperatures, derived from 5^^0smow data for the calcites and an assumed value of -l%o for the S^^OsMow of CO2 in equilibrium with water at the time of emplacement, suggest that the first generation calcites formed at 160-195°C. This temperature range is the same as that suggested by mineral stability relationships. The other carbonate samples analysed in the current study have much higher values for 5^^0smow (~+20%o) and a wide range of 8 Cpdb ^^^d values, reflecting variable input from meteoric water including pore water from the country rock. Calculated temperatures for these second generation carbonates suggest crystallisation at 51-73°C, which is consistent with the temperatures suggested by mineral stability relationships. All analysed prehnite samples have very similar values for both and 6D, indicative of formation from parental fluids that had the same origin and isotopic composition. These isotopic data imply that the parental hydrothermal fluids originated from meteoric water alone, or a mixture of meteoric and magmatic water. The upper part of a hydrothermal system can, be sealed by a caprock of minerals precipitated from the hydrothermal solutions. In the Prospect Intrusion both the carbonate rind described by Wilshire (1967) and the first-generation calcites were formed at higher temperatures from magmatic fluids, whereas the prehnite and second-generation carbonates deposited in the vesicles and fractures within the intrusion were derived from later, cooler fluids with significant meteoric input. Acknowledgments We thank Ross Pogson from the Australian Museum for supplying the samples for study. Darren Cann from Boral Quarries provided access to the quarry site, supplied core logs and drilling information, and core samples. Isotopic ratios were analysed at the Centre for Isotope Studies, a CSIRO-University consortium located at North Ryde, Sydney. References Allen, T.C., 1999. The Petrogenesis of the Jurassic Igneous Rocks of Southeastern Australia. PhD thesis, University of Sydney, Sydney (unpubl). Bird, D.K., Schiffman, P., Elders, W.A., Williams, A.E., and McDowell, S.D., 1984. Calc-silicate mineralization in active geothermal systems. Economic Geology, 79: 671-695. Carr, P.P., Pemberton, J.W., and Nunan, E., 1999. Low-grade metamorphism of mafic lavas, Upper Permian Broughton Formation, Sydney Basin. Australian Journal of Earth Sciences, 46: 839-849. Cho, M., Maruyama, S., and Liou, J.G., 1987. An experimental investigation of heulandite-laumontite equilibrium at 1000 to 20,000 bar P fluid. Contributions to Mineralogy and Petrology, 97: 43-50. England, B.M., 1994. Minerals of the Prospect Intrusion, New South Wales, Australia. The Mineralogical Record, 25: 185-194.

157 Evemden, J.F., and Richards, J.R., 1962. Potassium-argon ages in eastern Australia. Journal of the Geological Society of Australia, 9: 1-50. Herbert, C., 1997. Sequence stratigraphic analysis of Early and Middle Triassic alluvial facies in the Sydney Basin, Australia. Australian Journal of Earth Sciences, 44: 125-143. Liou, G. L., Maruyama, S., and Cho, M., 1987. Very low-grade metamorphism of volcanic and volcaniclastic rocks mineral assemblages and mineral facies. In: Frey, M. (ed). Low Temperature Metamorphism, Blackie, Glasgow and London: pp.59-113. McBimey, A.R., 1984. Igneous Petrology. Freeman, Cooper and Co., San Francisco. Noh, J.H., and Boles, J.R., 1993. Origin of zeolite cement from the Miocene sandstones in the North Tejon oil field, California. Journal of Sedimentary Petrology, 63: 248-260. Wilshire, H.G., 1967. The Prospect alkaline diabase-picrite intrusion. New South Wales, Austraha. Journal of Petrology, 8: 97-163. Zheng, Y.-F., 1993. Calculation of oxygen isotope fractionation in hydroxyl-bearing silicates. Earth and Planetary Science Letters, 120: 247-263.

158

Spatial resolution and the analysis of complex geometries in LA-MC-ICPMS Jon Woodhead, Janet Hergt and Roger Kemp School of Earth Sciences, The University of Melbourne, VIC 3010, Australia (i dwood@un i mel b. ed u. au) In typical LA-MC-ICPMS applications the requirement for large signals (ideally many volts, measured on Faraday cups), necessitates ablation using relatively large spot sizes, and comparatively rapid drill rates. Unfortunately, however, the size and/or geometry of many analytical materials precludes the use of such conditions. In such circumstances a number of altemative analytical approaches can be contemplated using a combination of appropriate ablation systems and time-resolved analysis (TRA) software.

Since ablation pits are generally many tens of microns in diameter and yet individual pulses may ablate only a few tenths of microns at a time, theoretically, depth profiling is capable of providing the best resolution and appears to be a viable technique as long as pit aspect ratios do not greatly exceed 2:1 (depthiwidth). This can be readily demonstrated using 'sandwiches' of standard materials stacked together to form a composite standard of appropriate dimensions. 0.284S

0.283-

0.282-

5 "

0.281 0.2800.279-

—— r —I— 100 —\— 120 80

1 s t I n t e g r a t i o n Summary Result: 1 Total Hf 2 176Lu/177Hf 3 176Hf/177Hf

160

—I— 140

Integrated from 61.4 (sees) to 92.8 (sees) Mean Error No. of Results 3.77 1.18E-01 147 of 158 0.001671 7.03E-05 158 of 158 0.282212 6.87E-05 148 of 153

2nd i n t e g r a t i o n Summary Integrated from 143.2 (sees) to 178.4 (sees) Result : Mean Error No. of Results 1 Total Hf 4.26 3.43E-02 170 of 177 2 176Lu/177Hf 0.001542 9.05E-05 177 of 177 3 176Hf/177Hf 0.282274 5.81 E-05 169 of 177

159 In many circumstances, however, (e.g. relatively large yet thin samples) depth profiling is not feasible and, in these cases line scans must be considered as an alternative analytical strategy. Line scans have their own inherent value in so much as they can provide robust checks on data quality if ablation of 'symmetrically zoned' materials are undertaken (e.g. under CL-control). In addition they often allow far greater spatial resolution than is possible using simple spot analyses. The use of excimer lasers also offers the potential for ablation of a 'slit' rather than a simple spot, thus maintaining high resolution while still allowing ablation of sufficient material for analysis—a feature of great utility for exhibiting very fine compositional zonation.

Time resolved results E

tS (>5

(>5

0.060-

(!o 0.056 0.052-

(0.70653±1 2lJ|ro.7091 61 ±5Q)

0.704

200

300

400

500

Finally, for samples with complex zonation, one further solution is to lower the spot size, increase the repetition rate, and ablate along a pre-digitised path using an appropriate translation

160 speed to achieve optimum signal intensity. Such an approach finds ready application, for example, in the interrogation of thin magmatic rims on many igneous zircons. In all these cases, adequate TRA software is essential to the task and yet, unfortunately, remains an area where individual users are often required to implement their own solutions. OutputCraphrCraphlWA,,

' © e o

nt Results

1st Int^ratfonpnd lnt«

bn_04 0.56-

0.40.3-

0.22 -

0 -

0.1 —T" 80

1— 100

—I—

0.0

—I—

160

1 0 -1

-2-3-

120

140

1.18

120

— I — 140

1 160

120

140"

160

0.284-

1.17-

0.283 -

1.161.15-

0.282

1.14-

0.281 -

1.130.280-

1.12—r— 80

— , 100

1 120

1— 140

160

3x10

2 1

0 -1 •

100

120

140

100

1st Integration Summary Result: 1 Total Hf 2 176Lu/177Hf 3 176Hf/177Hf

Integrated from 66 (sees) to 92 (sees) Error No. of Results Mean 2.13E-01 131 of 131 3.54 125 of 131 0.001921 6.39E-05 6.65E-05 124 of 131 0.282170

2nd I n t e g r a t i o n S u m m a r y Result: 1 Total Hf 2 176Lu/l77Hf 3 176Hf/177Hf

Integrated from 115.2 (sees) to 141.2 (sees) Mean Error No. of Results 4.11 2.53E-02 120 of 131 131 of 131 0.000540 3.03E-05 0.281249 124 of 131 5.45E-05

160

We have developed a potential solution to this problem for isotopic analyses on the Nu Plasma which allows visualisation of corrected isotope ratios versus time, and provides for user-defined baseline and 'on peak' integration windows. Examples will be shown of all these approaches in the analysis of a range of common geological and zoological materials.