OCEANIC PERIDOTITES ON MACQUARIE ISLAND: A UNIQUE RECORD OF MANTLE MELTING AND MELT PERCOLATION

Abstract

Macquarie Island (Southern Ocean) is a very rare occurrence of young (9 Ma) ocean crust exposed above sea level. As such, it is a key source of information about magmatic and tectonic seafloor spreading processes. Residual mantle peridotites are also exposed on the island. This unique location thus allows us to study both the residue and the product of mantle melting underneath a spreading centre, and allows us to test the co-genetic relation between oceanic mantle and crust. We will present a detailed petrographic analysis of the Macquarie Island harzburgites, and trace element geochemical as well as Nd and Sr isotopic data on fractions of clinopyroxene crystals. The Macquarie Island peridotites are among the most depleted peridotites from young oceanic basins worldwide in terms of modal composition and relatively compatible element concentrations, comparable to harzburgites from Hess Deep. At the same time, they show enrichment in light rare earth elements. The spoon-shaped trace elements patterns show a remarkable similarity to those in ophiolitic mantle peridotites, which are often interpreted as resulting from melting in supra-subduction zone settings. Macquarie Island thus acts as a key link between ophiolites and young ocean floor. Importantly, the spreading environment of Macquarie Island is well characterized and a supra-subduction zone setting can be ruled out unequivocally. On the basis of trace element modeling, we propose that the Macquarie Island harzburgites represent samples of upper mantle material that have undergone high degrees of accumulated melt extraction (21-22%). Modelling further shows that light rare earth enrichment was probably not the result of hydrous, supra-subduction style melting. In particular, observed mobile/immobile element ratios are inconsistent with such a model. Instead, a chromatographic melt percolation model produces a good match with the observed trace element patterns. The Macquarie Island harzburgites are thus examples of mantle rocks that have recorded cryptic metasomatism and re-equilibration associated with porous (unchanneled) flow of MORB-like melts. We also show that Nd isotope ratios are controlled by the isotope ratios in the percolating melts. Therefore, the similarity in isotope ratios between crustal and mantle rocks on the island cannot be used to infer that the harzburgites are the source residue of the oceanic crust. The high degrees of melting recorded by the harzburgites are inconsistent with the thin crust (c. 3 km), slow spreading rate (c. 2 cm/yr full rate) and the enriched character of the gabbros, dykes and lavas on the island, and argue against a simple co-genetic link between mantle and crust. The oceanic crust was probably produced from local highly enriched domains (‘plums’) within the mantle, with no significant contribution from the enclosing depleted mantle. A major implication of the study is that Macquarie Island- like, spoon-shaped trace element patterns in peridotites cannot be used as a ‘supra-subduction zone signature’ in ophiolites.