Geochemical variations in a depleted fore-arc mantle: The Ordovician Thetford Mines Ophiolite

Abstract

The ∼ 5 km thick mantle of the Ordovician Thetford Mines Ophiolite is dominated by harzburgite (≤ 5–6% clinopyroxene), cut by dunitic to orthopyroxenitic dykes. The plutonic crust is dominated by cumulates of boninitic affinity. The first lava cycle is dominated by boninites (and their differentiation products) with subordinate arc tholeiites; while the second cycle is exclusively of boninitic affinity. Boninites typically occur in fore-arc environments and are interpreted to be derived from previously depleted mantle as a result of decompression and influx of subduction-related fluids and melts. Thetford Mines Ophiolite mantle harzburgites are MgO-rich, and poor in CaO, Al 2O 3, TiO 2, Na 2O, and incompatible trace elements, with U-shaped primitive-mantle-normalized extended trace element profiles. Heavy rare-earth element abundances suggest that the Thetford Mines Ophiolite harzburgites are residues from ∼ 25 to 30% of fractional melting, or 30 to 35% critical melting (with 20% of each melt increment remaining in the source); as calculated from a fully-parameterized partition coefficient dataset. Tholeiites would represent 15–20% melting, while the boninites could represent the products of an additional 10–20% melting of the residues of tholeiite extraction. However, these models cannot explain the enrichment in light rare-earths (LREE) and large ion lithophile elements (LILE) seen in these otherwise refractory harzburgites or in the complementary lavas. A better match to the geochemical signatures is obtained if melting occurs synchronously with a continuous influx of slab-derived fluid and melt. About 0.005% of this arc component produces a much better fit to the Thetford Mines Ophiolite data. This amount is consistent with previously published Nd isotopic data on closely related rocks. However, bulk sediment contamination yields model residues that are too low in LILE Th, U, Nb, Ta, and LREE compared to Thetford Mines Ophiolite harzburgite, but produces model liquids that fit almost perfectly with Thetford Mines Ophiolite lavas. On the other hand, models involving contamination with 5% equilibrium melt of subducted continental margin sediment closely resemble Thetford Mines Ophiolite mantle, whereas the model liquids are too enriched in LILE, Th, U, Nb, Ta, and LREE compared to the lavas. One possible explanation for this conundrum is that melts generated at depth reacted with mantle rocks on their way to the surface, enriching the mantle and depleting the lavas in these elements. Despite an overall excellent fit to most data, some anomalies remain unexplained, for example, model residues show positive Ti peaks, whereas Thetford Mines Ophiolite mantle rocks do not.