Geochemical Modelling of Shallow Fractionation and Deep Mantle Melting Below Mauna Loa Volcano, Hawaii

Abstract

Hawaii is the archetypal example of intra-plate 'hotspot' volcanism, yet the mechanisms of plume formation, hotspot volcano formation, and the nature of chemical heterogeneity in hotspot lavas such as those at Hawaii remain in question. Of particular interest is understanding what physical and chemical processes lead to the distinctive 'double-track' geochemical trend, whereby the volcanoes on the south-west side of the Hawaiian volcanic chain ('Loa' track) vary systematically compared with those on the north-east side ('Kea' track) with respect to major elements, trace elements, and isotopic ratios. In order to make meaningful inferences about how the parallel distinction may arise, it is necessary to understand how the chemistry of magma evolves at Hawaii starting with the composition and melting of the mantle source, followed by emplacement of the magma into the crust with associated residence times and mixing, the progression of the crystallisation process as the magma cools, and finally the eventual eruption through the volcanic edifice. Here, major element, trace element, and Sr and Nd isotopic analyses of tholeiites from submarine Mauna Loa volcano are used alongside compiled literature data of Mauna Loa and Kilauea Volcanoes to investigate both shallow magmatic processes in the crust and deep mantle melting in the upwelling plume. Evidence from modal olivine, olivine forsterite content, glass compositions, and whole rock Ni suggest that none of the whole rock samples are direct parental magmas, but instead reflect an evolved magma in which olivine has remained entrained or become subsequently entrained to varying degrees. Computational geochemical modelling suggests the occurrence of two distinct fractionation sequences: ol + sp + opx + cpx + plag and another with ol + sp + cpx, which could be achieved with minor variations in the water content (0 - 0.5 wt. %). Furthermore, a compilation of experimental melt studies and trace element models are presented, showing that the source of Hawaii's lava is most probably a garnet-bearing peridotite at ~4 GPa with a major element composition similar to that of primitive mantle or Hawaiian pyrolite, contradicting some current hypotheses which argue for a pyroxenitic source. Trace element modelling reveals a the differences between Mauna Loa and Kilauea can broadly be explained by variable degrees of melting of a fertile, garnet peridotite, with Mauna Loa resulting from ~12% melting and Kilauea resulting from ~6%. It was shown that low-olivine pyroxenites cannot produce the required high Ni content of Mauna Loa lavas, and that the Ni may be better explained by the high temperatures and high MgO contents associated with the Hawaiian plume which cause an associated decrease in the bulk DNi. This negates the need for a low-olivine pyroxenite as the source. Despite this, a component with high radiogenic lead and 143Nd/144Nd with low 87Sr/86Sr is required in the Kilauea source, while the Mauna Loa source lies isotopically between Bulk Silicate Earth and Depleted MORB mantle.