Assessing magma storage beneath Mauna Loa, the world's largest active volcano, using combined barometric and microstructural constraints

The plumbing system beneath Mauna Loa, Hawai'i, has been understudied relative to its younger neighbor, Kīlauea. It is particularly interesting to ponder whether Mauna Loa’s larger size and greater maturity is reflected in its magma storage geometry. For example, prior work has suggested the presence of a deep (>18 km) storage system based on the presence of high Mg# (>84) Opx, from which magmas may ascend within days towards the surface. This would have very different implications for hazards and effective monitoring than at Kīlauea, where eruptions mobilize mostly from 1-5 km depth. To assess the possibility of deep storage, we examine crystal cargoes from picritic to aphanitic basalts erupted in 1852, 1855, 1868, 1949, 1950, 1984 and 2022 from the summit and rift zones, including harzburgitic xenoliths present in lavas erupted on the south flank. We use confocal Raman Spectroscopy to analyze fluid inclusions (small pockets of CO₂-rich fluid trapped within crystals), coupled with Energy Dispersive Spectroscopy to determine the chemistry of the host crystal. Fluid inclusion barometry provides a precise determination of the pressures at which fluids were trapped within crystals, and thus the depths at which crystals grew and/or stalled for prolonged periods. Analyses of ~300 fluid inclusions from seven effusive eruptions and ~80 from xenoliths show that magma is predominantly stored at 1-5.5 km depth, which is slightly deeper than similar measurements at Kīlauea. These results contrasts with past studies which have suggested that the high Mg# orthopyroxenes in the xenoliths formed at > 6 kbar based on pMELTS modelling of orthopyroxene stability. We show that different thermodynamic models (pMELTS, rhyoliteMELTS, MageMIN) disagree greatly on the depth of Opx stability, indicating that the high pressures of magma storage previously suggested for Mauna Loa are not required. The poikilitic textures within the xenoliths with rounded, resorbed olivine chadocrysts within large Opx oikocrysts suggest that the high Mg# Opxs within these samples formed from a reaction between a higher SiO₂ melt (~52 vs 49 wt%) and the olivine cumulate pile, at 3-5 km depth based on fluid and melt inclusion pressures from the olivine chadocrysts. The observed phase assemblage can be reproduced with melt-mush reaction models in rhyoliteMELTS. In summary, we suggest that the presence/absence of high Mg# Opx in shield Hawaiian volcanoes may instead reflect differences in bulk SiO₂ between the two chemical trends present in the Hawaiian hotspot (Loa and Kea trend), rather than drastic differences in the depth of magma storage.