Evidence of phenocryst growth during ascent in Yellowstone and Snake River Plain basalts: application of olivine-melt thermometry/hygrometry at the liquidus

Abstract

Geochemical and geophysical evidence from the literature shows that the Quaternary basalts erupted along the Snake River Plain (SRP) and adjacent to Yellowstone (YS) National Park have a mantle source that includes both the Yellowstone plume and subduction-modified lithosphere. Further evidence of the latter in the literature is seen in direct analyses of 1–3 wt% H₂O in olivine-hosted melt inclusions in YS/SRP basalts. With independent constraints on H₂O contents known, the goal of this study is to test application of a new olivine-melt thermometer and hygrometer (which is based on the simultaneous utilization of both D_Mg^oliv/liq and D_Ni^oliv/liq) to a suite of 17 SRP/YS basalts to evaluate if similar water contents are obtained. Previous studies show that accurate results are best obtained when the thermometer/hygrometer is applied at the olivine liquidus, thus providing temperature and water content in the melt at the onset of olivine crystallization. However, this requires that the most Mg-rich olivine analyzed in each sample passes equilibrium tests (based on Mg–Fe²⁺ and Mg–Mn exchange coefficients) when paired with a liquid composition that matches that for the whole rock. Sixteen of 17 YS/SRP samples from this study (94%) pass olivine-liquidus equilibrium tests, which shows that each erupted sample was initially crystal-free prior to growth of its observed phenocrysts. This observation, combined with textural evidence of rapid olivine growth, supports the hypothesis of phenocryst growth in the YS/SRP basalts during ascent. Application of the thermometer/hygrometer at the liquidus of each sample leads to temperatures that range from 1204 to 1126 °C and an average dissolved water content of 1.5 (± 0.8) wt%, which broadly matches H₂O analyses from the literature. A comparison of SRP/YS thermometry/hygrometry results from this study with those obtained by the same methodology applied to 45 basalts (all of which pass olivine-liquidus equilibrium tests) from four other volcanic fields (from diverse tectonic settings) lead to water contents that match those analyzed directly in olivine-hosted melt inclusions from the literature. Moreover, a compilation of all results shows a strong correlation between H₂O content and lnD_Ca^oliv/liq, consistent with evidence from the literature that the latter is strongly sensitive to dissolved water in the melt. One of the most surprising findings of this study is how many (16 of 17) of the SRP/YS basalts pass the olivine-liquidus equilibrium tests, given that the basaltic suite spans a range of composition (9–6 wt% MgO), reflecting the effects of crystal fractionation. This result raises the question of how efficient crystal fractionation occurred, leading to crystal-free melts prior to growth of observed phenocrysts. Here, it is proposed that fractionation may take place during ascent through sidewall crystallization along dike margins, and that this may be a relatively common mechanism by which basalts differentiate from high- to low-MgO compositions.