High-resolution Fe–Ti oxide thermometry applied to single-clast pumices from the Bishop Tuff: a re-examination of compositional variations in phenocryst phases with temperature

Abstract

Numerous analyses of multiple ilmenite, titanomagnetite, sanidine, plagioclase, biotite, orthopyroxene, and clinopyroxene crystals (when present) in each of 24 single-clast pumices from the Bishop Tuff (BT), CA are presented, together with whole-rock compositions on the same single clasts. Data from multiple Fe–Ti oxide grains in each sample, coupled with an assessment of analytical uncertainty, show that there is a single population of ilmenite and titanomagnetite in each of 20 samples and no more than two populations of ilmenite and/or titanomagnetite in the remaining four. A statistical evaluation of microprobe uncertainty (for beam conditions of 15 kV and 10 nA) constrains the minimum number of ilmenite (≥ 21) and titanomagnetite (≥ 56) analyses required to obtain Fe–Ti two-oxide temperatures with propagated uncertainties ≤ ± 10 °C. Thermometry results confirm a continuous range (~ 700–800 °C) across the BT deposit. Plagioclase displays a bimodal compositional pattern, where a predominantly sodic (An14–15) vs. calcic (An22–23) population is found in single clasts where pyroxene is absent vs. present; the latter coincides with geochemical evidence for mixing between high-SiO2 rhyolite and less differentiated melts as previously documented in the literature. When sanidine rims are paired with plagioclase, calculated two-feldspar temperatures (~ 690–780 °C) are strongly correlated (R2 = 0.94) with Fe–Ti two-oxide temperatures. Application of plagioclase–liquid hygrometry reveals a systematic decrease in melt water content (~ 7.2–4.1 wt%) with increasing temperature across the BT deposit, consistent with experiments from the literature that show cotectic crystallization of quartz–sanidine–plagioclase in rhyolite melt requires a concomitant decrease of ~ 3 wt% H2O between 700 and 800 °C at 200 MPa. Single-clast, whole-rock concentrations of some elements (e.g., Rb) correlate strongly with Fe–Ti oxide temperature (R2 = 0.95). The Mg# in biotite also correlates strongly with temperature (R2 = 0.97), as do the Mg# values in orthopyroxene and clinopyroxene with whole-rock Mg# (R2 = 0.6 and 0.8, respectively) and melt water content (R2 = 0.7 for both). Evidence points to biotite crystallization in all BT samples from high-SiO2 rhyolite melt, which was already continuously zoned in composition and temperature prior to mixing with less differentiated melts, as previously noted in the literature. In contrast, several lines of evidence, including hydrous phase-equilibrium experiments on rhyolites from the literature, show that both pyroxenes in BT pumices co-crystallized with calcic plagioclase (~ An21–29), in equilibrium with both titanomagnetite and ilmenite, after mixing between less differentiated melt(s) and relatively hot (≥ 755 °C) high-SiO2 rhyolite.