Search for a command to run...
Silicic magma systems are of great scientific interest and societal importance owing to their role in the evolution of the crust and the hazards posed by volcanic eruptions. MELTS is a powerful and widely used tool to study the evolution of magmatic systems over a wide spectrum of compositions and conditions. However, the current calibration of MELTS fails to correctly predict the position of the quartz + feldspar saturation surface in temperature, pressure and composition space, making it unsuitable to study silicic systems. We create a modified calibration of MELTS optimized for silicic systems, dubbed rhyolite-MELTS, using early erupted Bishop pumice as a reference. Small adjustments to the calorimetrically determined enthalpy of formation of quartz and of the potassium end-member of alkali feldspar in the MELTS calibration lead to much improved predictions of the quartz + feldspar saturation surface as a function of pressure. Application of rhyolite-MELTS to the Highland Range Volcanic Sequence (Nevada), the Peach Spring Tuff (Arizona–Nevada–California), and the late-erupted Bishop Tuff (California), using compositions that vary from trachydacite to high-silica rhyolite, shows that the calibration is appropriate for a variety of fluid-bearing silicic systems. Some key observations include the following. (1) The simulated evolutionary paths are consistent with petrographic observations and glass compositions; further work is needed to compare predicted and observed mineral compositions. (2) The nearly invariant nature of silicic magmas is well captured by rhyolite-MELTS; unusual behavior is observed after extensive pseudo-invariant crystallization, suggesting that the new calibration works best for relatively small (i.e. <50 wt %) crystallization intervals, comparable with what is observed in volcanic rocks. (3) Our success with rhyolite-MELTS shows that water-bearing systems in which hydrous phases do not play a critical role can be appropriately handled; simulations are sensitive to initial water concentration, and although only a pure-H2O fluid is modeled, suitable amounts of water can be added or subtracted to mimic the effect of CO2 in fluid solubility. Our continuing work on natural systems shows that rhyolite-MELTS is very useful in constraining crystallization conditions, and is particularly well suited to explore the eruptive potential of silicic magmas. We show that constraints placed by rhyolite-MELTS simulations using late-erupted Bishop Tuff whole-rock and melt inclusion compositions are inconsistent with a vertically stratified magma body.