The Origin of the Geochemical Zonation with Temperature in Early Bishop Tuff: Evidence of Rapid Time Scales Between Melt Segregation and Eruption?

Abstract:

High-silica rhyolites are the most evolved magmas on Earth and constitute some of the largest eruptions. An example is the Bishop Tuff (BT), which displays two distinct types of geochemical zonation: one related to magma mixing in “Late” Bishop pumices and the other unrelated to magma mixing and best displayed in “Early” Bishop pumices. The focus here is the origin of the geochemical zonation in Early Bishop pumices, in which elements such as Mn, Rb, Cs, Y, Hf, U, Th, Tb, Yb, and Lu all decrease with T °C, whereas other elements such as Ti, La, Ce, Nd, Ba, and Sr increase with T °C. Fe-Ti oxide thermometry was applied to hundreds of equilibrium pairs of ilmenite and titanomagnetite per pumice clast, permitting an average temperature to be obtained. When element concentrations are plotted against average T °C, strong correlations are observed (e.g., R² ≤ 0.90), which can be used to determine the relative compatibility and incompatibility of all elements. Because Early BT contains nine minerals, segregation from a crystal-rich mush that contains these same phases is modeled and their relative abundances are constrained by matching calculated bulk partition coefficients to the relative compatibility and incompatibility of the various elements obtained from the pumice clasts. The data show that the geochemical gradient in Early BT is readily explained by extraction of interstitial melt at various melt fractions/source depletion (proxied by temperature) from a leucogranite (qtz+K-spar+plag) that contained ~3% biotite, ~0.9% titanomagnetite, ~0.4% ilmenite, ~0.08% allanite, ~0.04% apatite and ~0.015% zircon. Melt segregation from this assemblage also explains the systematic increase in ∆NNO (i.e., Fe³⁺/Fe^T) with temperature, because Fe³⁺ is more compatible than Fe²⁺ in this crystalline assemblage. If this explanation of the geochemical gradient in Early Bishop pumices is correct, a new question emerges. Given the short time scales (~2-3 weeks) of Fe-Ti oxide re-equilibration, how is it that the temperatures from the Fe-Ti oxides remain so strongly correlated to melt composition, if the latter is set by conditions of melt segregation? One possibility is that the time scale between segregation, crystal growth, and eruption is extremely rapid, on the scale of months or less.