Rejuvenation of the Fish Canyon magma body: A window into the evolution of large-volume silicic magma systems

Voluminous, unzoned, phenocryst-rich pyroclastic deposits, considered as erupted batholiths, provide a unique opportunity to investigate magmatic processes in silicic magmas. The Fish Canyon Tuff, a well-documented example of these monotonous ignimbrites, displays evidence for simultaneous dissolution of feldspars + quartz and crystallization of hydrous phases during gradual near-isobaric reheating from similar to720 to 760 degreesC. These observations, along with a high crystallinity (45%) and near-solidus mineral assemblage, suggest that the Fish Canyon magma cooled to a rigid crystal mush before being partly remelted prior to eruption. Rejuvenation was triggered by intrusion of water-rich mafic magmas at the base of the Fish Canyon mush, but the mechanisms of heat transfer remain poorly understood. The growth of amphibole during reheating requires addition of mafic components, but the absence of any measurable gradients and the paucity of matic enclaves in the Fish Canyon magma rule out a reheating event dominated by convective mixing with a mafic magma. Closed-system processes, such as heat conduction and convective self-mixing, could not account for the transport of externally derived matic components. We performed numerical simulations of upward percolation of a hot, low-density H2O-CO2 fluid phase (gas sparging) through a crystalline framework saturated with rhyolitic melt to assess the efficiency of such a process in rejuvenating silicic mushes in open systems. Sparging by similar to20-40 km(3) of gas extracted from similar to3000 km(3) of matic magma is capable of reheating 7500 km(3) of silicic crystal mush by >40 degreesC in 150-200 k.y. Moreover, the vertical thermal gradient after 150 k.y. in most of the mush is small (similar to25 degreesC in the upper 65%). Gas sparging also produces an increase in the internal pressure of silicic crystal mushes and may lead to the formation of crystal-poor rhyolites by expelling interstitial melt. However, our simulations predict that filter pressing driven by sparging of externally derived gas could not solely account for the generation of the most voluminous rhyolites.