Catastrophic Caldera-Forming (CCF) Monotonous Silicic Magma Reservoirs: Geochemical and Petrological Constraints on Heterogeneity, Magma Dynamics, and Eruption Dynamics of the 3.49Ma Tara Supereruption, Guacha II Caldera, SW Bolivia.

The 5.65 to 1.80Ma Cerro Guacha Caldera Complex (CGCC) in the Altiplano-Puna Volcanic Complex of SW Bolivia, with >90% of its >2500km3 erupted volume consisting of crystal-rich dacite, has all the characteristics of a 'monotonous' magma system. However, it also records minor lithological heterogeneity. Such hand-sample scale heterogeneity is ubiquitous in dominantly 'monotonous' magmas, yet remains poorly investigated. Here we explore the heterogeneity in the CGCC, and its implications for the construction and evolution of 'monotonous' magma systems. We focus on the Guacha II Caldera (G2C), the younger of two calderas in the complex, because its pre- to post-climactic eruptive history is fully represented and, although the eruptive products are dominantly dacitic (66-72 wt % SiO2), the juvenile pyroclastic deposits and lavas erupted throughout the history of the G2C define a high-K, calc-alkaline suite of diverse compositions that range from andesite to high-Si rhyolite. The G2C cycle initiated with the effusive eruption of crystal-rich andesite lava. The subsequent explosive phase began with a short-lived plinian eruption of crystalpoor rhyolite pumice. This was immediately followed by the Catastrophic Caldera Forming (CCF) eruption at 3.49 ± 0.01Ma and the deposition of dacite-rhyolite and banded pumice within the >800km3 dense rock equivalent (DRE) Tara ignimbrite. A significant volume of magma remained and caused ~1.5 km of resurgent uplift. Three crystal-rich dacite-rhyolite post-climactic lava domes (Chajnantor Dome, Rio Guacha Dome, and Chajnantor Lavas) subsequently erupted from separate coexisting melt-rich 'pods' within the G2C's remnant mush. Whole-rock isotope ratios across all lithologies span a significant range in 87Sr/86Sr (0.709380-0.713159) and a relatively narrow range in 143Nd/144Nd (0.512179-0.512297) and δ18O(qtz) (+8.38 to+8.68%), best reconciled with a twostage assimilation-fractional crystallization (AFC) model. Stage 1 initiated with parental melts from the Altiplano-Puna Magma Body (APMB) fractionating and assimilating crustal lithologies in the upper crust (10-25km depth) to generate the magma compositions recorded in the andesite lava and the ignimbrite banded pumice. These magmas subsequently accumulated and underwent a second stage of AFC in the uppermost crust (~800-850 °C and 5-9km depth) to produce the most differentiated magmas recorded in the ignimbrite rhyolite pumices. Although the two-stage AFC model presented here is non-unique, it implies that the basement composition is temporally or spatially variable throughout the ~30km of upper crust beneath the G2C. The origin of some of the minor lithologies erupted from the G2C requires further explanation beyond AFC. Recharge, binary mixing, and additional crystal fractionation of the plagioclase, quartz, sanidine, biotite, and Fe-Ti oxides identified within the ignimbrite dacite-rhyolite pumice are required to explain the geochemical diversity of these minor lithologies. Integrating geochemical modeling with the volcanological framework we find that the andesite magma played a large role in the development and evolution of the G2C magma system. At least one episode of andesite magma recharge occurred in the upper crustal magma reservoir prior to the eruption of the G2C and fractionation of this andesite magma formed residual rhyolitic melt sampled by the plinian eruption. Andesitic recharge therefore not only sustained the pre-eruption magma reservoir's dominant dacitic composition throughout its history, but also may have acted as an eruption trigger. The minor lithologies erupted from a 'monotonous' magmatic system record processes that are otherwise lost in the inexorable march to homogeneity that characterizes the long-term evolution of many CCF magma systems. Our findings complement investigations that have identified heterogeneity at the crystal- and micro-scale, but emphasize that upper crustal processes are recorded as hand-sample scale heterogeneity that is serendipitously sampled by eruption before being integrated into a 'monotonous' whole. [ABSTRACT FROM AUTHOR]