Eruption sequence of the Otowi Member, Bandelier Tuff Formation, Valles Caldera, Jemez Mountains, New Mexico: lessons from the deposits

Models for caldera collapse are outstripping evidence to support or refute them based on detailed field studies. Continuing study of clast lithology is changing previously suggested ideas for the eruptive sequence of the Otowi Member (1.6 Ma) of the Bandelier Tuff. The eruption produced Plinian and co-ignimbrite fall deposits, outflow and intra-caldera ignimbrite (totaling a maximum of ~ 430 km3 DRE), and heralded the first major collapse of Valles caldera (24 km diameter). Previous work supports magma withdrawal from a high-silica-rhyolite chamber with systematic internal variations in incompatible trace element concentrations. More recent study, including determination of lithic-clast provenance, reveals new insights into the eruption sequence: 1) The first phase produced Guaje fall deposit unit A, after which column collapse produced coeval pyroclastic flows and fall deposits (units B-F) in a complex depositional pattern. In places the ignimbrite is intra-Plinian and in other places only fall units accumulated while pyroclastic flows by-passed the area. Pumice composition data suggests that fallout from vent-derived Plinian columns persisted late into the eruption. 2) Lithic provenance and abundance information from the ignimbrite combined with pumice composition data lead us to reject our previous idea of switching from an initial central vent to vents on caldera ring-fractures, at the time of deposition of co-ignimbrite lag-breccias. Observations are better explained by the central vent persisting longer into the eruption, and thus lag-breccia deposition was not proximal to the central vent. 3) 50 km from vent a preserved Otowi sequence includes a thin ignimbrite veneer, showing that at least some of the late-erupted pyroclastic flows were widespread and energetic. 4) Volumetrically, at least 2/3 of the erupted part of the chamber had been expelled before caldera collapse and possible vent-shift took place. These observations need to be included in a model for development of the first Valles caldera, which is hosted in ancient granitoid terrain covered by a thin sedimentary/volcanic sequence on the western flank of the Rio Grande Rift.