Geology of the Vilama caldera: A new interpretation of a large-scale explosive event in the Central Andean plateau during the Upper Miocene

The Vilama caldera is one of the very large-volume volcanic structures that formed during the ignimbrite flare-up that lasted from 10 Ma to 4 Ma on the central Andean plateau. Formerly, the Vilama structure was interpreted as a 65 × 40 km wide caldera with a two stage evolution. New field correlations, petrographical, geochemical and geochronological data lead to a substantially larger reinterpreted Vilama ignimbrite, whose outcrops cover more than 4000 km ². New and existing K/Ar and Ar/Ar dating shows that the dacitic Vilama ignimbrite erupted from the Vilama caldera at 8.4-8.5 Ma. The ignimbrite can be divided into extracaldera outflow and intracaldera deposits. The outflow has a mean thickness of ∼ 40 m and is separated into a restricted valley-ponded lower cooling unit, and a laterally extensive low aspect ratio upper cooling unit. The Vilama caldera, which is roughly rectangular (35-40 km × 15-18 km) in shape and has central coordinates of 22°24' S and 66°57' W, is considered to have formed in a single-stage collapse event. As the topographic rim is only seen on the western side, the extent of collapse and geometry of the caldera are incompletely known and inferred from indirect data. Possible collapse geometries include a slightly asymmetric single-block subsidence and non-chaotic multiple-block collapse. Estimated erupted volumes range from ∼ 1800 to 1200 km ³, or 1400 to 1000 km ³ in dense rock equivalents. The properties of the Vilama ignimbrite which include a crystal-rich and pumice-poor nature, a high degree of welding and induration and a prodigious volume, suggest that an external drive, rather than volatile overpressures, controlled and maintained the eruption. The best candidate is caldera subsidence triggered by the instability of a magma chamber roof above a batholith-scale magma body. Transtensive/distensive tectonic stresses resulted in the northwest elongation of the magma chamber and/or influenced the roughly rectangular/subelliptical shape of the subsided block. As most of the Vilama ignimbrite is within the caldera, subsidence must have started early in the eruptive history. Immediately before, or concomitantly with the onset of subsidence, the lower extracaldera unit was deposited from flows that formed during collapse of well-developed plumes with efficient convective phases. Once caldera collapse was completely established, the eruption dynamic changed dramatically, and the extracaldera and intracaldera facies erupted. During this phase, unstable and low plumes (boiling-over collapse fountains) preferentially collapsed towards the interior of the subsiding structure, causing the ignimbrite volume to be concentrated in the caldera. After much less than 1 Ma, resurgence, which might be linked to magma chamber recharge, domed the intracaldera facies and caused post-collapse volcanism to be channeled through subsidence and/or resurgence-related faults.