The Laguna del Maule volcanic field, straddling the Chile-Argentina border at 36° S, is currently the subject of a multi-disciplinary collaborative investigation supported primarily by the U.S. National Science Foundation Integrated Earth Systems program and the Observatorio Volcanológico de Los Andes del Sur (OVDAS) of SERNAGEOMIN. At least 50 post-glacial (younger than 20 ka) eruptions from more than two dozen vents encircling the 25x17 km lake basin have produced rhyodacitic-to-rhyolitic lava flows and ash deposits totaling > 30 km3, suggesting that a large, active, silicic magma reservoir fuels this system. Since 2007, GPS and InSAR geodesy reveal that Laguna del Maule has been experiencing rapid uplift at 20 to 25 cm/year centered within the ring of silicic vents. Moreover, a deformed paleo-shoreline that has been 36Cl-dated implies magma-driven surface uplift of > 60 m and that growth of this large shallow reservoir has occurred over at least the past 9,400 years. UW-Madison, Cornell, and OVDAS have deployed a seismic array covering ~450 km2 that surrounds the lake basin. The array consisted of 18 broadband stations in 2015, and was enlarged to 47 stations in 2016 (37 broadband, 10 short-period). The full array will remain in place until late March 2018. A variety of seismic studies are planned for the seismic array data, including body-wave tomography, surface-wave tomography, attenuation tomography, teleseismic tomography, receiver function analysis, seismic interferometry, and focal mechanism and moment tensor determination. The main goals are to detect the magma chamber underlying Laguna del Maule, characterize its dimensions and properties, and assess the state of stress of the system.

I will report on results from surface-wave tomography, and compare them to the results from other geophysical techniques. Due to the small array aperture (~30 km) and the limited frequency range of usable ambient noise, we combined three types of data for the surface-wave tomography: standard noise correlation analysis using pairs of stations within the array, correlation of earthquake coda at pairs of array stations, and differential dispersion for ambient noise for pairs of array stations correlated with remote stations. Somewhat unexpectedly, the Vs image shows evidence of a strong upper crustal low velocity anomaly along the southwest side of the lake, which does not extend under the entire lake basin. Depending on how the boundaries of the low-Vs anomaly are defined, we obtain estimates of ~25 to ~400 km3, and melt percentages of 5% to 8%. The position of the anomalous body is very close to the estimated source area for the uplift measured by InSAR and GPS. A Bouguer gravity low, interpreted to reflect the magma reservoir, is observed in the same area. In contrast, magnetotelluric results image a large low-resistivity zone near the north side of the lake and much smaller anomalies elsewhere.

Recent fluid dynamic models of bubble migration and accumulation have shown that volatiles can provide a mechanism whereby basaltic magma recharge could catalyze the eruption of rhyolite without imparting a substantial thermal or physical signature on the erupted lavas. This leads to the idea that during the Holocene, rhyolitic magma batches have been repeatedly extracted from a broad spatial footprint beneath the lake basin, but the magma batches that have fed the individual eruptions were likely of relatively modest volume, comparable to the geophysical results.