Automated first-motion focal mechanisms and local stress inversion at Laguna del Maule, Chile

Abstract:

The Laguna del Maule volcanic field located in the Chilean southern Andes is a large, actively deforming silicic magma system. The system has erupted over 40 km³ of rhyolite over the past 20 ka but has not erupted historically. Rapid surface inflation in the area drew attention to the system in 2008, and the inflation has continued at rates greater than 25 cm per year since. The Observatorio Volcanológico de los Andes del Sur (OVDAS) installed 6 seismic stations to monitor the area in 2011, and this network was augmented from January 2015 – April 2018 with an additional 38 temporary seismic stations. Processing large amounts of pre-recorded, three component seismic data from un-telemetered stations is a challenge, especially over long time periods or large networks. Automatic event picking offers a starting point, and under correct calibration can offer benefits when compared to strictly manual picking by a technician. We use the REST package by Steve Roecker to generate automatic P-wave (39,760) and S-wave (32,240) arrival picks. REST has the added benefit of determining polarities and event magnitudes for each event. We take these arrival picks and polarities, relocate the earthquakes in a 3D model from body-wave tomography, and use them as inputs for HASH (Hardebeck and Shearer, 2002) to produce focal mechanisms for 1757 local events automatically identified by REST. We compare the results from this “automated” method against a subset of manually analyzed events as a quality control measure and error assessment. With thousands of first-motion focal mechanisms we then invert for the local stress field. Though recent results of active faulting at LdM suggest that volcanism is controlled by this regional stress field, it has been poorly constrained locally because of a lack of reliable stress indicators.