Automated Earthquake Detection and Local Travel Time Tomography in the South‐Central Andes (32–35°S): Implications for Regional Tectonics. Issue 4 (10th April 2022)

• Record Type: Journal Article
• Authors: Ammirati, Jean‐Baptiste; Villaseñor, Antonio; Chevrot, Sébastien; Easton, Gabriel; Lehujeur, Maximilien; Ruiz, Sergio; Flores, María Constanza
• Is Part Of: Journal of geophysical research. Volume 127:Issue 4(2022)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2022-04-10
• Subjects: subduction -- crust -- Chile -- andes -- deep‐learning -- earthquake; Geomagnetism -- Periodicals; Geochemistry -- Periodicals; Geophysics -- Periodicals; Earth sciences -- Periodicals; 551.1
• DOI: 10.1029/2022JB024097
• Languages: English
• ISSNs: 2169-9313

Abstract

Abstract: In the South‐Central Andes, the crustal structures driving the tectonic evolution of the Andean Cordillera remain unresolved. So far, most seismological studies focused on the subduction interface, leaving crustal seismicity and its relationship with crustal deformation and Andean volcanism mostly unconstrained. However, because of their large number compared to higher magnitude events, the characterization of small‐magnitude crustal earthquakes is key to identify active structures and better constrain the tectonic models. In this work, we exploit 53 months of continuously recorded, three‐component waveforms from the permanent seismic network in central Chile using a deep‐learning approach to improve the detection of small‐magnitude earthquakes. To increase station coverage, we also use the seismic phases obtained from a previous temporary seismic deployment. We use the obtained seismicity catalog to refine tomographic models of that region, revealing a more detailed architecture of the Chilean forearc. Travel times calculated in the new 3‐D velocity model allowed us to locate ∼14, 000 earthquakes. Refined double‐difference relocations of ∼4, 900 events located beneath the West Andean Thrust suggest a large‐scale, west‐dipping structure which, together with the west‐verging tectonic front, likely contributed to the uplift and crustal deformation during the past ∼20 Myr. Plain Language Summary: The occurrence of earthquakes is closely related to plate tectonics and in particular, the creation of mountain ranges and their evolution over very long periods of time. In order to better understand how these processes work in the Andes of Central Chile, we analyze more than 4 years of earthquake records using a machine learning algorithm. This technique allows us to automatically and reliably detect very small earthquakes, that are, in general, missing in the seismicity catalogs. Since they are also more numerous, we can use them to improve the geophysical images and reveal more details of crustal structures. We find that most of the superficial earthquakes beneath the Andean Cordillera are aligned with a major structure, dipping toward the west. This structure accommodates processes of shortening and uplifting directly related to the construction of the Andes. Key Points: A large number of earthquakes detected using a deep‐learning approach are used for local earthquake tomography The obtained 3‐D seismic wave velocity structure is used to relocate the seismicity catalog Earthquake locations suggest a west‐dipping major structure that we interpret to have contributed to the uplift of the South‐Central Andes