Seismo-volcano source localization with triaxial broad-band seismic array

• Autores: Inza Callupe, Lamberto Adolfo, Mars, J. I., Métaxian, J. P., O'Brien, G. S., Macedo Sánchez, Orlando Efraín
• Formato: artículo
• Fecha de Publicación: 2011
• Institución: Instituto Geofísico del Perú
• Repositorio: IGP-Institucional
• Lenguaje: inglés
• OAI Identifier: oai:repositorio.igp.gob.pe:20.500.12816/3034
• Enlace del recurso: http://hdl.handle.net/20.500.12816/3034; https://doi.org/10.1111/j.1365-246X.2011.05148.x
• Nivel de acceso: acceso abierto
• Materia: Time-series analysis; Fourier analysis; Spatial analysis; Volcano seismology; Explosive volcanism; Volcano monitoring; http://purl.org/pe-repo/ocde/ford#1.05.00; http://purl.org/pe-repo/ocde/ford#1.05.04; http://purl.org/pe-repo/ocde/ford#1.05.07

Seismo-volcano source localization is essential to improve our understanding of eruptive dynamics and of magmatic systems. The lack of clear seismic wave phases prohibits the use of classical location methods. Seismic antennas composed of one-component (1C) seismometers provide a good estimate of the backazimuth of the wavefield. The depth estimation, on the other hand, is difficult or impossible to determine. As in classical seismology, the use of three-component (3C) seismometers is now common in volcano studies. To determine the source location parameters (backazimuth and depth), we extend the 1C seismic antenna approach to 3Cs. This paper discusses a high-resolution location method using a 3C array survey (3C-MUSIC algorithm) with data from two seismic antennas installed on an andesitic volcano in Peru (Ubinas volcano). One of the main scientific questions related to the eruptive process of Ubinas volcano is the relationship between the magmatic explosions and long-period (LP) swarms. After introducing the 3C array theory, we evaluate the robustness of the location method on a full wavefield 3-D synthetic data set generated using a digital elevation model of Ubinas volcano and an homogeneous velocity model. Results show that the backazimuth determined using the 3C array has a smaller error than a 1C array. Only the 3C method allows the recovery of the source depths. Finally, we applied the 3C approach to two seismic events recorded in 2009. Crossing the estimated backazimuth and incidence angles, we find sources located 1000 ± 660 m and 3000 ± 730 m below the bottom of the active crater for the explosion and the LP event, respectively. Therefore, extending 1C arrays to 3C arrays in volcano monitoring allows a more accurate determination of the source epicentre and now an estimate for the depth.