OKSP 2.0: Determinación de los parámetros de fuente sísmica para la comprensión de la mecánica de ruptura a lo largo de segmentos de falla en Costa Rica utilizando redes neuronales

Both the development of territorial planning plans and the robustness of the seismic construction code are fundamentally dependent on the local and regional seismotectonic characteristics (such as the type of fault, geology, geometry, extent, recurrence periods, etc.) of the country.

Understanding the seismic potential and the style of rupture along a fault system allows for better risk assessments in the event of earthquakes. In the Central Valley, for example, faults like the Aguacaliente, Cipreses, and Río Azul faults cross the most densely populated cantons of the Province of San José and Cartago, and little is known about their kinematics. A first-order estimate, based on their extension and their width at depth, suggests the possibility of events with a seismic moment magnitude greater than Mw = 5.5 at a shallow depth, but their exact nature is not known, nor is the period of recurrence or their physics To the south of the central canton in the Province of Cartago, the Navarro fault is responsible for generating important historical earthquakes, such as the Cartago Earthquake, with magnitude Mw = 6.1, which occurred on May 4, 1910, cataloged as the most destructive tremor in the history of Costa Rica, followed by the Cinchona earthquake, Mw = 6.2, on January 8, 2009, and the Valle de la Estrella earthquake, Mw = 7.7, on April 12, 1992. In order to anticipate large earthquakes like these, it is necessary to study in excellent detail the interseismic period of the faults that caused them, understanding the magnitude distribution and the frequency of events, as well as the style of faulting (focal or kinematic mechanism) and consequently the stress regime, friction coefficients and possible rotations of the stress field in time associated with the preparation of the fault before rupture. This research proposal aims to take advantage of the CENAT computational infrastructure to expand the use of the OKSP tool, developed for the identification and location of tremors through the use of seismological stations operated by OVSICORI-UNA, for the determination of the magnitude of the seismic moment and the calculation of focal mechanisms using neural networks. Currently, the calculation of focal mechanisms is based on two main methods: 1) that consider the determination of the polarity of the P-wave phase and 2) the calculation of the seismic moment tensor. So far, both cases are manual, and an automatic application based on Artificial Intelligence has not been developed that allows their estimation. Therefore, this proposal aims to take the field forward, developing an application that allows automating the calculation of focal mechanisms, reducing bias due to human intervention, improving seismotectonic understanding at different time and spatial scales, and generating a direct positive impact on seismological observatories worldwide.

1) Teoría y Metodologías en Computación
2) Aplicación de la computación en distintos dominios científicos, tecnológicos,
organizacionales y sociales