← Back to Sessions

Machine Learning-based Seismogenic Zones for Seismic Hazard Estimation in Mexico

Description: 

We explore a new approach for defining seismic zones intended to be the base of probabilistic seismic hazard calculations in Mexico. In order to reduce uncertainty in the PSH assessment, we evaluate the commonly used seismic zones in Mexico against the country's tectonic activity. We conclude that several discrepancies are observed, and a new approach is proposed to apply a machine learning algorithm designed to build seismic zones aligned with tectonic activity and geologic characteristics. To achieve the latter, we compiled focal mechanism data from an extended catalog by Franco et al. (2020), gravimetric data, and other geological features. The clustering algorithm HDBSCAN was selected for its suitability in handling the data distribution and density, providing optimal results for the task. Furthermore, we analyze b-values along a territorial grid to characterize seismicity and modify, evaluate, and validate the proposed zones. The results reveal a distribution of seismic zones that reflect the tectonic activity in Mexico, which brings more certainty to the seismic hazard estimations and opens a new direction in the definition of seismic zones.

Session: New Directions in Environmental, Seismic Hazard and Mineral Resource Exploration Studies [Poster]

Type: Poster

Date: 4/17/2025

Presentation Time: 08:00 AM (local time)

Presenting Author: Lilibeth

Student Presenter: Yes

Invited Presentation: 

Poster Number: 108

Authors