TerraeMotus - Earthquake recognition and prediction using deep neural networks

Authors

• Mustafa Al Ibrahim; Jihoon Park; Noah Athens
• {malibrah, jhpark3, nathens}@stanford.edu

Overview

Earthquake seismology is a major topic relevant to understanding hazards due to natural and induced earthquakes as well as understanding physical properties of the earth's crust. In the past decade, the number of seismic monitoring stations has increased dramatically, leading the field of research to transition from an observation-based science to a data-driven science. This project demonstrates the efficacy of applying deep learning to earthquake recognition and prediction problems.

Two binary classification problems:

• Given a seismic waveform, has an earthquake occurred?

• Given a seismic waveform, will an earthquake occur?

Packages required

• Keras
• Sklearn
• Obspy
• Numpy
• Matplotlib