From Echo to Epicenter: PhaseHunter's Detection & GENIE's Mapping of LA Basin's Hidden Depths

Artemii Novoselov, Asaf Inbal, & Gregory C. Beroza

Published September 10, 2023, SCEC Contribution #13210, 2023 SCEC Annual Meeting Poster #015

The Los Angeles Basin presents unique challenges for detecting and pinpointing deep seismic events. To tackle this, we turn to two powerful deep learning tools: PhaseHunter and GENIE. PhaseHunter offers an innovative approach to accurately detect seismic phase onset times, which are crucial for understanding when and where earthquakes begin. This method's advantage lies in its ability to not only detect but also provide a level of confidence in its readings, improving reliability. On the other hand, GENIE dives deeper into mapping the exact origins of these detections. Using a specialized network design called Graph Neural Networks, GENIE can connect seismic readings to their specific sources, even in areas as complex as the LA Basin. Together, PhaseHunter's detection and GENIE's precise mapping aim to provide a clearer picture of the seismic activity in the LA Basin's depths.

Key Words
LA Basin, PhaseHunter, Machine Learning, Genie

Citation
Novoselov, A., Inbal, A., & Beroza, G. C. (2023, 09). From Echo to Epicenter: PhaseHunter's Detection & GENIE's Mapping of LA Basin's Hidden Depths. Poster Presentation at 2023 SCEC Annual Meeting.

Related Projects & Working Groups
Seismology