← Back to Sessions

WITHDRAWN Automatic Seismic Monitoring Using Regional and Local Temporary Networks in Colombia

Description: 

WITHDRAWN Seismological networks, whether global, regional, or local, have the objective of monitoring seismic activity. This implies the detection of seismic events and determination of their location (latitude, longitude, depth and origin time) with an acceptable level of uncertainty. We apply these steps in three seismic networks to get an automatic catalog that shows the seismicity in the Colombian territory for 6 years (2016-2021). A regional seismic network (the Colombian national seismic network, with station separation of about 100 km), and two local and temporary networks (station separation 10-30 km approx.) in northern South America: the Middle Magdalena Valley Array, and the Caribbean-Mérida Andes seismic array.

To achieve this, continuous data of multiple stations needs to be processed to detect and pick seismic phases. We used the EQTransformer, a pre-trained Supervised Deep Learning model, to detect and pick seismic phases at each station. Then, we used GaMMA, an unsupervised machine learning method to associate phase detections originated by a common source. Then, we used the NonLinLoc algorithm to locate sources with a reliable hypocentral location. Finally, we estimated the local magnitude for each event. This process was fully automated, allowing it to be rolled out to other networks around the world.

The results show that this implementation is reliable enough to generate automatic seismic catalogs with the appropriate quality in terms of the event location errors, it is capable to define major tectonic structures and highlight regional crustal faults. Better yet, it can improve earthquake processing times and complement manual catalogs due to its good performance to detect small earthquakes and aftershocks.

Session: Opportunities and Challenges for Machine Learning Applications in Seismology [Poster]

Type: Poster

Date: 4/19/2023

Presentation Time: 08:00 AM (local time)

Presenting Author: German A. Prieto

Student Presenter: No

Invited Presentation: 

Authors