← Back to Sessions

Parameter Optimization of Automatic Phase Detection and Picking Algorithms - Application in Sao Paulo University Seismological Center and Colombian National Seismic Network

Session: Network Seismology: Keeping the Network Running While Integrating New Technologies II

Type: Oral

Date: 4/22/2021

Presentation Time: 02:45 PM Pacific

Description: 

Two methodologies, Grid-search, and Bayesian algorithm were used to optimize the automatic detection and phase picking parameters in SeisComP. These methodologies were tested using a set of stations selected from two seismological networks Sao Paulo University in Brazil (IAG/USP) and Colombia’s National Seismological Network (RSNC).

After comparing manual and automatic locations, we found numerous missing events and others with low-quality locations in automatic databases. We selected 508 manual events from 2017/07/01 to 2020/07/31 in Brazil, and 532 manual events from 2019/02/01 to 2019/02/15 near Bucaramanga Nest in Colombia, as training sets for the optimization process.

A code was implemented to use grid-search as an optimization methodology; this code does an iterative process that generates automatic picks modifying the parameters. On the other hand, the Optuna package was used to implement the Bayesian algorithm as optimization methodology. Selected events were used as a training set, and an iterative process according to the Bayesian method was used. The results of both methodologies were compared.

We reproduced automatically 70% of manual picks increasing automatic locations from 33% to 62% of manual events for IAG/USP using Grid-search. In the RSNC the number of automatic picks increases by 81% and the number of automatic locations doubled using the Bayesian methodology.

Both methodologies showed excellent results. Grid-search allowed us to perform a complete analysis of the results examining the entire space of parameters. Grid-search, however, increases the computing time by adding parameters involved in the optimization process. On the other hand, the Bayesian algorithm can be implemented using several parameters without increased computing time.

Seismological centers could implement methodologies such as Grid-search or Bayesian algorithm to improve their automatic processing systems. Besides, the standardization of these methodologies would help to make easier their implementation.

Presenting Author: Camilo E. Muñoz Lopez

Student Presenter: Yes

Authors