← Back to Sessions

A Cross-Correlation Algorthim for Detecting Small Earthquakes in Frequency Domain

Session: Waveform Cross-Correlation-Based Methods in Observational Seismology [Poster]

Type: Poster

Date: 4/21/2021

Presentation Time: 03:45 PM Pacific

Description: 

We present a cross correlation algorithm in the frequency domain to detect seismic waveforms. A known seismic event (master event) is correlated with a one-day recording. This longer time series is split into smaller time windows. The master event is shifted along each window and cross correlation is computed for each step. Cross correlation is computed in the frequency domain using the cross correlation theorem under a Cauchy-Schwarz normalization. Possible candidates are selected defining a threshold (0.8 in our case) for further analysis. Transforming the selected candidates back to the time domain, we obtain a seismic signal. We link our waveforms to an automatic P-phase picker, location algorithm, double-difference relocation code, and a coda magnitude calculator. We used 293 master events from a previous catalog in western Mexico. The dataset analyzed was a temporal Passcal experiment (CODEX) that was in operation from January 2006 to December 2007. It consisted of 22 broadband stations located around the Colima volcanic complex. Only the vertical component was taking into account. We could identify 1003 total events ranging in coda magnitude from 1.2 to 3.8 Mc. Our algorithm, as compared to free software codes; e.g. python (Obspy, Numpy, Scipy, matplotlib, pandas), has shown be more time efficient locating small magnitude events (down to ~1.2 Mc).Our code is written in free software and soon will be publicly available.

Presenting Author: Carlos R. P. Ramirez Pina

Student Presenter: Yes

Authors