Correlation Methods to Better Characterize Repeating Seismic Events
Abstract
The recent growth in continuous seismic data recording and improved access to long time records has provided new opportunities to take advantage of correlative signal processing algorithms. In cases where a prominent signal of interest is well known, such as a locally recorded small earthquake, a multi-station/-component template can be constructed and searched for through years of broadband data. However, this limits the search to relatively large sources that can produce signals across several seismic stations, and many of these events may not be repetitive in nature. To look for weaker sources that may be buried near the noise level, an auto-correlation technique can be designed to search for repetitive signals that can be enhanced through stacking of similar waveforms. For example, recent studies have found that tectonic tremor is primarily composed of a swarm of low frequency earthquakes, such that identifying individual low frequency earthquakes that repeat throughout the sequence can provide opportunities to improve source characterizations. We develop an auto-correlation technique capable of finding the most repetitive signals in multi-hour length waveforms, which we then apply to times of known seismic activity to identify individual repeating events. All of these events are grouped into families, and the stacked waveforms of these families are used as template waveforms that are cross-correlated over several years of recording to better characterize the time history of seismic events. This study seeks to refine the source locations by stacking families of similar events to enhance the signal to noise ratio and clarify P and S wave arrivals. Accurate time histories and source locations of event families can be achieved due to multi-station template matching algorithms being particularly sensitive to waveform shapes that are due to even small differences in source location. Since multi-station template matching does not depend on station amplitudes as more traditional techniques do, it will help to avoid a detection bias and clarify time histories. This approach is being applied to southern Cascadia, where there is limited seismic source location analysis due to both lower seismic activity and lower network density.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2013
- Bibcode:
- 2013AGUFM.S41B2439S
- Keywords:
-
- 7240 SEISMOLOGY Subduction zones;
- 8170 TECTONOPHYSICS Subduction zone processes;
- 7215 SEISMOLOGY Earthquake source observations;
- 7290 SEISMOLOGY Computational seismology