5:15 PM - 6:45 PM
[SCG40-P22] A possible new criterion to distinguish small-sized ordinary earthquakes and tremors in shallow subduction zone, based on waveform similarity
Keywords:Tremor, Nankai trough, slow earthquake
At least two types of seismic events occur in shallow subduction zones: ordinary earthquakes and tectonic tremors. They are distinguished by their waveforms and spectral shapes. Compared to ordinary earthquakes of similar magnitude, tremor signals are known to exhibit a long duration, unclear onset, and depletion of seismic energy above 10 Hz. The classification is relatively simple for large-sized seismic events, where the signals are observed at many stations and tremors show clearly long-duration signals. However, it becomes difficult for small-sized events, although such a classification should hold the key to the fundamental question in seismology: what causes the difference between ordinary earthquakes and tectonic tremors?
Recently, Toh et al. (2023) showed, through observations and modeling, that the local structure around the seismic source could be responsible for characterizing tremor waveforms. The hypocenter of tremors may be located in a strongly scattering medium. Furthermore, they demonstrated through synthetic waveform calculations that the waveforms change drastically with a slight (less than one-tenth of the seismic wavelength) shift in the source location when it is located within such a scattering medium. The synthetics suggest that, compared to ordinary earthquakes, tremors of similar waveforms would be detected less frequently. This characteristic may be useful for distinguishing between the two types of seismic events, including those of small size.
In this study, we examine whether the above-mentioned characteristics of tremors are valid in observations. In other words, we examine whether the two types of events could be distinguished by the frequency of finding other events with similar waveforms. Here, we analyzed a year-long vertical component seismogram of station KMB06 of DONET1 (Kaneda et al., 2015). As waveform templates, we arbitrarily selected 7 ordinary earthquakes and 9 tremors based on some previous publications (e.g., Toh et al., 2018). Then, by applying a matched filter technique (Yamaguchi et al., 2019), we investigated the frequency of detecting other events with similar waveforms per template. Overall, the results suggest that the frequency could become a new criterion for distinction. For example, we found more than 5 events with similar waveforms (cc > 0.7) for 4 out of 7 ordinary earthquakes, whereas no such event was found among the 9 tremors. Although 3 out of 7 ordinary earthquakes had either none or fewer than 2 events with similar waveforms in the year-long data, we might detect them by applying the matched filter technique to the data for a longer term. In the presentation, we will further discuss the potential of using waveform similarity as a new criterion for distinguishing between small-sized ordinary earthquakes and tremors.
Recently, Toh et al. (2023) showed, through observations and modeling, that the local structure around the seismic source could be responsible for characterizing tremor waveforms. The hypocenter of tremors may be located in a strongly scattering medium. Furthermore, they demonstrated through synthetic waveform calculations that the waveforms change drastically with a slight (less than one-tenth of the seismic wavelength) shift in the source location when it is located within such a scattering medium. The synthetics suggest that, compared to ordinary earthquakes, tremors of similar waveforms would be detected less frequently. This characteristic may be useful for distinguishing between the two types of seismic events, including those of small size.
In this study, we examine whether the above-mentioned characteristics of tremors are valid in observations. In other words, we examine whether the two types of events could be distinguished by the frequency of finding other events with similar waveforms. Here, we analyzed a year-long vertical component seismogram of station KMB06 of DONET1 (Kaneda et al., 2015). As waveform templates, we arbitrarily selected 7 ordinary earthquakes and 9 tremors based on some previous publications (e.g., Toh et al., 2018). Then, by applying a matched filter technique (Yamaguchi et al., 2019), we investigated the frequency of detecting other events with similar waveforms per template. Overall, the results suggest that the frequency could become a new criterion for distinction. For example, we found more than 5 events with similar waveforms (cc > 0.7) for 4 out of 7 ordinary earthquakes, whereas no such event was found among the 9 tremors. Although 3 out of 7 ordinary earthquakes had either none or fewer than 2 events with similar waveforms in the year-long data, we might detect them by applying the matched filter technique to the data for a longer term. In the presentation, we will further discuss the potential of using waveform similarity as a new criterion for distinguishing between small-sized ordinary earthquakes and tremors.