1:45 PM - 2:00 PM
[SCG45-25] Machine Learning-Based Detection and Localization of Tectonic Tremors in the Japan Trench
★Invited Papers
Keywords:Tectonic tremor, Machine learning, S-net, Japan Trench
Shallow tectonic tremors near trenches have been detected due to the advancement of offshore observation networks. Traditionally, tremors were identified by cross-correlating envelope waveforms between seismic stations (Obara, 2002). However, this method has struggled to differentiate tremor signals from earthquakes and sometimes missed tremors during active tremor episodes (Annoura et al., 2016). Addressing these challenges is crucial for monitoring tremors in seismically active regions, such as the Japan Trench. We developed a machine learning-based tremor monitoring system using a dense network of cable-type ocean-bottom seismometers (S-net) in the Japan Trench.
The system follows a three-step procedure. First, a CNN model, Discriminator for Earthquake and Tremor (DiET), classified input spectrograms into earthquake, tremor, or noise. Second, spectral clustering (von Luxburg, 2007), a graph theory-based technique, was used to associate stations where tremors were detected (Graph-based Associator with Signal Probability: GrASP). Finally, tremors were located every minute using a hybrid method combining envelope cross-correlation and seismic energy ratio.
The system analyzed continuous waveforms recorded from August 2016 to August 2024. Our analysis detected seven times more tremors than the previous study (Nishikawa et al., 2023) using envelope cross-correlation. The newly identified tremors expanded the known spatial distribution of tremor activity, both along the strike and dip, revealing a complementary spatial relationship between tremors and earthquakes. Additionally, our catalog improved temporal resolution, uncovering spatiotemporal patterns of tremors synchronized with slow slip events (Nishimura, 2021). Seismic energy rates of tremors were calculated, showing spatial variations along the strike and dip, with higher rates near asperities of large earthquakes. A positive correlation between energy rates and recurrence intervals was found, suggesting that these spatial variations reflect frictional heterogeneities on plate boundaries. The enhanced spatiotemporal resolution of our tremor catalog provides valuable insights into the relationship between slow and fast earthquakes.
Acknowledgements: We used S-net velocity waveform data (National Research Institute for Earth Science and Disaster Resilience, 2019). This research was supported by JSPS KAKENHI Grant Number JP21H05205 and JP21H05203 in Grant-in-Aid for Transformative Research Areas (A) “Science of Slow-to-Fast Earthquakes”.
The system follows a three-step procedure. First, a CNN model, Discriminator for Earthquake and Tremor (DiET), classified input spectrograms into earthquake, tremor, or noise. Second, spectral clustering (von Luxburg, 2007), a graph theory-based technique, was used to associate stations where tremors were detected (Graph-based Associator with Signal Probability: GrASP). Finally, tremors were located every minute using a hybrid method combining envelope cross-correlation and seismic energy ratio.
The system analyzed continuous waveforms recorded from August 2016 to August 2024. Our analysis detected seven times more tremors than the previous study (Nishikawa et al., 2023) using envelope cross-correlation. The newly identified tremors expanded the known spatial distribution of tremor activity, both along the strike and dip, revealing a complementary spatial relationship between tremors and earthquakes. Additionally, our catalog improved temporal resolution, uncovering spatiotemporal patterns of tremors synchronized with slow slip events (Nishimura, 2021). Seismic energy rates of tremors were calculated, showing spatial variations along the strike and dip, with higher rates near asperities of large earthquakes. A positive correlation between energy rates and recurrence intervals was found, suggesting that these spatial variations reflect frictional heterogeneities on plate boundaries. The enhanced spatiotemporal resolution of our tremor catalog provides valuable insights into the relationship between slow and fast earthquakes.
Acknowledgements: We used S-net velocity waveform data (National Research Institute for Earth Science and Disaster Resilience, 2019). This research was supported by JSPS KAKENHI Grant Number JP21H05205 and JP21H05203 in Grant-in-Aid for Transformative Research Areas (A) “Science of Slow-to-Fast Earthquakes”.