2:00 PM - 2:15 PM
[SSS03-02] Shallow tectonic tremor activity in Off Sanriku, near the Japan Trench: does tremor occur in the tsunami earthquake source regions?
★Invited Papers
Keywords:slow earthquake, tremor, OBS
The Japan Trench is a difficult place for monitoring slow earthquake activity in the shallow part of the subduction zone by onshore seismic networks. Very low-frequency earthquakes have been observed by the onshore broadband seismometers (Matsuzawa et al., 2015; Baba et al., 2020). Slow slip events have also been inferred from repeating earthquakes (e.g. Uchida et al., 2016). However, tectonic tremors had not been observed by the onshore seismic network due to attenuation during long propagation distances. However, recent Ocean Bottom Seismometer (OBS) observations shed light on tremor activities in the shallow portion of the Japan Trench. (Ito et al., 2015; Ohta et al., 2019; Tanaka et al., 2019; Nishikawa et al., 2019; 2023; Takahashi et al., 2024). In particular, the stable, long-term seafloor observation by S-net (Aoi et al., 2020) have figured out the spatiotemporal distribution of tremor in the whole Japan Trench, including the Off Sanriku region that is a northern neighboring region of the 2011 Tohoku-Oki coseismic slip region (e.g., Iinuma et al., 2012). However, the tremor activity near the trench axis has not been clear due to limited network coverage of the S-net. The slow earthquake activity in this region is a particular target of interest as it corresponds to the source region of past tsunami earthquakes such as the 1896 Meiji Sanriku earthquake (e.g., Satake et al., 2017) or the 2011 Tohoku-Oki earthquake (e.g., Kubota et al., 2023).
To monitor the possible slow earthquake activity in the source region of tsunami earthquakes, we performed an OBS observation campaign in Off Sanriku, near the trench axis of the Japan Trench. We deployed three mini-dense seismometer arrays and three single seismometers on the trench side of S-net stations to extend our monitoring capability toward the trench side. The campaign observation was conducted between October 2020 and January 2022. We used one OBS from each array and all single stations. 16 S-net stations covering known slow earthquake regions are also incorporated in our study.
To reveal slow earthquake activity near the trench, we detect and locate tremors based on the integrated OBS network. We apply the following procedures, which were also employed in our previous studies (Ohyanagi et al. in prep; 2022 AGU). First, we applied the modified envelope correlation method (Mizuno and Ide, 2019) to the OBS records filtered in 2 Hz - 8 Hz to detect candidate tremor events. After detecting the candidate tremor events, we removed detections that were supposed to be not tremors. First, we remove ordinary earthquakes by adopting Earthquake Transformer (Mousavi et al., 2020) and air-gun pulses generated by active source seismic surveys. Next, we remove detections with a signal-to-noise ratio of less than 2. At last, the tremor catalog is finalized by removing short-duration (< 20 sec) events and applying space-time clustering (e.g., Wech and Cleager. 2008).
The newly conducted OBS observation reveals the spatiotemporal distribution of tremors near the trench, which has not been resolved only by S-net. However, they are sparsely distributed and not as active as the main clusters distributed between 40 km to 60 km from the trench axis. Despite of observation by dense OBSs, those detected tremors did not exhibit episodic swarms as observed in the main clusters. This indicates tremor activity in the tsunami earthquake source is lower than in the main tremor swarm, irrespective of tremor detectability. According to the active seismic reflection and refraction survey (Tsuru et al., 2002; A .Ito et al., 2004), this sparse-tremor region corresponds to the thick sediment accreted on the toe of the over-riding plate. The number of tremors drastically increases as the structure transits to the upper crust from the accreted sediment. Such transition of the property of the over-riding plate may control the occurrence of the shallow, slow earthquake in the Off Sanriku region of the Japan Trench.
To monitor the possible slow earthquake activity in the source region of tsunami earthquakes, we performed an OBS observation campaign in Off Sanriku, near the trench axis of the Japan Trench. We deployed three mini-dense seismometer arrays and three single seismometers on the trench side of S-net stations to extend our monitoring capability toward the trench side. The campaign observation was conducted between October 2020 and January 2022. We used one OBS from each array and all single stations. 16 S-net stations covering known slow earthquake regions are also incorporated in our study.
To reveal slow earthquake activity near the trench, we detect and locate tremors based on the integrated OBS network. We apply the following procedures, which were also employed in our previous studies (Ohyanagi et al. in prep; 2022 AGU). First, we applied the modified envelope correlation method (Mizuno and Ide, 2019) to the OBS records filtered in 2 Hz - 8 Hz to detect candidate tremor events. After detecting the candidate tremor events, we removed detections that were supposed to be not tremors. First, we remove ordinary earthquakes by adopting Earthquake Transformer (Mousavi et al., 2020) and air-gun pulses generated by active source seismic surveys. Next, we remove detections with a signal-to-noise ratio of less than 2. At last, the tremor catalog is finalized by removing short-duration (< 20 sec) events and applying space-time clustering (e.g., Wech and Cleager. 2008).
The newly conducted OBS observation reveals the spatiotemporal distribution of tremors near the trench, which has not been resolved only by S-net. However, they are sparsely distributed and not as active as the main clusters distributed between 40 km to 60 km from the trench axis. Despite of observation by dense OBSs, those detected tremors did not exhibit episodic swarms as observed in the main clusters. This indicates tremor activity in the tsunami earthquake source is lower than in the main tremor swarm, irrespective of tremor detectability. According to the active seismic reflection and refraction survey (Tsuru et al., 2002; A .Ito et al., 2004), this sparse-tremor region corresponds to the thick sediment accreted on the toe of the over-riding plate. The number of tremors drastically increases as the structure transits to the upper crust from the accreted sediment. Such transition of the property of the over-riding plate may control the occurrence of the shallow, slow earthquake in the Off Sanriku region of the Japan Trench.