11:00 AM - 1:00 PM
[SGD01-P03] Detection of aseismic slip events along the plate interface after the 2011 Tohoku-oki earthquake inferred from geodetic data
Keywords:The 2011 Tohoku Earthquake, GNSS positioning, Slow slip event, Afterslip
It is important to estimate aseismic slip occurrence on a plate interface of a subduction zone for assessment of mechanical properties and strain releases on the plate-interface. Prior to the 2011 Tohoku earthquake, the following aseismic slip events have reported: afterslip following M7 class interplate earthquakes [Suito et al., 2011], a long-term transient event [e.g., Yokota et al., 2015; Mavrommatis et al., 2015], slow slip events (SSEs) on the shallow plate interface [Ito et al., 2015], and periodical SSEs associated with interplate earthquakes [Uchida et al., 2016]. As for after the 2011 Tohoku earthquake, it is difficult to detect small-scale aseismic slip events because large-scale postseismic deformation has been dominant. In this study, I tried to detect small-scale aseismic slip events by functional fitting of onland GNSS data (trajectory model [e.g., Bevis & Brown, 2014]).
I employed F5 solutions of GEONET at 213 sites locating 36°N–42°N and 139°E– as onland GNSS data. The data period is from Mar. 12, 2011 to Nov. 20, 2021. Common mode errors [Wdonwinski et al., 1997], coseismic offsets, and offsets due to maintenance of the antennas were eliminated from the GNSS data. I performed functional fitting to the GNSS time-series using the decaying functions for modeling postseismic deformation of the 2011 Tohoku earthquake obtained by Fujiwara et al. (2022) and an exponential function for modeling postseismic deformation of each M6– earthquake. As a result, I found systematic misfits after the early 2022; thus, I added a ramp function following Feb. 6, 2020 for the functional fitting. After this, the functional fitting generally well reproduced the observational time-series.
Residuals of the functional fitting, I found single spontaneous SSE without evident seismicity; that is the 2018 Boso SSE [e.g., Ozawa et al., 2019]. Thus, other aseismic slip events were accompanied by seismic events. Prior to 2013, the exponential functions for postseismic deformation of M6– earthquakes were in the trade-off relationship, and they cannot be separately evaluated. After 2013, remarkable postseismic deformations following the interplate earthquakes were found: [1] Mw 6.7 off-Sanriku earthquake on Feb. 17, 2015, [2] Mw 6.8 off-Miyagi earthquake on May 13, 2015, [3] Mw 7.1 off-Miyagi earthquake on Mar. 20, 2020, and [4] Mw 6.8 off-Miyagi earthquake on May 1, 2020. I estimated fault-slip distributions for the postseismic deformation of these events.
Among the above four events, afterslips of [2]–[4] were estimated along the coastline of Miyagi; their slip distributions are generally overlapped with the afterslip distribution of the 2011 Tohoku earthquake [e.g., Tomita et al., 2020]. Meanwhile, the aseismic slip distribution of [1] was broadly appeared including the shallow plate interface off Sanriku, and its moment magnitude was 6.94. Honsho et al. [2019] suggested a SSE on the shallow plate interface was triggered by the event of [1] and demonstrated that its displacements were measured by GNSS-A observations. Thus, I re-estimated the aseismic slip distribution considering the GNSS-A data [Tomita et al., this meeting] and found that large aseismic slip confined on the shallow plate interface off Sanriku with moment magnitude of 7.20. This aseismic slip distribution and its slip amounts are consistent with the repeating earthquake analysis [Honsho et al., 2019]. Since Honsho et al. [2019] did not estimate such a slip distribution, this study is the first outcome modeling this SSE in detail.
I also estimated an aseismic slip distribution for the ramp function component following Feb. 6, 2020. The estimated distribution is quite similar with the long-term transient event prior to the 2011 Tohoku earthquake [Yokota et al., 2015], and maximally 4 cm/yr of aseismic slip was estimated. However, we have to carefully discuss this component has been caused by the plate interface condition or not considering the future GNSS time-series data.
I employed F5 solutions of GEONET at 213 sites locating 36°N–42°N and 139°E– as onland GNSS data. The data period is from Mar. 12, 2011 to Nov. 20, 2021. Common mode errors [Wdonwinski et al., 1997], coseismic offsets, and offsets due to maintenance of the antennas were eliminated from the GNSS data. I performed functional fitting to the GNSS time-series using the decaying functions for modeling postseismic deformation of the 2011 Tohoku earthquake obtained by Fujiwara et al. (2022) and an exponential function for modeling postseismic deformation of each M6– earthquake. As a result, I found systematic misfits after the early 2022; thus, I added a ramp function following Feb. 6, 2020 for the functional fitting. After this, the functional fitting generally well reproduced the observational time-series.
Residuals of the functional fitting, I found single spontaneous SSE without evident seismicity; that is the 2018 Boso SSE [e.g., Ozawa et al., 2019]. Thus, other aseismic slip events were accompanied by seismic events. Prior to 2013, the exponential functions for postseismic deformation of M6– earthquakes were in the trade-off relationship, and they cannot be separately evaluated. After 2013, remarkable postseismic deformations following the interplate earthquakes were found: [1] Mw 6.7 off-Sanriku earthquake on Feb. 17, 2015, [2] Mw 6.8 off-Miyagi earthquake on May 13, 2015, [3] Mw 7.1 off-Miyagi earthquake on Mar. 20, 2020, and [4] Mw 6.8 off-Miyagi earthquake on May 1, 2020. I estimated fault-slip distributions for the postseismic deformation of these events.
Among the above four events, afterslips of [2]–[4] were estimated along the coastline of Miyagi; their slip distributions are generally overlapped with the afterslip distribution of the 2011 Tohoku earthquake [e.g., Tomita et al., 2020]. Meanwhile, the aseismic slip distribution of [1] was broadly appeared including the shallow plate interface off Sanriku, and its moment magnitude was 6.94. Honsho et al. [2019] suggested a SSE on the shallow plate interface was triggered by the event of [1] and demonstrated that its displacements were measured by GNSS-A observations. Thus, I re-estimated the aseismic slip distribution considering the GNSS-A data [Tomita et al., this meeting] and found that large aseismic slip confined on the shallow plate interface off Sanriku with moment magnitude of 7.20. This aseismic slip distribution and its slip amounts are consistent with the repeating earthquake analysis [Honsho et al., 2019]. Since Honsho et al. [2019] did not estimate such a slip distribution, this study is the first outcome modeling this SSE in detail.
I also estimated an aseismic slip distribution for the ramp function component following Feb. 6, 2020. The estimated distribution is quite similar with the long-term transient event prior to the 2011 Tohoku earthquake [Yokota et al., 2015], and maximally 4 cm/yr of aseismic slip was estimated. However, we have to carefully discuss this component has been caused by the plate interface condition or not considering the future GNSS time-series data.