5:15 PM - 6:45 PM
[U15-P06] Centroid moment tensor analysis for aftershocks of the 2024 Noto Peninsula earthquake (Mw7.5) using a 3-D seismic velocity structure model
Keywords:Centroid moment tensor, reverse faults, strike-slip faults, normal faults
At 16:10, January 1st, 2024 (JST), the 2024 Noto Peninsula earthquake with a Mw of 7.5 occurred. The large slip area of the mainshock extended from the northwestern part of the Noto Peninsula to off Sado. Accurate determination of the centroid location and focal mechanisms of the aftershocks is necessary for investigating the fault geometry of the mainshock and evaluating the stress state. However, for earthquakes occurring in a region that has a heterogeneous seismic velocity structure, such as a land-ocean boundary, the accurate estimation of the centroid location and focal mechanisms using a 1-D seismic velocity structure model may be difficult. Therefore, this study aims to Comprehensively and accurately determine the centroid moment tensors (CMTs) using a 3-D seismic velocity structure model.
We selected 955 earthquakes with JMA-scale magnitudes (MJMA) of 3.2–6.1 that occurred in the Noto Peninsula and the off Sado. We analyzed the earthquakes that occurred from January 1 to February 4, 2024. For CMT inversion, we used velocity seismograms of the vertical, north-south, and east-west components recorded by the broad-band velocimeters in F-net (NIED, 2019). We applied a band-pass filter with a period range of 20–50 s. We selected the stations within an epicentral distance of 50–300 km and 50–400 km, for earthquakes with MJMA<5.0 and MJMA>5.0, respectively. We used a simulation code based on the staggered-grid finite difference method (Maeda et al., 2017) to calculate the Green’s functions. The assumed 3-D model is based on the Japan Integrated Velocity Structure Model (JIVSM; Koketsu et al., 2012), including 3-D heterogeneous structures such as sedimentary layers. The assumed model included the JIVSM with the minimum S-wave velocity of 1.0 km/s, a topography model, and a seawater layer. To obtain the centroid location and time, a grid search was conducted to minimize the residuals between the observed and synthetic seismograms using the moment tensor solution obtained by the least square method.
We obtained 180 CMT solutions whose variance reductions were larger than 30%. The reason why the CMT solutions with variance reductions were less than 30% was mainly due to the low signal-to-noise ratio caused by the active seismicity immediately after the mainshock. The CMT catalog suggested that many earthquakes occurred within a depth range of 5–15 km and were characterized as reverse faults with a northwest-southeast compressive axis in a whole aftershock area. The spatial distribution of the CMT solutions indicated a southeastward dip trend beneath the Noto peninsula and a northwestward dip trend beneath the off Sado. This result is consistent with the dip angles of the assumed fault models suggested by the Japan Sea Earthquake and Tsunami Research Project. Moreover, we applied UMAP to the CMT catalog for a dimensionality compression (Kubo et al., 2023) and applied the hierarchical agglomerative clustering to investigate the characteristics of the CMT solutions. In the northern part of the Shiga town, most earthquakes were characterized as reverse faults with an east-west compressive axis, and centroid locations were obtained at depths shallower than 5 km and locally in the horizontal direction. Beneath the Noto Peninsula, we obtained strike-slip and reverse faults. This result is consistent with the stress field (Terakawa & Matsu’ura, 2010). Furthermore, off the northeast coast of Suzu city, several normal faults were obtained locally at a depth of ~5 km, with a Mw of ~3.5. For these earthquakes, the F-net MT solutions and the Hi-net focal mechanisms determined by using the first motion polarity also showed normal faults. The area where these normal faults were located was close to the region where the dip trend of the faults switched, which indicates that the heterogeneous stress field caused by the switching of the faults may induce the normal fault events.
We selected 955 earthquakes with JMA-scale magnitudes (MJMA) of 3.2–6.1 that occurred in the Noto Peninsula and the off Sado. We analyzed the earthquakes that occurred from January 1 to February 4, 2024. For CMT inversion, we used velocity seismograms of the vertical, north-south, and east-west components recorded by the broad-band velocimeters in F-net (NIED, 2019). We applied a band-pass filter with a period range of 20–50 s. We selected the stations within an epicentral distance of 50–300 km and 50–400 km, for earthquakes with MJMA<5.0 and MJMA>5.0, respectively. We used a simulation code based on the staggered-grid finite difference method (Maeda et al., 2017) to calculate the Green’s functions. The assumed 3-D model is based on the Japan Integrated Velocity Structure Model (JIVSM; Koketsu et al., 2012), including 3-D heterogeneous structures such as sedimentary layers. The assumed model included the JIVSM with the minimum S-wave velocity of 1.0 km/s, a topography model, and a seawater layer. To obtain the centroid location and time, a grid search was conducted to minimize the residuals between the observed and synthetic seismograms using the moment tensor solution obtained by the least square method.
We obtained 180 CMT solutions whose variance reductions were larger than 30%. The reason why the CMT solutions with variance reductions were less than 30% was mainly due to the low signal-to-noise ratio caused by the active seismicity immediately after the mainshock. The CMT catalog suggested that many earthquakes occurred within a depth range of 5–15 km and were characterized as reverse faults with a northwest-southeast compressive axis in a whole aftershock area. The spatial distribution of the CMT solutions indicated a southeastward dip trend beneath the Noto peninsula and a northwestward dip trend beneath the off Sado. This result is consistent with the dip angles of the assumed fault models suggested by the Japan Sea Earthquake and Tsunami Research Project. Moreover, we applied UMAP to the CMT catalog for a dimensionality compression (Kubo et al., 2023) and applied the hierarchical agglomerative clustering to investigate the characteristics of the CMT solutions. In the northern part of the Shiga town, most earthquakes were characterized as reverse faults with an east-west compressive axis, and centroid locations were obtained at depths shallower than 5 km and locally in the horizontal direction. Beneath the Noto Peninsula, we obtained strike-slip and reverse faults. This result is consistent with the stress field (Terakawa & Matsu’ura, 2010). Furthermore, off the northeast coast of Suzu city, several normal faults were obtained locally at a depth of ~5 km, with a Mw of ~3.5. For these earthquakes, the F-net MT solutions and the Hi-net focal mechanisms determined by using the first motion polarity also showed normal faults. The area where these normal faults were located was close to the region where the dip trend of the faults switched, which indicates that the heterogeneous stress field caused by the switching of the faults may induce the normal fault events.