09:45 〜 10:00
[SSS10-03] The complex rupture process of the 2021 MW 8.5 South Sandwich Islands earthquake
On August 12, 2021, the South Sandwich Islands (SSI) earthquake (MW 8.5) occurred in the shallow region of the South Sandwich subduction zone off the coast of the British South Sandwich Islands. At this subduction zone, the South American plate is subducting under the Sandwich plate at a rate of 70–78 mm/yr, in a WSW direction, and the trench axis has a convex curvature to the east. The U.S. Geological Survey (USGS) reported three earthquakes of MW 7.5, MW 8.1, and mb 6.7 in four minutes following the MW 7.5 event. During the sequence, the Global Centroid Moment Tensor (GCMT) Project reported only one thrust earthquake with a duration of 300 s and a magnitude of MW 8.3. These earlier analyses suggest the possibility of a complex process of the rupture sequence of the 2021 SSI earthquake. In this study, we estimate the rupture process of the 2021 SSI earthquake using the Potency Density Tensor Inversion (PDTI), which estimates slip rate and fault geometry simultaneously.
We used the vertical component of the teleseismic waveforms observed at 47 stations downloaded from the SAGE Data Management Center. We removed instrument responses from these waveforms and decimated the signal at 1.2-s resampling. We calculated Green’s functions with a 1-D structure model (CRUST1.0) in the source region. We use a non-planar and non-rectangular model plane based on the Slab2.0 model and the aftershock distribution within three days after the mainshock. A hypothetical rupture front, the rise time of each knot, and the total rupture duration are set to 2 km/s, 180 s and 280 s, respectively.
The result shows that the total moment tensor, calculated by space-time integration of the potency-rate density tensors, has a thrust focal mechanism, and the seismic moment is 6.47×1023 N m (MW 8.5). The thrust faulting is also dominating in the potency density tensors distribution estimated by temporal summation of the potency-rate density functions for each knot. The length of the slip area extends for ~400 km, exhibiting a peak slip at around 210 km SSE of the epicenter. The rupture begins near the hypocenter and propagates up-dip, with thrust faulting, bilaterally expanding toward SW and NE with the rupture front velocity at 2 and 3 km/s, respectively, until 30 s. From 30 s to 70 s, the rupture propagates towards SW along the trench axis at about 3 km/s, and it stalls at around 120 km SSW from the epicenter at 70 s. The rupture stagnates for approximately 75 seconds, before resuming propagation in a SW direction at a rupture speed of approximately 3 km/s at 145 seconds, which is almost identical to the occurrence time of the MW 8.1 earthquake reported by the USGS. The rupture ceases at 280 s.
From the resultant potency-rate density tensors, we extract the spatiotemporal variation of the slip vectors, which includes information of variation of the fault geometry and the rake angles. The azimuth of the slip vectors obtained from the nodal plane that is closer to the subducting plate interface is dominantly of ~70°, which is parallel to the direction of plate subduction, until 20 s. The azimuth then rotates clockwise towards ~85° until 30 s. During the 30–45 s interval the slip vectors rotate towards ~105°, diverging from the plate subduction direction. The azimuth of the slip vectors is about 85° after 145 s, which is close to the direction of plate subduction.
In this presentation we will discuss the variation of slip vectors in the context of a possible slip partitioning that may be associated with the convex South Sandwich subduction zone. In addition, we discuss the possible cause of the irregular rupture propagation with stalling, which should be associated with the apparent slow rupture velocity of the 2021 SSI earthquake.
We used the vertical component of the teleseismic waveforms observed at 47 stations downloaded from the SAGE Data Management Center. We removed instrument responses from these waveforms and decimated the signal at 1.2-s resampling. We calculated Green’s functions with a 1-D structure model (CRUST1.0) in the source region. We use a non-planar and non-rectangular model plane based on the Slab2.0 model and the aftershock distribution within three days after the mainshock. A hypothetical rupture front, the rise time of each knot, and the total rupture duration are set to 2 km/s, 180 s and 280 s, respectively.
The result shows that the total moment tensor, calculated by space-time integration of the potency-rate density tensors, has a thrust focal mechanism, and the seismic moment is 6.47×1023 N m (MW 8.5). The thrust faulting is also dominating in the potency density tensors distribution estimated by temporal summation of the potency-rate density functions for each knot. The length of the slip area extends for ~400 km, exhibiting a peak slip at around 210 km SSE of the epicenter. The rupture begins near the hypocenter and propagates up-dip, with thrust faulting, bilaterally expanding toward SW and NE with the rupture front velocity at 2 and 3 km/s, respectively, until 30 s. From 30 s to 70 s, the rupture propagates towards SW along the trench axis at about 3 km/s, and it stalls at around 120 km SSW from the epicenter at 70 s. The rupture stagnates for approximately 75 seconds, before resuming propagation in a SW direction at a rupture speed of approximately 3 km/s at 145 seconds, which is almost identical to the occurrence time of the MW 8.1 earthquake reported by the USGS. The rupture ceases at 280 s.
From the resultant potency-rate density tensors, we extract the spatiotemporal variation of the slip vectors, which includes information of variation of the fault geometry and the rake angles. The azimuth of the slip vectors obtained from the nodal plane that is closer to the subducting plate interface is dominantly of ~70°, which is parallel to the direction of plate subduction, until 20 s. The azimuth then rotates clockwise towards ~85° until 30 s. During the 30–45 s interval the slip vectors rotate towards ~105°, diverging from the plate subduction direction. The azimuth of the slip vectors is about 85° after 145 s, which is close to the direction of plate subduction.
In this presentation we will discuss the variation of slip vectors in the context of a possible slip partitioning that may be associated with the convex South Sandwich subduction zone. In addition, we discuss the possible cause of the irregular rupture propagation with stalling, which should be associated with the apparent slow rupture velocity of the 2021 SSI earthquake.