Marine terraces have long been a subject of paleoseismology to reveal the rupture history of megathrust earthquakes. However, the mechanisms underlying their formation, in relation to crustal deformation, have not been adequately explained by kinematic models. A key challenge has been that the uplifted shoreline resulting from a megathrust earthquake tends to subside back to sea level during subsequent interseismic periods. This study focuses on the remaining permanent vertical deformation resulting from steady plate subduction and examines it quantitatively using three plate subduction models. Specifically, we pay attention to the effects of irregular geometries in the plate interface, such as subducted seamounts. Besides a simplified model examination, this study employs the plate geometry around the Sagami trough, central Japan, to compare with surface deformation observation. The subduction models employed are the kinematic subducting plate model, the elastic/viscoelastic fault model, and the mechanical subducting plate model (MSPM). The MSPM, introduced in this study, allows for more realistic simulations of crustal displacements by imposing net zero shear stress change on the plate boundary. Notably, the presence of a subducted seamount exerts a significant influence on surface deformation, resulting in a concentrated permanent uplift above it. The simulation of earthquake sequence demonstrates that coseismic uplifts can persist over time and contribute to the formation of marine terraces. The results demonstrated that the geological observations of coseismic and long-term deformations can be explained by the influence of a subducted seamount, previously identified in seismic surveys.

The giant 2011 Tohoku earthquake (M9.0) could be expected to induce an 8-class outer-rise earthquake at the Japan Trench. In order to assess the risk of tsunamis from outer-rise events, we carried out tsunami simulations using 33 simple rectangular fault models with 60 degrees dips based on geophysical studies of the Japan Trench. The largest tsunami resulting from these models, a fault 332 km long producing a 8.66 normal-faulting event, had a maximum height of 27.0 m. We tested variations of the predictions due to the uncertainties in the assumed parameters. Because seismic observations and surveys show that the dip angles of outer-rise faults range from 45 to 75 degrees, we calculated tsunamis from events on fault models with 45-75 degree dips. We tested a compound fault model with 75 degrees dip in the upper half and 45 degrees dip in the lower half. Rake angles were varied by plus-minus 15 degrees. We also tested models consisting of small subfaults with dimensions of about 60 km, models using other earthquake scaling laws, and models including dispersive tsunami effects. Predicted tsunami heights changed by 5-10% for dip angle changes, about 5-10% from considering tsunami dispersion, about 2% from rake angle changes, and about 1% from using the model with subfaults. The use of different earthquake scaling laws changed predicted tsunami heights by about 50% on average for the 33 fault models. We emphasize that the earthquake scaling law used in tsunami predictions for outer-rise earthquakes should be chosen with great care.

The spatial variation of azimuthal S-wave phase velocity anisotropies caused by differential horizontal stress along the subducting plate at the Nankai Trough was analyzed to understand the stress state of the overhung block of the forearc region, off Kii Peninsula, Japan. We conducted controlled-source seismic surveys along the circumference of a 3 km diameter circle centered at each seismometer of a cabled earthquake observatory installed on the seafloor above the Kumano basin of the Nankai Trough subduction zone. We applied an anisotropy semblance method to estimate the orientation of fast and slow S-wave velocities of both shallow sediments and deep accretionary prism using the multi-azimuth seismic dataset acquired at each seismometer location. The estimated orientations of fast S-wave velocity are parallel to the convergent direction of the subducting place beneath the Kumano basin in the deeper accretionary prism while perpendicular to the convergent direction in the shallow sediments inside the Kumano basin. The orientations of these fast S-wave polarization show good agreement with those of horizontal maximum stress orientations estimated in situ borehole measurements in the observation area Then differential horizontal stress field in the Nankai Trough region was estimated from obtained S-wave anisotropy using a simple crack model. The azimuths of fast S-wave polarization and the derived differential stresses could be explained well by the tectonics of the Nankai Trough subduction zone. These results strongly suggested that the S-wave azimuthal anisotropy measurements could be used to monitor the subsurface stress field as a function of time.

We developed a mechanical subducting plate model and re-examined the crustal deformation history in the Sagami Trough subduction zone, central Japan, the northernmost convergence boundary of the Philippine Sea Plate. The elevation distributions and formation ages of the Holocene marine terraces, representing past coseismic and long-term coastal uplifts, have been thoroughly investigated in this region. However, no physically consistent formation scenario to explain them has been demonstrated. Surface deformations within subduction zones are typically calculated using kinematic elastic dislocation models, such as the back-slip model. However, these models cannot explain permanent deformation after an earthquake sequence. This study develops a mechanical subducting plate model that balances the slips of interplate shear stress and can produce permanent deformations caused by a local bump geometry. We modeled earthquake recurrences by shear stress accumulation and coupling patches. As a result, we successfully reproduced the averaged uplift rate distribution estimated from the Holocene marine terraces. The findings suggest that the subducted seamount significantly affects long-term deformation patterns. In addition, the discrepancy between the elevation distributions and formation ages of Holocene marine terraces, which previous geological studies have indicated, can be interpreted by the rupture delay of coupling patches. This study also demonstrates that the traditional assumption of the back-slip model on the plate boundary for long-term subduction possibly results in an oversimplified model.