10:45 AM - 12:15 PM
[SSS04-P10] Coseismic slip and stress drop estimation of the 2011 Tohoku earthquake in the off-Sanriku shallow region imposing mechanical conditions
It is important to estimate accumulation and release processes of slip deficit and stress in subduction zones, for quantifying strain energy as a driving force of megathrust earthquakes. Kubota et al. (2022) estimated the slip distribution of the 2011 Tohoku earthquake using tsunami data recorded inside the rupture area to discuss the stress drop and interplate mechanical state. However, they focused only on the off-Miyagi region where the large near-trench slip occurred but not the off-Sanriku region due to the limitation in the resolution because of the lack of observations.
When the Tohoku earthquake occurred, an ocean-bottom electro-magnetometer (OBEM) had beeen installed just outside of the Japan Trench off Sanriku (Ichihara et al. 2013). OBEM can record magnetic field changes induced by tsunamis. Recent studies have attempted to utilize the OBEM tsunami data for fault modeling analyses (Kawashima & Toh 2016; Yokoi et al. 2022). In addition, Saito and Noda (2022) proposed an inversion method to directly estimate the interseismic shear stress accumulation rate from surface velocity observations, combining the linear relationships between surface displacements and interplate slips and between interplate slips and shear stress changes. This approach has an advantage in the flexibility to incorporate mechanical conditions of the plate boundary as prior information, compared to the conventional slip inversion approach. In this study, adding the OBEM tsunami data and using the inversion method of Saito and Noda (2022), we attempt to estimate the slip and stress drop off Sanriku with a better resolution.
In the analysis, we divide the plate boundary into a set of triangular elements and give a unit stress drop amount to each element to calculate the surface displacements and tsunamis (i.e., Green's function). In the stress drop inversion, we do not use the non-negative constraint (i.e., allowing shear stress increase), which is often used in the conventional slip inversions to avoid normal-faulting slips. Considering the mechanical condition that the shallowest part of the plate boundary hosts a stable-sliding behavior and does not accumulate the shear stress (Scholtz 1998), we assume the following sets of stress drop elements: [1] no elements in the shallow plate boundary (z < ~10 km) but only in the deep part (z > ~10 km), [2] shallow elements in the off-Sanriku region in addition to [1], and [3] shallow elements in the whole part of the rupture area in addition to [1].
A large stress drop (> ~5–10 MPa) was estimated near the trench axis in the shallow part off Sanriku in the models [2] and [3], which was consistent with the slip inversion model by Kubota et al. (2022). These models reproduced tsunami data recorded off Iwate better than the model [1], which assumed no off-Sanriku shallow stress drop. In particular, the reproductivity of the maximum amplitudes and peak timings of the tsunamis recorded by the OBEM (Ichihara et al. 2013) was much better in the models [2] and [3]. The slip and stress drop distributions in the off-Miyagi region from the models [1] to [3] are similar to each other, as well as to those by Kubota et al. (2022). In the model [3], a large shallow slip and stress drop was estimated in the Off-Fukushima region, although such a large slip was not estimated in the slip inversion by Kubota et al. (2022), which reproduced the observations as reasonable as the model [3]. This may suggest this off-Fukushima shallow slip in the model [3] is due to the uncertainty of the estimation because of the lack of data to constrain it. Our result suggests the significant shear stress release at the shallowest plate boundary off Sanriku, although we need to evaluate the stress release process appropriately and quantitatively, considering the shallow heterogeneous structure near the trench and utilizing the coseismic seafloor displacement data near the trench axis based on the seafloor topography survey (Fujiwara 2021)
When the Tohoku earthquake occurred, an ocean-bottom electro-magnetometer (OBEM) had beeen installed just outside of the Japan Trench off Sanriku (Ichihara et al. 2013). OBEM can record magnetic field changes induced by tsunamis. Recent studies have attempted to utilize the OBEM tsunami data for fault modeling analyses (Kawashima & Toh 2016; Yokoi et al. 2022). In addition, Saito and Noda (2022) proposed an inversion method to directly estimate the interseismic shear stress accumulation rate from surface velocity observations, combining the linear relationships between surface displacements and interplate slips and between interplate slips and shear stress changes. This approach has an advantage in the flexibility to incorporate mechanical conditions of the plate boundary as prior information, compared to the conventional slip inversion approach. In this study, adding the OBEM tsunami data and using the inversion method of Saito and Noda (2022), we attempt to estimate the slip and stress drop off Sanriku with a better resolution.
In the analysis, we divide the plate boundary into a set of triangular elements and give a unit stress drop amount to each element to calculate the surface displacements and tsunamis (i.e., Green's function). In the stress drop inversion, we do not use the non-negative constraint (i.e., allowing shear stress increase), which is often used in the conventional slip inversions to avoid normal-faulting slips. Considering the mechanical condition that the shallowest part of the plate boundary hosts a stable-sliding behavior and does not accumulate the shear stress (Scholtz 1998), we assume the following sets of stress drop elements: [1] no elements in the shallow plate boundary (z < ~10 km) but only in the deep part (z > ~10 km), [2] shallow elements in the off-Sanriku region in addition to [1], and [3] shallow elements in the whole part of the rupture area in addition to [1].
A large stress drop (> ~5–10 MPa) was estimated near the trench axis in the shallow part off Sanriku in the models [2] and [3], which was consistent with the slip inversion model by Kubota et al. (2022). These models reproduced tsunami data recorded off Iwate better than the model [1], which assumed no off-Sanriku shallow stress drop. In particular, the reproductivity of the maximum amplitudes and peak timings of the tsunamis recorded by the OBEM (Ichihara et al. 2013) was much better in the models [2] and [3]. The slip and stress drop distributions in the off-Miyagi region from the models [1] to [3] are similar to each other, as well as to those by Kubota et al. (2022). In the model [3], a large shallow slip and stress drop was estimated in the Off-Fukushima region, although such a large slip was not estimated in the slip inversion by Kubota et al. (2022), which reproduced the observations as reasonable as the model [3]. This may suggest this off-Fukushima shallow slip in the model [3] is due to the uncertainty of the estimation because of the lack of data to constrain it. Our result suggests the significant shear stress release at the shallowest plate boundary off Sanriku, although we need to evaluate the stress release process appropriately and quantitatively, considering the shallow heterogeneous structure near the trench and utilizing the coseismic seafloor displacement data near the trench axis based on the seafloor topography survey (Fujiwara 2021)