5:15 PM - 6:30 PM
[SSS10-P12] Seeking for strong motion prediction model near the fault, considering the detailed fault structure shallower than seismogenic layer: An example from the 2016 Kumamoto earthquake
Keywords:Fault model, Shallow crustal earthquakes, Surface rupture, Kumamoto Earthquake
1.Introduction
In shallow crustal earthquakes caused by fault activity such as the Kumamoto earthquake, damage to structures is predicted due to 1) fault displacement and 2) strong ground motion in close proximity of the fault. Significant damage and huge economic loss caused by the rupture of active faults under the urban areas will be anticipated. For fault displacement, it is important to examine the distribution and deformation style of active faults, but for strong ground motions, the study of prediction methods for the vicinity of faults has not progressed sufficiently. Although the shallow structure is thought to be crucial is the case of the Kumamoto earthquake as pointed out by Irikura et al. (2016), no studies have been conducted considering the fault structure shallower than the upper end of the seismogenic layer. The purpose of this study is to construct a strong motion prediction method for shallow crustal earthquakes with articulate displacement and deformation nearby the active faults. Here, we report a study on how to handle a fault model shallower than the seismogenic layer.
2.Models for the shallow part of the fault
We have investigated the details of the fault model shape for predicting crustal deformation and strong ground motion in the nearby area of the fault (Norimatsu and Toda, 2020: Seismological Society of Japan Autumn Meeting). As the results, it was confirmed that the shape of the shallow part of the fault strongly affects the crustal deformation and strong ground motion near the fault. Prediction accuracy can be expected to improve by reflecting the complexity of branch faults playing as slip partitioning and the top shape for the fault model. In this study, the shape of the upper end of the fault model was set in detail based on the surface rupture of the Kumamoto earthquake by Kumahara et al. (2016). Employing the subparallel structure of the Idenokuchi fault and the Futagawa fault suggested by Toda et al. (2016), we found a possibility to explain the characteristics of building damage caused by the Kumamoto earthquake (Yoshimi and Toda, 2020: Japanese Society for Active Fault Studies). In particular, the areas where the fault shape is complicated and the large displacement observed are spatially correlated with the ones the building damage was significant.
3.Strong ground motion estimation for the shallower the seismogenic layer
Currently, the "strong ground motion recipe (The Headquarters for Earthquake Research Promotion, 2017)" is widely used as the standard method of strong ground motion evaluation. However, the "recipe" is a method that targets only in the seismogenic layer and no standard method has been established for the strong ground motion evaluation for areas shallower than the seismogenic layer. Hikima et al. (2012) and Tanaka et al. (2017) suggest that the shape of the slip velocity time function is different in the seismogenic layer and shallower than the seismogenic layer. In addition, Tanaka et al. (2017) propose to use the regularized Yoffe function (Tinti et al., 2005) as the slip velocity time function for shallower than the seismogenic layer. In this study, we tried to use the regularized Yoffe function into the strong ground motion simulation by the wavenumber integration method. The slip velocity-time function on the seismogenic layer was used as Nakamura-Miyatake function (Nakamura-Miyatake, 2000) with reference to the "recipe". By using the regularized Yoffe function in the area of shallower than the seismogenic layer, it is possible to reduce the overestimation of the calculated waveform.
4.Summary and future works
In this study, it has been shown that the structure of the shallow part of the fault may have a large influence on the strong ground motion and crustal deformation nearby the fault. In order to model a detailed faults structure shallower than the seismogenic layer, it is effective to utilize topographical and geological information such as surface rupture and active faults. This information is effective not only for setting a fault model that explains the past earthquakes, but also for assuming a strong ground motion prediction model based on active faults.
As a future work, it is necessary to study how to handle the seismic moments shallower than the seismogenic layer. There is currently no detailed and clear estimation standard for stress drop near the ground surface as well.
In shallow crustal earthquakes caused by fault activity such as the Kumamoto earthquake, damage to structures is predicted due to 1) fault displacement and 2) strong ground motion in close proximity of the fault. Significant damage and huge economic loss caused by the rupture of active faults under the urban areas will be anticipated. For fault displacement, it is important to examine the distribution and deformation style of active faults, but for strong ground motions, the study of prediction methods for the vicinity of faults has not progressed sufficiently. Although the shallow structure is thought to be crucial is the case of the Kumamoto earthquake as pointed out by Irikura et al. (2016), no studies have been conducted considering the fault structure shallower than the upper end of the seismogenic layer. The purpose of this study is to construct a strong motion prediction method for shallow crustal earthquakes with articulate displacement and deformation nearby the active faults. Here, we report a study on how to handle a fault model shallower than the seismogenic layer.
2.Models for the shallow part of the fault
We have investigated the details of the fault model shape for predicting crustal deformation and strong ground motion in the nearby area of the fault (Norimatsu and Toda, 2020: Seismological Society of Japan Autumn Meeting). As the results, it was confirmed that the shape of the shallow part of the fault strongly affects the crustal deformation and strong ground motion near the fault. Prediction accuracy can be expected to improve by reflecting the complexity of branch faults playing as slip partitioning and the top shape for the fault model. In this study, the shape of the upper end of the fault model was set in detail based on the surface rupture of the Kumamoto earthquake by Kumahara et al. (2016). Employing the subparallel structure of the Idenokuchi fault and the Futagawa fault suggested by Toda et al. (2016), we found a possibility to explain the characteristics of building damage caused by the Kumamoto earthquake (Yoshimi and Toda, 2020: Japanese Society for Active Fault Studies). In particular, the areas where the fault shape is complicated and the large displacement observed are spatially correlated with the ones the building damage was significant.
3.Strong ground motion estimation for the shallower the seismogenic layer
Currently, the "strong ground motion recipe (The Headquarters for Earthquake Research Promotion, 2017)" is widely used as the standard method of strong ground motion evaluation. However, the "recipe" is a method that targets only in the seismogenic layer and no standard method has been established for the strong ground motion evaluation for areas shallower than the seismogenic layer. Hikima et al. (2012) and Tanaka et al. (2017) suggest that the shape of the slip velocity time function is different in the seismogenic layer and shallower than the seismogenic layer. In addition, Tanaka et al. (2017) propose to use the regularized Yoffe function (Tinti et al., 2005) as the slip velocity time function for shallower than the seismogenic layer. In this study, we tried to use the regularized Yoffe function into the strong ground motion simulation by the wavenumber integration method. The slip velocity-time function on the seismogenic layer was used as Nakamura-Miyatake function (Nakamura-Miyatake, 2000) with reference to the "recipe". By using the regularized Yoffe function in the area of shallower than the seismogenic layer, it is possible to reduce the overestimation of the calculated waveform.
4.Summary and future works
In this study, it has been shown that the structure of the shallow part of the fault may have a large influence on the strong ground motion and crustal deformation nearby the fault. In order to model a detailed faults structure shallower than the seismogenic layer, it is effective to utilize topographical and geological information such as surface rupture and active faults. This information is effective not only for setting a fault model that explains the past earthquakes, but also for assuming a strong ground motion prediction model based on active faults.
As a future work, it is necessary to study how to handle the seismic moments shallower than the seismogenic layer. There is currently no detailed and clear estimation standard for stress drop near the ground surface as well.