11:45 〜 12:00
[SSS08-05] 様々な用途に活用可能な３次元海陸統合地下構造モデルを目指した南海トラフ域における多様な構造モデルの統合
キーワード：南海トラフ、３Dモデル
To predict strong ground motion, tsunamis, seismic hazards, and determine hypocenters with high-resolution, three-dimensional (3D) seismic velocity models must first be established. In the absence of a 3D structural model that can be commonly used for various applications, we proposed a 3D P-wave (Vp) velocity model merged models estimated by results based on regional offshore seismic refraction surveys and seismic observations as a first step (Nakanishi et al., 2018). However, some difficulties remained in evaluating the spatial reliability of the 3D model, because the 3D model is based on multi-scale velocity models with different resolutions and spatial coverages. There was not enough information on shallow structure including geotechnical layers as well as S-wave (Vs) velocity which is needed for many of the applications mentioned above. For onshore shallow structure in Japan, J-SHIS (https://doi.org/10.17598/nied.0012)and JIVSM (Koketsu et al.,2008, 2009, 2012) are the standard models for strong-motion calculations. There are also newly estimated structure models obtained from recent seismic refraction (e.g. Qin et al., 2021) and reflection surveys (e.g. Nakamura et al., 2022; Shiraishi et al., submitted) and seismic observations (Yamamoto et al., 2022, Matsubara et al., 2022). While preserving benefits of the complementary strengths of those models and information, we have updated the previous 3D Vp model to enhance it. Here we present the way how to seamlessly merge such models.
Simultaneously, Arnulf et al. (2022) made progress in creating a model using the 3D model of Nakanishi et al. (2018) as an initial model. Their model can explain the hypocenter parameters determined by the JMA as well as the controlled source (airgun shot) parameters the same as those used by Nakanishi et al. (2018). Therefore, we evaluate the updated 3D Vp model by showing the difference from the model of Arnulf et al. (2022) proposed by traveltime tomography. We compared the observed traveltime for some of the controlled source data used in Arnulf et al. (2022) with the calculated traveltime of the updated 3D Vp model constructed in this study. At present, the RMS travel time residuals look scattered from 0.32 sec to 0.42 sec. Validation with hypocenter determination shows that interplate events are well determined along the plate interface off the Kii Peninsula area while the hypocenters in the Hyuga-nada and Bungo Channel are sometimes dispersed or not well determined (Shiomi et al., 2024SSJ). More comparative data should be obtained and used for further updating the 3D model.
Based on the Vp model, feasible modeling method for the Vs and density model was devised and applied to the updated 3D Vp model. Although issues remain in the modeling method and the validation of the model, we have constructed a method for creating a 3D multi-parameter model. Then we constructed a 3D multi-parameter model as a result of our first challenge. Continuous modification and updates will be necessary to improve the accuracy of the model and make it as a reliable community model. Therefore, procedures and framework were also prepared to enable continuous work on model updating and evaluation at each stage.
This study is part of 'Research Project for Disaster Prevention on the great Earthquakes along the Nankai Trough' funded by Ministry of Education, Culture, Sports, Science and Technology, Japan. This work is partly supported by JSPS KAKENHI Grant Number 19H01982
Simultaneously, Arnulf et al. (2022) made progress in creating a model using the 3D model of Nakanishi et al. (2018) as an initial model. Their model can explain the hypocenter parameters determined by the JMA as well as the controlled source (airgun shot) parameters the same as those used by Nakanishi et al. (2018). Therefore, we evaluate the updated 3D Vp model by showing the difference from the model of Arnulf et al. (2022) proposed by traveltime tomography. We compared the observed traveltime for some of the controlled source data used in Arnulf et al. (2022) with the calculated traveltime of the updated 3D Vp model constructed in this study. At present, the RMS travel time residuals look scattered from 0.32 sec to 0.42 sec. Validation with hypocenter determination shows that interplate events are well determined along the plate interface off the Kii Peninsula area while the hypocenters in the Hyuga-nada and Bungo Channel are sometimes dispersed or not well determined (Shiomi et al., 2024SSJ). More comparative data should be obtained and used for further updating the 3D model.
Based on the Vp model, feasible modeling method for the Vs and density model was devised and applied to the updated 3D Vp model. Although issues remain in the modeling method and the validation of the model, we have constructed a method for creating a 3D multi-parameter model. Then we constructed a 3D multi-parameter model as a result of our first challenge. Continuous modification and updates will be necessary to improve the accuracy of the model and make it as a reliable community model. Therefore, procedures and framework were also prepared to enable continuous work on model updating and evaluation at each stage.
This study is part of 'Research Project for Disaster Prevention on the great Earthquakes along the Nankai Trough' funded by Ministry of Education, Culture, Sports, Science and Technology, Japan. This work is partly supported by JSPS KAKENHI Grant Number 19H01982