17:15 〜 19:15
[SSS11-P14] Three-dimensional attenuation structure in northeast Japan using combined strong-motion data of S-net, KiK-net, and K-NET
キーワード：Seismic tomography, S-net, Attenuation structure, Quality factor
Estimation of attenuation rate of seismic waves in the deep underground structure is important for predicting strong ground motions. In the present study, we tried to construct a wide-area three-dimensional attenuation structure of S waves (Qs) in northeast Japan by combining datasets from land (KiK-net and K-NET) and seafloor (S-net). Strong-motion records with maximum horizontal acceleration lower than 50 cm/s/s were selected at the S-net seafloor stations, whereas records with a maximum acceleration lower than 100 cm/s/s were chosen at the stations on land (K-NET and KiK-net). The period of data was from August 2016 to October 2023, and the area covered by the earthquakes and observation stations used were located within 138 to 148 degrees east and 32 to 45 degrees north. We calculated Fourier spectral amplitude in the frequency range of 1 to 10 Hz and applied a block inversion method (Nakamura and Uetake 2002), treating the source spectrum, Qs, and site amplification factors as unknown parameters. The block size was set as 0.2 by 0.2 degrees in the horizontal directions and 20 km along the depth direction. The stations were divided into 10 groups, assuming that each group had the same site amplification: the stations of K-NET and KiK-net were classified into eight groups following Nakamura & Shiina (2019) based on PS-logging data, and for S-net, because of the lack of measured S-wave velocities, stations were classified into two groups, buried and unburied, based on previous study following Dhakal et al. (2024). In the inversion process, we constrained the amplification factor of rock sites (Vs20 >= 1.0 km/s) to 2.0 at the free surface, thereby minimizing the effects of tradeoffs between the source and site factors following Nakamura (2009).
We obtained the following results, supported by checkerboard resolution analysis and previous studies. The obtained Qs structure was generally heterogeneous in both horizontal and vertical directions and in both land and offshore areas (see the accompanying figure below). A pronounced low-Qs belt extending roughly in the north-south direction from Cape Erimo was delineated in the depth range of 0 to 20 km in the Hokkaido region, similar to the results reported in Nakamura & Shiina (2019). Also, in the same depth range of 0 to 20 km, relatively lower Qs areas formed a nearly continuous pattern on the landward side of the Japan Trench, whereas the lower Qs areas were present in the volcanic regions in the land area. Another conspicuous structure found in this study is the presence of alternate bands of relatively high and low Qs values in the depth range of 20 to 40 km along the coastline and farther offshore, respectively, whereas a remarkably lower Qs belt was present in the backarc region. In the 40 to 60 km depth range, high Qs values were generally predominant in the offshore areas. However, a relatively weaker Qs structure appears along the coastline, extending north-south. If this weaker Qs structure is connected to the north-south low-Qs belt observed in the depth ranges of 20 to 40 km in the offshore region and also to the relatively lower Qs values in the depth ranges of 0 to 20 km near the Japan Trench, it can be interpreted as the effect of oceanic crust near the upper surface of the Pacific Plate. We plan to integrate the land and seafloor datasets further and examine the Qs structure in more detail in our future study.
References
Dhakal et al. (2024). Journal of Disaster Research 19(5):760-771
Nakamura (2009). PhD Dissertation, The University of Tokyo, 1-206.
Nakamura & Uetake (2002). Zisin 54, 475-488.
Nakamura & Shiina (2019). Earth, Planets and Space 71:114
Wessel and Smith (1998). Eos Trans AGU 79, 579.
We obtained the following results, supported by checkerboard resolution analysis and previous studies. The obtained Qs structure was generally heterogeneous in both horizontal and vertical directions and in both land and offshore areas (see the accompanying figure below). A pronounced low-Qs belt extending roughly in the north-south direction from Cape Erimo was delineated in the depth range of 0 to 20 km in the Hokkaido region, similar to the results reported in Nakamura & Shiina (2019). Also, in the same depth range of 0 to 20 km, relatively lower Qs areas formed a nearly continuous pattern on the landward side of the Japan Trench, whereas the lower Qs areas were present in the volcanic regions in the land area. Another conspicuous structure found in this study is the presence of alternate bands of relatively high and low Qs values in the depth range of 20 to 40 km along the coastline and farther offshore, respectively, whereas a remarkably lower Qs belt was present in the backarc region. In the 40 to 60 km depth range, high Qs values were generally predominant in the offshore areas. However, a relatively weaker Qs structure appears along the coastline, extending north-south. If this weaker Qs structure is connected to the north-south low-Qs belt observed in the depth ranges of 20 to 40 km in the offshore region and also to the relatively lower Qs values in the depth ranges of 0 to 20 km near the Japan Trench, it can be interpreted as the effect of oceanic crust near the upper surface of the Pacific Plate. We plan to integrate the land and seafloor datasets further and examine the Qs structure in more detail in our future study.
References
Dhakal et al. (2024). Journal of Disaster Research 19(5):760-771
Nakamura (2009). PhD Dissertation, The University of Tokyo, 1-206.
Nakamura & Uetake (2002). Zisin 54, 475-488.
Nakamura & Shiina (2019). Earth, Planets and Space 71:114
Wessel and Smith (1998). Eos Trans AGU 79, 579.