Ground motion estimation for the elevated bridges of the Kyushu Shinkansen derailment caused by the foreshock of the 2016 Kumamoto earthquake based on the site-effect substitution method
© The Author(s) 2016
Received: 11 July 2016
Accepted: 19 November 2016
Published: 1 December 2016
Beginning on April 14, 2016, a series of damaging earthquakes hit Kumamoto and Oita Prefectures in Kyushu, Japan. This series began with the Mw 6.2 event (April 14, 21:26 JST; hereafter referred to as the foreshock, although the source faults of the foreshock and the main shock are not identical), followed by the Mw 7.0 main shock (April 16, 1:25 JST). As of June 30, this sequence involved more than 1800 perceptible earthquakes (Japan Meteorological Agency (JMA) 2016). The entire sequence was named the “2016 Kumamoto earthquake” by the JMA.
List of strong motion stations
Derailment site of Kyushu Shinkansen (temporary earthquake observation site)
Kashima Town Office
K-NET Kumamoto [past] (temporary earthquake observation site)
K-NET Kumamoto (present)
Mashiki Town Office
To analyze the mechanism of this third derailment, it is crucial to evaluate the strong ground motion at the derailment site with high accuracy. For this study, strong motion estimation was carried out at the derailment site with consideration for empirical site amplification and phase effects. First, temporary earthquake observations were conducted at the derailment site. Then, ground shaking characteristics at the derailment site were evaluated based on the obtained records. In addition, strong ground motions at the derailment site during the foreshock were evaluated based on the site-effect substitution method (Hata et al. 2011). The same method was also applied to estimate strong ground motions at a nearby strong motion station, where the foreshock ground motion was observed, to investigate the applicability of this method to this particular earthquake. The estimated ground motions were highly consistent with the observed ground motions, which indicate the applicability of this estimation method. Finally, the response spectra of the evaluated strong motions at the derailment site were compared with the design response spectrum of the Specifications for Railway Structures (Railway Technical Research Institute (RTRI) 1999) based on an effect of very soft ground (G5 ground; RTRI 2012).
Observed ground motions
Conventional estimation methods
Conventional approaches to estimate strong ground motions after an earthquake at a site of interest based on moderate earthquake records fall into two categories. One of these categories is based on full strong motion simulation using fault models and moderate earthquake records as empirical Green’s functions (e.g., Suzuki and Iwata 2006). In these approaches, because the time history of strong ground motion is generated, one can obtain ground motion parameters such as peak amplitude and duration. However, the reliability of the results from these approaches depends on the quality of the fault model, and it may not always be possible to obtain a sufficiently reliable model.
For the second category, a practical estimation method for strong ground motions (called the “site-effect substitution method”) was proposed by Hata et al. (2011). This method is based on records of moderate earthquakes at both the site of interest and a nearby permanent strong motion observation station, and on a record of a large earthquake at the nearby station. Because this method is focused not only on the difference in site amplification factors but also on the difference in site phase effects between the site of interest and the nearby station, it can be used to compute time histories of strong ground motions at the site of interest with high accuracy (e.g., Hata et al. 2013, 2016b). Because JKM is located close to SNK, as shown in Figs. 1 and 2, this method is accepted as suitable for this study.
Another advantage of this method is its simplicity. Unlike full strong motion calculations, such as those based on the Stochastic Green’s Function Method (e.g., Hata et al. 2012), this method does not require a fault model for the large event. Therefore, it can be applied at an early stage of the response to a large event even if a reliable fault model is not yet available. Because the fault models for this earthquake are still under development, we have used the site-effect substitution method (Hata et al. 2011) to estimate strong ground motions in this study.
Temporary earthquake observation
During the 2016 Kumamoto earthquake sequence, in addition to the damage caused by the main shock, damage from the foreshock was significant. In particular, during the reconnaissance survey conducted on April 15, the location of the derailment was found just to the south of the JR Kumamoto Station (see Fig. 2).
Parameters for the foreshock and the observed earthquake events
April 14, 2016
April 15, 2016
April 15, 2016
April 15, 2016
Origin timea (H:Min:S)
NW Kumamoto Pref.
NW Kumamoto Pref.
NW Kumamoto Pref.
NW Kumamoto Pref.
M0 b (Nm)
1.74E + 18
4.48E + 14
4.25E + 14
5.52E + 14
(strike, dip, rake)b (deg.)
(212, 89, 164)
(33, 87, 162)
(207, 70, 178)
(281, 51, 59)
Site amplification factor
The horizontal site amplification factors for the K-NET stations, including the former K-NET Kumamoto station (KNK1), have previously been evaluated by Nozu et al. (2007) based on spectral inversion. However, the site amplification factors for the JMA stations, MLIT stations, and SNK have not been previously reported. In this study, the spectral ratio method (Hata et al. 2014) was applied to evaluate the horizontal site amplification factors at JKM, MKM, and SNK. This method was performed based on moderate earthquake records obtained at the reference station and the sites of interest simultaneously. The sites of interest include JKM, MKM, and SNK. KNK1 was selected as the reference station for this study.
The procedure of the spectral ratio method is summarized as follows. For each combination of a site of interest and the reference station, the spectral ratio of the Fourier amplitudes of the records from the reference station and the site of interest was calculated. Here, the moderate earthquake records from before the foreshock were used for JKM and MKM (Hata et al. 2016c, d). For SNK, the records after the foreshock were used (see Table 2). The effects of geometrical spreading and anelastic attenuation were considered as the path effect (Boore 1983) to correct the Fourier spectra. We assumed a Q value of Q = 104 f 0.63 (Kato 2001). The mean of the corrected spectral ratios (the site of interest/the reference station) was calculated. The site amplification factor at the site of interest was obtained as the product of the site amplification factor at the reference station and the spectral ratio. Here, the frequency range for the evaluation of the site amplification factor is from 0.2 to 10 Hz because the site amplification factor at the reference station is reliable within this range (Nozu et al. 2007). It should be noted that in this scheme, the site amplification factors at JKM, MKM, and SNK represent amplification from the seismic bedrock to the ground surface.
Shear wave velocity profiles at JKM, MKM, and SNK
Shear wave velocity (m/s)
Ground motion estimation
Then, the Fourier phase during the foreshock at each site of interest was approximated as the Fourier phase at the same site during EQ-2 (see Fig. 4), which occurred close to the foreshock. Finally, an inverse Fourier transform was conducted to obtain a causal time history (Nozu et al. 2009) of strong ground motions during the foreshock at each site of interest. To consider the effect of nonlinear ground response from the engineering bedrock at a depth of 26 (see Table 3b) and 35 m (see Table 3a) to the ground surface at MKM and SNK, respectively, not only linear calculations but also the equivalent linear calculations with the empirical dynamic deformation properties (Yasuda and Yamaguchi 1985; Yoshida et al. 2002; Hata et al. 2016b) were performed (see Fig. 6).
PGA and PGV values at JKM and SNK during the foreshock and main shock
PGA value (Gal)
PGV value (cm/s)
Figures 10a, b and 11a, b show the absolute acceleration spectra and the relative velocity spectra (damping: 5%) based on the observed foreshock records at JKM with respect to the design response spectrum of the Specifications for Railway Structures (RTRI 2012) and the surface ground conditions at JKM (see Table 3c). As shown in Figs. 10a, b, and 11a, b, at JKM, the observed acceleration and velocity response spectra recorded during the foreshock are almost included within the design response spectrum. This rough inclusion agrees well with the minor damage to the elevated bridge near JKM noted during the authors’ field reconnaissance.
Summary and conclusions
Although the JMA Kumamoto Station site and the derailed bridge site are closely located, the ground response characteristics at these sites differ, which indicates differences in the local site effects at these sites and the importance of considering local site effects in estimating strong ground motions. Therefore, the local site effects at the derailment site were evaluated based on temporary earthquake observations after the foreshock associated with this event.
A striking feature of the estimated acceleration and velocity waveforms for the foreshock at the derailment site based on the site-effect substitution method is that these waveforms include much larger amplitudes compared to the records at the permanent stations near the derailment site at the JMA Kumamoto and MLIT Kumamoto stations. The reliability of the estimations was confirmed by the fact that the same method reproduces strong ground motions at the MLIT Kumamoto station site accurately.
As a result of the estimation, it was found that around the natural period of the elevated bridge (0.7 s), the acceleration and velocity response spectra exceeded the design response spectrum of the Specifications for Railway Structures. Therefore, it is likely that the inertial force on the structure during the foreshock was one of the main causes of the derailment on the elevated bridge.
These results suggest that it is important to take into account site-specific characteristics of strong ground motions for reasonable safety assessment of anti-derailment devices for future large earthquakes. In future study, seismic response analysis of elevated bridges for bullet trains will be carried out using estimated strong ground motions.
Data and resources
K-NET and KiK-net strong motion data were provided by the National Research Institute for Earth Science and Disaster Resilience at www.kyoshin.bosai.go.jp (last accessed July 2016). Strong motion data from the JMA network were provided by the Japan Meteorological Business Support Center on CD-ROMs and by the JMA at www.data.jma.go.jp/svd/eqev/data/kyoshin/jishin/index.html (last accessed July 2016). Strong motion data at MKM can be obtained from the Earthquake Disaster Management Division, Road Structures Department, National Institute for Land and Infrastructure Management at www.nilim.go.jp/lab/rdg/ (last accessed July 2016). Strong motion data for the foreshock were provided by the local government office of Kumamoto Prefecture at www.data.jma.go.jp/svd/eqev/data/kyoshin/jishin/160414_kumamoto/index2.html (last accessed July 2016). The Centroid Moment Tensor (CMT) solutions of the F-net were obtained from the Full Range Seismograph Network of Japan (F-net) Web site at www.fnet.bosai.go.jp (last accessed July 2016).
YH conducted the observations and estimations. YH, MY, and AK drafted the manuscript. MY, AK, and YT acquired the strong motion data and performed the ground investigation. AK, HM, and YT participated in the reconnaissance survey of the seismic damage. HM, YT, and MA participated in the discussion and the interpretation. All authors read and approved the final manuscript.
The authors thank the anonymous residents of Nishi Ward, Kumamoto City for generously cooperating in conducting the temporary earthquake observations and the ground investigation. The authors also appreciate the assistance of Mr. Fumihiro Minato, a graduate student of Osaka University, during the observation. This study was carried out as an activity of the Subcommittee on Earthquake Damage Survey [Head: Associate Prof. Yoshikazu Takahashi (Kyoto University)] and the Subcommittee on Performance-based Seismic Design Method for Bridges [Head: Dr. Masaaki Yabe (Chodai Co., Ltd)], organized by the Earthquake Engineering Committee, Japan Society of Civil Engineers (JSCE) [Chairperson: Prof. Sumio Sawada (Disaster Prevention Research Institute (DPRI), Kyoto University)]. The manuscript was significantly improved based on comments from anonymous reviewers and Dr. Haruo Horikawa (guest editor).
The authors declare that they have no competing interests.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.
- Aoi S, Kunugi T, Fujiwara H (2004) Strong-motion seismograph network operated by NIED: K-NET and KiK-net. J Jpn Assoc Earthq Eng 4(3):65–74. doi:10.5610/jaee.4.3_65 Google Scholar
- Boore DM (1983) Stochastic simulation of high-frequency ground motions based on seismological models of the radiated spectra. Bull Seismol Soc Am 73(6A):1865–1894. http://www.bssaonline.org/content/73/6A/1865.abstract
- Hata Y, Nozu A, Ichii K (2011) A practical method to estimate strong ground motions after an earthquake based on site amplification and phase characteristics. Bull Seismol Soc Am 101(2):688–700. doi:10.1785/0120100142 View ArticleGoogle Scholar
- Hata Y, Ichii K, Nozu A (2012) Three-dimensional non-linear FEM analysis of a seismic induced crack at an airport runway. Soil Dyn Earthq Eng 42:105–118. doi:10.1016/j.soildyn.2012.06.009 View ArticleGoogle Scholar
- Hata Y, Ichii K, Nozu A, Maruyama Y, Sakai H (2013) Ground motion estimation at the farthest liquefaction site during the 2011 off the Pacific coast of Tohoku Earthquake. Soil Dyn Earthq Eng 48:132–142. doi:10.1016/j.soildyn.2013.01.002 View ArticleGoogle Scholar
- Hata Y, Nozu A, Ichii K (2014) Variation of earthquake ground motions within very small distance. Soil Dyn Earthq Eng 66:429–442. doi:10.1016/j.soildyn.2014.08.006 View ArticleGoogle Scholar
- Hata Y, Goto H, Yoshimi M (2016a) Preliminary analysis of strong ground motions in the heavily damaged zone in Mashiki Town, Kumamoto, Japan, during the main shock of the 2016 Kumamoto Earthquake (MW 7.0) observed by a dense seismic array. Seismol Res Lett 87(5):1044–1049. doi:10.1785/0220160107 View ArticleGoogle Scholar
- Hata Y, Kataoka S, Nozu A (2016b) Accuracy of methods to estimate strong ground motions using nearby strong motion records: the equivalent linear method versus the site effect substitution method. J Jpn Assoc Earthq Eng (Jpn Engl Abstr) 16(4):195–214. doi:10.5610/jaee.16.4_195 Google Scholar
- Hata Y, Yabe M, Kasai A, Matsuzaki H, Takahashi Y, Akiyama M (2016c) Evaluation of site amplification factors at the bridges damaged by the 2016 Kumamoto earthquake sequence based on temporary earthquake observation. In: Proceedings of the 19th symposium on performance-based seismic design method for bridges. Tokyo (in Japanese) JSCE, pp 127–134Google Scholar
- Hata Y, Murata A, Miyajima M (2016d) An approach on strong motion estimation in Kumamoto Castle during foreshock and main shock of the 2016 Kumamoto Earthquake: evaluation of site amplification factors at sites of seismic observation station by JMA. In: Proceedings of the 55th national conference of Japan landslide society, Kochi (in Japanese), pp 252–253Google Scholar
- Horika K (2013) Clarification of mechanism of Shinkansen derailment in the 2011 Great East Japan Earthquake and countermeasures against earthquakes. JR EAST Tech Rev 27:13–16. https://www.jreast.co.jp/e/development/tech/pdf_27/Tec-27-13-16eng.pdf
- Japan Meteorological Agency (2016) Monthly report on seismic and volcanic activities in June 2016. JMA Press Release on 8 July (in Japanese). http://www.jma.go.jp/jma/press/1607/08a/1606jishin.html
- Kato K (2001) Evaluation of source, path, and site amplification factors from the K-NET strong motion records of the 1997 Kagoshima-Ken-Hokuseibu earthquakes. J Struct Constr Eng (Jpn Engl Abstr) AIJ 543:61–68. http://ci.nii.ac.jp/naid/110004304043
- Kumamoto Prefecture (2015) Regional disaster prevention plan in Kumamoto Prefecture: edition for countermeasures against earthquake and tsunami disasters. (in Japanese) Google Scholar
- Ministry of Land, Infrastructure, Transport and Tourism (2016) Disclosure paper on 13th Shinkansen derailment countermeasure conference. MLIT Press Release on 25 May. (in Japanese). http://www.mlit.go.jp/report/press/tetsudo07_hh_000101.html
- Nakaaki S, Sakai K, Murono Y (2012) A comparative study on characteristics of earthquakes in 2011 Tohoku-Pacific Ocean Earthquake. J Struct Eng Earthq Eng (Jpn Engl Abstr) JSCE 68(4):161–168. doi:10.2208/jscejseee.68.I_161 Google Scholar
- Nishimae Y (2004) Observation of seismic intensity and strong ground motion by Japan meteorological agency and local governments in Japan. J Jpn Assoc Earthq Eng 4(3):75–78. doi:10.5610/jaee.4.3_75 Google Scholar
- Nozu A, Nagao T, Yamada M (2007) Site amplification factors for strong-motion sites in Japan based on spectral inversion technique and their use for strong-motion evaluation. J Jpn Assoc Earthq Eng (Jpn Engl Abstr) 7(2):215–234. doi:10.5610/jaee.7.2_215 Google Scholar
- Nozu A, Nagao T, Yamada M (2009) Simulation of strong ground motions using empirical site amplification and phase characteristics: modification to incorporate causality. J Struct Eng Earthq Eng (Jpn Engl Abstr) JSCE 65(3):808–813. doi:10.2208/jsceja.65.808 Google Scholar
- Ogura M (2006) The Niigata Chuetsu earthquake: railway response and reconstruction. Railw Transp Rev 43/44:46–63. http://www.ejrcf.or.jp/jrtr/jrtr43_44/pdf/f46_ogu.pdf
- Railway Technical Research Institute (1999) Specifications for railway structures and commentary: seismic design edition. Maruzen Co. Ltd, Tokyo (in Japanese) Google Scholar
- Railway Technical Research Institute (2012) Specifications for railway structures and commentary: seismic design edition. Maruzen Co. Ltd, Tokyo (in Japanese) Google Scholar
- Senna S, Adachi S, Ando H, Araki T, Iisawa K, Fujiwara H (2006) Development of microtremor survey observation system. Japan Geoscience Union Meeting, Makuhari, Japan (in Japanese), pp 14–18. http://ci.nii.ac.jp/naid/10018357183
- Suzuki W, Iwata T (2006) Source model of the 2005 west off Fukuoka Prefecture earthquake estimated from the empirical Green’s function simulation of broadband strong motions. Earth Planets Space 58(1):99–104. doi:10.1186/BF03351921 View ArticleGoogle Scholar
- Tokunaga M, Sogabe M, Tanimura H, Kiyoshi O (2015) Method for evaluating equivalent natural period with microtremor measurement. J Struct Eng Earthq Eng (Jpn Engl Abstr) JSCE 71(1):72–86. doi:10.2208/jscejseee.71.72 Google Scholar
- Uehara H, Kusakabe T (2004) Observation of strong earthquake motion by National Institute for Land and Infrastructure Management. J Jpn Assoc Earthq Eng 4(3):90–96. doi:10.5610/jaee.4.3_90 Google Scholar
- Yasuda S, Yamaguchi I (1985) Dynamic soil properties of undisturbed samples. In: Proceedings of the 20th annual Japan geotechnical society (JGS) conference (in Japanese), pp 539–542Google Scholar
- Yoshida N, Kobayashi S, Suetomi I, Miura K (2002) Equivalent linear method considering frequency dependent characteristics of stiffness and damping. Soil Dyn Earthq Eng 22(3):205–222. doi:10.1016/S0267-7261(02)00011-8 View ArticleGoogle Scholar