Aiming to acquire new knowledge by seismic motion evaluation from a new viewpoint, it was tried to create a site-specific earthquake ground motion evaluation model by machine learning using the earthquake ground motion records obtained in the past as learning data. The epicentral direction and the response duration time of the earthquake ground motion, which were not handled by the conventional attenuation relations formula, were also examined. As a whole, the observed values are well evaluated and modeled, within double to half of the observed values. The degree of influence of the epicentral direction on the response duration time is large, and in many cases, it is almost equal to or greater than each parameter of the conventional ground motion prediction formula.

Nonlinear response of subsurface structures during strong shakings has a large impact on results of strong motion predictions, but proper nonlinear parameters of soils should be selected and arranged to prediction models based on limited given information when we reproduce the nonlinear behavior theoretically. We propose a method to estimate an incident wave at the seismological bedrock and a horizontal amplification factor during strong shaking by a linear spectral analysis without nonlinear parameters, assuming the diffuse field concept for earthquake and the linearity of vertical amplification hold during a single strong earthquake. We first verified the proposed method by using weak earthquake motions. Second we investigated the applicability of the proposed method to strong motions recorded at several sites in Japan, and compared the estimated results by the proposed method with those estimated by the equivalent linear analysis which uses nonlinear parameters. In some cases, the diffuse field concept did not hold when we used strong motions started from the S-wave arrival time. However, when we used strong motions started from a later half of S-wave, horizontal-to-vertical spectral ratios and surface-to-borehole spectral ratios showed similar nonlinearity, especially in the high frequency range to those of strong motions started from the S-wave arrival, and the diffuse field concept is considered to be held in the later S-wave part. The seismic bedrock spectra estimated by the proposed method showed good agreements to those estimated by the equivalent linear analysis. We reproduced the strong motions observed on the ground surface by the equivalent linear analysis with the incident waves at the seismological bedrock which were made from the estimated spectral amplitudes by the proposed method and the spectral phases of the observed motions on the surface.

The 2011 off the Pacific coast of Tohoku Earthquake and the 2016 Kumamoto Earthquake have increased the interest in the seismic hazard assessment. As new knowledge has been revealed with every new major earthquake, it is important to consider the range of the uncertainty of the parameters which are needed to the assessment and to understand the impact of each earthquake by probabilistic seismic hazard analysis, in advance. The Ikata SSHAC project is the first attempt in Japan to apply SSHAC Level 3 to the probabilistic seismic hazard analysis. In this paper, we report on the results of the modeling of various seismic sources, which include the Median Tectonic Line fault zone which is a long active fault and trench earthquakes occurring in the Nankai Trough. In addition, we discuss the effectiveness of guidelines for SSHAC Level 3, the range of the uncertainty and on the impact on the seismic hazard, because the gained knowledge and know-how of the scope would be useful for subsequent studies.

In overseas countries, probabilistic risk assessment of nuclear power plant is based on probabilistic seismic hazard analysis (PSHA) applying the SSHAC (Senior Seismic Hazard Analysis Committee) Level 3. The Ikata SSHAC project is the first attempt in Japan to apply the SSHAC Level 3 seismic hazard analysis to a nuclear power plant, the Ikata Power Plant Unit 3. The characteristics of the Ikata site is that the Median Tectonic Line fault zone is located near the site and the hard rock site. Therefore, in order to take into account the epistemic uncertainty of the seismic motion, we introduced both ground motion prediction equation (GMPE) and fault rupture model simulation for evaluating ground motions, and analyzed the variations in ground motion near the seismic source. The results of this study are significant from the viewpoint of accurate and objective evaluation of uncertainty, and should be extended to subsequent PSHA.

We conducted a workshop to identify potential issues that could be caused by news media’s dissemination of the Nankai trough earthquake information (extra). The Nankai trough earthquake information (extra) was introduced in 2017 by the Japanese government, which will be issued when anomalous phenomena that suggest the occurrence of devastating earthquake along the Nankai trough are observed. However, the information could not state the specific date, location, and magnitude of the impending earthquake because of the limitation of earthquake science. We held a multi-stakeholder workshop to discuss how this ambiguous scientific information could be delivered and disseminated by news media to promote risk reduction as well as to avoid social disruption. Local news media including television, radio, and newspaper from four separate regions in Japan participated in the workshop along with researchers on seismology and disaster information studies. A “partial-area-rupture case” where a large-scale (magnitude 8 class) earthquake occurs on the western side of the Nankai trough was discussed. A scenario that describes the issuance of various warning and disaster information including the Nankai trough earthquake information was presented along the timeline, which then the participants discussed how the news media would respond to the sequence. Although it was a short workshop and the scenario was quite crude, several unrecognized issues were successfully found.