To assist the advancement of tsunami risk assessment methods for nuclear power plants, requirements for the generation of artificial tsunami waveforms crucial for close cooperation between the probabilistic tsunami hazard analysis and tsunami fragility analysis were first listed. Then, basic studies were conducted to model the phase and amplitude spectra of waveforms observed during the 2011 Tohoku earthquake tsunami. Consequently, the average values of the group delay times and amplitude spectrum for each period band were confirmed to be effective for each modeling. Finally, in accordance with these requirements and results of the basic studies, a technique for generating artificial tsunami waveforms using a statistical method was proposed.

The SMAC-M was a strong motion instrument that was developed in the 1970s and was active as a main instrument until the mid-1990s. The SMAC-M instruments have collected valuable strong motion records, including the strong motion record of the 1978 Miyagi-ken-oki Earthquake. Recently, a large number of records of the SMAC-M instruments stored at the Building Research Institute were re-digitized by a unified method. In this paper, we described the method of re-digitization, using the strong-motion data recorded at Tohoku University during the 1978 Miyagi-ken-oki Earthquake as an example. Furthermore, we proposed a method for detecting and eliminating the time-axis deviation of records obtained by multiple seismographs in the same building and verified its usefulness.

The Kumamoto Earthquake (Mj = 7.3) on 16^th April, 2016 caused damage to civil engineering structures and buildings in the epicentral region due to two strong seismic motions and liquefaction. Road embankments near Mashiki Town suffered subsidence of the backfill after the earthquake in a wide area and the subsequent level difference behind abutments caused severe disruption to post-earthquake traffic. In this study, field investigation and numerical analysis were conducted to examine the reality and causes of the damage, targeting Kiyama River flowing through Mashiki Town in view of the conspicuous level difference damage behind abutments. As a result, it was found that subsidence of the road embankment observed at the survey sites was mainly caused by liquefaction of the sandy ground under the road embankment and consolidation subsidence of the embankment due to dispersion of the excess pore water pressure. Moreover, the survey results in the longitudinal direction of the road embankment indicated that although the level difference generated behind abutments was caused by subsidence of the backfill near the abutments, the impact on road function was small over the entire banking section unless there was a factor causing a significant change of the amount of subsidence. Using a series of analysis results, issues for the aseismic design of a road embankment in anticipation of a Level 2 earthquake were identified and a proposal was made for an evaluation method to estimate the amount of subsidence in the longitudinal direction of a road embankment to be laid above liquefiable ground. The applicability of this method was then verified in comparison with actual examples of earthquake damage.

Modeling methods to obtain frequency-independent stress-strain relationships for complex modulus used in seismic response analysis of ground are discussed. In addition, a new complex modulus for the seismic response analysis of ground, named the YAS (Yoshida-Adachi-Sorokin) model, is proposed, which is designed so that the maximum stress and the hysteretic energy absorption agree with those of the cyclic shear test. It is shown that the complex modulus, in which the maximum stress and the damping ratio agree with those of the test result, is possible only when the damping ratio is less than 0.5. Then, three complex moduli, the Sorokin model used in the original SHAKE, the Lysmer model that is proposed to improve Sorokin model and is used widely in the equivalent linear method, and the YAS model, are compared and discussed. Obtained conclusions are as follows. The Sorokin model overestimates the maximum shear stress. The Lysmer model gives the same maximum stress as the cyclic shear test, but it underestimates the hysteretic energy absorption. Underestimation of energy absorption is less than 5 % for damping ratio less than 0.3 that are important damping ratio in practical use.

Five great interplate earthquakes such as the 1946 Showa-Nankai earthquake, the 1944 Showa-Tonankai earthquake, the 1854 Ansei-Nankai earthquake, the 1854 Ansei-Tokai earthquake and the 1707 Hoei earthquake have occurred along the Nankai trough over the past 300 years. In this study, we try to reproduce the tsunami trace heights of the five great earthquakes using characterized fault models based on the tsunami recipe (HERP, 2017)⁸⁾. Tsunami simulation has been carried out for the characterized fault models whose slip heterogeneities were expressed by two different slip areas, and the calculated maximum tsunami heights (T.P.) and the tsunami trace heights were compared along the coastline. The geometric average K, geometric standard deviation κ (Aida, 1978)²⁴⁾ and the residual sum of squares between the calculated maximum tsunami heights and the tsunami trace heights were used to evaluate the reproducibility of the models. We have obtained characterized fault models to reasonably explain the tsunami trace heights of the five great earthquakes, which indicate the effectiveness of the tsunami recipe (HERP, 2017)⁸⁾and probabilistic tsunami hazard assessment based on characterized fault models.

We propose a probabilistic model for evaluating the damage probability of a structure caused by liquefaction using the maximum acceleration of the ground surface as an index. This model is applied to the fragility curve, which evaluates the probability of function loss of roadbed and pavement structures that are directly affected by ground deformation. Parameters such as uncertainty are assumed from past studies on liquefaction and ground amplification. We evaluate the function loss probabilities due to liquefaction of roadbed and pavement structures such as runways and taxiways at Haneda Airport due to Level 1 and Level 2 earthquakes. In addition, the function loss probabilities before and after the liquefaction countermeasures are compared, and the effect of the countermeasures is considered from the viewpoint of the availability of the runway.

In Japan, large-scale earthquake disasters occur frequently in which serious damage is concentrated on vulnerable people. Based on the situation of past earthquake disasters, each municipality has created a list of vulnerable people's information. However, there are various issues such as high hurdles for listing due to personal information. In this study, we widely regard patients with orthopedic diseases as vulnerable people, and examine regional distribution using national health insurance data. In addition, alternate index of the distance to the nearest evacuation site is created for each town using GIS network analysis. Through this research, it became possible to clarify regional differences from the two viewpoints of number of orthopedic disease patients and the distance to evacuation sites.

Many grabens, which are special and strange, belt-like subsidences, formed in the caldera of the Aso Volcano during the 2016 Kumamoto Earthquake. These grabens caused severe damage to houses, roads, water and sewage pipes, rice fields, and river dikes. The authors conducted various site investigations, soil tests and analyses to determine the mechanism of the grabens. The measurement of the displacement by SAR from satellites showed horizontal displacements of about 2 to 3 m. In this area, an old lake was formed approximately 9,000 years ago. Borings and microtremor measurements demonstrated a special cohesive soil layer with diatom and pumice sediment in the lake, and the bottom of the lake is U-shaped. Soil tests on the lake sediment soil clarified that the shear modulus decreases drastically under cyclic loadings, similar to the liquefaction of loose sand. The authors conducted a seismic response analysis and a residual deformation analysis and concluded that the lake sediment soil deformed to wrap around the U-shaped bottom of the lake due to the sharp decrease in shear stiffness during the earthquake, resulting in horizontal tensile stress on the surface soil and the grabens near the edge of the lake.

Underground structures may suffer significant damage subjected to the fault rupture. This research clarified the failure mechanisms of RC box culvert subjected to fault displacement based on 2D FE analysis and fragility analysis was performed in terms of variation on concrete and soil strength, as well as soil stiffness. Based on the analytical results, it is found that the failure probability of the RC box culvert is significantly affected by the soil stiffness variation, rather than the strength variation of materials under the low-angle fault. On the other hand, the failure probability of the RC box culvert is affected by not only the soil stiffness variation but also the material strength variation under the high-angle fault.

In this study, we conducted building damage analysis using the damage survey data by the Uki City government, Kumamoto Prefecture due to the 2016 Kumamoto Earthquake, and constructed building fragility curves in combination with the estimated earthquake motion distribution. As the result of analysis on the tendency of damage ratios with respect to the structural material, construction period, roof types, and number of floors for wooden buildings, it was found that the major damage ratio of wooden buildings was larger than that of non-wooden buildings. The major damage ratio of wooden buildings in Uki City was significantly smaller than that of Mashiki Town for the same construction period and the same seismic motion level. The use of the both data from Uki City and Mashiki Town is suggested to establish unified fragility curves to cover a wide range of seismic motion.