From the perspective of establishing or verifying design-based input ground motions or ground motion prediction equations, it is important to estimate the hypothetical bedrock outcrop waves (bedrock ground motions) from ground motions observed in boreholes. In this study, we conducted numerous spectral ratio Qs inversions at a crystalline bedrock site with vertical array seismic observations down to the seismological bedrock at a depth of 2000 m. Seismometers were installed at depths of 5, 160, 500, and 2000 m. This allowed us to determine the frequency-dependent median spectral ratios of the bedrock ground motions at the surface relative to those at the seismological bedrock. We identified the subsurface structure model that explains this median curve well across broadband frequencies as the best estimate. Using the model, we read empirical amplification factor characteristics in 15 seismic records by removing the effects of shallower layers at depths of 5 m and 2000 m. The results showed no significant amplification from the seismological bedrock to the surface, and the dominant attenuation above 1 Hz. Additionally, the theoretical transfer function used to estimate bedrock ground motions at the seismological bedrock from observed seismic records exhibits the smallest uncertainty across broadband frequencies at a depth of 2000 m, whereas the smallest uncertainty below 4 Hz is estimated at a depth of 160 m.

This paper aims to perform real-time building damage estimation during earthquakes using a basic survey of city planning. This study verified the estimation accuracy of building damage by comparing actual damage and estimated damage for each building caused by the 2016 Kumamoto Earthquake in Mashiki Town. First, an analytical model for steel buildings is set, and damage estimation results are obtained using incremental dynamic analysis. Next, the damage estimates for all buildings within the target area are integrated with the estimated results of wooden, reinforced concrete, and steel buildings. As a result, the estimated damage distribution, while varying in severity, generally represents the actual damage distribution in certain areas. Building-level estimation is effective when the extent of damage varies within the same regional mesh. Because estimation at the regional mesh level using fragility curves cannot estimate this trend. Furthermore, this study estimated the distribution of road blockage based on the estimated building damage distribution and the actual damage distribution and compared it with aerial photographs taken immediately after the earthquake. The estimated road blockage distribution based on the estimated building damage matched the aerial photograph situation 79.6%, and the ratio of road blockage that was falsely estimated as passable was 11.3%. Factors contributing to misestimation include: overestimation of damage, failure of the debris flow area setting to accurately reproduce the actual phenomenon, and inability to reproduce blockages caused by structures not included in the estimation, concrete block walls, and retaining walls.

The 1983 Nihonkai-Chubu Earthquake (M7.7) marked its 40th anniversary in 2023. In Akita City, this earthquake caused three fatalities and damage to residential buildings, including 35 completely destroyed houses and 235 partially damaged houses. Akita City has also experienced earthquake damage from several past events, including the 1896 Rikuu Earthquake, the 1914 Akita Senboku Earthquake, and the 1964 Niigata Earthquake. This study focuses on the damage in Akita City caused by these four earthquakes. The characteristics and spatial distribution of the damage are compiled and examined, and the damage distribution is compared with historical topographic maps from the corresponding periods. Through this analysis, the historical record of earthquake damage in Akita City is outlined. Records of past earthquake damage and their spatial distribution constitute essential baseline data for urban disaster mitigation planning. Based on these results, the relationship between the historical earthquake damage record and the process of urban development in Akita City is discussed.

This paper investigates the mechanism by which significant lateral spreading occurs in flat ground due to liquefaction, using centrifuge model tests. Under conditions where a slope adjoins flat ground and a low-permeability surface layer exists above the liquefied layer, the excess pore water pressure varies even if the bottom of the surface layer is horizontal, excess pore water pressure varies due to differences in initial effective overburden pressure, resulting in a hydraulic gradient within the liquefied layer beneath the slope. After the end of shaking, the distribution pattern gradually changed so that the excess pore water pressure would equalize. As a result, water pressure increased beneath the toe of the slope and beneath the flat area, causing the excess pore water pressure ratio to exceed 1 by a significant margin. In the experiment designed to induce lateral spreading after the end of shaking, while horizontal displacement on the slope was small, horizontal displacement continued on the flat ground, resulting in lateral spreading. Observation of colored sand columns pre-installed in the ground confirmed that the lateral spreading occurred only in the surface layer of the flat ground. Based on the experimental results, we have proposed a mechanism for lateral spreading caused by liquefaction. After an earthquake, a water film forms beneath the low-permeability layer. Driven by the hydraulic gradient, the water in this film flows from beneath the slope to beneath the flat ground, causing the surface soil above the water film on the flat ground to flow laterally.

The 2016 Kumamoto earthquake resulted in significant casualties and building damage. To support safe and rapid assessment of building damage, this study develops a deep learning-based model for detecting completely collapsed buildings. The proposed approach uses a digital surface model (DSM) obtained from airborne LiDAR surveying together with post-earthquake aerial images. The model was trained using two different datasets. The results showed that the model achieved higher performance when using four-band images that combine the difference between pre- and post-earthquake DSM with post-earthquake aerial imagery. Based on this dataset, hyperparameter optimization using Optuna was further applied to improve the detection performance. The final model achieved an accuracy of approximately 81% and a recall of approximately 63%.

This study developed a building vulnerability curve for partial damage to wooden buildings in order to improve the accuracy of economic seismic risk assessment. In constructing the building fragility curve, municipal-level damage ratio data published by the General Insurance Rating Organization of Japan were utilized. Specifically, the building fragility curve for “minor damage”, based on disaster certificates, was adjusted using the municipal-level damage ratio data. Using the vulnerability curve for minor damage, building losses from the Tokyo Metropolitan Earthquake were calculated to be approximately ￥22.2 trillion. Of this total, building losses attributable to minor damage were estimated to exceed ￥6.7 trillion. These results demonstrate that even minor building damage significantly impacts the monetary value of building loss risk.

Fault displacements can cause structural damage and loss of functionality. The conventional framework for probabilistic fault displacement hazard analysis (PFDHA) estimates the characteristics of distributed ruptures as a function of distance from the principal fault. Recently, revised fault displacement prediction equations (FDPEs) have been proposed as a means to capture the attenuation characteristics of fault displacement in the far-field as a function of distance from the subsurface earthquake source faults, without classifying surface ruptures as principal or distributed. These FDPEs enable us to address many of the challenges identified during the first practical application of PFDHA in Japan, which was conducted at the Ikata site in northwestern Shikoku. In this study, we conducted a case study of PFDHA based on distance from the earthquake source faults at the Ikata site. Based on the results, we report the characteristics of the new FDPEs and insights gained from their practical application at the Ikata site.

We analyzed the residuals between the strong-motion data flat-file compiled from K-NET and KiK-net data (NIED, 2025), and a ground motion model (Morikawa and Fujiwara, 2013). We demonstrate that the generalized Pareto distribution, used in extreme-value theory, is more suitable than the conventional log-normal distribution for modeling the tail of the ground-motion intensity distribution, which strongly influences PSHA results at low exceedance probabilities.