In the Kanto district, Metropolitan Seismic Observation network (MeSO-net) of highly dense seismic network has accumulated continuous waveform data since 2008. It is expected that higher resolution of three-dimensional attenuation structures can be obtained by using these high-density observation records. However, seismometers of MeSO-net is installed at a depth of 20 m and there is a problem to handle each waveform as a record at a free surface. First, we examined characteristics of the spectrum of waveform records and it was found that troughs in the spectrum were formed due to underground installation. Next, we carried out simultaneous inversion of site amplification and three dimensional Q using MeSO-net, K-NET and KiK-net records. The stations are classified into eight groups based on the predominant period calculated from logging data for MeSO-net, K-NET and KiK-net stations were divided into two groups based on the predominant period estimated from AVS30 by J-SHIS. If the amplifications at all the observation points are added to unknown parameters, the trade-off will be caused with the attenuation structure. We grouped the stations to avoid this problem and assumed that the same group has a same amplification. As a result, the obtained attenuation structure was consistent with previous studies. It was confirmed that the effect of the underground installation could be obtained by properly considering the amplification factor. In addition, it was found that the amplification of the MeSO-net record is close to one at the free surface of hard rock site.
A reliable estimation of surface fault displacements is necessary for the safety of infrastructures and buildings. It is essential to consider various uncertainties in the fault rupture process for the reliable estimation. In this study, the effect of material uncertainties on fault displacements is studied by performing repetitive simulations using Latin hypercube sampling. The convergence of probabilistic responses versus the number of samples is analyzed by simulations with multiple sample sets. It is shown that the critical base slip that makes the rupture reach the ground surface are strongly affected by material uncertainties. The probabilistic distribution of the critical base slip can be approximated by the lognormal distribution well when the material properties are assumed to be log normal random variables. The mean and the standard deviation were estimated with a few cm errors from results of 230 simulations by conducting ten simulations when the material properties have coefficients of variance of less than 0.3.
High-densed microtremor measurements are made in Ikonobe area, Yokohama, in which a differential settlement between buildings occurred in 2014 and wooden houses were severely damaged during the 1923 Kanto earthquake. An empirical relationship between the predominant period of H/V spectrum (Tp) and subsurface-soil thickness (D) is obtained from microtremor measurements conducted at bore-hole sites. The estimated D is averagely 10-15m with the maximum depth of 20 m in the northern edge of the lowland along the northern plateau. The differential settlement spot in 2014 is located in the drastically D-changed area. Strong-motion spectral ratio (As) is estimated from the peak value of H/V spectrum (Ap). It is found that As median of Kouchi and Kawamukoh hamlets on Holocene deposits with wooden-house collapsed ratio over 30% is larger than 5 whereas that of Yabune hamlet mainly on Pleistocene deposits with the collapsed ratio lower than 10% is smaller than 3 suggesting that ground-motion amplification of soft subsurface soil deposits worsen the housing damage. However, the cause of the especially-high damage ratio of Kouchi hamlet cannot be explained in terms of ground-motion amplification.
In recent earthquakes, slope failure occurred with muddy collapse and flow at gentle slopes made of volcanic sandy soil. It seemed that the volcanic sand would be liquefied even in unsaturated condition in these damages. In this study, the unsaturated tri-axial tests using two kinds of volcanic sandy soils were conducted under the different initial conditions of suction and degree of saturation. Furthermore, numerical simulations were also performed in order to reproduce the experiments and make clear the factors which affect liquefaction strength in unsaturated condition. The experiments showed that the cyclic behaviors varied largely depending on kinds of volcanic soils and the unsaturated conditions determined by the values of suction and the degree of saturation. The numerical simulation which considers three phase (soil skeleton, pore water and pore air) reproduced the tendency of experiments. Furthermore, the parameteric study showed that the cyclic behavior of volcanic soils depend on the soil water characteristic curves.
Site amplification factors of ground motions are significantly large from the engineering bedrock to the ground surface in the Osaka plain, Japan. In the western Osaka region, including the Osaka Bay area, shallow soil layer is thick, while in the eastern Osaka region, high sensitivity clay layers are widely distributed. In this paper, we constructed a shallow soil model, including reverse layers, in the eastern Osaka region for the three-dimensional subsurface structure model. The proposed model, which considered both regions shallow portions, was verified by simulation analysis of small earthquake records using the 3-D finite difference method. Ground motions because of the hypothetical point source were examined by focusing on the coupling effects of shallow and deep subsurface structural influences in the Osaka plain, along with the effects of the reversed S-wave profile and one-dimensional ground amplification of shallow layers.
The procedure of evaluating fault parameters, so called Recipe, for predicting strong ground motions by the HERP [Headquarters for Earthquake Research Promotion, 2005], Japan, is often applied to assessment of design input earthquake motions for nuclear power plants in Japan. Due to this situation, the International Atomic Energy Agency  adopted this Recipe in Safety Report Series No. 85. However, this Recipe was validated only for Japanese earthquakes. Hence, we examined the validation of the Recipe by comparing the estimated strong ground motions based on the Recipe with those by the GMPE (Ground Motion Prediction Equation) obtained by Akkar et al.  from the Europe and Middle East crustal earthquakes. We concluded that the short-period level in the Recipe should be reduced to 50% for the second-stage earthquakes and that the stress drop equation of a circular crack should be altered. We also concluded that the short-period level should be reduced to 55% for the first-stage earthquakes under the condition that the stress drop equation of a circular crack should be maintained.
The 19 September 1985 Michoacan, Mexico earthquake (MW 8.0) and the 16 September 2015 Illapel, Chile earthquake (MW 8.3) were mega thrust earthquakes occurred under continental areas, and a number of seismic records were observed above the source regions.
We estimated source spectra, Q value and site amplification factors by spectral inversion analysis. Based on the source spectra, we estimated the short-period amplitude level and the stress drop.
The stress drop slightly showed focal depth dependency although the variation was large. The estimated short-period level of the 1985 Mexico earthquake is half of the level obtained from the conventional study, whereas that of the 2015 Chile earthquake corresponds to the average level.
For the purpose of evaluation and analysis of earthquake motion characteristics, the method of characteristics extraction for earthquake motion developed by the authors was applied to the time-frequency analysis of earthquake motion. For analysis of earthquake motion, the Gaussian mixture model was applied to the 99-dimensional feature vector representing the temporal characteristics of a strong motion on the basis of its Husid plot. In this study, the period-dependent feature vector based on the evolutionary power spectrum was defined. Furthermore, the Gaussian mixture model was applied to the period-dependent feature vector. The proposed method was applied to the earthquake motions observed from the Tokaido-oki earthquake that occurred at 23:57 September 5, 2004. The period-dependent feature vector and the Gaussian mixture model by proposed method were included of the characteristics of earthquake motion. For the dispersibility of surface wave found in the feature vector, the element model from the Gaussian mixture model was compared with the dispersion curve calculated from the soil profile.
Micro-tremor array data with limited wavelength range or inversion analysis without reference soil profile calls for a method to infer a simpler subsurface model. We have proposed the method to deduce surface layer thickness and shear wave velocity of a two-layered model by resorting to the quarter wavelength theory and the concept of wavelength-thickness ratio. The method was effective in numerical tests with PS logging data at K-NET and KiK-net observation stations with relevant wavelength-thickness ratio. The effectiveness of the method for observed dispersion curves is tested in this report.
During the 2016 Kumamoto earthquake, surface rupture of seismic faults occurred over a wide area. Simultaneously, large-amplitude pulse waves were recorded in the vicinity of the seismic faults. Herein, we present a theoretical evaluation of the near-fault seismic ground motions, which requires a fine division of the fault plane into many sub-faults. Numerical convergence of the spatial integration is examined using the Thin Layer Method (TLM). We found that permanent displacement and dynamic velocity converge using an integration spacing equal to approximately half of the fault shortest distance by comparison with rigorous methods. Finally, we applied the seismic fault model to the main shock of the 2016 Kumamoto earthquake to investigate the influence of differently divided sub-faults in the fault plane on the evaluation of strong ground motions adjacent to the seismic fault.
In order to predict tsunami height by submarine landslide due to earthquake, it is important to verify the evaluation method for submarine slope stability during earthquake. Finite difference method adopted two constitutive models was used to simulate the centrifuge model test results. The effect of water depth and crack on the top of slope to the slope failure was studied. The applicability of the constitutive models was also evaluated. Consequently, seismic coefficient at onset of failure were almost same irrespective of water depth and failure modes of slope with and without crack were almost same in the simulation. Simulation result using Mohr-Coulomb model with strength parameter considering shear strength under over-consolidated and low confining pressure condition was consistent with test result on seismic coefficient at failure and position of failure surface.
In this study, we conducted microtremor array observations and single-point-3-components observations to estimate the S-wave velocity structure of the Obama plain. Rayleigh wave phase velocities at each array site were calculated from the spatial autocorrelation method. The phase velocities at each site were inverted to a 1D S-wave profile using a genetic inversion. The subsurface structures down to the Tertiary bedrock were estimated as inversion problem based on the H/V spectra. The 1D S-wave velocity structures around the plain were interpolated into 3D structure by using Kriging techniques. The validity of the estimated structure from the microtremor observation was confirmed by comparing with the seismic prospecting and with the boring log data.
The 2011 Tohoku earthquake is an earthquake of the first M9 that many of the observed strong motion records are obtained. It is necessary to reflect the observation records in the earthquake resistant design of a building. This paper describes, first, quantitative investigation of the site-dependent strength demand using the earthquake observation records during the 2011 Tohoku earthquake in Sendai city. Then, ductility demand and seismic performance demand are quantitatively investigated for observed strong motion records with dominant period of 1.0(s). It is suggested that the investigated study is a basis for earthquake resistant design considering the ground motion characteristics, especially for a design to consider resonance of a building structure including moving resonance.
In order to verify seismic performance of arch dam, it is necessary to evaluate stresses within dam during earthquake. When evaluating against strong earthquake motion, it is necessary to set the values and to consider the nonlinearities of dynamic deformation properties. Here, we examined the effects of reduction of dynamic shear modulus on earthquake stresses of dam by 3-D dynamic analysis. As a result, the tensile stresses during earthquake generally showed a tendency to decrease as the value of dynamic shear modulus decreased, but the tensile stresses around the connection area between dam and thrust block increased. Since the increase of tensile stress during earthquake was shown in the part of dam body and abutment, it is necessary to pay attention to the stress increases during earthquake in the seismic performance evaluation.
Slope protection works are often constructed on railway embankment for the purpose of the prevention of seepage and erosion to rainfall. Thought the works sometimes cause increase of load acting to embankments, it is possible that the works contribute to the increase of seismic performance by keeping the embankment in unsaturated condition and connection with soil nails. However, these effects have not been studied clearly. In this paper, series of shaking table tests of model embankments with slope protection works and soil nails were conducted, with rainfall infiltration.
In this study, authors proposed a method to convert two artificial seismic waves which are both compatible with Japanese Seismic Code into a pair of waves which are compatible with bi-axial response spectrum. Secondly, in order to evaluate the seismic behavior of 20-story base isolated building, three-dimensional seismic response analysis was carried out by bi-axial response spectrum-compatible accelerograms with different acceleration trajectories. As a result, there found significant differences in the part of response values between uni-direction input and bi-direction inputs. It was shown that seismic performance evaluation considering actual behavior was possible by conducting analysis by horizontal bi-direction inputs. In addition, it was shown that installed cylinder-type dampers in two orthogonal directions is an effective method for seismic design against for uncertain bi-directional input.
Designers currently evaluate the seismic performance of civil engineering structures in sewage facilities by the seismic coefficient method. However, dynamic interaction between the structure and the ground should be properly considered for the sewage facilities at the soft ground. This study investigated the seismic behavior of it using the nonlinear finite element analysis in consideration of dynamic interaction. Appropriate seismic performance evaluation technique for civil engineering structures in sewage facilities was proposed.
In order to restore the electric power distribution system quickly after the huge earthquake, it is important to estimate not only the severe damage but also the minor damage which disturb the quick restoration. In this paper, a method of evaluating a seismic capacity of pole-mounted electric power distribution equipment using shaking table tests is established. In this method, the seismic capacity is defined as a damaged state considering with the restoration efforts and the damaged state is related with seismic intensity. Conducting the shaking table tests using real equipment, we can evaluate the seismic capacity of equipment virtually.