The author summarize the attenuation characteristics of strong-motion records during the 2011 Tohoku earthquake and discuss the applicability of Ground Motion Prediction Equations (GMPEs) to mega earthquakes. The major results are as follows: 1) GMPEs adopting linear Mw term tend to overestimate when applying to mega earthquakes. The source term must follow the scaling of omega-square model. 2) Some of the GMPEs using fault distance also need to revise their saturation rate of amplitude when evaluating mega earthquakes. 3) The different definition of source-to-site distance can cause not only the different attenuation characteristics but also the different source-term scaling in mega earthquakes, because the extensive fault of mega earthquake affect even in the distances farther than usually expected.

View full abstract

Many source models for the 2011 Tohoku earthquake are derived by dense observation of teleseismic, strong motion, geodetic, and tsunami open data. We review similarity and difference of source models from strong motion records. Long-period source models in a period range of longer than around 10 s inverted large slip near the hypocenter and along the Japan Trench. On the other hand, short-period source models in a period range of shorter than around 10 s provided several strong motion generation areas or strong-motion pulse generation areas with large stress drop near the hypocenter and along the down-dip edge of the fault. Both source models are partly overlapped in the Miyagi-oki region, and complementary in the Fukushima-oki and Ibaraki-oki regions. Source models from back projection, array analyses, envelope inversion, and seismic intensity inversion generally agreed with the above mentioned source characteristics. Alike the 2010 Maule, Chile, earthquake, the 2011 Tohoku earthquake could be consist of M9-calss long-period and M8-class short-period source models. However, new findings of complex rupture along dip direction and multiple rupture over the fault indicate the diversity of earthquake source. The resolving power of strong motion dataset along the Japan Trench is not enough for the 2011 Tohoku earthquake, and further offshore strong motion observation is highly expected.

View full abstract

This paper reviews studies aiming to understand the relation between the large acceleration observed during the 2011 Off the Pacific Coast of Tohoku Earthquake and the effect of surface geology at strong motion stations K-NET Tsukidate and KiK-net Haga where high acceleration and high seismic intensity was observed, as well as K-NET Furukawa and JMA Furukawa where heavy building damage had occurred through analysis of observed ground motion and microtremor data from temporary observations and numerical simulations. The effect of partial uplifting of the foundation of the seismometer which is claimed to have occurred at K-NET Tsukidate during the mainshock was also investigated. By considering the amplification of the surface geology, the effect of deep underground structure, the effect of heterogeneity and topological effects were taken in account of in order to explain the observed data. Despite using different approaches to determine the velocity structure at the target location, the results seems to become a common structure, but there are still not able to get a perfectly matched structure. Also, the fact that the spatial variation of ground motions are strongly affected by the heterogeneity of surface geology and deep structures were again indicated by the intense observations of strong motion and microtremors.

View full abstract

To understand the generating factor of strong ground motions for a period of 2 to 3 seconds observed in the western part of Tokyo Bay during the 2011 off the Pacific coast of Tohoku earthquake, we estimated shallow and deep S-wave velocity structures at 6 sites near the seismic stations by array microtremors observations. As a result, the existence of soft-soil sediments for Vs 150m/s to a depth of more than 50m and Vs 400m/s to a depth of 200m was confirmed in the JKPM (Jonanjima Seaside park) and UK1222 (Ukishimacho park) stations locating in reclaimed lands. Also, site amplification factors at 6 sites calculated using the estimated S-wave velocity structures were in good agreements with those observed using the seismic data of the main shock. In this study, therefore, we concluded that the strong ground motions for a period of 2 to 3 seconds, which were observed in the western part of Tokyo Bay during the 2011 off the Pacific coast of Tohoku earthquake, were due to the site amplifications of S-waves excited by deep underground structures to a depth of more than 200m.

View full abstract

We investigated the causes of large ground motions with the seismic intensity scale of 7 at K-NET Tsukidate and KiK-net Haga during the Tohoku earthquake. We showed that the vibrations pointed by Motosaka et al. as partial uplifting and slipping of the foundation of the seismometer and the nonlinearity of soil amplified 10 times at 4Hz for NS component and 15 times at 8Hz for vertical component at Tsukidate. We also pointed out that similar vibrations which are not ground motions amplified twice at 2 Hz for horizontal components and 10 times at 4 Hz for vertical component at Haga. The seismic intensity scale calculated from ground motions without such vibrations was estimated to be 6 upper at Haga by the equivalent linear analysis.

View full abstract

In Niigata-ken Chuetsu Area, the new seismic observation network consisting of 40 stations was established for the study of seismicity around the fault zone of western margin of the Nagaoka plain. The characteristics of the records observed in this network during the 2011 off the Pacific Coast of Tohoku Earthquake were studied. The several significant pulses were recognized in the acceleration records in Miyagi and Fukushima prefecture but there were no significant pulses in acceleration data of the network. The values of PGA were 10-20 cm/s² in a hill side and 30-40 cm/s² in the plain side. The acceleration Fourier spectra of the records were poor in higher frequency than 1 Hz. The spectrum amplitude in frequency of 0.1-2 Hz in plain side showed approximately five times larger than those in hill side. The spectrum amplitude in lower frequency than 0.1 Hz was almost the same in both sides. The common spectrum peaks were recognized at 0.04, 0.06 and 0.08 Hz in UD- and EW- component and 0.02 Hz in NS-component. Two clear pulses with period of about 20 seconds were recognized with the interval of about 50 s in the velocity waveforms of EW and UD. The values of PGA and PGV were large in plain side and small in hill side. Spatial distributions of PGA and PGV showed correlation to the topography and relative amplification factor to the reference station showed correlation to AVS30 those were calculated from PS-logging data. The relative amplification factor of the acceleration responses with 5 % damping at period of 5 and 10 seconds showed correlation to the depth of seismic bedrock not to topography. This also suggests that the effects of the thick sedimentary layers in this area on long-period ground motion amplification.

View full abstract

With The 1995 Kobe Earthquake as a turning point, social requests for ensuring safety of infrastructures at the time of large-scale earthquakes have become very high, so research and study on safety evaluation of various infrastructures against the largest-class earthquake (shown as ‘level 2 earthquake’ below), which is estimated from the current to the future at the relevant point, are conducted in many fields energetically. Furthermore, in order to evaluate level 2 earthquake at dam-site, attenuation equations based on acceleration records observed at dam foundation have been proposed.  In this paper, we report an outline of the latest attenuation relationships considering earthquake motion at The 2011 Tohoku Earthquake (M_W9.0), and results of an investigation of the applicability between calculated values and observed values.

View full abstract

We investigated the time-dependent change of soil response at three KiK-net vertical array sites for several years before and after the 2011 off the Pacific coast of Tohoku earthquake by means of a simple quantitative index of soil nonlinearity. The whole site showed significant nonlinear soil response during the mainshock of the 2011 Tohoku Earthquake. We focused on the soil response recovery process at these sites after the mainshock. The soil response characteristics at MYGH10 recovered to those before the mainshock faster than TCGH16, though both the sites were attacked by strong ground motions over 1 G during the mainshock. The soil response at FKSH14, where the spiky waveforms implying cyclic mobility were observed during the mainshock, also showed the slower recovery similar to TCGH16; the soil response recovery at these sites does not completed even three years after the mainshock. The difference of the soil response recovery at the three sites would be attributed to the difference of their soil characteristics.

View full abstract

To predict strong ground motions for future large earthquakes including huge subduction earthquakes, it is important to take into account the effects of soil nonlinearity. The author has been developing a simple method to simulate strong ground motions taking into account multiple nonlinear effects (e.g., Nozu and Morikawa, 2003). The method, however, has been validated only for limited amount of strong motion data, partly because there was only a limited amount of strong motion data affected by soil nonlinearity. Therefore, in this article, making use of strong motion data for the 2011 Tohoku earthquake and the source model developed for the same earthquake by the author (Nozu, 2012), strong motion simulation with considerations of multiple nonlinear effects was conducted and its effectiveness was studied. In particular, strong motion records with the evident effects of soil nonlinearity were selected and they were simulated using the source model and taking into account empirical site amplification and phase effects (Nozu et al., 2009). Soil nonlinearity was considered using the method of Nozu and Morikawa (2003). The method uses two parameters to represent the effects of soil nonlinearity; one representing the reduction of averaged shear wave velocity within the sediment (ν₁) and the other representing the increase of averaged damping factor within the sediment (ν₂). In the simulation, ν₁ was basically determined based on Wakai and Nozu (2013) and ν₂ was determined so that the observed ground motion could be simulated as accurately as possible. As a result, it was found that, the duration of strong ground motions tended to be overestimated if the parameter ν₂ was not used for the sites with the effect of soil nonlinearity, which is indicating the effects of multiple nonlinear effects. In each of the target sites, by using these two parameters, the simulation result was improved. Thus, the effectiveness of strong motion simulation with considerations of multiple nonlinear effects was confirmed. Based on the results, the application of the method for future earthquakes was also discussed.

View full abstract

The strong ground motion which has a large peak ground velocity of more than 100cm/s was observed at CCHG site in Small-Titan network during the 2011 off the Pacific coast of Tohoku earthquake. In this paper, we specified the interpretation of site amplification of CCHG site. First, S-wave velocity structure at CCHG site was estimated from the array observation of microtremors to represent the site amplification factor in case of linear soil using seismic data by small earthquakes. As a result, it is shown that the deep S-wave velocity structure to seismic bedrock is necessary to interpret the large amplification factor in CCHG site. Second, we compared the ground motion calculated by linear soil response analysis using the estimated S-wave structure to the strong ground motion observed at CCHG site during the 2011 off the Pacific coast of Tohoku earthquake. It is shown that the calculated Fourier spectrum in high frequency range (5-10Hz) is large and the dominant frequency move to the higher frequency compared to the observed spectrum. Therefore, the soil nonlinear response is also related to the origin of the observed ground motion at CCHG site.

View full abstract

During the 2011 Tohoku Earthquake, the peak ground accerelation (PGA) with high value of above 1000 cm/s² was observed at K-NET Hitachi and K-NET Hokota, located in the coastal region of Ibaraki prefecture which is close to the southern edge of the fault, and the PGA close to 1000 cm/s² was also observed at KiK-net Iwase, located in the boundary between Ibaraki and Tochigi prefecture. However, few records were observed in the area between these stations, referred to as Higashi-Ibaraki Terrace. Moreover, we have less information about subsurface structures in the area. Accordingly we performed aftershock observation of the Tohoku Earthquake and microtremor measurements to evaluate ground motion characteristics in the area. Aftershock records mostly show large accelerations in the eastern part of the terrace, but it is also large at the middle of the terrace. We then estimated the S-wave velocity profiles from microtremor explorations and we found that the shallow soil layers are deep in the middle of the terrace, while seismic bedrock is deep in the eastern part. Thus, the soil amplification has a peak at short period in the western part of the terrace, while it has a peak at about 1 second in the middle of the terrace. It has two peaks at short and long periods in the eastern part due to the shallow and deep soil structures. These amplifications showed similarity with the spectral ratios of the aftershocks.

View full abstract

The widespread liquefaction disaster that occurred in the reclaimed land around Tokyo bay area during the Great East Japan Earthquake was alarming. Validating this disaster by simulation is needed to prevent any liquefaction that may occur during likely huge future earthquakes. To this end, accurate soil properties and appropriate input wave as well as the finite deformation analysis that simulates the reappearance of liquefaction should be comprehended. In this study, soil-water coupled analysis code GEOASIA with SYS Cam-clay model, which can reproduce the function of soil structure, was used. Firstly, elasto-plastic properties of Urayasu soils were identified from the results of boring survey and laboratory testing. Then, by means of transfer method, both NS and EW upward waves at the foundation ground were simultaneously estimated from K-NET recorded motions at Urayasu.

View full abstract

We conducted the MASW and Microtremor explorations at a Port and Airport Research Institute (PARI) site at Sendai, where borehole earthquake observation have been conducted and P-wave and S-wave (PS) logging in the boreholes had been released to the public. Under the PRENOLIN Project in France we got new boring data including PS logging and soil nonlinearity. Most of the MASW lines across the sites showed a homogeneous one-dimensional (1D) structure with a soft horizontal layer down to several (6∼7) meters. Also most of the microtremor data showed quite common horizontal-to-vertical (H/V) spectral ratios except for sites close to the office building, which also suggest the existence of a 1D structure at the site. Then we conducted 1D inversion analysis to obtain the best S-wave velocity structure that explains both observed H/V ratios and surface-to-borehole spectral ratios of weak earthquake ground motions. Although it is not the optimized structure, we obtained three quite similar structures based on three different initial models and searching ranges derived from previous PS logging, MASW, and new PS logging.

View full abstract

The non-linear response of the ground in the Hitachi-Naka thermal power station during the 2011 off the Pacific coast of Tohoku earthquake was studied by comparing with the underground records of KiK-net in the nearby area. PGA ratios were decreasing in proportion to the increase of amplitude of bedrock motion. The values of response spectral ratio in short period range calculated from the main shock and largest aftershock data were smaller than that of other events. The Fourier spectral ratios showed almost the same characteristics to the response spectral ratios. During the main shock, the spectral ratios were varied in proportion to the amplitude of input motion. These characteristics show the non-linearity of the ground response. Peak shift of H/V spectral ratio of the ground in the Hitachi-Naka site suggests that the Vs in surface layers decreased to 70 % during the main shock.

View full abstract

In this report, using the strong motion records during the 2011 off the Pacific coast of Tohoku Earthquake at MLIT strong motion stations, we estimated strong ground motions at nearby K-NET stations with two methods, namely, the equivalent linear method and the site effect substitution method. In the latter method, the difference of the site effects below the engineering bedrock is considered, while in the former method, it is neglected. The estimated ground motions were compared with the observed ground motions at K-NET stations and the accuracy of the methods was discussed. The results indicate that, in the frequency range for which the difference of the site effects is significant between the stations, the accuracy of the equivalent linear method may not be quite satisfactory.

View full abstract

Author has been developing temporary strong ground motion observation site around Aomori prefecture with a specific purpose for each site. During the 2011 off the Pacific coast of Tohoku earthquake, six accelerograms were obtained in four area, those are Higashidori-mura, Tohoku-machi, Aomori city, and Hirosaki city. The maximum PGA of observed data is 168cm/s², equivalent seismic intensity defined by JMA is 5.0. Though these values are not significant, obtained ground motions meet purposes. In this paper, those accelerograms are introduced and characteristics of ground motions at each site are explained with estimated ground structure.

View full abstract