Frequency dependences of near-surface attenuation of S-waves based on the vertical array records were examined to clarify their physical mechanisms using the data collected at the KiK-net Narita observatory site. The attenuation was estimated as a function of frequency by matching the SH-wave transfer function to the observed spectral ratio using the subsurface velocity structure derived from the PS logging profile. In addition to these examinations, numerical experiments with synthetic pseudo spectral ratio, which was calculated by use of the subsurface model constructed from the velocity and attenuation structure estimated at the Narita observatory, were performed to make clear the cause of frequency dependences of S-wave attenuation from an analytical point of view. The estimated S-wave attenuation of the sedimentary layer-basement system at the Narita observatory has a constant value of about 0.004 - 0.005 (h=0.4 - 0.5%) with its lower limit at frequencies higher than 4 - 5Hz. On the other hand, the attenuation strongly increases with decreasing frequency below around 2 - 3Hz. It is quite likely that selection of the time window for picking out the S-waves portion may lead to artificial frequency dependence of attenuation at these lower frequency ranges. The attenuation value and its frequency dependence tend to decrease as the length of time window increases. In other words, the S-wave attenuation evaluated from the vertical array records seems to have some artificial frequency dependences, especially at lower frequencies. It is therefore important to carefully check errors in estimating attenuation.

View full abstract

We presented an empirical predictive model for the long-period (1-15sec.) ground motion and examined the validity of the model by comparing with the observations. We computed the acceleration response spectrum of 5 and 1 percent damping using the strong ground motion records retrieved by the K-NET, KiK-net, and JMA Seismic Intensity Observation Network. We defined the shakeability at a site by the spectral ratio of the response spectrum observed at a site to the ones at seismic baserock (average VS ~ 2.4 km/s) predicted using the empirical attenuation model proposed by Yuzawa and Kudo (2011). Shakeability maps for long-period (1-15 sec) ground motion in Japan were provided and they were examined using typical observed records.

View full abstract

In this article, a characterized source model is proposed for the 2004 Mid Niigata Prefecture, Japan, earthquake (M_J=6.8) by using the stochastic Green's function method, taking into account the site specific amplification and phase characteristics. The strong motions are simulated based on the source model with four asperities, referring to the result of waveform inversion. The simulated ground motions are well consistent with the observed records. It is found from the simulation that the large-amplitude velocity-pulses observed at K-NET NIG019 are generated by small asperities nearby the hypocenter. Furthermore, validity of the strong motion simulation technique using site specific phase characteristics is studied. The result indicates the importance of using the site-specific phase characteristics for reliable response analyses of structures.

View full abstract

A new data logger is developed to use for microtremor measurement. Especially, to know the deep structure, we need to handle microtremors in low frequency range: that is, lower than 1 Hz. In this frequency range, not only power of microtremors is very small, but the sensitivity of sensor is generally critical. A data logger with very low noise and high resolution is required for this purpose. Of course, we can find very expensive systems with high performance in noise and resolution, though, they are not so suitable for the microtremor survey, usually. We, thus, introduce some traditional but new features and complete a new system. Basic performances of the new systems are examined and discuss the aplicability to microtremor observations. Consequently, the noise is reduced to about 30 to 36 dB (5 to 6 bits) and the overall dynamic range is more than 170 dB (29 bits).

View full abstract

We estimated aged deterioration and strength difference by the seismic design code revision of wooden houses using fragility curves classified by age. First, we calculated strength distribution in each age by seismic response analyses using a wooden house cluster model, and we got relation between age and a base shear coefficient. Next, we separated relations between age and a base shear coefficient to the aged deterioration and seismic design model. We found that strength degraded 30∼40% for 30 years by aged deterioration, but after that it became constant and that revision of seismic design code in 1981 made about 20∼30% difference, whereas revision in 1959 had almost no influence.

View full abstract

The empirical equation for estimating the site amplification factor of ground motion by the average shear-wave velocity of ground (AVS) is examined. In the existing equations, the coefficient on dependence of the amplification factor on the AVS was treated as constant. The analysis showed that the coefficient varies with change of the AVS for short periods. A new estimation equation was proposed considering the dependence on the AVS. The new equation can represent soil characteristics that the softer soil has the longer predominant period, and can make better estimations for short periods than the existing method.

View full abstract