Earthquake engineering importance of the observation of long-period microtremors as an effective tool to elucidate deep ground characteristics during an earthquake has been confirmed. For achieving better reliability on their engineering applications, however, understanding of them as wave events has been thought indispensable and been attempted.
The long-period microtremors were interpreted as an ensemble of dispersive surface waves (Rayleigh and Love waves) in the preceding paper. As a continuation of this line of study, a more detailed examination on Rayleigh wave component is carried out and reported in this paper. Phase velocity, particle motion, and deposit/basement spectral amplitude ratio are analysed, by which Rayleigh wave mode structure is studied.
Concluding remarks on this examination are as follows.
1. Rayleigh waves in the long-period microtremors are mainly composed of M₁₁ mode. M₂₁ mode is also detected but is not a predominant mode.
2. Rayleigh wave period-ellipticity relationship, known as a useful means to evaluate underground structures, can be disclosed employing three-component records of the long-period microtremors.
3. A detailed and further examination on Love waves is also desired, since the long-period microtremors are composed of both Rayleigh and Love waves.

View full abstract

P-wave velocity structure down to 150km in Kanto district is determined from 418 P-wave arrival time data observed at 10 stations for 55 seismic events that occurred in the region concerned. Simultaneous determination of velocity structure, hypocenter parameters and station corrections is made by using a procedure of successive iteration of linear inversion.
The estimated P-wave velocity structure in Kanto district is 5.5km/sec at the depth of 0-5km, 6.3km/sec at 5-30km, 7.86km/sec just below the M-discontinuity, 8.14km/sec at about 60km, 8.0km/sec at 70-100km, and 8.1km/sec at 100-150km.

View full abstract

It is one of the most difficult problems in the automatic seismometry system to determine the arrival time of the P-wave and S-wave.
In this paper, a new method to detect earthquakes and determine the arrival time is presented.
The algorithm should be simple, because the quantity of data is very large and rapid analysis is required. In this situation, the most suitable mathematical model of the seismic wave may be an autoregressive (AR) model.
The arrival time of P-wave and S-wave is determined by the log-likelihood ratio function for AR model. In this method, both amplitude and spectral information are effectively utilized and a backtracking process is included which simulates the way taken by the specialist in the manual operation.
The suitable order of the AR model and the numbers of the data in the calculation of AR model are determined from the practical point of view.
The experimental results show that this method is highly reliable and practical.

View full abstract

For the earthquakes occurring in depths less than 35km under the Hidaka mountain region in Hokkaido, two remarkable phases with sharp onsets and large amplitudes are sometimes recorded on the only seismograms at station IWN. Station IWN is one of the six telemetering seismic stations of Research Center for Earthquake Prediction of Hokkaido University located in the Hidaka mountain region about 80km from east to west and about 100km from north to south. These stations have observed more than 700 earthquakes occurring in the region from 42.25°N to 43.40°N and from 142.50°E to 144.20°E during the period from June 1976 to October 1978, among which eight earthquakes were selected as producing the two remarkable phases at station IWN which are referred to as the I₁ and I₂ phases, respectively.
The two phases are always recorded at the station as clear later arrivals; that is, the I₁ phase arrives in 1.7-2.6sec after the direct P wave arrival and the I₂ phase, in 0.8-1.1sec after the direct S wave arrival. In this paper, the analysis made on wave motions and travel times of these two phases is presented.
The principal component analysis applied to the phases recorded in three component seismograms shows that the prevailing wave motion of the I₁ phase is in the longitudinal direction while that of the I₂ phase is in the transverse direction and the I₁ phase arrives with much smaller incident angle. In addition to the wave motions, the spectra of the phases were obtained: the spectra show that the amplitudes die out at frequencies lower than about 15Hz for the I₁ phase and at frequencies lower than about 3Hz for the I₂ phase. The wave motions and spectra thus obtained strongly suggest that both phases can safely be identified with a kind of reflected waves associated with a boundary possibly existing in the lower crust. The I₁ phase may be assumed as a converted wave which is generated by S wave originated from the source reflecting as P wave at the boundary and the I₂ phase, as a purely reflected S wave. From travel time analysis applied to both phases, the boundary related to these phases was inferred to exist in depths of 30km beneath a point about 20km south of station IWN and to dip roughly toward the south with the surprisingly large dip angle of 33°. Spatial extent of the boundary with such a large dip angle may be limited probably in a small portion under station IWN, because the other five stations in the region considered have not observed such remarkable later phases.

View full abstract

Velocity inversion in the upper crust has been discussed mainly in the interpretation of a relevant strong phase on seismograms obtained in an epicentral distance range less than 150km. Synthetic seismograms constructed on a basis of models with or without velocity inversion were made and they were compared to the observed seismograms obtained by explosion works in the southwestern part of Japan. Appreciable differences were not found in a feature of synthetic seismogram record section according to the change of models in a distance range less than 150km. A feature of the record section varies, however, considerably as the change of models in a distance range greater than 250km. Taking the fact into considerations, velocity inversion, if it exists, would not be a well developed one at least in the southwestern part of Japan.

View full abstract

Modes and degrees of the damages caused by the Onikobe earthquake (M=4.9) of July 5, 1976 were investigated in special reference to the predominant direction of the ground motion. The directions of falling and slip of gravestones near the focal region were restricted to a narrow range of directions and these data were found useful for estimating the predominant direction of the ground motion. On the basis of the seismological data such as the hypocenter location, the focal mechanism and aftershock distribution, various dynamicsource models were tested if they could explain the predominant direction near the focal region. As a result, precise determination of the hypocenter location became possible, and the finally selected model explained the systematic pattern of the predominant directions observed near and around the focal region.

View full abstract

An observation system was constructed which can record long period component of seismic waves generated from near earthquakes. A seismograph of vertical component with natural period 5 seconds was used for this purpose. It was confirmed by repetitious rigorous tests that the seismograph works quite stably against various conditions in a period range much longer than the natural period. A system was designed so as to make the observation of longer period component as effectively as possible within a specific dynamic range taking characteristics of earthquake source, environmental condition, seismograph, amplifier, and recorder into considerations. It is expected, by using this observation system, that seismic waves of period up to 30-50 seconds can be observed for earthquakes magnitude smaller than 4 and at epicentral distance nearer than 200km.

View full abstract

The stability condition for the TTT (b) type triple junction of McKenzie and Morgan (1969) has not been known. A stability condition for this triple junction which was not defined by McKenzie and Morgan is presented.
If each plate moves parallel to the lateral trench attached to itself and subducts at the trench facing its leading edge, the configuration of the triple junction is maintained stably. This is the only stable condition for this type of triple junction.

View full abstract