In the previous papers it was confirmed that the long-period microtremors change their spectral character very sensitively to deep soil deposits and also the amplification characteristic of earthquake strong motions is, specially in the long-period range, correlated to that of the microtremors. This clarified the engineering importance of observing the long-period microtremors. However, observational and analytical processes in those papers were concentrated to the elucidation of the spectral peaks of the microtremors and their relation to the underground structures, that is the microtremors were treated as if they were standing waves. For further improvement of their applicability in earthquake engineering, considerations from wave-theoretical point of view are better to be introduced. The best research way is to understand their generation, propagation, and amplification due to ground layers as a series of events.
In this paper, as the first step of reserch in this way, a chiefly concerned investigation with the propagative characteristics of the long-period microtremors was carried out. To do this, a temporay array observation was prepared putting six sets of three component seismometers in an equal space of 60m, and the sequential observations were performed along three independent lines with an angle of 45° each other.
Analysis was begun by obtaining apparent propagation velocities of the microtremors by means of phase correlation method, and by determining the direction of the wave propagation. Then the true phase velocities resulted as a function of frequency. Group velocities of the microtremors were roughly estimated by considering flows of the wave energies figured out on running power spectra at the two observation points far away from each other. Next, these phase and group velocities were compared with those calculated by knowing the underground structure at the site. Amplitudes of the microtremors in horizontal and vertical components were investigated in relation to their particle orbits at the ground surface.
The results by these analyses are summarized as follows.
1. The long-period microtremors are interpreted as an ensemble of dispersive waves. In the period range as we concern, the M₁₁, and M₂₁ modes of Rayleigh waves are dominant. Adding to these, the fundamental mode of Love waves is also mixed.
2. The propagation direction of the long-period microtremors is, though slightly time-variant, decisive and mainly from open sea side (Pacific ocean).
Finally, we would like to add that further experimental observations for the second step of the wave-theoretical analysis of the long-period microtremors are on schedule.

View full abstract

Recently, changes in V_p/V_s ratio have been found by Iizuka to occur before huge earthquakes with hypocentral region depths of 30-40kms below seabottom off Japan.
It will be difficult to explain the occurrence of earthquakes with hypocentres of 30-40kms by only the dilatancy mechanism, since there will be some limits to the downward penetration of water, though original water included in the rocks may play an important role. In contrast, the deeper the location of hypocentres, the more significant will be the existence and behaviour of heat.
In the case of 30-40kms depth earthquakes around Japan, for example those taking place beneath the Pacific Ocean just off the Japanese islands, it has been suggested that, following the plate tectonics theory, the subduction of the Pacific plate will act to drag down the margins of the main Japanese islands as part of the overlying plate, and that earthquakes will occur along the plane due to the great force at work there. The great earthquakes are said to occur in low heat flow zones. However, is it true that such earthquakes always take place only in low heat flow regions?
To explain this matter, we put the locations of epicentral areas on the heat flow map (Fig. 5). From this figure, it is clear that the huge earthquakes have not always taken place only in the low heat flow zones. The Kanto (1923), Enshunada (1854), Tonankai (1944), Nankaido (1946), and Hyuganada (1968) earthquakes, all of them of the order of M=8 (Richter's scale), occurred in comparatively high heat flow zones of 2.0 HFU or more beneath the Pacific Ocean just off Japan in the western half of the Japanese islands.
On the otherhand, the major earthquakes occurring beneath the Pacific Ocean just off the northeastern half of Japan are located in the abnormally low heat flow zone averaging 1.0 HFU. However, it is worthy of note that the depths of hypocentres (or energy accumulations) in this area have been located at depths greater than those of the huge earthquake zones in the southwestern half of Japan. This means, even in the case of those earthquakes of the low heat flow zones, that high temperature can be expected at the hypocentral depths.
The relation of temperature against pressure and the thermal gradient curves with melting temperature dada are given in Fig. 6. In this figure M. T. refers to the melting temperature of wet basalt, while (M. T.) indicates the melting temperature of dry basalt. In the case of granite, the melting temperature M. T. g is lower.
In the case of the earthquakes of Kanto, Enshunada, Tonankai, Nankaido and Hyuganada, the depths of hypocentres are roughly 30kms, and the heat flow values recorded on the sea floor 1.5-2.5 HFU, therefore, from the thermal gradient curves it is obvious that the temperature in the hypocentral zone will approach the melting points of wet basalt. Under such a situation, assuming that there is little pressure decrease due to dilatancy and temperature increase due to the existance of the huge force acting on the rocks, partial melting or even phase transitions will be possible in the hypocentral zone.
Even in the case of the earthquakes of the low heat flow zones, occurring beneath the Pacific Ocean just off the northeastern half of Japan at depths of 50-60kms, we can see from the thermal gradient curve starting at 1.0 HFU on sea floor, the temperature at the hypocentral zone will meet the melting point of granite or will be 100°C less than the melting point of wet basalt. Therefore in this case, also, it can be presumed that partial melting will occur, assuming that once again there is little pressure decrease and temperature increase due to the existence of huge force acting on the rocks.
The melting or partial melting will create a volume increase, consequently stresses will be originated, then such forces will be added to the original forces, causing earthquakes to take place.
During the ti

View full abstract

The Kalapana tsunami was generated off the south coast of Hawaii Island, accompanying the earthquake of magnitude M_s=7.2 (NOAA), at 14h 48m (GMT), Nov. 29, 1975. Some features of this tsunami are investigated on the basis of tide gauge records of NOAA and Japan, adding the reports of the US field investigation.
Tsunami magnitude of the Imamura-Iida scale is decided as m=2, judging from the tsunami heights observed near and distant fields. The initial motion of the tsunami waves was in an upward direction at the whole Hawaiian stations, suggesting the uplift of the sea-bottom in the tsunami source area. The source dimension of tsunami is inferred to be 70km along the south coast of Hawaii Island, and the area is 2.2×10³km². By applying the corrections for the refraction and shoaling from the inundation heights at the Hawaii Island, the average vertical displacement of 1.2m would be occurred in the source area. The tsunami energy of approximately is 1.6×10²⁰ ergs. According to statistical relation, the present tsunami is large compared with the earthquake magnitude.
The tsunami fronts arrived in NE Japan at about 7h 40m after the occurrence of the earthquake. The maximum double amplitude is 20-30cm with the period of about 15min and the amplitude is relatively higher than that in SW Japan. It is noticed that the wave rays emitting the tsunami source concentrate in the Kuril Islands.

View full abstract

There is a remarkable discrepancy in the previous results on the temporal variations in seismic wave velocities related with the Matsushiro earthquake swarm that have been studied by many investigators. A result from the V_p/V_s method shows co-seismic anomaly, while all the data from the P-residual method indicate precursory anomaly.
To make clear the cause of this discrepancy, temporal changes in V_p/V_s ratio was reexamined by using the different set of data source from previous studies. Two kinds of velocity anomalies, namely, precursory and co-seismic or post-seismic anomaly were detected as the same as in the previous investigations.
Combining the present data with the data from previous studies, it was inferred that the precursory anomaly in seismic velocities might occur mainly in the zone of lower crust to the uppermost mantle, while the co-seismic or post-seismic anomaly might occur in the upper part of crust whose depth is not greater than 10km, underneath at and adjacent area of the Matsushiro earthquake swarm. In other words, the anomalous zone must have migrated from the deeper parts to the upper crust about the time when the swarm began.
The origin of these two anomalies were interpreted by a new model which takes thermal effects into consideration for earthquake occurrences.

View full abstract

This is the second paper dealing with the patterns of static deformation fields due to the faults. In the previous paper (MATSU'URA & SATO, 1975b), displacement fields due to some typical fault models were shown systematically, and the general relationships between a fault model and the corresponding deformation fields were also pointed out. In the present study, strain and tilt fields caused by some typical dimensional faults in a semi-infinite medium are given systematically, by exhibiting vector maps for principal strain and tilt, and contour maps for other strain components, which provide a helpful guide to grasp gross feature of focal mechanism from the observed zero-frequency data.
The change of patterns of deformation fields with variation of the dip-angle is remarkable for pure dip-slip fault (Figs. 7 and 8), but not for pure strike-slip fault (Figs. 3 and 4). For the fault with arbitrary slip direction, it is found that extention (or dilatation) fields are mainly controlled by the dip component of slip vector, while shearing strain fields by the strike component.
Computations and mappings are carried out by the computer program developed by SATO & MATSU'URA (1974), by which all of the components of deformation fields (displacement, strain and tilt) can be automatically calculated and illustrated for a given fault model.

View full abstract

Programs for automatic reading of seismic P wave onset by a digital computer are introduced. These programs are composed of a series of simple processes as follows. A low cut recursive filter is used to increase S/N amplitude ratio. The arrival times of P waves are read by means of comparison with a threshold level corresponding to noise level before the P waves. Discrimination between P arrivals and noises are made by counting the number of points where amplitude of the wave exceeds the threshold level during a pre-set time interval. Then, the earlier of the closest two reading times among the three independently obtained from the three component data, is selected as the best P wave onset time.
As the result, 66% of microearthquakes observed at the Iwatsuki deep borehole observatory could be read within the time difference of 0.1sec between the automatic readings and ordinary visual readings.
Furthermore, an effective indicator of reliability of the finally adopted P onset times is discussed.

View full abstract

Hypocenters and focal mechanisms of microearthquake clusters were determined by using data obtained through the telemetering observation system of the Tottori Microearthquake Observatory. The events which occurred during the period April 24-September 30, 1976 were analysed.
The linear dimension of source region of a microearthquake cluster is found to be about several hundred meters; in a particular case, 200-300m. Focal mechanisms of the events in a certain source region are similar to one another. Fault plane and geometry of a source region is discussed from the view point of hypocentral distribution and focal mechanism.
Time sequences of respective clusters are of swarm type except for one cluster which shows a main shock-aftershock sequence. There seems to occur migration of activity from a certain source region to another.

View full abstract