Three methods of analysis of earth tides-Lecolazet's, least square's and Fourier integral's-were compared with each others using one month's data obtained with an Askania gravimeter in Kyoto, Japan. Both a modified Lecolazet's method to exclude systematic errors of analysis and the least square's method assumed to be present seven components of tides, give excellent results in gravimetric factor and phase lag, using hourly values covering the duration of one month. Observational data obtained in Kyoto (December 1967-March 1968) and those in Mizusawa (November 1967-February 1968) were analyzed continuously by means of the modified Lecolazet's method. These analyses show a similar variation in gravimetric factor and phase lag, and it is considered to be a part of seasonal variation in comparison with the previous observation made in Kyoto (August 1959-August 1960). Further, the present results for gravimetric factor give greater values, in general, than the previous ones. This suggests secular change or long period variation in gravimetric factor.

View full abstract

Since the International Geophysical Year, gravimetric tidal observations were made at many stations in the world. But the obtained values of tidal factor and phase lag of gravity show fairly large fluctuations with time even at the same station, and they are too large to be attributed only to observational errors. In order to investigate the causes of such fluctuations, simultaneous observations of earth tides with four Askania gravimeters and two LaCoste & Romberg gravimeters were made from August 30 to November 17, 1968 at the International Latitude Observatory of Mizusawa. In succession, simultaneous field survey including two other LaCoste & Romberg gravimeters was carried out for the purpose of calibrating the scale constant of the gravimeters along the route from Mizusawa to Hachinohe (gravity difference is approximately 210 mgal). The results obtained through the present investigations are as follows:(1) There appears to be the discrepancy among the gravimeter type.(2) Large fluctuations of the tidal factor recognized by the Askania Gs-11 gravimeters amount to about ±35%, and it may be considered that they are mainly due to non-linearity of the recording system.(3) It is to be noted that trustworthy results cannot to be expected so far as continuous observations of earth tides are made by a single gravimeter.

View full abstract

Strain steps associated with earthquakes have been observed with super-invar-bar extensometers at Iwakura, Amagase and Donzurubo Observatories for earthquakes of magnitude 3.2-7.9. The stability of these instruments for vibration was confirmed experimentally by two methods. Strain step amplitude dependance upon distance seems to be R^-2.4, and based on a few assumptions, fault length is related to earthquake magnitude by the following equation, M = -8.2+2.2log10 Lwhere L is fault length in cm.

View full abstract

If the earthquake is the result of the accumulation of strain energy in the crust and the upper mantle, it might be expected that the necessary period of energy storage will be related to the scale of the resulting earthquake. For shallow earthquakes, it is expected that the crustal deformation accompanied by the strain accumulation may be observed by frequent surveys of geodetic nets in the area concerned. Although it is quite difficult at the present stage to find geodetic data which clarified the period of duration of the preseismic crustal deformation, the writer looked for following three examples which might have the possibility of suggesting the preseismic duration.Name Year and Mag. Assumed duration T Nankaido Earthquake 1946 M 8.1 92 years Kitamino Earthquake 1961 M 7.0 12 years Omi Earthquake 1967 M 5.0 0.3 years Considering that these three data show a linear relation between M and log T, the writer calculated the coefficient of log T = bM+a by using method of least squares as follows;log T = bM+a = 0.79 M-4.44. (1)Further he checked this equation by applying it to some examples which suggest the durations of preseismic deformations in somewhat insufficient features comparing the above mentioned data, and found considerable consistency (Table I and Fig. 5). If the earthquake may occur when the linear strain per unit length of the rock of the earth's upper layer attained to some limiting value and Eq. (1) is concluded, then we may have the following equation;eT = const. (2)where e is the velocity of strain accumulation per unit length per year. This shows that the small earthquake will be caused by quicker accumulation of strain than large one, which is favourable condition for the purpose of earthquake prediction. From Eq. (1), the duration of crustal deformation of the earthquake of M8.6 (which may be the maximum to be expected) is about 220 years. If we assume that, in the volume surrounding the epicentral region which stores most part of seismic energy, the average strain per unit length will be about 5 x 10-5 immediately before the earthquake, then Eq. (2) becomeseT≈5 x 10-5, (3)and for the earthquake of M8.6e≈2.2 x 10-7 per year.Let us assume furthermore that the horizontal average diameter of this seismic volume be about 200 km, then the crustal movement of the circumference of this area will be about 4.5 cm/year, which nearly corresponds to the value suggested by the theory of continental drift and ocean bottom spreading. The writer assumed the time interval To and the block (described as a seismic province in the following) which covers so-called a seismic cycle and contains one complete sequence of earthquakes including one earthquake of maximum scale. Let the number of earth-quakes of almost same magnitude per year be N, and T be the period of preparation which might be considered as duration of progressive strain accumulation, then in one seismic province the following equation might be obtained, assuming statistically the sequential collation of duration periods for earthquakes of nearly the same magnitude, ToNT = To, as ToN is the number of earthquakes in To years. Then NT =1. (4)After Gutenberg and Richter log N = β M+α. (5)From Eqs. (1), (4) and (5), b = - β and α = - a. (6)According to the estimation by Gutenberg and Richter, the coefficients 3 and a for Japan area β=-0.80 and α=+5.5. (7)Comparing (1) with (7), the first condition of (7) is almost satisfied and the second may also hold, if we assume about ten seismic provinces for Japan. To confirm the reality of these relations, more data and investigations are expected.

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract