Spatial distribution of earthquake foci for the period from 1926 to 1968 in and near Japan were expressed in 16 maps each for a layer of 20-40km in thickness lying between 0 to 600km in depth. The maps show definitely the boundary lines and centers of local activities distinguishable in seismicity from each other. We named them seismic blocks. The centers of the shallow seismic blocks which crowd in a zone along the west side of the Japan trench can be traced down to 600km in depth. At first such centers move with increasing depth toward two junction areas of the island arcs of Kuril, NE-Japan and Izu-Bonin, that is, the Hokkaido and the Kwanto districts. From there they continue sinking to north for Hokkaido and to west for Kwanto making a dipping angle of 30°. Then their directions change abruptly by 90° at a depth of about 300km. From North Hokkaido one branch of the locus of the center goes into the Okhotsk Sea and the other into the Japan Sea along the coast of Primorye. On the other hand, the locus of the block center moving toward the west Kwanto turns to the south at that depth under Gifu Pref. for the Shikoku Basin, although a shorter branch goes northward into the Japan Sea. It is notable that few deep shocks occur in the middle of the Japan Sea. Considering the other geophysical and geological phenomena such as (1) the negative zone of isostatic anomaly of gravity lying along the Japan trench branches out to the junction areas in land, (2) the movement of the triangulation stations for the last 60 years or so amounts to about 3m to the north at Aomori Pref. in comparison with the Kui Peninsula, showing that two forces are acting in the respective areas in opposite directions, (3) the axis of mountain ranges made of metamorphic rocks which encounter the locus of the block center in Hokkaido and Central Honshu underwent severer bending, we conclude that the locus of the seismic block center expresses the location of the mantle current coming from the Pacific. itself.

View full abstract

The writer tried to find out whether any secular variations occur in the value of β in the expression of relation between the number N of earthquakes and their magnitude logΔ, logN=α-βlogΔ, using, in place of magnitude M, the value logΔ, a common logarithm of the maximum felt distance, Δ (km), of the earthquake.
The areas investigated in the present paper are Regions E, F, G, and H shown in Fig. 1. The values of Δ were sorted out for the years from 1920 to 1967 and divided into four groups: Δ=180km(100km≤Δ<250km), 320km (250km≤Δ<400km), 500km (400km≤Δ<600km), and 800km (600km≤Δ<1000km). With regard to the number N of earthquakes during the period of every five years (for Regions E and F) or every ten years (for Regions G and H), the coefficients α and β in each region and each period were determined by the method of least squares.
The obtained values of α and β are shown in Tables 1 to 4, together with their probable errors, and the variations of the value of β are shown in Figs. 2 to 5, respectively. In each region, it is certain that the value of β shows secular variation. At the top of each figure, the total energy released by the earthquakes during the statistical period is described. It is noticeable that release of great seismic energy took place when the value of β was becoming minimum. The relationship between the seismic energy (erg) per 10⁴km² per year and the values of β is, as shown in Fig. 6, expressed by the form, logE=23.97-1.07β. If we in this expression let β→0, we find that E≈10²⁴erg/10⁴km², and it means that the maximum energy released by a earthquake is about 10²⁴erg. This estimated value may be reasonable.
In Region E, the value of β for the aftershocks of '68 Tokachi-Oki Earthquake was obviously greater than the value of β for the earthquakes during five years just before the Earthquake.

View full abstract

Earthquake disaster is dependent upon both seismic intensity and strength of establishment. The higher the strength, buildings become more safe, which in turn increases the cost of investment. Therefore there exists an optimum strategy to minimize loss (=investment/life). The Monte Carlo simulation of construction and destruction of buildings is carried out by generating a time series of earthquake occurrence. The strength of buildings is expressed by the magnitude M and is assumed to decrease with time according to exponential decay law. Effects of partial damage and repair are included. It is concluded that the initial strength corresponding to M=7.5-8.5 is optimum. However, there exists no optimum solution if the effect of half damage is significant.
Economic merit of earthquake prediction and of counterplan is also studied. It is found that the merit is sensitively dependent upon a reliablity of prediction which is expressed by (damaged area/area where damage is predicted). If this ratio is 50, any counterplan appears to be of no use.

View full abstract

In general, the Ishimoto-Iida's Relation between the maximum trace amplitude a and the number of earthquakes n(a)
n(a)=ka^-m
holds good for earthquakes observed at an observation, where k and m are constants.
In order to determine if the parameter “m” changes with time and with a range of observed earthquake magnitude, this study was carried out in case of the Matsushiro earthquake swarm at Sanada, about 15km southeast from Matsushiro, by means of an observation in four dynamic ranges.
The period of observation is 82 days from June 15 to September 9, 1966, with interuption from August 15 to August 19.
The method of least-squares was used to determine the parameter “m”. For a calculation, the running means of 3 days, 6 days, and 8 days were adopted as daily number of earthquakes. We could not use the data of the fourth channel, because the number of earthquakes was not large enough for statistical analysis.
Some conclusions were derived from this analysis; 1, The parameter “m” for the three channels changes with time and these changes have statistical significances respectively. 2, Each value of the parameter “m” deduced from each channel is different, and they are estimated as 1.267±0.04, 1.877±0.153 and 1.901±0.067 from the first, second and third channels, respectively. 3, The difference between the values of the parameter “m” derived from the first and second channel has a statistical significance throughout the observation period with few exception. As for the difference between the values for the second and third channel, however, the statistical significance cannot be ascertained, for it changes with time for the observation period.

View full abstract

The loss of gravitational energy on core formation is calculated for Venus, Mars, and Mercury. The energy is 1.02×10³⁸ ergs, 1.10×10³⁶ ergs, and 1.34×10³⁶ ergs for Venus, Mars, and Mercury, respectively. This is equivalent to an average rise of temperature of 1600°, 130°, and 310° for Venus, Mars, and Mercury, respectively.
The results show that the energy released by the formation of iron core of Venus may play an important role in the thermal structure.

View full abstract