1970 年 79 巻 3 号 p. 115-138
In earlier days most Japanese wanted eagerly to have means of earthquake prediction in order to escape earthquake disaster taking refuge in advance in open spaces since they were almost powerless in encountering the menaces. We are now enabled to long for the mitigation of the disaster by the advancement of science and technology, although we have not yet succeeded to predict the coming of a disastrous earthquake. Our safety thus depend entirely on our effort for the countermeasures against the calamity. But we have only limited allowances for such preparations economically and in time available. It is therefore inevitable to start the preparation from the most important and imminent localities in a country, from the national point of view.
In Japan, according to the writer's witness on the imminence of Southern Kwanto District at the Diet in July, 1964, such preparation work for Tokyo Metropolis was started the next month by a new Earthquake Division of the Disaster Prevention Council of the Metropolis, and several important hazard possibilities of the imminent earthquake have been revealed with the result that similar committees have also been made in other prefectures and cities in the same area as well as in the Central Government of the country.
In this paper the writer intended to give the ground for his assertion of the imminence of a disastrous earthquake giving proofs of the periodicity of such calamities on the statistical and historical analyses, and made some comments on the possible consequences and major problems in the countermeasures to cope with the conceived disaster as revealed by the studies made on the national and metropolitan levels.
In view of the importance, interest and the space allowance, only the proofs of the periodicity in the recurrence of destructive earthquakes in the district concerned are summarised in this abstract.
Historical earthquakes in and around Kwanto District as shown in the map (Fig. 2.1) were examined and those which were destructive at Kamakura or presumable as such from the standard intensity-distance-magnitude curve were selected and listed in Table 1.1. Periodicities of these earthquakes were analysed statistically. Similar analyses were also applied to the earthquakes experienced in Tokyo which are listed in Table 2.2. A definite period τ of 69 years (besides a few integral multiples of this period) was clearly found. The methods used in the above analyses were by means of (1) Fourier transforms (Fig. 1.1), (2) autocorrelation functions (Figs. 1.3 and 1.4) and (3) historical examinations of the number of cases when no destructive earthquakes took place within the standard deviation +√ ξ2 from the year tp when a destructive earthquake was expected to occur from the periodicity. Test of persistence of the period was also made by means of the theory of random walks consisting of Fourier amplitudes in successive periods (Figs. 1.5 a and b). Rayleigh-Schuster's criterion (Formula 1.3) that is the probability of obtaining the actual Fourier amplitude on the assumption of no periodicity or the resultant distance of above mentioned random walks gave the values of the probability equal to 0.06% and 0.03% for Kamakura and Tokyo respectively. In comparison to usual value of 5 or 10% for recognizing periodicity in geophysical phenomena, the smallness of the above values is more than enough to disprove the non-existence of the periodicity of 69 years in the recurrence of destructive earthquakes in the districts under consideration.