Abstract
As a fundamental study on the prediction of intraplate earthquakes in the Japanese islands, the author examined the reliability of the reported seismic precursor data on shallow earthquakes occurring in the Japanese islands and their close proximity. Linear regression analyses of the precursor data proved the necessity of re-confirmation of these data because of their lack of reliability.
In order to identify as objectively as possible some precursory anomalies in the seismicity patterns and some geophysical and geochemical observations obtained at stations in the Tokai area, computer programs were developed on the basis of the Fuzzy set theory.
Based on the results obtained by using these programs, the prediction and success rates of earthquake prediction were estimated, and the study showed that the rates are as low as 10 to 50 %. The statistical study showed that the rates for precursory anomalies in the seismicity pattern vary remarkably from region to region, too.
The use of various kinds of data for identifying precursory anomalies improved both the prediction and the success rates. This seems to suggest that the method is useful to discriminate precursor from noise.
A computer system for monitoring changes in the seismicity pattern for the prediction of earthquakes occurring in the Japanese islands was made by using the above-mentioned various computer programs.
