An approach to prediction of magnitude and occurrence time of earthquakes

Abstract

The probability that an earthquake, when it is assumed to occur in a specified area, will fall within a given magnitude and time range is estimated with the aid of statistical seismicity and a few premonitory effects such as anomalous crustal deformation, anomalous geomagnetic change and foreshock activity. Much stress is put on estimating changes in probability when observation of these specialized phenomena is added one by one. As far as the area centred on Tokyo is concerned, prediction of magnitude seems fairly reliable, but no prediction of high reliability is possible for occurrence time. The rating of expected earthquake danger areas as proposed by the Japanese Geodetic Council, into “areas for special observation” and “areas for concentrated observation”, can, without being biased by subjective thinking, be made on the basis of the probabilities estimated in a way presented in this paper.