抄録
Analysis of earthquake precursor data so far obtained in Japan makes it possible to determine the dependence of maximum detectable distance of a precursor (Dmax) on main shock magnitude (M). The M-Dmax relation thus obtained is different from discipline to discipline of precursor.
When a precursor-like phenomenon is observed at an observation point, we can draw a circle centering at that point with a radius of Dmax peculiar to that discipline provided a certain value of M is assumed for the main shock. If there are a number of observation points in an earthquake-threatened area and if anomalies are found for various precursor disciplines, we may have many circles with different radius. In that case it is likely that the epicenter of the soon-to-occur earthquake is located in the area, which is common for all the circles. When a magnitude, which is too small, is assumed, the circles do not overlap, so that the lower limit of magnitude can be derived. On the contrary, the overlapped area, which may be regarded as the epicentral one, becomes so wide that it cannot be accepted as the actual area over which the seismic energy is accumulated when an unreasonably large magnitude is assumed. In such a way we may have a rough idea about the main shock magnitude.
As examples, the method is applied to the precursors of the Izu-Oshima Kinkai (M=7.0, 1978) and Niigata (M=7.5, 1964) earthquakes achieving a remarkable success in predicting the epicentral area and main shock magnitude, which are more or less the same as the actual ones.
