| Popis: |
Macro-anomalies preceding a number of large earthquakes in Japan are analyzed in relation to maximum detectable distance, D max , from the epicenter and to precursor time, T , between anomaly appearance and main shock occurrence. These anomalies include: anomalous animal behaviour, strange detonation, unusual light in the sky, change in underground water and hot springs; and are all things which can be sensed by humans without using scientific instruments. The number of anomalies analyzed amounts to 910. As a result an approximate relation between main shock magnitude, M , and D max is brought out. Given an M , therefore, it is possible to draw a circle, having a radius equal to D max , centering at the spot where an anomaly is observed. In that case the epicenter of coming earthquake must lie within that circle. When there are multiple anomalies, the epicenter should be located somewhere in an area where all the circles overlap. It will be shown that such a D max method leads us to an assessment of the approximate epicentral area and main shock magnitude. It appears that macro-anomalies begin to be observed about 100 days before the main shock, anomaly numbers tending to increase at around T = 10 days in most cases. The maximum frequency of appearance usually occurs around 1 day prior to the main shock. The distribution of frequency can well be approximated by a Weibull distribution. Making use of parameters of the distribution, changes in probability of earthquake occurrence in a specified time interval can be evaluated when an anomaly is observed. When many anomalies are observed successively, time changes in the synthetic probability can be evaluated. It appears that such a probability tends to increase, reaching a value close to 1 as the main shock occurrence approaches and the frequency of anomaly appearances increases. Although no actual pre-earthquake prediction based on macro-anomaly data has as yet been conducted, it is hoped that the technique developed in this paper will be applied to anomaly data related to coming earthquakes, once a well organized data acquisition system comes into operation. The authors are of the opinion that macro-anomalies may be applied, at least to some extent, to earthquake prediction, despite the traditional view that macro-anomalies are so unreliable and contaminated by noise that they cannot be used for prediction. |
