Abstract
In the earthquake prediction problem the estimation of the occurrence probability of major earthquakes by using multiple precursory data is desirable. A relation between the synthetic probability and individual probabilities for the earthquake prediction relying on a Markov chain model is discussed. A historical earthquake sequence in northern China of about two thousands years and the results from the observation of the radon content anomaly which covered a period of about ten years are used. The following linear relation is assumed:
P*7(D)=αP(M)*3(D)·P(R)*3(D),
where α shows a coupling constant, a factor concerning degree of dependence of two precursory phenomena, P*7(D), P(M)*3(D), P(R)*3(D) are the probabilities of the earthquake occurrence normalized by the interval of the time segment in a Markov chain, when both precursors, moderate-size earthquakes, a radon content anomaly appear as precursory phenomena, respectively. Analysis shows that this relation fits well to the actual data. The value of α is found to be about 150-170, independent of the interval of the time segment and radon content anomaly recurrence interval.
