Prediction of the next earthquake in the Mexican subduction zone and NAFZ using the predictive distribution

Abstract

The estimation of the time of the next earthquake, in a given seismic region, is one of the most useful and difficult tasks for scientists who study and predict earthquakes. In this study, we examine the previous times between earthquakes in two seismic regions, Michoacán in Mexico and the North Anatolian seismic region near Izmit, Turkey. The Kolmogorov–Smirnov test shows that the recurrence times follow the exponential distribution. Based on this finding the predictive probability distribution of the time for the upcoming earthquake is derived. The mean and median of the predictive distribution are used to estimate the time for the next earthquake; furthermore prediction intervals for the time of the upcoming earthquake are derived. It is found that the median estimator is a better estimator than the mean.