Probabilistic prediction of the next large earthquake in the Michoacán fault-segment of the Mexican subduction zone

Abstract

Estimation of the time interval Δt until the next strong earthquake to be expected in a seismic source region is a difficult problem. In the conventional method of time-interval prediction, given some distribution of observed interval times between large earthquakes and knowing the elapsed time t since the 1ast large earthquake, the probability of a new seismic event in an interval time Δt may be estimated. In this paper, we reverse the approach and we estimate the interval time for the occurrence of the next large seismic event assuming that the conditional probability of earthquake occurrence is a maximum, provided that a large earthquake has not occurred in the elapsed time t since the last large earthquake. We assume the Weibull distribution, the Rayleigh distribution or the Pareto distribution for the earthquake recurrence time intervals.In the Michoacán seismic region and using a list of historical large earthquakes in this seismic area, we found that the Pareto model predicts a damaging earthquake (M ≥ 7) before the year 2014.99, or before December 2014 ± 1.76(yrs.).

Keywords

• Probabilistic prediction
• conditional probability of earthquake occurrence
• Weibull
• Rayleigh and Pareto models
• estimation of interval times of occurrence.