Short-term Probabilistic Prediction of Earthquake Occurrence in Southwestern Nigeria
*Adepelumi, A. A., Onibiyo, O.S and Isogun, M. A
Department of Geology, Obafemi Awolowo University, Ile-Ife, Nigeria.
*Corresponding author. email: aadepelu@oauife.edu.ng
Accepted on November 15, 2010.
Abstract
Prediction of the time interval when the next strong earthquake or tremor would occur in a seismic source region is a difficult and ill-posed problem. In this study, a priori information on past records of earthquake occurrences, distribution of observed interval times between earthquake events and the elapsed time since the last earthquake occurred are the information needed for the prediction of a new seismic event. A well-tested, time-dependent statistical model was employed to predict the probabilistic occurrences of earthquakes. The Earthquake Recurrence Model uses two predictors based on concepts of fracture and earthquake recurrence rate. This probability model takes the mean recurrence intervals and standard deviation of historic earthquake events in the study area in order to determine the probability of earthquake occurrence for the predicted years. The earthquake records used showed that at least twelve minor earthquakes/tremors of magnitude (MS) varying between 1 and 5.4 had occurred in the region since 1914. A time-window of Year 2008-2029 was considered for the modelling exercise taking into consideration the Ijebu-Ode region, south-western Nigeria. It is thus assumed that the sequence of events recorded in this region represents a set of an unknown number of consecutive phases, and that the observations follow the Poisson distribution process. The results of the model showed that the probability of earthquake occurrence in the study area between the Year 2009 and 2028 increased from 2.8% to 91.1%. The result also showed that the probability of 2 events occurring has the highest likelihood within the predicted years. It was also observed that the Weibull probability density model predicts a damaging earthquake (M≥ 5) before Year 2020. However, there is inherent danger in using purely empirical predictors like the one employed in this study; therefore the result should be treated as preliminary.