Probabilistic Earthquake Hazard Zonation in Kerman using Statistical Analysis and Artificial Neural Networks

The science of earthquake prediction has not yet reached its desired level of development. It has not yet led to a successful prediction of an earthquake based on physical principles. Therefore, the prediction of the occurrence of an earthquake has always been an interest for researchers. Three parameters of earthquakesincluding: time, location and magnitude, are the most important uncertainties in earthquake prediction. Uncertainties of earthquakes have the fundamental role in assessment of reliability of structures. In reality, the mechanism of earthquakes makes predicting them more difficult. Earthquake prediction will have a special validity if random variables consider. Therefore, nowadays statistical hypothesis methode are used to determine the probability of an earthquake. There are many studies in prediction of earthquake parameters, but there is not any particular research in the probabilistic earthquake hazard zonation. Kerman province has experienced destructive earthquakes such as Bam, Golbaf and Zarand as the one of the most seismic regions of Iran. In this study, the probability of future earthquakes location with the magnitude greater than 4.5 Richter has been predicted in Kerman by using artificial neural networks and statistical analysis. The most probable event of earthquake has been predicted 38.6% in south of Kerman.