Seismicity and Water Level change in the Caspian Sea, an Explicit Function Based on Genetic Algorithm

Unique water level variations make the Caspian Sea a suitable case for investigating seismic responses, analogous to the problem of Reservoir Induced Seismicity (RIS) which is an interesting issue in civil, geotechnical and earthquake engineering. An analysis using past data demonstrates that the seismicity of the Caspian Sea region changed with changing sea water level and changes in the b-value of the Gutenberg-Richter Relation have had an inverse correlation with sea level changes. It is conventionally assumed that, the b-value can show the stress level in a region, hence, additional loads on the earth crust, represented by sea level changes, affect the b-value and change seismic regime. It is not possible to calculate b-values for all time periods due to limitations in the past data and therefore, an explicit non-linear function is proposed to approximate the obtained correlation using a genetic algorithm and an RBF (Radial Basis Function) neural network. This function can estimate b-values where there is not enough data for a definitive evaluation of the b- value..

Keywords: Seismicity; Sea level changes; Function approximation; Genetic algorithms; Neural network