Real-time information from seismic networks

In this thesis, the performance of a novel methodology for the real-time estimation of seismic source parameters is evaluated which can be used in earthquake early warning systems. The method, PreSEIS, uses artificial neural networks to map seismic observations from a network of sensors onto likely source parameters. The method is veryfied using two real earthquake datasets, as well as a synthetic earthquake dataset to evaluate the performance of the Istanbul early warning system.