Publication

Abstract : Tsunami and earthquake are most dangerous natural disaster. Hence an early and accurate warning is necessary in such fields. Earthquakes are hard to predict but its resulting tsunami’s can be predicted from seismogram. The available tsunami warning systems are not much effectively used in practical situations and they are time consuming since the time taken for data processing and modeling is high. In this proposed work, the characteristics of seismograms are used to distinguish the tsunamigenic earthquake from nontsunamigenic earthquakes. Feature extraction is done by applying wavelet decomposition and Shannon entropy method. Wavelet decomposition is chosen for determining the energy while Shannon method retrieves the entropy value. A classifier is defined and its performance level is tested. Accuracy up to 93% is achieved and time consumption is low when compared to other tsunami warning system. This result may contribute major part in the assessment of tsunami early warning system.