Investigating the Tsunamigenic Potential of Earthquakes from Analysis of the Informational and Multifractal Properties of Seismograms

Revealing the tsunamigenic potential of an earthquake is very challenging in regards to minimizing the casualties a tsunami can provoke. Thus, development of methodologies that can reliably furnish a early warnings of a tsunami is crucial. In order to accomplish this aim it is important to preliminarily identify the characteristics of seismograms that can be used to distinguish tsunamigenic (TS) earthquakes from non-tsunamigenic (NTS) earthquakes. In this paper P-wave time dynamic of 17 seismograms of TS earthquakes and 26 NTS seismograms are analysed by means of two advanced statistical tools: the Fisher-Shannon method and the multifractal detrended fluctuation analysis (MFDFA). Both methods are well suited to disclosing the inner time properties of complex signals, as seismograms appear to be. Using these two methods jointly, we defined a classifier, the performance of which was tested by means of the receiver-operating characteristic curve that plots true positive rate versus false positive rate. This classifier shows a discrimination power that can be considered acceptable in comparison with the devastating effects caused by a non-alarmed tsunami. Our findings indicate that proper choice of the classifier's threshold allows correctly identification of approximately 69 % of the NTS seismograms and approximately 76 % of the TS seismograms. The presented results presented may be helpful in addressing the complex problem of early tsunami warning.