Aristotle University Of Thessaloniki (AUTH)
Συμβολή στην εκτίμηση της σεισμικής επικινδυνότητας με τη χρήση πολυμεταβλητών σεισμικών χρονοσειρών
Abstract
dc:descriptionThe investigation of the complex seismicity behavior constitutes a major scientific challenge and an indispensable component in improving our knowledge concerning seismogenesis and seismic hazard assessment. The strong earthquakes (e.g. M>6.0) as well as the aftershocks of them can cause catastrophic consequences with losses of human lives. The purpose of this doctoral thesis is the estimation of the future seismic activity for the Greek area, using the multivariate seismic time series. In the second and third chapter of thesis, the estimation of the future seismic activity is attempted through the construction of either a correlation network or a recurrence plot. The node of the seismic correlation network is each variable of the multivariate seismic time series that represents a seismic zone as the Greek territory is divided. Each observation of multivariate seismic time series captures either the sum of the earthquakes that occurred or the cumulative seismic moment that released at each node, respectively. The connections of the seismic correlation network among of its nodes are given by the Pearson correlation coefficient. On the contrary, the node of the recurrence plot is each earthquake and the connections among of its nodes are given when the difference of magnitudes is less than an arbitrary threshold. Having formed the seismic network, the estimation of the future seismic activity is achieved either through the use of network measures or through the study of its non-trivial topological properties, such as the small-world and scale-free structure. Therefore, the estimation of seismic activity is attempted by studying either the dynamic evolution of network values or the dynamic evolution of network structure in sliding time windows either before or after from main shocks or strong aftershocks, respectively, that occurred in the Greek territory spanning the period 1999-2018. It is revealed that the values of network measures are, in many cases, statistically significant as well as that the seismic network presents the structure of the small-world shortly before the occurrence of main shocks. In the fourth chapter of the thesis, the estimation of the future seismic activity is attempted through the canonical correlation analysis which calculates the correlation between multivariable time series by finding their strongly correlated linear combinations. The variables of the multivariate time series represent the magnitude of earthquake and its time succession, in days, until the occurrence of next earthquake. The canonical correlation analysis is used to examine whether there is statistical significance, using the appropriate statistical test, of the correlation between the last successive five earthquakes, per pair, that occurred before the occurrence of main shocks in the Greek territory spanning the period 1964-2018 for each seismic zone. The results reveal that the statistical significance of the correlation is present between the last five successive earthquakes, per pair, that occurred before the occurrence of main shocks in each seismic zone. In addition, the map for the probabilities of occurrence of the upcoming main shocks (Μ≥5.5) for each seismic zone, is presented, to help the estimation of the future seismic activity for the Greek area.
Degree
thesis:*- Grantor dc:publisher
- Aristotle University Of Thessaloniki (AUTH)
- Year dc:date
- 2020
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Chorozoglou, Dimitrios-Efstathios
- Χορόζογλου, Δημήτριος-Ευστάθιος
Subjects
dc:subject × 12Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/47713