Supervisors info:
Γεώργιος Καβύρης, Επίκουρος Καθηγητής, Τμήμα Γεωλογίας και Γεωπεριβάλλοντος, Εθνικό και Καποδιστριακό Πανεπιστήμιο Αθηνών
Summary:
The splitting phenomenon (during which a shear – wave is divided in two orthogonally polarized components travelling with different velocities) is attributed to the anisotropy of the propagation medium. Its observation is relying on identifying two main parameters: the polarization direction of the fast shear – wave (φ) and the time – delay between the travel – times of the two components (td). The current MSc thesis aims to develop a software suite (in the Python programming language) for determining the splitting parameters from local events, in the upper crust. Even though there are various programs available, they do not offer the integrated use of manual, semi- and fully automatic methods. The developed software (Pytheas) aims to harmonize the implementation of these techniques. User - friendliness is a basic feature of the program. The application incorporates three methods: (a) visual inspection of particle motion diagrams, (b) eigenvalue method and (c) rotation – correlation method While (a) is a manual technique and requires the complete involvement of the user, (b) and (c) are semi – automatic and depend upon the manual determination of an analysis window. This selection is fully automated by using Cluster Analysis, in combination with either (b) or (c). The Pytheas software connects all these techniques through a friendly and intuitive graphical user interface. For the evaluation of the measurements, a grading system was developed, based on the noise in the signal and the errors of the two splitting parameters.
For the evaluation of Pytheas, local data from three different areas were used: (i) the Western Gulf of Corinth (tectonic regime) in Greece, (ii) the Santorini Volcanic Complex (volcanic regime) in Greece and (iii) the Wellington Oil Field in Kansas, USA (induced seismicity). Cluster analysis was utilized in selecting the optimal analysis windows for the eigenvalue method, in over 2,400 event – station pairs, with a maximum angle of incidence equal to 35°. A total of 516 measurements with acceptable grades were determined. Results, compared with corresponding ones from previous studies that used the manual method, showed that the automated scheme is significantly reliable. However, the more intense scatter in both φ and td must be investigated, possibly by refining the automatic grading algorithm.
In addition, the Shear – Wave Splitting Tomography method was examined, by using manual measurements of past studies (due to their increased number of observations) in the Western Gulf of Corinth and the Santorini Volcanic Complex. Results showed that, to adequately interpret the tomography, additional information is required, given that the determined anisotropic structures do not agree with static geologic and tectonic features of the upper crust. The investigation of the stress distribution (by pressure measurements in wells or the inversion of focal – mechanisms) could be used towards this goal, to relate the tomography to dynamic phenomena.
Keywords:
Seismic anisotropy, Shear-wave splitting, Upper crust, Cluster analysis
