VIBRIS - PhD thesis

The seismic tomography is a tool allowing to explore the Earth's interior. My PhD project consists in imaging the geologic structures located between 0 and 100 km depth below the Massif Armoricain and the Golfe de Gascogne. The interpretation of these images will allow a better understanding of the geodynamic evolution of western France. The technique used is the surface wave tomography by noise correlation. The data are obtained by using the seismic noise created by the oceans and continuously recorded by an array of seismic stations (see Fig. 1). The seismic noise correlation recorded at two stations allows, in good conditions, the emergence of a consistent signal which its characteristics are directly linked to the properties of the media located between the two stations (see Fig. 2). The compilation of all the measures obtained between each pair of station leads, after an inversion process, to a seismic velocity model of the studied region interpretable in term of geologic structures.

Figure 1: Regional array of broadband stations deployed on the Massif Armoricain. The red arrows indicate the pair of stations used to compute the noise correlation functions shown in Fig. 2. The direction of the arrows indicates the mean direction of the energy which contributes to the emergence of a signal in the positive times of the noise correlation functions shown in Fig. 2.

Figure 3: Symmetry of noise correlations shown in the time-frequency domain. The symmetry is estimated by the difference between the signal to noise ratio computed on the signal part corresponding to the negative times of the noise correlation and the one computed for the positive times (see Fig. 2). The calculations were done on the vertical components for the PY41-PY48 pair (see Fig. 1). A value equal to 0 corresponds to a perfect symmetry. The arrows indicate rapid changes of symmetry localized in time ( ~1 day) and in frequency (between 0.25 and 0.1 Hz) which occur the 28th of October and the 1st of November 2012. These observations show that the dominant direction of the noise energy flux (see Fig. 2) can by briefly disturbed by a noise energy flux coming from an opposite direction.