The velocity structures of southern and central part of Sweden have been derived with the local tomography (LET) method. The region has been divided into two study areas and the datasets come from the P- and S-wave traveltimes recorded by the Swedish National Seismic Network (SNSN). Man-made explosions and earthquakes occurring over the period of 5 years and 10 years, respectively, within the study areas have been used. One-dimensional starting models were derived based on an a priori model obtained from the SNSN, that were later used for starting models in the inversion for the 3-D crustal structures of the study areas. Attempts were also made to invert for Moho topography in the areas. The study areas are found to have an upper-crustal thickness of approximately 20 to 25 km and the Moho boundaries vary from 42 to 46 km in depth. The Vp/Vs ratios varies from about 1.68 to 1.78. The LET method appear to resolve the different between the Sveconorgwegian and Svecofennian orogen regions, but the stations and sources are too sparsely distributed for higher resolution models. The seismicity in the study areas are distributed in two distinctive depth ranges. The focal depth of the SNSN catalogued earthquakes concentrated in approximately 5 km and 15 - 20 km depth. Relocations of the earthquakes using a global search method reduced this tendency. The results also show that using 3-D models produces less biased results than using 1-D models with the same relocation method.

Induced seismicity is a common phenomenon that occurs as soon as the stress state in the subsurface is externally altered in a way that faults are destabilized. It is especially problematic in stable tectonic regions where the area is not used to earthquakes; the infrastructure is not built to withstand ground movement and thus when the induced seismicity occurs damage can follow. In this thesis, mining-induced seismicity has been studied at the Dannemora mine, located in central Sweden, with the aim to locate the seismicity and gain understanding of its occurrence and behavior. The mining company, Dannemora Mineral AB, provided with blasting locations and times, as well as maps over the mine's orebodies and stopes. Seismic data acquired between 01 July 2014 - 25 March 2015 from 4 temporary seismic stations, deployed in the summer of 2014 surrounding the mine, along with 8 SNSN stations was analyzed. The project encompassed field work and processing of the data, which involved different methods to investigate the characteristics of the mine's seismicity: Statistics were kept to record the activity rate of the seismicity over time; spectral analysis was used to study the frequency content of the seismicity; particle motion plots were constructed to identify body-phases in the seismicity; Local Earthquake Tomography was used to upgrade the velocity model of the mine and to relocate the induced seismicity with more accuracy; cross-correlation was used to find events originating from similar sources; and finally, magnitude analysis was used to compare the different types of seismicity within the mine. Three main types of induced events were observed in the mine: low-frequency events with clear first arrivals, emergent events with long duration, and high-frequency events that could either have clear first arrivals or emergent-like with long durations. Through the analysis of their characteristics, they were linked to different types of rockbursts. The low-frequency events were linked to both reactivation of fault zones triggered by the mine activity, and rockbursts within the mine directly related to the mining. The emergent and high-frequency events were also linked to rockbursts directly related to the mine activity, e.g. ejection of rock from the tunnel walls or arch collapses in stopes.

La sismicité sous la mer Tyrrhénienne méridionale et le sud de l'Italie est essentiellement attribuée à la subduction de la lithosphère océanique Ionienne qui s'enfonce dans le manteau Tyrrhénien jusqu'à 500 km de profondeur. Les tomographies proposées à ce jour n'ont pas une résolution suffisante pour distinguer les différentes structures géologiques à grande profondeur, laissant donc encore de nombreuses questions ouvertes concernant la géodynamique régionale. Dans ce travail on a en développé des aspects méthodologiques qui mettent en œuvre une nouvelle technique de post-processing appelée WAM (Weighted Average Model). Cette méthode nous a permis d' obtenir des modèles de vitesse à haute résolution de la région sud Tyrrhénienne en réalisant une tomographie sismique 3D des ondes P et S. Avec WAM on a ainsi pu reconstruire la géométrie 3D du slab Ionien et proposer un nouveau scenario géodynamique pour la région basé sur de considérations pétrologiques.