V41D-02:
Linking the dynamics of the 2000 dike intrusion at Miyakejima to the focal mechanisms and the statistics of the induced seismicity

Thursday, 18 December 2014: 8:15 AM
Luigi Passarelli1, Eleonora Rivalta1, Simone Cesca1 and Yosuke Aoki2, (1)Helmholtz Centre Potsdam GFZ German Research Centre for Geosciences, Potsdam, Germany, (2)University of Tokyo, Bunkyo-ku, Japan
Abstract:
Faulting in volcanic environment is the result of the interaction of pressurized fluid-filled conduits and cracks with the host rocks. Faulting styles are also influenced by the local and regional tectonic setting. Seismicity in volcanic areas usually shows complex patterns difficult to decipher and there is a need of physics-based models linking observations to the mechanics of fluid-filled sources. Magma-filled dikes often induce abundant seismicity in form of swarms thought to occur close to the propagating tip (or edges, in 3D) of the dike, where stresses are concentrated, and in areas where pre-diking seismicity was high. The spatial distribution and focal mechanisms of the earthquakes bear information on the interaction of the dike stress field and the tectonic setting of the area. Here we use data of the 2000 dike intrusion at Miyakejima volcano (Izu Archipelago, Japan) to study the relation between the shape of the dike and the faulting style of the induced seismicity. We find that the strike and rake angles of the focal mechanisms are correlated with each other and are consistent with the dike shape and optimally- oriented faults according to the expected Coulomb stress changes from the 3D dike-induced stress field. We perform a clustering analysis on the FM solutions and relate them to the dike stress field and to the scaling relationships of the earthquakes. We calculate the frequency-size distribution of the clustered sets finding that focal mechanisms with a large strike- slip component are consistent with the Gutenberg-Richter relation with b-value about 1. Conversely, events with large normal faulting components deviate from the Gutenberg-Richter distribution with a marked roll-off on its right-hand tail suggesting a lack of large magnitude events (M>5.5). This can be interpreted as resulting from a limited thickness or localized weakness of the layer of rock above the dike, where normal faulting is expected.