Seismotectonics and seismic hazard modelling for Southern Bulgaria

Abstract

The seismotectonics of southern Bulgaria and its surroundings have been reviewed and analysed and a catalogue has been compiled for the region. We show that for southern Bulgaria the seismicity characteristics, focal mechanisms and stress inversions: (a) confirm the hypothesis that the neotectonics of this region are similar to that of northern Greece and the northern Aegean Sea with presently N-S extension; (b) indicate that the seismic energy release is mostly related to normal faulting, with possibly some small strike-slip component in the Struma fault zone; (c) result for this region in a frequency magnitude relation that suggest characteristic magnitudes around M S = 7.0; and (d) indicate that seismic activity occurs in crustal fault zones rather than crisp fault planes. A probabilistic seismic hazard analysis (PSHA) for southern Bulgaria represents a typical case of seismic hazard for a tectonically complex region with large uncertainties in model parameters. We show, using a Monte Carlo approach, that large uncertainties in seismic characteristics have relatively little effect on the PSHA output, especially when compared with uncertainties associated with the attenuation function. Uncertainties in the seismotectonic model emerge strongest for the regions around Pazardjik and Jambol, the western and eastern tails of the Maritsa fault zone. For six important economic centers in southern Bulgaria we present the seismic hazard and their uncertainties with respect to attenuation and seismicity modelling. Our PSHA maps show, for example, regions with significant seismic hazard that exceed (with 50% confidence) a 1% annual probability of the peak ground acceleration exceeding 0.3 g. Future PSHA improvements for the region can be obtained by (in order of decreasing significance): (a) carefully mapped site responses or more accurate regional attenuation models; (b) additional constraints on the seismic zones; and (c) better constrained magnitude-frequency distributions (for example with independent regional strain rate information).