Search

International audience; A seismic hazard map proposed as part of a new building code for Romania is presented here on basis of the recommendations in EUROCODE 8. Seismic source zones within an area of about 200 km around Romania were constructed considering seismicity, neotectonics and geological development. The probabilistic seismic hazard assessment in terms of intensities is performed following Cornell (1968) with the program EQRISK (see Mc Guire, 1976), modified by us for use of intensities. To cope with the irregular isoseismals of the Vrancea intermediate depth earthquakes a factor ? is introduced to the attenuation law (Kövesligethy, 1907). Using detailed macroseismic maps of three earthquakes ? is calculated by fitting the attenuation law to observed intensities, i.e. to local ground conditions. Strong local variation of ? is avoided by a gridding of 0.5° in longitude and 0.25° in latitude. The contribution of the Vrancea intermediate depth zone to the seismic hazard at each grid point is computed with the corresponding representative ?. A seismogenic depth of 120 km is assumed. The final seismic hazard is the combination of both contributions, of zones with crustal earthquakes and of the Vrancea intermediate depth earthquakes zone. Calculations are done for a recurrence period of 95, 475 and 10000 years. All maps show the dominating effects of the intermediate depth earthquakes in the Vrancea zone, also for the capital Bucharest.

High-frequency records of nine low magnitude shallow earthquakes of the Vrancea (Romania) seismic region are inverted for the seismic moment tensor (MT). An approach is suggested regarding how to obtain at least a rough estimate of the MT when the information on the structure of the crust is poor. Here simple 1-D layered models are used in the Green’s function synthesis despite the fact that the structure of the region is undoubtedly very complex. Different 1-D models were used for different source-station paths to approximate lateral variations. Record of a station located on a ray path which crosses a structure differing substantially from a 1-D model may however bias the retrieved MT essentially. Therefore, we did not collectively process all records, but subsets of stations separately. We check the consistency of the MTs resulting from these individual ‘‘bootstrap’’ solutions, and reject those which differ substantially, assuming that this is due to the oversimplification of the forward modeling. Thereafter, the averaged moment tensor yielded by the consistent subsets is accepted. Moreover, the distribution of the T, N and P axes from the moment tensors used for averaging provides a rough estimate of their reliability. Following this simplistic procedure, we found seven acceptably constrained solutions among nine events processed. Their P axes are compared with the general trend of the stress in the area: three comply with it, the others should be attributed to the complex stress field active in the region.