Faire avancer la sûreté nucléaire

La Recherchev2


Simulating strong ground motion with the 'k(-2)' kinematic source model: An application to the seismic hazard in the Erzincan basin, Turkey.



Email :

Mot de passe :

Berge-Thierry, C; Bernard, P; Herrero, A. JOURNAL OF SEISMOLOGY, 5: (1) 85-101.

Type de document > *Article de revue

Mots clés > séismes, aléa sismique, séismes

Unité de recherche > IRSN/DEI/SARG/BERSSIN

Auteurs > BERGE-THIERRY Catherine

Date de publication > 01/01/2001


We present a seismic hazard application of a kinematic broad-band rupture model. This model is based on the k-square dislocation distribution concept (Herrero and Bernard, 1994). Synthetic seismograms are calculated in the far-field approximation with a layered velocity medium for the 13 March 1992 Erzincan earthquake. With a parametrization of the source constrained by other studies, the far-field contribution correctly fits the recorded strong ground motion, which presents a 0.5 to 2 Hz dominant frequency range. As the k-square model is particularly well adapted to synthetize realistic strong-motion at short distances from the fault, it is a reliable tool for calculating seismic hazard maps around active faults. We thus present a synthetic peak ground acceleration map associated with the 13 March 1992 activated fault, for a 60 km x 60 km region around the epicenter taking into account a smoothed velocity structure of the basin in agreement with the absence of significant site effects related to 1D resonance deduced from the aftershock records study. This map is compared with several post seismic reports: macroseismic intensities, detailed distribution of damage, and soil cracking and liquefaction. Our model shows that the values of the peak acceleration and velocity can explain the dominant spatial distribution of these effects, which concentrates in a narrow band along the activated segment fault, and in particular at its southern extremity. These results enable us to present such maps for hypothetical future earthquake ruptures, located on the major visible or inferred active fault segments in and around the basin. The effects of these potential sources are analyzed in relation to the 1992 event effects in order to eliminate unknown site responses. We show that the southern part of the basin is particularly exposed because of the presence of strike-slip faults, and that the western part of the basin would suffer a significantly higher strong motion level than during the 1992 event with the activation of moderate sized normal faults evidenced on the southwestern edge of the basin.