1095 Bulletin of the Seismological Society of America, Vol. 92, No. 3, pp. 1095–1101, April 2002 Stochastic Simulation of Strong-Motion Records from the 15 April 1979 (M 7.1) Montenegro Earthquake by Zafeiria Roumelioti and Anastasia Kiratzi Abstract Acceleration time histories, recorded during the destructive 15 April 1979 (M 7.1) Montenegro earthquake, have been simulated using a stochastic mod- eling technique for finite faults proposed by Beresnev and Atkinson (1997, 1998a). In this approach, the ground-motion amplitudes are simulated as a summation of stochastic point sources. The length of the fault was taken as 70 km and its width as 25 km, and the fault plane was divided into 13  5 elements. The applied meth- odology is tested against its ability to predict site-specific strong-motion records by the incorporation of mean frequency–dependent site-amplification factors, based on a gross characterization of the site class. The results show that the overall agreement between stochastic and recorded waveforms and spectra is quite satisfying. Never- theless, significant discrepancies exist at certain stations, implying that site-amplifi- cation functions play an important role in the simulation process. A repetition of the simulations combined with the use of site-specific amplification function estimated by the horizontal-to-vertical ratios technique improved the fit to the observed time histories and spectra. Introduction On 15 April 1979, at 06:19:40 UTC, a strong earthquake of M 7.1 occurred near the coast of Montenegro (in the for- mer Yugoslavia). According to several authors (e.g., Kociaj, 1980; Boore et al., 1981; Kanamori and Given, 1981; Ka- rakaisis et al., 1984; Baker et al., 1997), the fault-plane so- lution of the mainshock and the spatial distribution of its aftershocks imply that this earthquake involved thrust fault- ing on a plane striking northwest–southeast, nearly parallel to the Adriatic coast and dipping northeast. The earthquake caused extended damage in Montenegro and in the northern district of Albania, especially along the Adriatic coasts. Ninety-four people were killed in Montenegro and 35 more in Albania, while the injured reached 1172 and 328, respec- tively. After the earthquake, 80,000 people were left home- less (Papazachos and Papazachou, 1997). Strong-motion instruments located mainly within the state of Montenegro recorded the mainshock, as well as a significant number of foreshocks and aftershocks. These strong-motion data were recently disseminated by Ambra- seys et al. (2000) as part of a European strong-motion data- base. The availability of these data, as well as the destruc- tiveness of the 1979 event, provided the stimuli for simulating the recorded acceleration time histories. Since the information concerning the recording sites was rather lim- ited, as is the case for most European strong-motion stations (Ambraseys et al., 2000), we concentrated on the represen- tation of the local site responses. Therefore, we first used mean amplification factors based on a gross characterization of the site geology, and next we employed site-specific amplification functions estimated by horizontal-to-vertical (H/V) ratios. The technique used in this study for the simulations of strong-motion records is the stochastic finite-fault technique of Beresnev and Atkinson (1997, 1998a, 1999). In this method, the finite source is represented by a rectangular plane, which is subdivided into a certain number of elements (subfaults). Each subfault is treated as a point source, and each subevent has an x-squared spectrum. The ground mo- tion at an observation point is obtained by summing the con- tributions over several subfaults. A simple kinematic model of the Hartzell (1978) type is used to simulate the rupture propagation, which is assumed to start at the hypocenter and radially propagate from it. Propagation effects are empiri- cally modeled by using the observed regional dependence of ground-motion amplitudes and duration on distance. Data Data used in the present study include the rock and stiff soil records of the 1979 Montenegro event contained in the European strong-motion database (Ambraseys et al., 2000). Twenty-four accelerograms are used, obtained from five three-component stations installed at rock sites and three sta- tions at stiff soil sites. The classification of the sites is based