عنوان مقاله [English]
By increasing the number of earthquake records, it has been made obvious that the dynamic characteristics of ground motion between different stations (even stated in a specific zone) are considerably different. This distinction is further highlighted in surrounding areas due to the epicenter. Seismic events all over the world have shown that ground motion in the vicinity of causative faults (within a distance of 15 km from the fault) may be categorized as a large-velocity pulse and a large-displacement fault, which have the potential to cause considerable structural damage. Consequently, the main cause of long-period pulse formation is the cumulative effect of shear wave propagation along the fault rupture, according to seismological investigations. Near-fault ground motion is severely affected by fault mechanisms, rupture propagating directions relative to the site and, finally, the permanent deformation of the ground. These parameters create two significant effects; directivity and the fling step, which should be taken into account when estimating ground motion in the vicinity of causative faults. Forward-directivity depends on the rupture mechanism and slip direction relative to the site. It is characterized by a large pulse occurring at the initiation of the record and oriented in a perpendicular direction relative to the fault plane. In contrast, the fling-step is affected by tectonic deformation in the fault and commonly generates permanent static displacement, which occurs parallel to the strike of the fault for strike-slip events, and normal-to-fault direction for dip-slip earthquakes. In this study, an innovative mathematical model is developed for the prevalent pulse modeling of near-fault records during the 1978 Tabas and 2003 Bam earthquakes. The proposed model is capable of simulating the long-period portion of near-fault records with a high level of precision. Simulated pulses for the ground motion components of the Bam earthquake and their elastic response spectra have good compatibility with that of actual records. Furthermore, the generated elastic response spectra for the simulated pulse and corresponding characteristics for actual records of the Tabas earthquake indicate good agreement just for long-period areas.