抄録
Near-fault strong motions from the asperity model are evaluated by the Kamae and Irikura's (KM) and Takemura and Ikeura's (TI) method. KM explicitly considers the asperity, whereas TI assumes uniform heterogeneity on the fault. In the period range shorter than 1 sec, both methods give the same amplitude of response spectra. On the contrary, remarkable difference can be found at the predominant period of 2 sec, in particular, in the direction of rupture propagation. This difference is attributed to the rupture directivity effect caused by the asperity. It is crutial to understand the difference when the stochastic methods are applied to the prediction of strong motions.