ABSTRACT
The lifelines such as gas mains, oil pipelines, bridges, dams etc., extend over important distances so that their multiple supports undergo different seismic movements during earthquakes. This phenomenon, called spatial variability of seismic excitations, was the motivation of numerous studies which revealed that it was the result of the contribution of four mechanisms: (the effect of loss of coherency of seismic waves, the wave-passage effect, the attenuation effect, and the site-response effect) through a proposition of several generic models of functions of coherency. Because spatial variation of ground motion is influenced by earthquake source, propagation path and site condition, the coherency function models from seismic records of different arrays or different seismic records
of the same array are very different (Somerville et al. 1988, 1991; Santa-Cruz et al. 2000; Ding and Song 2010). However, it is very difficult for engineers to make sure which one is more appropriate for design, because most of these models are only valid for sites where the array accelerograms are employed. In 1997, Zerva and Harada simplified horizontal stochastic layers at a site as a single-degree-of-freedom system with random characteristics to study the effect of a soil stochasticity on the coherency function. They pointed out that the effect of soil layer’s stochasticity should also be incorporated in spatial variation models because the variability in the soil characteristics will reduce the coherency function at the stochastic layer predominant frequency. Their analysis is in agreament with the coherence-hole phenomena observed by Cranswick (1988) in his geophysics research in the
1980s of Liao and Li In the the effects of in ground motions are addressed; however, without sufficient details to address practical applications in engineering design and therefore acceleration time histories should be used in dynamic analysis.
