Model dimensionality effects on the amplification of seismic waves

Despite recent advances in numerical methods and computer architectures that make it ever more practical to obtain computationally the surface response to idealized or realistic seismic events, while fully accounting for three-dimensional effects due to topography or to heterogeneities, reliance on one-dimensional models persists. As discrepancies between computed and recorded responses still remain, in this study we aim at highlighting the effect the model dimensionality choice has on the discrepancies, in the presence of topographic features and/or heterogeneity. First, we briefly discuss the components of an integrated seismic-motion simulator that deploys best-practice tools for the study of wave amplification in arbitrarily heterogeneous sedimentary basins, while also accounting for topography. Then, we report numerical experiments in two and three dimensions for various prototype topography-endowed and layered domains, and compare the motion amplification/de-amplification patterns against one-dimensional simulations, in order to quantify the effects model dimensionality has on surface motion. We conclude that one-dimensional models greatly underestimate the effects of topography and heterogeneity on the amplification of seismic waves; two-dimensional models fair better, but, in general, they too underestimate the response. It appears that, in the presence of topography and complex stratification, there is no suitable alternative other than three-dimensional models to account for reasonable estimates of motion amplification to guide the design of earthquake-resistant structures.