National Observatory of Athens, Greece
It has long been known that ground motion at lower frequencies can be amplified due to stiffness (shear-wave velocity) contrasts and geometry of near-surface layers. Rock, on the contrary, is usually assumed to have an ideal response, exhibiting negligible amplification or attenuation.
Recently, a more complete picture has been built. At higher frequencies, ground motion on hard rock can he stronger than on soft rock, if the reduction in attenuation (damping) is stronger than the amplification from contrasts. Stiffness itself is not a good proxy for damping in rock, and rock sites typically considered as similar can exhibit significant variability. In short, we are no longer justified in assuming all rock sites behave the same.
The implicit treatment of rock response as ideal has strong implications for seismology (reference station definition, seismic source issues), cascading to engineering seismology (empirical ground-motion prediction/adjustment, simulations, probabilistic and deterministic hazard), and earthquake engineering (critical infrastructure safety, input motions for structural analysis).
This talk gives an overview of recent advances in characterising attenuation at high frequencies. It illustrates the effect that differences in attenuation can have on the response spectral features of rock sites belonging to the same class. While this leads to the need for site-specific design in critical structures, it is suggested that even for standard design we need to revisit and possibly refine rock-site classification in codes.