ALBERT

Description: We analyze the broadband recordings from a borehole station installed in the bedrock underneath the Po basin (Northern Italy) and from a second station temporarily deployed on the surface at the same site to determine the surface sedimentary layer response and test the robustness of the different empirical techniques used to estimate local site effects. We investigate the site amplification by evaluating the surface-to-downhole spectral ratios and the horizontal-to-vertical spectral ratios on both weak-motion events and microtremors. All the different empirical methods, except the horizontal-to-vertical spectral ratios on weak motions, reveal the main amplification peak at about 0.8 Hz, although the amplification factors are not always in agreement; the 0.8 Hz peak is the fundamental mode of the site, due to the upper 130-m Quaternary alluvium column (V (sub s) approximately 300 m/sec). We compare these experimental spectral functions with the theoretical 1D site response including and not including the borehole site response: first we evaluate the theoretical transfer function at the surface then at depth and take the smoothed ratio. The shallow velocity and attenuation structure at this site is assumed from previous studies. We attempt to account for the effect of the reflected wave field in the borehole recordings in two different ways: first we include the reflected wave field in the theoretical response function taking the theoretical transfer function's ratio (Hst/Hbt) and comparing it to the experimental earthquake ratios of horizontal components recorded at surface and at depth (Hs/Hb). Alternatively we multiply the Hs/Hb earthquake ratio by the coherence and compare it to the surface response function. This second approach does not provide a good fit to the amplitude of model response, even if the shape of the ratio is in agreement with the theoretic response, suggesting that multiplying by the coherence is not a good way to correct for the effect of reflected phases. To further understand the goodness of our theoretical 1D model (Hst/Hbt theoretical evaluations), we simulate the propagation through the shallower soil structure of the seismic waveforms recorded by the borehole sensor for a couple of events, obtaining reasonably good agreement with the surface time histories both in phase and amplitude. This test site and the results of this study are useful for understanding the influence of sedimentary layers on the ground shaking in the Po basin, the largest and most populated alluvial basin in Italy.