SEAR · Site Effects Assessment for seismic Regulations by developing and validating physically based methods

Site effects assessment for seismic regulations by developing and validating physically based methods

Given the impact on human loses and injuries and the economical impact from major earthquakes, seismic risk mitigation is a primary objective of research activities in the seismological and earthquake engineering fields. Surface geology can drastically exacerbate damage (e.g. Mexico City in 1985). These local effects, so called site effects, depend on the geological conditions, both lithological and structural and may produce a strong modification in amplitude level, spectral content and time duration of earthquake ground motions. These effects may increase the seismic hazard level even in areas with moderate seismicity, thus improving the way such local hazard modifications are accounted for in seismic risk estimation is a major concern. The challenge is to develop inexpensive, simple though reliable methods to estimate site effects. Thus the first SEAR project goals were to improve the site characterisation together with the site effect assessment. Indeed site characterisation might be directly used for site effect assessment. The most used physical parameter is Vs30, the time-average shear wave velocity of the first 30 m. A new parameter is rising to be used for site effect assessment; it is the fundamental resonance frequency f0. My collaborative work with D. Boore and E. Thompson (Boore et al. 2011) allows to compute the classical Vs30 from smaller penetration depth (i.e. between 5 and 25 m). In the same time, the analysis of Greek data (Cadet and Savvaidis 2011) also results in interesting relation to derive Vs30 from the dispersion curve. This dispersion curve depicts the variations of the surface wave velocity with frequency and can generally be measured only over a limited frequency band. In the Greek study, the dispersion curves were obtained for ambient vibration recordings. The analytical function in this case was derived from empirical data. In a second step, numerical data were studied in Cadet and Bard (2011) to derive a more robust relationship to get Vs30 from the dispersion curve. Those different studies reply to the first goal of the SEAR project towards a better site characterisation. Then the second goal was fulfilled thanks to two different approaches: either to express the site effect through the combination of Vs30 with f0 or by using directly the dispersion curve. Both solutions are innovative because they propose to assess site effect from easily available information, while keeping the frequency information that is usually neglected in the simple seismic code regulation. The first solution is based on Japanese data and fully described in two articles (Cadet et al. 2012a and b). After explaining the huge treatment realised in the KiK-Net data, an analytical function is proposed to assess site effect from only two parameters Vs30 and f0. It was proved that this solution is considerably more accurate than the Euro-code 8 solution. The second proposition is more original as the dispersion curve was never yet directly used for such a purpose. This proposition was present to a national conference (Cadet et al. 2011). It shows that only few high frequency samples are needed to obtain a site effect assessment as good as the one obtained with the classical Vs30 proxy. However the development of an analytical relationship between the dispersion curve and site effect is still under development.

Data: CORDIS, © European Union