Parametric analysis of the seismic response of a 2D sedimentary valley: implications for code implementations of complex site effects

A detailed 2D model has been constructed and validated for Euroseistest valley, in northern Greece. We take advantage of this model to investigate what parameters, in addition to surface soil conditions (obviously the most important parameter), can be used to correctly characterize site response in a 2D structure. Through a parametric analysis using 2D numerical simulations for SH waves, we explore the differences between the computed ground motion for different simplifications of the valley's structure. We consider variations in the velocity structure within the sediments, and variations of the shape between sediments and bedrock. We also compare the results from different 1D models reflecting current approaches to the determination of site response. Our results show clearly that, in the case of Euroseistest, site response owes fundamentally to its closed basin shape because it is largely controlled by locally generated surface waves. Thus, in terms of predicting site response, a rough idea of its shape ratio and of the average mechanical properties of the sediments are better than a very detailed 1D profile at the central site. Although the details of ground motion may vary significantly between the models, the relative amount of surface waves generated in the 2D models seems to be relatively constant. Moreover, if we quantify the additional amplification caused by the lateral heterogeneity in terms of the 'aggravation factor' introduced by Chávez-García & Faccioli [7], a roughly constant factor between 2 and 3 seems to appropriately take into account the effects of lateral heterogeneity. Of course, a correct estimate of the overall impedance contrast is necessary to correctly predict the maximum amplification, a caveat that also applies to 1D models. In this sense, Euroseistest rings an alarm bell. In this valley the more significant impedance contrast lies at about 200 m depth, and it is missed both by consideration of the average shear wave velocity of the first 30 m (the Vs30 criterion) or using the detailed velocity profile down to a depth where a shear wave velocity larger than 750 m/s is found. Our conclusions indicate that, in order to improve current schemes to take into account site effects in building codes, the more to be gained comes from consideration of lateral heterogeneity, at least in the case of shallow alluvial valleys, where locally generated surface waves are likely to be important.