Blind prediction of the site effects at Ashigara Valley,Japan, and its comparison with reality

Weak and strong ground motions were numerically predicted for three stations of the Ashigara Valley test site. The prediction was based on the records from a rock-outcrop station, one weak-motion record from a surface-sediments station, and the standard geotechnical model. The data were provided by the Japanese Working Group on the Effects of Surface Geology as a part of an international experiment. The finite-difference method for SH waves in a 2-D linear viscoelastic medium (a causalQ model) was employed.Comparison with the real records shows that at two stations the predictions fit better than at the third one. Strangely, the two better predictions were for stations situated at larger distances from the reference rock station (one station was on the surface, the other in a borehole). The strong ground motion (the peak acceleration of about 200 cm s^–2) was not predicted qualitatively worse than the weak motion (8 cm s^–2). A less sophisticated second prediction (not submitted during the experiment), in which we did not attempt to fit the available weak-motion record at the sedimentary station, agrees with the reality significantly better.