Nonlinear Soil Response and Its Effects on Strong Ground Motions during the 2003 Miyagi-Oki Intraslab Earthquake

Abstract

We investigate nonlinear soil responses based on strong motion records at a large number of observation sites during the 2003 Miyagi-oki intraslab earthquake (Mw 7.0). First, we examine the efficiency of S-H/V (horizontal-to-vertical spectral ratio for the S-wave portion at the ground surface) method to identify the nonlinearity by comparing the results with those by the standard S-wave surface-to-borehole spectral ratio method. In the examination we propose a new quantitative index to measure the degree of the nonlinear soil response, DNL, which evaluates the gap between the spectral ratio for strong ground motion and that for weak ground motion. The DNL values by the S-H/V method as well as those by the surface-to-borehole spectral ratio method increase with observed peak ground acceleration (PGA) values at surface (100∼1000 cm/s2), reflecting the increase of the degree of nonlinearity. However, both DNL values at sites with large S-wave velocity of the surface layer (>300 m/s) do not show any increase even for large PGA values (∼800 cm/s2), indicating linear site response. From these facts we conclude that the S-H/V spectral ratio method is also efficient to identify the nonlinearity. Second, in order to examine the effects of nonlinear soil response on strong ground motions, we make the broadband strong motion simulation for the 2003 Miyagi-oki earth.quake by means of the empirical Green’s function (EGF) method. The synthetic waveforms for the horizontal components at the ground surface significantly overestimate the observed ones at stations with large DNL values; the synthetic PGA values are about two times greater than the observed ones. We confirm that the overestimation is attributed to the ignorance of the nonlinear soil response in the EGF simulation. Finally we briefly discuss the potential of the DNL value for studying the nonlinear soil response.