Inversion of equivalent linear soil parameters during the 2011 Tohoku earthquake, Japan
Résumé
On the afternoon of March 11th, 2011, an undersea megathrust earthquake of magnitude Mw 9 occurred off the Pacific coast of Tohoku, Japan, with an epicenter approximately 70 kilometers east of the Oshika Peninsula of Tohoku. Thanks to various Japanese networks, excellent quality data have been recorded. Particularly, many strong-motion data have been recorded with long durations and high peak ground accelerations and velocities (PGAs and PGVs, respectively) in Miyagi, Ibaraki, Tochigi, and Chiba Prefectures. Such long durations and high PGAs are probably characteristics of records lengthened and amplified by site effects (i.e., modification of the waves coming from the source due to the impedance contrast between the bedrock and the soft sediment basins). Owing to the large PGVs observed, the strain level sustained by the soft sediment is large and the response of the sediments is no longer linear. Soil nonlinearity is, however, a very complex phenomena and research on this topic is essential for the future seismic design input. Thanks to the data recorded during the 2011 Tohoku earthquake, different types of nonlinear soil behavior have been observed ranging from classic high-frequency de-amplification to liquefaction. Our study has for objectives to: - Check the validity of the linear elastic soil profiles of some borehole stations (e.g., KiK-net) by comparing their theoretical 1D spectral ratio with observed spectral ratios using weak motions. - Try to answer to these important questions by performing inversions of KiK-net and K-NET data: What is the soil profile of a site in terms of shear-wave (S-wave) velocity and damping factors to match the observed nonlinear response? What is the depth of sediments experiencing nonlinear response? Is this reduction consistent with standard laboratory tests (shear modulus reduction and damping ratio curves)? Is there any healing after nonlinear effects have passed? - Discuss the inversion results obtained by two different theories (i.e., surface-to-downhole spectral ratio and optimal H/V spectral ratio) and conclude if the optimal H/V spectral ratio (Kawase et al., 2011) technique gives satisfactory results. If so, a clear step forward would be done on the quantification of nonlinearity (through equivalent linear parameters). This would also validate the use of H/V ratio technique for inverting equivalent linear parameters and spread the domain of inversion over the free surface networks that would lead to a better knowledge of the underground, and a better assessment of the degree of nonlinear behavior.