Earthquake Ground-Motion Simulation including Nonlinear Soil Effects under Idealized Conditions with Application to Two Case Studies

Abstract

Nonlinear soil behavior in sediments remains one of the most relevant and challenging aspects in site characterization. Geotechnical engineers began studying soil changes in the shear modulus and damping ratio with increasing strain amplitude in the late 1960s and early 1970s (e.g., Seed and Idriss, 1969; Hardin and Drnevich, 1972). However, the influence of soil nonlinearities on the ground motion, in places other than those involving liquefaction, was not fully accepted by seismologists until observations made after the 1989 Loma Prieta and 1994 Northridge earthquakes, and from monitoring projects such as the SMART1 and SMART2 arrays in Taiwan (Field et al. , 1998; Aki, 2003). These events and arrays showed that soil nonlinearities caused permanent deformations away from the source, reductions in soil‐to‐rock ratios by average factors of 2 with respect to those of smaller events, and reductions in shear moduli and shear‐wave velocities of up to 80% and 50%, respectively (e.g., Wen, 1994; Beresnev et al. , 1995; Field et al. , 1997). Soil nonlinearities seem to be more prominent in frequency ranges from 0.5 to 4 Hz (e.g., Darragh and Shakal, 1991; Field et al. , 1997; Beresnev, 2002), and have been associated with peak ground acceleration values above 0.1–0.2 g and shear strains larger than 10−5 to 10−4 (e.g., Chang et al. , 1989; Chin and Aki, 1991; Beresnev and Wen, 1996; Trifunac and Todorovska, 1996). Modeling and simulation of nonlinear soil behavior and effects also date back to the 1960s and 1970s (e.g., Idriss and Seed, 1968; Schnabel, Seed, and Lysmer, 1972; Joyner and Chen, 1975). Early 1D and 2D studies were mostly done using equivalent linear models. Although these models do not capture all the characteristics …