Records of Extreme Ground Accelerations during the 2011 Christchurch Earthquake Sequence Contaminated by a Nonlinear, Soil–Structure Interaction

Abstract

ABSTRACT Ground-motion records are critical for seismic hazard assessment and seismic design of buildings and infrastructures. Large (>1g), asymmetric vertical accelerations (AsVAs) have been observed at strong-motion stations during recent earthquakes. However, it is not clear whether all of the observed AsVAs reflect actual ground shaking or the interaction of a building structure and underlying ground. Here, we investigate the cause of large AsVAs recorded at several seismic stations in Christchurch, New Zealand, during the 2011 Mw 6.2 Christchurch earthquake. We first define three metrics and quantify the degree of waveform asymmetry in all available records from nearby M>3 earthquakes. Histograms of the metrics show greater waveform asymmetry for larger accelerations at these stations, which is consistent with the prediction of a nonlinear, soil–structure interaction associated with the elastic collisions of a foundation slab onto the underlying soil. We then use finite-element models to examine the occurrence of the nonlinear, soil–structure interaction at these stations during the Mw 6.2 mainshock and Mw 5.6 aftershock of the 2011 Christchurch earthquake. The parameters of the numerical models are constrained by site investigation of selected stations. We find that numerical simulations closely reproduce the large AsVAs recorded at stations HVSC and PRPC, suggesting that these ground-motion records were contaminated by the nonlinear, soil–structure interaction. Seismic sensors located near the corner of a concrete slab are shown to be more prone to this phenomenon. Our results further suggest that artificial recording of large AsVAs due to the nonlinear, soil–structure interaction can be mitigated if a seismic sensor is placed closer to the center of a foundation slab. The analytical procedure presented in this study may be useful in identifying the occurrence of AsVAs elsewhere and in assessing whether AsVAs are caused by the nonlinear, soil–structure interaction.