Comparison of Measured and Computed Lateral Spread Displacements for Mw8.8 Maule Chile Earthquake Case Histories

Abstract eng:
Because earthquakes larger than Mw8.0 are relatively rare, empirical models for predicting lateral spread displacement have generally focused on earthquakes of lesser magnitude. As a result, these models have not been calibrated for larger magnitude events. Nevertheless, engineers are often required to estimate lateral spread displacements for Mw8+ events and some engineers have extrapolated in employing current lateral spreading models for predicting displacements when designing utilities, bridges, and piers. In this paper, two lateral spread case histories from the Mw8.8 Maule Chile earthquake in 2010 are used to help understand the strengths and weaknesses of various empirical models for predicting lateral spread displacements. In addition, recommendations for improving the accuracy of these models are suggested in some cases. In this paper, predictions from five empirical models commonly used in engineering practice are compared with measured displacements. The model that best matched the measured displacements from the Maule Chile case histories uses local attenuation relationships to predict ground motions rather than simple magnitude and distance terms. This approach makes it easier to apply the model to any seismic region. In contrast, models that use magnitude and distance directly are often rather difficult to apply for cases involving subduction zone earthquakes where the closest horizontal distance to the zone of energy release may simply be zero over large area of an affected region producing poor discrimination of ground motion. In addition, appropriate site-to-source distances can vary greatly between different seismic regions and for different faulting mechanisms. For this reason, models that depend on an internal source-to-site distance show less promise with large subduction zone earthquakes throughout the world. Models with site-to-source distances are most accurate in the western United States and Japan because the case histories for these models came from these countries. Nevertheless, soil properties such as the T15 thickness, the fines content, and the mean grain size all seemed to be useful in predicting displacement. Models that use shear strains from lab data typically over-predicted measured displacements but using a strain-reduction factor with depth significantly improved the accuracy of the predictions. Similarly, current CPT-based empirical equations generally over-predicted measured displacements by more than a factor of two and strain reduction factors based on depth were necessary to improve agreement with measured results.

• Export as: BibTeX | MARC | MARCXML | DC | EndNote | NLM | RefWorks
• Add to your basket