Influence of Ground Conditions to the Time-Varying Frequency Content of Earthquake Ground Motions

Abstract

This study aims to characterize the influence of local site conditions on the non-stationary frequency content of earthquake-induced ground motions, as captured by the wavelet-based alpha index. Local site conditions are defined as the combined effect of soil stiffness, represented by the shear wave velocity of the soil stratigraphy, and the severity of earthquake shaking. The research involves constructing comprehensive databases of computer-simulated and field-recorded ground motions for a wide range of ground conditions, correlating alpha with conventional ground motion parameters widely used in earthquake engineering, and quantifying differences in the non-stationary frequency content between bedrock and surface ground motion time-series.

The methodology involves conducting site response analysis using DEEPSOIL software to simulate the propagation of artificial bedrock accelerograms through various soil profiles with differing properties. Additionally, recorded accelerograms were carefully selected from the KiK-net vertical seismometer array in Japan to complement the simulated dataset. Statistical regression analyses were then applied to ground motion properties, including the alpha index, extracted from the combined databases of simulated and recorded bedrock and surface accelerograms to achieve the study's objectives.

The study reveals key insights into the influence of ground conditions on the temporal nonstationary frequency content of earthquake-induced ground motions, quantified using the alpha index. Alpha is established as a meaningful scalar metric for capturing the average rate of temporal change in the mean instantaneous period during the strong motion phase of ground motions. For computer-simulated ground motions generated through DEEPSOIL site response analyses, the findings show that surface alpha increases with softer soils and higher peak bedrock acceleration, with greater variability in alpha observed under these conditions. The results also highlight that record-to-record variability has a more pronounced effect on alpha than uncertainties in soil properties. Statistical analyses of recorded ground motions from the Japanese KiK-net database demonstrated that surface alpha does not correlate with PGA but exhibits a positive correlation with average period, peak ground velocity, and soil shear strain proxy, with the significance of its correlation with soil stiffness increasing under higher seismic intensity.

Overall, the herein reported results and findings support the hypothesis that local soil/ground conditions contribute to the potentially increased seismic demands on structures not only by amplifying the motion of the ground and/or shifting its average frequency content, but also by changing the temporal evolution of ground motion frequency content.

Date of Award	9 Dec 2024
Original language	American English
Supervisor	Agathoklis Giaralis (Supervisor)