Monte Carlo Simulation to Study the Spatial Variation of Ground Motion Associated with Basin Heterogeneities

Abstract

Basin effects cause a complex wave interference inside a basin which can be attributed to basin versus bedrock material contrast and edge effect. This will have a significant impact on spatial variation, duration, and intensity of surface ground motion (SGM) during an earthquake. While important, the lack of sufficient information about material properties and stratigraphy of a basin prevents accurate simulation of the phenomena, particularly in high frequency regime. Stochastic analysis and the Monte Carlo technique are suitable approaches to address this issue, where basin material is represented by a correlated random field. In this study, We use a 2D finite element analysis of an idealized-shaped basin subjected to a vertically propagating SV plane wave and investigate the spatial variation of SGM associated with basin effects by assuming a correlated random field to represent basin material. We generate a random medium by adding perturbations to a homogeneous domain with various correlation lengths, coefficient of variations, and autocorrelation functions to evaluate their contribution to SGM. Our results show a difference between the output of homogeneous and stochastic models, where we conclude that the former would not represent basin response, especially in the high-frequency regime correctly. Among the parameters we consider, the coefficient of variation has the most influential impact on surface acceleration. We observe that increasing this parameter decreases the mean value of surface amplification while its standard deviation increases. In addition, correlation length affects the standard deviation of surface acceleration, but it does not significantly impact the mean amplification. As for the autocorrelation function, where we consider von Karman, Gaussian, and exponential, the results show that the trend of surface amplification does not change by choosing a different autocorrelation function. Finally, by comparing the 2D basin versus 1D layered medium, we show that one cannot accurately capture basin response by using a 1D analysis for seismic hazard quantification.

Files

Additional details