GEOSCAN Search Results: Fastlink


TitleBayesian inversion of ambient seismic noise for earthquake site response
AuthorDosso, S E; Molnar, S; Cassidy, J F
Source10th International conference on theoretical and computational acoustics, abstracts volume; 2011 p. 41-42
LinksOnline - En ligne
Alt SeriesEarth Sciences Sector, Contribution Series 20100270
Meeting10th International Conference on Theoretical and Computational Acoustics; Taipei; TW; April 24-28, 2011
Mediaon-line; digital
Subjectsgeophysics; earthquakes; earthquake studies; earthquake mechanisms; r waves; seismic waves; wave propagation; seismic interpretations; seismology
ProgramTargeted Hazard Assessments in Western Canada, Public Safety Geoscience
AbstractThis paper applies Bayesian inversion to Rayleigh-wave dispersion data derived from array measurements of ambient seismic noise to estimate the near-surface shear-wave velocity profile, an important property for characterizing earthquake site response as required for seismic microzonation. The Bayesian inversion algorithm, based on an efficient implementation of Metropolis-Hastings sampling of the posterior probability density (PPD), provides the most-probable shear-wave velocity profile together with a quantitative uncertainty analysis. Nonparametric estimation of the data error covariance matrix from residual analysis is applied, with rigorous a posteriori statistical tests to validate this covariance estimate and the underlying assumption of a Gaussian error distribution. The most appropriate model parameterization is determined using the Bayesian information criterion, which provides the simplest model consistent with the resolving power of the data. Parameterizations considered vary in the number of layers, and include layers with uniform, linear, and power-law gradients. Seismic noise data were collected at two sites in southwestern British Columbia, the area of highest seismic risk in Canada, to study the ability to recover the shear-wave velocity profile in different geologic settings. One site, on the Fraser River delta near Vancouver, is characterized by unconsolidated sediments several hundred metres thick, while the other site, in Victoria, has thin sediments over bedrock. Invasive measurements from seismic cone penetration and surface-to-downhole methods are used to assess the reliability of the noise inversion results, with excellent agreement at both sites. To extend the uncertainty analysis from shear-wave velocities to microzonation, the PPD sample obtained via Bayesian noise inversion is mapped into probability distributions for a variety of properties used to characterize site response, including building code site class, peak ground velocity/acceleration, and amplification and resonance spectra.