GEOSCAN Search Results: Fastlink


TitleBayesian trans-dimensional inversion of surface-wave dispersion for earthquake site response estimation in Northern British Columbia, Canada
AuthorGosselin, J M; Cassidy, J F; Dosso, S E; Brillon, C
SourceSeismological Research Letters vol. 87, no. 2B, 2016 p. 449
Alt SeriesEarth Sciences Sector, Contribution Series 20160056
PublisherSeismological Society of America
MeetingSeismological Association of America Annual Meeting; Reno, NV; US; April 20-22, 2016
Mediapaper; on-line; digital
File formatpdf
Subjectstectonics; Science and Technology; s waves; earthquake studies; earthquakes; modelling; array seismology; Bayesian inversion; seismic noise
ProgramWestern Canada Geohazards Project, Public Safety Geoscience
LinksDownload entire publication / téléchargement de la publication au complet (21.7 MB)
AbstractKnowledge of the shear-wave velocity (Vs) profile over the upper 10s of metres of the soil column is important for characterizing the expected ground response to earthquake shaking at a specific site. Non-invasive and passive methods based on recording ambient seismic noise are increasingly popular for estimating Vs structure with minimal cost and site disruption. This study applies an advanced Bayesian inversion methodology to estimate Vs structure from surface-wave dispersion measurements processed from passive seismic array recordings using frequency-wavenumber analysis. A fully non-linear trans-dimensional (trans-D) Bayesian inversion based on parallel tempering is used to estimate Vs models and associated model uncertainties. In trans-D inversion, the model parameterization (e.g., number of layers of the Vs profile) is included as an unknown which is sampled probabilistically, accounting for the uncertainty of the model parameterization in the uncertainty of the model parameter estimates. Shear-wave velocity profiles and uncertainties are obtained from the velocity-depth marginal distribution of the trans-D posterior probability density. These probabilistic results for the Vs profile allow for probabilistic estimates of site response factors such as peak ground velocity/acceleration and amplification/resonance spectra, which ultimately quantify the confidence in the site characterization. This methodology is applied to passive seismic array recordings collected at five sites in Kitimat, British Columbia, Canada. Inversion results from several sites are compared to active-source (vibroseis) reflection data collected nearby.
Summary(Plain Language Summary, not published)
This study examines earthquake site response at a number of locations in Kitimat, British Columbia, where soft soils are expected to cause significant variations in earthquake shaking. We use a small seismic array recording ¿background noise¿ to determine shear wave velocity profiles ¿ a key parameter used in the National Building Code for seismic hazard calculations. We develop and apply an improved ¿inversion methodology¿ that provides not only shear wave velocity profiles, but also a greatly improved estimate of uncertainty in the derived profiles. This allows us to quantify the confidence in the earthquake site characterization. We also compare our inversion results to active-source (vibroseis) reflection data collected nearby.