Postprint version. Published in 2014 Geo-Congress Proceedings: Atlanta, Georgia, February 23, 2014, pages 773-782.
The definitive version is available at https://doi.org/10.1061/9780784413272.075.
A linear array consisting of 24 1 Hz geophones at 10 m spacing was used to measure the passive surface waves at the UTexas1 site. This paper describes analysis of this dataset using three different methods: ESPAC (extended spatial autocorrelation), f-k (frequency-wavenumber), and ReMi (refraction microtremor). Dispersion curves were developed using each method and the median trends as well as the uncertainty about the medians are compared. The dispersion curves were then individually inverted to estimate shear wave velocity profiles. The inversion results are compared, and a bounded best-estimate shear wave velocity profile is presented. For the site conditions, specific recording equipment, and array geometry the ESPAC method was the most consistent between recordings and able to resolve the lowest frequency Rayleigh waves. The shear wave velocity of all three methods were in close agreement in the upper 20 m above a stiff layer, but were increasing disparate as depth increased. All three methods resolved a velocity inversion, a stiff layer, at roughly 20 m to 55 m depth overlying softer material. As is expected with surface wave methods, and particularly with passive methods that are measuring ambient noise along single linear array, the uncertainty in the dispersion curves increased with decreasing frequency, and the uncertainty in the shear wave velocity profiles increased with increasing depth.
Civil and Environmental Engineering