Attenuation Estimation With Uncertainty for Seismic Noise Interferometry: Application to a Dense 3C Array in Groningen, Netherlands

Date: 4/24/2019

Time: 05:15 PM

Room: Pine

Seismic wave attenuation information is important for shallow subsurface structure characterization because it is related to fracture, temperature, composition and fluid content. It is also important for efforts to predict strong ground motion We use a method based on noise cross-correlations of linear triplet of stations (Liu et al., 2015) to compute the relative amplitude decay between different paths with respect to frequency. The focusing/defocusing effect for the ambient noise should be common for the co-linear triplet of stations under the assumption of a remote noise source. We further assume constant Q over each narrow frequency band considered. Combined with uncertainty quantification for the stacked noise cross-correlation (Liu and Beroza2019), we estimate Q from these relative amplitude measurements and propagate the amplitude uncertainty to the uncertainty of Q value through linear least-square estimation. We apply this method to the Loppersum dense 3C array in Groningen, Netherlands. The Loppersum array features 415 stations at ~250 m station spacing. We treat each station as a virtual source, and construct amplitude maps with uncertainties for both the fundamental and first higher Rayleigh wave modes. We remove errant amplitude measurements using a combination of statistical stability over time and spatial coherency. Stacking the results from different virtual sources yields an averaged attenuation map for each narrow frequency band. We convert the attenuation maps at different frequencies to attenuation with depth and construct a 3D attenuation model for the shallow crust from these results.

Presenting Author: Xin Liu

Authors