Spatial Variation and Frequency Dependence of Lg Wave Attenuation With Site Response Correction Along the CCSE Array in Central California, US

Abstract

We estimate lateral Lg wave attenuation (Q) structure at four center frequencies (0.75, 1, 2 and 2.75 Hz) along the Central California Seismic Experiment array in western US crossing the San Andreas Fault and Central Valley. We take two steps in constructing the site-response-corrected Lg Q model: (a) we compute relative site responses at each station using the reverse two-station method, and (b) we estimate Q values based on the two-station method after removing the site term. Removal of the site response in the Q model allows to probe laterally varying Q properties at mid-to-lower crustal depths. Our model follows a power-law frequency dependence as Q(f) = (81 ± 8)f^(0.62±0.11), reflecting the active tectonic setting and the presence of fluids in the region. A change in lithology from softer sediments near Pacific coast to harder basements near Sierra Nevada correlates well with the increasing trend of the Lg Q values towards east. Our laterally varying estimates at lower frequencies generally follow the variation of shear-wave velocities at deeper crustal depth and Moho temperature, whereas those at higher frequencies mostly follow the shear-wave velocity variation at shallow depth. Positive site responses obtained by reverse two-station method are found at 34 stations out of total 46 stations examined, and their responses are mostly correlated with surficial lithology (i.e., sedimentary rocks) along the profile, rather than the thickness of the sediments. The site responses also exhibit a strong negative correlation to the V_(S30) data.