An Evaluation of ${{\varvec{V}}}_{{{\varvec{SZ}}}}$ Estimates from the P-wave Seismogram Method for Sites in California

Abstract

The P-wave seismogram method estimates the average shear wave velocity over a representative depth ($V_{SZ}$) from earthquake recordings at a site. The $V_{SZ}$ is computed from the amplitudes of the radial and vertical P-wave arrivals on the earthquake recordings and an estimate of the seismological ray parameter ($p$). The ray parameter is estimated from the depth of the event, the epicentral distance, and the regional crustal velocity model. We evaluated the P-wave seismogram approach to estimating $V_{SZ}$ at 153 seismic recording stations in California for which shear wave velocity profiles are available and tested the effect of different crustal models on the estimated ray parameter and the resulting $V_{SZ}$. Across all the sites, the estimated $V_{SZ}$ values were, on average, about 24% larger than the measured $V_{SZ}$, although the difference was negligible for softer sites and as large as 45% at stiffer sites (>1000 m/s). Two crustal velocity models for California were considered: a simplified four-layer crustal model for the entire state and a set of more detailed crustal models used for different parts of the state. The effect of the assumed crustal velocity profile was not significant for earthquake events with focal depths greater than about 3–5 km, but for shallower events the detailed crustal velocity model produced $V_{SZ}$ values significantly smaller than the $V_{SZ}$ from the simplified crustal velocity model due to the effect of the shallow low velocity layers and large gradient on the ray parameter.