Estimation of VS30 at Transportable Array Stations by Inversion of Low-Frequency Seismic Noise

Abstract:

The main source of seismic noise below 0.05 Hz is usually atmospheric pressure variations, especially when pressure variations are large. The amount of ground deformation under surface pressure changes reflects the characteristics of shallow elastic structure. If a surface station is co-located with seismic and pressure sensors, shallow rigidity can be estimated by adopting Sorrells’ half-space model, under a propagating pressure waves assumption. We estimated half-space elastic structure at 784 Transportable Array (TA) stations in a previous study. Furthermore, we developed an inversion approach to estimate layered structure using seismic and pressure data at different frequencies below 0.05 Hz. In the inversion scheme, we utilize surface observables η(f)=S_Z/S_P, where f is frequency and S_Z and S_P are the power spectral densities of vertical seismic data and of surface pressure data. A vertically heterogeneous medium is assumed beneath a station where density, P wave velocity, and S wave velocity change with depth. Using numerical differentiation, we derive depth sensitivity kernels for η(f) with which we invert η(f) for shallow structure, just like we invert phase velocity of surface waves for the Earth structure. Depth sensitivity is typically controlled by elasticity in the uppermost 50-100 meters. We estimated layered structure at 9 co-located stations in the Piñon Flat Observatory. Our estimates of V_S30 agree with both an independent estimate of V_S30 on the site (Yong et al., 2016) and are consistent with knowledge of local geology. Although estimated layered structures are much smoother compared to the velocity profile obtained by the active-source survey, our method is still sufficient for estimating V_S30 because V_S30 is an averaged quantity for the uppermost 30 meters. We will present V_S30 at many TA stations estimated from the inversion approach.