Body-wave retrieval and imaging from ambient seismic fields with very dense arrays

Abstract:

Correlation-based analyses of ambient seismic wavefields is a powerful tool for retrieving subsurface information such as stiffness, anisotropy, and heterogeneity at a variety of scales. These analyses can be considered to be data-driven wavefield modeling. Studies of ambient-field tomography have been mostly focused on the surface waves, especially fundamental-mode Rayleigh waves. Although the surface-wave tomography is useful to model 3D velocities, the spatial resolution is limited due to the extended depth sensitivity of the surface wave measurements. Moreover, to represent elastic media, we need at least two stiffness parameters (e.g., shear and bulk moduli). We develop a technique to retrieve P diving waves from the ambient field observed by the dense geophone network (~2500 receivers with 100-m spacing) at Long Beach, California. With two-step filtering, we improve the signal-to-noise ratio of body waves to extract P wave observations that we use for tomography to estimate 3D P-wave velocity structure. The small scale-length heterogeneity of the velocity model follows a power law with ellipsoidal anisotropy. We also discuss possibilities to retrieve reflected waves from the ambient field and show other applications of the body-wave extraction at different locations and scales. Note that reflected waves penetrate deeper than diving waves and have the potential to provide much higher spatial resolution.