Enhancing noise sources in stationary-phase zones for accurate phase-velocity estimation of high-frequency surface waves

Abstract

Passive surface-wave methods are appealing tools for their ability to noninvasively obtain shear-wave velocity with high accuracy and low costs. A linear array of stations is widely adopted in urban areas for its convenient deployments and efficient utilization of ambient noise energy from stationary-phase zones (SPZs), which contributes the most to surface-wave retrieval. However, noise sources in nonstationary-phase zones may threaten the reliability of velocity estimation when they possess strong energy. To obtain reliable phase-velocity estimation of high-frequency surface waves (>1 Hz), we enhance the contribution of noise sources in SPZs by using the multichannel-coherency-weighted stack (MCWS) method for the stacking of noise crosscorrelation functions and dispersion spectra. We model a nonuniform source distribution with strong offline sources. The seismic interferometry (SI) and passive multichannel analysis of surface waves (PMASW) methods overestimate velocities at lower frequencies (<5 Hz), whereas phase velocities of short wavelengths are relatively accurate. The true velocities of short wavelengths are then set as the velocity scanning range for MCWS, and the waves whose apparent velocities lie within the scanning range would be emphasized by MCWS. Noise sources in SPZs are enhanced because the apparent velocities of waves from other noise sources are higher and not in the scanning range. After using MCWS, the phase-velocity estimations of SI and PMASW are consistent with the theoretical dispersion curve. A field data example at a crossroads demonstrates that uneven source distributions might cause serious artifacts and the feasibility of our method is again confirmed. In addition, the enhancement of noise sources in SPZs is verified through noise source distribution imaging by a matched field processing technique.