Determination of surface-wave phase velocities by zeros of Aki’s spectrum of active-source records. Application to the dense array in Tongzhou, China

The spatial autocorrelation (SPAC) (also known as the Aki’s spectral method) of ambient seismic noise has been widely adopted in surface wave phase velocity extraction. In two-dimensional cases, the surface wave velocity can be calculated by fitting the SPAC coefficients with the zero-order Bessel function of the first kind or using the zeros of the Aki’s spectrum. This method has also been extended to active-source records. This study examined the application of the zeros of Aki’s spectra on active-source records using theoretical analysis and numerical simulation. We show that the zeros of the Aki’s spectrum should be associated with the zeros of the cosine function instead of the zeros of the zero-order Bessel function when extracting the phase velocity of the surface wave, considering the data acquisition and processing of the active-source records. The proposed method was then applied to the active-source data from methane explosion experiments collected using a dense array in Tongzhou, the subcenter of Beijing, for extracting the phase velocity of Rayleigh wave. The extracted dispersion curves were integrated with those obtained by beamforming the ambient noise to yield broadband dispersion curves at 0.3–6 Hz. This provides insightful results at high frequencies, at which less information can be obtained through the passive-source beamforming. The combing phase velocities from active-source with those obtained from ambient noise provide a better constrain on the shallow structure. Based on the combined fundamental mode dispersion curves at 28 excitation points, the S-wave velocity structure below the dense array is obtained by depth inversion. Due to the constraints imposed by the high-frequency information from active source, the estimated Vs30 are more reliable and can be used to the site classification.