Abstract
TThe noise cross-correlation function (NCF) derived from continuous seismic records approximates the Green's Function (GF) of the far-field surface wave, which is crucial for S-wave velocity tomography. A critical step is stacking short-term NCFs over extended periods to improve the signal-to-noise ratio (S/N) and enhance the imaging quality of seismic data. However, incoherent noise significantly affects the accuracy of NCF calculations. Moreover, strong directional sources in the wavefield pose a challenge to conventional linear or phase-weighted stacking (PWS), often leading to overestimates of surface-wave phase velocity due to unresolved azimuthal effects. Selecting stationary phase zone (SPZ) data and correcting surface wave propagation paths based on source orientation requires a 2D array setup, which imposes stringent layout constraints. To address these challenges, we propose a novel data selection workflow for a 1D linear array based on the probability density distribution of surface-wave phase velocity. Firstly, we divide the noise data into different frequency bands and calculate the probability density curve of the phase velocity distribution for each frequency band. Then, based on the differences in phase velocities between SPZ data, directional source data, and incoherent noise, we select the probability density segments of SPZ data in each frequency band for stacking, effectively reducing the impact of incoherent noise and strong directional sources. The typical synthetic data and distributed acoustic sensing (DAS) field data examples demonstrate that the proposed workflow significantly improves the S/N in NCF and increases the useful bandwidth of the surface- waves dispersion curve, which furnishes reliable dispersion curve and waveform data, thereby facilitating subsequent S-wave velocity tomography.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call