A Stockwell-Bessel Transform Based Method for Extracting Broadband Dispersion Curve From Seismic Ambient Noise Data

Abstract

Ambient noise tomography has been widely used to image subsurface velocity structures, with a critical step being the extraction of surface-wave dispersion curves from noise cross-correlation functions (NCFs). However, obtaining reliable broadband dispersion data, especially at low frequencies, remains challenging. In this study, we introduce a novel method for extracting surface-wave dispersion curves from NCFs of each station pair based on the Stockwell-Bessel (S-J) transform, which requires no parameter settings. Compared with conventional methods, the new method overcomes limitations imposed by the 2–3 wavelength station distance. Synthetic tests demonstrate that the new method can accurately extract Rayleigh wave dispersion curves, with a relative error of less than 1$\%$, even when the inter-station distance is as small as one wavelength. Field applications in Madagascar further confirm that the S-J transform produces more stable and broader-band dispersion curves than conventional methods, facilitating higher-resolution imaging of subsurface velocity structures. Moreover, this method allows for the extraction of multi-modal dispersion curves through the superposition of S-J spectra from multiple station pairs, offering a practical approach for deriving regionally averaged velocity structure.