Improving the Resolution of Array-Based Surface-Wave Phase Velocity Tomography by Correcting for Array-Averaging Effects

Abstract In recent years, the development of array-based surface-wave technology has brought ambient noise surface-wave imaging into the era of overtone surface waves. Overtone surface-wave dispersion curves offer additional constraints in inversion, significantly enhancing depth resolution, especially for low-velocity zones. However, these methods require noise cross-correlation functions from dozens of stations for each dispersion extraction. Previous imaging work used dispersion curves extracted from the subarray to represent the average structure beneath the subarray, obtaining a 1D velocity structure at the center point of different subarrays. This approach can lead to observed phase velocity values and distributions being influenced by the subarray’s size due to the averaging effect. In this work, we try to establish the relationship between the phase velocity distribution and the observed dispersion curves using simple approximations. Then, the averaging issue can be effectively addressed through a spatial phase-velocity further inversion, for which the phase velocity distribution is treated as the inverted model and the phase velocities obtained by subarrays as the observed data. Synthetic tests demonstrate that this inversion effectively enhances the horizontal spatial resolution of the array surface-wave method and corrects the observed absolute velocity values. We also apply this method to the eastern South China block with our previously developed surface-wave method, the frequency–Bessel transform method, supplementing more essential details for the midcrustal low-velocity zones. Furthermore, the inversion enables a chessboard test to estimate the horizontal resolution of the inverted phase velocity distribution, addressing a key concern in array-based surface-wave imaging.