Enhancing Dispersion Imaging with the Frequency-Bessel Method Using the Kolmogorov-Smirnov Subarry Partition Test: A Case Study of SGB Dense Array

Recent studies on the frequency-Bessel transform (F-J) method demonstrate the ability of the array-based method to extract higher-mode surface wave dispersion curves from ambient noise. Presently, the subarray selection of F-J method for three-dimensional imaging predominantly relies on regular partition shapes. However, a crucial tradeoff arises concerning partition size: a larger size improves dispersion curve quality at lower frequencies but deviates from the assumption of a flattened velocity structure required by F-J method. Conversely, a smaller size ensures compliance with the flattened structure assumption but limits usability to higher frequencies. This tradeoff becomes particularly challenging in scenarios with relatively shallow depth scales, such as fault imaging, where subsurface velocity structures exhibit significant variations. To address this dilemma, we propose a novel strategy employing dispersion curves from small partitions as probes. We use the Kolmogorov-Smirnov (K-S) test to assess the flattened velocity structure within a larger partition by utilizing these probes. This process streamlines the elimination of arrays within the larger partition that diverge from the assumption of a flattened structure. Consequently, when utilizing the larger partition, we can maximize the compliance of the flattened structure assumption.        In this study, we leverage noise cross-correlation functions from a dense array at the Sage Brush Flat (SGB) site, located in the intricate trifurcation area south of Anza in Southern California. The array comprises geophones arranged in a 600m×600m configuration. We hope that the implementation of this partition strategy will enhance the feasibility of array-based surface wave tomography methods, especially in dense arrays within regions characterized by complex shallow velocity structures.