Linear array analysis of passive surface waves combined with mini-Sosie technique

Abstract

SUMMARY Mini-Sosie, which is convenient and environmentally friendly with high energy output, is gaining increasing attention as a seismic exploration source. Passive surface wave survey using ambient noise have become a powerful tool for exploration in urban areas due to the advantages of being efficient and non-destructive. A 2-D or pseudo-1D array that can attenuate phase velocity overestimation due to directional noise are the optimal choice, but such arrays are limited by the complex environment of the experimental site on the one hand and the need to ensure that the medium beneath the array meets the assumption of lateral isotropy on the other. However, the fully 1-D linear array as an alternative lacks the ability to suppress the directional effect of the noise source. In this study, we develop a novel approach for linear array analysis of passive surface waves combined with mini-Sosie technique, called LAPSS. We use the mini-Sosie technique with fixed-frequency impact in the in-line direction to provide the theoretical phase velocity values at several frequency points as a reference for the biased dispersion image obtained from the linear array analysis, so as to estimate the azimuth of the noise source and to perform the correction of the biased dispersion image. We present the detailed workflow of LAPSS and compare the performance of LAPSS with PLAS (which has proven to be superior to the frequency–wavenumber method, spatial autocorrelation method, refraction microtremor and multichannel analysis of passive surface waves) in achieving unbiased dispersion image with different noise source distributions through synthetic tests. The results show that the accuracy of the two methods is comparable, but LAPSS greatly improves the computational efficiency, has stronger generalizability and depends on weaker assumptions. Finally, a field experiment is conducted to verify the feasibility and effectiveness of LAPSS in obtaining unbiased dispersion images using a fully linear array.