Comparisons between non-interferometric and interferometric passive surface wave imaging methods—towards linear receiver array

Abstract

Summary Passive seismic methods in highly populated urban areas have gained much attention from the geophysics and civil engineering communities. Linear arrays are usually deployed for passive surface-wave investigations because of their high convenience, and passive surface wave imaging methods commonly used for linear arrays can be grouped as non-interferometric methods (e.g., passive multichannel analysis of surface wave, refraction microtremor) and interferometric methods (e.g., multichannel analysis of passive surface waves and spatial autocorrelation). It is well known that the seismic interferometry method is able to retrieve Green’s function between inter-station pairs based on passive seismic data and that is how interferometric methods work. Although non-interferometric methods are also popular and effective in near-surface seismic imaging, particularly in the geotechnical industry, there is no theoretical proof to clarify the accuracy and/or the bias of these methods. In this study, we employ numerical derivations and simulations to demonstrate the underlying physics for both non-interferometric and interferometric methods, under two common noise source environments including a homogeneous source distribution and a dominant in-line source distribution. We also prove the strength of interferometric methods for accurate dispersion imaging over the non-interferometric methods, and provide a way to estimate the biases in non-interferometric measurements. Finally, we present comprehensive comparisons between different passive surface wave methods with three typical field examples considering various observation systems.

High‐Resolution Near‐Surface Imaging at the Basin Scale Using Dark Fiber and Distributed Acoustic Sensing: Toward Site Effect Estimation in Urban Environments

“…According to Cheng et al. (2023b), passive‐source surface wave dispersion analysis methods can be broadly grouped into non‐interferometric methods (e.g., ReMi and PMASW) and interferometric methods (e.g., MAPS and SPAC). The former directly extract phase velocity dispersion spectra from ambient surface wave records (Louie, 2001; C. Park et al., 2004), while the latter calculate interferograms before dispersion estimation is attempted; in this case interferograms are either EGFs (e.g., Cheng et al., 2016) or SPAC coefficients (e.g., Asten, 2006; Chávez‐García et al., 2006).…”

“…We follow Cheng et al. (2023b) to briefly introduce the data processing involved in the three dispersion measurement techniques. For MAPS, we apply the phase‐weighted slant‐stacking algorithm (Cheng et al., 2021) on the retrieved EGFs enhanced by bin‐offset stacking, and measure the dispersion spectra E ( f , v ) using $$$E(f,v)=\omega (f,v)\left\vert \sum\limits _{j=1}^{N}\mathrm{exp}\left(i2\pi f{x}_{j}/v\right){C}_{j}(f)\right\vert $$$ where, ω ( f , v ) denotes the phase‐weight for the slant‐stacking enhancement, which is calculated by the instantaneous phase of the slant‐stacking complex spectra (Equation 5 in Cheng et al.…”

“…According to Cheng et al (2023b), passive-source surface wave dispersion analysis methods can be broadly grouped into non-interferometric methods (e.g., ReMi and PMASW) and interferometric methods (e.g., MAPS and SPAC). The former directly extract phase velocity dispersion spectra from ambient surface wave records (Louie, 2001;C.…”

“…We follow Cheng et al (2023b) to briefly introduce the data processing involved in the three dispersion measurement techniques. For MAPS, we apply the phase-weighted slant-stacking algorithm (Cheng et al, 2021) on the retrieved EGFs enhanced by bin-offset stacking, and measure the dispersion spectra E(f, v) using…”