3D Wave-Equation Dispersion Inversion of Distributed Acoustic Sensing Data to Reveal the Shallow Near-Surface Geology

Abstract

Distributed Acoustic Sensing (DAS) is an emerging seismic acquisition technology that utilizes optical fibers as sensing media. Compared to traditional geophones, DAS offers high spatial resolution, cost efficiency, suitability for large-scale deployment, and adaptability for long-term continuous monitoring, making it well-suited for 3D seismic exploration. However, DAS data collected in the field often exhibit a low signal-to-noise ratio (SNR) influenced by factors such as ground coupling, and DAS is sensitive only to axial strain along the fiber. Conventional wave-equation inversion techniques, including 2D/3D Full Waveform Inversion (FWI) and Wave Equation Dispersion (WD) inversion, rely on particle velocity data from geophones. Applying these methods to DAS data requires converting strain components to velocity, which can introduce numerical errors. This study presents a 3D wave-equation dispersion inversion method based directly on DAS strain data (3D-DAS-WD). The objective function minimizes the squared sum of wavenumber differences across frequencies in the fundamental mode dispersion curve of the DAS data. This approach avoids numerical errors associated with the strain-to-velocity conversion and addresses convergence issues in FWI with low-SNR DAS data. Synthetic model tests validate the method's stability and applicability in complex geological settings. At the same time, field data results indicate that 3D-DAS-WD enables high-resolution 3D subsurface imaging, effectively identifying features such as subsurface voids.