Wave Equation Dispersion Inversion of Distributed Acoustic Sensing Data

Abstract

AbstractDistributed acoustic sensing (DAS) is a new dense sampling seismic observation technique rapidly developed to image near‐surface velocity structures. Compared with traditional seismometer surveys, DAS can record the strain rate data with a high spatial and temporal resolution and low maintenance cost observation. However, the strain variation recorded by the DAS system is the horizontal component with a low signal‐to‐noise ratio because the coupling between the fiber and the earth is poor. In addition, conventional velocity inversion algorithms such as 1D dispersion curve inversion and 2D/3D full‐waveform inversion (FWI) are developed for particle velocity components. Hence, their application in DAS strain‐rate data requires conversion to particle velocity data, which would bring some numerical error. In this work, we propose the DAS wave‐equation dispersion inversion (DAS‐WD) method to invert the S‐wave velocity model directly from DAS strain rate data. The WD method does not require 1D dispersion curve inversion layered assumptions and overcomes the problem that the FWI is challenging to converge because the recorded horizontal DAS component data is noisy. Meanwhile, the strain rate DAS‐WD theory avoids the error caused by velocity and strain rate conversion by scaling of apparent velocity followed by time integrations. The typical synthetic and offshore DAS field data demonstrate that the proposed DAS‐WD method has great potential for a high‐resolution velocity tomogram.