Abstract:Shallow-seismic full-waveform inversion (FWI) is becoming increasingly popular for the reconstruction of the shallow-subsurface model in near-surface geophysics. Because Rayleigh waves dominate the vertical component of the shallow-seismic recording, FWI mainly fits Rayleigh waves in the observed data, while the utilization and fitting of P waves in the data are usually overlooked. The appropriate use of the body-wave signal in shallow-seismic FWI remains a problem. We propose herein a wavefield-separated full-waveform inversion (WSFWI) method to make better use of the P-wave signal in the shallow-seismic data. The WSFWI method mainly contains two steps: (1) separating the P wave from the observed data and applying an acoustic FWI to it for the reconstruction of the P-wave velocity model, and (2) fixing the P-wave velocity model and applying an elastic FWI to the entire recording for the reconstruction the S-wave velocity model. We show that Rayleigh and P waves are sensitive towards different areas of the model, and therefore, WSFWI can better utilize the sensitivity kernel of P waves to improve the accuracy of the reconstructed P-wave velocity model. A synthetic example shows that WSFWI outperforms the FWI that uses the whole recording simultaneously, especially in the reconstruction of the P-wave velocity model. It also proves that WSFWI is able to avoid (mitigate) crosstalk imposed from the S-wave velocity to the P-wave velocity models. We apply both the conventional FWI and WSFWI to a field data set acquired in Olathe, Kansas, USA. The field example shows that the S-wave velocity model can be reconstructed with high robustness in both the FWI and WSFWI results. Both the conventional FWI and WSFWI nicely fitted the observed Rayleigh wave, while WSFWI also fitted the P waves in the observed data and reconstructed the P-wave velocity model with relatively higher reliability.

GPU Accelerated 3D P-wave Source Free Adaptive Wavefield Reconstruction Inversion with an application to experimental VSP physical modeling data

Acceleration of full waveform inversion based on random super-Shot and GPU

Accelerating 3-D Acoustic Full Waveform Inversion Using a Multi-GPU Cluster

Multi-GPU based two-level acceleration of full waveform inversion

An Accelerated Proximal Gradient Algorithm for Source-Independent Waveform Inversion

Inexact Augmented Lagrangian Method-Based Full-waveform Inversion with Randomized Singular Value Decomposition

CUDA‐based Acceleration of Full Waveform Inversion on GPU

Physics-Consistent Data-driven Waveform Inversion with Adaptive Data Augmentation

Full-waveform Inversion with Randomized Space Shift

Robust Wavefield Inversion via Phase Retrieval

An Approach of Full Waveform Inversion on GPU Clusters and Its Application on Land Datasets

Wavefield-Separated Full-Waveform Inversion of Shallow-Seismic Rayleigh Waves

Algorithmic analysis towards time-domain extended source waveform inversion

Waveform inversion with structural regularizing constraint based on gradient decomposition

Inexact Augmented Lagrangian Method-Based Full-waveform Inversion with Randomised Singular Value Decomposition

A graphics processing unit implementation of time-domain full-waveform inversion

Full waveform inversion with nonlocal similarity and model-derivative domain adaptive sparsity-promoting regularization

Elastic Full Waveform Inversion with Angle Decomposition and Wavefield Decoupling.

Time-domain wavefield reconstruction inversion

Velocity Model Building by Waveform Inversion of Early Arrivals and Reflections: A 2D Case Study with Gas-Cloud Effects