Joint inversion of first-break and early-arrival wave for near surface imaging Available to Purchase

Complex near-surface modeling is one of the most challenging problems in seismic exploration. First-arrival traveltime tomography exhibits low dependency on the initial model and adapts well to complex surfaces. However, due to the limitations of high-frequency approximations, the modeling accuracy and resolution are insufficient. In western China’s piedmont zone and loess plateau areas, there are complex near-surface issues, such as dramatic changes in the surface elevation, the development of low velocity zones, and high-velocity anomalies near the surface. Conventional methods based on first-arrival traveltime tomography struggle to accurately calculate the complex structures and high-velocity gravel boundaries in the shallow layers. The limitations in accuracy for near-surface modeling in complex exploration areas severely constrain the exploration of deep and ultra-deep petroleum resources. In this study, we propose a traveltime-waveform joint inversion method with total variation regularization, which combines first-arrival traveltime with early-arrival waveform data. This approach effectively avoids the high-frequency approximation assumption in traveltime tomography and the limitation of low-wavenumber information in full-waveform inversion. It also better captures the sparsity of the velocity model gradient, thereby preserving important underground structural boundaries and details during the inversion process. Additionally, we enhance the stability and accuracy of the inversion using memoryless BFGS to approximate the search direction of the Hessian matrix based on the hybrid conjugate gradient method. The effectiveness of our proposed approach is verified through inversion experiments on thrust structure models and real data.