High-Precision Sub-Seafloor Velocity Building Based on Joint Tomography and Deep Learning on OBS Data in the South China Sea

Abstract

This study combines data-driven DL with physics-driven tomography inversion to construct a more accurate subseafloor P-wave velocity model and applies it to OBS data from the SCS. The experimental results show that adding SSIM to the U-net can enhance the network’s ability to capture data details and improve learning efficiency. On the real data, this method shows good effectiveness and reliability in identifying stratigraphic interfaces and complex geological structures. Furthermore, applying the cosine transform for data preprocessing extracts key features, further improving neural network efficiency and effectiveness, and offering a feasible solution to the issue of limited sample size. Although some progress has been made, the accuracy improvement is still limited, and future research will incorporate physical constraints. By constructing an objective function containing physical constraints, the subseafloor velocity inversion results can follow the laws of geophysics and improve the accuracy and interpretability of the model.

References

[1] BianA F, ZouZ H, ZhouH W, et al.. Evaluation of Multi-Scale Full Waveform Inversion with Marine Vertical Cable Data. Journal of Earth Science, 2015, 26(4): 481-486

[2] ChenG X, ChenJ X, JensenK, et al.. Joint Data and Model-Driven Simultaneous Inversion of Velocity and Density. Geophysical Journal International, 2024, 237(3): 1674-1698

[3] ChenG X, WuX, LiJ, et al.. Initial Model Building Method Based on Iterative Deep Learning in Sparse Transform Domain. Annual Meeting of Chinese Geoscience Union (CGU), Xiamen, 2024(in Chinese)

[4] ChenG XAccurate Background Velocity Model Building Method Based on Iterative Deep Learning in Sparse Transform Domain, 2024

[5] ChengX Q, LiuQ H, LiP P, et al.. Inverting Rayleigh Surface Wave Velocities for Crustal Thickness in Eastern Tibet and the Western Yangtze Craton Based on Deep Learning Neural Networks. Nonlinear Processes in Geophysics, 2019, 26(2): 61-71

[6] DouJ, XiangZ L, XuQ, et al.. Application and Development Trend of Machine Learning in Landslide Intelligent Disaster Prevention and Mitigation. Earth Science, 2023, 48(5): 1657-1674(in Chinese with English Abstract)

[7] HanM, ZouZ, MaR. Deep Learning-Driven Velocity Modeling Based on Seismic Reflection Data and Multi-Scale Training Sets. Oil Geophysical Prospecting, 2021, 56(5): 935-946(in Chinese with English Abstract)

[8] HuJ J, DingYT, ZhangH, et al.. A Real-Time Seismic Intensity Prediction Model Based on Long Short-Term Memory Neural Network. Earth Science, 2023, 48(5): 1853-1864(in Chinese with English Abstract)

[9] HuangF M, ChenB, MaoD X, et al.. Landslide Susceptibility Prediction Modeling and Interpretability Based on Self-Screening Deep Learning Model. Earth Science, 2023, 48(5): 1696-1710(in Chinese with English Abstract)

[10] LiC F, SongT R. Magnetic Recording of the Cenozoic Oceanic Crustal Accretion and Evolution of the South China Sea Basin. Chinese Science Bulletin, 2012, 57(24): 3165-3181

[11] LiuY T, LiC F, WenY L, et al.. Mantle Serpentinization beneath a Failed Rift and Post-Spreading Magmatism in the Northeastern South China Sea Margin. Geophysical Journal International, 2021, 225(2): 811-828

[12] SambolianS, GorszczykA, OpertoS, et al.. Mitigating the Ill-Posedness of First-Arrival Traveltime Tomography Using Slopes: Application to the Eastern Nankai Trough (Japan) OBS Data Set. Geophysical Journal International, 2021, 227(2): 898-921

[13] WamriewD, ChararaM, PissarenkoD. Joint Event Location and Velocity Model Update in Real-Time for Downhole Microseismic Monitoring: A Deep Learning Approach. Computers & Geosciences, 2022, 158: 104965

[14] XieY H, YeY F, HuangX G, et al.. Advancements and New Frontiers in Offshore Seismic Exploration Technology. Journal of Earth Science, 2024, 35(5): 1749-1757

[15] YangF S, MaJ W. Deep-Learning Inversion: a Next-Generation Seismic Velocity Model Building Method. Geophysics, 2019, 84(4): R583-R599

[16] YangH C, LiP, MaF, et al.. Building Near-Surface Velocity Models by Integrating the First-Arrival Traveltime Tomography and Supervised Deep Learning. Geophysical Journal International, 2023, 235(1): 326-341

[17] ZhangJ Z, ZhaoM H, DingW W, et al.. New Insights into the Rift-to-Drift Process of the Northern South China Sea Margin Constrained by a Three-Dimensional Wide-Angle Seismic Velocity Model. Journal of Geophysical Research: Solid Earth, 2023, 128(4): e2022JB026171

RIGHTS & PERMISSIONS

China University of Geosciences (Wuhan) and Springer-Verlag GmbH Germany, Part of Springer Nature