Multi-constraint two-step intelligent prestack seismic inversion and its application: A case study in the Junggar Basin, Northwest China

Abstract

Prestack seismic inversion serves as a bridge connecting seismic observations to subsurface rock properties, enabling quantitative estimation of elastic parameters, such as P-wave velocity ( V_p), S-wave velocity ( V_s), and density ( ρ), which provides direct evidence for reservoir prediction and fluid identification. However, simultaneously inverting V_p, V_s, and ρ from prestack data is a highly nonlinear and ill-posed problem. Traditional inversion methods often struggle to achieve an optimal balance among accuracy, stability, and physical consistency. In recent years, deep learning has offered new insights through its powerful nonlinear mapping capabilities. However, purely data-driven models rely heavily on large labeled datasets and often overlook physical laws, leading to inversion results that lack geological consistency. To overcome these limitations, this paper proposes a multi-constraint two-step intelligent prestack inversion method. First, near-angle seismic data were used to robustly invert P-wave impedance ( I_p). Then, employing TransUNet as the core network, a multi-constraint joint loss function was constructed to systematically integrate four types of prior information: (i) seismic data matching to ensure consistency with observed data. (ii) Physical relationship constraints linking I_p, V_p, and ρ. (iii) Empirical statistical relationships from well logs to regularize the ill-posed ρ inversion. (iv) Well data fitting to realize the matching of inversion results at well locations. This achieved high-precision inversion of V_p, V_s and ρ under the dual guidance of data driving and physical mechanisms. Tests on the Marmousi 2 model and actual shale reservoir data from the Junggar Basin demonstrated that the proposed method significantly improved inversion accuracy, stability, and noise resistance for all three parameters, particularly ρ, validating its potential for practical applications.

Keywords

Prestack seismic inversion / Multiple constraints / TransUNet / Two-step method

References

[1] Zong Z, Yin X, Wu G. Geofluid Discrimination Incorporating Poroelasticity and Seismic Reflection Inversion. Surv Geophys . 2015; 36(5): 659-681. doi: 10.1007/s10712—015—9330—6

[2] Tarantola A. Inverse Problem Theory and Methods for Model Parameter Estimation . Philadelphia, PA: Society for Industrial and Applied Mathematics; 2005. doi: 10.1137/1.9780898717921

[3] Buland A, Omre H. Bayesian linearized AVO inversion. Geophysics. 2003; 68(1): 185-198. doi: 10.1190/1.1543206

[4] Cooke DA, Schneider WA. Generalized linear inversion of reflection seismic data. Geophysics . 1983; 48(6): 665-676. doi: 10.1190/1.1441497

[5] Downton JE. Seismic parameter estimation from AVO inversion . Dissertation. Calgary, AB: University of Calgary; 2005.

[6] Guo Q, Zhang HB, Han FL, et al. Prestack Seismic Inversion Based on Anisotropic Markov Random Field. IEEE Trans Geosci Remote Sens. 2018; 56(2): 1069-1079. doi: 10.1109/TGRS.2017.2758800

[7] Aki K, Richards PG. Quantitative Seismology: Theory and Methods . San Francisco, CA: W.H. Freeman; 1980.

[8] Shuey RT. A simplification of the Zoeppritz equations. Geophysics. 1985; 50(4): 609-614. doi: 10.1190/1.1441936

[9] Zhi LX, Chen SQ, Li XY. Amplitude variation with angle inversion using the exact Zoeppritz equations — Theory and methodology. Geophysics. 2016; 81(2): N1-N15. doi: 10.1190/geo2014—0582.1

[10] Alemie W, Sacchi MD. High—resolution three—term AVO inversion by means of a Trivariate Cauchy probability distribution. Geophysics. 2011; 76(3): 420-424. doi: 10.1190/1.3554627

[11] Theune U, Jensås IØ, Eidsvik J. Analysis of prior models for a blocky inversion of seismic AVA data. Geophysics . 2010; 75(3): C25-C35. doi: 10.1190/1.3427538

[12] Yuan M, Miao Z, Pan L, et al. Generalized Gaussian distribution—based prestack inversion: A case study of channel sands in the Lianggaoshan Formation, Sichuan Basin. J Seism Explor. 2026; 35(1): 200-220. doi: 10.36922/JSE025450105

[13] Haas A, Dubrule O. Geostatistical inversion—a sequential method of stochastic reservoir modelling constrained by seismic data. First Break. 1994; 12(11): 561-569.

[14] Azevedo L, Soares A. Geostatistical Methods for Reservoir Geophysics . Berlin, Germany: Springer; 2017.

[15] Sen MK, Stoffa PL. Bayesian inference, Gibbs’ sampler and uncertainty estimation in geophysical inversion. Geophys Prospect. 1996; 44(2): 313-350.

[16] Francis A. Understanding stochastic inversion: part 1. First Break . 2006; 24(11): 69-77. doi: 10.3997/1365—2397.2006026

[17] Srivastava RP, Sen MK. Stochastic inversion of prestack seismic data using fractal—based initial models. Geophysics . 2010; 75(3): R47-R59. doi: 10.1190/1.3379322

[18] Hamid H, Pidlisecky A. Multitrace impedance inversion with lateral constraints. Geophysics . 2015; 80(6): M101-M111. doi: 10.1190/geo2014—0546.1

[19] Li C, Liu G, Deng Y. Nonstationary phase—corrected full—waveform inversion with attenuation compensation in viscoacoustic medium. J Geophys Eng . 2022; 19(4): 724-738. doi: 10.1093/jge/gxac046

[20] Zhang G, Wang Z, Chen Y. Deep learning for seismic lithology prediction. Geophys J Int. 2018; 215(2): 1368-1387. doi: 10.1093/gji/ggy344

[21] Yang L, Fomel S, Wang S, et al. Denoising of distributed acoustic sensing data using supervised deep learning. Geophysics. 2023; 88(1): WA91-WA104. doi: 10.1190/geo2022—0138.1

[22] Gao Y, Zhao D, Li T, et al. Deep learning vertical resolution enhancement considering features of seismic data. IEEE Trans Geosci Remote Sens . 2023; 61: 1-13. doi: 10.1109/TGRS.2023.3234617

[23] Das V, Pollack A, Wollner U, et al. Convolutional neural network for seismic impedance inversion. Geophysics . 2019; 84(6): R869-R880. doi: 10.1190/geo2018—0838.1

[24] Chen Y, Zhang G, Bai M, et al. Automatic waveform classification and arrival picking based on convolutional neural network. Earth Space Sci. 2019; 6(7): 1244-1261. doi: 10.1029/2018EA000466

[25] Wu B, Meng D, Wang L, et al. Seismic impedance inversion using fully convolutional residual network and transfer learning. IEEE Geosci Remote Sens Lett . 2020; 17(12): 2140-2144. doi: 10.1109/LGRS.2019.2963106

[26] Alfarraj M, AlRegib G. Semisupervised sequence modeling for elastic impedance inversion. Interpretation. 2019; 7(3): SE237-SE249. doi: 10.1190/INT—2018—0250.1

[27] Liu X, Wu B, Yang H. Multitask full attention U—Net for prestack seismic inversion. IEEE Geosci Remote Sens Lett . 2023; 20: 1-5. doi: 10.1109/LGRS.2023.3303698

[28] Vaswani A, Shazeer N, Parmar N, et al. Attention is all you need. In: Advances in Neural Information Processing Systems. Vol 30. 2017: 5998-6008.

[29] Aleardi M, Salusti A. Elastic prestack inversion through Discrete Cosine Transform reparameterization and Convolutional Neural Networks. Geophysics. 2021; 86(1): R129-R146. doi: 10.1190/geo2020—0313.1

[30] Chen T, Zou B, Wang Y, et al. Prestack Seismic Inversion Driven by Priori Information Neural Network and Statistical Characteristic. IEEE Trans Geosci Remote Sens. 2024; 62: 1-14. doi: 10.1109/TGRS.2024.3394523

[31] Biswas R, Sen MK, Das V, et al. Prestack and poststack inversion using a physics—guided convolutional neural network. Interpretation. 2019; 7(3): SE161-SE174. doi: 10.1190/INT—2018—0236.1

[32] Wang S, Liu C, Song C, et al. Prestack AVO Inversion Based On Physics—constrained Deep Learning Method. Appl Geophys. 2025: 1-19. doi: 10.1007/s11770—025—1179—y

[33] Karpatne A, Atluri G, Faghmous JH, et al. Theory—Guided Data Science: A New Paradigm for Scientific Discovery from Data. IEEE Trans Knowl Data Eng . 2017; 29(10): 2318-2331. doi: 10.1109/TKDE.2017.2720168

[34] Raissi M, Perdikaris P, Karniadakis GE. Physics—informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations. J Comput Phys. 2019; 378: 686-707. doi: 10.1016/j.jcp.2018.10.045

[35] Liu Z, Zhang J, Chen Y, et al. From Physics Constraints to Trustworthy Bayesian Reasoning: Synergies of PINN, iPINN, and iBPINN in Prestack AVO Inversion. IEEE Trans Geosci Remote Sens. 2025; 63: 1-14. doi: 10.1109/TGRS.2025.3636416

[36] Zhang J, Zhao X, Chen Y, et al. Domain knowledge—guided data—driven prestack seismic inversion using deep learning. Geophysics. 2023; 88(2): M31-M47. doi: 10.1190/geo2021—0560.1

[37] Zhao L, Zhu J, Qin X, et al. Joint geochemistry—rock physics modeling: Quantifying the effects of thermal maturity on the elastic and anisotropic properties of organic shale. Earth—Sci Rev. 2023; 247: 104627. doi: 10.1016/j.earscirev.2023.104627