FaciesGAN: A conditional GAN framework for realistic facies scenario generation as an efficient alternative to multiple-point statistics

Abstract

Facies are rock bodies that reflect specific depositional environments and play a central role in reservoir characterization. Accurate facies modeling is a key challenge in generating realistic geological scenarios that honor sparse well data while capturing geological uncertainty. This study introduces FaciesGAN, a novel deep learning framework based on conditional generative adversarial networks (cGANs). The method employs a hierarchical structure of generators and discriminators that progressively refine coarse estimates into high-resolution facies models, ensuring consistency with well data and depositional patterns at each stage. FaciesGAN was validated using the limited Stanford Earth Science Data dataset, demonstrating strong performance even under data scarcity. The quantitative evaluation employed multidimensional scaling and yielded an intersection over union index of 99.96% relative to the conditioning well data. These results confirmed the model’s ability to generate diverse scenarios with high fidelity while preserving statistical distributions. Compared with a traditional multiple-point statistics implementation, FaciesGAN produced more realistic and varied geological realizations with significantly greater computational efficiency. These results indicate that cGAN-based approaches, such as FaciesGAN, represent a promising direction for subsurface modeling, offering robust tools for data augmentation, improved uncertainty assessment, and enhanced reservoir characterization.

Keywords

Conditional generative adversarial network / Facies / Hard data / Geostatistical simulations / Seismic inversion

Funding

This research was funded by the project “4D Integrated Quantitative Characterization of Reservoirs” (Contract No. FEESC/PETROBRAS/4600671602) and the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior-Brasil (CAPES)-Finance Code 001.

References

[1] Sheriff RE, Geldart LP. Exploration Seismology. Cambridge: Cambridge University Press; 1995. Available from: https://books.google.com.br/books?hl=es&lr=&id=wRYgAwAAQBAJ&oi=fnd&pg=PA1901&dq=Sheriff+RE,+Geldart+LP.+Exploration+Seismology.+Cambridge:+Cambridge+University+Press%3B+1995&ots=wCZUKyTUNq&sig=ELLBEg-lVBcYuxFxazP0aboszHY&redir_esc=y#v=onepage&q=Sheriff%20RE%2C%20Geldart%20LP.%20Exploration%20Seismology.%20Cambridge%3A%20Cambridge%20University%20Press%3B%201995&f=false Last accessed on 025 Nov 06].

[2] Grana D, Mukerji T, Doyen P. Seismic Reservoir Modeling: Theory, Examples, and Algorithms. United States: John Wiley and Sons; 2021.

[3] Sangree JB, Widmier JM. Interpretation of depositional facies from seismic data. Geophysics. 1979; 44(2):131-160. doi:10.1190/1.1440957

[4] Lee J, Mukerji T. The Stanford VI-E Reservoir: A Synthetic Data Set for Joint Seismic-EM Time-Lapse Monitoring Algorithms. 25th Annual Report, Stanford Center for Reservoir Forecasting, Stanford, CA: Stanford University; 2012.

[5] Selley RC. Elements of Petroleum Geology. United States: Gulf Professional Publishing; 1998.

[6] Veeken PCH. Seismic Stratigraphy, Basin Analysis and Reservoir Characterisation. Vol. 37. Netherlands: Elsevier; 2006.

[7] Grana D, Mukerji T, Dvorkin J, Mavko G. Stochastic inversion of facies from seismic data based on sequential simulations and probability perturbation method. Geophysics. 2012; 77(4):M53-M72. doi:10.1190/geo2011-0417.1

[8] Kemper M, Gunning J. Joint impedance and facies inversion-seismic inversion redefined. First Break. 2014; 32(9):89-95. doi:10.3997/1365-2397.32.9.77968

[9] Pyrcz MJ, Deutsch CV. Geostatistical Reservoir Modeling. Oxford: Oxford University Press; 2014.

[10] Slatt RM. Geologic controls on reservoir quality. Dev Petrol Sci. 2013; 61:229-281. doi:10.1016/B978-0-444-56365-1.00006-7

[11] Petrovic A, Aigner T, Pontiggia M. Facies heterogeneities in a ramp carbonate reservoir analogue: A new high-resolution approach for 3D facies modelling. J Petrol Geol. 2018; 41(2):155-174. doi:10.1111/jpg.12698

[12] Slatt RM. Basic principles and applications of reservoir characterization. Dev Petrol Sci. 2013; 61:1-38. doi:10.1016/B978-0-444-56365-1.00001-8

[13] Strebelle S. Conditional simulation of complex geological structures using multiple-point statistics. Math Geol. 2002; 34:1-21. doi:10.1023/A:1014009426274

[14] Roksandić MM. Seismic facies analysis concepts. Geophys Prospect. 1978; 26(2):383-398. doi:10.1111/j.1365-2478.1978.tb01600.x

[15] Dembicki H Jr. Practical Petroleum Geochemistry for Exploration and Production. Netherlands: Elsevier; 2017. doi:10.1016/C2014-0-03244-3

[16] Journel AG. Fundamentals of Geostatistics in Five Lessons. Vol. 8. Washington, DC: American Geophysical Union; 1989.

[17] Al-Mudhafar WJ. Geostatistical Simulation of Facies and Petrophysical Properties for Hetero-Geneity Modeling in a Tidal Depositional Environment: A Case Study from Upper Shale Member in A Southern Iraqi oil Field. In: Unconventional Resources Technology Conference, 26-28 July 2021; 2021. p. 2627-2642. doi:10.15530/urtec-2021-5551

[18] Journel AG. Nonparametric estimation of spatial distributions. J Int Assoc Math Geol. 1983; 15:445-468. doi:10.1007/BF01031292

[19] Strebelle SB, Journel AG. Reservoir Modeling using Multiple- Point Statistics. In: SPE Annual Technical Conference and Exhibition?; 2001. p. SPE-71324. doi:10.2118/71324-MS

[20] Mariethoz G, Caers J. Multiple-Point Geostatistics: Stochastic Modeling with Training Images. Wiley-Blackwell; 2014. Available from: https://books.google.com.br/books? hl=es&lr=&id=Fe7sBQAAQBAJ&oi=fnd&pg=PA85&dq =Multiple%E2%80%90Point+Geostatistics:+Stochastic+ Modeling+with+Training+Images&ots=P9LR5qUq 9B&sig=98MN9Mk5KPtkMgfniK_ECGVnUSQ& redir_esc=y#v=onepage&q=Multiple%E2%80%90 Poi nt%20Ge ost at i s tics%3A%20Stochas tic%20 Modeling%20with%20Training%20Images&f=false

[21] [ 2025 Nov 06]

[22] Harding A, Strebelle S, Levy M, et al. Reservoir facies modelling: New advances in MPS. In: Geostatistics Banff 2004. Berlin: Springer; 2004. p. 559-568. doi:10.1007/978-1-4020-3610-1_57

[23] Zhang TF, Tilke P, Dupont E, Zhu LC, Liang L, Bailey W. Generating geologically realistic 3D reservoir facies models using deep learning of sedimentary architecture with generative adversarial networks. Pet Sci. 2019; 16(3):541-549. doi:10.2523/IPTC-19454-MS

[24] Feng R, Grana D, Mosegaard K. Geostatistical facies simulation based on training image using generative networks and gradual deformation. Math Geosci. 2025; 57:1021-44. doi:10.1007/s11004-024-10169-y

[25] Tian L, Wang Z, Liu W, Cheng Y, Alsaadi FE, Liu X. A new GAN-based approach to data augmentation and image segmentation for crack detection in thermal imaging tests. Cognit Comput. 2021; 13(5):1263-1273. doi:10.1007/s12559-021-09922-w

[26] Goodfellow IJ, Pouget-Abadie J, Mirza M, et al. Generative adversarial nets. Adversarial Neural Information Processing Systems; 2014. p. 27.

[27] Bowles C, Chen L, Guerrero R, et al. Gan Augmentation: Augmenting Training Data using Generative Adversarial Networks. [arXiv preprint]; 2018. doi:10.48550/arXiv.1810.10863

[28] Patel M, Wang X, Mao S. Data Augmentation with Conditional GAN for Automatic Modulation Classification. In: Proceedings of the 2nd ACM Workshop on Wireless Security and Machine Learning; 2020. p. 31-36. doi:10.1145/3395352.3402622

[29] Shaham TR, Dekel T, Michaeli T. Singan: Learning a Generative Model from a Single Natural Image. In: Proceedings of the IEEE/CVF International Conference on Computer Vision. 2019. p. 4570-4580. doi:10.48550/arXiv.1905.01164

[30] Mosser L, Kimman W, Dramsch J, Purves S, La Fuente Briceño A, Ganssle G. Rapid Seismic Domain Transfer: Seismic Velocity Inversion and Modeling using Deep Generative Neural Networks. In: 80th Eage Conference and Exhibition 2018. Vol. 2018. p. 1-5. doi:10.3997/2214-4609.201800734

[31] Creswell A, White T, Dumoulin V, Arulkumaran K, Sengupta B, Bharath AA. Generative adversarial networks: An overview. IEEE Signal Process Mag. 2018; 35(1):53-65. doi:10.1109/MSP.2017.2765202

[32] Douzas G, Bacao F. Effective data generation for imbalanced learning using conditional generative adversarial networks. Expert Syst Appl. 2018; 91:464-471.doi:10.1016/j.eswa.2017.09.030

[33] Liu M, Nivlet P, Smith R, BenHasan N, Grana D. Recurrent neural network for seismic reservoir characterization. In: Bhattacharya S, Di H, editors. Advances in Subsurface Data Analytics. Ch. 4. Netherlands: Elsevier; 2022. p. 95-116. doi:10.1016/B978-0-12-822295-9.00010-8

[34] Miele R, Azevedo L, Grana D. Iterative Stochastic Seismic Inversion Method using GAN with Spatiallyadaptive Normalization. In: Third EAGE Conference on Seismic Inversion; 2024. p. 1-5. doi:10.3997/2214-4609.202438056

[35] Gulrajani I, Ahmed F, Arjovsky M, Dumoulin V, Courville AC. Improved training of wasserstein gans. Advance Neural Information Processing Systems; 2017. p. 30.

[36] Tran TT. Improving variogram reproduction on dense simulation grids. Comput Geosci. 1994; 20(7-8):1161-1168. doi:10.1016/0098-3004(94)90069-8

[37] Cox Michael AA, Cox TF. Multidimensional scaling. In: Handbook of Data Visualization. Berlin: Springer Berlin Heidelberg; 2008. p. 315-347. doi:10.1007/978-3-540-33037-0_14

[38] Miele R, Levy S, Linde N, Soares A, Azevedo L. Deep generative networks for multivariate fullstack seismic data inversion using inverse autoregressive flows. Comput Geosci. 2024; 188:105622. doi:10.1016/j.cageo.2024.105622