Skip to main content

Geological realism in Fluvial facies modelling with GAN under variable depositional conditions


This study investigates generative adversarial networks (GANs)’ capacity to model multi-facies distributions of meandering systems. Earlier works showed that GANs outperform geostatistical methods in reproducing complex geometry, like the shapes of fluvial channels. However, the reproduction of geological complexity and geological realism remains an issue when modelling fluvial depositional systems. Meandering systems deposit multiple facies and change facies shape following the migration of rivers. Sand accretes at the inner bank of channels, forming the point bar and erodes the plain at the outer bank to create sediments. Channel fills with mud or sand at the bottom after abandonment due to avulsions or meander cut-offs. Those sedimentary processes yield complex geological patterns. This paper proposes further developing a GAN model, Fluvial GAN, to learn complex multi-facies fluvial patterns across depositional variability. We create a set of meandering facies models by a process-based model, FLUMYTM, for training a GAN and assessing how well it can learn fluvial facies distributions representing sedimentary processes. Fluvial GAN has three distinct enhancements: (i) a One-Hot Encoder for better handling of multi-facies distribution, (ii) a Hybrid-discriminator for better learning geological patterns, and (iii) an improved loss function to prevent mode collapse. We compare Fluvial GAN performance with two more standard configurations using qualitative and quantitative geological features assessments. Fluvial GAN vastly reduces the occurrence of a typical unrealistic feature, channels forming isolated loops, which we called ‘closed channel’ in this study. We analyse the diversity of Fluvial GAN generations via a dimensionality reduction algorithm, UMAP, that plots the training dataset and Fluvial GAN generations together in a 2D space. Fluvial GAN provides good coverage of the uncertainty space represented by the training dataset.

This is a preview of subscription content, access via your institution.

Data Availability

The training dataset used in this study is available at [43]. Supporting data of the results are available at

Code Availability

Codes are available at


  1. Grimaud, J.-L., Ors, F., Lemay, M., Cojan, I., Rivoirard, J.: Preservation and completeness of fluvial meandering deposits influenced by channel motions and overbank sedimentation. J. Geophys. Res. Earth Surface, 2021–006435

  2. Laloy, E., Hérault, R., Jacques, D., Linde, N.: Training-image based geostatistical inversion using a spatial generative adversarial neural network. Water Resour. Res. 54(1), 381–406 (2018)

    Article  Google Scholar 

  3. Chan, S., Elsheikh, A.H.: Parametric generation of conditional geological realizations using generative neural networks. Comput. Geosci. 23(5), 925–952 (2019)

    Article  Google Scholar 

  4. Zhang, T.-F., Tilke, P., Dupont, E., Zhu, L.-C., Liang, L., Bailey, W.: Generating geologically realistic 3D reservoir facies models using deep learning of sedimentary architecture with generative adversarial networks. Pet. Sci. 16(3), 541–549 (2019)

    Article  Google Scholar 

  5. Song, S., Mukerji, T., Hou, J.: Geological facies modeling based on progressive growing of generative adversarial networks (GANs). Comput. Geosci. 25(3), 1251–1273 (2021)

    Article  Google Scholar 

  6. Maharaja, A.: Tigenerator: object-based training image generator. Comput. Geosci. 34(12), 1753–1761 (2008)

    Article  Google Scholar 

  7. Remy, N., Boucher, A., Wu, J.: Applied Geostatistics with SGeMS: A User’s Guide. Cambridge University Press, Cambridge (2009)

    Book  Google Scholar 

  8. Bubnova, A.: On the conditioning of process-based channelized meandering reservoir models on well data. PhD thesis, Université Paris sciences et lettres (2018)

  9. Lopez, S., Cojan, I., Rivoirard, J., Galli, A.: Process-Based Stochastic Modelling: Meandering Channelized Reservoirs. Analogue and Numerical Modelling of Sedimentary Systems: From Understanding to Prediction, pp 139–144. Wiley, Oxford (2009)

    Google Scholar 

  10. Rongier, G.: Lessons learned from simulating fluvial deposits using process-based models and generative adversarial networks. In: 21st Annual Conference of the International Association for Mathematical Geosciences. Beyond Gaussianity: GANs, MPS, Cumulants or Copula approaches? International Association for Mathematical Geosciences (2022)

  11. Bogaart, P.W., Van Balen, R., Kasse, C., Vandenberghe, J.: Process-based modelling of Fluvial system response to rapid climate change—I: model formulation and generic applications. Quat. Sci. Rev. 22 (20), 2077–2095 (2003)

    Article  Google Scholar 

  12. Azevedo, L., Paneiro, G., Santos, A., Soares, A.: Generative adversarial network as a stochastic subsurface model reconstruction. Comput. Geosci. 24(4), 1673–1692 (2020)

    Article  Google Scholar 

  13. Song, S., Mukerji, T., Hou, J.: Bridging the gap between geophysics and geology with generative adversarial networks. IEEE Trans. Geosci. Remote Sens. 60, 1–11 (2021)

    Google Scholar 

  14. Zhang, C., Song, X., Azevedo, L.: U-net generative adversarial network for subsurface facies modeling. Comput. Geosci. 25(1), 553–573 (2021)

    Article  Google Scholar 

  15. Goodfellow, I., Pouget-Abadie, J., Mirza, M., Xu, B., Warde-Farley, D., Ozair, S., Courville, A., Bengio, Y.: Generative adversarial networks. Commun. ACM 63(11), 139–144 (2020)

    Article  Google Scholar 

  16. Hancock, J.T., Khoshgoftaar, T.M.: Survey on categorical data for neural networks. J. Big Data 7(1), 1–41 (2020)

    Article  Google Scholar 

  17. Radford, A., Metz, L., Chintala, S.: Unsupervised representation learning with deep convolutional generative adversarial networks. arXiv:1511.06434 (2015)

  18. Ioffe, S., Szegedy, C.: Batch normalization: Accelerating deep network training by reducing internal covariate shift. In: International Conference on Machine Learning, pp 448–456. PMLR (2015)

  19. LeCun, Y., Bottou, L., Bengio, Y., Haffner, P.: Gradient-based learning applied to document recognition. Proc. IEEE 86(11), 2278–2324 (1998)

    Article  Google Scholar 

  20. Metz, L., Poole, B., Pfau, D., Sohl-Dickstein, J.: Unrolled generative adversarial networks. arXiv:1611.02163 (2016)

  21. Niculescu-Mizil, A., Perlich, C., Swirszcz, G., Sindhwani, V., Liu, Y., Melville, P., Wang, D., Xiao, J., Hu, J., Singh, M., et al.: Winning the KDD cup orange challenge with ensemble selection. In: KDD-Cup 2009 Competition, pp 23–34. PMLR (2009)

  22. Park, T., Liu, M.-Y., Wang, T.-C., Zhu, J.-Y.: Semantic image synthesis with spatially-adaptive normalization. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp 2337–2346 (2019)

  23. Wang, T.-C., Liu, M.-Y., Zhu, J.-Y., Tao, A., Kautz, J., Catanzaro, B.: High-resolution image synthesis and semantic manipulation with conditional GANs. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp 8798–8807 (2018)

  24. Yu, F., Koltun, V.: Multi-scale context aggregation by dilated convolutions. arXiv:1511.07122 (2015)

  25. Wang, P., Chen, P., Yuan, Y., Liu, D., Huang, Z., Hou, X., Cottrell, G.: Understanding convolution for semantic segmentation. In: 2018 IEEE Winter Conference on Applications of Computer Vision (WACV), pp 1451–1460. IEEE (2018)

  26. Dumoulin, V., Visin, F.: A guide to convolution arithmetic for deep learning. arXiv:1603.07285 (2016)

  27. Sun, C., Demyanov, V., Arnold, D.: Comparison of popular generative adversarial network flavours for Fluvial reservoir modelling. In: 82nd EAGE Annual Conference & Exhibition, vol. 2021, pp 1–5. European Association of Geoscientists & Engineers (2021)

  28. Isola, P., Zhu, J.-Y., Zhou, T., Efros, A.A.: Image-to-image translation with conditional adversarial networks. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp 1125–1134 (2017)

  29. Miyato, T., Kataoka, T., Koyama, M., Yoshida, Y.: Spectral normalization for generative adversarial networks. arXiv:1802.05957 (2018)

  30. Lim, J.H., Ye, J.C.: Geometric GAN. arXiv:1705.02894 (2017)

  31. Thanh-Tung, H., Tran, T., Venkatesh, S.: Improving generalization and stability of generative adversarial networks. arXiv:1902.03984 (2019)

  32. Heller, P.L., Paola, C.: Downstream changes in alluvial architecture; an exploration of controls on channel-stacking patterns. J. Sediment. Res. 66(2), 297–306 (1996)

    Google Scholar 

  33. Slingerland, R., Smith, N.D.: River avulsions and their deposits. Annu. Rev. Earth Planet. Sci. 32, 257–285 (2004)

    Article  Google Scholar 

  34. Willems, C.J., Nick, H.M., Donselaar, M.E., Weltje, G.J., Bruhn, D.F.: On the connectivity anisotropy in Fluvial hot sedimentary aquifers and its influence on geothermal doublet performance. Geothermics 65, 222–233 (2017)

    Article  Google Scholar 

  35. Loshchilov, I., Hutter, F.: Decoupled weight decay regularization. arXiv:1711.05101(2017)

  36. King, P.: The connectivity and conductivity of overlapping sand bodies. In: North Sea Oil and Gas reservoirs—II, pp 353–362. Springer (1990)

  37. Renard, P., Allard, D.: Connectivity metrics for subsurface flow and transport. Adv. Water Resour. 51, 168–196 (2013)

    Article  Google Scholar 

  38. Pirot, G., Joshi, R., Giraud, J., Lindsay, M.D., Jessell, M.W.: loopui-0.1: indicators to support needs and practices in 3D geological modelling uncertainty quantification. Geosci. Model Dev. 15(12), 4689–4708 (2022)

    Article  Google Scholar 

  39. McInnes, L., Healy, J., Melville, J.: Umap: Uniform manifold approximation and projection for dimension reduction. arXiv:1802.03426 (2018)

  40. McKie, T., Audretsch, P.: Depositional and structural controls on triassic reservoir performance in the Heron Cluster, ETAP, Central North Sea. In: Geological Society, London, Petroleum Geology Conference Series, vol. 6, pp 285–297. Geological Society of London (2005)

  41. Caers, J.: Modeling Uncertainty in the Earth Sciences. Wiley (2011)

  42. Li, C, Wand, M: Precomputed real-time texture synthesis with markovian generative adversarial networks. In: European Conference on Computer Vision, pp 702–716. Springer (2016)

  43. Sun, C, Demyanov, V, Arnold, D.: GAN River-I: A process-based low NTG meandering reservoir model dataset for machine learning studies. Data in Brief 46, 108785 (2023). Elsevier

    Article  Google Scholar 

  44. Caers, J: Modeling Uncertainty in the Earth Sciences. Wiley (2011)

Download references


The work contained in this paper contains work conducted during a PhD study undertaken as part of the Natural Environment Research Council (NERC) Centre for Doctoral Training (CDT) in Oil & Gas [grant number NEM00578X/1]. This project is sponsored by Heriot-Watt University via their James Watt Scholarship Scheme whose support is gratefully acknowledged. We would like to thank MINES ParisTech for their software FLUMYTM, Process-based channelized reservoir models. CopyrightⒸMINES PARIS-PSL / ARMINES. Free download from We would like to thank Dr Andy Gardiner for his help on sedimentary geology, thank Prof. Ahmed ElSheikh for his suggestions on result evaluation, and thank Dr Zeyun Jiang, Dr Guillaume Rongier and two anonymous reviewers for their reviews that helped improve the paper.


The work contained in this paper contains work conducted during a PhD study undertaken as part of the Natural Environment Research Council (NERC) Centre for Doctoral Training (CDT) in Oil & Gas [grant number NEM00578X/1]. This project is sponsored by Heriot-Watt University via their James Watt Scholarship Scheme.

Author information

Authors and Affiliations


Corresponding author

Correspondence to Chao Sun.

Ethics declarations

Conflict of Interests

The authors have no conflicts of interest to declare that are relevant to this article.

Additional information

Supplementary information

Our codes and results are available at The training dataset used in this study is available at [43].

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Vasily Demyanov and Daniel Arnold contributed equally to this work.


Appendix A: Details of Fluvial GAN model

1.1 A.1 Generator

Table 6 Generator architecture
Table 7 NN Block architecture

1.2 A.2 Hybrid-discriminator

Table 8 CNN-based PatchGAN discriminator
Table 9 Dilated CNN-based PatchGAN discriminator

1.3 A.3 Training

We use a benchmark GPU, RTX3090, to test our Fluvial GAN model’s training and generation speed. Training 200 epochs take around 23 hours for a single RTX3090 and the pre-trained Fluvial GAN can simulate about 16000 realisations per minute. Figure 18 shows the loss curves during training.

Fig. 18
figure 18

Loss curves of Fluvial GAN during training

Appendix B: Details of making training dataset

Table 10 FLUMYTM parameters to create 3D simulations

Rights and permissions

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and Permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Sun, C., Demyanov, V. & Arnold, D. Geological realism in Fluvial facies modelling with GAN under variable depositional conditions. Comput Geosci 27, 203–221 (2023).

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: