Volumetric Delineation of Paleochannel Bodies within an Interval of Interest from Seismic Data Based on Synthetic Training of a Neural Network

Delineation of paleochannel bodies from seismic data remains one of the practical tasks in predicting the lithofacies heterogeneity of terrigenous reservoirs. The paper presents a method for layer-by-layer delineation of paleochannel objects within a study interval bounded by two reflecting horizons using a neural network trained on synthetic examples, followed by construction of a volumetric probability model of paleochannel occurrence. The target interval is divided between the reference horizons into proportional slices; for each level, a composite is generated in which the R, G, and B color channels are filled with amplitudes from adjacent levels. Importantly, this approach better reflects the vertical variability of the seismic wavefield than frequencycolor composites obtained by spectral decomposition. A probability map of channel-type objects is then calculated. The training dataset is created by superimposing templates of modern river morphology prepared from open sources onto real seismic backgrounds, with variations in scale, contrast, brightness, inversion, and noise. At the application stage, global and local predictions are compared; test-time augmentation, several variants of local contrast enhancement, a sliding window with a Hann weighting function, and morphological post-processing are used. The result is a series of probability maps of channel bodies that can be used for preliminary tracking within the reservoir interval. The method is considered a support tool for seismic interpreters and sedimentologists rather than a standalone replacement for geological and geophysical analysis.

1. Alekseeva P.A., Kalugin A.A., Kiryanova T.N. (2022). Delineation of Paleochannels in the Tyumen Formation Deposits Using a Neural Network Based on 3D Seismic Data. Geofizika, (3), pp. 9–15. (In Russ.) DOI: 10.34926/geo.2022.44.85.001

2. Bergen K.J., Johnson P.A., de Hoop M.V., Beroza G.C. (2019). Machine learning for data-driven discovery in solid Earth geoscience. Science, 363(6433), DOI: 10.1126/science.aau0323

3. Boyd R., Dalrymple R.W., Zaitlin B.A. (2006). Estuarine and incisedvalley facies models. In: Posamentier H.W., Walker R.G. (eds.), Facies Models Revisited. SEPM Special Publication, 84, pp. 171–235. https://doi.org/10.2110/pec.06.84.0171

4. Catuneanu O., Galloway W.E., Kendall C.G.St.C., Miall A.D., Posamentier H.W., Strasser A., Tucker M.E. (2011). Sequence stratigraphy: methodology and nomenclature. Newsletters on Stratigraphy, 44(3), pp. 173–245. DOI: 10.1127/0078-0421/2011/0011

5. Chopra S., Marfurt K.J. (2007). Seismic Attributes for Prospect Identification and Reservoir Characterization. Tulsa: Society of Exploration Geophysicists, 456 p. (SEG Geophysical Developments Series No. 11). https://doi.org/10.1190/1.9781560801900

6. Daroya R., Rowley T., Flores J., Friedmann E., Bennitt F., An H., Simmons T., Hughes M.J., Kluetmeier C.L., Kica S., Vélez J.D., Esenther S.E., Howard T.E., Ye Y., Turcotte A., Gleason C., Maji S. (2025). RiverScope: High-Resolution River Masking Dataset. DOI: 10.5281/zenodo.15376394

7. Dramsch J.S. (2020). 70 years of machine learning in geoscience in review. Advances in Geophysics, pp. 1–55. https://doi.org/10.1016/bs.agph.2020.08.002

8. Hart B.S. (2008). Channel detection in 3-D seismic data using sweetness. AAPG Bulletin, 92(6), pp. 733–742. https://doi.org/10.1306/02050807127

9. Kontorovich A.E., Kontorovich V.A., Ryzhkova S.V., Shurygin B.N., Vakulenko L.G., Gaideburova E.A., Danilova V.P., Kazanenkov V.A., Kim N.S., Kostyreva E.A., Moskvin V.I., Yan P.A. (2013). Jurassic paleogeography of the West Siberian sedimentary basin. Russian Geology and Geophysics, 54(8). pp. 972–1012. (In Russ.)

10. Kontorovich A.E., Ershov S.V., Kazanenkov V.A., Karogodin Yu.N., Kontorovich V.A., Lebedeva N.K., Nikitenko B.L., Popova N.I., Shurygin B.N. (2014). Cretaceous paleogeography of the West Siberian sedimentary basin. Russian Geology and Geophysics, 55(5–6), pp. 745–776. (In Russ.) DOI: 10.15372/GiG20140504

11. Kupriyanova E.K., Saifutdinov R.Sh., Moiseev S.V., Kerusov I.N., Alekseeva P.A., Vorobyova L.F. (2020). A New Methodological Approach to the Assessment of Paleochannel Body Characteristics Based on Attribute Analysis in Various Geological Conditions. Geology and Mineral Resources of Siberia, 3(43), pp. 23–29. (In Russ.) DOI: 10.20403/2078-0575-2020-3-23-29

12. Lin T., Goyal P., Girshick R., He K. & Dollar P. (2017). Focal Loss for Dense Object Detection. 2017 IEEE International Conference on Computer Vision (ICCV), pp. 2999–3007. https://doi.org/10.1109/ICCV.2017.324

13. Miroshnichenko D.E., Alekseeva P.A., Kiryanova T.N., Kerusov I.N., Barantsev A.A. (2024). Experience in studying the structure of channel systems using convolutional neural networks. EAGE Conference & Exhibition “Saint Petersburg 2024”, Saint Petersburg, April 8–11, 2024.

14. Olneva T.V., Zhukovskaya E.A. (2016). Seismic Vision of Geological Processes and Phenomena: Channel Deposits of Continental Sedimentary Environments. Geofizika, (2), pp. 2–9. (In Russ.)

15. Olneva T.V., Zhukovskaya E.A., Murtazin D.G. (2018). Influence of Paleochannel Meandering Type on the Potential Volume of Lithological Trap and the Efficiency of Geological Exploration. Geofizika, (4), pp. 75–81. (In Russ.)

16. Olneva T.V., Oreshkova M.Yu., Butorin A.V., Egorov A.S. (2024). Morphometric Approach to Quantitative Assessment of Thickness of Sediments Meandering Paleochannels of the Tyumen Formation of the West Siberian Oil and Gas Basin. Georesursy = Georesources, 26(3), pp. 143–150. (In Russ.) https://doi.org/10.18599/grs.2024.3.15

17. Pekel J., Cottam A., Gorelick N. & Belward A.S. (2016). High-resolution mapping of global surface water and its long-term changes. Nature, 540(7633), pp. 418–422. https://doi.org/10.1038/nature20584

18. Posamentier H.W. (2001). Lowstand Alluvial Bypass Systems: Incised vs. Unincised. AAPG Bulletin, 85(10), pp. 1771–1793. https://doi.org/10.1306/8626d06d-173b-11d7-8645000102c1865d

19. Posamentier H.W., Davies R.J., Cartwright J.A. & Wood L. (2007). Seismic geomorphology - an overview. Geological Society, London, Special Publications, 277(1), pp. 1–14. https://doi.org/10.1144/GSL.SP.2007.277.01.01

20. Ronneberger O., Fischer P. & Brox T. (2015). U-Net: Convolutional Networks for Biomedical Image Segmentation. Lecture Notes in Computer Science, pp. 234–241. https://doi.org/10.1007/978-3-319-24574-4_28

21. Sakhnyuk V.I., Novikov E.V., Sharifullin A.M., Belokhin V.S., Antonov A.P., Karpushin M.U., Bolshakova M.A., Afonin S.A., Sautkin R.S., Suslova A.A. (2022). Machine learning applications for well-logging interpretation of the Vikulov Formation. Georesursy = Georesources, 24(2), pp. 230–238. (In Russ.) https://doi.org/10.18599/grs.2022.2.21

22. Wu X., Liang L., Shi Y. & Fomel S. (2019). FaultSeg3D: Using synthetic data sets to train an end-to-end convolutional neural network for 3D seismic fault segmentation. Geophysics, 84(3), IM35–IM45. https://doi.org/10.1190/geo2018-0646.1