Construction of a structural model of the geological environment based on modern airborne geophysical survey and 2D CDPM seismic data

The construction of a structural-tectonic model of the environment in the early stages of exploration is associated with significant uncertainties due to the sparse network of seismic profiles. The choice of one or another algorithm for interpolation of times and velocities of seismic waves without involvement of additional a priori information may significantly affect the result of structural constructions.
Aim. With the purpose of detailing the structural model and minimizing uncertainties and geological risks at early stages of geological exploration, we developed an approach to reconstruct the morphology of horizons in the interprofile space of seismic data, using as a priori information of modern airborne geophysical methods.
Materials and methods. To solve the problem, we used methods of machine learning with a teacher. On the model data we tested the most popular algorithms of machine learning. It was concluded about the applicability and limitations of the used algorithms. The main stages of solving the problem are described in detail. The main categories of uncertainties that accompany structural prediction, the possibilities of their identification and quantification are considered. Approaches to analyze the influence of features on prediction results are described, including for flexible machine learning models. An approach based on a heuristic search algorithm is proposed to automate poorly formatted procedures, such as feature selection and optimization of hyperparameters of machine learning models. All computations were performed in the Python programming language, using open source libraries.
Results. Using examples of model and real data, we have demonstrated a significant refinement of structural plan of the horizons, based on the prediction, taking into account a priori information, in comparison with the classical interpolation algorithms.
Conclusion. The results allow the conclusion about high efficiency of involvement of remote geophysical methods at the stage of structural analysis of the early stages of geological exploration. Such a comprehensive analysis allows to obtain a more reliable geological model, to focus the attention on the most promising objects with less geological risks when planning further detailed exploration works.

1. Sheriff R., Geldart L. Seismorazvedka. Obrabotka i interpretacia dannih. [Seismic Exploration. Data processing and interpretation]. Т. 2. Moscow: Mir, 1987, 400 p. (In Russ.)

2. Boganik G.N., Gurvich I.I. Seismorazvedka. Uchebnik dlya vuzov. [Seismic prospecting: Textbook for high schools]. Tver: Publishing house of AIS, 2006, 744 p. (In Russ.)

3. Gardner G.H.F., Gardner L.W., Gregory A.R. Formation velocity and density — the diagnostic basics for stratigraphic traps // Geophysics, 1974, vol. 39, pp. 770–849.

4. Rozhkov A.S. Determination of density, velocity and boundaries of layered geological medium of near-wellbore space with curvilinear interfaces. // Geofizicheskie issledovaniya [journal of Geophysical Research], 2016, vol. 17, no. 4, pp. 67–77. (In Russ.)

5. Shreibman V.I., Zhdanov M.S., Vitvitsky O.V. Korrelyacionnie metodi preobrazovaniya i interpretacii geofizicheskih anomalii [Correlation methods of transformation and interpretation of geophysical anomalies]. Moscow: Nedra, 1977, 237 p.

6. Kolmakov A.V. Machine learning methods in tasks of complex interpretation of potential fields and seismic data [GeoEvrazia-2021 Geologorazvedka v sovremennih realiyah], 2021, vol. II, pp. 76–80. (In Russ.)

7. Kolmakov A.V., Trusov A.A., Meisner A.L., Grigoryev G.S. Forecasting of the reflectors morphology in the interprofile space by potential field data by deep learning methods [ GeoEvrazia-2020. Sovremennie tehnologii izucheniya i osvoeniya nedr Evrazii], 2021, vol. III, pp. 76–80. (In Russ.)

8. Shklyaruk A., Kuznetsov K., Atutyunyan D. Lygin I. Algorithms for constructing structural surfaces by geophysical data based on neural networks // Engineering and Mining Geophysics, 2021, pp. 1–6. (In Russ.)

9. Gulin V.D., Ananyev V.V., Grigoryev G.S., Zaitsev S.V. et al. Application of machine learning algorithms to reconstruct the structural framework from potential field data in areas with no seismic data. [Proceedings of the 48th session of the D.G. Uspenskiy–V.N. Strakhov. International Scientific Seminar], 2022, pp. 90–94.

10. Blokh Y.I. Interpretaciya gravitacionnih i magnitnih anomaliy. Uchebnoe posobiye dlya vuzov. [Interpretation of Gravitational and Magnetic Anomalies. Textbook for students and universities], 2009, 232 p

11. Geo I/O for Subsurface. Random grids [Electronic Resource]. Access: https://code.agilescientific.com/gio/userguide/Random_grids.html

12. Andrew Trask, Felix Hill, Scott Reed, jack Rae, Chris Dyer, Phil Blunsom. Neural Arithmetic Logic Units. Advances in Neural Information Processing Systems, 2018, vol. 31, pp. 8046–8055

13. Scott M. Lundberg and Su-In Lee. A unified approach to interpreting model predictions. Advances in Neural Information Processing Systems, 2017, pp. 4765–4774.

14. Lloyd S Shapley. A value for n-person games. Contributions to the Theory of Games 2.28, 1953, pp. 307–317.

15. Roberts D. R., Bahn V., Ciuti S., Boyce M.S., Elith j., Guillera-Arroita G., Hauenstein S., Lahoz-Monfort j.j., Schroeder B., Thuiller W., Warton D.I., Wintle B.A., Hartig F., Dormann C.F. Cross-validation strategies for data with temporal, spatial, hierarchical, or phylogenetic structure. ECOGRAPHY, 2017, vol. 40 (8), pp. 913–929.