Interpretable intelligent gas-bearing reservoir prediction using time–frequency analysis and manifold-regularized semi-supervised GAN

Abstract

As hydrocarbon exploration advances toward deep and complex reservoirs, the identification accuracy of traditional time–frequency (TF) analysis is constrained by strongly heterogeneous geological conditions. Concurrently, while deep learning has shown great potential, mainstream supervised models commonly face the dual challenges of scarce labeled samples and the lack of interpretability in their black-box decision-making processes. To address these challenges, this study proposes an innovative, intelligent prediction framework integrating high-precision TF analysis, manifold-regularized semi-supervised generative adversarial networks, and SHapley Additive exPlanations (SHAP) for interpretability. First, the Fourier-based synchrosqueezing transform was employed to extract two-dimensional TF features with superior energy concentration, effectively overcoming the resolution limits imposed by the Heisenberg uncertainty principle. Subsequently, the manifold-regularized semi-supervised generative adversarial network was developed. By incorporating manifold regularization constraints, the discriminator captures the intrinsic topological structure of large-scale unlabeled samples, effectively leveraging data geometry to significantly enhance generalization capability under sparse-label conditions. Finally, the SHAP method was utilized to conduct a post hoc interpretation. Experimental results on the Marmousi II model demonstrate a remarkable testing accuracy of 98.4%. In a real-world application to deep marine reservoirs in the Sichuan Basin, the framework achieved an 85.0% testing accuracy using only 5% labeled samples. Compared to baseline models, the semi-supervised strategy and manifold regularization contributed accuracy gains of 18.8% and 5.0%, respectively. SHAP analysis further confirms the model’s adaptive capability to extract geophysical features, enabling it to deconstruct the tuning-effect patterns in synthetic data and the low-frequency enhancement/high-frequency attenuation patterns in real data, respectively. This validation of geophysical consistency provides a theoretical foundation for the application of artificial intelligence in complex hydrocarbon exploration.

Keywords

Gas-bearing prediction / Semi-supervised learning / Synchrosqueezing transform / Generative adversarial network / SHapley Additive exPlanations

Funding

This work was supported by the National Natural Science Foundation of China (Grant 42330813 and Grant 42030812), and the Natural Science Foundation of Sichuan Province (Grant 2026NSFSC1138).

Conflict of interest

The authors declare they have no competing interests.

Author contributions

Conceptualization: Shuying Ma, Junxing Cao, Rong Wang Formal analysis: Jun Wang, Lingsen Zhao, Hong Li, Xin Tang Investigation: Junxing Cao, Xudong Jiang Methodology: Shuying Ma Writing–original draft: Shuying Ma Writing–review & editing: Shuying Ma, Junxing Cao, Rong Wang

Availability of data

Data supporting the findings of this study are available from the corresponding authors upon reasonable request.

References

[1] Hammond AL. Bright Spot: Better Seismological Indicators of Gas and Oil. Science. 1974; 185(4150): 515-517. doi: 10.1126/science.185.4150.515

[2] Brown AR. Dim spots: Opportunity for future hydrocarbon discoveries? Lead Edge . 2012; 31(6): 682-683. doi: 10.1190/tle31060682.1

[3] Backus MM , Chen RL. Flat spot exploration. Geophys Prospect. 1975; 23(3): 533-577. doi: 10.1111/j.1365—2478.1975.tb01547.x

[4] Korneev VA , Goloshubin GM , Daley TM , Silin DB. Seismic low—frequency effects in monitoring fluid—saturated reservoirs. Geophysics . 2004; 69(2): 522-532. doi: 10.1190/1.1707072

[5] Silin DB , Korneev VA , Goloshubin GM , Patzek TW. A Hydrologic View on Biot’s Theory of Poroelasticity . Office of Scientific and Technical Information (OSTI); 2004. doi: 10.2172/822181

[6] Mitchell JT , Derzhi N , Lichman E , Lanning EN. Energy absorption analysis: A case study. In: SEG Technical Program Expanded Abstracts 1996. 66th SEG Annual International Meeting; January 1, 1996; Denver, Colorado, USA. ASME; 1996: 1785-1788. doi: 10.1190/1.1826480

[7] Santos LF , Silva RMGE , Gattass M , Silva AC. Direct hydrocarbon indicators based on long short—term memory neural network. In: SEG Technical Program Expanded Abstracts 2019. 89th SEG Annual International Meeting; August 1, 2019; San Antonio, Texas, USA. ASME; 2019: 2373-2377. doi: 10.1190/segam2019—3215628.1

[8] Júnior DAD , Batista da Cruz L , Otávio Bandeira Diniz J , et al. Detection of potential gas accumulations in 2D seismic images using spatio—temporal, PSO, and convolutional LSTM approaches. Expert Syst Appl. 2023; 215: 119337. doi: 10.1016/j.eswa.2022.119337

[9] Ma S , Cao J , Liu Z , Jiang X , Su Z , Xue Y. Gas—bearing prediction of deep reservoir based on DNN embeddings. Front Earth Sci . 2023; 11. doi: 10.3389/feart.2023.1117797

[10] Jiang X , Cao J , Zu S , Xu H , Wang J. Detection of hidden reservoirs under strong shielding based on bi—dimensional empirical mode decomposition and the Teager—Kaiser operator. Geophys Prospect. 2021; 69(5): 1086-1101. doi: 10.1111/1365—2478.13073

[11] Gao J , Song Z , Gui J , Yuan S. Gas—Bearing Prediction Using Transfer Learning and CNNs: An Application to a Deep Tight Dolomite Reservoir. IEEE Geosci Remote Sensing Lett. 2022; 19: 1-5. doi: 10.1109/lgrs.2020.3035568

[12] Yang J , Lin N , Zhang K , Zhang D , Wang D , Zhang J. Variable Selection and Oil—Gas—Bearing Evaluation Using Multicomponent Seismic Data and AMPSO—ML Approaches. SPE J. 2023; 28(03): 1183-1201. doi: 10.2118/214298—pa

[13] Yang JQ , Lin NT , Zhang K , Cui Y , Fu C , Zhang D. Deep learning CNN—APSO—LSSVM hybrid fusion model for feature optimization and gas—bearing prediction. Pet Sci. 2024; 21(4): 2329-2344. doi: 10.1016/j.petsci.2024.02.012

[14] Yang X , Song Z , King I , Xu Z. A Survey on Deep Semi—Supervised Learning. IEEE Trans Knowl Data Eng. 2023; 35(9): 8934-8954. doi: 10.1109/tkde.2022.3220219

[15] Goodfellow I , Pouget—Abadie J , Mirza M , Xu B , Warde—Farley D , Ozair S , et al. Generative adversarial networks. Commun ACM. 2020; 63(11): 139-144. doi: 10.1145/3422622

[16] Odena A. Semi—supervised learning with generative adversarial networks. arXiv. Preprint online 2016. doi: 10.48550/arXiv.1606.01583

[17] Liu M , Jervis M , Li W , Nivlet P. Seismic facies classification using supervised convolutional neural networks and semisupervised generative adversarial networks. Geophysics. 2020; 85(4): O47-O58. doi: 10.1190/geo2019—0627.1

[18] Song L , Yin X , Zhang R , Li J , Zhang J , Li J. Reservoir Discrimination Based on Physic—Informed Semi—Supervised Learning. IEEE Trans Geosci Remote Sens. 2024; 62: 1-14. doi: 10.1109/tgrs.2024.3409578

[19] Song L , Yin X , Yin L. Reservoir Lithology Identification Based on Improved Adversarial Learning. IEEE Geosci Remote Sens Lett . 2023; 20: 1-5. doi: 10.1109/lgrs.2023.3281545

[20] Daubechies I , Lu J , Wu HT. Synchrosqueezed wavelet transforms: An empirical mode decomposition—like tool. Appl Comput Harmon Anal. 2011; 30(2): 243-261. doi: 10.1016/j.acha.2010.08.002

[21] Lundberg SM , Lee SI. A unified approach to interpreting model predictions. In: Guyon I, Von Luxburg U, et al., editors. Advances in Neural Information Processing Systems 30. 31st International Conference on Neural Information Processing Systems (NIPS 2017); December 4—9, 2017; Long Beach, CA, USA. Curran Associates Inc; 2017: 4766-4777.

[22] Huang Z lai , Zhang J , Zhao T hu , Sun Y. Synchrosqueezing S—Transform and Its Application in Seismic Spectral Decomposition. IEEE Trans Geosci Remote Sens. 2016; 54(2): 817-825. doi: 10.1109/tgrs.2015.2466660

[23] Mao X. A Concentrated Time—Frequency Method for Reservoir Detection Using Adaptive Synchrosqueezing Transform. IEEE Geosci Remote Sens Lett. 2022; 19: 1-5. doi: 10.1109/lgrs.2022.3160930

[24] Belkin M , Niyogi P , Sindhwani V. Manifold regularization: A geometric framework for learning from labeled and unlabeled examples. J Mach Learn Res. 2006; 7(85): 2399-2434.

[25] Ni Y , Koniusz P , Hartley R , Nock R. Manifold Learning Benefits GANs. In: Proceedings of the 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). June 19—24, 2022; New Orleans, Louisiana. IEEE; 2022: 11255-11264. doi: 10.1109/cvpr52688.2022.01098

[26] Lecouat B , Foo CS , Zenati H , Chandrasekhar VR. Semi—Supervised Learning with GANs: Revisiting Manifold Regularization. arXiv. Preprint posted online May 23, 2018. Last accessed November 4, 2024. doi: 10.48550/arXiv.1805.08957

[27] Oberlin T , Meignen S , Perrier V. The fourier—based synchrosqueezing transform. In: Proceedings of the 2014 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). May 4—9, 2014; Florence, Italy. IEEE; 2014: 315-319. doi: 10.1109/icassp.2014.6853609

[28] Salimans T , Goodfellow I , Zaremba W , Cheung V , Radford A , Chen X. Improved Techniques for Training GANs. arXiv. Preprint posted online June 10, 2016. Last accessed November 1, 2024. doi: 10.48550/arXiv.1606.03498

[29] Ioffe S , Szegedy C. Batch normalization: Accelerating deep network training by reducing internal covariate shift. In: Proceedings of the 32nd International Conference on Machine Learning. International Conference on Machine Learning; July 7—9 2015; Lille, France. PMLR; 2015; 37: 448-456.

[30] Martin GS , Marfurt KJ , Larsen S. Marmousi—2: An updated model for the investigation of AVO in structurally complex areas. In: SEG Technical Program Expanded Abstracts 2002. 72th SEG Annual International Meeting; January 1, 2002; Salt Lake City, Utah, USA. ASME; 2002: 1979-1982. doi: 10.1190/1.1817083

[31] Ma S , Cao J. A Two—Branch Neural Network for Gas—Bearing Prediction Using Latent Space Adaptation for Data Augmentation—An Application for Deep Carbonate Reservoirs. IEEE Geosci Remote Sens Lett. 2024; 21: 1-5. doi: 10.1109/lgrs.2024.3436836

[32] Zhang K , Lin N , Yang J , Zhang D , Cui Y , Jin Z. An Intelligent Approach for Gas Reservoir Identification and Structural Evaluation by ANN and Viterbi Algorithm—A Case Study From the Xujiahe Formation, Western Sichuan Depression, China. IEEE Trans Geosci Remote Sens. 2023; 61: 1-12. doi: 10.1109/tgrs.2023.3247183

[33] Song Z , Li S , He S , Yuan S , Wang S. Gas—Bearing Prediction of Tight Sandstone Reservoir Using Semi—Supervised Learning and Transfer Learning. IEEE Geosci Remote Sens Lett. 2022; 19: 1-5. doi: 10.1109/lgrs.2022.3177314

[34] Ma S , Cao J. An Efficient Transformer Model Enhanced by S—Transform and Transfer Learning for Predicting Gas Distribution in Deeply Buried Reservoirs. IEEE Trans Geosci Remote Sens. 2025; 63: 1-14. doi: 10.1109/tgrs.2025.3558731