Abstract
Modeling the semivariogram to characterize spatial continuity requires expert geostatistical knowledge and domain expertise about the spatial phenomenon of interest. Moreover, although practitioners may have experience in semivariograms, their interpretations may vary due to experimental semivariogram noise and ambiguity. In general, modeling semivariograms remains highly subjective. This paper presents a data-driven, deep learning-based automated semivariogram modeling method known as automatic semivariogram modeling with convolution-based deep learning (ASMC) that improves the utilization of available spatial information to reduce the subjectivity of semivariogram modeling. Training models are generated by sequential Gaussian simulation (SGS) and labeled with their associated semivariogram parameters (i.e., maximum correlation length, aspect ratio of major and minor direction ranges, and azimuth of major direction). ASMC consists of two convolutional neural networks (CNNs). The first CNN model maps the sparse spatial samples to the exhaustive SGS-derived spatial models, and the second CNN maps the SGS spatial model to the semivariogram parameters. Both CNNs are trained with realistic spatial training data, and their validity is also checked with validation data withheld from training. Two-dimensional synthetic, but realistic, case studies demonstrate that the first CNN successfully learns the spatial characteristics among spatial data and generates realistic subsurface model estimates. The second CNN learns the spatial context of the estimated subsurface model and successfully predicts the semivariogram parameters with greater than 96% accuracy. The proposed machine, deep learning-based workflow improves the utilization and objectivity of spatial data in semivariogram-based spatial continuity modeling. With the optimal design of the experiment for training and tuning of model hyperparameters, this method may be generalized for application for a wide range of spatial modeling projects.
Similar content being viewed by others
Code and Data availability
We carried out the neural network aspects of our proposed workflow using Tensorflow and performed geostatistical analysis using the GeostatsPy Python package (Pyrcz et al. 2021) reimplementation of GSLIB (Deutsch and Journel 1998). The code and data will be made available via the authors' repositories https://github.com/GeostatsGuy/GeostatsPy and https://github.com/whghdrms.
References
Bergen KJ, Johnson PA, Maarten V, Beroza GC (2019) Machine learning for data-driven discovery in solid Earth geoscience. Science 363(6433).
Chilès J, Delfiner P (2012) Geostatistics: modeling spatial uncertainty, 2nd edn. Wiley, New York
Cho Y, Cao Y, Zagayevskiy Y, Wong T, Munoz Y (2020) Kriging-based monitoring of reservoir gas saturation distribution using time-lapse multicomponent borehole gravity measurements: Case study, Hastings Field. J Pet Sci Eng 190:107054
Dasassis N, Renard D (2013) Automatic semivariogram modeling by iterative least squares: univariate and multivariate cases. Math Geosci 45:453–470. https://doi.org/10.1007/s11004-012-9434-1
David M (1977) Geostatistical ore reserve estimation. Elsevier, Amsterdam
Deutsch CV, Journel A (1998) GSLIB geostatistical software library and user’s guide. Oxford Press, New York
Dimitrakopoulos R (1993) Artificially intelligent geostatistics: a framework accommodating qualitative knowledge-information. Math Geol 25(3):261–279
Dowd PA, Sarac C (1994) A neural network approach to geostatistical simulation. Math Geol 26(4):491–503
Emery X (2010) Iterative algorithms for fitting a linear model of coregionalization. Comput Geosci 36(9):1150–1160
Guo Y, Liu Y, Oerlemans A, Lao S, Wu S, LewMS, (2016) Deep learning for visual understanding: a review. Neurocomputing 187:27–48. https://doi.org/10.1016/j.neucom.2015.09.116
Jensen J, Lake LW, Corbett PW, Goggin D (1997) Statistics for petroleum engineers and geoscientists. Gulf Professional Publishing, New Jersey
Jo H, Santos JE, Pyrcz MJ (2020) Conditioning well data to rule-based lobe model by machine learning with a generative adversarial network. Energy Explor Exploit 38(6):2558–2578
Jo H, Pan W, Santos JE, Jung H, Pyrcz MJ (2021) Machine learning assisted history matching for a deepwater lobe system. J Pet Sci Eng, 109086.
Journel AG, Huijbregts CHJ (1978) Mining geostatistics. Academic Press, London
Kerry R, Oliver MA (2007) Comparing sampling needs for variograms of soil properties computed by the method of moments and residual maximum likelihood. Geoderma 104(4):383–396. https://doi.org/10.1016/j.geoderma.2007.04.019
Larrondo PF, Neufeld CT, Deutsch CV (2003) VARFIT: A program for semi-automatic semivariogram modeling. In: Deutsch CV (ed) CCG Annual report 5. University of Alberta, Edmonton, p 17
LeCun Y, Boser B, Denker JS, Henderson D, Howard RE, Hubbard W, Jackel LD (1989) Backpropagation applied to handwritten zip code recognition. Neural Comput 1:541–551
Li Z, Zhang X, Clarke KC, Liu G, Zhu R (2018) An automatic variogram modeling method with high reliability fitness and estimates. Comput Geosci 120:1150–1160. https://doi.org/10.1016/j.cageo.2018.07.011
Liu W, Pyrcz MJ (2020) A spatial correlation-based anomaly detection method for subsurface modeling. Math Geosci. https://doi.org/10.1007/s11004-020-09892-z
Loffe S, Szegedy C (2015) Batch normalization: Accelerating deep network training by reducing internal covariate shift. In: Paper presented at the 32nd International Conference on Machine Learning, Lille, France, 6–11 July 2015
Maldonado-Cruz E, Pyrcz MJ (2021) Tuning machine learning dropout for subsurface uncertainty model accuracy. J Pet Sci Eng, 108975.
Mardia KV, Marshall RJ (1984) Maximum likelihood estimation of models for residual covariance in spatial regression. Biometrika 71(1). https://doi.org/10.2307/2336405
Mosser L, Dubrule O, Blunt MJ (2017) Reconstruction of three-dimensional porous media using generative adversarial neural networks. Phys Rev E 96(4):043309. https://doi.org/10.1103/PhysRevE.96.043309
Nelder JA, Mead R (1965) A simplex method for function minimization. Comput J 7(4):308–313. https://doi.org/10.1093/comjnl/7.4.308
Oliver MA, Webster R (2014) A tutorial guide to geostatistics: computing and modelling variograms and kriging. CARENA 13:56–69. https://doi.org/10.1016/j.catena.2013.09.006
Pan W, Torres-Verdín C, Pyrcz MJ (2021) Stochastic Pix2pix: a new machine learning method for geophysical and well conditioning of rule-based channel reservoir models. Nat Resour Res 30(2):1319–1345
Pardo-Igúzquiza E, Mardia KV, Chica-Olmo M (2009) MLMATERN: A computer program for maximum likelihood inference with the spatial Matérn covariance model. Comput Geosci 35:1139–1150. https://doi.org/10.1016/j.cageo.2008.09.009
Park J, Datta-Gupta A, Singh A, Sankaran S (2021) Hybrid physics and data-driven modeling for unconventional field development and its application to US onshore basin. J Pet Sci Eng, 109008.
Pyrcz MJ, Deutsch CV (2014) Geostatistical reservoir modeling. Oxford University Press, New York
Pyrcz MJ, Gringarten E, Frykman P, Deutsch CV (2006) Representative input parameters for geostatistical simulation. In: Coburn TC, Yarus JM, Chambers RL (eds) Stochastic modeling and geostatistics: Principles, methods, and case studies, volume II: AAPG Computer Applications in Geology 5, pp 123–137. https://doi.org/10.1306/1063811CA53230
Pyrcz MJ, Jo H, Kupenko A, Liu W, Gigliotti AE, Salomaki T, Santos JE (2021) GeostatsPy. https://github.com/GeostatsGuy/GeostatsPy
Ronneberger O, Fischer P, Brox T (2015) U-Net: convolutional networks for biomedical image segmentation. arXiv preprint https://arxiv.org/abs/1505.04597. Accessed 6 May 2020
Salazar JJ, Lake LW (2020) The physical meaning of the koval factor. Math Geosci 52:1017–1033. https://doi.org/10.1007/s11004-020-09883-0
Santos JE, Xu D, Jo H, Landry CJ, Prodanović M, Pyrcz MJ (2020) PoreFlow-Net: A 3D convolutional neural network to predict fluid flow through porous media. Adv Water Resour 138:103539. https://doi.org/10.1016/j.advwatres.2020.103539
Santos JE, Yin Y, Jo H, Pan W, Kang Q, Viswanathan HS, Prodanović M, Pyrcz MJ, Lubbers N (2021) Computationally efficient multiscale neural networks applied to fluid flow in complex 3D porous media. Transp Porous Media. https://doi.org/10.1007/s11242-021-01617-y
Srivastava N, Hinton G, Krizhevsky A, Sutskever I, Salakhutdinov R (2014) Dropout: a simple way to prevent neural networks from overfitting. J Mach Learn Res 15:1929–1958
Wang Y, Arns CH, Rahman SS, Arns J (2018) Porous structure reconstruction using convolutional neural networks. Math Geosci 50:781–799. https://doi.org/10.1007/s11004-018-9743-0
Wilde BJ, Deutsch CV (2012) Automatic variogram modeling from censored variogram volumes. In: Deutsch CV (ed) CCG Annual report 14. University of Alberta, Edmonton, p 127
XuejiaoW, Qiuyan T, LianghaoW, Dongxiao L, Ming Z (2015) Deep convolutional architecture for natural image denoising. In: Paper presented at the 2015 international conference on wireless communications & signal processing (WCSP), Nanjing, China
Yasojima C, Protzio K, Meiguins B, Neto N, Morais J (2019) A new methodology for automatic cluster-based kriging using k-nearest neighbor and genetic algorithms. Information 10(11):357. https://doi.org/10.3390/info10110357
Yeh RA, Chen C, Lim TY, Schwing AG, Hasegawa-Johnson M, Do MN (2016) Semantic image inpainting with deep generative models. Paper presented at the 30th IEEE conference on computer vision and pattern recognition, Honolulu, Hawaii, USA, 21–26 July 2017. Institute of Electrical and Electronics Engineers Inc, Los Alamitos, pp 6882–6890
Zhang K, Zuo W, Chen Y, Meng D, Zhang L (2017) Beyond a Gaussian Denoiser: residual learning of deep CNN for image denoising. IEEE Trans Image Process 26:3142–3155. https://doi.org/10.1109/TIP.2017.2662206
Acknowledgements
The authors thank the two reviewers for their critical feedback and their help to improve the quality of the manuscript. The authors also appreciate the support of the DIRECT consortium for subsurface data analytics and machine learning at The University of Texas at Austin. The corresponding author is also thankful to the Hildebrand Department of Petroleum and Geosystems Engineering at The University of Texas at Austin, Texas, USA.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Jo, H., Pyrcz, M.J. Automatic Semivariogram Modeling by Convolutional Neural Network. Math Geosci 54, 177–205 (2022). https://doi.org/10.1007/s11004-021-09962-w
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11004-021-09962-w