Introduction
The Scarab Field lies within the West Delta Deep Marine (WDDM) concession, 50 to 100 km offshore the northwestern Nile Delta in 250 to 850 m of water. Gas is trapped in Pliocene channelized sandstones of the El Wastani Formation, deposited as slope-channel canyon fills. The primary reservoirs are two vertically stacked, unconnected channel systems with rapid lateral and vertical facies variation.
The field shows a strong bright-spot response, but flat-spot expression at the gas-water contact (GWC) is inconsistent. Amplitude-based interpretation alone has difficulty separating lithology within the channel fill and placing fluid contacts with confidence. Earlier work on Scarab relied on interpreter-driven seismic analysis. This study applies a Principal Component Analysis (PCA) and Self-Organizing Map (SOM) workflow to classify lithofacies, delineate gas-bearing channels and quantify GWC accuracy against well control.
Data
- One 19 km, 2D post-stack time-migrated seismic line (4 ms sample interval, ~25 Hz dominant frequency, 10 to 80 Hz bandwidth), provided by the Egyptian General Petroleum Corporation (EGPC)
- Five wells: Scarab-1, Scarab-2, Scarab-Da, Scarab-Dd and Scarab-De, tied to seismic with correlation coefficients of 0.72 to 0.78
- Electrofacies defined from gamma ray, shale volume, effective porosity and deep resistivity logs
Method
Attribute generation. Twenty-eight seismic attributes were computed across instantaneous, geometric (curvature) and spectral decomposition families, including CWT and maximum-entropy spectral attributes, sweetness and thin-bed indicators.
Attribute selection with PCA. The first 10 principal components, which account for about 85% of cumulative variance, were retained. Attributes contributing at least 70% to a retained eigenvector were selected, reducing the set to 15. These include envelope amplitude, CWT spectral magnitude and voice from 25 to 85 Hz, instantaneous frequency, thin bed, curvature structure, instantaneous Q, Hilbert and sweetness.
SOM classification in Paradise. The 15 selected attributes were classified with SOM in the Paradise AI workbench. Topologies of 4×4, 6×6, 8×8 and 10×10 neurons were tested. The 8×8 topology (64 neurons) gave the best balance between geological detail and noise: 4×4 under-resolved the channel fill, and 10×10 over-segmented it. Training converged after 76 epochs with no unused neurons.
Calibration and validation. Neurons were assigned to lithofacies by the overlap of probability density functions between neuron-associated petrophysical responses and well-defined lithofacies. Classification accuracy was then measured with confusion matrices against independent well data, and SOM-derived GWCs were compared with log-defined contacts.
Results
Six lithofacies. The SOM resolved gas-prone sandstone (Sandstone-1), lower-quality or water-bearing sandstone (Sandstone-2), clayey sand, clayey limestone, siltstone and claystone. Sandstone-1 corresponds to neurons 56, 63 and 64, and coincides with low gamma ray and high effective porosity at Scarab-1, Scarab-Da and Scarab-De.
Classification accuracy. Sandstone-1 was identified with 0.85 accuracy on the confusion-matrix diagonal. Overall balanced accuracy across the six classes was about 70%.
Gas-water contacts. At Scarab-1, the SOM-derived GWC falls at ~1,824 m MD against a log-defined contact at 1,821 m MD (~3 m misfit). At Scarab-Dd, the SOM places the regional GWC at 1,850 m MD, matching the log-defined depth. Amplitude-only interpretation produced misfits of about 20 m. The SOM workflow reduced GWC depth error by 85 to 92.5%, depending on the benchmark.
Perched water at Scarab-Dd. The SOM also resolved a second contact at 1,872 m MD, about 22 m below the regional GWC. Effective porosity stays high and shale volume stays low across it, while resistivity drops sharply. The authors interpret this as perched water held by stratigraphic compartmentalization rather than a transition zone or a shale increase. Scarab-Dd was plugged back at 1,850 m MD, where the SOM shows the onset of non-reservoir facies.
Channel-edge heterogeneity. Sandstone facies correlate with observed bright and flat spots near Scarab-2. Abrupt sand-to-clay transitions at channel margins indicate stratigraphic barriers that may limit lateral connectivity.
Conclusions
- PCA reduced 28 seismic attributes to 15 while retaining more than 85% of cumulative variance.
- An 8×8 SOM in Paradise classified the seismic samples into six lithofacies, calibrated to well logs.
- SOM-derived gas-water contacts matched log-defined contacts within 0 to 3 m, compared with about 20 m for amplitude-only interpretation.
- The workflow distinguished a perched-water interval from the regional contact, which bears directly on reserve estimates and completion design.
- The study used a single 2D line, so channel width, sinuosity and lateral connectivity were not mapped. The authors note the workflow applies directly to 3D seismic volumes.
Citation
Khalifa, T., Eleslambouly, A., Khalil, A. E. & Ismail, A. Unsupervised neural clustering and seismic attribute analysis for gas channel delineation in the Scarab field. Scientific Reports 16, 38097 (2026). https://doi.org/10.1038/s41598-026-75298-2
Published open access under a Creative Commons Attribution 4.0 International License. Data courtesy of the Egyptian General Petroleum Corporation (EGPC). SOM analysis performed with the Paradise AI workbench from Geophysical Insights.
- Ghassal, B. I., Atfy, E., Sachse, H., Littke, R. & V. Source rock potential of the Middle Jurassic to middle Pliocene, onshore Nile Delta basin, Egypt. Arab. J. Geosci. 9, 744 (2016).
- Ismail, A., Ewida, H. F., Al-Ibiary, M. G., Gammaldi, S. & Zollo, A. Identification of gas zones and chimneys using seismic attributes analysis at the Scarab field, offshore, Nile Delta, Egypt. Pet. Res. 5, 59–69 (2020).
- Ghoneimi, A., Farag, A., Bakry, A. & Nabih, M. A new deeper channel system predicted using seismic attributes in Scarab gas field, West Delta Deep Marine concession, Egypt. J. Afr. Earth Sc. 177, 104155 (2021).
- Samuel, A., Kneller, B., Raslan, S., Sharp, A. & Parsons, C. Prolific deep-marine slope channels of the Nile Delta, Egypt. AAPG Bull. 87, 541–560 (2003).
- Abuamarah, B. A., Nabawy, B. S., Shehata, A. M., Kassem, O. M. & Ghrefat, H. Integrated geological and petrophysical characterization of Oligocene deep marine unconventional poor to tight sandstone gas reservoir. Mar. Pet. Geol. 109, 868–885 (2019).
- Turner, R. et al. Structural and stratigraphic controls on reservoir architecture: A case study from the lower Oligocene Vicksburg formation, Brooks County, Texas. Mar. Pet. Geol. 160, 106627 (2024).
- Cross, N. E. et al. Three-dimensional seismic geomorphology of a deep-water slope-channel system: The Sequoia field, offshore west Nile Delta, Egypt. AAPG Bull. 93, 1063–1086 (2009).
- Mohamed, I. A., El-Mowafy, H. Z. & Fathy, M. Prediction of elastic properties using seismic prestack inversion and neural network analysis. Interpretation 3, T57–T68 (2015).
- Mokhtar, M., Saad, M. & Selim, S. Reservoir architecture of deep marine slope channel, Scarab field, offshore Nile Delta, Egypt: application of reservoir characterization. Egypt. J. Petroleum 25, 495–508 (2016).
- Farouk, S. et al. Petrographical, petrophysical, and geomechanical characterization of the Pliocene sandstone reservoirs of the Scarab Gas Field, deepwater Nile Delta, Egypt: Inferences on reservoir development. J. Afr. Earth Sc. 214, 105259 (2024).
- Abd El-Gawad, E. A. et al. Static reservoir modeling of El Wastani formation, for justifying development plans, using 2D seismic and well log data in Scarab field, offshore Nile Delta, Egypt. J. Afr. Earth Sc. 158, 103546 (2019).
- Nabih, M. et al. Rock physics analysis from predicted Poisson’s ratio using RVFL based on Wild Geese Algorithm in Scarab gas field in WDDM concession, Egypt. Mar. Pet. Geol. 147, 105949 (2023).
- Mohamed, I. A. Neural Network Analysis versus Pre-Stack Seismic Inversion for Reservoir Characterization, Scarab Field, Offshore Nile Delta, Egypt. In LIPE & AAPG European Region Northern Arabia Geoscience Conference & Exhibition, Lebanon, May 27–29, 2014.
- Mohamed, I. A., El-Mowafy, H. Z., Kamel, D. & Heikal, M. Prestack seismic inversion versus neural-network analysis: A case study in the Scarab field offshore Nile Delta, Egypt. Lead. Edge 33, 498–506 (2014).
- Ismail, A., Ewida, H. F., Al-Ibiary, M. G. & Zollo, A. Application of AVO attributes for gas channels identification, West offshore Nile Delta, Egypt. Pet. Res. 5, 112–123 (2020).
- Othman, A., Fathy, M. & Mohamed, I. A. Application of artificial neural network in seismic reservoir characterization: A case study from offshore Nile Delta. Earth Sci. Inform. 14, 669–676 (2021).
- El-Dabaa, S. A., Metwalli, F. I., Maher, A. & Ismail, A. Unsupervised machine learning-based multi-attributes analysis for enhancing gas channel detection and facies classification in the Serpent field, offshore Nile Delta, Egypt. Geomech. Geophys. Geo-Energy Geo-Resources 10, 185 (2024).
- Song, C. et al. Unsupervised seismic facies analysis with spatial constraints using regularized fuzzy c-means. J. Geophys. Eng. 14, 1535–1543 (2017).
- Roden, R., Smith, T. & Sacrey, D. Geologic pattern recognition from seismic attributes: Principal component analysis and self-organizing maps. Interpretation 3, SAE59–SAE83 (2015). https://doi.org/10.1190/INT-2015-0037.1
- Lubo-Robles, D., Bedle, H., Marfurt, K. J. & Pranter, M. J. Evaluation of principal component analysis for seismic attribute selection and self-organizing maps for seismic facies discrimination in the presence of gas hydrates. Mar. Pet. Geol. 150, 106097 (2023).
- Aly Ismail, A., Abdel Kader Boukhary, M. & Ibrahim AbdelNaby, A. Subsurface stratigraphy and micropaleontology of the Neogene rocks, Nile Delta, Egypt. Geologia Croatica 63, 1–26 (2010).
- El-Kahawy, R. M., Aboul-Ela, N., El-Barkooky, A. N. & Kassab, W. G. Bio-sequence stratigraphy of the Neogene: an example from El-Wastani gas field, onshore Nile Delta, Egypt. J. Micropalaeontology 42, 147–169 (2023).
- Makled, W. A., Mandur, M. M. & Langer, M. R. Neogene sequence stratigraphic architecture of the Nile Delta, Egypt: A micropaleontological perspective. Mar. Pet. Geol. 85, 117–135 (2017).
- Kamel, H., Eita, T. & Sarhan, M. Nile Delta hydrocarbon potentiality. In Proceedings of the 14th Egyptian General Petroleum Corporation Exploration and Production Conference, Cairo, Egypt, 485–503.
- Younes, M. A. A. Natural Gas Geochemistry in the Offshore Nile Delta, Egypt. Advances Petrochemicals, 27–40 (2015).
- Farouk, S., El-Din, A. A., Shennawy, E., Al-Kahtany, T., Hu, Y. & K. Integration of seismic and well log data for structural evaluation of the West Delta Deep Marine, Egypt: implications for petrophysical analysis of the Messinian Abu Madi Reservoir. Mar. Geophys. Res. 46, 16 (2025).
- Aal, A. A. et al. Tectonic evolution of the Eastern Mediterranean Basin and its significance for hydrocarbon prospectivity in the ultradeepwater of the Nile Delta. Lead. Edge 19, 1086–1102 (2000).
- Abd-Allah, A. M., Aal, M. H. A. & Ghandour, A. Structural characteristics and tectonic evolution of the northwestern margin of the Nile Delta, Egypt. J. Afr. Earth Sc. 68, 82–95 (2012).
- Barber, P. M. Messinian subaerial erosion of the proto-Nile Delta. Mar. Geol. 44, 253–272 (1981).
- Pigott, J. D. & Abdel-Fattah, M. I. Seismic stratigraphy of the Messinian Nile Delta coastal plain: Recognition of the fluvial regressive systems tract and its potential for hydrocarbon exploration. J. Afr. Earth Sci. 95, 9–21 (2014).
- Khairy, A. et al. Geochemical assessment and hydrocarbon potential of Oligocene–Pliocene source rocks from northeast onshore Nile Delta, Egypt. Int. J. Coal Geol. 274, 104283 (2023).
- Vandré, C., Cramer, B., Gerling, P. & Winsemann, J. Natural gas formation in the western Nile delta (Eastern Mediterranean): Thermogenic versus microbial. Org. Geochem. 38, 523–539 (2007).
- El-Mowafy, H. Z., Ibrahim, M. & Dunlap, D. B. Unlocking gas reserves in bypassed stratigraphic traps in a deepwater brownfield using prestack seismic inversion: A case study from offshore Nile Delta, Egypt. Lead. Edge 37, 502–509 (2018).
- Eleslambouly, A., Zeynalli, M., Moncada, A., Fathy, A. & Rouxel, S. Advancing carbon storage in offshore Louisiana: Evaluation and modeling potential of two major depleted reservoirs in Vermilion_014 Field. In Carbon Capture, Utilization, and Storage Conference and Exhibition, 198–217.
- Szabó, N. P. & Dobróka, M. Extending the application of a shale volume estimation formula derived from factor analysis of wireline logging data. Math. Geosci. 45, 837–850 (2013).
- Larionov, V. Radiometry of Boreholes 127 (Nedra, 1969).
- Schlumberger. Log Interpretation Principles/Applications (Schlumberger Educational Services, 1991).
- Tiab, D. & Donaldson, E. C. Petrophysics: Theory and Practice of Measuring Reservoir Rock and Fluid Transport Properties (Elsevier, 2024).
- Eleslambouly, A., Khalifa, T., Aldhanhani, O., Zeynalli, M. & Abdelmaksoud, A. Integrated field characterization and static hydrocarbon reserve estimations of the Penobscot Field, Nova Scotia, Canada. Energy Geoscience, 100448 (2025). https://doi.org/10.1016/j.engeos.2025.100448
- Yahia, L., Toni, M., Metwalli, F. I., Mansour, M. & Ismail, A. Integration of 3D Structural Modelling and Well Logging Analysis in the Evaluation of October Oil Field, Gulf of Suez, Egypt. Trends Adv. Sci. Technol. 1, 5 (2024).
- Eleslambouly, A. K., Zeynalli, M., Al-Shalabi, E. W. & Alsuwaidi, M. Probabilistic approach for static carbon storage capacity estimation: A case study on the VR014 depleted gas field in offshore Louisiana, USA. Geoenergy Sci. Eng. 260, 214397 (2026). https://doi.org/10.1016/j.geoen.2026.214397
- Chopra, S. & Marfurt, K. J. Seismic attributes: A historical perspective. Geophysics 70, 3SO–28SO (2005). https://doi.org/10.1190/1.2098670
- Eleslambouly, A., Aldhanhani, O., Fathy, A., Zeynalli, M. & Alsuwaidi, M. Geothermal Plays Feasibility and Potential in a Transform Basin: A Case Study from Southern Vienna Basin. In Abu Dhabi International Petroleum Exhibition and Conference, D031S100R003 (SPE).
- Chopra, S. & Marfurt, K. J. Seismic Attributes for Prospect Identification and Reservoir Characterization (Society of Exploration Geophysicists and European Association of…, 2007).
- Chopra, S. & Marfurt, K. J. Spectral decomposition and spectral balancing of seismic data. Lead. Edge 35, 176–179 (2016).
- West, B. P., May, S. R., Eastwood, J. E. & Rossen, C. Interactive seismic facies classification using textural attributes and neural networks. Lead. Edge 21, 1042–1049 (2002).
- Ismail, A., Ewida, H. F., Al-Ibiary, M. G. & Zollo, A. Integrated prediction of deep-water gas channels using seismic coloured inversion and spectral decomposition attribute, west offshore, Nile Delta, Egypt. NRIAG J. Astron. Geophys. 9, 459–470 (2020).
- Torrado, L., Mann, P. & Bhattacharya, J. Application of seismic attributes and spectral decomposition for reservoir characterization of a complex fluvial system: Case study of the Carbonera Formation, Llanos foreland basin, Colombia. Geophysics 79, B221–B230 (2014).
- Chopra, S. & Marfurt, K. Seismic attributes: a promising aid for geologic prediction. CSEG Recorder 31, 110–120 (2006).
- Oumarou, S. et al. Seismic attributes in reservoir characterization: an overview. Arab. J. Geosci. 14, 1–15 (2021). https://doi.org/10.1007/s12517-021-06626-1
- Xu, Z. et al. Software defect prediction based on kernel PCA and weighted extreme learning machine. Inf. Softw. Technol. 106, 182–200 (2019). https://doi.org/10.1016/j.infsof.2018.10.004
- Zhang, Y. et al. Unsupervised seismic facies analysis technology based on SOM and PSO. Chin. J. Geophys. 58, 3412–3423 (2015). https://doi.org/10.6038/cjg20150933
- Kohonen, T. Self-Organizing Maps Vol. 30 (Springer Science & Business Media, 2012).
- Kohonen, T. Essentials of the self-organizing map. Neural Netw. 37, 52–65 (2013).
- Coléou, T., Poupon, M. & Azbel, K. Unsupervised seismic facies classification: A review and comparison of techniques and implementation. Lead. Edge 22, 942–953 (2003). https://doi.org/10.1190/1.1623635
- de Matos, M. C., Osorio, P. L. & Johann, P. R. Unsupervised seismic facies analysis using wavelet transform and self-organizing maps. Geophysics 72, P9–P21 (2007).
- Zhao, T., Jayaram, V., Roy, A. & Marfurt, K. J. A comparison of classification techniques for seismic facies recognition. Interpretation 3, SAE29–SAE58 (2015). https://doi.org/10.1190/INT-2015-0044.1
- Roden, R. & Chen, C. W. Interpretation of DHI characteristics with machine learning. First Break 35(5) (2017). https://doi.org/10.3997/1365-2397.35.5.88069
- Assenza, A., Valle, M. & Verleysen, M. A comparative study of various probability density estimation methods for data analysis. Int. J. Comput. Intell. Syst. 1, 188–201 (2008). https://doi.org/10.2991/ijcis.2008.1.2.9
- Davis, J. Justification of probability density function for resolution in statistical models of overlap. Chromatographia 44, 81–90 (1997). https://doi.org/10.1007/BF02466520
- Montanari, A. & Calò, D. G. Model-based clustering of probability density functions. Adv. Data Anal. Classif. 7, 301–319 (2013). https://doi.org/10.1007/s11634-013-0140-8
- Masalmeh, S. K., Shiekah, I. A. & Jing, X. Improved characterization and modeling of capillary transition zones in carbonate reservoirs. SPE Reservoir Eval. Eng. 10, 191–204 (2007).
- Abreu, V., Sullivan, M., Pirmez, C. & Mohrig, D. Lateral accretion packages (LAPs): an important reservoir element in deep water sinuous channels. Mar. Pet. Geol. 20, 631–648 (2003). https://doi.org/10.1016/j.marpetgeo.2003.08.003
- Hulea, I. N. Perched water contacts: Understanding fundamental controls. Petrophysics 60, 438–449 (2019).



















