EGU24-17893, updated on 11 Mar 2024
EGU General Assembly 2024
© Author(s) 2024. This work is distributed under
the Creative Commons Attribution 4.0 License.

Reservoir and seal characterization of deep marine sediments using seismic facies analysis with machine learning techniques

Tural Feyzullayev1, David Lubo-Robles2, Beatriz Benjumea1, Heather Bedle2, Estefanía Llave1, Francisco Javier Hernández-Molina3,4, and Zhi Lin Ng5
Tural Feyzullayev et al.
  • 1Instituto Geológico y Minero de España (IGME-CSIC) Madrid, Spain (
  • 2School of Geosciences, the University of Oklahoma, Norman, OK, USA
  • 3Instituto Andaluz de Ciencias de la Tierra (IACT), CSIC-UGR, Granada, Spain
  • 4Dept. Earth Sciences, Royal Holloway Univ. London, Surrey, UK
  • 5School of Marine Science, Sun Yat-sen University, Zhuhai, China

This work describes key aspects of the methodology for subsurface characterization of Late Miocene deep marine sedimentary systems of the Gulf of Cádiz. In particular, we focus on the products of alongslope bottom currents processes, known as contourite systems, and mixed deposits developed by the interaction between contourite and downslope turbidite systems. Both these systems offer prospects for CO2 storage for their high reservoir potentials. In addition, hemipelagic sediments and fine-grained contourites present in the area could act as seals. The objective of this study consists of using seismic attributes and machine learning techniques for conducting a seismic facies analysis to distinguish between various Late Miocene deep marine deposits in a 3D seismic volume. The first step is to restrict the dataset to the deposits of interest in order to avoid irrelevant sediments or structures such as the allochthonous unit of the Gulf of Cádiz or salt domes or diapirs. This adjusts the dynamic range of the clustering to focus on our targets. The second step is the testing of the seismic attributes to improve their selection criteria, in order to maximize the differences between the distinct seismic facies. Finally, we apply an unsupervised clustering algorithm for the selected seismic attributes to perform an automatic seismic facies analysis that facilitates both reservoir and seal imaging. This study will ultimately help to assess the socio-economic impact of Late Miocene sediments developed by bottom currents on climate change mitigation and energy transition. This research and the Grant PRE2022-102745 were funded by MCIN/ AEI/10.13039/501100011033 and they are linked to the ALGEMAR project (PID2021-123825OB-I00). This work is partly supported by SEASTORAGE project (TED2021-129816B-I00), funded by MCIN/ AEI/10.13039/501100011033/PRTR-C21 and by the European Union NextGenerationEU.

How to cite: Feyzullayev, T., Lubo-Robles, D., Benjumea, B., Bedle, H., Llave, E., Javier Hernández-Molina, F., and Lin Ng, Z.: Reservoir and seal characterization of deep marine sediments using seismic facies analysis with machine learning techniques, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-17893,, 2024.