Petroleum Geoscience - Volume 31, Issue 1, 2025

Image-based deep learning methods, especially convolutional neural networks (CNNs), are gaining traction in seismic interpretation, but their application still demands manual validation. This study compares a U-Net structured CNN, called Fault-Net, with conventional edge-enhancing seismic attributes of variance and chaos that serve as a scientific baseline. We adopted two seismic interpretation workflows: (1) conventional attributes to enhance fault features; and (2) a deep learning-based workflow for fault segmentation using CNNs. Both workflows were applied to a high-resolution 3D seismic dataset from the structurally complex deep-water Orange Basin (offshore South Africa). While deep learning-based software packages are commercially available, it is unclear whether they are suitable for the Orange Basin and for use in an academic setting due to their proprietary architectures and generally closed training data. This study provides public evidence of the feasibility of automated structural interpretation in complex seismic datasets using deep learning, revealing both key benefits and limitations. Where high-quality labelled data are available, the deep learning approach is faster and tends to produce a cleaner and more accurate depiction of larger faults compared to conventional methods. The open availability of Fault-Net makes deep learning-based interpretation particularly advantageous for academic settings, offering significant time and resource efficiency while enhancing the understanding of complex subsurface structures.

Carbonate ramp systems present significant seismic interpretation challenges due to their pronounced facies heterogeneity, which frequently results in chaotic seismic outputs that obscure the underlying geological structures. The Porto Badisco Calcarenite in Salento, southern Italy, an Oligocene carbonate ramp, serves as the case study for this research, offering an analogue for understanding similar geological systems. By integrating fieldwork, laboratory analysis and MATLAB modelling, this study pioneers the use of detailed petrophysical data to construct innovative velocity models based on the velocity ranges of the different lithofacies analysed. These models distinctly illustrate the impact of facies heterogeneity on seismic velocities, providing fresh insights into acoustic impedance and variable propagation velocities across different facies constituting the carbonate ramp. Through advanced high-resolution synthetic seismic modelling conducted on carefully fine-tuned unmigrated stack sections, the research demonstrates how variations in petrophysical characteristics within measured ranges reflecting carbonate textures can dramatically alter seismic imaging. The innovative models, based on propagation velocity ranges, not only deepen the understanding of the seismic representation of lithofacies but also act as a potent tool for probing the subsurface architecture of complex carbonate systems, providing an interpretative key for the analysis of seismic images. This approach signifies a substantial advancement in seismic modelling that is aimed at refining interpretations and enhancing exploration strategies in carbonate ramp environments globally.

The geochemical data of the gas sourced from the Majiagou Formation indicate low organic content and high thermal maturity, which make the evaluation of the hydrocarbon generative potential difficult. However, the low organic abundance and high thermal maturity of the Majiagou Formation make it difficult to evaluate its hydrocarbon generation potential. The study of the genesis of pyrite can provide a better understanding of a source-rock depositional environment and its evolution. In this paper, we report on a significant amount of pyrite linked to the hydrocarbon source in the Majiagou Formation. Light and electron microscopy were used to observe the morphology of the pyrite, while sulfur, carbon and oxygen isotopes were used to investigate the geochemical characteristics of pyrite and the surrounding rock matrix. The results showed that macroscopic stellate pyrite filled cracks and is found in veins or along the laminar surface, as well as agglomerated pyrite; microscopically, pyrite is dominated by autotypic pyrite in the form of cubes, pentagonal dodecahedrons and columns. The 34SCDT ranged from 4.6 to 27.5‰, 13CPDB values ranged from −3.4 to −2.3‰ and 18OPDB values ranged from −10.5 to −7.3‰. Our findings, combined with the geological background and thermal history of the region, suggest that the pyrite in the Majiagou Formation was formed by thermochemical sulfate reduction, which indicates that large-scale hydrocarbon generation events have occurred. The rocks of the Majiagou Formation had the ability to generate hydrocarbons. Our research can be utilized as supporting material for the exploration and development of natural gas fields in the Ordos Basin.

The Aptian Dariyan (Shuaiba) Formation, a major Cretaceous reservoir in the Middle East, remains poorly understood regarding the influence of depositional facies and diagenetic processes on reservoir quality. This research addresses the gap through an integrated analysis of facies, petrophysics and geochemistry on a continuous, 104.5 m-long core from the Salman oil/gas field in the eastern Persian Gulf. Employing fully automated techniques, we identified hydraulic flow units (HFUs). We classified nine carbonate facies into three distinct facies associations, arranged from shallowest to deepest: inner ramp (lagoon and shoals), shallow open-marine mid-ramp and deep open-marine (outer ramp and intrashelf basin). These facies associations exhibit a stacking pattern delineating five third-order transgressive–regressive sequences. The identified HFUs include the barrier unit (HFU1), the baffle unit (HFU2) and the normal unit (HFU3), assessed based on lithological and petrophysical attributes. The normal unit, characterized by good storage capacity but poor to moderate flow capacity, highlights the complexity of reservoir quality. The Dariyan Formation is predominantly composed of mud-supported textures formed in warm, tropical waters. Additionally, late diagenetic cementation severely obstructed pore spaces, altered primary rock characteristics and reduced effective flow capacity.

The oil and gas industry faces challenges in estimating reservoir properties in regions without well data. To overcome this, a combination of seismic attributes and drilled well information is used to predict the unknown drilled portion of the reservoir. Integrating these datasets enhances geological and flow models, leading to better reservoir predictability and improved strategies for reservoir development and production. This study focuses on a lacustrine carbonate environment and employs a stratigraphic-sedimentological modelling approach. The aim is to understand sedimentation times and propose a growth model for carbonate build-up based on prior integrated studies using seismic attributes. The modelling methodology was applied to the BM-C-33 Exploratory Hydrocarbon Block in the Campos Basin, offshore Brazil. The Macabu Formation was chosen as the target unit. Seismic attribute-driven carbonate mound features were identified and categorized into different sedimentation periods. Geological processes were then modelled, incorporating parameters such as topographical and bathymetric surfaces, lake-level variations, subsidence mapping, and rates of carbonate deposition. Four depositional domains for lacustrine carbonates were considered: a high-energy sediment domain, a build-up domain, a low-energy sediment domain and a clayey sediment domain. By integrating seismic attributes and well data, more reliable growth models of the carbonate mounds were developed. The results demonstrate the efficiency of the methodology in improving the understanding and representation of carbonate reservoirs, facilitating the characterization of the studied region, and mitigating associated project risks such as identifying new drilling locations.