Advanced Geobody Interpretation Workflow in SubsurfaceAI

Seismic interpretation plays a crucial role in subsurface exploration by identifying geobodies—three-dimensional features derived from seismic attributes. These geobodies often correspond to key reservoir structures such as channelized deposits, carbonate buildups, and turbidite fans. Extracting, classifying, and analyzing these geobodies is essential for hydrocarbon exploration and reservoir characterization.

SubsurfaceAI offers a suite of advanced machine learning-driven workflows designed to enhance geobody identification, classification, and volumetric quantification. By leveraging seismic attributes, facies classification, and interactive AI-assisted interpretation, SubsurfaceAI enables geoscientists to achieve high-resolution geobody extraction and volumetric estimation, facilitating more accurate reservoir evaluations.

Geobody Interpretation Methods in SubsurfaceAI

1. Thresholding from Seismic Attributes

• One of the most fundamental methods for geobody extraction involves thresholding based on specific seismic attributes. This technique is particularly effective when geobodies exhibit distinct seismic signatures that differentiate them from surrounding formations.

Key Attributes for Thresholding:

• Impedance Values: Many geobodies, such as gas-charged sands and carbonate reservoirs, display distinct impedance contrasts, making impedance a valuable attribute for differentiation.
• Predicted Gamma Ray and Sonic Volumes: If well log data is available, AI-driven inversion techniques can predict gamma-ray and sonic responses across the seismic volume. These properties aid in delineating potential reservoir zones when applied with thresholding.
• Amplitude and Instantaneous Attributes: Attributes such as coherence, sweetness, and RMS amplitude highlight geobody structures effectively.

• By defining appropriate threshold values, geoscientists can isolate voxels representing potential geobodies, allowing for precise visualization and volumetric quantification.

2. Interactive Geobody Tracking on Seismic Sections

Geobody interpretation is a critical step in seismic analysis, influencing reservoir characterization, prospect evaluation, and subsurface modeling. Traditional methods often rely on static thresholds or manual delineation, which can be time-consuming and less adaptable to complex geological settings.

A more dynamic and interactive approach allows users to define seed points and adjust window size and threshold values to track geobodies across seismic sections in real time. This method enhances flexibility, enabling geoscientists to refine geobody boundaries interactively, ensuring higher accuracy in interpretation.

One of the key advantages of interactive tracking is real-time visualization. Tracked geobodies can be displayed in both 2D and 3D windows, allowing users to perform immediate quality control (QC) by comparing the results against well data. This iterative approach ensures that interpretations remain geologically consistent and data-driven.

By integrating interactive geobody tracking into seismic workflows, interpreters can achieve:

• Greater precision in geobody delineation
• Faster iteration with real-time parameter adjustments
• Improved validation using multi-dimensional visualization and well data cross-checking

As geophysical software continues to evolve, interactive tools like these are becoming essential for maximizing the accuracy and efficiency of seismic interpretation.

3. Converting from a Seismic Facies Volume

• SubsurfaceAI allows users to classify seismic attributes into facies volumes through both unsupervised and supervised machine learning techniques. These classified facies volumes can then be converted into geobody representations.

Seismic Facies Classification Techniques:

• Unsupervised Learning: Clustering algorithms such as K-means and self-organizing maps (SOM) group seismic data into clusters based on similar patterns, helping reveal geological features.
• Supervised Learning: Users can train machine learning models using labeled data (from well logs, core samples, or manual interpretations) to classify seismic facies based on multiple attributes.

Extracting Geobodies from Seismic Facies:

• Once seismic facies are classified, contiguous facies samples that exhibit geobody-like characteristics (such as continuous channel-like structures) can be extracted as discrete geobodies. This workflow is particularly valuable for:
• Identifying stratigraphic traps
• Mapping lateral facies variations
• Delineating depositional environments

4. AI-Assisted Interpretation with InterpAI

• InterpAI, a dedicated module within SubsurfaceAI, enhances geobody interpretation through interactive machine learning workflows. This AI-assisted approach significantly accelerates interpretation compared to traditional manual picking methods.

Workflow Steps:

• Step 1: Interactive Training on Seismic Sections and Stratal Slices
  Users manually label representative seismic sections or stratal slices that contain the targeted geobody features. These labeled sections serve as training data for the AI model.
• Step 2: AI-Based Pattern Recognition
  The AI model generalizes user-defined labels across the entire seismic volume, automatically detecting similar patterns and anomalies. It refines its predictions iteratively based on user feedback, ensuring improved geobody delineation.
• Step 3: 3D Extraction and Visualization
  Once geobodies are identified, they can be extracted into 3D objects for further analysis. The resulting 3D geobody models provide enhanced visualization and interpretation, streamlining the overall workflow.

Geobody Quantification: Estimating OOIP and OGIP

• Beyond detection and visualization, SubsurfaceAI incorporates volumetric computation tools to estimate key reservoir parameters, including:

• Original Oil in Place (OOIP)
• Original Gas in Place (OGIP)

• Using extracted geobodies, the software calculates:

• Volume Metrics: Determines bulk rock volume and net pay thickness.
• Porosity and Saturation Estimates: Integrates well log data or AI-predicted property volumes to compute effective porosity and fluid saturation.
• Hydrocarbon Volume Estimates: Utilizes industry-standard equations to derive OOIP and OGIP, enabling geoscientists to assess reservoir potential with greater accuracy.

SubsurfaceAI revolutionizes geobody interpretation by integrating advanced AI techniques into subsurface modeling workflows. Its diverse methodologies—including thresholding, seismic facies conversion, and AI-assisted interpretation—allow geoscientists to extract and quantify geobodies with unprecedented precision. By incorporating volumetric calculations, SubsurfaceAI enhances reservoir evaluation, making it an invaluable tool for modern geoscience applications.