Easy, Integrated Workflow for Interpretation Geophysicists – and How It Compares to Pre-Stack Gathers

Amplitude Variation with Offset (AVO) is a powerful way to connect seismic amplitude behavior to changes in rock and fluid. In many teams, though, AVO can feel “specialist-only” or complicated to run end-to-end.

SubsurfaceAI is designed to change that.

The AVO tools in SubsurfaceAI are built so that interpretation geophysicists—not just quantitative specialists—can:

• Run AVO on everyday seismic projects (2D or 3D, time or depth)
• Use partial angle or offset stacks, with up to 5 stacks
• Set simple angle ranges per stack in a guided dialog
• Get intercept (A) and gradient (B) volumes with just a few clicks
• Explore AVO behavior using interactive cross plots and density maps
• Define AVO classes, extract geobodies, and visualize them instantly in 2D, maps, and 3D
• Build fluid stack and lithology stack volumes directly from the cross plots
• Save AVO-based results as facies volumes/grids/surfaces
• Resample these facies and properties directly to the static reservoir grid

All of this is tightly integrated with the rest of SubsurfaceAI—RGT volumes, horizons, faults, wells, machine-learning tools—so AVO becomes just another part of your interpretation workflow, not a separate science project.

1. Inputs: What the AVO Tool Accepts (Without Making You a Specialist)

SubsurfaceAI’s AVO from partial stacks is designed around the data you actually get from processing, and the way an interpreter likes to work.

1.1 Time or Depth Domain

You can run AVO on:

• Time-domain seismic (e.g., post-stack time migrated volumes)
• Depth-domain seismic (e.g., depth migrated surveys)

You don’t have to do anything special—just select the appropriate volumes. SubsurfaceAI keeps track of whether you’re in time or depth and keeps everything consistent for later mapping and resampling.

1.2 Partial Angle or Partial Offset Stacks

The AVO tool accepts:

• Partial angle stacks (near / mid / far by incidence angle)
• Partial offset stacks (grouped by offset)

In a simple configuration panel, you:

• Choose each stack volume from a dropdown,
• Specify whether it’s angle- or offset-based,
• Enter the angle range for that stack (e.g., 0–10°, 10–20°, 20–30°).

No equations needed—the UI walks you through the required fields.

1.3 Up to Five Partial Stacks

You can use:

• 2 stacks (near / far)
• 3 stacks (near / mid / far)
• 4 or 5 stacks, if your processing provides them

SubsurfaceAI supports up to 5 partial stacks per AVO run. More stacks, within reasonable angle coverage, usually mean a more robust AVO estimate.

2. AVO Workflow in SubsurfaceAI – Designed for Interpreters

Here’s how an interpretation geophysicist would typically use the AVO workflow in SubsurfaceAI.

Step 1 – Basic QC and Setup

Before you open the AVO tool, you simply confirm:

• The partial stacks are co-registered and have consistent processing.
• You know (or can read from the processing report) the angle ranges for each stack.

Then you open the AVO from partial stacks function.

A guided dialog walks you through:

1. Picking the analysis object (2D line, 3D volume, strata-grid, or horizon).
2. Choosing the partial stack volumes from dropdown lists.
3. Entering the angle range for each stack in a simple table.
4. Selecting which attributes to compute (A, B, and optionally a few derived attributes).

You don’t need to script anything or leave the main interpretation environment.

Step 2 – Running the Calculation

When you click Run:

• SubsurfaceAI processes the data block-by-block behind the scenes.
• For each sample and each tile, it reads the stack amplitudes, uses the angle ranges you gave it, and calculates:
  • Intercept (A) – zero-offset reflectivity estimate
  • Gradient (B) – how amplitude changes from near to far angles

The code is optimized for large datasets (including OpenVDS), so you can run AVO on full 3D surveys. You can keep working elsewhere in SubsurfaceAI while it runs—no special cluster jobs needed for most projects.

Step 3 – Visual QC of A and B Volumes

Once computed, A and B show up in your project like any other seismic attributes:

• You can add them as overlays on your existing 2D sections.
• You can display them as horizon maps or time/depth slices.
• You can combine them with RGT slices, faults, or other attributes.

This makes QC very intuitive:

• Do A and B show coherent patterns where you expect reservoir or gas-charged zones?
• Is noise level acceptable, or do you see striping, edge artifacts, or problematic areas that point back to processing issues?

If needed, you can easily rerun AVO with slightly different stack selections or angle ranges—same tool, same dialog.

Step 4 – Cross Plotting AVO Attributes

Next, you move from raw AVO attributes to actual interpretation.

In the cross plot tool, you select:

• X-axis: Intercept (A)
• Y-axis: Gradient (B)

You then choose which samples to include:

• A full volume (for a global view),
• A specific time or depth window,
• One horizon or a set of strata layers,
• A small box around a well or prospect.

SubsurfaceAI plots all the selected samples as points in A–B space and offers options to:

• Color points by horizon, depth, or another attribute.
• Show a density map, where clusters stand out clearly.
• Overlay well-based facies (e.g., “brine sand,” “gas sand,” “shale”) if you have those defined.

You get an immediate visual sense of:

• Background shale trends,
• Possible brine and gas sand clusters,
• Outliers or noisy data.

This is all done visually—no scripting or external plotting tools required.

Step 5 – Defining AVO Classes and Geobodies

From the cross plot, you can turn clusters into geological concepts.

In the cross-plot window, you:

• Draw polygons around clusters that look like distinct AVO behaviors.
• Give each polygon a class label (e.g., “Class III-like gas sand candidate,” “background shale,” etc.).

SubsurfaceAI then:

• Assigns each seismic sample an AVO class based on where its (A, B) point falls.
• Extracts geobodies in 3D for each class.
• Shows those geobodies live in:
  • 2D sections (overlaid on amplitude)
  • Basemap (colored polygons or outlines)
  • 3D views (transparent or solid bodies)

The key point: when you edit the polygon on the cross plot, the geobodies update immediately. You can see in real time how tightening or loosening class boundaries changes the geological picture.

Because this all happens inside SubsurfaceAI, you can also:

• Cross-reference with fault frameworks and horizons,
• Compare to well logs, synthetic gathers, or RGT-based features,
• Use the same class definitions across multiple intervals or volumes.

Step 6 – Fluid Stack and Lithology Stack (Still Easy to Use)

SubsurfaceAI also lets you make fluid-sensitive and lithology-sensitive stacks from the A–B cross plot, without forcing you into heavy theory.

How it works in practice:

1. In the cross plot, you identify directions that look fluid-driven vs rock-driven, using wells or rock physics as a guide:
   • One axis (line) that separates brine sands from gas sands → fluid direction
   • One axis that follows the shale-to-sand trend → lithology direction
2. You set these directions with simple interactive tools:
   • Drag/rotate the axes until they line up with the trends you want.
   • SubsurfaceAI shows you how points project along each axis.
3. When you’re happy, you click to generate:
   • A fluid stack volume – a single attribute volume that “lights up” fluid-sensitive anomalies.
   • A lithology stack volume – a volume that responds mainly to lithology changes (e.g., sand vs shale) with reduced fluid influence.

These stacks appear in your project just like other attributes. You can:

• View them in sections and maps,
• Use them to seed geobodies,
• Combine them with other SubsurfaceAI attributes or ML workflows.

The underlying math is there, but you don’t have to see it—the UI keeps it conceptual and visual.

Step 7 – Facies and Property Volumes from AVO

At this point, you now have:

• AVO classes (from polygons on the cross plot),
• Fluid stack and lithology stack volumes.

SubsurfaceAI lets you turn these into:

• Facies volumes (3D), facies grids (for strata-based objects), or facies surfaces (for horizons) – all based on AVO classes.
• Continuous property volumes:
  • Fluid-stack-based property (fluid indicator)
  • Lithology-stack-based property (reservoir quality / sandiness indicator)

These outputs plug directly into other SubsurfaceAI functions and standard reservoir characterization workflows.

Step 8 – Resampling to the Static Reservoir Grid (End-to-End)

Because SubsurfaceAI is built to connect interpretation and modeling, it includes tools to:

• Resample AVO-based facies volumes from seismic grid → static reservoir grid.
• Resample fluid and lithology property volumes onto the same static grid.

You can choose upscaling strategies:

• Majority facies per grid cell (for facies volumes),
• Average, median, or custom aggregation (for property volumes).

SubsurfaceAI handles:

• Domain issues (time vs depth),
• Spatial alignment between seismic cube and reservoir grid.

The end result:

• Reservoir engineers receive grid-scale facies and property fields that are directly derived from your AVO interpretation, with no messy intermediate exports or manual transforms.

3. Partial Stacks vs Pre-Stack Gathers – For the Interpretation Geophysicist

SubsurfaceAI’s workflow focuses on partial stacks because they’re usually more practical for interpreters. Still, it’s useful to understand how this compares to doing AVO on full pre-stack gathers.

3.1 Partial Stacks – Why They Work So Well in Practice

Strengths:

• Fast and field-scale
  • Smaller data volumes mean you can run AVO across an entire 3D survey, not just a small area.
  • Computations fit nicely into everyday interpretation workflows.
• Cleaner signal-to-noise
  • Stacking within angle ranges improves S/N, often giving smoother AVO trends.
  • Less sensitive to individual noisy traces.
• Easy to interpret and integrate
  • A and B volumes fit naturally into the same SubsurfaceAI environment you use for structure, RGT, faults, and attributes.
  • Cross plots, class definitions, geobodies, fluid and lithology stacks are all in one interface.

Limitations:

• You see one average amplitude per angle range, not the full amplitude vs angle curve.
• You rely on processing quality for each partial stack—bad angle binning or amplitude balancing will carry through.
• Partial stacks are not ideal for detailed elastic inversion (Vp, Vs, density) or azimuthal AVO.

For most interpretation geophysicists, these trade-offs are acceptable—and often desirable—because partial stacks let them run AVO efficiently and use the results every day.

3.2 Pre-Stack Gathers – When You Still Need Them

Working on pre-stack gathers has its place, particularly when:

• You need detailed elastic inversion to Vp, Vs, and density.
• You want to do AVAz or anisotropic AVO.
• You have high-quality, well-conditioned gathers and the time/computing power to run more advanced workflows.

But for many projects:

• Gathers are noisy,
• Data volumes are huge,
• And gather-based AVO/inversion ends up limited to small zones or specialized studies, rather than being part of routine interpretation.

SubsurfaceAI’s partial-stack workflow is meant to fill that gap: give interpreters a practical AVO toolset that covers the entire field, and that easily feeds into reservoir models.

4. Summary – AVO as a First-Class Interpretation Tool in SubsurfaceAI

For interpretation geophysicists, SubsurfaceAI turns AVO into a natural extension of standard seismic interpretation:

• You work with the data you already have: partial angle or offset stacks, in time or depth, up to five stacks with simple angle range inputs.
• You run AVO through a guided, easy-to-use interface, no scripting required.
• You see intercept and gradient volumes side-by-side with your amplitudes, horizons, and faults.
• You use interactive cross plots and density maps to define AVO classes visually.
• You see AVO geobodies update in real time in 2D, map, and 3D views.
• You create fluid and lithology stacks directly from the cross plot with intuitive axis tools.
• You convert AVO results into facies and property volumes, and
• You resample them seamlessly to the static reservoir grid, ready for geomodeling and simulation.

The key theme is ease-of-use and integration. You don’t have to be an AVO or inversion specialist. SubsurfaceAI lets you bring AVO into your everyday interpretation—closing the loop from seismic amplitude to reservoir-scale facies and property models in one consistent environment.