Reconciling Reality: Building OGMP 2.0 Level 5 Inventories from Real-World Methane Measurements

At the American Geophysical Union Annual Meeting in Dec 2025, Arolytics presented a statistically rigorous framework for OGMP 2.0 Level 5 inventories that reconciles disparate methane measurements by combining multi-technology data with probabilistic inference: producing transparent, auditable, and defensible measurement-informed emissions inventories (MII).

The Industry Challenge: Measurement Without Reconciliation

Oil and gas operators are expanding methane detection technology portfolios (satellites, aircraft, trucks, CEMS, optical gas imaging, and more) to meet regulatory expectations and voluntary frameworks like Oil and Gas Methane Partnership 2.0 (OGMP 2.0). However when multiple measurement programs are deployed, how can you derive a single, measurement-informed inventory?

Different technologies:

  • Observe different parts of the emissions distribution
  • Operate at different times and frequencies
  • Have different performance metrics such as detection limits and probabilities of detection

The result is a fragmented picture of methane emissions. Adding more data can increase disagreement, not reduce it. For OGMP 2.0 Level 5, where inventories must be measurement-informed, this creates a challenge:

How do you reconcile incomplete and conflicting measurements into a single, defensible inventory?

Measurement-Informed Emissions Inventories (MII): The Core Concept

OGMP 2.0 Level 5 is built on the principle that inventories should be informed by direct measurement, not solely by emission factors or engineering estimates. However, measurement-informed does not mean measurement-only. No single dataset represents emissions reality. Every dataset is partial evidence of a shared, latent emissions truth.

Some emissions are routinely detected. Others, such as intermittent events or operational activities, remain largely invisible to surveys. Reconciliation is therefore not about choosing which dataset is “correct.” It is about combining imperfect evidence in a statistically consistent way.

A Statistical Framework for OGMP 2.0 Level 5 Reconciliation

During the AGU presentation, Jeff Neubeker outlined a framework developed by Arolytics in collaboration with the FluxLab at St. Francis Xavier University, working closely with Dr. David Risk and postdoctoral researchers Martin Lavoie and Nadia Tarakki. Key framework steps are described below.

Step 1: Annualizing Measurement Data

Field measurements are episodic snapshots. To be useful for inventories, they must be translated into annualized emissions estimates. Each measurement program is scaled based on:

  • Spatial coverage
  • Survey frequency
  • Detection thresholds
  • Technology-specific performance characteristics

The output is not a single number, but an emissions distribution with uncertainty for each technology.

Step 2: Probabilistic Reconciliation of Multi-Technology Data

Measurement technologies do not miss emissions randomly, but systematically. The Arolytics approach explicitly models:

  • Probability of detection
  • Technology-specific biases
  • Temporal and spatial sampling limitations

Using a Bayesian inference framework, multiple datasets are reconciled into a single posterior distribution representing consolidated measured emissions. This distribution reflects:

  • What was observed
  • What was likely missed

The result is a transparent, auditable reconciliation process aligned with OGMP 2.0 Level 5 expectations.

Step 3: Integrating Invisible Emissions

Even with multiple surveys, some methane sources remain largely undetected. Examples include:

  • Intermittent operational events
  • Recurring activities that rarely coincide with surveys

Rather than ignoring these sources, the framework integrates them probabilistically using:

  • Engineering estimates
  • Emission factors
  • Constraints informed by reconciled measurements

These sources are not layered on independently. They are constrained by what the measurements tell us. Bias is reduced, and uncertainty is more realistic.

The Outcome?

The final output of this framework is a site-level, OGMP 2.0 Level 5-consistent inventory that includes:

  • Explicit uncertainty propagation
  • Documented assumptions
  • Traceable logic
  • Year-over-year repeatability

Lessons Learned from Real-World Implementation

One of the most important messages from the presentation was pragmatic: statistics alone are not enough. Data quality matters more than model sophistication. Successful reconciliation depends on:

  • Clean, structured inventory metadata
  • Strong understanding of site infrastructure
  • Clear documentation of assumptions

Reconciliation is not a one-time calculation. It is a learning system.

Each year of data improves the inventory and measurement program design and assumptions.

From Inventory to Action: Why Reconciliation Matters

A reconciled inventory becomes a feedback mechanism that identifies which sources drive emissions for targeted mitigation, reveals where measurement programs should evolve, and connects emissions back to operational signals.

In this way, reconciliation closes the loop between:
measurement → inventory → action

At Arolytics, we view measurement and reconciliation as pathways to methane reduction and methane abatement.

Key Takeaways

Jeff Neubeker’s AGU presentation described a repeatable, transparent, and defensible code-based process to build measurement-informed inventories. Measurement-informed does not mean measurement-only. It means statistically reconciling reality—acknowledging uncertainty, integrating imperfect evidence, and building inventories that can be defended, improved, and acted upon.

Are you an oil and gas company currently building measurement-informed inventories, or working on reconciliation processes aligned with OGMP 2.0? We would love to hear from you! Email us at info@arolytics.com.