Building the Integrity Infrastructure for mCDR Modeling

Sherry Lippiatt

In order to achieve broad support and get to scale, climate interventions like marine carbon dioxide removal (mCDR) must meet a high bar for integrity, transparency, and safety. This means that carbon credit market actors, regulators, and the broader scientific community all need to have access to and confidence in the data underpinning carbon removal claims. 

In other regulated industries, this is accomplished by what is called an audit trail: a transparent and interpretable record of what was done, when, and by whom. Consider the financial records of a publicly traded company. Every number ties back to an auditable process or “paper” trail.

When we apply that same concept to mCDR, especially to the ocean modeling that underpins a carbon removal claim, we encounter a number of challenges. Namely, ocean models:

• require input data and forcing files from multiple sources,

• contain myriad configuration options and parameters that affect outcomes,

• generate massive, complex output datasets, and

• need to be run on high performance computing platforms. 

Further, the requirements for what must be tracked and retained are still evolving as the mCDR market matures. Third party verifiers (also called Validation and Verification Bodies, or VVBs) are typically hired by carbon credit registries (entities that issue, track, and manage carbon credits) to audit carbon removal claims. During an audit, a VVB assesses the integrity, quality, and accuracy of a carbon removal claim against the registry’s protocol. Today, mCDR verifiers must have deep expertise in ocean modeling in order to understand the approach that was taken and the assumptions that were made. And it’s virtually impossible for a less experienced audience to evaluate whether an ocean model was configured appropriately for a given mCDR deployment, or even identify what decisions were made in the process.

C-Star and the Blueprint Concept

At [C]Worthy, we are building tools to streamline and standardize the process of configuring, running, managing, and sharing the results of an ocean model simulation. C-Star, our flagship software now in development, is built on open-source ocean models, and is designed with data management, reproducibility, and auditability at its core. C-Star is designed to:

• encapsulate and share ocean model configurations, pointers to associated input datasets, and information about the way the model was run;

• orchestrate the running of a model simulation; and

• output metrics and data packages formatted to align with registry protocols (e.g., Isometric).

C-Star will bring consistency and transparency to ocean modeling, leading to more verifiable carbon removal claims. These are requirements of a high-integrity carbon removal market based on the best available science. 

One essential element of the C-Star system is the Blueprint, a lightweight framework to encapsulate everything required to run or reproduce an ocean model simulation. Blueprints are structured text files that contain information about:

• regional model domains (e.g. the boundaries of the area of the ocean the model will simulate and the grid used for numerical simulation);

• links to forcing files and initial conditions;

• ocean model parameters and configuration options; and

• links to the source code and repository.

The relationship between C-Star and Blueprints can be compared to that of a record player and a record. C-Star orchestrates the running of an ocean model simulation based on the information contained in a Blueprint. That information is in a structured format to facilitate the review of model setup parameters or reproduction of model results during an audit.

C-Star Blueprints can be created by anyone, shared by multiple users, and hosted in an open-source Blueprint Registry. This allows for independent replication of simulations, which is a key aspect of transparency. The C-Star Blueprint Registry will contain:

• a curated, open-source repository of C-Star Blueprints;

• searchable metadata;

• documentation of model skill and scientific validation against observations; and

• instructions for users to access a Blueprint in order to replicate the associated model experiment.

With these features, C-Star Blueprints and the Blueprint Registry can serve as critical elements of the mCDR crediting audit trail. They will translate the complexity of the ocean models that are used to quantify carbon removal in an mCDR deployment into a standard format more suitable for inspection and interpretation by regulators, verifiers, and buyers. Further, the Blueprint infrastructure offers a pathway for open collaboration with the research community, promoting the sharing and iterative improvement of model domains and configurations. 

Marine CDR will only succeed if the science behind every credit is transparent, reproducible, and trusted. To that end, C-Star’s integrity infrastructure will lay important groundwork for a trustworthy, high-quality carbon removal market that stands up to market, regulatory, and scientific scrutiny. To learn more:

• Follow us on LinkedIn or sign up for the [C]Worthy newsletter via the link below for updates as these and other C-Star features are launched. 

• Read more about Blueprints on the C-Star Github.