A registry is the ledger. A standard is the rulebook. For carbon dioxide removal (CDR), they are what turns “we removed a tonne of CO₂” into a serialized credit that someone can buy, count, and retire. A standard defines what counts as removal and how it must be measured. A methodology applies those rules to one specific process. Independent auditors check the project against the rules, and the registry issues the unit and tracks who owns it. When a buyer asks whether a credit is trustworthy, they are really asking whether each of these layers did its job. When a lender asks whether a project is bankable, much of the answer depends on whether a recognized standard has a methodology that fits the project’s process.

The mechanics

Standards and registries. In CDR these are often the same organization, but they do different jobs. Puro.earth runs the Puro Standard and issues CO₂ Removal Certificates. Isometric publishes removal-only protocols and runs its own registry. Verra and Gold Standard come from the broader voluntary carbon market, where avoidance and reduction credits dominate. Each has added removal-specific methodologies on top of that base.

Methodologies. A methodology is where the substance lives. It specifies:

  • the system boundary
  • the baseline
  • which emissions must be subtracted (energy, transport, feedstock, embodied materials)
  • how durability is demonstrated
  • what data must be collected, and how often

The accounting choices matter a great deal. A critical review of life cycle assessment for CDR technologies found that boundary and allocation decisions can shift reported net removal substantially. Those decisions are exactly what a methodology fixes in place.

Why one pathway can have several methodologies. Standards weigh the underlying trade-offs differently:

  • how long carbon must stay stored
  • how much measured data is required versus modeled data
  • how uncertainty is discounted

Two biochar credits from different standards are therefore not automatically equivalent. The same applies to two direct air capture (DAC) credits.

Validation and verification. These are separate steps, usually done by accredited third-party validation and verification bodies (VVBs):

  • Validation happens before or early in operation. It checks that the project design matches the methodology.
  • Verification happens after the fact. It checks the monitoring data and confirms how many tonnes were actually delivered in a reporting period.

The standard owns the rules. The auditor applies them. Keeping these roles apart is the main defense against a developer effectively certifying itself.

Issuance and retirement. These are different events:

  • Issuance creates serialized credits on the registry after verification.
  • Retirement permanently cancels a credit on behalf of a named beneficiary. Only retirement supports a claim.

Credits can sit issued but unretired, or be transferred several times before retirement. Some buyers also sign offtake agreements for tonnes that have not been issued yet. That is a forward contract, not a credit.

Market trackers. Tools such as CDR.fyi sit alongside the registries but are not registries. They compile publicly reported purchases and deliveries across developers and standards, which makes the gap between contracted and delivered tonnes visible. They do not issue credits or adjudicate quality.

How it plays out across pathways

Biochar. Biochar is the most mature pathway for methodologies, and the one where methodology differences are easiest to see. The key question is persistence: how much of the carbon in the char stays stable over century-plus timescales. The literature on biochar in climate change mitigation supports long-term stability for much of the carbon, but decay modeling assumptions still drive how many credits a tonne of char earns. Biochar producers show how feedstock and technology change what a methodology must handle:

  • Dark Earth Carbon processes roughly 8,000 tonnes of wood waste in a rotary kiln in Tanzania.
  • Alcom Carbon Markets runs rice-husk gasifier/biochar facilities in the Philippines, where energy co-production raises allocation questions.
  • Varaha operates across South Asia and parts of Africa, where distributed production puts pressure on monitoring requirements.

Enhanced weathering. Here verification is the bottleneck, not the production event. Carbon removal happens slowly in soils and waterways and has to be inferred from sampling and models. Metalplant pairs olivine weathering with nickel phytomining in Albania. That adds a co-product which a methodology must decide how to treat.

Direct air capture. Measuring captured CO₂ is comparatively straightforward. The contested inputs are the energy source and the integrity of storage. Examples include:

  • Climeworks, which runs Orca (4,000 t/yr) and Mammoth (36,000 t/yr) in Iceland with geological storage.
  • Heirloom, which uses limestone-based sorbents.

For both, the credit depends on storage-site monitoring as much as on capture.

Bioenergy with carbon capture and storage (BECCS). BECCS adds biomass sourcing to the energy and storage questions above. Reverion captures carbon from biogas while generating electricity. Carbon America develops Class VI geological storage in the US. Methodologies must decide how the biomass carbon debt and the electricity output are allocated.

Ocean CDR. For ocean approaches such as Liquid Trees, the measurement science is still being built. The US National Academies’ research strategy for ocean-based CDR treats monitoring and verification as an open research priority.

Agricultural and soil carbon. Grow Indigo and Klim sit at the boundary where reversal risk and durability make “removal” labels contested.

The hard parts

Durability equivalence. The analysis of what net zero means argues that emissions and removals must be matched for like-for-like durability. Standards still differ on how many years of storage count as “permanent.” They also differ on whether shorter-lived removals should be discounted or kept in a separate category.

Who pays the auditor. VVBs are typically paid by the developers they audit. Separating the auditor from the rule-maker reduces this conflict but does not remove it.

Measurement versus modeling. Measurement-heavy protocols cost more per tonne. Model-heavy ones risk over-crediting. No standard has settled this trade-off for open-system pathways such as enhanced weathering or ocean CDR.

Fragmentation. With several standards per pathway, buyers must compare methodologies, not just registry brands. Trackers improve transparency, but they report what registries and developers disclose rather than assessing quality.

Liability for reversal. Buffer pools, replacement obligations, and insurance vary by standard. How they will perform under a large reversal event is largely untested.