Carbon credits come in two fundamentally different kinds — removals and avoidance. Knowing the difference is the single most useful skill a credit buyer can learn.
Every carbon credit represents the same nominal thing — one tonne of CO₂ equivalent — but credits get there by two profoundly different routes. An avoidance credit is issued when an emission that would have happened does not: a forest that was going to be cleared stays standing, a coal power plant runs less because a wind farm exists. A removal credit is issued when carbon dioxide is actively pulled out of the atmosphere and stored: a degraded mangrove is restored and builds new carbon in its biomass and soils, or a machine captures CO₂ and injects it underground.
The distinction sounds technical. It is, in fact, the single most consequential question a buyer can ask, because the two credit types do different jobs, carry different risks, and belong in different parts of any serious climate strategy. This article explains both, where each fails, and what the difference means specifically for ocean and coastal projects.
An avoidance credit certifies a counterfactual: emissions would have been higher without this project. The classic examples are avoided deforestation (REDD+ projects), renewable energy that displaces fossil generation, and methane capture at landfills. The atmosphere genuinely benefits when the counterfactual holds — a forest left standing is a real service to the climate.
The weakness lives in that word "counterfactual". You are paying for something not to happen, which means the entire credit rests on a prediction about a world that did not occur. If the deforestation was unlikely anyway — the land was already protected, or clearing was uneconomic — the credit represents little or no real climate benefit. This is the additionality question in its sharpest form, and years of investigative reporting and academic analysis have found that a material share of the voluntary market's avoidance credits, particularly some avoided-deforestation projects, overstated their baselines. Overstated baselines mean overcrediting: more tonnes sold than tonnes actually saved from the atmosphere.
None of this makes avoidance credits worthless. Preventing the loss of a carbon-rich old-growth forest or a mature mangrove is far cheaper and faster than regrowing one, and the biodiversity protection is immediate and real. But it does mean avoidance credits deserve the hardest scrutiny of baseline claims, and that buyers should think of them as preventing harm rather than cleaning up their own emissions.
A removal credit certifies something physical: CO₂ came out of the atmosphere and went somewhere durable. Removals divide by durability into loose tiers:
Removal credits answer the question that net-zero commitments ultimately need answered: given that some emissions will remain even in a deeply decarbonised economy, how do we deal with them? You cannot neutralise a tonne of emitted CO₂ by not cutting down a tree somewhere else, in the strict physical sense. You neutralise it by taking a tonne back out. That is why standards bodies writing net-zero guidance increasingly insist that avoidance credits can support "beyond value chain" contributions, but only removals can count toward neutralising residual emissions.
Nature-based removals reintroduce a problem avoidance credits do not have to the same degree: the carbon can come back. A restored mangrove can be destroyed by a cyclone; a replanted forest can burn. Removal credits from ecosystems therefore need permanence mechanisms and buffer pools — insurance structures where a share of credits is set aside against future losses, with monitoring obligations over decades. Well-designed ecosystems also build durability through diversity: a restored kelp forest with mixed age classes, local genetics and healthy herbivore communities is more resilient to marine heatwaves than a monoculture, but no living system can promise geological permanence, and honest projects say so.
The market, for all its faults, understands the difference. Avoidance credits trade at low prices; high-quality nature-based removal credits command several times more; durable technological removals trade at an order of magnitude above that — rough, shifting bands rather than fixed figures. A buyer seeing very cheap credits should hear the question the price is asking: what kind of tonne is this, really? Price is not proof of quality, but persistent cheapness usually reflects a counterfactual-tonne supply that is far more abundant than genuine additional removals.
Blue carbon projects can be either type, and the difference matters for how buyers should use them:
There is also a middle category worth naming: emission reductions within your own value chain — a fishery cutting fuel use, a farm reducing fertiliser. These are not credits at all but internal abatement, and they should come first in any strategy. Offsets, of either species, are for what remains.
The full due-diligence checklist — registries, vintages, methodologies, safeguards — is in our blue carbon buyer's guide, and the broader integrity framework in what nature-based solutions integrity actually requires.
Not morally, and not always practically. Protecting a forest about to be cleared can deliver more climate benefit per dollar than planting one. "Better" depends on the claim being made: for neutralising emitted CO₂, removals are the physically correct instrument; for maximising near-term climate and biodiversity benefit per dollar, well-vetted avoidance can be excellent value.
Because they are harder. Growing new carbon stocks or operating capture machinery takes years of measured, monitored work, whereas an avoidance credit requires demonstrating a plausible threat and changing a plan. Scarcity also matters: genuine, verified removals are a small fraction of credit supply.
Under most credible net-zero standards, avoidance credits can complement a net-zero strategy but cannot neutralise residual emissions — that role belongs to removals, with durability matched to the lifetime of the emitted CO₂. Check which standard your target is set under.
Removal: trees physically draw CO₂ from the air as they grow. But it is a reversal-prone removal, which is why permanence monitoring and buffer pools exist, and why planting is not a licence to keep emitting.
Kelp projects are removal-type, but the creditable carbon is the verified share of kelp carbon exported to durable sinks — deep sediments and the deep ocean — not the standing forest itself, which turns over seasonally. Conservative, pathway-based accounting is essential, and buyers should ask projects which specific pathways their credits rest on.
Additionality is the make-or-break test of every carbon credit: did the credit cause the climate benefit? How the counterfactual is tested — and faked.
Priya Raman ·
Blue carbon credits can fund real coastal restoration — or expensive stories. A practical due-diligence checklist for corporate buyers before they sign.
Priya Raman ·
Nature-based climate solutions can fund real restoration or sell attractive stories. The integrity checklist that separates credible projects from greenwash.
Priya Raman ·