Burying Carbon at Sea: Is mCDR a Real Solution, or a New Kind of Hope-Peddling?

For the past few years, a concept has been drifting from the margins of climate discussions toward the center: **marine Carbon Dioxide Removal**, or **mCDR**. Investment is growing fast; "ocean carbon credits," which numbered only a few thousand a year not long ago, topped 340,000 last year. Major companies like Microsoft are pre-ordering hundreds of thousands of credits for trials whose results are not yet known. It sounds appealing: the ocean already absorbs carbon from the atmosphere, so what do we lose by "speeding up" that process a little?

The answer is far more complicated than it seems. This piece draws on three sources: the Ocean & Climate Platform's policy brief published in June 2026 ("Navigating Hopes and Threats"), the underlying npj Ocean Sustainability perspective article signed by nearly thirty scientists ("Three Challenges to Marine Carbon Dioxide Removal"), and the Stockholm Resilience Centre's coverage that brought the topic to public attention. My aim is not alarmism, but to share the realistic picture the evidence actually shows.

1. What mCDR is, and what it isn't

mCDR is the umbrella term for a group of techniques designed to artificially strengthen the carbon-absorption process the ocean already carries out. There are three main approaches:

• Biological approaches: These rely on photosynthesis. Triggering phytoplankton blooms through iron fertilization or artificial upwelling, or cultivating seaweed (macroalgae) and sinking the biomass into deep water, fall into this category.
• Chemical approaches: These directly alter the chemistry of seawater. Adding alkaline substances such as crushed olivine or lime to raise pH (ocean alkalinity enhancement), or using engineered systems to extract CO2 directly from seawater (direct ocean carbon capture), belong here.
• Physical approaches: Artificially pumping CO2-rich surface water down into deeper layers (artificial downwelling) tries to mimic the ocean's own "biological pump."

One distinction here is critical: blue carbon — the conservation and restoration of mangroves, salt marshes, and seagrasses — sits in a different category from these three approaches. It is technologically mature, relatively safe, and provides real co-benefits for ecosystems. But its potential is limited: at most 0.1–1 Gt CO2 per year, roughly 1% of current global emissions. All the other techniques remain experimental, unproven, and — if deployed at scale — carry the potential to affect vast stretches of ocean. This piece focuses on that second group, excluding blue carbon.

2. The measurement problem: "How would we even know it's working?"

This may be the least discussed but most critical point in the mCDR debate. Proving that a technique genuinely captures carbon, holds onto it for a long time, and does so without being offset elsewhere (additionality) requires a robust Monitoring, Reporting, and Verification (MRV) system. On land, for forest-based CDR, this is relatively more tractable. At sea, two fundamental obstacles stand in the way:

First, the signal-to-noise problem. The ocean already naturally absorbs more than 10 Gt of CO2 every year. Distinguishing the marginal contribution of an mCDR project from within this enormous natural cycle is, given the scale of regional natural variability, often close to practically impossible. The existing uncertainty in the carbon cycle can exceed the very effect a given mCDR technique claims to produce.

Second, uncertainty about permanence. Claims of storage "for centuries, even millennia" are common, but some independent assessments suggest the realistic storage duration could be as short as 50 years. It's easy to say buried carbon will "stay there" in the deep sea; but the deep ocean isn't an empty vault — it's a living ecosystem, and verifying whether carbon truly stays put for centuries is nearly impossible.

A third problem lies in the verification mechanism itself. There is still no commonly agreed standard for MRV protocols today. Worse, the companies that verify carbon credits earn revenue based on the volume of credits they verify — meaning their claimed independence is undercut by a structural conflict of interest.

3. Possibilities for sleight of hand: Science, or marketing?

With measurement this uncertain, there is nonetheless considerable commercial activity underway. To date, roughly 1 million ocean carbon credits have been sold; only around 30,000 of them are recorded as "delivered." Most of the market is concentrated among four companies. Some companies sell future credits before a trial's outcome is even known — the clearest example being a deal in which Microsoft agreed to purchase up to 350,000 credits from a single company.

This picture creates three distinct risks:

• A disconnect between reward and outcome: If credits are sold before an experiment concludes, exaggerated or unproven results risk being rewarded. The credibility of the carbon market is already contested; this makes it worse.
• A "tragedy of the commons": Most of the financing flows toward field trials driven by commercial interest. Profits are privatized while the potential ecological and social risks are borne by the public.
• Mitigation deterrence: Presenting mCDR as a "technical savior" risks diverting political attention and financial resources away from the real priority — emissions reduction. Some companies, and even some climate policies, are already declaring "net zero" targets based on unproven CDR potential — if that potential fails to materialize, an extremely difficult gap will remain.

On top of this comes "forum shopping": because the regulatory landscape is so fragmented (more on this below), companies can choose whichever country or region offers the loosest oversight. This raises the likelihood that risk will shift toward countries with weaker regulatory capacity — disproportionately in the Global South.

4. Risks: Ecological and social dimensions

The risk profile varies by technique, but the general picture is as follows:

• Phytoplankton fertilization: Harmful algal blooms, oxygen depletion, disruption of nutrient cycles and food webs.
• Ocean alkalinity enhancement: Sudden local shifts in pH, trace-metal pollution (e.g., nickel), and the deforestation and water pollution caused by land-based mining of materials like olivine.
• Biomass-sinking techniques (e.g., sunk macroalgae): Oxygen depletion on the deep-sea floor, sediment acidification, and smothering of benthic life.
• Artificial downwelling: Disruption of deep-sea habitats, altered stratification.

Beyond ecological risks, the social and economic dimension cannot be ignored. Coastal communities may face conflicts with these projects over fisheries and marine-space rights. The principle of "free, prior, and informed consent" for affected communities exists in theory, but is frequently sidelined in practice.

Finally, there is a governance gap. The international frameworks that govern marine areas — the UN Convention on the Law of the Sea, the London Protocol, the Convention on Biological Diversity, the new BBNJ Agreement, the UNFCCC — none of them were designed specifically for mCDR. The 2013 amendment to the London Protocol, adopted specifically to regulate marine geoengineering, still has not entered into force more than a decade later: only 9 of 56 contracting parties have ratified it, and key players such as the United States are not among them. This fragmented structure creates both an oversight gap and fuels the forum-shopping described above.

Conclusion: Even the optimistic IPCC doesn't open this door

Here I want to underline a point from my own field. As I often write, IPCC scenarios sit structurally on the optimistic side — they largely exclude "tipping-point dynamics" such as permafrost methane release, potential Amazon dieback, or abrupt shifts in the AMOC. In other words, the IPCC's official projections do not represent the realistic worst case; they represent the realistic optimistic bound.

And even that optimistic framework does not present mCDR as a solution. IPCC scenarios include CDR at varying levels — in some, nearly zero — but none impose mCDR as a defined target or plan. On the contrary, the IPCC flags these techniques as immature and risky. Even as the Seventh Assessment Report is being prepared, the literature on mCDR's efficacy and risks remains limited — precisely because so few experiments have been designed to measure them.

At this point, it's worth pausing to ask: if a scientific consensus that is already structurally cautious — one that represents the optimistic bound rather than the worst case — flags a technique as "not yet ready," how responsible is it to present that same technique to the public today as a climate solution? The answer is clear: it isn't. Urgent, drastic emissions reduction remains the one proven path we actually have. Elevating mCDR as a "technical escape hatch" is both scientifically premature and politically dangerous.

This is, in fact, a new version of a familiar disease in climate policy: treating mitigation and protection (adaptation) as sequential steps rather than parallel ones, and using one as an excuse to defer the other. Deferring today's mitigation burden onto an unproven removal technology is just as dangerous as deferring adaptation until "mitigation fails first" — and that deferral falls hardest on the vulnerable countries and communities that can least afford to pay its price.

This is not pessimism. This is realism grounded in data. Cautious, well-designed, non-commercial small-scale research — yes, absolutely; we need it to close the remaining knowledge gaps. But sacrificing the ocean prematurely to the promise of an unproven "commercial solution" — no.

Sources: 

• Ocean & Climate Platform (2026), "Navigating Hopes and Threats: How Precaution Should Guide Marine Carbon Dioxide Removal Research and Governance"; 
• Brun, V. et al. (2026), "Three Challenges to Marine Carbon Dioxide Removal," npj Ocean Sustainability; 
• Stockholm Resilience Centre news coverage (June 2026).*