Marine carbon removal is moving from theory into real-world field trials. Are the science and governance keeping pace?
In a new Comment in Nature Climate Change, Finn Ross (Visiting Fellow to the department) and colleagues argue that the key question is not whether marine carbon dioxide removal should move forward despite uncertainty, but what level of uncertainty is acceptable at each stage of development.
Climate pathways consistent with the Paris Agreement suggest we may need to remove around 7–9 billion tonnes of CO₂ from the atmosphere each year by 2050. Novel carbon-removal methods currently deliver only a tiny fraction of that. The ocean could make a material contribution, but it is not simply a very large container for carbon. Ocean currents can carry altered chemistry, nutrients and biological effects well beyond a project boundary. Benefits may occur in one location while ecological consequences emerge somewhere else, potentially much later. Even measuring the additional carbon removed can be extremely difficult against the natural variability of the ocean. That is why we argue the choice should not be framed as “deploy now” versus “wait until everything is known”. Instead, precaution should scale with the intervention:
- Laboratory studies and bounded field experiments should be enabled where they can resolve important uncertainties.
- Larger and longer trials should face progressively stronger evidence, monitoring and governance requirements.
- Research programmes need predefined stage gates, with clear criteria to advance, modify, pause or stop.
- The ability to stop an intervention should not be confused with the ability to reverse its ecological effects.
Delay is not necessarily the risk-free option either. One US modelling study estimated that postponing novel carbon removal until mid-century could require carbon prices 59–79% higher and lead to US$128–220 billion in losses from prematurely retiring fossil-fuel infrastructure.
The responsible path is therefore neither paralysis nor a rush to commercial deployment.
It is to learn at scales where risks can be managed, while raising the evidence threshold as interventions become larger and more consequential.
Read the full article, Marine carbon removal at a crossroads, in Nature Climate Change here: https://www.nature.com/articles/s41558-026-02740-8.
