Ocean iron fertilisation revisited: from Planktos to the new marine carbon removal

For a few years around 2007, it seemed possible that the first large-scale act of deliberate climate engineering would be carried out by a small American company with a second-hand ship and a plan to sell carbon offsets. It did not happen. Five years later, something close to it did, off the coast of British Columbia. The rules written in response still shape how the oceans are treated as a carbon sink, and a new generation of companies and researchers is now testing them.

The idea: rust as fertiliser

Phytoplankton, the microscopic algae at the base of the marine food web, need iron in tiny amounts to photosynthesise. In large parts of the ocean, notably the Southern Ocean, the equatorial Pacific and the subarctic North Pacific, surface waters are rich in nitrate and phosphate but grow surprisingly little plankton. Oceanographers call these "high-nutrient, low-chlorophyll" regions.

In the late 1980s, the oceanographer John Martin, at Moss Landing Marine Laboratories in California, argued that iron was the missing ingredient. His bottle experiments suggested that adding iron to water from these regions set off rapid growth. He also linked the idea to the ice ages: ice cores showed that glacial periods were dustier, and wind-blown dust carries iron, so perhaps a dustier world had fertilised the oceans and drawn carbon dioxide out of the air. Martin is remembered for a lecture-room quip to the effect that half a shipload of iron could start an ice age. He died in 1993, before the open-ocean tests of his hypothesis.

The climate logic was simple. Add iron, grow plankton, and when the plankton die or are eaten, some of their carbon sinks into the deep ocean, where it stays for centuries.

A decade and a half of experiments

Between 1993 and 2009, research teams carried out about a dozen open-ocean iron experiments. The IronEx cruises in the equatorial Pacific in 1993 and 1995 came first; SOIREE, south of Australia in 1999, was the first in the Southern Ocean. Others followed off Japan, in the north-east Pacific and in the South Atlantic sector of the Southern Ocean.

The results were consistent on one point and murky on the one that mattered. Adding iron almost always produced a bloom, often visible from satellites. Showing that much of that carbon reached the deep ocean, rather than being eaten and breathed back out near the surface, was far harder. One 2004 experiment, written up in Nature in 2012, reported that a substantial part of its bloom sank to great depths. Others found little export. The Indo-German LOHAFEX expedition in 2009 produced a bloom that was largely eaten by small crustaceans, with modest carbon sinking. That expedition also became a political fight in Germany, where the environment ministry objected on the grounds of international biodiversity commitments and the research ministry let it proceed after outside reviews.

Scientists also raised side effects. Some experiments saw growth of diatom species that can produce a neurotoxin. Models suggested that fertilising one patch of ocean could rob nutrients from waters downstream, lowering productivity there, and that decomposing plankton could increase emissions of other greenhouse gases such as nitrous oxide. By the late 2000s, the mainstream view among ocean scientists was that iron fertilisation was useful as an experimental tool for understanding ocean biology and a poor bet as a climate fix.

The entrepreneurs arrive

Markets moved faster than the science. Planktos, a California company, announced in 2007 that it would spread iron over a patch of the Pacific and sell carbon offsets based on the plankton bloom, initially naming waters near the Galápagos. Its ship, the Weatherbird II, left Florida in late 2007.

The objections came from many directions: oceanographers who doubted the carbon would stay down, conservation groups worried about one of the world's most famous marine ecosystems, and regulators who had never had to consider a commercial ocean-dumping scheme sold as climate action. The scientific groups of the London Convention and Protocol, the treaties that govern dumping at sea, issued a statement of concern in 2007. In early 2008 Planktos suspended its operations, blaming a hostile campaign and a lack of investors.

The episode prompted the first formal responses. In 2008 the parties to the London Convention and Protocol agreed that ocean fertilisation activities, other than legitimate scientific research, should not be allowed. The same year, the Convention on Biological Diversity asked its parties to ensure that ocean fertilisation did not take place until there was an adequate scientific basis, with an exception for small-scale research in coastal waters. In 2010 the London treaties adopted an assessment framework for deciding what counted as legitimate research.

Haida Gwaii, 2012

In July 2012, the Haida Salmon Restoration Corporation, a company backed by the village council of Old Massett on Haida Gwaii, released roughly 100 tonnes of iron-rich material into the Pacific a few hundred kilometres west of the islands. One of the people involved was Russ George, who had led Planktos. The stated aims were to restore salmon runs by boosting the plankton at the bottom of their food chain, and potentially to earn carbon credits.

The dump became public in October 2012 through press reports, and satellites had picked up a large bloom. Environment Canada opened an investigation. The project's backers later pointed to a very large pink salmon return in 2013 as evidence of success; fisheries scientists said the connection could not be established, given how much salmon numbers vary from year to year.

What made Haida Gwaii important was not its scale but its origin. This was not a Northern start-up working in someone else's waters. It was a community with real grievances about declining salmon, persuaded that an unproven intervention could help it and pay for itself. That combination is likely to recur.

The 2013 amendment

The London Protocol's parties responded in October 2013 by amending the treaty itself. The amendment adds a new article and two annexes. Marine geoengineering activities listed in the new annex may only be carried out under a permit, and a permit may be granted only for legitimate scientific research that passes the assessment framework. Ocean fertilisation is the only activity listed so far. The design allows other techniques to be added later.

The amendment is still not in force. It needs acceptance by two-thirds of the Protocol's parties, and only a handful have ratified it. The London Protocol itself has far fewer parties than the older London Convention. So the rule exists, has political weight, and binds almost nobody in legal terms.

The second wave

Interest in iron faded after 2013. Interest in the ocean as a carbon sink did not. Since around 2020, as companies began buying carbon removal for their net-zero claims, a broader set of "marine carbon dioxide removal" approaches has attracted money:

• Ocean alkalinity enhancement: adding crushed minerals or alkaline substances to seawater so that it can hold more carbon dioxide as dissolved bicarbonate, a slow version of the natural weathering of rocks.
• Electrochemical methods: using electricity to strip carbon dioxide from seawater or to make it more alkaline, then returning the water to the sea.
• Growing and sinking seaweed or other biomass, so that the carbon it captured ends up on the deep seabed, a marine chapter in the wider push to turn plant matter into a climate commodity.
• Artificial upwelling and downwelling, moving water between depths to change nutrient supply or carbon storage.

Each raises the questions iron raised. How much carbon is really stored, and for how long? How do you measure it in an open, moving ocean? What happens to the organisms living where the material is added? The business side has been bumpy. Running Tide, a US company that sank wood and other material at sea for carbon credits, shut down in June 2024, saying there was not enough demand for carbon removal. Advance-purchase coalitions such as Frontier, set up in 2022 by technology and consulting firms, have kept some ocean projects funded, but buyers are few and the credits depend on measurement methods still being developed.

Local politics have not gone away either. A proposed alkalinity trial off Cornwall drew organised local opposition in 2023 from residents and people who use the bay. Iron itself is also back on the research agenda: a US-led consortium called Exploring Ocean Iron Solutions has argued for new, closely monitored field studies in the north-east Pacific, presenting them as the careful science the earlier experiments lacked.

Who decides now

The treaty bodies have tried to keep pace. In 2022 and 2023 the London Convention and Protocol's governing bodies singled out several newer techniques, including alkalinity enhancement and the sinking of seaweed, for priority evaluation, and urged parties to treat proposals with caution in the meantime. Whether any of them will be added to the 2013 amendment's annex is an open question, and the amendment would still need to come into force to make the listing bite.

Other forums now overlap. The Convention on Biological Diversity's 2010 decision on geoengineering, discussed in our guide to solar geoengineering, covers marine methods too. The high seas treaty adopted in 2023, which passed its sixtieth ratification in September 2025 and entered into force in January 2026, requires environmental impact assessments for certain activities beyond national jurisdiction, which could include large ocean interventions. In coastal waters, decisions fall to national regulators who may have never handled such a request before.

Five questions keep coming back, whatever the technique:

1. Who issues the permit? A coastal state, a flag state, a treaty body, or nobody, if the work is done on the high seas from a ship registered where no one asks.
2. Who checks the carbon? If credits are sold, the seller has an interest in generous accounting. Independent verification in the ocean is expensive and still immature.
3. Who is told in advance? The opposition in Cornwall and the solar experiments halted in 2024 suggest that projects which surprise the people living nearby tend to fail, whatever their merits. Haida Gwaii showed the opposite risk: local enthusiasm with almost no outside scrutiny.
4. Who bears the side effects? Nutrient robbing and toxic blooms would affect fisheries downstream, often far from the people who chose to intervene.
5. Who can stop it? Research may be stopped; a commercial operation with paying customers has every reason to continue.

These are the same questions raised about gene drives, another technology designed to spread beyond the place where it is released. They are also, in a sense, the questions the 2008 decisions were meant to settle. Those decisions drew a line between legitimate research and everything else. The newer wave of ocean carbon removal sits uneasily across that line: much of it is framed as research, funded by people hoping to sell credits, and carried out in waters where nobody has clear authority to say no.

Martin's hypothesis has held up well as science. Iron does limit plankton in large parts of the ocean. What the last two decades showed is that a bloom on a satellite image is not the same as carbon stored for centuries, and that the hard part of using the ocean as a climate tool is not chemistry but consent.