The pair suggest that the pipes could be either free-floating or tethered, 100–200 metres long, 10 m in diameter and contain a one-way flap valve at the bottom end to enable pumping by wave movement. The resulting algae bloom would “pump down carbon dioxide and produce dimethyl sulphide, the precursor of nuclei that form sunlight-reflecting clouds,” they say.

Lovelock and Rapley admit that the approach could fail on engineering or economic grounds and that the impact on ocean acidification would need to be taken into account.

“But the stakes are so high that we put forward the general concept of using the Earth system’s own energy for amelioration,” they say. “The removal of 500 gigatonnes of carbon dioxide from the air by human endeavour is beyond our current technological capability. If we can’t ‘heal the planet’ directly, we may be able to help the planet heal itself.”

The field of stimulating algae growth to sequester carbon is a controversial one. Research published in Science in May found that little of the additional biomass produced by an eddy that enriched surface waters actually sank to the ocean floor – instead it was recycled at the surface. "We find that you have to be careful about this strategy as making a phytoplankton bloom occur does not always equal carbon export," said Claudia Benitez-Nelson of the University of South Carolina.

In June, US company Planktos announced plans to use iron dust to fertilize a 36 square-mile area of the Pacific Ocean in international waters around 350 miles west of the Galapagos Islands. Following delays, Planktos’ ship Weatherbird II was due to depart from Florida in September to carry out the trial.