The chorus of people calling for scientists to investigate geoengineering schemes, like seeding the seas with iron or the clouds with aerosols, as emergency backup plans to fend off runaway global warming, continues to swell.

Last year Britain's prestigious Royal Society published a special issue of its journal suggesting the idea should be taken seriously in case governments fail to act, and now an even bigger player has joined in. John Holdren, President Obama's science adviser, told the AP that the administration would have to at least consider geoengineering plans: "We don't have the luxury ... of ruling any approach off the table," he said.

After Holdren's comments ignited a freak out, the White House clamored to clarify its position—Holdren never said the administration would pursue any of these schemes, only that nothing would be absolutely ruled out. However, a new study suggests that even if politicians were serious about iron seeding, one of the most popular geoengineering ideas, it might not do any good anyway. Or, at least, it would take a lot more thinking than just dumping metal into the sea and hoping for the best.

The iron hypothesis, which dates back two decades, speculates that seeding the sea with iron could prompt blooms of plankton, and that those plankton would take up atmospheric carbon. They go through the food chain, and eventually the carbon is sequestered at the bottom of the sea. But the study by Department of Energy oceanographers Todd Wood and Jim Bishop found that while recent iron seeding experiments made plankton bloom, much of the carbon never made it to the sea floor.

During the great Southern Ocean Iron Experiment of 2002, a test of the iron hypothesis in the Southern Hemisphere, Wood and Bishop deployed their autonomous Carbon Explorer down 800 meters or more where the seeding had taken place, and also in non-seeded areas as a control. The Carbon Explorer transmitted the carbon levels they measured 24 hours a day for more than a year. The researchers initially ublished a paper that seemed to support iron seeding with its data, but taking this second look at the numbers these years later found that the carbon situation was a lot more complex.

Specifically, they say, the most likely reason the Carbon Explorer didn't pick up carbon when lush plankton bloom was above it, was that the experiment hadn't accounted for the rapid mixing of different levels of seawater in that area of the Southern Ocean. Currents pulled down the surface phytoplankton to a depth that was too dark for them to grow; larger zooplankton couldn't snack on them. Thus, the zooplankton population has died down by the spring, when it gets lighter. The phytoplankton rebounded, but there weren't enough predators to eat them and the chain was broken.

None of this means that iron seeding couldn't work in some places—it did prompt a plankton bloom like it was supposed to. Only, as usual, the complexity of an ecosystem overpowered our desire to control it. Little effects—like some mixing water—have big biological consequences.