I have some extended quotes from David Archer on the subject below the fold. But if you are interested, read the whole thing.
In spite of public relations claims by Planktos representatives in comments, it appears that most of the scientific community does not think highly of the Planktos claims.
Marinov et al (2006) showed that a stimulation of phytoplankton production in one part of the ocean usually acts to depress production elsewhere. So what’s the point of paying for a carbon offset to fertilize a water parcel now, when nature would fertilize it soon anyway? That’s against the rules of offsets; it has to be something that wouldn’t happen anyway.
The one part of the ocean where fertilization of the ocean does not depress the fertility elsewhere is the deep Southern Ocean. Here the water sinks to the abyss, rather than taking a leisurely tour through the upper ocean. But now the practical picture looks different. Instead of the benign tropics, you have sea ice, waters mixed to hundreds of meters down (bad for phytoplankton) and total darkness for much of the year. Fertilize that!
Modelers have long ago concluded that iron fertilization of the ocean can play only a small role in managing the carbon cycle in the coming century. Part of the issue is that the Southern Ocean also covers only a very small area of the surface ocean, just a few percent. Model experiments where the Southern Ocean is completely fertilized show a drawdown of maybe 15 ppm by the year 2100 [Zeebe and Archer, 2005]. We could change a light bulb and do better than that.
But the change in carbon chemistry of the ocean and ultimately the atmosphere need to be transparently documented, also, if we are to trade carbon offsets based on iron fertilization. Documenting a change in carbon content of surface waters might be possible in the tropics, but it would be a nightmare in the Southern Ocean, probably impossible to do reliably. Ocean chemistry data is generally cleaner than land data, less susceptible to local variability. In tranquil, well-behaved parts of the ocean like near the Galapagos, it would be probably easier to document changes in the carbon content of the upper ocean than it would be on land. On the other hand, the ocean moves around a lot more than the land does, in general. The Southern Ocean, in particular, is a maelstrom. Tracking a plume of fertilized water to measure the change in carbon content would be a mite trickier.
Southern Ocean surface water also has a harder time changing the CO2 concentration of the atmosphere, because it gets mixed into the interior so quickly. Ultimately it would take centuries to bring the atmospheric CO2 to a new equilibrium value. You would have to wait until your fertilized water filled up the entire deep ocean. I think the long time scale also means that a ton of carbon removed from Antarctic surface waters does not translate to a ton of carbon removed on some reasonable timescale from the atmosphere. The efficiency is much lower than that, and difficult to document.
I would put ocean fertilization on the avoid list, along with planting trees. It’s too hard to pin down the actual amount of CO2 removed from the atmosphere by your actions. It’s also not a long-term solution, since the ocean leaks. Humankind would have to keep fertilizing the ocean indefinitely in order to preserve the claimed CO2 drawdown. If you’re concerned about climate change, build a windmill. Ocean fertilization does not seem to me suitable to be the basis for a reliable financial commodity, or a practical tool for geo-engineering climate.