[I think that as a climate-saving strategy geo-engineering is somewhere between a dead end and a hoax — why would you choose chemotherapy that might make you sicker if your doctors told you diet and exercise would definitely work (see "Geo-engineering remains a bad idea" and "Geo-engineering is not the answer")?

The likely new science advisor, John Holdren, has written, "The ‘geo-engineering’ approaches considered so far appear to be afflicted with some combination of high costs, low leverage, and a high likelihood of serious side effects (PDF)." And the new head of NOAA is someone "who would put oceans first," whereas absent a successful effort to stabilize at 450 ppm or below, most geo-engineering schemes would put oceans last, leaving them acidified and inhospitable to most current ocean life, possibly for hundreds of thousands of years. But do our children and their children and the next 5000 generations really need a livable ocean if it means we don’t have to reallocate about 1% to 2% of our wealth today (see "Absolute must read report: Debate over, further delay fatal, action not costly")?

Yet desperation drives some people to contemplate extreme things, and climate scientists are increasingly desperate to prevent global warming (see "Desperate times, desperate scientists"). Jeff Goodell files this reports dispatch from the AGU’s annual meeting.]

On Wednesday, in the Q & A session after Jim Hansen’s talk about the dire state of the earth at the AGU meeting, eminent Rutgers University professor Paul Falkowski asked Hansen: "The genie is out of the bottle now — What do you think of geoengineering as a way to deal with global warming?"

I half-expected Hansen to throw his laser-pointer at Falkowski. After all, geoengineering — deliberate, large-scale manipulation of the earth’s climate — has long been a taboo subject in the climate debate. Only crackpots brought it up.

But Hansen didn’t miss a beat. He said it would make sense to try "soft" geoengineering first, such as no-till agricutlure and afforestation. But as a last resort, Hansen admitted, more aggressive geoengineering schemes might be necessary. Call it prudence. Or desperation.

Geoengineering comes to two flavors: the less controversial variety is really just carbon management, such as dumping iron in the oceans to stimulate plankton blooms or building stand-alone machines that can strip CO2 out of the air. The bolder variety is more accurately described as albedo engineering, since it refers to technologies that modulate the amount of short wave radiation (aka sunlight) that hits the planet. One popular idea is to mimic the natural process of a volcano and inject tiny particles into the stratosphere to deflect sunlight. Another is to seed clouds to brighten them and increase reflectivity. In theory, you can offset the warming caused by a doubling of CO2 by reducing the amount of sunlight that hits the earth by just 1 – 2 %. And, unlike carbon management, which works slowly due to the earth’s thermal inertia, albedo engineering is something you could do to lower the temperature of the planet in a hurry.

There have always been three big objections to geoengineering: first, that it’s a "quick fix" that diverts attention from the push to cut CO2 pollution; second, that cooling the earth by reducing sunlight does nothing to solve other urgent problems caused by high CO2 levels, such as ocean acidification; third, that we are mucking around with a system we don’t understand and will undoubtedly screw it up. And if the climate system crashes, there is no restart button.

These objections remain. But they are increasingly downed out by rapidly melting ice, our abject failure to cut CO2, and fears of big climate surprises in the near future. As David Keith, Director of the Energy and Enviromental Systems Group at the University of Calgary, put it in his talk at AGU on Thursday morning: "Uncertainty + Inertia = Danger."

But Thursday’s sessions at AGU reinforced the idea that geoengineering is fraught with unexpected complexity. Keith sketched ideas about how you might loft particles into the stratosphere (dump them out of jet aircraft, push them up a long hose connected to a high-altitude balloon), as well as the possibility of engineering particles to optimize mass and light-scattering qualities. But Richard Turco, professor of atmospheric sciences at UCLA, argued that it is extremely difficult to control the size of particles in the stratosphere, thus making it hard to know how much to inject and what the precise impact will be. "It will be like trying to control the clouds," Turco said.

As for environmental impacts, one big concern is changes in global precipitation patterns. Although some climate modelers have argued that a geoengineered climate might be more like our own than a high-CO2 climate without geoengineering, one of the mostly hotly debated issues right now is how brighter clouds or particles in the stratosphere will impact rainfall in the tropics and sub-tropics. One modeling study presented at AGU showed severe drying in the Amazon basin; another was inconclusive.

"Maybe geoengineering is a crazy idea," one veteran climate scientist told me after Thursday’s session. "But it’s even crazier not to consider it, because the sad truth is, the way things are going, we may need it."

[Geo-engineering is one of the most controversial areas of climate science and policy. Even Jeff and I don’t see eye-to-eye on it. For instance, I think there are many more than three big objections. And as a practical matter, I think geo-engineering is unlikely to be a viable or rational strategy, especially as a pure alternative to very, very, very robust mitigation. Explaining why will take some effort, so early next year I will do a multipart series on the subject. As always, I very much welcome your thoughts.]