Geoengineering, says scientist David Keith, "is like chemotherapy. It's something nobody should like."

But if you can't avoid cancer, chemotherapy may be your best option. And, if it becomes evident that the earth can't avoid the catastrophic impacts of climate change, it is not merely possible that governments will turn to geoengineering.

Some people believe that it is all but certain.

Geoengineering, as you probably know, is the deliberate large-scale manipulation of the planet to counter global warming. It can take a number of forms, as the graphic below shows, some perhaps still to be discovered. Long a taboo subject, geoengineering is being talked about openly these days by scientists, environmentalists and policy thinkers.

The National Academy of Sciences held a workshop on geoengineering in June. Influential books including SuperFreakonomics and Whole Earth Discipline, by longtime environmentalist Stewart Brand, argue that it's time to take geoengineering seriously. A congressional subcommittee held its second hearing on geoengineering just last week.

Among those testifying was Keith, who directs the energy and environmental systems group at the University of Calgary and, interestingly, also leads a team of engineers who are developing a technology to capture CO2 from ambient air. I heard him speak a week ago during a six-hour workshop on geoengineering organized by the Environmental Defense Fund, a nonprofit known for its pragmatism. EDF invited me to attend, on the condition that I seek permission from the scientists before quoting them.

Geoengineering is not a new idea -- it was mentioned in a 1965 report on the environment delivered to President Lyndon Johnson. But until recently, environmentalists have avoided talking about it because they worry that a focus on geoengineering will divert attention and resources from their attempts to get governments and business to curb carbon emissions -- attempts which, it must be said, have had limited success so far.

Nor is geoengineering entirely unproven. Experts say solar radiation management (SRM), the form of geoengineering that has drawn the most attention lately, can be achieved by adding light-scattering aerosols to the upper atmosphere or increasing the reflectivity of clouds below.

What makes scientists think it will work? When the Mount Pinatubo volcano in the Philippines erupted in 1991, spewing fine particles of sulfur dioxide into the stratosphere, enough sunlight was reflected back into space that the earth was cooled by about 0.5 degrees C, at least for a time.

The trouble is, solar radiation management surely will have other consequences as well. Some are known -- less precipitation and less evaporation, which is bound to affect agriculture -- and others are not.

"The concerns, really, are the unknown unknowns," says Keith.

The EDF workshop was itself a sign that geoengineering is moving closer to the mainstream. It was organized for EDF's trustees and senior staff during a board meeting at Cavallo Point in Sausalito, Ca.; the organization hasn't decided yet whether to support further research into geoengineering but, to its credit, it is open-minded about the idea. Listening to the presentations, I found myself appalled at times and thrilled at others. This is a fascinating subject, one that raises many more questions than there are answers.

One useful way to think about geoengineering in general and SRM in particular is to compare them to mitigation, the current approach to climate change. Mitigation means reducing carbon emissions, most importantly by replacing the burning of fossil fuels (coal and oil) with low-carbon energy sources such as wind, solar, nuclear power, so-called cleaner coal and biofuels -- on a vast scale. Mitigation requires enormous expenditures of capital and takes a very long time to work because CO2 emitted today persists in the atmosphere for decades. Even if we could arrange for an international agreement to curb emissions, which we cannot, it will take decades to reverse the rising concentrations of carbon in the atmosphere.

By contrast, solar radiation management is arguably "fast, cheap and imperfect," said Keith -- particularly if it is done crudely and without proper governance, oversight and testing. As little as $5 to $10 billion a year could pay for a short-term program, scientists estimate. By email, Keith put it this way: "The raw cost of implementation is less than 10% and probably less than 1% of the cost of cutting emissions when you average costs over 100 years." Most of the technology required is within reach.

"It's pretty clear that you could do it if you wanted to, and you could do it now," Keith said. "If we put a lot of reflective aerosols in the upper atmosphere, it gets colder and it gets colder quickly."