Today President Obama said that climate change skeptics are being pushed to the margins, but that may have been wishful thinking.

Poll results from the Pew Research Center for the People and the Press released yesterday say that the number of people who believe "there is solid evidence that the earth is warming" dropped from 71 percent in April of 2008 to 57 percent now. Only 36 percent said there was good evidence warming is due to human activity, down from 47 percent in April of 2008. Only 35 percent say climate change is a serious problem.

The numbers of climate change believers have been declining for the last few years among Democrats, Independents and Republicans. For independents, for example, 79 percent believed there was solid evidence in 2006, compared to 53 percent now. It might not be a coincidence that Al Gore's film, "An Inconvenient Truth," came out in 2006.

Now here's the really odd thing. In spite of these low numbers, 50 percent of Americans believe there should be limits on carbon emissions, even if this causes energy prices to rise. Only 39 percent oppose it.

There's an uncharitable interpretation--that Americans are being inconsistent. But there's also a more hopeful interpretation. Climate change models are full of uncertainties. No one really knows just how much the Earth will warm, or what impact this will have, particularly on regional weather patterns. Maybe Americans are learning about these uncertainties, hence the lower numbers siting "solid evidence," yet concluding that the risk is high enough that we should do something to avoid the worst possible scenarios.

The sequel to Freakonomics, the best-selling book that uses economics to uncover surprising facts about the world, came out today. Superfreakonomics, cowritten by Steven Levitt, a professor of economics at the University of Chicago, and Stephen Dubner, a journalist, is an attempt to outdo the original, and it does this in part by taking on a huge, controversial, and very important topic--climate change.

Unfortunately, the authors' solution to climate change, which they say is simple, cheap, and safe, is actually dangerous--a cure that could be worse than the disease. (This part of the book has already generated plenty of debate online.)

The authors set up their chapter on climate change as a challenge to global-warming orthodoxy--saying that "the movement to stop global warming has taken on the feel of a religion," putting climate-change claims in the context of past errors by scientists, and suggesting that climate models are less reliable than risk models for financial institutions that failed in the recent waves of bank closures.

So it's a little disorienting to discover that the chapter actually argues for the development of radical solutions to global warming. It argues that not enough has been done to curb greenhouse gas emissions and warns of catastrophic events like the melting of ice sheets in Greenland and Antarctica.

The solution that Levitt and Dubner put forward is geoengineering. More specifically, they advocate a scheme that would inject particles into the upper atmosphere to block a small percentage of incoming sunlight and so cool the earth--an idea that's been around since at least the 1970s. The scheme would mimic the action of big volcanic eruptions, which also inject particles into the stratosphere and have been shown to have a cooling effect.

Historically, Levitt and Dubner say, the main problem with this idea was that proposals for injecting the particles have been too expensive. They add that there might be some sort of vague environmental concerns, but label them as religious objections, not practical, science-based ones. The "moralism and angst" of these environmentalists make it hard for them to see what the authors call a "fiendishly simple" and "startlingly cheap" solution to global warming. They then describe a scheme for delivering sulfur dioxide (which will form sulfate particles) to the stratosphere and declare that it would cost $250 million for the first year and $100 million thereafter, compared to $1.2 trillion a year for reducing carbon emissions. A bargain.

Other than dismissing the potential for damage to the ozone layer, the authors don't talk about the real environmental concerns that come with sulfate injection to the stratosphere. But there are serious and specific concerns.

Scientists studying the impact of a fairly recent, large volcanic eruption--the Mount Pinatubo explosion in the Philippines in 1991--have found that not only did the layer of sulfates it produced cool the earth, it also led to a "huge change in precipitation," says Gavin Schmidt, a climate scientist at the NASA Goddard Institute for Space Studies. By decreasing direct sunlight, the event cut down on evaporation, leading to the "lowest rainfall amount over land since 1948," the earliest year that good records are available, says Kevin Trenberth, a climate scientist at the National Center for Atmospheric Research in Boulder, CO. The change in precipitation caused severe droughts that damaged crops and limited drinking water, he says. Schmidt says the potential for drought must be considered before any geoengineering is done. "What good does it do to save the Arctic if you cause the failure of the Indian monsoon on a regular basis?" he says. "That's billions of people."

The change in precipitation isn't the only known adverse affect. Shading the earth does nothing about the levels of carbon dioxide in the air. This has some benefits--plants grow better with more carbon dioxide--but it also makes the ocean more acidic, which can lead to the destruction of coral reefs around the world and prevents some shellfish and crustaceans from developing, cutting off an important source of food for fish and whales, and ultimately destroying important food sources for humans.

And then there are potential unanticipated consequences. Volcanoes inject sulfates into the stratosphere sporadically. No one knows what will happen if the sulfates become a permanent part of the stratosphere. It could very well be that major problems won't become obvious until many years or decades into a sulfate injection project. Levitt and Dubner argue that we could simply stop if problems arise. But this could be disastrous. All of the warming that's been prevented by the sulfates over the years would happen suddenly, far too fast for people to adapt.

If nothing is done to curb greenhouse gas emissions, the sulfate injection scheme will have to be kept up year after year, potentially for well over a hundred years, given the lifetime of carbon dioxide in the atmosphere. As concentrations of the gases mount, ever more sulfate will be needed to offset the warming effect, increasing costs. And the dangers of stopping the program--due to war or economic hardship or a shift in the political winds--would mount. The same holds true for another scheme the authors mention--cloud whitening, an approach that may not work and that could also lead to severely reduced precipitation over land. It is not, as they suggest, "geoengineering that the greenest green could love."

Geoengineering by shading the earth is simply not an alternative to curbing greenhouse gas emissions. In some extreme case--the impending collapse of major ice sheets, or the realization that the world is warming far faster than anticipated--it might be used to buy a little time. But even this is a risky proposition, not just because of the environmental concerns, but because of political ones, since some countries would be harmed more than others. The authors point out--in passing--that one can "imagine the wars that might break out over who controls the dials," that is, who selects how much the earth should be cooled. Oddly, they don't seem to consider this a serious objection to geoengineering.

But although the authors may be wrong in failing to point out the significant hazards of shading the earth (let alone some annoying side effects, such as obscuring the view from ground telescopes and reducing the power output from some solar power systems), they may be right that geoengineering may prove necessary. They point out that changing people's behavior is notoriously difficult, and that the uncertainty of climate predictions makes it particularly hard to set up and enforce government policies, particularly those that require international agreements. For poor countries, the uncertain cost of climate change may seem small compared to the cost of forgoing cheap electricity, at least until cheap carbon sequestration or renewable energy is available.

Donald Johnston, the former secretary general for the Organisation for Economic Co-operation and Development (OECD), has said that political realities may make strong international emissions controls impossible: "I foresee a situation about 10 years from now where the world will be warming, the new targets for greenhouse gases set [at the December 2009 United Nations climate change meeting] in Copenhagen will be ignored by many big emitters as they have in the past, and desperation will force the world to consider reducing the penetration of the sun's rays through geoengineering."

If we reach that point, we'd better have a clear idea what geoengineering might entail, so we can choose the best methods and prepare for the inevitable bad side effects. That means research must be funded to create ever more sophisticated computer models of geoengineering and to run some small- and perhaps even large-scale experiments. Also, governments need to start talking about geoengineering policy. How do you decide--and who decides--how much to cool the earth? How do you decide how to reimburse people who suffer from negative side effects? How will lawsuits be handled? What's to be done if a country decides to undertake geoengineering on its own?

This research and planning should be accompanied by continued efforts to reduce greenhouse gas emissions and, eventually, to start pulling carbon dioxide out of the atmosphere. The goal should be to shade the earth for as short a time as possible--or not at all. The only way to drive these changes is to be as clear as possible about the dangers of both global warming and geoengineering. That's going to be a lot harder with Levitt and Dubner making geoengineering sound like a panacea.

If we really wanted to decrease gas consumption, we'd push state and federal governments to make driving more expensive by doing things like increasing gas taxes and tolls on highways. That would force consumers to buy more-efficient cars, move closer to work, or even use public transit. But politicians know that these taxes and tolls are unpopular, so no matter how much sense they make, they tend to vote against them.

It just happened again, this time in Massachusetts. Faced with government spending that's far higher than what the commonwealth is taking in, legislators decided that they needed to increase taxes by about a billion dollars. But instead of raising them in a way that would actually do some good, legislators decided to increase the sales tax by 25 percent, according to the Boston Globe, while dismissing a proposal from the governor to increase gas taxes. The sales taxes will also stave off a proposed highway toll increase, which would have made it more expensive to drive.