When a report on climate change hit the U.S. president's desk, the suggestion was not to cut greenhouse gas emissions. Rather, scientific advisors counseled intervention via technology in the climate system itself—a practice now known as geoengineering. And the president was not Barack Obama, George W. Bush or even Bill Clinton—it was Lyndon Johnson in 1965.

"This generation has altered the composition of the atmosphere on a global scale through…a steady increase in carbon dioxide from the burning of fossil fuels," President Johnson told Congress in February of that year. To address the problem, his science advisors suggested spreading reflective particles over 13 million square kilometers of ocean in order to reflect an extra 1 percent of sunlight away from Earth.

Today, with climate change accelerating and little being done to curb the greenhouse gas emissions, some scientists have resurrected the idea of "deliberate large-scale manipulation of the planetary environment," as the U.K.'s Royal Society puts it. After all, it's an idea nearly as old as the understanding of the physical principles behind global warming itself. Swedish chemist Svante Arrhenius thought that global warming would be a boon to humanity and therefore fossil fuel burning should be encouraged, after calculating by hand the likely temperature impact of continued coal-burning and rising carbon dioxide (CO2) concentrations in the late 19th century—roughly matching the United Nations Intergovernmental Panel on Climate Change and their computer models more than a century later.

That's why 175 scientists and other interested folks (including companies looking to profit from geoengineering) gathered in the Asilomar conference center near the end of March to try to repeat the success of molecular biologists who gathered there in 1975 to reassure a skeptical public about genetic engineering. Ultimately, the gathered would-be geoengineers released a statement calling for, among other things, "further research in all relevant disciplines to better understand and communicate whether additional strategies to moderate future climate change are, or are not, viable, appropriate and ethical."

The list of unintended consequences of human manipulation of natural systems is long: concrete jungles creating urban heat islands, vast oceanic dead zones resulting from fertilizer use on inland agricultural fields, and intentionally introduced species, such as the cane toad in Australia, that then wreak havoc on ecosystems, among others. Whether the idea is to mimic a volcano's cooling impact on climate by continuously pumping sulfates into the stratosphere or to brighten clouds via crewless ships spewing water vapor, possible problems range from shutting down rainfall in certain regions to unilateral declarations of war.

As the Royal Society noted in its 2009 report on geoengineering: "The safest and most predictable method of moderating climate change is to take early and effective action to reduce emissions of greenhouse gases. No geoengineering method can provide an easy or readily acceptable alternative solution to the problem of climate change."

Nevertheless, humans are already managing the climate, even with actions intended to improve the environment. A recent decision to cut sulfate pollution from cargo ships will, in effect, further warm the globe as more cooling particles are removed from the sky.

spoke to climate modeler and geoengineering expert Ken Caldeira of the Carnegie Institution's Department of Global Ecology at Stanford University, who coined the term "solar radiation management" for efforts to dim the sun (though he now prefers "climate intervention"), about why humans might want to get smart about planetary management.

[An edited transcript of the interview follows.]

First off, what is geoengineering?
Geoengineering is a word that means many different things to many different people. Typically what people call geoengineering is divided into two major classes. There are approaches which attempt to reduce the amount of climate change produced by an increase in greenhouse gas concentrations and there are approaches that try to remove greenhouse gases that have already been released to the atmosphere.

The Earth is warmed by sunlight and the heat that is absorbed by the Earth is later re-radiated back to space. Greenhouse gases make it more difficult for the Earth to radiate energy to space. So the two main ways you can cause Earth to cool are either to create conditions such that Earth absorbs less sunlight or make it easier for the Earth to radiate heat energy back to space.

The first category of approaches typically includes things like: putting giant satellites in space to deflect sunlight away from Earth, putting tiny particles in the stratosphere, whitening clouds over the ocean, or perhaps whitening roofs or planting lighter [colored] crops. They are all attempts to deflect sunlight away from Earth.

The second allows more heat energy to escape.

There is one more category that some people propose: that we might take heat that exists near the surface of the Earth and stuff it down deep into the ocean. This hasn't been looked at very much. But it's another way of altering Earth's surface temperatures.

Why do we even need to think about this?
If we froze greenhouse gas concentrations at current levels the Earth would continue to warm for many decades if not centuries. We did a study showing that if we want to stabilize temperatures through emission reductions, they would need to be cut to zero. Emissions would need to be eliminated entirely. Under every emission scenario considered by the [United Nations Intergovernmental Panel on Climate Change], temperatures continue to increase throughout this century.

Given all of the inertia in the physical climate system, in our energy infrastructure, and our political system, there's really no practical way that emission reductions can reduce the pace of climate change or greatly reduce the amount of climate risk. Emission reductions cannot start cooling the Earth this century, especially if we also control sulfur emissions from power plants, which exert a cooling influence today.

Near the end of this century, if current trends continue, almost every summer in the tropics will be hotter than the hottest yet experienced. That presents the possibility that there could be widespread crop failures and famines. If these kinds of terrible conditions start becoming commonplace we would be facing a situation where many people are starving with the prospect of continued warming for decades and possibly centuries into the future.

Given the enormous stakes and the essential irreversibility of warming through greenhouse gas emission reductions this century, it's only responsible to think about what we would do in the face of a climate emergency. Part of this involves thinking about geoengineering. But we also need to be thinking about developing crop varieties that can grow in hot and dry conditions. We need to be thinking about how to help [developing nations] that can't grow food [in the future] to industrialize so they can get foreign exchange to buy food.

How did you first start studying geoengineering?
I first heard about these ideas in 1998 from [physicist] Lowell Wood, who was a protégé of [physicist and hydrogen bomb–maker] Edward Teller. Teller was pessimistic about human nature and optimistic about technology…. In the mid-1980s, Teller started thinking about climate change. He didn't trust human institutions to develop the capability to reduce greenhouse gas emissions. So he wondered are there technical means to address the climate change problem which wouldn't require changes in human institutions or human nature? He came across the idea of geoengineering.

Geoengineering has deep historical roots. A 1965 report to President Johnson said warming from greenhouse gases could pose a risk to the U.S. It suggested that we might spread reflective particles across the surface of the ocean and this would offset the warming. The concept of reflecting sunlight to space to address climate change has deeper historical roots than reducing greenhouse gas emissions. Nobody suggested that to Johnson.

In 1998, [physicist and geoengineer] David Keith and I were at a meeting with Lowell Wood talking about ways to address climate change. Lowell suggested that we put a bunch of particles high in the stratosphere. I remember thinking, 'Oh, this will never work,' because greenhouse gases work day and night. They work at the North Pole and the equator and in winter or summer. But sunlight is strongest in the day, at the equator and in summer.

So I made an effort to show Lowell Wood was wrong. Back at the office, I worked with colleagues to do simulations that reflected sunlight away. Much to our surprise, reflecting sunlight offset most of the climate effect of increased CO2 both regionally and seasonally.

We set out to show it wouldn't work and our simulation ended up indicating that it basically would…. The resulting climate is pretty similar to the preindustrial climate. It's not exactly the same. You offset 90 percent of the temperature change and maybe 70 percent of the hydrological change.

… [But] CO2 is chemically active and in the oceans forms carbon acid, which attacks the shells and skeletons of marine organisms. These approaches won't do anything to help ocean acidification.

What are the other risks?
There are two main types of risk associated with these climate intervention approaches. One has to do with environmental science and intended or unintended consequences. The other is social, political or even military risks.

In the case of environmental risks, the offsetting of greenhouse gases by increasing the reflection of sunlight is not going to be perfect. Some people, potentially a small minority, will get less rainfall. There is concern about what particles might do to the ozone layer.

In 1991, a volcano in the Philippines known as Mount Pinatubo erupted and sent a huge amount of material into the stratosphere. It reflected two percent of sunlight back to space and Earth cooled by half a degree Celsius. That material fell out of the atmosphere after a year or so but had that material been maintained it would have been more than enough to offset all the global warming expected this century.

After Mt. Pinatubo there was a three percent reduction in the amount of ozone in the atmosphere. This loss of ozone looks like it might be more in springtime and more in northern latitudes. People living in northern Europe could be adversely affected.

… The offsetting of sunlight and CO2 is not perfect. We would expect there to be some impact on ocean circulation. Ocean currents could change dramatically.

We're basically entering into uncharted territory here. There's a host of potential bad things that could happen…. Any time you try to intervene in a complex system, you have unexpected results. But, with greenhouse gases, we are already intervening in a big way in a very complex system.

Environmental consequences are among the least of our worries…. Imagine a scenario where the world has gotten much hotter and China goes into a period of deep drought. The Chinese leaderships says, 'Our people are suffering. We're having a famine. Let's put aerosols in the stratosphere to restore our climate.' It's hard to imagine they would resist deploying that system.

Imagine they do this and then the U.S. goes into a decade or two of similar deep drought. Whether the Chinese intervention was the cause of the American drought or not is almost irrelevant. The U.S. population is highly likely to blame the Chinese for the descent into arid conditions. The potential for great political tension and possibly even military action is high.

I think it's highly likely that as a result of any climate intervention there will be winners and losers. In a nuclear-armed world, a world with terrorism and where losers have the ability to strike back at winners, the potential for the kind of political or military risk to overwhelm any environmental benefits is very real.

The clearest path to environmental risk reduction is greenhouse gas emission reductions. But emissions and global temperatures keep going up and up and up. We need to think about what we'll do if bad things happen…. If we were already on a trajectory of reducing greenhouse gas emissions and helping the most vulnerable develop their economies, I would feel very comfortable about developing these options. If we're still building coal-fired power plants and gas-guzzling SUVs at the same time we're developing these options? You wonder to what extent are these options facilitating bad behavior.

Can these risks be overcome?
The best way from the environmental perspective would be to kind of tiptoe into climate intervention: ramping it up gradually so if bad things started happening you could ramp it back down again. Unfortunately, that's probably the most politically difficult thing to do. We're never going to get consensus to deploy anything until a real climate crisis in which case the motivation will be to turn on the spigot at once to avert catastrophe. From an environmental perspective, that's the riskiest thing you can do.

Do we need to test these ideas? Or just model them?
It's premature to be doing field-testing of deployment systems. [But] a lot of the science you need to do around climate intervention is the same science you need to understand climate change, like understanding how do particles in the stratosphere affect stratospheric chemistry? How do particles in the lower atmosphere affect cloud formation?… It's really research that climate scientists should be doing anyway.

Where things get more controversial is when you start doing tests that are so big that they start affecting people across international borders or you want to conduct these tests in the oceans. These tests are controversial enough that it would be good to wait until there's a little more consensus that they need to be done.

Who can be trusted with this? There are companies looking to invest in this research, can they be trusted?
It's a slippery slope. Take the cloud whitening approach. It might be if you had a bunch of ships lightening clouds off the coast of Los Angeles and San Diego, it would blow cooler, moister air over the Desert Southwest of the U.S. There could be a big economic value to producing lighter clouds and there might be a role for private companies in that endeavor. But it's important to develop consensus about how these different issues will be handled before companies go out there doing their own field tests.

… I would like to see government investment. My druthers would be for the [U.S. National Science Foundation] to take leadership in this scientific investigation because I think a lot of the best science is really driven by university-based researchers motivated by the pursuit of truth and not any single outcome.

But can we trust the government?
I don't think we can trust the government. It's willing to kidnap and torture people and engage in secret wars…. [But] peer-reviewed open inquiry promoted by the normal scientific method is an institution that is really wonderful. It's not perfect in that bad papers get published and sometimes results are kept secret that should be more open.  Keeping this research in publicly funded, non-classified, open journals is very good.

What is the significance of the recent Asilomar conference on geoengineering?
… The Council on Foreign Relations held a meeting two years ago in [Washington], D.C., on governance of these technologies. Another group held a meeting in Lisbon last year on similar issues. The Royal Society is holding a meeting this fall. There is an ongoing conversation about how to govern these technologies. I see this meeting as a continuation and broadening of that discussion. I have no issue with what happened at the meeting at all. [Caldeira did not attend, because of concerns about the financial interests of the organizers.] These issues are complex and are not going to be resolved in a few days.

Ultimately, do you think we'll ever use these ideas?

Nobody can foresee the future. I have friends who think it's almost certain that these kinds of systems will someday be deployed. My hunch is that we will never deploy these systems.

But I think the stakes are high enough that we need to understand whether these systems can help in the event of a climate crisis…. Experts are very bad at predicting the future. All we can do is develop options that we think can reduce risk under a wide range of scenarios. There's a non-zero chance that terrible outcomes might happen [as a result of climate change]. We should prepare for them. Because the cost of catastrophe is so large and the cost of developing [geoengineering] is so relatively small.

Whether it's ever deployed? I hope not.