Here is the plan. Customize several Gulfstream business jets with military engines and with equipment to produce and disperse fine droplets of sulfuric acid. Fly the jets up around 20 kilometers—significantly higher than the cruising altitude for a commercial jetliner but still well within their range. At that altitude in the tropics, the aircraft are in the lower stratosphere. The planes spray the sulfuric acid, carefully controlling the rate of its release. The sulfur combines with water vapor to form sulfate aerosols, fine particles less than a micrometer in diameter. These get swept upward by natural wind patterns and are dispersed over the globe, including the poles. Once spread across the stratosphere, the aerosols will reflect about 1 percent of the sunlight hitting Earth back into space. Increasing what scientists call the planet’s albedo, or reflective power, will partially offset the warming effects caused by rising levels of greenhouse gases.

The author of this so-called geoengineering scheme, David Keith, doesn’t want to implement it anytime soon, if ever. Much more research is needed to determine whether injecting sulfur into the stratosphere would have dangerous consequences such as disrupting precipitation patterns or further eating away the ozone layer that protects us from damaging ultraviolet radiation. Even thornier, in some ways, are the ethical and governance issues that surround geoengineering—questions about who should be allowed to do what and when. Still, Keith, a professor of applied physics at Harvard University and a leading expert on energy technology, has done enough analysis to suspect it could be a cheap and easy way to head off some of the worst effects of climate change.

According to Keith’s calculations, if operations were begun in 2020, it would take 25,000 metric tons of sulfuric acid to cut global warming in half after one year. Once under way, the injection of sulfuric acid would proceed continuously. By 2040, 11 or so jets delivering roughly 250,000 metric tons of it each year, at an annual cost of $700 million, would be required to compensate for the increased warming caused by rising levels of carbon dioxide. By 2070, he estimates, the program would need to be injecting a bit more than a million tons per year using a fleet of a hundred aircraft.

One of the startling things about Keith’s proposal is just how little sulfur would be required. A few grams of it in the stratosphere will offset the warming caused by a ton of carbon dioxide, according to his estimate. And even the amount that would be needed by 2070 is dwarfed by the roughly 50 million metric tons of sulfur emitted by the burning of fossil fuels every year. Most of that pollution stays in the lower atmosphere, and the sulfur molecules are washed out in a matter of days. In contrast, sulfate particles remain in the stratosphere for a few years, making them more effective at reflecting sunlight.

The idea of using sulfate aerosols to offset climate warming is not new. Crude versions of the concept have been around at least since a Russian climate scientist named Mikhail Budkyo proposed the idea in the mid-1970s, and more refined descriptions of how it might work have been discussed for decades. These days the idea of using sulfur particles to counteract warming—often known as solar radiation management, or SRM—is the subject of hundreds of papers in academic journals by scientists who use computer models to try to predict its consequences.

But Keith, who has published on geoengineering since the early 1990s, has emerged as a leading figure in the field because of his aggressive public advocacy for more research on the technology—and his willingness to talk unflinchingly about how it might work. Add to that his impeccable academic credentials—last year Harvard lured him away from the University of Calgary with a joint appointment in the school of engineering and the Kennedy School of Government—and Keith is one of the world’s most influential voices on solar geoengineering. He is one of the few who have done detailed engineering studies and logistical calculations on just how SRM might be carried out. And if he and his collaborator James ­Anderson, a prominent atmospheric chemist at Harvard, gain public funding, they plan to conduct some of the first field experiments to assess the risks of the technique.

Leaning forward from the edge of his chair in a small, sparse Harvard office on an unusually warm day this winter, he explains his urgency. Whether or not greenhouse-gas emissions are cut sharply—and there is little evidence that such reductions are coming—”there is a realistic chance that [solar geoengineering] technologies could actually reduce climate risk significantly, and we would be negligent if we didn’t look at that,” he says. “I’m not saying it will work, and I’m not saying we should do it.” But “it would be reckless not to begin serious research on it,” he adds. “The sooner we find out whether it works or not, the better.”

The overriding reason why Keith and other scientists are exploring solar geoengineering is simple and well documented, though often overlooked: the warming caused by atmospheric carbon dioxide buildup is for all practical purposes irreversible, because the climate change is directly related to the total cumulative emissions. Even if we halt carbon dioxide emissions entirely, the elevated concentrations of the gas in the atmosphere will persist for decades. And according to recent studies, the warming itself will continue largely unabated for at least 1,000 years. If we find in, say, 2030 or 2040 that climate change has become intolerable, cutting emissions alone won’t solve the problem.

“That’s the key insight,” says Keith. While he strongly supports cutting carbon dioxide emissions as rapidly as possible, he says that if the climate “dice” roll against us, that won’t be enough: “The only thing that we think might actually help [reverse the warming] in our lifetime is in fact geoengineering.”

The Experiment

David Keith clearly sees the world through the eyes of an experimental physicist. During his time as a graduate student in the MIT lab of David Pritchard, he spearheaded a project that built the first atom interferometer. Keith and his coworkers outcompeted some of the world’s top atomic-physics labs, including one at Stanford led by Steven Chu, who later won a Nobel Prize and served as the U.S. secretary of energy. Everyone knew the interferometer would be a breakthrough, recalls Pritchard, but Keith displayed a rare combination of creativity and the ability to “blast ahead” through the frustrations and difficulties of building and testing it. Keith, however, says his remarkable achievement caused him to “walk away from [atomic] physics,” in part because one of the most obvious applications for atom interferometry was in highly accurate gyroscopes for submarines carrying ballistic missiles.

Soon, Keith had moved on from the esoteric world of atomic physics to energy problems. In 1992, he published a paper called “A Serious Look at Geoengineering,” one of the first rigorous scientific reviews of the topic. Almost no one cared.

Indeed, the field of geoengineering remained more or less dormant for much of the next decade. A handful of serious scientists wrote occasional papers and the field attracted a robust fringe of fanatics, but academic discussion of the subject—let alone actual research—remained somewhat taboo. Many felt that discussing geoengineering as a realistic option would take attention away from the urgency of cutting greenhouse-gas emissions. Then, in 2006, Paul Crutzen, one of the world’s leading climate scientists and a winner of the 1995 Nobel Prize in chemistry for his work on atmospheric ozone depletion, published a paper called “Albedo Enhancement by Stratospheric Sulfur Injections: A Contribution to Resolve a Policy Dilemma?”

In the paper, Crutzen acknowledged that the “preferred way” to address climate warming was to lower emissions of greenhouse gases, but he concluded that making sufficient cuts was only “a pious wish.” Not only did he give his blessing to the idea of geoengineering, but he singled out the use of sulfate aerosols in particular as worthy of research, even though it’s well known that the particles can facilitate the chemical reactions that lead to ozone loss. He pointed to the eruption of Mount Pinatubo on an island in the Philippines in 1991 as evidence that sulfate particles can effectively cool the planet. The giant volcano spewed some 10 million metric tons of sulfur into the stratosphere. Subsequent analysis showed that the world’s temperature decreased by an average of 0.5 °C for a couple of years.

At a time when many experts were increasingly frustrated with the lack of progress in cutting greenhouse gases, the paper permitted the topic of intentional climate alteration to be more openly discussed. In subsequent years, geoengineering gained still more attention, including high-profile reviews by the U.K.’s Royal Society and the Washington-based Bipartisan Policy Center, both of which recommended further exploring SRM. (Keith helped write both reports.) Endless modeling and computer simulations have followed. But now Keith is anxious to conduct field experiments.

That idea is highly controversial. Many climate scientists still consider field experimentation premature, and critics of geoengineering tend to believe it would be the first step in what would turn into an inexorable move toward full-scale deployment. Last year, a public outcry led by several international environmental groups helped shut down a simple experiment that a team of British researchers had proposed. The group wanted to pump water to a height of one kilometer through a thin hose held aloft by a helium balloon. The object would have been to test whether a similar system could someday be used to inject sulfur particles into the stratosphere at an altitude of 20 kilometers.

The experiments Keith and Anderson are considering would be far more ambitious. Their goals: first, to test how sulfuric acid should be distributed to optimize the size and longevity of the resulting particles, and second, to measure how sulfur affects ozone at the altitude and under the conditions associated with SRM.

Anderson, who helped unravel the chemistry behind the ozone hole that appeared in the Antarctic during the 1980s, says the “demonic system” that implicates sulfate particles in ozone destruction is highly sensitive to the levels of water vapor in the air. So in one set of experiments, using a scheme based on Anderson’s earlier work, the group would send a helium-filled balloon to the lower stratosphere, use a Kevlar thread to lower canisters filled with water vapor and sulfur, and release small amounts of the test samples. Then the researchers would drop down miniature laser-based analytic instruments to monitor the chemistry in the small “seeded” area. The setup, says Anderson, provides “exquisite control” and a way to precisely monitor the effect of different amounts of sulfur and water vapor.

Anderson stresses that the experiment would have no conceivable impact on the stratosphere: it would use only “micro-amounts” of sulfur and would be confined to a very small region. And he says it is critical to study the reactions under the conditions “where they actually take place” and not in the confines of the lab.

Still, while he is keen to test SRM, Anderson says that adding sulfates to the stratosphere worries him “tremendously” because of the potential impact on ozone. He points to a study his group published last year in Science showing that increasingly intense summer storms over the United States—triggered by climate warming—are injecting more water vapor into the stratosphere. That, he says, could speed the ozone-destroying reactions: “If nature is adding increased water vapor to the stratosphere and we’re adding sulfates, it is a very lethal cocktail for ozone loss.”

Keith appears more sanguine. “The uncertainties are substantial,” he says. “You could get very bad [ozone] outcomes, but there are also ways where you could have no impact, or even a positive impact, on ozone.” In any case, he says, it is “just crazy” not to begin conducting experiments on solar geoengineering to find out. Nearly all the work done on SRM is based on computer modeling, and Keith says we need to move to “perturbation experiments” to learn whether we can use it to safely and effectively intervene in the climate. The field “really needs to grow up” and begin experiments in “the real world,” he says.

Barking Mad

Critics of SRM—and even its advocates—note that the technology has numerous limitations, and that no one is entirely sure what the consequences would be. Sulfate aerosols reflect sunlight in the upper atmosphere, thus directly cooling the planet. But greenhouse gases operate very differently, trapping long-wave infrared radiation escaping from Earth’s surface and thus warming it. While sulfates would be likely to offset warming, it’s not clear exactly how they would counteract some of the other effects of greenhouse gases, particularly changes in precipitation patterns. And SRM would do nothing to reduce the acidification of the oceans caused by rising levels of carbon dioxide in the atmosphere.

“The term ‘solar radiation management’ is positively Orwellian,” says ­Raymond Pierre­humbert, a geophysicist at the University of Chicago. “It’s meant to give you a feeling that we really understand what we would be doing. It’s a way to increase comfort levels with this crazy idea. What we’re really talking about is hacking the planet in a case where we don’t really know what it is going to do.” In delivering the prestigious Tyndall Lecture at the annual American Geophysical Union meeting last December, he said the idea of putting sulfate aerosols in the stratosphere was “barking mad.”

Pierre­humbert also rejects the value of doing field experiments. “The whole idea of geoengineering is so crazy and would lead to such bad consequences, it really is pretty pointless. We already know enough about sulfate albedo engineering to know it would put the world in a really precarious state. Field experiments are really a dangerous step on the way to deployment, and I have a lot of doubts what would actually be learned.”

The fundamental problem with albedo engineering, says Pierrehumbert, is that once we start using it, we’ll need to continue indefinitely. Since it only offsets warming, once the process stops, temperature changes caused by greenhouse gases will manifest themselves suddenly and dramatically. “If you stop—or if you have to stop—then you’re toast,” he says. Even using it as a temporary Band-Aid doesn’t make sense, he argues: “Once you get to the point in terms of climate changes that you feel you have to use it, then you have to use [SRM] forever.” He believes that this makes the idea a “complete nonstarter.”

Besides, Pierrehumbert says, our climate models “are nowhere near advanced enough for us to begin thinking of actually engineering the planet.” In particular, computer models don’t accurately predict specific regional precipitation patterns. And, he says, it’s not possible to use existing models to know how geoengineering might affect, say, India’s monsoons or precipitation in such drought-prone areas as northern Africa. “Our ability to actually say what the regional climate patterns will be in a geoengineered world is very limited,” he says.

Alan Robock, meanwhile, has a long list of questions concerning SRM, at the top of which is: can it even be done? Robock, an expert on how volcanoes affect climate and a professor of environmental sciences at Rutgers University, cautions that while the Pinatubo eruption confirmed the cooling effect of sulfate aerosols, it injected a massive amount of sulfur dioxide into the stratosphere over a few days. Solar geoengineering would use far less sulfur but disperse it continuously over an extended period. That could be a critical difference. The optimal way to achieve SRM is with sulfur particles only about half a micrometer in diameter. Sunlight reflects off the surface of the particles, and smaller particles have more surface area than larger ones, making them far more efficient at blocking the sun. Robock worries that as sulfur is continuously injected and concentrations build up, the small particles will clump together into large ones, necessitating far more sulfur than some current proposals assume.

These details of aerosol chemistry could help determine the viability of SRM. “David [Keith] thinks it is going to be easy and cheap, and I don’t agree,” says Robock. He estimates that several million tons of sulfur would have to be injected into the atmosphere annually to offset doubled levels of carbon dioxide, but if the particles clump together, “it could be many times that.”

Research so far shows that producing a cloud in the stratosphere—Robock’s preferred description of SRM—”could cool the climate,” he says. “But you would have a very different planet, and other things could be worse.” He points out, for example, that in the aftermath of Mount Pinatubo, rainfall decreased significantly in some parts of the world. Robock supports more modeling on solar geoengineering, but “right now, I don’t see a path in which it would be used,” he says. “I don’t see how the benefits outweigh the negatives.”

Still, climate scientists differ widely in the way they interpret the research on those risks. Phil Rasch, for one, who is chief scientist for climate science at the Pacific Northwest Laboratory in Richland, Washington, cautiously says the models do not yet indicate “showstoppers” that would preclude consideration of certain SRM strategies.

Rasch, who published a paper with Crutzen in 2008 on using sulfate aerosols for geoengineering, says research shows that the particles will cause some ozone depletion—”it is absolutely something we need to pay attention to”—but that the loss of ozone is somewhat tempered by the ability of the sulfate particles to block ultraviolet radiation. As for rainfall, he says, models tend to agree that SRM “leads to a [future] world that’s closer to the present day with respect to precipitation than if you don’t geoengineer.” Overall, says Rasch, SRM would stave off some effects of climate change, though “some parts of the planet are more strongly affected than others, and there are many issues that remain unexplored.”

A Moratorium

The scientific uncertainties and the prospect of winners and losers among different parts of the world make it almost unfathomably difficult to envision how SRM might be appropriately implemented and controlled. How could we fashion the international system of governance that would eventually be needed? Who would decide how and when to implement the technology? Who would monitor and control it? Who would set Earth’s thermostat and at what temperature? If anything, the questions about who would make the decisions on solar geoengineering are more daunting than the questions about the science itself.

While the need for international governance is still years in the future, Keith and several close collaborators, including Edward Parson, a law professor at the University of California, Los Angeles, are already thinking about how such a system might evolve. Research on the technology is key, Parson says, to achieving a better understanding of what solar geoengineering can do and what the risks are. Without such knowledge, he says, “you don’t know what you need to govern.”

The controversy over field experiments, such as the ones Keith and Anderson are designing, is emerging as an early battleground for the social and political issues. Keith is adamant that work will not go forward unless he and his colleagues receive public funding and approval from established scientific agencies. Indeed, he and his collaborators see the experiments as an early test not only for the technology but also for how a governance system can work. The hope, says Parson, is that the funding and approval process could provide an opportunity to establish “norms” that will help shape longer-term discussions—standards such as transparency, public review, and open disclosure of the results.

No one thinks that field experiments involving tiny amounts of sulfur would be physically dangerous, says Parson. “What concerns people,” he says, “is the political and social consequences of the research going ahead, followed by bigger and bigger experiments—and then you’re on the slippery slope all the way to full-scale deployment.” These worries should be taken seriously, he says: “You need to encourage small-scale research, but you need some kind of limited governance to mitigate the risk of a slide to deployment.” Established scientific funding agencies could probably take care of that, he believes. And he suggests that early experiments must be strictly limited, and researchers need to clearly state that no one is going to do anything big for the time being.

Keith and his collaborators are pushing fellow researchers to sign an agreement that would “function like a moratorium” on deploying solar engineering. That, Keith believes, could calm fears that some are rushing ahead on the technology—worries that he concedes are “not ungrounded,” since there are, in fact, no international laws or regulations barring anyone from implementing geoengineering schemes. By signing a moratorium, he hopes, scientists could “help free up research” on the risks and efficacy of SRM.

Switching It On

For very brief spells, Keith sometimes lapses into animated annoyance with SRM critics. A moment later, however, he is calmly and logically countering the criticism with responses he has developed after years of thinking and writing about geoengineering. He sketches a graph showing that, in fact, sulfur injection could be rationally ended a century or less after it’s begun; while the underlying climate changes it was masking would return, the rate of change affecting ecosystems and humans would have been slowed and managed. The idea that initiating SRM would commit us to continuing it indefinitely “is just not true,” he states with characteristic self-confidence.

Even many of the strongest advocates of SRM research say the technology would be a nearly unthinkable last resort for a desperate world facing climate changes so destructive that the risks would be worth taking. Keith, however, has a far less apocalyptic vision. “If we’ve actually found something that could substantially reduce the risk of climate change over the next century and save a lot of lives, that’s nothing to be upset about,” he says. “It’s something to celebrate.” In fact, he says framing the case for geoengineering as a last resort in a climate emergency is “a bit of a rhetorical trick”: it leaves undefined what a “climate emergency” is, and “there is no simple definition.”

The approach Keith proposes is at once more deliberate and far more radical: “In my view, we should begin real research now, and if it bears out that [SRM] could meaningfully reduce climate risks without too many risks of its own—which may or may not turn out to be true—then we should actually begin doing this relatively soon, but with a very slow ramp.” He believes the technology could be ready to be deployed as early as 2020 (or, more realistically, 2030) and would involve levels of stratospheric sulfur “practically” within normal ranges for the first decade. The process could be monitored and evaluated, and because the amounts of sulfur injected into the stratosphere would be relatively small, “the chances of a big problem are pretty close to zero.”

It is often assumed that SRM would be “turned on with a big switch,” says Keith. “But there’s no reason you can’t ramp it up.” And that ability to turn on the system slowly and with minimal risk is behind his “willingness to take geoengineering seriously,” he says: “If it was a one-time decision, I would be much more skeptical about doing it. It would be very hard to persuade me that it was sensible.” Given the possibility of a more deliberate approach, “I lean pretty strongly, I got to say, to doing it.”

Listening to Keith’s logical arguments and careful descriptions of how SRM might be carried out, it’s just possible to start believing that intentionally adjusting the climate wouldn’t be an extreme action. But it would. It would create a different planet—even the color of the sky would be whiter. And it would almost certainly be driven by desperation. On the other hand, the buildup of greenhouse gases is already altering the atmosphere and climate in an unprecedented and uncontrolled manner. How big a leap is it to intentionally “engineer” ways to begin counteracting that? And Keith is surely right that climate researchers should explore solar geoengineering to determine whether it would actually work and how safe it would be, and that political scientists need to start thinking about how we might implement such an unprecedented planetary project. All that will be left then is for society and governments to face the impossibly difficult task of deciding whether to do it.