It's a highly controversial view. After all, geoengineering is fraught with risks. Things could go badly awry. And Keith is upfront about those dangers — indeed, he has called for an international moratorium on deployment until we understand the technology better. But geoengineering may be the best way to limit some of the damage from the carbon-dioxide we've already put in the atmosphere. At the very least, he says, scientists need to start researching the idea more thoroughly.
Keith, who left the University of Calgary in 2011 for a position at Harvard's School of Engineering and Sciences, talked with me this week about his new book.
Brad Plumer: So, what is the basic case for geoengineering?
David Keith: First — and this is going to sound very pedantic and academic — but it's important to say what you actually mean by "geoengineering."
I would prefer if we mainly talked about the ideas that are called "solar radiation management," or "solar geoengineering," which involve cooling the earth by reflecting away some sunlight. All of these ideas are scientifically quite similar; they're all risky; they all act very fast; and they're all cheap.
There's another set of ideas around geoengineering the carbon cycle [i.e., figuring out ways to take carbon out of the atmosphere]. I'm not saying this approach is better or worse, but it's completely different and it doesn't have the same acute challenges of governance.
BP: Okay. So, reflecting sunlight to cool the planet -- why should we consider this?
DK: The best case for taking solar geoengineering seriously is that the balance of scientific evidence we have — from the same kind of climate models and other science that we use to understand climate change — suggests that these technologies could, if used carefully, significantly reduce climate risk. Full stop.
These technologies appear to provide a pathway by which we could substantially reduce climate risks over the next half-century. That means reducing the risks of sea-level rise, reducing the risks of stress for the crops of people in the poorest and hottest parts of the world.
This would be complementary to emissions reductions. Nothing changes the fact that in the long run, the only way to manage carbon risk is to stop emitting carbon-dioxide. But, similarly, nothing we know about cutting carbon-dioxide emissions says that's going to help us deal with the risk of CO2 that's already in the atmosphere, or deal with climate risks in the very short term.
So you can look at cutting emissions as a long-term solution and geoengineering as more of a short-term solution.
BP: There are all sorts of ways we could reflect sunlight to cool the planet. Volcanoes do this by putting sulfate particles in the atmosphere. What are the most realistic options here?
DK: There are a very broad range of things you could put in stratosphere. There are sulfates, which mimic nature, so we have lot of confidence about how they would act, though we also know a lot of the disadvantages there. And then there are other types of engineered particles that might work better, but we also know less about them.
The other idea that's probably most prominent is that we could add fine sea salt spray to certain kinds of marine clouds and make them a little whiter.
Often people get caught up in scientific reporting about all the different methods here. But they're all kind of similar. The hard questions aren't technical but rather about how to manage the risk, about who makes the decisions to use these technologies.
BP: Let's talk about those risks. What are the biggest ones?
DK: We can say what the technical risks are. Putting sulfates in the stratosphere can accelerate the depletion of ozone that comes from the chlorine that we've already put there from CFCs. It could change atmospheric circulation in ways that are hard to predict exactly. The sulfates could also make their way down to the lower atmosphere, where they'll contribute to air pollution. It's a small contribution proportionally, but that doesn’t make someone who's sick because of it feel better.
The bigger risks have to do with misuse. People often talk about using these technologies to return temperatures to pre-industrial levels. If you did that, that would be a dramatic climate cooling, with bad consequences, like reducing precipitation a lot.
And the most fundamental risk is that the technology has enormously high leverage, by which I mean it's cheap and a single small nation could use it by itself. We lack even the basics of how to create a norm of behaviors around this technology, let alone a treaty to make decisions about how to set the thermostat.
BP: Isn't there a concern that focusing on geoengineering will distract people from the task of cutting carbon emissions? What do you say to those critics?
DK: First, I think that's a completely valid concern. I think I was the first person to introduce the term "moral hazard" into this debate some 15 years ago. If we get to the point where these technologies are used and seem to be working, I think it is likely that it will distract somewhat from cutting emissions.
But while it’s a concern, I don't think it's a reason to say no. When you introduce seat belts, people tend to drive a little bit faster. When you introduce AIDS drugs, people may have riskier sex, at least some of them. But I think it would be perverse to say that because there are these risk-compensation behaviors that you shouldn't introduce some sort of risk-reducing technology.
It's important to be clear about the moral imperative here. Nothing plausible we do to reduce emissions in the next, say, quarter-century is going to materially reduce the risks for real people, especially some of the poorest and most vulnerable on our planet from climate change.
So, yes, the potential moral hazard is a major problem. But the fact that it’s a major problem is hardly an argument for foregoing a technology that might substantially reduce risk for those living now.
BP: The Intergovernmental Panel on Climate Change's most recent report made a similar point. It said that even if we cut emissions drastically today, the benefits will take awhile to materialize. Sea levels will keep rising for the next 50 years no matter what. Our only choice is between a big rise or an even bigger rise:
DK: And will continue for half a millennium. Even if we stopped all emissions today, you're looking at a change that goes on for centuries.
BP: Right. So is the basic idea that geoengineering would help us deal with some of the short-term climate problems, while curtailing carbon emissions is necessary to deal with the broader set of long-term problems?
DK: Yes. But "dealing with it" doesn't mean dealing with it perfectly. For example, solar radiation management technology does nothing to manage the carbon that's changing the acidity of ocean waters, or so-called ocean acidification.
It's also possible that over a centuries-long timescale we could find safe technological means to remove carbon-dioxide from the atmosphere at scale. (And full disclosure: I work on a company related to that.) That does deal with what you might call the root problem. But solar geoengineering technologies don't do that. They only reduce near-term risks, or at least they seem to.
BP: So your argument is that we should at least start researching geoengineering more seriously?
DK: There are some very thoughtful reasons to be cautious here. But I think ignorance is dangerous. If you choose not to research, that means we may later make decisions in ignorance and in a rush, and that’s typically how you make bad decisions. My view is that we would be better off to know more.
BP: At the same time, you've also called for a moratorium on large-scale deployment of stratospheric spraying. Why?
DK: I think there are very legitimate concerns that there's going to be this inevitable slide from research to deployment. That there will be an institutional lock-in. And we need to think about ways to stop that, and one of those ways is an international moratorium against deployment.
BP: So what does geoengineering research actually look like?
DK: A lot of the research program would look like an application of normal science, in the sense that there's a huge amount of scientific knowledge about climate and about aerosols and fine particles from air pollution research and other research. So part of it is taking that body of knowledge and applying it to this problem.
Currently there aren't any organized federal research programs here, although there are some programs in Europe, which is surprising given our expectation that America is more likely to embrace technological fixes than Europe.
I also think there's a danger of group-think, especially with a highly uncertain technology like this. It would be very unwise to have a single coordinated research program. Right now, there's a small group of people who work on this topic, the "geonerds," and we've convinced ourselves of a broad set of facts. But we might be wrong! And if I were organizing a research program, I would work hard to get a red team, blue team effect. Have researchers who are trying to figure out exactly why this wouldn't work.
BP: How would we conduct actual experiments into geoengineering?
DK: There's a host of experiments that could be done at a very small scale. They don’t actually attempt to perturb the climate, but they do attempt to investigate some of the key processes that we need to know about to understand how this would work.
So, Professor James Anderson and I here at Harvard are working on developing experiments — and they're not yet funded, so they're not necessarily going ahead — that would look at ways in which sulfate aerosols interact with ozone chemistry in the stratosphere. But the amount of sulfur would be in the kilograms, and the climate impact would be less than that from a few minutes of an airplane flight.
BP: Can you actually get useful data from a small experiment like that? Or will there be things we'll never know unless we actually try it on a large scale?
DK: Yes and yes. We think there is very useful data to be drawn about exactly how these chemicals do or don't amplify ozone loss. But no amount of research is going to remove all uncertainties. There will always be uncertainties.
BP: You mentioned that the biggest questions are around the politics of geoengineering. Who would control the technology? Who gets to set the thermostat? How does this play out?
DK: We had an essay in Science recently about this question. The answer is that people really don’t know. This is a new technology. It can be used cheaply; anyone can deploy it. It can be used for war. It's a lot like the type of power you get with a nuclear weapon.
But there a lot of specific ideas for governance. If we’re ever going to have actual deployment — and we're quite a ways away from that — then we're probably going to need governance that's at level of the U.N. Security Council or something like that. Whether or not the U.N. is the right mechanism, nobody knows.
There are obvious questions about who gets to weigh in on decisions to use geoengineering. But that's true of any high-leveraged technologies. New genetic technologies or nuclear weapons raise many of the same issues. But that doesn't mean we necessarily fail. We have had nuclear weapons for more than half a century now. And we have had some very close calls, but we haven't blown it. We have managed smallpox stocks and various other diseases without disasters. So I think there are realistic prospects for getting it right.
BP: At the beginning you mentioned technologies that could, conceivably, suck carbon out of the atmosphere. That sounds much less controversial. Why can't we just focus on that?
DK: I think you can more or less say that doing that quickly is impossible. To take carbon out of the atmosphere, you have to materially manage this enormous volume of material. We're talking hundreds of billions of tons of carbon. I think it’s close to impossible to think there will be some magic easy way to do that.
Interview has been lightly edited for length and clarity.