Skip to Content
Climate change and energy

This Scientist Is Taking the Next Step in Geoengineering

Harvard’s David Keith explains why it’s time to move forward with outdoor experiments and broader research programs.
July 26, 2017

Harvard professor David Keith has done as much as any single researcher to push the touchy topic of geoengineering toward the scientific mainstream (see “A Cheap and Easy Plan to Stop Global Warming”).

He was among the first to seriously assess potential ways of altering the climate to ease global warming, and he has undertaken some of the most detailed research on a promising approach known as stratospheric injection. He also wrote a book on the subject, A Case for Climate Engineering, and co-manages a Bill Gates–backed energy and climate fund that has supported research in this area. This year, Keith helped launch Harvard’s Solar Geoengineering Research Program, and announced plans with a colleague to carry out what would be among the earliest outdoor experiments in the field (see “The Growing Case for Geoengineering”).

The basic idea behind stratospheric injection is that spraying particles high above the Earth could help reflect more heat back into space, offsetting rising temperatures. It would mimic a natural phenomenon that occurs when large volcanic eruptions blast sulfur dioxide into the atmosphere, which nudges down global temperatures in the months that follow.

For the proposed experiment, Keith and fellow Harvard professor Frank Keutsch plan to launch high-altitude balloons that would spray small amounts of materials such as sulfur dioxide, alumina, or calcium carbonate into the stratosphere. They would then employ sensors to measure the reflectivity of the particles, the degree to which they disperse, and how they interact with other compounds. Initial test flights could occur as early as next year.

Listen to a conversation with David Keith on using geoengineering to address the threat of climate change.

Keith, a professor of applied physics and public policy at the university, sat down with MIT Technology Review earlier this year to discuss the upcoming experiments, and broader issues surrounding geoengineering. We’ve included highlights from that interview in the audio player below; the text follows.

Keith will also be a speaker at the publication’s EmTech 2017 conference, November 6–9 at the MIT Media Lab in Cambridge, Massachusetts, where we’ll continue the conversation on geoengineering and other means of addressing the growing risks of climate change.

Why is it time to move forward with geoengineering field trials?

I think it’s time to move forward with a broad research program on solar geoengineering, one that is open-access and international and transparent. And that’s because there is a real chance—we don’t know—that it might be able to substantially reduce climate risk this century. But it raises hard governance challenges, and we don’t know the risks or how well it works. We need to know that so we can give the information to the next generation so they can make good decisions. That’s the fundamental reason to have a research program. 

The answer to why do outdoor experiments is because that’s a normal part of research. In any kind of environmental research, you have to do a mixture of building careful theoretical models and then going out and making careful controlled experiments to understand where those models go wrong. The big issue here is really to understand how we’re wrong, and you do that ultimately by setting up carefully controlled experiments that are quantitative, where you can see errors in your prediction.

And from my point of view, it’s wrong to think about this as a trial of geoengineering. What a field trial would mean to me is that you had the full system you wanted to deploy, and you were beginning to deploy it in some trial mode to see if it worked. That is absolutely not what we’re doing. At this point, it’s much too early to think about engineering a complete system for deployment. I would oppose that. We’re trying to do science to help us understand how well some solar geoengineering ideas might work, how they might fail, what their risks might be. 

Even geoengineering criticsor at least credible geoengineering critics, I guessseem to agree that [at this small scale], the trials being discussed don’t pose any significant environmental risk. But one of the main criticisms that remains, and the one that you hear about a lot, is the idea of “moral hazard” — or I think the term you prefer is “risk compensation.” This basic idea that people and policy makers will see these serious, well-credentialed scientists doing this work and say, “Oh, well, they’ve got this,” and therefore we don’t need to take greenhouse gas reductions as seriously. What's your response to that argument? 

First of all, it’s important to say that that argument doesn't apply to experiments any more than it does to speech. I think at some level it’s a realistic risk that if the scientific community overstates the case that solar geoengineering really could work, there is a chance that it will be used politically to weaken efforts to cut emissions. I think that’s a sensible thing to worry about. But I don’t think there’s any bright line between a small science experiment and, say, a movie or a big public talk or, for that matter, the former chief advisor to the [UN] secretary general starting a major effort on governance. All those things make this look more serious, and I don’t think these concerns apply in a special way to experiments, which in some sense don’t physically do anything different than existing scientific experiments. 

In thinking about the moral-hazard argument, you have to weigh two things. Yes, there is some legitimate possibility that it would reduce commitments to cut emissions, but there’s also some possibility that it could substantially reduce climate risk over this century. And in the end, we don’t make that decision on either thing. This is about giving more information to the next generation that will make the serious decisions about this—not us. 

You made an interesting point in your book on this issue. You basically said that the moral-hazard argument is probably correct, as you were just saying, but that we should pursue research anyway because the risks of climate [change] are so high. And specifically those risks, which can be kind of treated as sort of an abstraction in the moral-hazard argument, basically boil down to a lot of death and destruction in poor countries. Can you explain that point? 

Sure. One of the reasons I’m particularly interested in developing the possibility of solar geoengineering is that it does appear that the benefits are most felt by the poorest. And that’s because the biggest climate impacts—particularly impacts from extreme heat and extreme precipitation events like tropical cyclones—fall on the world’s poorest. And there is now pretty clear evidence that solar geoengineering would be remarkably effective in reducing some of those risks, and the relative benefits [would] actually go more to the poor than to the rich. For me it’s a fundamental ethical reason that we do need to develop the technology to do it, and that we need to engage deeply with people in the developing world, in getting their input into what this development looks like and in diffusing the technology. This may be being developed here, but it’s open-access, and I don’t actually think it’s likely that the U.S. will be the country that deploys. It’s much more likely to be poor countries. 

The argument that resonates the most with me, not against research but against eventual deployment, is just the idea that the climate system is massive and complex and uneven, and we know our climate models aren’t perfect. So we might be pretty sure [that stratospheric injection] will work, and we might be pretty sure that the negative consequences won’t be that bad or can be limited, but we can’t really know for sure what the full effects will be. I’m curious for you yourself, when you get to the point of deployment in your own imagination, does that give you pause? 

Yes. But I think that you have to think both sides of the argument through. If you don’t do it, then you have what’s called radiative forcing, which is the cumulative effect of all that CO2 in the atmosphere pushing the climate.

We have uncertainty on both sides. But the question is: which uncertainty is larger? A kind of most obvious scientific answer, balance of evidence, is that the uncertainty goes up with the total radiative forcing. The harder you push the climate system, the more you get these unexpected, nonlinear, frightening outcomes. So yes, it’s true that we don’t know for sure what the response to solar geoengineering is; we never will. But we also don’t know what the response to CO2 is, and we never will. And the issue is: which combined state of the world is less risky?

With whatever [level of] CO2 we peak at, the question is: Would you rather have that peak alone and not know for sure what’s going to happen, or have that peak alone and a little bit of solar geoengineering, so it is a little less total climate forcing? And we’ll never know. Both are uncertain. But it sure looks like it’s less risky to have a little bit of solar geoengineering. 

Since it is and always will be this kind of a risk-to-risk decision, does that ultimately mean that full-scale deployment doesn’t make sense [outside] of the context of unfolding disasters? Or at least right up to the edge of what we think is about to be an unfolding disaster?

Climate is a slow-moving beast, so there will be individual things that are locally disastrous, like Category 5 [hurricanes]. And now it might be possible with enough CO2 to get Category 6 hurricanes, or a heat wave that kills thousands or tens of thousands of people. Those things, when you’re in them, are disasters.

But all that climate change or solar geoengineering does is alter the probability of those things slowly and gradually over time. While there are political tipping points, it’s not so obvious there really are strong physical tipping points in the overall Earth’s climate.

In general, I think if you’re going to use solar geoengineering, it makes more sense to begin using it very slowly and gently, so you can look for errors and look for problems as you gradually increase it, and you can begin to reduce risks slowly. So I don’t think it’s that useful as an emergency response. But at some level I don’t really know what a climate emergency is. It’s partly just an effective turn of speech. 

Even if that’s the best route scientifically to take, is the thing I just described the more likely politically feasible thing? 

Yeah, yeah. The way politics works in democracies, and at some level globally, I think, is you get these things moving along under the surface, where some subset of the, you know, elite has a set of different issues around tax reform, or obscure pieces of IP reform, or solar geoengineering. Those things kind of stay beneath the surface of people fighting their turf wars, and thinking about different regulatory systems. Then some event suddenly brings it to the surface, and then the politicians act. So, if you’re talking about school bus safety, maybe there’s some perfectly good arguments, but then one day there’s an awful school bus accident, and that happens to be when legislators act.

I’m sure the same will be true here—that action around solar geoengineering is likely to be triggered by some kind of big event, like a massive killing heatwave. But that doesn’t mean the action is just a panicked response to that event. What I hope, and the reason to have research, is that there will be as much background knowledge as possible to make informed decisions when that happens.  

Deep Dive

Climate change and energy

Super-efficient solar cells: 10 Breakthrough Technologies 2024

Solar cells that combine traditional silicon with cutting-edge perovskites could push the efficiency of solar panels to new heights.

The race to get next-generation solar technology on the market

Companies say perovskite tandem solar cells are only a few years from bringing record efficiencies to a solar project near you.

How one mine could unlock billions in EV subsidies

The Inflation Reduction Act is starting to transform the US economy. To understand how, we tallied up the potential tax credits available as the nickel from a single mine flows through the supply chain.

Heat pumps: 10 Breakthrough Technologies 2024

Heat pumps are a well-established technology. Now they’re starting to make real progress on decarbonizing homes, buildings, and even manufacturing.

Stay connected

Illustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at with a list of newsletters you’d like to receive.