In this episode, host Daniel Raimi talks with David Keith, a physics and public policy professor at Harvard University and one of the world’s leading experts on solar geoengineering. Keith outlines prominent examples of solar geoengineering technologies in development—from ambitious ideas, such as sending aerosols to the stratosphere, to more operational solutions, like painting roofs white to reflect more sunlight into space. An advocate for expanding solar geoengineering, Keith discusses how coalitions among like-minded nations and clearer guidance from policymakers could expand deployment of promising technologies.
Listen to the Podcast
- Big shifts in climate are not the goal of solar geoengineering: “My view is, at least in this period where emissions are still positive and concentrations of CO₂ are going up, the more sensible way to think about these technologies is to limit the rate of warming, or to halt warming and other large-scale climate changes, but not to attempt to drive the climate back toward pre-industrial [levels].” (8:06)
- Why sweeping unilateral action in major geoengineering projects remains unlikely: “If you were a country … highly motivated to implement these technologies for self-protection, because of these sharp governance challenges and because of the interconnectedness of global affairs, I think you'd be unlikely to actually want to be purely unilateral, because you could expect blowback. Instead, you'd likely work out coalitions of like-minded countries that also wanted to act. Indeed, if you wanted this to be stable, which it would be in your self-interest to do … you'd want to think about a coalition that could have some legitimacy. On the other hand, I think the idea of a pure consensus—in which all countries agree—is very unlikely.” (20:13)
- Policy cannot keep ignoring solar geoengineering: “A reasonable position is to say there should be a global moratorium on [solar geoengineering] ever happening. It's not a position I agree with, but there’s reason for [that] point of view ... I think what is becoming increasingly untenable is the idea that we think about climate policy over the next decades in a way that just pretends solar geoengineering doesn't exist. I think that is an increasingly implausible position. We need to bring it into the core of the debate about what we do about climate risk.” (27:48)
The Full Transcript
Daniel Raimi: Hello, and welcome to Resources Radio, a weekly podcast from Resources for the Future. I'm your host, Daniel Raimi. This week, we talk with Harvard University professor David Keith about solar geoengineering. David will describe to us the variety of ways that solar geoengineering could work, as well as some of its risks at local, regional, and global scales. We'll also talk about the current state of play with regard to recent small-scale experiments, and what might be needed to deploy a larger scale research program. Last but not least, we'll talk about public policies related to the potential deployment of these technologies, including the substantial issues surrounding governance and geopolitics. And one quick note, this episode was recorded well before the extent of the coronavirus pandemic had become clear. Stay with us.
Okay. David Keith from Harvard University. Thank you so much for joining us today on Resources Radio.
David Keith: Thanks. Great to be here.
Daniel Raimi: David, we're going to talk today about geoengineering, and particularly a type of geoengineering called solar radiation management. We'll define those terms in just a minute and get into the details. First, we always like to ask our guests how they got interested in environmental issues in the first place. Can you tell us a little about your background and how you got into this world?
David Keith: I guess through family. My father and stepmother were environmental scientists, biologists, and I actually went out and did field work with them as a kid. My dad had a tie back to the kind of high world of international environmental policy. He went to the Stockholm meeting in '72, the original global environmental conference that led to Rio and so on.
Daniel Raimi: Very cool. What type of field work would you do with them? Were they gathering specimens or something?
David Keith: This was about DDT and organochlorines in migratory birds. And so one thing we would do is go out to a little Island in the middle of Lake Ontario, that was a herring gull colony, and collect eggs and specimens.
Daniel Raimi: Wow, very cool. Let's get into the topic at hand, which is geoengineering. As some of our listeners might know, you've been sort of all over the place in terms of podcasts lately. I know you've had lots of practice answering the first question, which is: When people say “geoengineering,” what do they typically mean, and do you have kind of a preferred definition of that term?
David Keith: This could get difficult. I don't know what people mean, it's moving. Nobody owns the words. There's no right answer. What I would say is there are four broad categories of things that humans can do to manage the climate problem. The first is decarbonization or mitigation, it's sometimes called, but basically all the ways we break the link between economic activity and emissions reductions. The second is carbon removal, or if you like carbon geoengineering, or negative emissions. That's a whole bucket of things, reasonably heterogeneous, that are all the ways in which humans might engineer a negative emissions kind of carbon banking.
Then, the third is what I call solar geoengineering, but it also could be called climate radiation intervention, or solar radiation modification, which is what the IBCC calls it. The fourth are local adaptation measures. Historically, there was a period where people lumped what I would call carbon removal, and what I would call geoengineering, into a single category called geoengineering. My view is that is unhelpful. It's not about whether they're good or bad, it's that they are just not tightly linked in either the science, the technology, or the public policy implications. In fact, carbon removal is much more tightly linked to mitigation. Parts of our removal are almost inseparable for mitigation or decarbonization, and solar geoengineering probably stands on its own as this weird thing. If it's linked to anything, it's probably more linked to adaptation.
Daniel Raimi: Great. Let's talk now a little more about solar geoengineering. I want to kind of start by focusing on some technical elements of what the technology could look like, and then get to the related and very complex, and potentially fraught questions of how and whether it might be deployed. Can you start off by giving us some idea of the technologies that are associated with solar geoengineering, and also some sense of their associated costs?
David Keith: Sure. I guess, moving down from far out in space to the surface, the first technology would be space-based systems. Perhaps, the idea that humans could actually build a shield, a giant structure between the earth and the sun, to block out say some half a percent or percent of the sunlight. This would be at the L1 Lagrange point between Earth and sun. The second would be putting aerosols, fine particles, in the upper atmosphere, the stratosphere. The third would be modifying cirrus clouds, the thin high clouds. It turns out that cirrus clouds tend to mostly warm the planet, because they have an infrared heat trapping, as well as a solar band reflection that tends to cool upon it. But the heat trapping can dominate, and there are ways one might reduce the total amount of cirrus clouds in some locations, that could let more infrared light out.
The fourth would be so-called, "marine cloud brightening." The idea that some kinds of marine stratus clouds, the kind of clouds you'd find, say off Seattle or off England, that those clouds could be seated, if you like, with fine sea salt aerosols, in a way that might make them brighter, or last a bit longer. Then, I guess the fifth category would be all the ways that we would manipulate the reflectivity of the land surface, from painting roofs white to putting reflective materials on sea ice.
Daniel Raimi: Right. Yeah. For me, when I'm kind of casually reading about this topic, the one that comes up the most frequently seems to be the aerosol-based approach. Am I right in thinking that's been where most of the research has been, or has it been all over the place?
David Keith: Yeah, most of the research I would say has been on stratospheric aerosols. To be clear, cirrus thinning and marine cloud brightening also involve adding aerosols. All three of those are aerosol-based approaches. Stratospheric aerosols have certainly been the largest research focus.
Daniel Raimi: Great. This next question is kind of a timing question. It's about how we should think about the relevant timeframes that we need to keep in mind when we're thinking about solar geoengineering. You might want to only refer to one or two of the specific technologies that you just mentioned, or maybe you want to give us a view on all five of them. Let's say, for example, that it was decided to undertake some type of major solar geoengineering intervention. If it was stratospheric aerosols, for example, if the technology were deployed, and it worked exactly as it was designed to, how quickly could something like that reduce temperatures, and how long could the effect be sustained?
David Keith: First of all, to somewhat push back on the framing of your question, nothing will ever work exactly as it was designed. There are no versions of any of these, where people will ever know precisely what the answers are. It gets back to kind of risk trade-offs and uncertainty. To add to the question for stratospheric aerosols, at least some methods of stratospheric aerosols—say putting sulfuric acid in the stratosphere—those methods could be implemented pretty quickly. I think a technologically advanced country could begin to put materials in the stratosphere in just a few years if it really wanted to.
Your question supposed that the right answer is to drive temperatures downward. Of course, nobody knows what the right answer is in a big range of views. My view is, at least in this period where emissions are still positive and concentrations of CO2 are going up, the more sensible way to think about these technologies is to limit the rate of warming, or to halt warming and other large-scale climate changes, but not to attempt to drive the climate back towards pre-industrial.
Daniel Raimi: Another sort of time frame question, that I often hear about when people talk about stratospheric aerosols, is the potential that the technology would need to be sort of continuously deployed in order for the effect to be sustained. Is that a useful way of thinking about it?
David Keith: All of these technologies are ways to alter the radiated balance of the earth. All of them inherently have a kind of short lifetime, a lifetime that ranges from hours to a day for cirrus thinning and marine cloud brightening, to a year or so for stratospheric aerosols. All of them have relatively short timescales. One would need to keep doing it in order to keep a given level of radiated force.
Daniel Raimi: Got it. Next question on the technical side of solar geoengineering, and then we'll get to some of the questions about governance and concerns that people have about it: How much do we know about the technical effectiveness of some of these approaches, such as stratospheric aerosol? Maybe you want to talk about some of the others. You mentioned a moment ago that, of course, we don't know with precision exactly how these things would turn out if they were deployed in the real world. How well is our understanding bounded about the potential range of effects of deploying some of these technologies?
David Keith: It's useful to think of that question in two really different parts. One part is understanding the risks and uncertainty in making the radiative forcing, and the actual trail between putting some material in the stratosphere, and some resulting radiative forcing, or putting some material in marine clouds and resulting radiative forcing. The second part of the question is understanding what is the climate response to that forcing.
On the first part, that varies a lot between technologies. For cirrus thinning or marine cloud brightening, we actually really don't know very well how much radiative forcing would get made by some interventions. It could even be in some cases radiative forcing has the opposite sign. On the other hand, lots of small experiments, much too small to have a climate impact, could quickly tell us a lot about that. We could monitor it in a kind of feedback sense. For stratospheric aerosols, I think there's actually very high confidence that stratospheric aerosols will make a kind of given radiative forcing, and there's ways that would feed back, one could, I think, effectively control the amount of radiative forcing. I think what I see as a good thing for governance about stratospheric aerosols, is that radiative forcing inherently, if you like, kind of wants to be pretty even, very even East to West, and pretty even North to South. It's relatively easy to get a pretty uniform global radiative forcing from stratospheric aerosols. That's the radiative forcing side.
The second part of this is what's the climate response? I think there's really very strong evidence from essentially every major client model that's been run on this, and most of them have. If one used a pretty globally uniform homogeneous radiative forcing, as could be achieved or nearly achieved with stratospheric aerosols, and you did it in a way that wasn't trying to drive the climate back to pre-industrial, but just to shave the peak of climate risk—to reduce the amount of risk—then the evidence seems to be that essentially every [area] in the world would see many, or nearly all of the major climate hazards, like extreme temperatures, changes in water availability, et cetera, see those changes reduced. It looks like that could be done without any large areas being made significantly worse off. That may or may not be true, but that's what we see in quite a range of different models.
Daniel Raimi: That leads nicely into the next question that I wanted to ask you, which again, I know could have a wide variety of potential answers. I welcome you answering it in the way or ways that you think are most appropriate. When we think about either one or multiple of these technologies, some of them—at least at first blush—would appear to be relatively low risk in terms of unintended consequences for the physical environment. If I think about painting white roofs, again at first blush, I can't imagine that would have the same types of potential unintended consequences that stratospheric aerosols could have. I might be wrong about that. Can you tell us a little about some of the key unknowns, in terms of unintended consequences for the physical environment, if we were to undertake one or more of these interventions?
David Keith: Yeah, that's a great question. It's really interesting to try and think about how to compare these interventions that happen at different scales. It's of course true that if you do roof painting or marine cloud brightening at a local scale, that has much less global risk than a large scale application of say, stratospheric aerosols. That's not comparing apples to apples, that's comparing something that has a very small local effect, and also very small local or global benefit, to something that has a global large benefit of reducing overall climate risks.
The better way to think about it is to compare the ratio of benefits to harms. For the ratio of benefits to harms, I think things look quite different. For example, take an idea that has now got some traction of adding reflective silicate scatterers to Arctic sea ice. That, on the one hand, seems like something that doesn't have big global implications. But if you do it in a very local place that also doesn't have any global benefit. If you really did that large scale, enough to really affect Arctic sea ice, then you can think about comparing that to doing an equivalent amount of say, stratospheric aerosols, to achieve the same impact on Arctic sea ice.
I think if you compare those two things, and ask which one has the largest environmental risks, I think the answer might very well be putting reflective aerosols in the sea ice. That involves an immense industrial machinery to move much more material, thousands of times more material, onto the sea ice, with equivalent heavy industrial emissions, and a big infrastructure in the Arctic to do that—whereas, the impacts of a stratospheric aerosols might way be well be much less.
Similarly, with marine cloud brightening or cirrus thinning, they're inherently local. While it's possible to locally adjust the radiative forcing, it's not possible to have an only local effect on climate. The world's climate is inherently interconnected by flows of heat and momentum. If you alter the radiative forcing and local temperature, say in one place, you will for sure make distant climate changes that may be unexpected, and may be harmful to people in distant places. That's why these inherently local technologies may actually be a bigger challenge for governance because they will be more likely to produce an unequal winners and losers outcome, and more likely to result in a situation where one nation or region are going to get pitted against another.
Daniel Raimi: That's so interesting. Can you give us maybe an example or two of how a deployment of one of those technologies could have some at least theoretical negative global consequence, that could lead to disputes between nations, for example?
David Keith: Sure. I mean, let's say that the Chinese become worried about weakening of their monsoon strength. A monsoon arises from the contrast between the warm summertime land and the relatively cold ocean. Maybe they believe that cooling the ocean off the China sea, using some kind of Marine cloud brightening technique, would help, and maybe it would, or maybe it wouldn't. The point is there's a sort of “push me pull you” between the Chinese and Indian monsoons. The flows are linked by the large-scale circulation of the atmosphere. It's perfectly plausible that that might actually make the Indian monsoon worse. To be clear, I don't know if that actually makes it worse or better. I don't think anybody does. But if the Indians thought that it made it worse, that would set up a pretty sharp confrontation on issues that relate directly to food security between two of the world's largest and nuclear armed powers.
I think it's intuitive for people to think that these local interventions, like cirrus thinning or marine cloud brightening, are somehow much safer and easier to govern than a more global intervention like stratospheric aerosols. There's certainly aspects in which that's true. I mean, individual local actions are in a sense easier to govern. We don't have coordinated global governance. On the other hand, the difficulty of coordinating local actions is exactly why we have a public goods problem in climate. I fear that there's a bit of a trap there. That, in fact, some of these local interventions both really will have local harms. They involve fluxes in materials, they create an atmosphere of chemical changes, and so on. Also, all the local things inevitably produce at least some non-local climate response. That non-local response means there's a kind of inherent inequality, a “push me pull you” quality.
While it is of course true that if all we're talking about is very local, short-term use of these methods, for example, if we're just using temporarily marine cloud brightening to limit peak temperatures in the Great Barrier Reef, which is being considered, that's really an adaptation measure. I can't speak to whether it works or not, but it doesn't have big governance consequences. If you're talking about larger-scale, continuous use of these technologies to really cool some area, that will cause some climate changes somewhere else. I think that is inherently pretty hard to govern and seems more likely to spiral out. Whereas, putting aerosols in the stratosphere feels somehow more scary, because it's global. If it's right, doing that produces a much more even result, where the distribution of winners and losers is much flatter, it looks like. In fact, most areas win. It's actually much easier to quantify what the impacts are, including the health impacts. That may actually be easier to govern.
Daniel Raimi: That is so fascinating. Many of these concepts are relatively new to me, as a pretty, I would say, casual observer of the research on these topics. It's really great to dig into some of these complexities with you. One related concern that often gets raised, again, to a fairly casual observer of these topics is, because some of the technologies are relatively cheap to implement, and could have a fairly meaningful effect on either local or potentially global temperatures, people, there's some concerns that are raised, that a single nation, or maybe a single wealthy individual could undertake a program of global consequence without collaboration from the broader global community. How much do you worry about that type of scenario? What are some of the nuances that you see within it?
David Keith: I definitely worry about it. I think that this sort of fundamental reality, that some of these technologies could be implemented by, say in principle, a single nation, will drive what happens. I think the idea that it's actually a single actor doing it, in a pure unilateral way, seems to be quite unlikely. When people think about it, they often migrate to one of two poles. One is the kind of unilateral action, paying no attention to the rest of the world. The other is the idea that it can only happen or should only happen with a global consensus, presumably through some kind of new UN treaty with notice agreement. I think both of those are highly unlikely to be what actually occurs.
I think even if you were a country that had maybe suffered some killing heat wave, maybe a tropical country that really was highly motivated to want to implement these technologies for self-protection, because of these sharp governance challenges, because of the interconnectedness of global affairs, I think you'd be unlikely to actually want to be purely unilateral. You could expect blowback. Instead, you'd likely work out coalitions of like-minded countries that also wanted to act. Indeed, if you wanted this to be stable, which it would be your self-interest to do—that is that the deployment keep happening in a way that was safe—you'd want to think about a coalition that could have some legitimacy, so the action was stable.
On the other hand, I think the idea of a pure consensus, all countries agree, is very unlikely. Even if all countries did broadly share the same potential benefits and risks, it might well be that a country would choose to do the kind of hostage taking thing that happens at any consensus based process, where you refuse to join the consensus in order to get some concessions or benefit on another element of negotiations. I think neither the global consensus or the pure unilateralism is very likely to happen. What's much more likely to happen is a coalition. The interesting thinking is all about the dynamics of those coalitions, and what combinations of countries and interests would make those coalitions most or least legitimate.
Daniel Raimi: Yeah. Do you have any thoughts that you could expand on in that particular area, about what types of coalitions you think might be most sort of likely to emerge, and how coalitions might differ?
David Keith: Yeah, I think that's a fair question. I mean, I can't pretend to see what will happen, but I think a key access of differentiation is the role of the superpower. I would say, clearly, if say China and the U.S. both want to do solar geoengineering in some form, or both definitely want it not to happen, then that probably determines the outcome. If in fact the superpowers don't have a really strong view or interest one way or the other, then you can imagine scenarios where this is at least partly drawn by a coalition of smaller powers, that somehow span some of the poor South and some smaller democracies in the North. That's certainly my best hope in a sense. You can come up with a hundred of these scenarios. I think the truth is we just have no idea the way this will play out.
Daniel Raimi: David, in talking to you about this, one of the things that I've learned is that when we think about solar geoengineering, we shouldn't just be thinking about temperature. There are other sort of environmental factors at play. Can you talk a little more about that?
David Keith: Yeah. There's a kind of common misconception that somehow solar geoengineering is intended to alter global temperatures, and that other things are side effects, or that it works for temperatures, but not for precipitation. That's sort of not true. If anything, it's the opposite of the truth.
I mean, the sort of scientific outcome is that solar geoengineering is in a relative sense actually more effective at reducing, say peak precipitation, than it is at reducing temperatures. The evidence from really all climate models is that it could reduce many of the key physical climate hazards, meaning changing water availability, changes in extreme storms, changes in peak temperatures, and changes of global averages to global temperatures. It could do that in a reasonably uniform way, and assertions that its only goal is global temperatures are kind of rhetorical. Nobody owns what the goal is, but the physical fact is that it will reduce a range of climate hazards, but not exactly equally, and there will be some trade-off between them.
Daniel Raimi: That's so fascinating. Can I ask you now to talk a little about the sort of state of research when it comes to real world experiments that are going on, either with your research group or others that you're aware of?
David Keith: Sure. Experiments only make sense in the context of the larger research effort. My view is that the central question now is not really about whether we should deploy geoengineering, but it's about what a serious research program should look like. My most important view is that there's enough potential for these technologies to be used, whether people like it or not, and enough potential for them to do good, that we should have a much bigger research effort. That research effort would have lots of components.
One component of it is trying to understand some of the atmospheric processes that are relevant to assessing the large-scale risks and efficacy of solar geoengineering. In understanding those processes better, one thing you may need to do is field experiments, which may or may not involve releasing material, but field experiments that are focused on particular processes that are important in climate science, that are poorly understood, and that really matter for solar geoengineering. There are ideas for experiments moving forward, including in our group, and there are experiments that have happened, laboratory experiments involving those processes in our group, and a group in Cambridge, and elsewhere. There have been some outdoor experiments, an experiment called E-PEACE, that was performed off the Pacific coast a few years ago. There's development experiments for marine cloud brightening, and for what we're doing.
I think a common misconception is that these experiments are somehow tests of solar geoengineering, or they're a step towards deployment. I think that really misunderstands the gigantic gap between these tiny experiments and deployment. It misunderstands what the experiments are designed to achieve. They won't tell you whether this will work or not. They will simply tell you some little chunk of information about some kind of nerdy inner process in the way the climate system works, that will slightly improve our knowledge of the overall process.
Daniel Raimi: David, we've touched on several different areas related to solar geoengineering over the last 20 or 25 minutes. Can you kind of give us your view of just what's the state of play right now, whether in terms of governance, or research, or other areas that you think are relevant?
David Keith: Yeah, that's a good question. It's really changing very fast. As of right now, the total global spending on solar geoengineering research is something like $20 million a year maybe—maybe not quite that. It's a bunch of little research programs in Australia, China, India, Norway, Germany, our program, whatever. These are all tiny uncoordinated efforts. But at the same time, there's much higher level of talk about larger efforts and about how to govern them. Just in the last month or so, there's been the first US federal budget allocations that are significant, that are in the direction of research on these topics. I think that's quite an important development. At the same time, there's been the first beginning of real discussions at a high level at the United Nations. There's been just a much larger international engagement.
Following an effort last year at the United Nations, the environment assembly, several countries have got much more involved in thinking through governance at a high level, so maybe Australia and France for examples. Broadly, I would say the international dialogue on this is much more substantive than it was a year or two ago. I don't think it's as substantive as it needs to be, but I think we're going in direction to do what I feel must be done, which is to bring solar geoengineering into the center of climate policy. That's true whether or not one thinks it should ever be done.
A reasonable position is to say that there should be a global moratorium on it ever happening. It's not a position I agree with, but there’s reason for a point of view of that. That would be one way to think about its role in global policy. I think what is becoming increasingly untenable is the idea that we think about climate policy over the next decades in a way to just pretend solar geoengineering doesn't exist. I think that is an increasingly implausible position. We need to bring it into the core of the debate about what we do about climate risk.
Daniel Raimi: Your presence on this podcast, and other podcasts and elsewhere, I think, is certainly helping us do that, and helping us start to think about these issues in a new and critical way. David, now let me ask you the last question that we ask all of our guests on the show, which is: What's at the top of your literal or metaphorical reading stack? So something you've read, or watched, or heard recently related to the environment, or maybe related to solar geoengineering, that you think is really interesting and that you'd recommend to our listeners?
We've mentioned podcasts a couple of times, and I'll briefly mention the new podcast from Harvard's program in environmental economics, led by Rob Stavins, who is a friend of RFF’s and a friend of the show. It's called Environmental Insights. David, you were a guest on that show recently. I learned a lot from your conversation with Rob. I learn a lot from other conversations that Rob has on that podcast. So I'd recommend people check it out.
David Keith: Rob's terrific.
Daniel Raimi: Yeah. How about you, David? Anything in particular that you're reading or enjoying?
David Keith: Well, if you want to know what's literally on my bedside, it's a book called Inner Ranges, a mountaineering literature by Geoff Powter. That's not totally incidental. I spend a lot of time climbing and in the climbing world. I had been struck by the fact that there has been this long connection between climbing and environmental activism.
There's actually a recent book by Harvard University Press, Pilgrims of the Vertical, that explores that. From the foundation of the Sierra Club, to the foundation of a bunch of environmental movements in the '60s, there was this link to that world. I've been puzzling about why that's true more, whether it's risk-seeking people who care about the environment end up being more inclined to environmental activism. In any case, what's actually on my reading list is Inner Ranges.
Daniel Raimi: Great. Sounds really interesting. Okay, one more time, thank you, David, so much for joining us on Resources Radio. This has been a fascinating conversation.
David Keith: Thanks a lot. Bye bye.
Daniel Raimi: You've been listening to Resources Radio. If you have a minute, we'd really appreciate you leaving us a rating or a comment on your podcast platform of choice. Also, feel free to send us your suggestions for future episodes. Resources Radio is a podcast from Resources for the Future. RFF is an independent nonprofit research institution in Washington, DC. Our mission is to improve environmental energy and natural resource decisions through impartial economic research and policy engagement. Learn more about us at RFF.org.
The views expressed on this podcast are solely those of the participants. They do not necessarily represent the views of resources for the future, which does not take institutional positions on public policies. Resources Radio is produced by Elizabeth Wason, with music by me, Daniel Raimi. Join us next week for another episode.