Nuclear Energy: What are the real risks? with David Keith

James Lawler (00:07):

You are listening to Climate Now. I’m James Lawler.

Katherine Gorman (00:10):

And I’m Katherine Gorman. Today we’re going to be talking about the potential and the risks of expanding nuclear fission as a carbon neutral alternative energy source. Currently, the United States produces about 100,000 megawatts of nuclear generated electricity annually. That’s almost 20% of the national electricity need, but this number has been declining since 2000. According to the US energy information association, it’s actually projected to continue decreasing at least through 2050, despite our desperate need to replace fossil fuels.

James Lawler (00:42):

We spoke with David Keith of Harvard University to get a better understanding of why nuclear energy is not being prioritized in the effort to decarbonize and whether it should be. Professor Keith is an award-winning experimental physicist, notably carrying out groundbreaking research in carbon capture and storage and geoengineering. Dr. Keith has appointments in applied physics and public policy at Harvard and has worked at the interface of climate science, energy technology, and public policy for more than 25 years. David, thank you so much for joining us today.

David Keith:

Thank you. Great to be here and great you’re doing this.

James Lawler:

I’d love to begin by setting the stage for discussion with a little history, could you tell us the story of nuclear energy in the United States? How did it develop? And what’s happened over the decades since we started using nuclear energy?

David Keith (01:25):

The story, no question, starts with the world war and the development of nuclear weapons, that was central. The people who developed nuclear weapons ended up becoming the leading science advisers. And there was an era where that really seemed to be the most important thing in defining security and defining the age. And I think probably was too much influence from that nuclear grandees, if you like, in science policy, not just in the US, but around the world. So that started with nuclear weapons, but then moved to the idea that there would be ability to make relatively cheap commercial power with these claims. You probably heard they were sometimes made of too cheap to meter. The real focus on developing reactors was first, and this is I think kind of important, on developing Naval reactors, reactors for submarines, where it’s a unique use case. So unlike providing power, where nuclear power is maybe better for climate and worse in some other ways and so on, but it doesn’t do anything uniquely different from a coal or gas fired power plant; for underwater propulsion obviously it does because you don’t have oxygen to burn things.

David Keith (02:22):

So there was this focus on nuclear power for submarines, and the civilian reactors developed out of that. And the submarine reactors were developed to have relatively high power densities that have a lot of power, a lot of watts per square meter or kilogram of core. And that was a design that made sense for the nuclear Navy, but probably didn’t make sense for civilian reactors. And one of the big questions is how we progress beyond that initial set of reactor designs that really came out of the nuclear Navy. So anyway, then sort of jumping to commercialization, there was this boom of commercialization that really started in the sixties, with the peak of the reactor builds kind of being in the seventies. And that really ends with Three Mile Island, with this nuclear accident at Three Mile Island in Pennsylvania in 1979.

David Keith (03:11):

After that, there were no new reactor starts, that is new commercial decisions to proceed with building nuclear reactors in the US. Up until these most recent Westinghouse AP1000s in Georgia, where there were four starts, but now two of them have been canceled, if I’ve got my memory right. You mentioned that a hundred thousand megawatts, but I think it’s easier to say gigawatts about a hundred gigawatts of capacity. So a big chunk of that, uh, hundred gigawatts was built over roughly a 15 year period kind of centered there in the seventies. And then with essentially no build since then. And that of course, you know, bears on the question of how quickly we could build more nuclear power if it was sensible to do that for climate.

James Lawler (03:52):

Very interesting. So it only took about 15 years to build sufficient nuclear capacity to meet about 20% of the current electricity needs in the United States. Can we put that into perspective globally? You know, what, what percentage of global electricity use is nuclear?

David Keith (04:07):

So, 10% of global electricity is nuclear. And that’s down from about 17%, which is about the global peak in the mid-nineties.

James Lawler (04:16):

So, we’re talking about 10% as a fraction of the total energy use for global electricity production?

David Keith (04:21):

10% of output. The most useful, simple thing to do is count output, count electricity delivered to the grid. Where you’re comparing wind, solar, natural gas, coal, nuclear, et cetera. And on that score, nuclear delivers about 10%.

James Lawler (04:37):

And so how is that distributed globally? Can you give us a sense of the differences that exist in how much nuclear is used in different parts of the world?

David Keith (04:44):

The world certainly doesn’t look like the global average. So the US would be higher. France would be much higher, one of the countries with the highest fraction, but falling, and Asia has been low, but rising. The nuclear output is growing, but as a fraction, it may be about flat.

Katherine Gorman (05:00):

That is really great context for the state of things, you know, where we are today and how we got there. I’d love to explore where we should be going. How should we be expanding nuclear as a part of our approach to reach net zero emissions, and to start thinking about this, maybe you can do a little comparison for us. Could you walk us through what the impacts are to some of our various paths to decarbonize? That is, what we need to consider as we’re choosing to use biofuels or hydroelectric power or solar versus nuclear.

David Keith (05:30):

Yeah. So when it comes to sort of the primary energy is going to energize civilization, it’s useful to think in terawatts, thousands of gigawatts and, you know, global primary power. I feel like I keep having numbers that are out of date now, but let’s call it 17 or so terawatts, I’m sure some reader will correct me, that’s not quite right. And if we’re hoping that the poorest, billions get more access to electricity and energy, you’d expect that number to go up. But on the other hand, as we shift to be more electricity and less thermal, as a reason for it to go down, but I think you, you’re thinking that, you know, if you want a high energy society, you’re talking about something like 20 terawatts pretty quickly. And then you look around, you say, how could we supply that in a way that’s carbon free with low environmental impacts.

David Keith (06:12):

So, I think there’s some things you can cross off almost right away. The global hydro capacity is of order a couple terawatts and that’s it. There just isn’t more hydro available. And, you know, I think it’s easy to argue that most of that hydro has very high environmental impacts. Most obviously measured in land use. It destroys the land, and it destroys some of the bottom lands that are most ecologically valuable. Biofuels, it is, I think theoretically possible to imagine supplying all that energy from biofuels, but doing so would require something that I really think isn’t an overstatement to call, it would be kind of environmental Holocaust. I mean, you would just need to take all available land and run it, and kind of a high efficiency, biofuel production system. One useful way to think about this, I’ve been saying, talking about terawatts, another way to think about it, sorry for all these units, is watts per square meter, the amount of power you get out to supply society per square meter of landscape that you disturb, and biofuels are of order are less than a watt per square meter of real delivered, useful power typically.

David Keith (07:14):

And what that means is if you want to supply 20 terawatts, you end up taking quite a few percent of the total land surface, you know, culling the equivalent to current crop land or larger, depending on how you adjust the numbers. So, in my view, I kind of dispensed, I think, with biofuels and hydro. In my view, I’ll jump to what I think are the only two really serious contenders, and they are solar and nuclear. I think only solar and nuclear can plausibly be scaled to 20 or 30 or 40 terawatts of primary energy with reasonably low environmental footprints. I mean, to be clear, anything we do at this scale has an environmental footprint. There’s no such thing as perfectly clean energy, but they can certainly be effectively carbon free, and they can have low environmental footprint, certainly low as measured by percentage of land disturbed, and also by mobilization of toxics, and so on.

David Keith (08:01):

Wind power is a funny intermediate case. Wind power has a much lower energy density, if you count the overall area of the wind power installation, not just the area under the turbines, but that’s complicated because of course you can farm and do things under the turbines, but wind power also alters the climate subtly in ways that matter though, under big wind turbine arrays. So, there’s not a black and white threshold for wind power, it’s certainly possible to generate 20 terawatts of wind power. So, you could run the whole world on wind power, but my overall judgment is doing that has pretty high environmental impact compared to solar or nuclear.

James Lawler (08:38):

Well, is solar that low impact? I mean, when, when you think about deployment of solar, you know, it kind of involves this image of tessellating solar panels across huge swaths of land, doesn’t it?

David Keith (08:48):

Well, a huge needs a number, but in that same watts per square meter, I gave you, solar power, real practical output, not theory from active sites can easily be of order 10 watts per square meter. And that’s an annual average number. And so at that number, you’re looking at of order, 1% of global land surface to supply that kind of energy needs. I think that’s a small number. I mean, obviously there’s a lot of opinion here, but I think that there are some objective facts. I think that if we’re really talking about something like 20 or more terawatts, a decarbonized high energy society. You quickly winnow it down. If you want to have something that has a low footprint in terms of land, a low footprint in terms of mobilizing obvious toxics, you get down to solar and nuclear as the key choice, not to say it’s one or the other, it can be a mixture.

James Lawler (09:37):

I would love to transition to talking about some of the arguments that have been put forth against the use of nuclear energy technology. There are more or less four to my understanding. Number one is operational, so the possibility of meltdown. Number two is waste disposal. Number three is cost, very expensive technology. And fourth, you hear this concern that as we use more nuclear energy, there’ll be more nuclear material available, and you know, more awareness of the technology. And so perhaps, an increased proliferation of nuclear weapons.

David Keith (10:13):

Yeah, well that list of four is the right list. And I’ll kind of jump to the end and say, I think the ones that I take most seriously, but everybody has a different take, are costs and proliferation. And I’ll argue that waste disposal and operational risks, measured by reasonably objective risk measures, are pretty small, but of course, as we’ll get to, that’s not in practice, how many people perceive risks. And us experts, who sometimes think of risk as being probability times consequence, don’t own the way risk is defined. And the reality is many people perceive a very high level of threat from nuclear power. But, so let’s talk about operational risk. And I think my view is that compared to other energy sources in ways that matter, the operational risks, including meltdowns, are very low. And I want to give you some sense of what that means.

David Keith (11:07):

So at Fukushima, the most recent, really severe nuclear accident that has occurred. At Fukushima, something like a handful, Wikipedia says six, but a handful or so of the most exposed workers got a radiation dose that was bigger than say a quarter of a sievert. People often will read this in millisieverts, obviously millisieverts that thousands of a sieverts. So above 250. So, you know, for many listeners that will mean absolutely nothing. It’s just gobbledygook. What does it mean to get a quarter of a sievert dose? The answer as to what that means, is it means it increases your 30 year chance of dying from cancer by something like half a percent. There’s a fair error bar on that, but that’s the round number, but the uncertainties at that level are not gigantic, it’s of that order. So it’s important to say just when you look at those workers, I’m very glad I don’t have to face that kind of risk in my day job, but compared to lots of other industrial risks, having a big industrial process where there’s a bad accident and you have six people who have a kind of some fraction of a percent increase in getting a fatal cancer in 30 years.

David Keith (12:19):

I mean, if I was their child, I’d think that was awful, but just compared to other industrial technologies, when you think about the total amount of power, this is a very small risk. And it may actually be a small risk compared to lots of, I mean, you think about renewable power, and you may think, oh, that’s a good thing, but if you look at, say, the accident stats in collecting wood, you would almost certainly find them much higher number of, of severe accidents per unit time. All of these risks are about comparisons. And if you look at the risk to the population outside the plant, the outside population has received much smaller doses doses, typically less than a millisievert, it is really unclear what the risk is cause that, cause we actually don’t have solid epidemiological data to say what the risks is of these low exposures.

David Keith (13:05):

We don’t directly observe them, but if we assume this thing called linear no threshold, and as we assume that we have this data from high exposures, from the nuclear weapons the Americans used on Japan and from industrial actions over the years, we have data on high exposures and the probability of that causing fatalities for cancers. If we extrapolate down to zero linearly, then you can get an estimate of what the risks are to the whole population. And my view is that those risks from accidents are still low compared to other industrial technologies when you do fair comparisons, and then there’s a whole separate set of questions about whether this linear no threshold model is actually correct. I think there are reasonably persuasive arguments that it is exaggerating the risks of low exposure to, to nuclear, uh, radiation. And the risks are probably lower than that. They may actually be near zero for these low doses, but I think the truth is we simply don’t know.

Katherine Gorman (13:59):

I think that it’s that uncertainty that makes people nervous, isn’t it? I mean, even if it is probably close to zero risk, it’s still more than zero, right?

David Keith (14:07):

I guess I’m talking risk. And I feel it’s, like, really important to get people comparison. So, I’m sitting in Western Canada right now. The air quality that I measure outside my house has often in the last few days been over 200 micrograms per cubic meter. This is air from smoke, just to give you a sense of it. What the heck does that mean? Well, roughly every 20 or 30 micrograms per cubic meter, if you lived in it all year, it takes a year off your life. When you look at the risks from the existing fossil fuel power fleet, for example, you get total estimate of fatality from that dose response relationship, which we actually know pretty well for air pollutants that suggest, you know, something of order, a few tens of thousands of premature deaths in the US, and that’s even after the enormous cleanup by the Clean Air Act. So, I’m saying this to kind of give people a sense that there is no zero risk technologies, but that the technologies we’re talking about here, I think when you compare them to water and energy technologies, the meltdown risk is pretty low.

Katherine Gorman (15:05):

So what about the concern around waste disposal? The prospect of dealing with huge amounts of waste, the that’s radioactive, and is dangerous for thousands of years. It’s definitely unappealing. How accurate is that idea though? I mean, is this a huge problem?

David Keith (15:22):

At some level, the answer is we don’t know, because if you’re asking how safely it can be disposed of over a thousand years, almost by definition, we have no idea because you need a thousand years of history to know that. I have a couple different answers, I mean, my, my bottom line answer is I don’t think nuclear waste disposal is truly a central problem. I think there are ways in which people who have other legitimate reasons for not liking nuclear power have found that that was the most effective way to try and stick a wheel in the spokes and stop it. And you also have to understand this is a case where the US situation is a little different from the rest of the world. The US made a formal congressional level decision to put the Nuclear waste in Nevada. A bill that was called the Screw Nevada Bill by, I forget who was the most, there was a famous Nevada congressperson who said that one of your readers will know, and this was the same Nevada that had been the subject with, of outdoor nuclear weapons testing.

David Keith (16:12):

And I think there was a set of political decisions there. And I said, technical decisions about the way the US built this Yucca mountain disposal site, which I think, in hindsight, is a bad design. And it’s important to say that other countries, Finland, Canada seemed to be moving forward with deep disposal, using a different design in deep water, saturated, stable rock formations that appears both to be technically a lot lower risk, but all of these risks are very small, technically, but also appear to be happening in societies where there’s maybe a higher level of trust in government and science and appear to be more socially acceptable. The way I think about it, is about risk trade-offs. So, if we are choosing, say between running a gigawatt year of nuclear power or a gigawatt year of hydro or wind power or solar power, or what have you, or regular old fossil fuel, all of those things pass risks down to the future.

David Keith (17:06):

Regular fossil fuel most obviously passes the risk of the CO2 in the atmosphere down in the future. But constructing, say, solar power is not without risks because it liberates a bunch of toxic metals. I don’t want to overstate that, I’m actually very much pro solar, but it’s a reality. It’s a little hard to quantify because it’s changed as manufacturers have cleaned up their acts. But if you attempt to do this on a lifecycle basis per watt delivered, it is significant compared to the nuclear waste from nuclear power. And all of these things are just tiny compared to the actual air pollution risk of the existing fossil fuel power system.

James Lawler (17:46):

Well, let’s shift the conversation then and consider the concerns you mentioned. You felt were legitimate when we’re talking about nuclear energy expansion, and that those issues were the costs of nuclear energy and the risk of proliferation of technology that can produce nuclear weapons. So what should we know about these issues? Maybe we, maybe we can start with risks around proliferation.

David Keith (18:04):

I think there is a real sensible case to be made to say, we just shouldn’t develop large-scale nuclear power because of, of nuclear weapons. To be clear, there’s not a one-to-one link. You can develop nuclear weapons without civilian nuclear power. In general, they have been developed more or less independently in a technical sense, but I think it is pretty much a fact that if we really did develop large scale nuclear power, as part of the climate problem, developing many terawatts of new nuclear capacity over the next three decades, you spread knowledge, technology and the key materials around the world in a way that fundamentally makes it easier for states or subnational entities, small groups to develop weapons. I just don’t see there’s a way that you avoid doing that. And how you think about the risks of that really depends on who you are. I don’t think there’s an easy answer.

David Keith (18:52):

I mean, one answer to say nuclear weapons are out there anyway. It’s not going to fundamentally change it that much. We have to learn how to live with them, otherwise we’re done. But I think it is an underlying reason that the way there’s no, there’s no quantification because this involves political judgments about conflict, where we can’t quantify how they’ll be used in future conflicts. So to me, that’s one of the most sort of substantive or deep concerns. And I think it also, it’s driven a bunch of the concerns. So a bunch of the anti-nuclear concern, came out of movements against nuclear war, which I personally very much support and against the excessive proliferation of nuclear weapons.

Katherine Gorman (19:31):

So, unlike the risk of a meltdown or the risk of waste leakage, which you’ve described in quantitative terms, we can’t really calculate this kind of risk, right? It’s hard to compare whether this risk is, say, greater or less than the risks associated with solar, hydro, fossil fuels.

David Keith (19:50):

Yeah. I don’t know a useful way to do that. I mean, I guess if I was scoring points, I could say, well, there’s never been a terrorist nuclear weapon. And so historically their risk is zero, but obviously, that’s not very convincing. I think the answer is we really don’t know, and it depends on kind of global decisions, but it really depends on the politics of, of who is in control of how it was done. There are so many examples where what is done ends up being deeply irrational because of just the reality of politics. So, for example, France has a system of when you have a uranium reactor, you end up making some plutonium, it’s a process called breeding. You can separate that plutonium and make a plutonium oxide fuel. It has some advantages, if you believe we’re running short of uranium, which some people did. If you were doing that system and you were rational, you would have done it all in one facility with guards and guns around it because in the place where you actually make one part of that process, relatively pure plutonium, that’s a material that can be diverted for weapons, but there’s a jobs program.

David Keith (20:49):

So, in fact, one of the facilities, I forget which one, the fuel fabrication facility is like in Marseilles, and the separation facility, somewhere in Northern France, and there’s truck driving back and forth between them. Like that’s just nuts, and nobody would ever plan that, but it’s the way it happened and more of that’s going to happen. And I think nuclear advocates have to be honest that if we really develop terawatts of nuclear power, there’ll be a lot of dumb decisions like that made along the way. Even if coldly rational, people would not have made their decisions that way.

James Lawler (21:20):

Well, let’s move to cost, I understand that when we’re talking about the cost of electricity, there’s more than one kind of cost that one needs to consider. Can you outline what those costs are?

David Keith (21:28):

One is basically the cost of the capital. So, you spend money to build the plant, and then the fact that you have to kind of pay back the bank over the operation of the plant for that capital, that’s the capital costs. And the other are the operational costs, which can be divided into fuel costs and non-fuel operating costs, which are mostly salaries for most of these plants. So, you add those all together and you get the overall levelized cost. And in general, for nuclear reactors, as they have operated in the US, capital costs in general have been the dominant contributor to levelized cost for nuclear.

James Lawler (22:01):

And what is this cost? Is it something that’s decreased with time due to technological development?

David Keith (22:06):

I think the right answer to that question is there’s too little data to know. And the data is all so contingent, that it’s really not like solar power, where there’s a mass market for commodity devices, and there’s well understood pricing in that market. And so, you can just see the way solar module prices have declined over time, but for nuclear power, the thing I said about no new reactor starts after ‘79, that was the case in the US, but more or less, not exactly that was the case in the rest of the western world. And so, so more or less we had this early set of builds, and a guy called Jonathan Koomey, who I really trust, who really works to get the data right, I think isn’t sort of pro- or anti-nuclear, he did a very nice retrospective analysis on the US plant builds because there was, that data in the end was pretty much all public to estimate the capital cost of those builds.

David Keith (22:57):

And in today’s dollars, a big fraction of the total US nuclear fleet was built around 2,500 bucks a KW. And if you could do that today, that’s cheap. And if you do that and can operate them in a reasonable way, that’s very cheap, low carbon power. And there’ve been several other analyses recently of the Korean experience and the French experience, but all of them are very discontinuous. And for all those, so one way to tell a story as costs have gone up over time. And I think there’s some weight to that. I mean, so clearly in the US, if you look at the two reactors that are now actively under construction, the two AP1000s in Georgia. Vogtle, there, I’ve lost track of where costs are, but they’re, you know, north of $15,000 per KW, I think is probably a fair number.

David Keith (23:40):

And that’s true of some of the other builds around the world too, but there are also examples of builds that have come in at more like 3000 bucks a KW. Recently, my understanding is some of the Korean reactors, including the builds in UAE, were done for firm contracts at prices of that order. So, my sense is, in fact, depending on your sort of pro- or anti-nuclear stance, you can define quite different views of the world. One where the costs are really out of control. And the other one where they’re actually factors that do produce low costs, and we know what those factors are, and we can build them low cost.

Katherine Gorman (24:14):

So are you saying with the cost, it doesn’t actually have to be a major barrier for nuclear power?

David Keith (24:19):

It’s clear that it is possible to build out a lot of nuclear power at costs that are very much competitive with other low carbon power sources. So, I would say we know that it’s doable in a kind of existence proof sense, but in some kind of organizational sense, we don’t know how to do it. That is, I don’t think there’s any government, combination of industries, and NGOs who can say we’ve got an effective – You know, technology isn’t just hardware. Technology is a combination of industrial knowledge and a regulatory system. All of which are able to work together with a financial system to gain public acceptance, you know, to be a technological ecosystem. We do not have a working technological ecosystem that allows us to produce large quantities of nuclear power plants now. I think we could have one, but we don’t have one now.

James Lawler (25:06):

Okay. So clearly this is not a straightforward issue, but to wrap up our discussion, I’d love to get your opinion. On the whole, how does nuclear net out for you? Do you think we should be ramping up our construction nuclear plans? Or should we just go for solar, which was the other option you considered the most feasible?

David Keith (25:21):

I, I’m uncomfortable answering this as a ‘should’ question. It requires a kind of omniscience about human politics that I certainly don’t have. I think that I would vote for as a citizen, a serious effort to build out nuclear power, but it would require something that will, it really was a reboot. If we’re going to do this in a way that’s useful for climate, we need to find a way, and that requires leadership, requires leadership from the environmental community, from the environmental NGOs, requires leadership from the top of government to say that there’s going to be some major change, changing the way we site these things, and the way we settle disputes about siting, changing the way we contract for them, changing the way we regulate them. And that’s a big deal. It’s doable, but it’s not going to happen as a kind of sideshow. I would say, we’re not going to do this unless nuclear power rises at some point to near the very top of the political decision pile.

David Keith (26:13):

But I think that if we could figure out how to really build out a substantial amount, terawatts of nuclear power, collectively as a species – not in a way that’s perfect, there’s going to be accidents, but if we can figure out how to do that, I think that would likely get us to a better way to deal with a climate problem than if we don’t. Because it would give us a lot of 7/24 dispatchable, low carbon power with low environmental footprints. And that’s something I care about a lot. And I think if we end up solving this problem all with some combination of, of the other things I mentioned, it, well, including actually fossil with CCS, we end up overall with a bigger environmental impact. And my view is that there are ways to do this that don’t really alter the large-scale proliferation risk that much, and that we separately have to, you know, figure out how to learn to live in a world with the knowledge of nuclear weapons, or we’re done. That is the fundamental truth that’s been known since 1945, and whether or not we do nuclear reactors I think doesn’t change that in a fundamental way.

Katherine Gorman (27:21):

Well, David, thank you so much for being part of this conversation. And for our listeners, if you’d like to learn more about how to assess the relative costs and benefits of the various alternative energy technologies, you can check out David Keith’s online course, Energy within Environmental Constraints. We’ve got a link included in the transcript on climatenow.com. And that is it for this episode of the podcast. You can check out other interviews, watch our videos, and sign up for our newsletter at climatenow.com. And if you want to get in touch with us, email us at contact@climatenow, or tweet at us @weareclimatenow, and we hope you’ll join us for our next conversation.