When it comes to the carbon footprint of food, most people have worked out that reducing your carbon footprint involves comparisons. Nothing in a supermarket is carbon neutral ... unless perhaps it is offset somehow ... but some foods have a much lower carbon footprint than others.
The same is true of energy sources. And, as with foods, some people find comparison results counter intuitive.
I gave a presentation a couple of years back in which I spent considerable time comparing the mining requirements for wind, solar and nuclear power. Despite abundant evidence to the contrary, during the discussion time following the presentation, a member of the audience stood up and proclaimed: "But the big problem with nuclear is the mining. It takes far more energy to produce uranium than you get back."
It doesn't. Why do people just say stuff without bothering to check?
A multitude of variants of this falsehood have been circulating around environmental circles for decades. Sometimes the claim adds in one or more of enrichment, waste management and decommissioning. The permutations vary, but the spirit of the falsehood remains.
A very mild form of the claim was made recently on twitter by Professor Colin Butler, an ANU epidemiologist. You don't get to be an epidemiologist without seriously good logical and mathematical skills and I've met Colin, so I know he has such skills. So I want to examine his claim carefully:
There are three problems with Colin's claim.
First, carbon neutrality isn't the right way to measure the quality of energy sources.
Saying something is carbon positive, negative or neutral is a very crude scale. Rankings are far more important than some absolute and probably unobtainable qualification like being carbon neutral. Why rankings? Why not just quantify the impacts precisely? Carbon accounting is always tough and particularly bad for energy sources. Is the aluminium in your solar panel produced in a smelter powered by coal or gas, or by nuclear power or hydro? Aluminium from the later two has a far lower carbon cost than aluminium from the former pair. The energy required to make the aluminium doesn't change but the emissions sure do.
These considerations lead naturally to a better measure of the intrinsic efficiency of an energy source: EROEI ... energy return on energy invested. One of the largest uranium enrichment plants in the world used to be the Georges Besse I in France. It was powered not by coal but by nuclear power, so the enrichment process at that plant produced no emissions ... other than those associated with mining the fuel for the reactors. We'll discuss those below. That plant closed in 2012; replaced by something 50 times more energy efficient.
While theoretically attractive, EROEI is still tough to measure. A switch to electric mining vehicles on the supply chain of some energy source will at first glance leave the EROEI unchanged, until you start thinking about the resources used to make the batteries. The energy required is considerable. The emissions from this don't exceed the tailpipe emissions that they replace, but they are significant. So much so that it's hard to see even electric vehicles being clean enough to fit any realistic emissions budget. Compare the green and blue chunks in the image below from a European study.
Still, many people have done EROEI studies on energy sources. Google it, you'll find plenty of material. But it's still something where different people, acting in good faith, get different answers, because the answers depend not just on the technology in a generic sense, but on the technical details of the implementation.
Nonetheless, even if these measures are tough to evaluate precisely, we can probably rank energy sources; remembering that what is critical is to minimise our impacts ... there is no free lunch.
We can probably rank variants of energy sources as well. Rooftop solar, for example, is far more costly in energy terms (and $) per kilowatt hour than utility scale solar. Small wind turbines have a lower EROEI than big ones. Economies of scale matter.
So in focusing on carbon, and carbon neutrality Colin is looking at the wrong measure.
But wait there's more
Secondly, I think Colin's claim carries an implication that mining/processing and decommissioning are some kind of achilles heel when it comes to nuclear power. He may not believe the utterly baseless falsehood about using more energy to make nuclear fuel than it delivers, but he clearly believes something is rotten in the nuclear power cycle. His use of the generic word "processing" rather than "enrichment" is probably an indication that he's not read deeply on the topic.
The truth, as I will demonstrate below, is that nuclear fuel production in all its forms involves far less mining and embodied energy than wind and solar systems. The title of a recent study in Nature says it all: "Renewable energy production will exacerbate mining threats to biodiversity".
The article notes that
... future production [of renewable technologies] will also escalate demand for many metals[refs]. It is unlikely that these new demands will be met by diverting use from other sectors or from recycling materials alone[refs].
When I challenged Colin on his claim, he responded with two refererences (here and here). The first paper is by Gavin Mudd and Mark Diesendorf and the second by Diesendorf alone; both well known as being heavily anti-nuclear.
Both papers discuss the nuclear fuel cycle; the production and disposal of uranium fuel. Let's focus on mining.
The first paper, by Mudd and Diesendorf, contains a table of greenhouse gas emissions per tonne of uranium ... including two interesting lines relating to the Olympic Dam mine in South Australia. The first line is "Olympic Dam (100%)" and the second is "Olympic Dam (20%)". The authors explain that Olympic dam produces other things besides uranium, so the 100 percent line corresponds to allocating ALL energy use and consequent CO2 emission to the uranium. This is obviously wrong, so why put it in the table? And why omit the relative production ratios of uranium to the other items (copper, gold and silver)? Possibly because it's only obviously wrong if you know the relative production ratios!
In a typical recent year, Olympic Dam produces about 200,000 tonnes of copper and about 4,000 tonnes of uranium. So basically, they crush 8 or 9 million tonnes of rock and get 204,000 tonnes of useful stuff (including a tiny amount of gold and silver ... measured in oz rather than tonnes). So despite only 2% of the result of the mining being uranium, Mudd and Diesendorf "generously" don't allocate 100% of the emissions to the 2%, but only 20% of those emissions. Why 20%? Mudd and Diesendorf use this percentage because it's the revenue ratio.
Suppose during some mining operation I happen to unearth a diamond worth as much as the rest of the output that year. Do I allocate half the carbon emissions of the operation to the diamond? That would be simply dishonest.
What about the rest of the Mudd and Diesendorf table? The other figures in the table seem reasonable and consistent with a Canadian study looking at emissions from uranium mining as a function of different ore-grades. With high grade ore, you have to crush less rock, so the emissions per energy unit go down.
But some methods of uranium production don't involve crushing any rocks! That mucks with the M&D table.
The in-situ leaching method pumps slightly acidic (or alkaline, depending on the geology) water into the ground to dissolve the uranium for extraction. There is no big hole in the ground and no crushing of rocks. In the year 2000, this method only produced 16 percent of global uranium ... but it is now used for 57 percent of production. If it wasn't for the copper at Olympic Dam, that uranium might be mined quite differently, or not at all. Pumps certainly use energy, but this is very different from crushing vast quantities of rock!
The Canadian study found that mining and milling contributed about 1 gram of Co2 per kwh. There was no in-situ mining in Canada at the time of the study.
So the global shift to in-situ mining will reduce an already tiny number to something truly insignificant.
I doubt Colin or anybody would be too worried about 1 gram ... but Mudd and Diesendorf don't put anything in context and didn't mention in-situ mining. They assiduously avoid comparisons. Comparisons would, at least, rank mining requirements of nuclear power with other sources and provide a valid resource for decision making. Decades of such comparo-phobic propaganda has left many people with the impression that uranium mining is a big deal, while with wind and solar, there are no mines, no processing factories, no waste, just green and sunny fields of shiny glass and mesmerising floating white spinning turbines.
The third problem with Colin's claim is also an implication that things like mining, processing and decommissioning aren't usually "accounted for". I've lost track of the number of times I've heard this canard. It's false. Utterly and completely. A 2015 study (pdf) looked at 25 studies of nuclear carbon footprints. Every study included all of these issues ... and more. I'm always flabbergasted when people say things like "But they never consider mining, waste and decommissioning in claims about nuclear!". This is typically an indication that the speaker hasn't read a single study on the life-cycle impacts of nuclear power.
Some missing context ...
But dodgy uranium mining numbers is only the start of the problems with the Mudd and Diesendorf papers. The really big omission in both the papers and Colin's tweet ... is, as I've already said, the lack of context. You can forgive this perhaps in a tweet, but not in the published papers.
So where do the emissions associated with uranium mining, milling and enrichment, waste disposal sit when compared to other energy sources?
Before looking at that, let's look briefly at enrichment. I won't deal with waste in this blog, but will do so in future (as I have in the past!). Most, but not all, nuclear reactors run on enriched uranium; not the natural uranium oxide which is generally shipped from a mine. The fission process in the reactor relies on a particular sub-type of uranium ... U-235. Natural uranium is low in U-235 so needs to be enriched in U-235. Put simply, enrichment takes a pile of uranium and removes some of the non-U-235 material, which boosts the ratio of U-235 in what is left. Consider an analogy ... dissolve a teaspoon of sugar in a cup of water. You can increase the sweetness of the mix if you can remove some pure water. Removing non-U-235 is considerably more difficult than boiling off some water from a sweetened mixture, but the goal is similar. Enrichment has traditionally been a very energy intensive process. But the carbon footprint of the process will depend on where that energy comes from. It's the electric vehicle issue all over again. If the energy comes from another reactor, then the footprint is smaller than if it comes from a coal plant.
But I said "traditionally" above because methods have changed. I mentioned the French enrichment plant above ... which closed in 2012. The method of enrichment it used (gaseous diffusion) uses 50 times more energy than the newer method; using centrifuges. Over the past 20 years, the percentage of uranium enriched by diffusion has dropped from 50 percent to zero.
While it can indeed take the nuclear industry decades to replace an obsolete technology, the anti-nuclear movement has never been known to replace an obsolete argument.
Even better than efficient enrichment is no enrichment. It is possible to design reactors to use natural uranium. Canadian CANDU reactors have been running on natural uranium for decades. But decisions about enrichment levels are complex, both in physical and economic terms. If enrichment was really a big issue, everybody would have gone the CANDU route; but it isn't.
Now let's get back to the context missing in the Mudd and Diesendorf papers as will as Butler's tweet.
Renewables and mining
Scientific papers often omit crucial context and lean on the excuse that context is irrelevant. They say in effect: "This is a study of X so why should Y get a mention?" Believe that if you like.
Here's a graph showing the various material required by various technologies. All of which is required to be originally mined and typically then transformed by various heavy duty industrial processes. It was produced from a table is a US Government technology review (p.390).
It's self explanatory ... nuclear uses far less mining than renewables. Not 20% less, or 50% less but 5-8 fold less. But what about fuel? Is there some duplicity in excluding it? Not for nuclear. Please keep reading, I promise I'll quantify it!
But what about the grid?
The above graph doesn't just ignore the fuel, it ignores transmission; getting electricity to the people who need it. Europe's population is predicted to decline over the next 30 years, but wind and solar power are driving a vast expansion of the electrical grid in Europe. Tens of thousands of extra kilometres of high voltage power cables will have to rolled out over the continent to handle the spatial and temporal variability in renewable energy sources. This grid expansion will come at a cost, not just in more mining, but more direct destruction of habitat. High voltage power cables are typically either copper or a mix of copper and aluminium ... not to mention the assorted plastics.
Technological intuition
When you tell people that renewables need far more mining than nuclear power, they tend to look at you like you are batshit crazy. That's because decades of propaganda have associated "nuclear" with "big" and "renewable" with "small". The reverse is true. Our intuitions have been twisted and mislead. They need to be reset by facts.
On the left (below) is the Opal research reactor at Lucas Heights in Sydney. This reactor is a research reactor, with a tiny output (as reactors go), and the pool is about 4m across. The fuel is right at the bottom; invisible if you don't know where to look! The image is a little distorted by a fish-eye lens, but you can see from the size of the people how small it is. It can produce about 20 megawatts of electricity ... and run 24x7 if required. There is a ladder into the pool, just in case anybody fall in ... which wouldn't be dangerous ... unless they couldn't swim.
On the other hand a Korean APR1400 power reactor produces 70 times more power ... 1400 megawatts ... also 24x7. How big is the pressure vessel ... the part when the fission reaction occurs? Roughly the same size as the Opal pool. But instead of being largely filled with water, it is packed with fuel rods.
So the output is massive, but the actual reactor is still quite small.
When people talk about "massive nuclear plants", they are actually talking about the generators, cooling towers and perhaps the concrete containment building. All of which are still tiny compared to renewable plants of similar output.
On the right of the image are the foundations of a 7 megawatt E126 wind turbine. It's maximum output, in a suitable wind, is about a third of the output of the little Opal. But it is truly massive. You can judge it's width by the people, but it goes down below ground for quite a distance. A wind turbine is like a tree, you need roughly as much mass below ground as above or it will collapse in a big wind.
During the course of a year, the little Opal research reactor can output 7 times the electricity of the much bigger wind turbine; because the wind turbine only produces energy when the wind blows. The foundations of the wind turbine weigh in at about 2,500 tonnes. And the tower holding the actual turbine is 135m high and weighs even more. A steel tower that high would buckle, so it is filled with concrete! You can see why they power air-craft carriers and ice-breakers with nuclear reactors and not wind turbines.
So we can see already a massive difference in the resources used to generate electricity. A nuclear reactor is intrinsically small. When people talk about "large" reactors, they are talking about the output, or the electrical infrastructure around it; not the actual working parts. The big cooling towers are the same for any thermal power plant, whether its heat is produced by coal, uranium, gas or solar power. Yes, you can have solar thermal power plants and if you had one with a similar output to an APR1400, its cooling towers would be similar. Air cooling can be used in any thermal power plant. The myth about nuclear power needing water is based on not understanding thermodynamics. Any thermal power plant ... a plant which drives a generator with steam ... works more efficiently with water cooling, but it isn't a necessity.
Small Modular Reactors (SMRs) are even smaller than the Opal reactor ... because they aren't packed with scientific equipment. They just have fuel. You can put them underground and they don't need the huge concrete containment structure that surrounds something like an APR1400.
SMRs are small high tech engineering and scientific masterpieces, whereas wind and solar farms are brilliant examples of state-of-the-art logistics applied to the assemblage of vast amounts of resources mined in big mines all over the planet and then transformed in large factories. The renewable industrial complex is far bigger than the nuclear industrial complex because it needs to be ... to assemble and process all those mountains of stuff. And the mountain is even bigger when you add in batteries.
The waste stream from a reactor is also small, concentrated and its management is well understood; the waste stream from wind and solar farms is scattered over large areas and poses considerable logistic and scientific headaches. The backing sheets of solar panels often contain fluorine, and the panels themselves (like those at Nyngan) can contain cadmium; both are toxic forever. Household panels tend not to contain cadmium, but can still contain toxic backing sheets.
Uranium mining
So finally ... how much uranium do you need for a "large" nuclear reactor, like an APR1400? You need to mine about 250 tonnes of uranium per year. So over a 60 year lifetime, the reactor would need about 15,000 tonnes of fuel. This is tiny compared to the mining for the concrete and steel in the 545 E126 wind turbines (requiring about 2.7 million tonnes of concrete and steel) you'd need to generate the same amount of electricity + the mining for the battery backup + the construction of the chemical plants to turn that material into actual batteries + the extra grid requirements. Miners of many kinds love renewable energy.
A nuclear plant, even a "big" one, will fit entirely in your field of view. But it's equivalent, in wind and solar farms, will be spread over vast areas of what should be wildlife habitat. Wind and solar farms are the ultimate in ecologically unfriendly energy ... starting with mining and ending with land clearing.
This blog is already too long to discuss waste, but a "big" reactor ... and now you know why I've put in the quotes ... will produce about 30 tonnes of it per year. It's management is hampered by the simple fact that nobody in the nuclear industry wants to throw it out. That's as stupid as sending aluminium cans to land fill. They want to use it as fuel in fast reactors. What are these?
Growth trajectories
The difference in output between cutting edge solar panels in research labs and those is solar farms is small. The difference in output, per tonne of uranium, between the best experimental nuclear reactors ("fast" reactors) and the most common reactors around today is almost 100-fold. The waste quantity goes down also, but not as much. The potential for better reactors, once the world gets behind the industry, is immense. But dealing with anti-nuclear propaganda has proved incredibly difficult. The anti-nuclear movement perfected the art of fear and fabrication long before the anti-vaxxers and 5G nutters. I became anti-nuclear in the pre-internet era and like to excuse my idiocy by the difficulties of fact-checking in those early days. I actually took the claims of Helen Caldicott at face value. These days, people can easily fact check propaganda ... but only if they want to.
The anti-nuclear movement is even having success in France ... despite France's decades of success in low emission nuclear electricity. The following graph says it all. The per person electricity generation emissions in Australia, France and Sweden over the past 3 decades. The Economist a few weeks back had a fanciful article about what would have happened if the world had embraced nuclear power in the 70s, like France. They suggested that climate change would be a minor technical issue for experts. I think they are wrong about that, but it is most certainly true that the infatuation with renewables (and red meat) has cost us decades in our efforts to roll back our damage to the climate.
French madness on display
But the French aren't immune from anti-vaxxers or anti-nukkers. During the 2020 Tour De France, the camera panned over the solar farm below.
It doesn't take too much imagination to imagine the land as it was before being trashed for this plant. SBS's Matthew Keenan read out from his notes that the 200 hectare plant could supply energy for 10,000 homes. He sounded mightily impressed. Happily he's a bike race commentator, not an energy analyst.
The Civaux I and II nuclear plants, on the other hand, occupy about 100 hectares and can power almost 2,000,000 homes. And they run 24x7; they don't shut down at night.
Donald Trump knows the value of a lie told frequently. Despite Joe Biden condemning the violence in response to policemen killing black people, Trump just keeps accusing Biden of ignoring the violence. Over and over again, he tells the same lie. So it is with the anti-nuclear movement, they just keep banging on with the same lies about mining and waste and decommissioning (and more) and those lies will get taken up by people without the time and inclination to check the facts. Honest people can be seduced into repeating and believing all manner of falsehoods.