I’ve been following with interest how some nuclear power advocates are suggesting that building anything else than nuclear power is sidetracking us from the climate goals. These advocates claim that variable, non-dispatchable renewables will not be ultimately capable of delivering a deeply decarbonized energy system, and therefore we shouldn’t waste time or money on building them because we’d only have to replace them with dispatchable sources of low-carbon power – which in practice means nuclear.

It should be noted that this is very much a minority position among those who support nuclear power as a weapon in the climate fight. The anti-nuclear establishment’s most common strawman argument against nuclear supporters is to claim we want to solve climate change with nuclear only. This is not true, and there are only a handful of pro-nuclear activists who see no need at all for anything else. However, the 100% nuclear argument is not as far-fetched as it may seem at first, and before arguing why I don’t quite believe in it I must first explain what it is all about.

The key issue at the root of the problem is that variable renewables begin to cannibalize their own profitability long before the energy system will be decarbonized. As a rule of thumb, increasing the market share (penetration) of variable renewable energy sources (VREs) beyond their capacity factor will be increasingly difficult.

Why? Because when the market share of a VRE begins to equal its capacity factor, the occasional full output of all these VRE generators – say, during a sunny or windy day – will begin to exceed the total demand on the electricity grid.

Figure 1: Production from wind and solar in Germany + Austria on one ordinary day if solar generation is increased 4x and wind generation 8x from current. Black line denotes demand.

Because of the way electricity is valued, this will cause the price of all electricity produced during such times to fall close to the lowest marginal cost of production – which is close to zero with VREs. At worst, inflexibilities in grid operation may mean that producers have to pay to users to consume the excess electricity. At this “inflection point,” as John Morgan has dubbed it here, adding more of the said VRE will not be economic any longer. If, for example, grid electricity will be almost free when the sun shines, where is the incentive to install additional solar panels?

Figure 2: What electricity can be sold at non-zero prices. Note in particular how little solar electricity has any value at all. For more thorough discussion, see my post here.

(See also here the excellent essay by Jesse Jenkins and Alex Trembath examining this very problem. I’ve also written about the issue here.)

The problem stems from the fact that realistic capacity factors for wind may remain below 40% and for solar, 15% is already optimistic in the northern latitudes. Furthermore, these figures aren’t truly additive: because wind turbines often produce while the sun is shining and vice versa, we cannot simply add up the capacity factors of various energy sources to determine what is the maximum total the electricity grid can economically manage. It may turn out that in large-scale grids, a 50% total market share for VREs could already present substantial practical difficulties. Smaller – that is, national – grids could easily exceed this, provided they have neighbors willing and able to both absorb occasional excess production and provide backup when weather isn’t cooperating. However, high VRE penetration in one region would also mean that its neighbors cannot economically build as much VREs themselves.

The limit will also bend depending on how much we can assume energy storage and demand management. In theory, with economic and scalable energy storage (or perfect demand flexibility), the problem could be solved easily. However, it is still unknown whether we can rely on these technologies to develop the way we want – and there may be other difficulties that reduce the overall limit, such as existing low marginal cost generation, problems in building enough transmission lines, and so forth. Nevertheless, whether the economic limit will be 40, 50, 60 or 70% is not very relevant. As long as the limit is not close to 100% the question is where do we take the rest from?

It is very likely that if we exclude hydropower, which cannot be realistically expanded in most developed countries, the least-cost option for covering the rest of the demand will be by burning stuff – and there aren’t that much carbon neutral fuels to go around, not even if all the waste and all the sustainable biomass is burned for energy. The likely end result will be blown carbon budgets and a failure to prevent dangerous climate change.

Thus, say the 100% nuclear advocates (who typically do not believe that energy storage and/or demand management will develop fast enough to matter), shouldn’t we focus on low-carbon power plants that deliver power up to 90% of the time? And if we have low-carbon power that delivers with a 80-90% capacity factor, then variable sources do not really add anything we actually need from an emissions standpoint. Aren’t wind turbines and solar panels therefore superfluous, and isnt the “all of the above” energy strategy doomed from the beginning?

Why do we nevertheless need all of the above?

I don’t buy the 100% nuclear argument, and in the following I try to present several reasons why. Due to the complexity of the issues, the following four points are necessarily more like thought experiments than anything you should take as a gospel, and I’d value your feedback.

1. All low-carbon energy reduces fuel burn and conserves hydro reservoirs.

The first reason why I support all low-carbon energy sources right now is because the world energy system still runs mostly on fossil fuels. This is true even for electricity system, which is in theory the easiest to decarbonize.

Therefore, with some important local exceptions, all new low-carbon energy can reduce fuel burn and conserve water in hydro reservoirs, which in turn helps to reduce fuel burn during peak demand. In most places, we are still so far from the point where new low-carbon energy will not reduce burning that the question “which low-carbon energy we should build” is largely academic. Anything helps.

2. Totally discounting technological change is unwise.

The problems of large variable renewable market shares outlined above are real. However, they may be solvable with a combination of partial solutions. So far, the pessimistic analyses I’ve seen have focused on some proposed solution, typically energy storage, and proceeded to “disprove” its feasibility by quoting high prices, low energy returns (EROI), lack of critical raw materials, or similar limitations.

These cautionary messages have their place, but just as very few people are claiming we should solve the climate/energy problem with nuclear alone, very few people are saying some single solution would solve the problem of intermittency.

More likely and more feasible is a combination of solutions. Many different forms of energy storage, from batteries to gas generators to heat and cold storage will help; some demand response will do its bit; grid interconnections and “smart” grids do theirs; and optimizing renewable energy generators for maximum capacity factor instead of maximum annual energy production will increase capacity factors and thus raise the rule of thumb limit described above. (For solar panels, orientation can make quite a bit of difference in capacity factor, while wind turbines can be optimized for low wind speeds at the expense of annual generation. The reason this hasn’t been widely done yet is because many subsidy schemes reward peak generation, not firmness of supply.)

Furthermore, price trends in e.g. batteries do make the more pessimistic assumptions outdated even as we speak. It may be that battery prices will not follow the most optimistic trends, but there is still clearly room for improvement, and such improvements are happening. Ditto for EROI issues, and for determining whether we should install some particular source or not, EROI is a problematic metric anyway.

That said, we should be careful not to draw the opposite conclusion – that very real technical issues will be solved just because solving them would be very beneficial for the society. Demand alone does not guarantee a technological solution; this is one of the key takeaways from my forthcoming PhD thesis that examines the technological response to resource scarcities. We shouldn’t discount the capabilities of our scientists and engineers nor the power of suitable incentives, but we also shouldn’t rely on them.

3. Electricity price crashes build necessary niches for critical decarbonization technologies.

Related to point #2, I actually believe that we need electricity price volatility to develop the necessary components for a truly decarbonized energy system.

Decarbonizing current electricity system alone is relatively easy, and in developed countries, it might be possible solely with renewables or solely with nuclear. However, electricity accounts for only about a third of global carbon emissions. To get to the rest, we urgently need technologies that can reduce fossil fuel use and emissions in other sectors. Transport is particularly important, but other sectors, such as construction and industry, also need to be decarbonized.

The simplest and, probably, the most reliable way of getting from here to there is by electrifying just about everything. That is, we need to learn how to use electricity, either directly or indirectly, for tasks that now need fossil fuels. For example, we need to learn how to transport people and goods with electricity, and how to make iron with it.

Most of these tasks are likely to require technologies that, in essence, convert electricity to some other form of energy or to some other material, or transfer demand for electricity from one moment to next. For example, we might turn towards battery-powered electric cars, and make iron in a hydrogen process where hydrogen is separated from ordinary water via electrolysis. Likewise, we could design heaters and coolers to operate only when electricity is plentiful.

At the same time, such technologies could help us solve the problem of electricity overproduction. Given cheap enough methods for storing electricity for later use – the hydrogen route seems one of the more promising ones, either directly or after conversion to methane – the problems outlined in the introduction evaporate. Excess electricity production would simply be absorbed in storage systems, to be released when production flags.

But the problem right now is that these storage systems are prohibitively costly. Likewise, electrification of fossil fuel-demanding industrial activities proceeds only slowly, because electric alternatives are usually more expensive.

To get the prices of these technologies down to where we need them to be, we are very likely to need substantial innovation. This, in turn, is likely to require what is known as “niches” in technological transitions literature. Niches are specific market conditions or other safe havens for innovations to occur and be improved before having to compete with more established technologies. Absent such niches, it is unlikely that even good ideas can be developed to mature enough levels.

Quite obviously, technologies that require us to store or convert electricity would benefit from a niche where electricity is very cheap – or where the grid operator might actually pay for users to spend excess power.

This is increasingly the situation we’re seeing in grids characterized by large percentage of variable renewables or relatively inflexible (for economic reasons, mind) nuclear. What’s more, I suspect that increase in solar power in particular will be crucial to the emergence of low-cost electricity niches: this is because household solar PV is soon becoming a fairly attractive proposal, particularly in markets where transmission costs are high. It may be that new houses in particular soon begin to come with solar panels as standard, and even though they might generate excess electricity, the absence of transmission costs alone may make them profitable – at least until transmission cost structure changes to account for this. Such installations could, in the next ten years or so, provide quite a bit of low-priced electricity gluts that help energy storage, conversion and demand management technologies to grow.

I believe we need these low-carbon technologies no matter what the 2050 electricity grid will look like. Even though nuclear power hits the capacity factor wall much later than variable renewables, a 100% nuclear energy system would also need energy storage, conversion, and demand management. Therefore, having regular electricity price crashes has its upsides as well.

A caveat worth keeping in mind, however, is that electricity glut may be used for other purposes as well. It is quite conceivable that alongside, and perhaps instead of storage and conversion, we’re simply going to see electricity being used more when it is cheap for needs we don’t even know we have right now. An example I’ve toyed with is that cheap electricity might well be utilized to mine for bitcoins; this could be a much more profitable activity than storing the electricity for later use, particularly if the electricity grid is firmed up by the least-cost solutions – fossil fuels. And this last, very real possibility is in my mind the most important reason to support all low-carbon energy sources now.

4. Both 100% RE and 100% nuclear by 2050 are politically and technologically unrealistic pipe dreams.

The recent U.S. elections showed that climate change is still one of the least-motivating political issues in most developed countries. In our environmentalist bubble, where we’ve been sharing climate change news and worries for the last decade or so, this may sound all wrong. But this is simply the reality where we seem to live in.

Going either 100% renewable or 100% nuclear would require an enormous public program to build what would be, in effect, nationalized electricity grids. This would be necessary simply because the required build rates are so immense that there is no hope to achieve them on a commercial basis alone. In case of nuclear industry that has been allowed to wither, the required build rates for 100% nuclear would present significant bottlenecks in manufacturing capability alone. Furthermore, the economics of low-carbon power would require state intervention even for major decarbonization of national electricity grid. This has been done before with nuclear in France and to lesser extent in other countries, but the only real justification for going this route today would be avoiding dangerous climate change.

Compared to other scenarios, 100% nuclear has an additional problem: nuclear is by far the most hated energy source there is. That this hate and fear is based on decades of misinformation is regrettable, but immaterial. Mobilizing opposition to any plans to build an energy system that’s even mostly nuclear would be laughably easy. To be fair, right-wing populists today have begun to oppose renewable energy and wind power in particular with the same blatant disregard for facts, logic and research the shrillest anti-nuclear advocates perfected and weaponized in the 1980s, but this does little to make nuclear more popular, even though some do support nuclear simply to spite “the Greens”. Those of us who continue the fight against low-carbon energy sources they dislike are studiously ignoring the elephants in the room: that the most likely alternatives, and the largest beneficiaries from this fight, are fossil fuels. Meanwhile, the fight between low-carbon sources takes our time, saps our resources and results to personal animosities between those genuinely concerned about climate change. All these are regrettable enough outcomes of infighting in the best of times, and deadly threats today.

My interpretation of the world is that particularly right now, we environmentalists do not have the slightest hope of building the 2050 energy system we might really want. But we might have some hope to influence decisions so that we can reduce emissions instead of increasing them wildly. In other words, when we see proposals that might help, we should rally to help them come to fruition.

Barring several technological miracles, the Paris targets of 1.5°C, tenuous from the beginning, are almost certainly unachievable as the United States now puts her might against climate change mitigation. But we need to remind ourselves that this fight isn’t about 2°C or bust; every tenth of a degree counts, as there are degrees in any catastrophe.