This is part 3 of my informal series about our energy infrastructure. My last post was about addressing concerns about nuclear energy, but really can only be understood in the context of our overall energy plan. The comments have been quite fruitful, and I would like to thank all the commenters who provided useful resources for further information, much of which I will synthesize here. That was exactly what I was hoping for, so again, thanks.

I won’t rehash the assessment of nuclear power, but just summarize my position. I am not saying that nuclear is the answer, only that something like it is necessary, and we should not take it off the table. Nuclear is relatively safe, we have plenty of fuel (enough to last centuries), we can deal with the waste, and the Gen IV reactors are extremely promising. But even for those who acknowledge those points but still reject nuclear, a common theme emerged. That theme is – we don’t need nuclear because X is a better option. This approach, however, is fatally flawed for two important reasons.

The first has to do with the economics of power utilities, which ironically was often raised as a point against nuclear – it’s too expensive. The best reference to address this issue is this lecture by Jesse Jenkins, a Harvard environmental fellow.  He addresses this, plus another common theme that emerged in the comments – we no longer need baseload production; that is an antiquated notion. I encourage you to watch the entire lecture, but here is the quick version.

There are three basic types of energy production and demand that we can use to balance the grid, to match production with demand moment to moment.

1- We have intermittent energy sources, mainly wind and solar. Their advantage is that they are renewable and zero carbon. Their disadvantage is that they are intermittent and cannot be controlled.

2 – There is “firm” energy production (similar concept to baseload). There are sources of power that run at a constant rate and are slow to ramp up or down. This does not mean they cannot be varied at all, just not quickly. We might, for example, plan on turning off a reactor during a time of day when we know solar production will peak. In this category are nuclear, hydroelectric, geothermal, and natural gas with carbon capture. We might also add to this category strategies for long term, massive, cheap energy storage.

3 – Rapid response strategies. This include sources of energy that can quickly be turned on and off, mostly natural gas. It also includes rapid storage options, like batteries, that can provide instant energy. On the demand side this category would also include strategies like shifting demand, such as charging your electric car overnight during minimal energy demand to smooth out the demand curve.

Traditionally we have managed our grid with options 2 and 3, but now we are having increasing penetration of option 1. This has lead many people to conclude that all we need are options 1 and 2, and we can dispense with 3. Wind and solar are becoming cheaper, and are already the cheapest option for adding capacity, so let’s just go with that. We’ll keep some natural gas in the mix until battery technology can take over. This superficially sounds reasonable, but I think Jenkins convincingly argues (with lots of data and simulations) that this won’t work.

The problem is that in order to make the grid work with intermittent sources of energy you need several things simultaneously. First, you needs lots of overcapacity. The way to make intermittent sources work is to spread out production over a wide area and share energy. This way the wind is blowing or the sun is shining somewhere. But this strategy means you need to have a lot of redundant capacity, to cover areas when the wind is not blowing or the sun is not shining. In order to build all this extra capacity, wind and solar need to become much cheaper.

Second, we need to upgrade the grid. We essentially would need a continent-wide grid capable of efficiently sharing energy across long distances. And finally we need lots of storage, which means much cheaper batteries, but also batteries with a longer lifespan, with more available and less toxic materials.

Even if we get everything on our wish list, with this strategy the greater the penetration of intermittent sources the lower the value of each added solar panel or wind turbine. When you go past 60% penetration the relative costs go up geometrically. This is because at higher penetrations you are adding overcapacity which by definition will be adding relatively less energy to meet demand.

If you compare this strategy to one in which we add back in firm energy, everything is easier. Even though firm energy sources, like nuclear, might be much more expensive, the cost of the overall energy production is lower when you cover the floor of 40-50% with firm energy production, and then use renewable sources and rapid response strategies to cover the rest. So even though nuclear is more expensive, it makes the whole system cheaper than if we tried to use only renewables and storage.

So right now, with very low renewable penetration, renewables are the best investment. However, we cannot extrapolate this condition all the way to high or total penetration, because we get diminishing returns. We’ll hit an economic ceiling somewhere around 60% penetration, and then utilities will want to put firm production into the mix. But here’s the thing, if we wait until that moment, our options will be limited. This is exactly the problem that Germany ran into – they got to that point, and had to build coal fired plants. They were already phasing out their nuclear plants, and didn’t have time to build new ones.

We need to plan for the predictable eventuality now, and plan on replacing our current coal fired plants with other options of firm energy production. We should not wait until something relatively quick like coal is the only option. In fact, we need to be phasing out our existing coal fired plants, but not by reducing our firm production (at least not completely), but rather by replacing it with zero carbon production.

The second fatal flaw with the single answer approach is practicality. All sources of power have strength and weaknesses. The problem is – the weaknesses get profoundly exacerbated when we try to scale them up. For each energy source there is, in essence, low hanging fruit – optimal locations or applications that we should be (and are, mostly) exploiting first. But once we go through the optimal options, we get diminishing returns and increasing problems if we try to push further and further.

For example, hydroelectric plants work best when the geography is optimal – a large source of water with sufficient vertical height to extract power from the potential energy. We are already using the optimal locations for hydroelectric power and pumped hydro storage. Pumped hydro is a great option, where it is a great option. Some pointed to this recent study concluding that there are potentially 530,000 sites around the world where pumped hydro can theoretically work. But this comes with the same problems. First, this is a theoretical study without ground work to back it up. But even if we assume its conclusions are accurate, that does not mean this is a feasible strategy.

Here is a good analysis showing that the numbers just don’t add up. In order for pumped hydro to be cost effective, you need at least one of the two water basins (you need two basins, one much higher than the other with little horizontal distance) already in existence. Creating water basins sufficiently large is a major civil engineering project. If we try to make this our one solution, we push the concept into less and less ideal locations, with higher expenses. Even figuring out where all the water will come from becomes an issue.

Some argue that geothermal is the one answer, but the problem is the same. We are already building or operating geothermal plants where they are feasible. We need locations where the magma table is high, there is porous rock, and a large water supply. In order to extend geothermal, we need advanced geothermal. This uses essentially fracking in order to create the necessary conditions for geothermal in other locations. However, this runs the risk of earthquakes. This strategy will likely work to extend geothermal locations, but is unlikely to run civilization on geothermal.

I wouldn’t even put all our eggs in the nuclear basket. Then we get into issues of fuel, waste disposal, locations, and sourcing all the concrete necessary to build the containment vessels.

Even solar panels have a tremendous drawback. There is toxic waste from producing solar panels, and difficulty with disposing of them at the end of their life. We also become overly dependent on rare earths and other raw material.

The bottom line is, that it makes no practical sense to have one magical solution as the only major source of energy. Let’s use all of the above. We should pick the low-hanging fruit from every option, and spread out the risks and downsides. Maximize hydroelectric, geothermal, continue to research carbon capture, and build Gen IV nuclear. Continue to research long term energy storage, and keep driving down the cost of solar, wind, and battery technology. We also need to pursue energy conservation strategies.

Jenkins also points out that this minimizes risk of failure, or maximizes our chances of success. What we are talking about is deep decarbonization of our energy infrastructure in time to avoid the worst outcomes of global warming. The worst case scenario is that scientists are correct to fear that there are tipping points, and if we pass a tipping point we will change the climate in ways that cannot be reversed on human timescales. But even without this, we want to minimize the impact of climate change, and we really need a strategy to get us to the end of this century. We also cannot count on game-changing technological breakthroughs (like fusion). We need to be conservative and assume incremental advances, at least for now.

The more different options we pursue, the greater our chance of success. Counting on renewables only plus storage is actually not a good gamble. Too many things have to come together. We are more likely than not to end up having to burn coal to make up the last 40% of our energy (or something like that), and that would be a disaster. And hey, if renewables plus storage does end up working out because of technological progress, then fantastic. We can phase out nuclear and other options at that time.

Framing our choice as between nuclear and your favorite renewable source of energy is the wrong choice – it is not the one we actually face.

First – we are rapidly expanding our renewable portfolio. That is happening, driven by economics. We are also incorporating battery storage, also because for now it is cheaper than other options, like building another natural gas plant. But this strategy will not get us across the finish line.

The choice we are actually facing is this – do we phase out fossil fuels as renewables increase, or do we phase out nuclear? It’s really a choice between coal and nuclear. This assumes we will maximize other options like geothermal and hydroelectric, but not entertain fantasies that these or something else will be the sole solution. Germany made the wrong choice, and ended up increasing their carbon footprint by having to build coal fired power plants. They admit that now – they should have phased out coal first, and worried about nuclear later, but they caved to popular opinion driven mostly by fear.

A far better plan is to bet on all of the above – let’s do everything, continue to research, and develop the optimal uses of each option – with the highest priority of phasing out all fossil fuels and deeply decarbonizing our energy infrastructure. Nuclear needs to be in the mix if we want to quickly displace coal. The Gen IV designs also have many advantages. One I particularly like is that some designs produce hydrogen in addition to electricity. Hydrogen can be burned cleanly to produce energy, and so adds to the rapid response portfolio. So these Gen IV plants are safer, cheaper, produce less waste (and can even burn waste from older reactors) and are more flexible, and are also a double win. They contribute to two of the three legs of our three-legged energy stool. I acknowledge they may not work out, which is why we need to bet on other options also.

There is one thing that almost everyone (except climate change deniers) agrees one – we basically have one shot at this. If our plans for the next 50 years don’t succeed, we don’t get another chance. We also can’t wait until we are arbitrarily certain about our climate models. We need a comprehensive strategy now. Just letting economic pressures determine what happens is also not the best option. We need to anticipate future realities. We need a 20-30 year plan, based on current knowledge and technology, designed to minimize risk of failure. There is no one magic solution – we need everything.