As oil prices recede from all-time dollar highs and some of the hot air gets let out of energy policy debates, it’s a good time to remember that here’s a key concept missing from almost every popular discussion of the subject: energy density. Specialist economists get it, but almost nobody else does. It is important to understanding why most forms of “alternative energy” are mirages, and what a sane energy policy would actually look like.
The background to this is that the few technologies we have for storing electricity (batteries, pump-fed ponds above hydroelectric turbines) are lossy and don’t scale well. Worse, power transmission is significantly lossy as well. These mean several things, all of them bad.
Absence of a decent storage technology means we can’t really time-shift electricity demand. When more electricity is needed (for example, to run air conditioning during the day in the American Southwest) more power plants have to be running and feeding power to the grid in real time. There’s no way to run plants at night and store the generated power for daytime use.
Transmission losses mean our ability to space-shift demand is limited, too, though not as severely. Electricity-intensive industries (the classic example is aluminum smelting) need their own dedicated power plants nearby.
The combination of these problems means that household energy conservation is mainly a way for wealthy Westerners to feel virtuous rather than an actual attack on energy costs. Household conservation slightly decreases the maximum capacity needed locally where the conservation is being practiced, but has little impact further away, where demand has to be supplied by different plants. Industrial efficiency gains are far less visible; but, because the scale of industrial energy use is so much larger, they matter a lot more.
The combination of these problems also means we cannot, practically speaking, aggregate lots of very small flows of electricity into one big one. It’s not just total volume of energy production that matters, but the energy density available to high-volume consumers at a given place at and at a given time. This may sound like a dry technical point, but it has huge and nasty implications.
One is that the most touted forms of so-called “alternative energy” and are largely (though not entirely) useless. Solar and wind power are both time-variable and low-density. Lacking good ways to time-shift and aggregate electricity, this means you can’t count on them to run factories and hospitals and computer server farms. The best you can hope for is that they can partially address low-density usage, running climate control and appliances for homes and some purpose-designed office buildings.
Biomass (including processed forms like ethanol) seems more promising because, like fossil fuels, you can burn it when and where you choose in order to to match your electricity production to demand. The problem is that biomass has much lower practial energy density than coal or oil. This means you have to transport and burn a lot more of it, with much larger pollution and life-cycle costs. One of those is much higher CO2 emissions, which are a significant problem for other reasons even if (like me) you don’t believe they’re driving global warming.
Hydroelectric power is, in most respects, ideal — nonpolluting, renewable, all those good things. The trouble is that, at least in the U.S. and elsewhere in the developed world, we’re not going to get any more of it. All the good sites for high-density hydropower are taken. There’s some potential for low-density hydropower, especially in rural areas to run farms.
Geothermal is like hydropower, economically speaking, but requires unusual geology. Basically the only place it can work on a large scale is in Iceland, home of a full third of the world’s active volcanoes.
Accordingly, hydropower and geothermal are not going to support any larger share of what energy economists call the industrial base load — the day-in, day-out demand for high-density power from all those factories and hospitals and server farms — and printing presses, and food-processing operations, and everything else.
The industrial base load is the life blood of technological civilization; without it, we’d have a hideous global population crash, and then revert to pre-1750 conditions in which the economy is almost entirely subsistence farming and life is nasty, brutish, and short. The first question any energy-policy proposal has to address is how to sustain an industrial base load equivalent to today’s — much higher than today’s actually, if we don’t want to condemn the Third World to perpetual poverty. But most advocates of “alternative energy” evade this question, because they don’t like the answers they get when they look at it squarely.
In the real world, there are only three base load sources that matter: coal, oil, and nuclear (hydropower would be a fourth if it weren’t already maxed out). What they have in common is that you can get lots of energy per gram out of the fuel, thus lots of both energy volume and energy density out of one power plant.
Of these three, nuclear has the highest density, then oil, then coal. Both economic arguments and historical evidence tell us that you can’t have an industrial civilization without a fuel that has an energy density at least as high (and thus a cost per unit of energy as low as) coal. Higher density is better, because it means lower cost.
Those costs are not denominated just in money; low-density energy sources are more labor-intensive to operate and that causes more illness and death. Compare annual deaths from coal mining to annual deaths in the petroleum industry to the annual deaths associated with nuclear power; the trend is dramatic and favors higher-density sources, even if you ignore chemical air pollution entirely.
Nothing on offer from advocates of low-density “alternative energy” even comes close to coal as an industrial baseload source. let alone oil or nuclear. Ethanol and hydrogen look like it, until you consider life-cycle costs; basically, making either costs a lot more than mining coal, both in money and in input energy.
Another key point is that, for transportation, oil is basically the only thing we have that will do. For fixed-location power plants, the energy yield per gram of fuel matters a lot and the weight of the plant machinery only a very little. On the other hand, for automobiles and ships and airplanes, power-to-weight ratio matters a lot. Nuclear and coal basically cannot make that cut, cost-insensitive military applications of nuclear notwithstanding.
For fixed-location power plants, however, nuclear is the clear winner. Coal and oil have lower density and serious pollution costs. They are also much less safe. Yes, I said less safe; the historical evidence is extremely clear on this.
There are some kinds of non-polluting fixed-location plants that might become available in the future: notably tidal generators and atmosphere towers. They will probably be good replacements for nuclear down the road, but they’re not an answer to the transportation problem. Oil is non-renewable; the price is rising and eventually (though not in the near term and not as rapidly as predicted by peak-oil-collapse hysterics) it will run out.
And no, electric cars aren’t the answer either; the power to run them has to come from somewhere. The best case is that people will charge them off the grid at night. This will require power plants to be burning just as much additional fuel as if the cars themselves were doing it, perhaps more given transmission losses. What electric cars can do, at best, is give us fuel flexibility, replacing direct oil-burning with nuclear plants and coal. But that’s not going to net out to lower pollution and lower costs unless we build a lot of nuclear plants very quickly. Thanks to decades of scare-mongering and NIMBYism, this probably isn’t possible in the U.S.
The pressing question, then, remains: What’s going to replace oil?
Let’s draw up a specification for the ideal replacement. We’d like a fuel with the energy density of oil, or better. We’d like the only per-unit cost to be sunlight, because that’s the only thing that’s 100% renewable and (unlike tidal, hydropower, and geothermal) available everywhere. Ideally, we’d like the stuff to not require a huge, expensive conversion job on our energy infrastructure.
Happily, I think this spec can be filled. There are demonstration plants making synthetic oil from algae at a per-unit cost not far above that of oil, and plenty of venture capital looking to fund more. As this technology scales up, algal-synfuel costs will drop below that of oil. At that point, the free market will have solved the problem.
It’s largely forgotten now, but we’ve actually been through a similar transition before. In the mid-19th century whale oil was heavily used for lighting and as an industrial feedstock. Prices rose as whales were hunted to near-extinction; fortunately, the stuff proved to be replaceable by petroleum. Yes, that’s right; big oil saved the whales. A century and a half later, pond scum looks likely to save civilization.