Presenting information: an energy example

(return to Environment chapter)

The purpose of this piece is to show how presenting information can help decision-making, for instance before voting. Most of this piece provides background, the sort of thing a voter could look up by following links in voter education material. Ideally, information could be pursued to any depth desired. I’ve tried to write in that spirit with links to original or at least authoritative sources, but it’s beyond my resources to do that to the standard I would hope to see in a government agency or a watchdog organization. The idea here is just to give a general example, not an exhaustive one. The actual example of what might appear at the top level, the least “effortful” one, is the diagram at the end. The rest is a second layer of more detailed information, and some of the links lead to yet deeper layers. Let’s assume a broad policy decision is needed about which forms of energy should receive the most development and resources. The calculus in these cases is always to get the most benefit for the least cost. (It is, interestingly enough, very difficult to find comparable numbers on the total cost of different forms of energy. That right there is an example of why good information is essential to good decisions, and an uncomfortably clear indication of why we have so many bad decisions on these issues.) So let’s get to it. The cost – benefit factors for each form of energy are constrained by the total energy available from a given source and the amount that can be recovered in practice, by the cost of its infrastructure, and by its environmental and medical consequences. (An earlier and more detailed summary of energy pros and cons is here.)

Projected Needs

First, which sources of energy are large enough to satisfy current and projected needs? There is, after all, no point devoting disproportionate resources to a source that’s prevented by the laws of physics from delivering benefits that exceed the costs. The world’s current energy budget is around 14 terawatts per year (e.g. MIT Energy Research Council, 2006 (pdf)). However, since this exercise is to decide on a course of action for the future, projected needs are even more important. There are a number of different variables to consider. Does one include improved conditions for the world’s poor? If so, they’ll be using more energy and the estimate goes up. How much growth is expected? Or should we be talking about how much is desired? The Rolls Royce estimate of energy needs in 2050, the one assuming growth for everyone unlimited by the availability of energy or irritations like global warming, is around 60TW per year (pdf). The “let’s all huddle over a cold grate” estimate is only somewhat more than we use now. The estimate currently considered acceptable assumes a 2% growth rate, which is supposed to provide some improvement in developing countries, allow for expected population growth, and maintain standards in the developed world. (IEA, World Energy Outlook 2004 (pdf), and OECD, “Energy: The Next Fifty Years” (pdf), Belcher & Nocera, MIT lecture, and one of the linchpin articles: Hoffert et al., 1998, in Nature as well as a popular take on it.). Those estimates converge on approximately 28 terawatts needed in 2050. Looking beyond 2050 and assuming continued growth, the need is higher. Projected needs including efficiency improvements. But — and this is really important, so I’ll put it in caps — BUT if energy efficiency were applied to the extent possible in transportation, buildings, and industrial processes, the estimate of energy needed drops. In that case, assuming 2% growth for everyone, including the poor, projected needs for 2100 fall to 7TW. (Also see e.g. Rosenfeld (pdf), and American Solar Energy Society, 2008(pdf).) Seven terawatts. For a better standard of living than we have now. Our biggest “source” of energy is simply not wasting it. (At 2050 we’d still be in transition because of how long it takes to replace old inefficient buildings. Demand would be at its peak, 21TW, which even then is less than 28TW.) So the range of estimates of energy required a few decades hence ranges from 7TW to more than 60TW, depending on the path chosen.

Fossil fuels

These are the most familiar forms of energy, and the ones whose pros and cons require little discussion. In terms of their ability to provide the needed energy, right now only oil is capable of providing for any and all energy needs, with a strong assist from coal and natural gas. Unlike other sources of energy, the physical plant is in place for fossil fuels and there’s no long construction lead time as there is in varying degrees for newer types of sources. On the minus side, all fossil fuels have associated environmental coss. Fossil fuels are not sustainable, so their cost has to include the cost of moving to something that does not mortgage the future. An often overlooked point is that costs of fossil versus alternative fuels are generally calculated on a non-comparable basis. “The cost of conventional fuel is calculated on a marginal basis while alternative fuel costs are calculated on a fixed cost basis. Meaning the cost of roads, trucks, and mining equipment is not factored into the price of each piece of coal, only the marginal cost of producing each ton of coal. For solar and wind energy systems, the cost to construct the system is factored into the total cost while the marginal cost of producing electric is virtually free.” In order to make a real comparison of all fuels, the fixed costs of fossil fuels need to be included. Such as all the roads which are essential to their distribution, for instance. Otherwise one is comparing apples and cheese doodles, and has no real sense of the actual price paid for a technology. It’s not enough to say that all the fossil fuel infrastructure is already paid for and no longer matters. At some point it all has to be replaced, repaired, repaved. I’ve included a factor for environmental and health costs, but not things like highway and rail infrastructure since I couldn’t find any estimates for that. It’s worth remembering that the cost-benefit ratio for fossil fuels has nowhere to go but worse, as bad as it is. If we had fair pricing now, one that budgeted for the minuses as well as the pluses, fossil fuels would be nearly the most expensive choice, not the cheapest.

Nuclear Energy

Nuclear fuel, contrary to popular image, has the same problem of unsustainability as fossil fuel. (A more detailed discussion of the issues with nuclear energy, with links to sources, is here.) Uranium is a finite resource. Using it on the scale required to supply much of the world’s energy means it would run out within several decades. So nuclear energy, too, would have to include the cost of transition. It would also have to include the cost of construction, of decommissioning, of waste storage, of security measures, and of medical expenses. Using breeder reactors is sometimes mooted as a solution to the sustainability problem, but the price would have to include much larger funds for waste storage, environmental damage, decommissioning, medical expenses, and the effects of proliferation. Those radiation-related problems are well-known, even when ignored. It’s less appreciated that nuclear energy has too long a lead time to be a practical way out of the energy crisis. The US consumes about a quarter of the world’s energy, so its share is around three and a half terawatts. One seventh of the current requirement, i.e about 14%, is half a terawatt, or 500 gigawatts. For the US to supply that rather small amount of current needs from nuclear power the country would need to build a new gigawatt nuclear plant every month or so for the next thirty five years. Since reactors last some thirty years, replacements would have to start being built before the project was finished. The construction and decommissioning schedule, and their expense, would have to be continued until the uranium ran out. All that effort would yield a decreasing proportion of overall power as needs increased over time. (Nuclear energy, at least so far, has not had a pricing structure that rewards small-scale, distributed methods like efficiency. Even though the laws of physics don’t make them mutually exclusive, efficiency is a low priority in nuclear-driven scenarios. Hence the assumption that need would increase with time.) Since nuclear reactors cannot produce more than a fraction of the energy needed, even allowing for an improbably brisk pace of construction, they cannot be more than a small and temporary part of the energy picture. At that level of usage — supplying one seventh of energy needs — the uranium which it’s practical to extract would be gone in about fifty to one hundred years, assuming no new finds of large deposits. After the uranium ran out, the cost of decommissioning and waste storage would continue. Decommissioning and construction of waste facilities are both very expensive, but after that the costs wouldn’t be large in any given year, barring accidents. However, they would continue for tens of thousands of years. (Maybe less after a recent breakthrough, although that would involve other construction and decommissioning costs.) Given an obligation on the part of the people who get the energy to pay the costs, those would all come due over the thirty year life of the reactor.


Biofuels have some ramp-up costs and, like any combustible fuel that produces secondary compounds, have environmental implications. Let’s assume for the sake of argument that they’re produced only from agricultural waste or ecologically responsible algal plants, so that we can avoid considering the social costs of taking land away from food production or the environmental costs of destroying natural areas. The social and environmental costs are the very opposite of trivial under current methods of biofuel production. The theoretical maximum of energy obtainable from biofuels is equivalent to burning every formerly living thing on the planet, so it’s irrelevant. The technically feasible maximum depends on what’s acceptable. If biofuels are made from crops and all arable land is used, the limit is about seven terawatts. Depending on how important it is to eat, the practically recoverable portion is correspondingly less. If only agricultural waste or algal products are used, then it’s much less. Another tangential factor is the amount of fresh water needed to make biofuels, since fresh water is also in increasingly short supply. Biofuels, therefore, can only be a small part of the energy budget. That doesn’t mean they’re useless, but it does mean there’s no point hoping biofuels could solve, for instance, the transportation fuel problem by themselves.

“Minor” renewables

Other renewables, wind, ocean power, and hydroelectric, also have maximum limits on the amount of energy they can deliver. Those limits can be pushed to some extent as we get better at extracting the available energy, but they’re all about an order of magnitude away from actually satisfying current needs. Wind, for instance, has a theoretical maximum around 870TW according to one estimate, or 300TW (ppt) according to another. The practically extractable amount, however, is about 2-4TW. That number could rise as we learned to extract energy from lower speed winds [2015: link not working, but here is a similar one from the same approximate year], but it would not rise tenfold or by enough to allow wind power to be a universal solution to energy needs. The limit is based on the total wind energy available in places where we could use it. The numbers for some other sources are taken from the same powerpoint presentation linked above by Lewis of Caltech. Ocean power, including currents, tides, and waves: practically recoverable maximum of around 3TW. Hydroelectricity: 4.6TW theoretically available, 1.5TW practically recoverable. So, again, these renewables can be partial solutions, and may be the best solutions in given cases, but they can’t be the whole solution. Geothermal is one renewable source that does have the theoretical potential (42TW) to meet reasonable future needs, but there are a number of caveats on that figure. The ocean’s potential is 30TW of that. Land-based potential is closer to 12TW. If the heat is bled off too quickly, the site needs time to warm back up, so it’s renewable, but not inexhaustible. Geothermal comes in two flavors: small scale shallow heat pumps that could be used in many places, and much higher yielding deep boreholes that depend on heat from past or current volcanic processes. The water used in the latter will wind up with lots of dissolved, corrosive minerals and a goodly amount of toxics such as sulfur compounds, mercury, or arsenic. The boreholes go over 10km deep, which is currently too expensive to drill economically just for heat (as opposed to oil). It would require yet-to-be-invented breakthroughs to make such drilling feasible on an industrial scale. There’s also the disadvantage that these are large, complex, not easily accessible construction projects with long lead times. Again, in practice, geothermal may provide some useful power in some places, but it’s a long way from being an easy or universal solution.


And finally we come to the elephant in the room: solar energy. Its theoretical maximum is over 12,000 TW per year, the theoretically recoverable portion is on the order of 600TW, and the practically recoverable amount given current technology is around 60TW. It could meet all current and projected energy needs several times over. It is the only source of energy that can do that. In the near term, using it would require significant construction, revamping of transmission technology, and significant changes to storage technology. None of these are trivial, but neither are they overwhelming. Solar energy does have environmental consequences because of the toxic and nonrenewable compounds currently used in the manufacture of photovoltaic panels, transmission lines, and batteries. Those can be mitigated by controlling pollution at the plants, devoting research to organic-based PV (which is already showing plenty of promise), and improved battery technology and recycling. Solar is also the only energy source (besides efficiency) which can be unobtrusive enough to live with. It can be produced right where people need it. That makes its social effects unusually benign in terms of distributing work and wealth. Logically, it seems to me, the type of price one pays should also figure into the decision of which one it’s preferable to pay. Or, to put it another way, I’d rather pay with money than by getting cancer. One argument against rooftop solar is that there aren’t enough roofs to go around. Lewis, for instance, points out that covering all single family homes would only provide 10% of our energy. But single family homes aren’t the only flat surfaces around. There are also multifamily and apartment blocks, malls, parking lots, warehouses, factories. Also, as the capability to use thin films and light-trapping dyes improves, windows and walls can become photovoltaic (Science, 2008). We’re not limited to roofs. Distributed solar is said to be less efficient, and it is, purely in terms of amount of sun striking a cell and wattage produced. However, add in ease of construction, maintenance, and recycling, as well as how much easier it is to distribute the effort among millions of people, and the systemwide total cost of ownership is lower. The utilities, whom regulations reward for centralizing power, are the ones who argue against distributed solutions. Regulations, however, are not laws of nature. Unlike those, regulations can be changed. In some cases, centralized power is not just a profit center; it’s actually necessary. Ore smelting for instance, requires a large amount of concentrated energy and therefore a potentially a centralized solution. Solar thermal turbines are one way to do that. Alternatively, much better distribution and storage technology could provide large delivery of power to huge consumers, too. I haven’t mentioned hydrogen because it’s basically a storage technology. It is, in effect, a replacement for batteries and, as such, shares the cost-benefit profile of its energy source. Hydrogen made from coal (the current dominant method) is really a fossil fuel. Hydrogen made by fuel cells using solar energy (not yet near becoming a production technology) is stored solar energy and shares that cost – benefit profile.

Summary diagram

Putting all the information together yields a range of choices that provide more or less energy per unit cost, each with their own characteristics that may make a more expensive option a good choice under specific circumstances. The per kilowatt numbers are necessarily approximate because they fluctuate based on factors other than energy source. For instance, there was an interesting article in the NYTimes, Jan. 8, 2007 discussing how price for electricity can be anywhere from one cent to fifty cents per kwH, and that’s these days, with our boring narrow choice of energy sources. However, the modal or “overall” cost based on source is still a vitally important number. All the other fluctuations are on top of that, so a generally cheaper source will always be cheaper than an expensive source under the same conditions. That overall cost-benefit distribution of energy sources is a direct implication of the physics involved: of the theoretical maximum available energy, of its distribution, of the processes required to make it usable, and of the construction costs due to those characteristics. Since we’re talking about orders of magnitude differences in costs and energy production, exact estimates aren’t even important. An energy source could move quite a bit within its quadrant, but short of new ways around the laws of physics, it won’t move out of it. A grasp of that overall picture would faciltate making decisions. Figure 2. Relative costs and possible total energy yields.
Solar energy and efficiency provide the most bang for the energy buck, with increasingly favorable cost-benefit into the future.  Nuclear provides a limited amount of the most expensive energy, with increasing costs for decreasing energy output into the future.  Fossil fuels don't fare much better due to environmental and health costs and global effects.

Notes. Maximum recoverable energy over time determines placement on yield scale. Any finite source has relatively low maximum energy compared to some non-finite sources. The cost-benefit of any energy source will, of course, change relative to others over time (e.g. NREL). This has been visually expressed in the case of nuclear and solar.

Nuclear: more than 30¢ / kwH. Components: mining, processing, transport, construction, operation, power distribution, liability insurance, immediate environmental costs, long term environmental costs, medical costs of carcinogens, decommissioning, waste processing, waste storage. (Severance, 2007, 2009) Costs will continue even when no power is produced, at which point each kwH is infinitely expensive. No chance of providing more than a fraction of needed energy, hence low on the energy scale.

Fossil fuels. Finite, hence can’t provide needed energy into the future, so low on the energy scale.
coal: 3¢ / kwH charged for fuel costs. With all components, more than 15¢ / kwH? More than 20¢ / kwH? And up. Components: mining, environmental damage from mining, processing, coal sludge tailings,transport, environmental costs of acid rain, mercury pollution, global warming, air pollution, medical costs of asthma, cardiovascular disease, carcinogens. Effective CO2 capture would include costs of capture, pipeline and transport, storage and sequestration, liability insurance for suffocating leaks and explosions. CO2 capture alone adds 3¢ / kwH to cost. Planet-wide damage by global warming is a potentially de facto infinite cost.
oil: 5¢ / kwH for fuel costs. See coal for hidden subsidies.
natural gas: 9¢ / kwH for fuel costs. Components: mining, pipelines, transport, global warming. CO2 capture: see coal.

Biofuels: 10¢ / kwH ±5¢ depending on source. Currently uses significant petroleum inputs to produce, although if biofuel is produced from waste that is less of an issue. Other components: transport, processing, air pollution, waste. If made from grain: massive social impact which puts biofuel in the upper left zone.

Geothermal: 5¢ – 8¢ / kwH for “low power” geothermal (i.e. not using deep drilling or exotic heat exchange fluids). (NREL)(pdf) Components: construction, if in a volcanic zone then environmental costs. Low power geothermal is limited to areas with significant near-surface heat, but there are many of them. “High power” geothermal would require breakthroughs in drilling technology to make it feasible on an industrial scale. It would also have more potential environmental effects.

Other renewables. Maximum energy recoverable in theory is a fraction of the energy needed. Wind: 10¢ / kwH at good sites. Environmental effects for people living nearby, and potentially for birds or bats.
Hydro: 5¢ / kwH but varies depending on site. High power hydro, other components: Construction, distribution, environmental damage.
Tidal: (I haven’t seen cost estimates. My guess: around 10 – 15¢ / kwH, at least initially, falling with time as technology improves.)

Efficiency. 1 – 3¢ / kwH. Other costs: replacement of old stock that can’t be retrofitted for efficient use, therefore the expense, energy and pollution costs of the manufacturing involved. However, the need for goods and services also increases overall GDP. Well-managed efficiency in some situations has actually made rather than cost money.

Solar. 10 – 30¢ / kwH, currently commonly around 15¢ / kwH. (Includes all types: photovoltaic, thermal, concentrated.) (NREL)(pdf) Other costs: environmental if sited in pristine wilderness, pollution from rare earths in silicon-based PV technology. Energy storage is a major consideration. If widely applied, distribution can be a major consideration. Unlike any other energy source, has the capacity to provide for all current and projected energy needs.

(return to Environment chapter)

So far, I can't make comments appear on archive pages. Please click on the post title or permalink to go to the post page, if you'd like to comment.