There is now little question that humans have caused a worldwide rise in carbon dioxide concentrations which, projected into the rest of this century, will result in catastrophic climate changes around the world. Are we doomed? Or are there solutions on the horizon?
An excellent overview published last year is Innovative Energy Strategies for CO2 Stabilization, edited by Robert G. Watts, based on a 1998 summer workshop at the Aspen Global Climate Change Institute. It reviews the CO2 problem itself, and covers essentially the entire range of possible technical and policy solutions. Energy efficiency, traditional renewables such as hydro power, wind, biomass; more speculative options like tidal power and salinity gradients; even fission power, according to the analyses here, can all contribute something to the problem, but individually or even together seem insufficient to handle the full problem. On the other hand, there are a few options available that could possibly solve the whole energy/environment problem in themselves: solar power, on earth or in space (in particular the Moon); fusion whenever it becomes available; and geoengineering in one form or another (carbon sequestration, stratospheric aerosols, or space-based solar shields). The main question is whether they can solve it in sufficient time and at low enough expense to avoid global economic disaster.
Anybody needing an overview of the entire energy question along with related environmental, economic, and social policy issues, would learn a lot from this book. It's not in an easy style to read, but, as it says on the back cover blurb: "essential reading for all those interested in the development of 'clean' energy technologies, including engineers and physicists of all kinds (electrical, mechanical, chemical, industrial, environmental, nuclear), and industrial leaders and politicians dealing with the energy issue.
The individual chapters that make up this book have been thoughtfully selected to cover essentially the full spectrum of options available, and all address the same global problem on a 50 to 100-year timescale. But the discussion is not entirely coherent: for example energy units vary across the chapters - total energy (for a year, or available from a given resource), can be seen here in terawatt-hours, terawatt-years, or exajoules, in some cases referring to thermal, in other cases electric power requirements. At least the book has consistently abandoned the imperial 'btu', still used elsewhere for example in tables from the US Energy Information Administration, and there's no reference to equivalent barrels of oil or cubic feet of natural gas.
More significantly, each chapter brings the biases of the authors to bear, which actually turns this book into an interesting debate, with each option seen trying to put its best foot forward, backed by real data.
The opening and closing chapters are not so much about energy, but about CO2 - specifically (in the first chapter) the severity of the climate change problem and projections from current trends under a variety of economic/energy model scenarios. And (in the last chapter) the range of "mitigation" approaches available, other than reduction of fossil fuel use (which the middle of the book is all about). This really is a very serious problem for this century. The last chapter points out, however, that we may actually have some technical/engineering tools to handle it. In fact, we have already been inadvertently reducing the warming effects through emission of sulfate aerosols - (from p. 428):
anthropogenic sulfate aerosols in the troposphere currently influence the global radiation budget by around 1 W/m^2 - enough to counter much of the effect of current anthropogenic CO2. [... on side effects ...] one of the many interesting valuation problems posed by geoengineering: How much is a blue sky worth?
But the middle chapters are the key to the puzzle, because whatever we can do about the CO2 effects, we also only have finite resources of fossil fuels available. Much of humanity suffers from lack of access to low-cost energy already. Is there hope?
The first part of the problem, discussed in chapter 2 and repeated with variations in some of the subsequent chapters on detailed energy solutions, is understanding just exactly how much energy the world will need, and under various scenarios, how much CO2 that can be expected to produce. The Intergovernmental Panel on Climate Change (IPCC) has one set of future scenarios; the World Energy Council has another from IIASA (International Institute for Applied Systems Analysis); various others are also cited.
The basic parameters that go into the scenarios are captured in the so-called "Kaya identity"; the rate at which carbon is released is the product of the following terms:
Population growth is a factor, but the largest plausible changes in population expected this century make little difference to the total; we have two far more urgent problems reflected in these numbers:
In other words, the range of values for factor 1 don't make much difference. Factor 2 will, most scenarios expect, increase greatly this century. The overall product has to drop; that means the focus has to be on factors 3 and 4: reducing primary energy intensity (conservation and economic shifts), and reducing carbon intensity (reduction in fossil fuel use).
The third chapter is the only one that really goes into depth on the economic issues; this, by Robert Lempert and Michael Schlesinger (of RAND and the University of Illinois respectively), discusses the basic economic tradeoffs between the possible initial strategies "do a little", and aggressive emissions controls. They find optimal an adaptive policy strategy that includes both taxes (carbon taxes or general energy taxes to promote efficiency) and "technology incentives". The technology incentives are targeted to bring down the costs of emissions reductions - these would include supporting R&D, training people, standardization and certification, funding demonstrations and infrastructure, disseminating information, market liberalization, and tax credits and subsidies to encourage adoption and suitable economies of scale. Under the adaptative scheme, tax rates and technology incentives are adjusted in light of climate damage expectations, economic growth, and technology adoption targets; incentives would expire after a period of time for technologies not making sufficient progress.
If the authors of the various chapters disagree on anything, they clearly disagree on the degree to which we can expect improvements in the third factor: energy intensity. Clearly, there have been tremendous improvements in energy intensity in the past century, and before. But how much further can it go? Most experts (according to Watts, chapter 2) see continued improvements between 0.8 and 1.4% per year; some argue (in particular Hassol, Strachan, and Dowlatabadi in chapter 4) that the numbers could be considerably higher; however in practice efficiency improvements often go to increased capability rather than reduced energy use - for example, average house size has increased significantly in the US, and new houses are loaded with many more devices and appliances.
Including efficiency/intensity improvements then, the first three factors in the Kaya identity determine a target total energy requirement for the world. Given the need for significant reduction in CO2 output (and the fact that fossil fuels will eventually run out), that leaves a target number for non-fossil energy that the various alternatives must meet over the next 100 years. In every scenario this means a dramatic increase in non-fossil energy production, so no matter what else happens, we clearly need these technologies.
First a note on forms of energy. Actually useful energy (work) usually is applied to our devices in a mechanical or electrical form; some industrial processes depend on energy in a chemical form as well. The one major exception is heating, which requires just low quality thermal energy (although heating can be done with greater than one-to-one conversion through use of heat pumps). Wind and hydro plants use mechanical energy; solar photovoltaics generate electrical energy from light which is itself electromagnetic. Interconversion between mechanical and electrical energy through motors and generators is very efficient (90% or more), so they can be viewed as pretty much interchangeable forms. But going from heat to electrical or mechanical energy involves considerable losses; typical steam turbine generators run at about 33% conversion efficiency - i.e. only 1/3 of the thermal energy gets converted to electricity. This factor of 3 between thermal and electrical energy is used quite routinely in the literature; in fact hydroelectric production is typically multiplied by a factor of 3 to give a thermal energy equivalent. The rough rule of thumb then is that an equivalent electrical energy requirement is 1/3 of the thermal requirement.
Interconversion between electrical and chemical energy is intermediate in efficiency - for rechargeable batteries typically about 70-80% (round trip), for fuel cells it can range from 40% to 90% or so. Burning chemicals typically generates just thermal energy, although in a vehicle internal combustion engine or in a gas turbine generator the combustion gases produce mechanical energy directly with (for combined cycle gas turbines) efficiencies of 50% or more. Generally we want to minimize the number of these interconversions between primary energy supply and the final use; for example, burning coal to produce electricity to electrolyze hydrogen to be used in fuel cells to produce electricity to power a vehicle is not likely to be a good use of the primary coal energy.
The actual world energy requirement numbers that come out, finally, both in terms of energy capacity, and capital investment, are enormous. Current (2003) total production of thermal energy worldwide is about 14,000 GW (14 terawatts). With the factor of three rule of thumb, that's equivalent to just under 5000 GW electric. That's 14 TW for a year, every year right now - in energy quantities (multiplying by the number of hours or seconds in a year) that comes to about 120,000 TWh (thermal) which is equivalent to 450 x10^18 joules, or 450 exajoules (EJ) of thermal energy. With the 1/3 rule of thumb, that translates to 40,000 TWh or 150 EJ of electrical energy, every year. Actual world electricity consumption is about 1/3 of that again, or 13,000 TWh per year, right now (the remaining energy consumption is primarily transportation and home and industrial direct use of fossil fuels).
The picture for 2050, then, is an increase in total energy use of between 33% and 140%; 19 to 33 TW (thermal) total. Criswell's scenario (chapter 9; more below) sees a need for up to 60 TW (thermal) by 2050, to address world poverty in an adequate fashion. 60 TW (thermal) was seen as a likely requirement by 2100, under the high growth IIASA scenarios. The main point here is non-fossil fuels have to comprise at least 9 to 10 TW (thermal) of supply by 2050, under any of the scenarios. That's well over half of current energy usage. So the question for the remainder of the book: what (if any) of the nuclear or renewable options available can actually meet this enormous requirement, in a sustainable fashion and at a cost that won't cripple global growth?
9 to 30 TW (thermal), translates to 26,000 - 87,000 TWh/yr or 95 - 320 EJ/yr (electric) of non-fossil energy by 2050. Chapter 5 (Short and Keegan) summarizes the potential from various renewable sources in a table on p. 145: current global renewable use is about 8,000 TWh (electric)/yr - mostly from biomass burning in developing countries, the remainder from hydro power. Long-term "technical potential" for pure solar power is over 280,000 TWh/yr (electric), for biomass over 140,000 TWh/yr, for hydro and wind perhaps 14,000 TWh/yr each, and for geothermal and ocean energy perhaps 2000 TWh/yr. The potential they see economically exploitable by 2025 is primarily in biomass (8-15,000 TWh/yr) and hydro (4-6000 TWh/yr); solar and wind could contribute 1000 to 2000 TWh/yr each. In other words, these renewable sources may be barely sufficient to meet the 2050 demand, but biomass and hydro would continue to be the primary contributions.
For solar and wind we have a serious problem, discussed in chapters 5 and 6 - the intermittency issue. A utility can't rely on the power to be there, and so has to build in (and order ahead of time) sufficient capacity to meet peak needs without taking them into account. Similar issues apply with transmission of power; solar and wind generators would have to pay the capital costs of power lines without being able to make full 24x7 use of them. This currently limits these to well under 20% of supply in most utility systems. The main way to mitigate all this is with power storage: pumped hydro or compressed air are traditional methods, but are very location dependent, among other failings. Other proposed storage methods (flywheels, batteries, electrolysis and fuel cells with hydrogen storage) are relatively costly. Something like 5 TWh of overnight storage would be needed for a system with just 3000 TWh/yr (electric) of solar/wind; at current prices of $100/kWh or more, that amounts to at least $500 billion worth of batteries, which would need replacing every 5 years or so.
Intermittency can also be mitigated by spreading the load across many different supply locations; doing so would require much longer-distance power transmission - a worldwide superconducting grid for example. But capital costs for that also could easily be multiple trillions of dollars, without significant cost improvements.
Two other problems remain with solar and wind - capital cost of the photovoltaics and turbines themselves, and land usage. Actually, land usage for solar may not be worse than for coal, when the area destroyed by coal mining is counted; of course existing coal mines are one of the "lock-in" features of our current fossil fuel dependency, so it's still a problem. But capital costs are the real show-stoppers: for photovoltaics, at $3/peak watt current prices (solar thermal systems are roughly the same), even in an ideal location, 2000 TWh/yr requires about 1 TW peak capacity, or $3 trillion capital investment. For wind, prices are now about $1/peak watt, and capacity factors somewhat better, so the 1000 TWh/yr for wind may require "only" $300 billion investment (but addresses only about 4% of the minimal renewable requirement). Continued cost improvements as manufacturing scales up should cut these costs somewhat.
Annual expenditures on energy systems are already close to $1 trillion/yr, however, so these numbers, while immense, shouldn't be impossible. Doubling or tripling world hydro capacity, as this scenario calls for, would also involve trillions of dollars of capital investment.
But there are three other major energy options that need to be considered to help fill this need for non-fossil energy by 2050, one or all of which may end up being more cost effective and thus less harmful to global economic growth: nuclear fission (chapter 7), fusion (chapter 8), and solar power collected in space rather than on Earth's surface (chapter 9).
What fission, fusion, and space solar all have in common is their ability to directly replace base power supply currently provided by coal (fission already supplies about 0.5 TW of base power worldwide). These are not intermittent, as are wind and terrestrial solar. Even hydro and biomass have seasonal supply variations. Coal has the highest carbon intensity of any fossil fuel; replacement of utility base power generation should be the top priority to combat global warming.
Krakowski and Wilson (chapter 7) give an amazingly thorough review of the situation for nuclear fission energy. The first concern is fuel supply itself. Fission, like fossil energy, relies on a fuel material whose supply may be somewhat limited. Known reserves of uranium (other than low concentrations in granite and seawater) are actually roughly equivalent in energy content to estimated fossil fuel reserves. However, on the fossil fuel side, that equivalence is dominated by oil shale and sub-sea "clathrates", which may not be physically (or environmentally soundly) recoverable. Total fossil fuel is close to a million EJ or about 500 years of the long-run thermal energy requirement in the high-growth scenario (at 60 TW(thermal) by 2100), but not including clathrates and oil shale, that is reduced to about 50 years of reserves at the 2100 rate (i.e. we're not even going to make it to 2100).
But the uranium number also depends on use of breeder reactors and fuel reprocessing to make use of the full energy content of the U-238, as well as the initial U-235 that supplies energy the first time through. Krakowski and Wilson go into a lot of details on proposals to make that as safe as possible and reduce the threat of terrorists or rogue states getting hold of the intermediate plutonium. They wrote, however, before September 11, 2001; even these proposals may not be viable any longer. With just once-through processing, the high-grade uranium known would only last us about 5 years if it were to supply the world's energy at 2100 levels. They speculate that there is a lot of uranium ore still waiting to be discovered, however, if prices were to rise. Reprocessing, use of thorium, and extraction from low concentrations in rock or seawater would increase the supply to thousands of years worth. Whatever the solution it is clera that fission energy, to meet world energy needs, would have a much bigger impact on the world than it has up to now.
Aside from fuel supply, four cardinal issues for expansion of nuclear energy are of concern:
1. Safety - North American reactors have a good safety record; nevertheless the Chernobyl accident demonstrated the devastation that is potentially there. Systems are designed to have a probability of less than one in ten thousand for a core meltdown in any given year; but that could mean one every 5 years if nuclear supplied 2 TW of power, or one per year at the 10 TW or higher level (with roughly 10,000 nuclear plants worldwide). Most scenarios for future fission have kept total supply below 1.5 TW through 2050 for this and other reasons.
2. Waste disposal - there are actually some solutions to this. Sweden apparently has settled on a publicly agreeable disposal solution for their reactors. Fuel reprocessing can extract the worst isotopes and send them back into the fuel cycle to have their energy actually do some good. The remaining fuel could potentially be less hazardous than the original uranium ore. The current situation in the US with once-through fuel cycles and hundreds of supposedly temporary waste storage sites is not reassuring to anybody, though.
3. Proliferation of nuclear weapons - basically eternal vigilence is the key. Will that be possible? Particularly as the industry has to cut costs to be competitive? Proliferation is less of an issue with the once-through cycles, however, since the resultant plutonium stays embedded with other highly radioactive wastes.
4. Costs. In the 1970's-1980's, primarily due to burdensome safety reviews and what the authors term "over-regulation", nuclear power plants in the US came in with costs on the order of $4000/kW or more. Those costs may not be really representative; a new reactor in Taiwan is being built for about $1690 per kW (electric). That still means 1.5 TW of nuclear power (supplying roughly 12,000 TWh (electric) per year of energy) would have a capital cost of $2.5 trillion or so.
Fusion - the picture painted here is one of long-term promise, but probably not helpful by 2050. Two major fusion research reactors are being built over the next decade - the international ITER magnetic confinement reactor (for $5 to 10 billion) and the US National Ignition Facility (NIF - $2 to 5 billion) to study "inertial confinement". Both projects are suitable for scientific and engineering discovery of extreme properties, but disturbingly, neither turns out to be close to a good design for a commercial reactor. Supply of fuel, however, is not a problem: it seems that water by weight contains roughly 100 to 300 times as much energy as gasoline, for the various fusion reactions that may be feasible "early in the third millenium", as the authors phrase it.
So - short story - fusion will likely not be helping much, if at all, by 2050.
Chapter 9 is David Criswell's take on space solar power, in particular heavily promoting his Lunar Solar Power proposal. I'm most familiar with the situation here, and find I have to quibble with some of the numbers he uses; on the other hand, he makes a very persuasive case.
Unlike all the other strategies outlined in the book, Criswell's lunar solar power is (at least according to him) scalable and affordable enough to not only meet all world energy needs as currently projected, but to allow for significant expansion in global world product without environmental harm.
Now, like all the others, Criswell's scheme is a trillion-dollar scale proposal. Unfortunately, unlike the others, it's hard to do it in small steps; this is one giant project. While it seems likely it could be the most economically efficient of all of them, working towards it in smaller chunks seems the only way to make it actually happen. What would those smaller chunks be? Criswell discusses more traditional solar power satellites (in geosynchronous orbit); he dismisses them on a number of grounds (for example, their large size would make the number required for his 20 TW electric rather excessive - however, at a smaller scale they make a lot of sense) and I believe we can now do better than some of his numbers for regular satellites. But he makes a lot of good points.
I actually met Criswell last May - I had invited him as our Saturday keynote speaker at the National Space Society annual meeting in San Jose. His congressional testimony from last fall is an eloquent summary of his position. Of course he's been talking about this for nearly 20 years; what's different now seems to be (1) we need the energy now more than ever, and (2) the space program is really looking for a new goal. Is there any chance something will come of this? We'll see.
Finally, to sum it all up:
(1) Human-generated CO2 and the associated global warming is a big problem for the coming century, although there are some engineering strategies that could (with other side-effects) mitigate it.
(2) We're going to be running out of fossil fuels anyway in the next few centuries; without alternatives, global economic prosperity will be endangered much sooner than that.
(3) Depending on how far efficiency improvements can get us, the mid-century energy requirement from non-fossil sources is between 9 and 30 TW(thermal), or 3 - 10 TW (electric), year-round.
(4) No current renewable technology can provide that power level for less than about $10 trillion in capital investment.
(5) The best plan seems to be an adaptive one: introduce a carbon tax and technology incentives of all sorts for the renewable options, and then adjust both taxes and incentives in response to changing assessments of CO2 damage and non-fossil technological promise.
(6) Wind may be ready for large scale installation; however investments are needed in energy storage and transmission technologies to make it really practical. Biofuels are already in large-scale use: R&D investments to improve their efficiencies, perhaps including genetically engineered crops, should be supported. Solar is a little further away, but R&D there should be strengthened because of the huge potential.
(7) Nuclear fission will be around - we need to decide whether to try to make it a big part, or a small part, of our energy future (i.e. choosing between once-through and breeder fuel cycles).
(8) Fusion likely won't help by mid-century. But the long-term payoff may be large; we should continue to invest moderately in the technology.
(9) Space solar power, whether or not on the Moon, has enormous theoretical potential. Technology incentives to prove its capabilities seem warranted - investments and demonstration projects at least for photovoltaic capabilities, light-weight space construction, space launch, and wireless power transmission. all seem well justified by this and spinoff applications.