by Vaclav Smil
The overall power density of fossil fuel production (roughly 1.85 TW, claiming roughly 7,500 km2) is relatively low, about 250 W/m2. Imports and the processing of fossil fuels (dominated by shipments of crude oil and refined products and by the refining of domestic and imported crude) make a relatively small spatial claim (less than 500 km2), a total smaller than an unavoidable margin of error in estimating fuel extraction claims. America's fossil fuel-fired electricity generation (fuels needed to energize it have already been accounted for) occupies less than 300 km2, and nuclear generation (including the fuel supply) takes up a similar amount of space-and the land requirements of both of these industries are dwarfed by the land flooded by reservoirs used for hydroelectric production and by ROWs of high-voltage transmission lines.
The country's dominant fossil fuel-nuclear-hydro energy supply system-whose domestic primary power output reached about 2 TW in 2010 and whose land claim was on the order of 55,000 km2-had an overall power density of nearly 40 W/m2 as its low-density components (hydroelectric generation and pipeline and transmission ROWs) overwhelmed the high power densities of fossil fuel extraction and processing. Net fossil fuel imports added about 750 GW to the domestic production, and so the power density of the entire system would be about 50 W/m2. As expected, the overall power density of the nascent energy supply delivered by new conversions of renewable energy sources is much lower: the growing triad of wind turbine-generated electrocity, solar electricity, and liquid biofuels reached a bit over 60 GW in 2010, and even after counting only the land actually occupied by wind turbines and their infrastructure and excluding all transmission ROWs the new renewable system delivers with an overall power density of just 0.4 W/m2, less than 1/100th of the currently dominant arrangements.
Additional annual land claims cannot be estimated by simply applying appropriate power densities to specific expansions of fuel extractions and electricity generation or to an extension of pipelines and high-voltage lines. This is why such simplistic extrapolation would end up with substantial errors: in some localities the land disturbances created by surface coal mining are more than matched by the mandatory reclamation of old abandoned wasteland; directional drilling of multiple wells from a single well pad reduces the specific land claims of new wells; new refinery capacities and new natural gas-fired electricity generation can be accommodated within the sites of existing facilities; and some new transmission lines and some new pipelines can use, fully or in part, existing ROWs. As a result, my best estimate is that recent net annual additions (for land actually transformed by new energy projects and for new ROWs) have been less than 500 km2.
I will close this section in the same way as I did the previous one, by contrasting the land claims of US energy production with other important land uses (fig. 7.8). Here they are in descending order: protected areas in forested areas are just over 3,250,000 km2 (33% of the nation), and arable land and permanent crops occupy 1,630,000 km2 (about 17% of the total), while the recently harvested area of annual crops has averaged less than 1,400,000 km2. As I have already explained, urban areas (including also grassy and treed surfaces) take up about 150,000 km2, and impervious surface areas (just buildings and paved surfaces) amount to about 50,000 km2.
Figure 7.8
Comparisons of US land use in 2010. Carl De Torres Graphic Design.
This means that the US fossil fuel-nuclear-hydro energy system (including all ROWs) is slightly more extensive than the country's impervious surfaces, and that the land taken up by cities (including their green surfaces) is roughly three times as large as the dominant energy supply system consisting of fossil fuels and nuclear and hydroelectricity. Inversely, this means that for every 3 m2 of urban areas (where most of the modern energies are used) there is roughly 1 m2 of land that is either transformed or whose other uses are largely preempted by the extraction, conversion, transportation, and transmission components of the US fossil fuel-hydro-nuclear energy system.
In contrast, in 2010 ethanol delivered an equivalent of only about 5% of the country's motor gasoline supply but the production of its primary feedstock (grain corn) required roughly four times as much land as the entire US fossil fuel-hydro-nuclear energy system. This brings me to the closing chapter of the book, in which I will look at how much land these new renewables would require if they were to entirely displace our current reliance on fossil fuels and nuclear fission. In other words, I will examine the unfolding energy transition in terms of shifting power densities.
Unfolding energy transition changes a system dominated by fossil fuels and thermal electricity generation to new arrangements in which renewable energy conversions become more important. In this chapter I look at this transition through the prism of power densities and offer some conclusions about the import and consequences of these new realities.
We are at the beginning of another epochal shift to new sources of energy. The timing of the first shift, to using energy other than our muscular exertion, cannot be determined with accuracy as the earliest confirmed dates of the controlled use of fire have been steadily receding and now appear to be at least 105 years ago (Smil 2013a). The second shift began to take place more than 7,000 years ago with the domestication of draft animals: their greater power was a critical component in the evolution of traditional agriculture, land transport, and construction. Only millennia later did the settled societies added the limited use of wind and water power harnessed by sails, windmills, and waterwheels. Some combination of these traditional energy sources characterized all preindustrial societies. This pattern began to change only in the early modern era with the rising combustion of coal in the UK and in a few European regions; after 1860 hydrocarbons were added to the rapidly expanding use of coal, and the introduction of internal combustion engines and thermal electricity generation accelerated the transition from phytomass to fossil fuels.
Modern civilization is thus a material and intellectual embodiment of converting fossil fuels into the useful energies of heat, electricity, motion, and chemical potential. While photosynthesis remains the planet's most important energy conversion, as it powers all life (with the exception of deep-sea organisms aggregating near hot thermal vents), our relative reliance on phytomass fuels has been steadily declining. In 1850 coal, the only commercially extracted fossil fuel, supplied less than 10% of the world's primary energy, and during the closing years of the nineteenth century the gross energy content of traditional phytomass fuels (wood and charcoal and straw) and fossil fuels became even.
At the beginning of the twenty-first century, fossil fuels (with crude oil slightly ahead of coal, followed by natural gas) supplied just over 80% of the world's primary energy (including noncommercial biomass burned by poor households), and, because of their more efficient conversions in boilers, furnaces, and engines, they provided more than 90% of all useful energy (Smil 2010b). The worldwide share of fossil fuels in commercial energy supply was almost 87% in 2013, when fossil energies supplied 93% of Japan's, 90% of China's, 89% of Russia's, 86% of the United States', and 83% of Germany's primary energy demand, with the largest share of nonfossil energies originating either in nuclear power (in the United States and Russia) or in hydropower (in China and Japan).
New kinds of renewables-dominated by wind turbines and PV cells in Germany and China and by liquid biofuels in the United States-reached the highest share of the overall supply in Germany, where preferential policies had pushed their share to 9.1% of all primary energy by 2013 (BP 2014). Elsewhere the contributions of new renewables were much lower, amounting to 2.6% in the United States, 2% in Japan, and 1.5% in China, while in Russia (where they accounted for a mere 0.014% of the total supply) these alternatives were essentially absent. These realities make it clear that the transition from fossil fuels to new renewables, much like all previous energy transitions, will be a gradual, protracted affair that will take place over many decades (Smil 2010b).
Two arguments are advanced in favor of an accelerated shift away from fossil fuels: conc
erns about their future supply and worries about the longterm environmental consequences of their combustion. The first concern has been around for generations but has received more attention recently, thanks to many publications claiming the imminent arrival of peak oil, peak coal, and peak everything (Smil 2010a; 2013b). Production realities do not reflect any convincing signs of an imminent global peak for any fossil fuel extraction, and the best appraisals of remaining coal and hydrocarbon resources indicate a sufficient supply for decades to come: neither the leading government agencies (USDOE 2013a) nor large energy companies (ExxonMobil 2013b) assume otherwise.
The argument can then shift to the quality and cost of fossil fuels. Running out of fossil fuels (of any quality and producible at any cost) may be a slow process with no imminent end, but diminishing reserves in many of the most accessible deposits in the richest coal seams and hydrocarbon reservoirs in the areas that have been explored and exploited for generations are an undeniable reality. One of the latest impressive demonstrations has been the high cost of bringing new crude oil to the market: in 2013 no new major oil project (with at least 300 Mb in estimated lifetime output) added in the previous two years had a breakeven cost below $70 per barrel (Goldman Sachs 2013).
The environmental argument for an accelerated shift away from fossil fuels-the need to slow down anthropogenic global warming-has dominated the quest for new renewables since the late 1980s. The only two technical solutions that would obviate global decarbonization would be the massive development of nuclear power or an equally large-scale effort to capture and sequester CO2 generated by fossil fuel combustion. Both of these options have their strong advocates, but there is no evidence of any commensurate deployment. Nuclear power is either stalled (United States) or on the way out (EU), with more than 40% of all reactors (66) under construction in China and 75% of them being built in just four countries, China, Russia, India, and South Korea (IAEA 2012; Schneider et al. 2013). And most of the proposed carbon capture and sequestration projects have been shelved or postponed (Global CCS Institute 2014). There is a third option, climate engineering, but its efficacy, practicality, and acceptance are even more uncertain (Keith 2013).
By 2012 nearly 140 countries had policy targets for an energy supply provided by renewable sources, ranging from modest shifts to profound changes: in the UK the target is 15% of all energy from renewable sources by 2020; in Germany it is 30% by 2030 and 80% by 2050 (REN21 2013). Moreover, some studies have argued that it would be perfectly possible to have all new energy demand supplied from renewable sources (wind, water, and solar) by 2030 and the world's total energy demand coming from renewable sources by 2050 (Jacobson and Delucchi 2011). The latest incarnation of this plan sees the state of New York completely energized by nothing but renewably generated electricity by 2030 (Jacobson 2013), a plan found by Dodge (2013) to lack any technical credibility as it assumes unrealistic performance factors and the mass deployment of unproven conversion techniques.
I have shown in a great detail that no worldwide energy transition has ever been-indeed, cannot be-so rapid (Smil 2010b), and that since the rise of coal none has ever resulted in complete domination by a single energy source. Apparently, such realities do not matter where theoretical calculations reign. Of course, a theoretical option might consider only electricity generation: if it were exceptionally inexpensive, that electricity could be used as the foundation of a new hydrogen economy, a solution that has been advocated by some enthusiasts for decades (Ball and Wietschel 2009; Dickson, Ryan, and Smulyan 1977).
In reality, we could not convert every convertible industrial, transportation, residential, and commercial energy use to electricity in just 15 years, and we would still need fuels for iron smelting, for feedstocks (needed to replace fossil hydrocarbons that now dominate many chemical syntheses), and, of course, to energize our land, water, and air transportation: the world's fleets of heavy trucks or container ships appear unlikely to run on electricity or fuel cells anytime soon, and electric or cell-powered commercial jetliners are pure science fiction. That is why any successful transition to new renewable energy sources would have to deliver both electricity and fuels. And, obviously, it would have to do so on multi-gigawatt scales in large countries and on a terawatt scale globally. I will examine first the spatial implications of displacing all liquid fossil fuels by biofuels and all metallurgical coke by charcoal and then take a closer look at wind and solar power electricity in the unfolding energy transition before closing the chapter with quantitative illustrations of what it would take to create an energy system based completely on renewable sources.
Biofuels
Proponents of new renewable energy sources argue that this challenge can be met by relying, as we had for millennia, on phytomass-but doing so by cultivating it with much higher yields, through using a wider variety of plants and converting them to fuels with higher efficiencies. But in in no other instance of displacing fossil fuels by renewable energy sources is there such a mismatch as between the power densities of producing oil-based liquid fuels and the power densities of producing biofuels: the extraction of crude oil and the harvest of phytomass feedstocks differ by at least two, most commonly by three, and in many cases by four orders of magnitude. Moreover, the scale of the needed substitution (roughly 2.5 Gt of transportation liquids in 2012) is enormous: even if we assume no increase or even a slight decrease in the US demand (USEIA 2013a) and restrained growth of the current liquid fuel requirements in Asia, the annual supply of liquid fuels would have to amount to at least 3 Gt (that is, 4 TW) by 2040 (ExxonMobil 2013b).
Many studies have endeavored to show that such mass-scale biofuel futures are possible. When Berndes, Hoogwijk, and van den Broek (2003) reviewed all published long-range forecasts of biomass energy contributions they found that the maxima for the year 2050 ranged from less than 3 TW to nearly 13 TW, the latter total being higher than the world's total primary energy supply (TPES, including traditional biofuels) in the year 2000. Three years later Moreira (2006) claimed that phytomass can eventually supply about 32 TW, which is more than twice as much as all fossil fuels produced in 2012. But all of these high claims rest on overly enthusiastic constructs that assume unrealistically high phytomass yields and exceptional conversion efficiency maxima, posit an abundance of requisite land, and fail to examine a number of other critical constraints. In reality, biofuel production would be constrained by many factors, and the consequences of shifting from fossil fuels to biofuels would be profound. Power densities help to explain some of these outcomes.
I have always been skeptical about any claims of grandiose phytomass futures. In 1983 I concluded my book on biomass energies by writing that "there are countless better uses for plants than to burn them directly or to use them as feedstocks to make fancier fuels out of them" (Smil 1983, 417), and during the past three decades nothing has taken place to make me change that conclusion. Phytomass will continue to be a non-negligible contributor to global energy supply for generations to come, but it cannot provide large fractions (a third, one-half, two-thirds) of that rising total, and differences between theoretical appraisals of potentially available phytomass and realistic opportunities for its use should always be kept in mind.
For example, in 2005 a study commissioned by the US Department of Energy concluded that the country has an annual supply of one billion dry tons (that is, 900 Mt) of agricultural and forest phytomass (USDOE 2005), enough to displace about 30% of oil consumption at that time. But no component (crops, wood, residues, wastes) was restricted by the cost of cultivation and harvesting, and all referred to phytomass on farms or in forests and excluded all transportation and storage costs, handling losses, and quality deterioration. This resulted in a misleadingly huge mass that shrinks on closer examination. The study's update acknowledged those shortcomings and offered a more rigorous assessment (USDOE 2011a). Or, as Sinclair (2009, 407) put it, "while biofuels can be a contributor to the energy needs of the future, realistic assessments of the prod
uction challenges and costs ahead impose major limits."
Liquid Biofuels
Crude oil extraction proceeds mostly with densities of 102-103 W/m2, and in the largest fields the rate goes up well into 104 W/m2-while the dominant feedstocks for the production of liquid biofuels are harvested with power densities of 10-1 W/m2 and the rates are further reduced by their processing. A power density of 0.23 W/m2, typical for US corn-based ethanol, demonstrates the inherent limits of America's principal alternative automotive fuel. If all of America's gasoline demand in 2012 (a total of 16.96 EJ, or 537.87 GW) were to be supplied by corn-based ethanol produced with that power density, then the United States would have to be growing corn for ethanol on 234 Mha, an area nearly 75% larger than that of all recently cultivated land and a third larger than the country's total cropland (USDA 2013a).
Ethanol's low power density also causes extensive qualitative environmental changes associated with the fuel's production (Howarth and Bringezu 2009; Smil 2010a). Corn cultivation's environmental impacts include soil erosion from the cultivation of a row crop; the demand for heavy applications of nitrogen, averaging more than 150 kg/ha and surpassing 200 kg/ha in the Corn Belt; the ensuing nitrogen leaching, causing the eutrophication of coastal waters and an expanding dead zone in the Gulf of Mexico; the depletion of aquifers for irrigation; the expansion of monocultures as traditional rotations with soybeans and alfalfa; and a net increase in greenhouse gas emissions. In addition, ethanol distilleries discharge large volumes of wastewater (10-13 times the volume of the produced fuel), and those discharges increase oxygen demand in streams and water bodies.
