by Vaclav Smil
Reservoir pressure must be maintained by keeping sufficient volume of cushion (base) gas (typically about 50% of the entire storage). That gas remains unrecoverable and the capacity of reservoirs is measured only in terms of working gas that can be withdrawn according to demand. Large volume of gas that can be held in depleted gas reservoirs makes them perfect for securing seasonal requirements, with injections done during summers and gradual withdrawals (these storages cannot deliver sudden, rapid outflows) continuing between November and March. This means that these reservoirs have usually only once-a-year turnover unlike salt caverns, the storages used to supply peak demand whose turnover is only weeks or even just days. The US now has about 350 storages in depleted reservoirs, mostly in seasonally cold Northeast and Midwest.
Salt caverns are the second best option for storing large volumes of gas: their impermeability and structural strength guarantee their integrity and longevity, but they must be first formed by drilling wells and injecting water to dissolve enough salt to create suitably sized spaces inside natural salt domes at depths mostly between 500 and 1,500 m. This is costly, but it creates a storage that has the highest reliability: as long as a cavern contains enough cushion gas (about a third of a cavern’s total capacity) to keep the entire volume under sufficient pressure, the needed gas can begin to flow rapidly (in less than an hour) and at a fairly high rate, and hence, they are best used to cover peak load needs; that is why their numbers have been steadily growing along with the expansion of gas-fueled electricity generation (see the next chapter). At the same time, gas volumes stored in salt caverns are too small compared to those held by depleted reservoirs and are not used for base load storage. The United States now has about 50 salt-cavern storages, with the largest number in Texas and Louisiana, the two states with numerous salt domes.
In the regions that are far from any existing or former gas-producing fields and that are also devoid of any natural salt formation, gas must be stored in aquifers. These water-holding strata are obviously porous and permeable, but it is not easy to ascertain their overall size and porosity and to find their potential storage volume. Aquifers chosen for gas storage must have their water-bearing sedimentary formation overlaid with an impermeable rock cap in order to prevent gas escape into aquifers tapped for drinking water. Aquifer storage also requires much larger volumes of cushion gas (up to 80% of the total capacity) in order to maintain desirable rate of outflow; in some cases, high-pressure injection must be used to displace water by gas; in virtually all cases, the withdrawn gas must be dehydrated; and there may be considerable long-term losses as injected gas seeps out of the aquifer layers.
US data offer weekly or monthly monitoring of all key storage variables including capacity, storage by type, base, and working gas inventories (USEIA, 2014d). In 2012, depleted reservoirs accounted for about 85% of total storage volume, aquifers for 10%, and salt caverns for the rest; the analogical shares of working volume were about 81, 9, and 10%; and in terms of actual delivered gas, the shares were 75:10:15 (USEIA, 2014d). The total storage capacity in the United States reached to more than 240 Gm3 by 2013, equal to roughly a third of annual demand in that year, and about half of it was the working storage. Since the year 2000, actual stored volume fluctuated between October maxima of about 200 Gm3 and summer minima of about 135 Gm3.
Gas liquefaction made it possible to set up aboveground storages of natural gas: LNG tanks can store only limited volumes of the gas, but it can be instantly available. In the United States, first such facility was built in Cleveland in 1941 (for more, see Chapter 5), and by the year 2000, the United States had about 100 LNG storages (with or without liquefaction), concentrated in regions lacking large underground storages (New England and coastal Mid-Atlantic) and used to cover peak demand. In 2012, total additions to these LNG storages amounted to less than 1 Gm3, a tiny fraction of underground storages but an important option for peaking demand.
3.3 CHANGING PRODUCTION
For decades, associated gas accompanying crude oil production was an unwelcome and unwanted by-product. Before the advent of long-distance pipelines, it could have only limited local uses: in the United States, this began to change before WWII, but in the Middle East this was true even in the early 1970s. Faisal al-Suwaidi, the former CEO of Qatargas, recalls how in 1971 the discovery of the world’s largest natural gas field, the North Dome, was met with dismay: why Qatar has not been so lucky as the neighboring oil-rich Saudis? At that time, with commercial gas liquefaction in its earliest stages and with no large global market for LNG, gas associated with the extraction of crude oil could be used locally for enhanced oil recovery or burned for heat or electricity generation, but huge volumes of nonassociated gas were more of a problem than an opportunity (in the fifth chapter, I will note how al-Suwaidi engineered the gas-based transformation that made Qatar the richest state in the world).
In the absence of major local demand, the extracted gas had to be simply vented or flared, which is piped into tall stacks and burned. This wasteful and environmentally damaging practice (see Chapter 7) has been greatly reduced since the 1970s, but there are still considerable volumes of stranded gas, and the annual volume of global flaring is still unacceptably high. The first detailed global account of natural gas flaring ended up with 165 Gm3 in 1970, with Iran (about 30 Gm3), Venezuela (about 18 Gm3), and the United States (nearly 14 Gm3, and just ahead of Saudi Arabia and the USSR) wasting most of their gas (Rotty, 1974). Despite of the industry’s rapid post-1970 growth, the annual rate of flaring is now less than four decades ago: in 2010, it was estimated at 134 Gm3, mostly in giant Western Siberian oil fields, in Nigeria, and in Iran, but that is still an equivalent of almost 20% of the US gas use (Roland, 2010; GGFR [Global Gas Flaring Reduction], 2013). Flaring in the United States was reduced by 75% during the 1970s and then fluctuated at a new low level, but it has nearly doubled between 2000 and 2012 due to the rapid expansion of shale gas production (CDIAC, 2014; Figure 3.10).
Figure 3.10 Gas flaring in Pennsylvania.
© Corbis.
And it is not only due to flaring that in almost all cases the volume of the marketed natural gas is substantially lower than the gross extraction (labeled gross withdrawal in the US statistics). In recent years in the United States, the average difference was almost exactly 20%, with (as already noted) venting and flaring wasting less than 1% and about 11% used to repressurize wells; 3% were nonhydrocarbon gases separated from the alkanes, and about 4% were losses during the production (USEIA, 2013a). As already explained, further (relatively small and variable) losses take place during the long-distance transportation of natural gas and its distribution, and gas actually delivered to consumers may be no more than 75% of initial withdrawals even in countries that have virtually eliminated flaring.
Global production of natural gas began to matter only after WWII. The total volume was less than 4 Gm3 in 1890 and only about 7 Gm3 in 1900, nearly all of it in the United States and Russia (UNDESA [United Nation Department of Economic and Social Affairs], 2013). By 1930, the total rose to 56 Gm3, and in 1934, when the World Power Conference published its first Statistical Yearbook, the main producer of natural gas outside North America was Russia (about 1.6 Gm3), and small volumes were extracted in Austria, Poland, Argentina, Brunei, Sarawak, and the Dutch East Indies (WPC, 1934). By 1945, the global output rose to 120 Gm3 (largely due to the 50% wartime increase of US consumption), and then it nearly doubled in just 5 years to 220 Gm3 and doubled by 1960 to 440 Gm3.
Then a more rapid and fairly linear increase brought it to 1.2 Tm3 by 1975 and to just over 2.4 Tm3 by the year 2000 (Smil, 2010a; BP, 2014a). Annual global production thus increased roughly 340-fold during the twentieth century, and I have integrated the best available data to come up with roughly 65 Tm3 of cumulative gas production between 1900 and 2000 (Smil, 2010a). In energy terms, that was about 2.3 ZJ (zeta = 1021 J) compared to almost 5.4 ZJ for coal and 4 ZJ for crude oil, whic
h means that natural gas contributed about 20% of all fossil energies converted by high-energy civilization during the twentieth century, compared to 46% supplied by coal and 34% coming from refined liquid fuels. Cumulative totals for the second half of the twentieth century show crude oil at 39%, coal close behind with 38%, and natural gas at 23%.
During the first dozen years of the twenty-first century, natural gas extraction rose by just over 40% to nearly 3.4 Tm3, and in 2012, gas supplied nearly 28% of all fossil energies and 24% of all primary commercial energy, exclusive of traditional biomass fuels (BP, 2014a; Figure 3.11). This expansion of absolute output has been accompanied by significant diversification of supply. Pre-WWII extraction of natural gas was an overwhelmingly American affair: in 1900, the US production accounted for all but a tiny fraction (<2%) of the global output; by 1950, the share was 75%; and it was still 60% by 1970. Then came the rapid rise of Soviet gas extraction (it had more than quadrupled in two decades between 1970 and 1990), European output (mainly Groningen and the North Sea gas from Norwegian and British waters) doubled during the 1970s before it stabilized, and the Saudi production had tripled during the 1990s. This brought the US share of the global output down to 25% by 1990.
Figure 3.11 Global natural gas extraction.
Afterward, the output in the countries of the former USSR fell and stagnated before it began to recover after the year 2000, but a number of old producer multiplied their extraction: between 1990 and 2010, the absolute increases were 2.2-fold in Australia, nearly four-fold in Malaysia, more than six-fold in China and Iran, nine-fold in Trinidad and Tobago, and 18-fold in Qatar, and by 2010, the US share of the global natural gas production was just below 19%. Another way to illustrate the diversification process is to trace the number of major (annual output >20 Gm3) natural gas producers: only five countries were in that category in 1970 (the United States, Canada, the Netherlands, Romania, and the USSR), eight in 1980 (adding Mexico, the United Kingdom, and Norway), 18 in 1990 (with Russia, Ukraine, Uzbekistan, and Turkmenistan dividing the previous Soviet total), 24 in the year 2000, and 27 in 2012.
In that year, more than 50 nations (i.e., every fourth country worldwide) had nonnegligible natural gas industry (with annual output >1 Gm3), and this number will grow further due to recent discoveries. New offshore discoveries have already put Mozambique’s gas reserves on par with those of Indonesia, while new fields in the Mediterranean (Tamar, discovered in 2009, and Leviathan, discovered in 2010) have made Israel a new gas exporter: in 2014, it signed deals with the Palestinian Authority and with Jordan. Moreover, for about half of the world’s two dozen major hydrocarbon producers, natural gas is now more important (in total energy terms) than crude oil. In 2012, energy content of the two fossil fuels was nearly equal in Russia and Algeria, but gas was more important than oil in Argentina, Australia, Egypt, India, Indonesia, Malaysia, Netherlands, Norway, Qatar, Turkmenistan, and Uzbekistan.
This is an apposite place to note the changes in energy return on investment (EROI). This ratio—energy returned from one unit of energy invested in an energy-producing activity—has been promoted by some of the practitioners of ecological economics, but it has been ignored by most energy economists (Hall, 2011). As a result, we have only a relatively small number of EROI values for natural gas. Guilford, Hall, and Cleveland (2011) estimated EROI for the US oil and gas production during the twentieth century: the ratio shows a slowly declining trend from about 20:1 in 1919 to 8:1 in 1982 (the year of the peak drilling), followed by a recovery to 17:1 between 1986 and 2002, but in the first decade of the twenty-first century, the mean ratio fell rapidly to just 11:1.
And Grandell, Hall, and Höök (2011) estimated that EROI of Norwegian oil and gas production rose from 44:1 in the early 1990s to a maximum of 59:1 in 1996 and then it declined to about 40:1 by 2008. But as these studies combine the returns for two distinct fuels, their EROI ratios can be used only as indicators of a basic (and expected) secular trend of diminishing, but still relatively high, returns. But we have a specific study of Canadian natural gas production during the last decade of the twentieth century and the first decade of the twenty-first century: according to Freise (2011), EROI in Canada’s conventional natural gas has entered the era of permanent decline after falling from 38:1 in 1993 to 15:1 in 2005.
4
Natural Gas as Fuel and Feedstock
Natural gas was not the first gaseous substance with widespread commercial use: its combustion by household, industries, and commercial establishments was preceded for nearly a century by the burning of manufactured gas. This fuel was first generated during the latter half of the eighteenth century as a by-product of coking. The first dedicated facilities producing the gas for lighting (by carbonization of bituminous coal, i.e., by high-temperature heating of the fuel in ovens with limited oxygen supply) began to deliver gas for industries, workshops, and street illuminations in 1812 in London, 1816 in Baltimore, 1818 in Brussels, 1825 in New York, 1829 in Boston, and 1849 in Chicago (Webber, 1918). By the middle of the nineteenth century, every major city in Europe and the United States had its gas works, and in 1855, the introduction of Robert Bunsen’s burner (mixing the gas and air in correct proportion to produce safe flame for cooking and heating) opened the way for mass-scale domestic adoption of town gas, and between 1860 and 1900, the production and distribution of town gas became established as one of the leading industries of new industrial era (Hughes, 1871; Figure 4.1).
Figure 4.1 New York 1900: light and cook with gas.
© Corbis.
Ammoniacal liquor and tar liquid fraction had to be separated from the generated coal gas—actually a mixture of gases dominated by hydrogen (48–52%), with 28–34% of CH4, 13–20% of CO2, and 3% of CO with energy density no higher than 20 MJ/m3 (<60% of typical natural gas)—whose combustion provided a very inefficient source of exterior and interior lighting. But because the earliest incandescent lights were also very inefficient and because mass adoption of electricity required first the extension of transmission lines and interior wiring, coal gas continued to be used in many cities for decades after the first urban power plants were built in the United States and Europe during the 1880s.
In many large cities, the town gas industry was active throughout the first half of the twentieth century, and in some places, gas works operated even into the 1950s and 1960s. Town gas lost first its lighting market to light bulbs (and, starting in the 1930s, to fluorescent lights), but it still competed with electricity for household and institutional cooking. Once natural gas became available (in many US cities before WWII, in Europe widely during the 1970s), town gas era ended, but not its long-lasting environmental legacy of persistent tars that are still contaminating soils and water near former gas works, with the United States having more than 50,000 such sites (Hatheway, 2012).
Because natural gas combines three attributes that are very desirable for a fuel—clean combustion, convenient regulation of the combustion process, and (once the pipelines are in place) on-demand delivery—it has been a favorite (and often superior) form of energy for a variety of uses. Not surprisingly, the extent of these uses has changed with time, and it has been also influenced by the presence or absence of infrastructures (natural gas could be readily distributed in Western cities where houses were connected for decades to receive municipal coal gas), assured availability (as demonstrated by location of many large petrochemical plants close to major gas fields), and affordability (despite technical advances, LNG remains inherently more expensive than pipeline gas).
Traditionally, by far, the most important use of natural gas as fuel was to generate heat or steam required for a large variety of industrial processes, from heavy metallurgy to fine manufacturing. Emergence of this market during the closing decades of the nineteenth century was followed by the adoption of natural gas as a leading fuel for space heating and cooking in households as well as in many commercial and institutional facilities, and since the 1980s, the last two sectors have also u
sed natural gas for central cooling. Another major market for gas as a fuel came with the growth of cleaner electricity generation as methane replaced coal, and also fuel oil, in large centralized power and created new possibilities for more decentralized power based on gas turbines.
A relatively small share of global natural gas extraction that is not burned but serves as a feedstock is not a good indicator of the fuel’s importance as a raw material. Production of the world’s most important fertilizer now depends overwhelmingly on inexpensive natural gas: there are other ways to synthesize ammonia, but none is as economical as using methane both as a feedstock and as a fuel to energize the synthesis. And, in this world suffused by plastic materials, methane and natural gas liquids (ethane, propane, butane, and pentane) are the simplest precursors of so many complex compounds that the material circumstances of modern society would be quite different if those hydrocarbons could not supply the foundation of a large part of the modern petrochemical industry.
In addition, methane is the most important starting material for the synthesis of methanol (methyl alcohol) which, in turn, is an important chemical feedstock used in the production of formaldehyde, acetic acid, and other intermediates that are eventually turned into resins, plastics, paints, adhesives, and silicones. And methanol production is, at least for now, the most important process of transforming gas to liquids, a class of conversions designed to produce more valuable fuels that are also easier to transport. Production of gasoline, kerosene, and diesel oil has been the ultimate goal, and although technically feasible it faces many engineering and economic challenges. Before looking at major consumption sectors, I will summarize the US and foreign pricing in a few paragraphs.