by Vaclav Smil
Clearly, the choice has the greatest appeal in lower-income countries where CNG helps to ease specific challenges: shortage of oil refining capacity in Iran, need to reduce expensive oil imports in Pakistan and Bolivia, and high air pollution in China’s cities. But even there, the switch may not ease the greenhouse gas emissions. Alvarez et al. (2012) concluded that replacing gasoline cars by CNG light-duty cars would bring no reductions of radiative forcing for 80 years and that for heavy-duty diesel vehicles there would be no gain even after more than 100 years. But CNG might offer an alternative to shipping smaller volumes of stranded gas. Smaller carrying capacity of CNG-fueled tankers proposed by Japan’s Kawasaki Kisen Kaisha (equivalent of just 12,000 t of LNG) would be outweighed by less expensive fuel preparation (compression to 20 MPa vs. liquefaction to −162°C) and cheaper vessel construction (“K” Line, 2014).
7.3 NATURAL GAS AND THE ENVIRONMENT
Why should we worry about methane in the environment? The gas has been a part of the biosphere for billions of years; it is produced by a variety of natural processes and then is effectively removed from the atmosphere by oxidation. Moreover, as stressed in the opening chapter, methane generates less CO2 per unit of useful energy than does the burning of any kind of wood, coal, crude oil, or refined oil products, and hence, it ranks as the best carbon energy source in the age concerned about global warming and rising carbon emissions. But this great advantage comes with an undesirable property: methane is the only fossil fuel that is also a greenhouse gas and, indeed, a more potent one than CO2. Every molecule of CH4 absorbs more outgoing long-wave radiation than does a molecule of CO2, and approximately one–fifth of the increase in anthropogenic radiative forcing during the past 250 years has been due to rising atmospheric methane concentrations.
7.3.1 Methane Emissions from Gas Industry
By the end of the twentieth century, the global atmospheric burden of the gas had more than doubled compared to preindustrial era (to about 1,730 ppb); then it remained nearly constant for nearly a decade before a strong growth resumed in 2007 (Dlugokencky et al., 1998; Kai et al., 2011; Nisbet, Dlugokencky, and Bousquet, 2014). Assessing the role of natural gas in this increasing burden requires a stepwise approach: we must first establish the total of CH4 emissions associated with the natural gas industry, compare it first to other anthropogenic source of methane and then to all biospheric fluxes of the gas, and, finally, quantify its contribution to the overall global warming potential (GWP). Only after going through this sequence we could determine the degree of concern attributable to methane emissions associated with gas industry. Obviously, it would be counterproductive if the industry producing the fossil fuel with the lowest specific CO2 emissions would negate much, even most, of that desirable effect due to large volumes of methane that it let escape into the atmosphere.
If all natural gas would be recovered and burned, then its contribution to global warming would come only indirectly, due to the oxidation that transforms the simplest alkane into CO2 and water. But it is inevitable that some methane escapes to the atmosphere before it can be burned, during the drilling of hydrocarbon wells, during their often decades-long operation, from the gathering pipelines in gas fields, and during the processing and long-distance transportation and local distribution to industrial users and households. Moreover, unwanted gas produced without any market access has been always deliberately flared, and this wasteful and environmentally damaging practice still continues at an unacceptable scale. And, finally, coal mining, thanks to China recently the most rapidly increasing kind of fossil fuel production, is also a source of uncontrolled releases of methane.
Because CO2 remains by far the most important anthropogenic greenhouse gas, it has been a standard practice to express the warming impact of other gases in terms of CO2 equivalent (CO2e). This involves the concept of the GWP, a measure that expresses equivalent effects of various gases over a 100-year period. This standardization could be done for shorter (20 years) or longer (500 years) periods, and it is necessary due to different lifetimes of greenhouse gases in the atmosphere: CO2 molecule may remain aloft for up to 200 years, methane’s average is just 12 years, while nitrous oxide (N2O) persists for more than a century. According to the IPCC’s Second Assessment Report, the GWP for the simplest chlorofluorocarbon (CCl3F) is 3,800 CO2e; for N2O, it is 310; and for methane, it is 21 (IPCC, 2007; 1 Mt CH4 = 21 Mt CO2e). For comparison, methane’s potential for a 20-year period would be about 72, while that for 500-year span would be less than 8. I will quote published estimates of CH4 emissions in Mt/year, and for comparisons with other greenhouse gases, I will use the standard mass conversion of (1 CH4 = 21 CO2e).
Natural gas industry generates CH4 in several ways: as the gas escaping during the fuel’s extraction; due to deliberate venting, while flaring (controlled combustion of escaping gas) produces CO2 and is only a minor CH4 source; and as losses during gas processing and pipeline transportation. CH4 emissions from stationary combustion sources are highest for residential wood combustion (300 g/GJ) and (at just 1 g/GJ) negligible for either gas- or coal-based electricity generation (USEPA [United States Environmental Protection Agency], 2008). Our best understanding of production emissions comes from a large number of direct measurements at the US onshore sites (Allen et al., 2013). They included gas from well completion flowbacks (gas dissolved or entrained in liquids that must be removed from wells before the beginning of extraction), unloadings (lifting of accumulated liquids that restrict gas flow from producing wells), and routine operation of well sites (including wellheads, separators, controllers, and tanks).
Assuming that those measurements were representative of nationwide operations, the annual CH4 emissions from these operations would be 957,000 (±200,000) t, while the estimate for comparable emissions categories in the national inventory was 1.2 Mt (USEPA, 2014). This is a fairly close agreement for inherently uncertain estimates of this kind. After adding USEPA’s estimates for other source (not measured directly in their study), Allen et al. (2013) ended up with the total of 2.3 Mt or 0.42% of the nation’s gross natural production, compared to the USEPA’s 2011 total of about 2.55 Mt or 0.47% of the gross output. Processing losses are usually no higher than 0.2%, and leakage and venting during transmission and distribution range mostly between 0.3 and 1.0% of gross production.
Not surprisingly, natural gas losses from the world’s longest pipelines connecting Western Siberia with Central Europe have received particular attention. According to Reshetnikov, Paramonova, and Shashkov (2000), during the early 1990s, their losses (established by inlet–outlet difference) amounted to 47–67 Gm3 or 6–9% of total extraction. That would have been an extraordinary waste as the volume of 50–60 Gm3 was, at that time, larger than Italy’s total gas imports. In contrast, several Russian, American, and German measurements of the mid-1990s indicated losses of only about 1% of the gas produced, but they were questioned due to a small number of surveyed sites. The uncertainty was resolved by a comprehensive German–Russian measurement at compressor stations and along export pipelines in 2003. Their results indicated annual losses of 3.4 Gm3 or an equivalent of 0.6% of produced gas or 0.7% when underground storage is included (Lechtenböhmer et al., 2007). The authors put the 95% confidence interval at 0.5–1.5% or about 7% of the total volume of energy used by Gazprom for pipeline operations.
The last source of emission from traditional oil and gas industry is methane flaring (see also Figure 2.2). Flaring that takes place in refineries and gas plants, as well as during well tests, is a minor source of emissions compared to the flaring associated (solution) gas produced along with crude oil or bitumen. Those emissions range between 1 and 4 kg CH4/t of burned gas, and satellite observations are used to estimate annually flared volumes on national and global scales. The global level between 2009 and 2012 was around 140 Gm3 or about 95 Mt/year (GGFRP, 2014). This means that flaring would be emitting at least 95,000 and as much as 380,000 t CH4 a year, a negligible c
ontribution that is far smaller than the error inherent in estimating emissions from all activities associated with natural gas production. The most realistic range of losses associated with properly operated natural gas production and transportation is thus between 1 and 2.5%, and with the 2010 global production of 3.2 Tm3, that would prorate to roughly 32–80 Gm3 or 22–54 Mt of CH4.
Historic estimates of these emissions show that the industry remained a negligible source of CH4 until after WW I, with annual rates increasing from less than 0.5 Mt CH4 in 1900 to less than 8 Mt in 1950 and to about 33 Mt by 1990 (Stern and Kaufmann, 2001). Höglund-Isaksson (2012) put the global CH4 emission from gas production and transportation in the year 2005 at about 29 Mt (with about 60% due to leaks during long-distance transmission and distribution), and calculations by the Global Methane Initiative (2014) have the 2010 emissions from the global oil and gas industry at roughly 65 Mt. But natural gas industry is just one-half a dozen anthropogenic sources of CH4, and its emissions must be compared to those emanating from other activities, above all from agriculture.
Human intervention in the methane cycle began tens of thousands of years ago with deliberate burning of forests and grasslands; later (sometime between 8,200 and 13,500 years ago) came domestication of rice (Molina et al., 2011) that led to expanded anaerobic fermentation in water-covered soils and domestication of ruminants, starting with sheep in the Southwest Asia about 11,000 years ago, followed by goats and cattle at about 8,600 BCE (Harris, 1996; Chessa et al., 2009). Methane from coal mining began to add significant amounts of CH4 during the eighteenth century, from urban landfills and municipal waste water during the nineteenth century.
Stern and Kaufmann (2001) prepared the most comprehensive inventory of historical anthropogenic emissions of CH4 between1860 and 1994. Their estimates for 1900 show the total of about 114 Mt, with nearly half from wet fields (mainly rice cultivation) and almost a third from livestock and roughly 10% each from the burning of phytomass and coal mining. By 1950, the total was up to nearly 180 Mt, with more than a third from wet fields, nearly 30% from livestock, 12% from coal mining, and still only less than 5% from natural gas production and flaring. The series ends in 1994 with the total of 371 Mt CH4, with rice at 30%, livestock at 27%, and natural gas emissions at 9%.
Other studies of anthropogenic CH4 emissions put the global total at the end of the twentieth century at between 300 and 350 Mt, and the USEPA’s values (expressing the total in terms of CO2e with GWP = 21) were just over 300 Mt for the year 2000 and roughly 343 Mt for 2010 (USEPA, 2012). Höglund-Isaksson (2012) calculated 323.4 Mt in 2005, and the Global Methane Initiative (2014) put the 2010 emissions at about 327 Mt, with agriculture (mainly enteric fermentation of livestock and rice cultivation) contributing 50%, oil and gas industry 20% or about 65 Mt CH4, and coal mining about 20 Mt or 6%. Recent estimates thus show a fairly good agreement given the multiple uncertainties inherent in quantifications of this kind and allow the following conclusions: in 2010, anthropogenic emissions of CH4 were most likely no higher than 350 Mt, with natural gas responsible for no more than 70 Mt.
In turn, anthropogenic emissions of CH4 are added to a few relatively large natural source that have been emitting the gas for hundreds of millions to billions of years. Beerling et al. (2009) published an interesting modeling exercise of tropospheric CH4 levels during the past 400 million years. They scaled a wetland emission estimate for the middle Pliocene (3.6–2.6 million years ago) by the relative rate of coal basin deposition. This indicated extremely high peaks of about 12,000 parts per billion (ppb) during the Permo-Carboniferous era when tropical swamplands reached their largest extent and lows of just 100 ppb during the Triassic (coal-gap age). Prehuman biogeochemical cycle of CH4 was dominated by emissions from natural wetlands (about 100 Mt CH4/year), with termites and emissions from anoxic marine sediments each an order of magnitude lower.
The most comprehensive review of global methane sources and sinks during the recent decades (Kirschke et al., 2013) ended up with the following totals: natural emissions (mostly from wetlands) of 218–347 Mt CH4/year (for, respectively, top-to-bottom and bottom-up accounting approach), anthropogenic emissions of 335–331 Mt, and 96 Mt from all fossil fuels (Figure 7.6). This means that human activities now account roughly for between 50 and 60% of all methane additions to the atmosphere, fossil fuels account for 14–18% of the total influx, and natural gas production and transportation for no more 10–13%.
Figure 7.6 Global methane emissions.
But it should be noted that the range for global wetland emissions of methane (175–217 Mt/year) is as large as it was in 1974 when the world’s first global methane budget was published, largely because annual natural methane fluxes can vary by a factor of two or more (Christensen, 2014). And a recent discovery of hundreds of ocean floor seeps leaking methane off the East Coast of the United States is yet another illustration of our inadequate understanding of natural CH4 emissions. Skarke et al. (2014) identified about 570 gas plumes at depths of 50–1,700 m between Cape Hatteras and Georges Bank, a finding that suggests existence of as many as 30,000 of such vents worldwide—and their emissions have not been properly accounted for in any global CH4 inventories.
Keeping that in mind, the final step in assessing the contributions of natural gas-related CH4 emissions to global warming is to assign the share of radiative forcing attributable to the gas. In 2011, forcing’s net value (after taking into account cooling effect of sulfates, nitrates, and organic carbon) was 2.83 W/m2 (IPCC, 2013), with CO2 contributing 1.82 (64%), CH4 being a distant second with 0.48 W/m2 (17%), and N2O in the third place with 0.17 W/m2 (6%). Assigning between 10 and 15% of methane’s radiative forcing to emission from natural gas production and transportation makes the industry responsible for just between 0.05 and 0.07 W/m2 or merely on the order of 2% of aggregate radiative forcing in 2011.
To complete the account of radiative forcing contributions attributable to natural gas, we must add CO2 emissions from the fuel’s combustion: they will, of course, make a much larger contribution than those attributable to methane. Until 1845, they accounted for less than 5% of the fossil-fuel total and hence were smaller than the inevitable calculation error (due mainly to changing carbon content of coal). By 1975, combustion of natural gas produced 13% of all CO2 emissions from fossil-fuel combustion; by the year 2000, the share rose to 19% and by 2010 declined marginally, mainly due to rising carbon emissions from China’s increasing coal combustion.
But because CO2 has a relatively long atmospheric lifespan, we should consider the cumulative contribution as well. Between 1750 and 2010, combustion of fossil fuels, gas flaring, and cement production emitted the grand total of about 350 Gt C into the atmosphere (CDIAC, 2014) of which about 50 Gt C came from the burning of natural gas, the share roughly rounded to 15%, and the same share would be the most appropriate multiplier to estimate the carbon burden attributable to CO2 emissions from natural gas. Multiplying carbon dioxide’s rate of 1.82 W/m2 by 0.15 yields 0.27 W/m2, and combined with the methane’s forcing share, it produces the combined contribution of about 0.3–0.35 W/m2 or about 10–12% of the total anthropogenic radiative forcing, clearly an excellent performance for the fuel that now provides nearly 25% of the world’s TPES!
At the same, it may be too optimistic to expect that increased use of natural gas will bring substantial decarbonization of the global energy use. Indeed, simulations by McJeon et al. (2014), based on five state-of-the-art integrated assessment models of energy–economy–climate systems and independently forced by an abundant natural gas scenario, show large increases in the combustion of gas (up to 170% more by 2050) with the resulting impact on overall CO2 emissions ranging from a modest decline of 2% to an increase of 11% and with a majority of the models indicating a small gain (range of −0.3 to 7%) in climate forcing associated with the increased reliance on natural gas.
7.3.2 Methane from Shale Gas
New concerns about atmospheric impacts of natural gas (and oil) production have been created due to America’s rapid expansion of hydraulic fracturing: its opponents realized that high methane losses during hydraulic fracturing, well completion, and transportation could reduce, or even negate, the benefits of gas as a fuel with much lower GWP than coal. I will offer a brief, and largely chronological, summaries of some recent claims and counterclaims about CH4 emissions from shale gas extraction and transportation in the United States: these findings show how difficult it is to come up with any confident generalizations of fractional methane losses (be they from fracking operations or from conventional gas extraction) and how wide the range of uncertainty remains.
A study done at the National Energy Technology Laboratory used a detailed life-cycle model for natural gas that included about 30 items of processes encompassing extraction, processing, transport, and conversion, and it concluded that average natural gas (a mixture of conventional onshore, offshore, associated, tight, shale, and coal bed methane) base-load electricity generation has life-cycle greenhouse gas emissions 53% lower than average coal base-load power generation (Skone, 2011). Four other life-cycle analyses found similar, or even higher, benefit when comparing gas to coal (Burnham et al., 2011a; Hultman et al., 2011; Jiang et al., 2011; Stephenson, Valle, and Riera-Palou, 2011).
In contrast, Howarth, Santoro, and Ingraffea’s (2011) study was the first that found the very opposite: they concluded (on the basis of limited measurements) that venting and leaks from shale gas production will produce, over the lifetime of a well, methane emissions at least 30% higher and perhaps more than twice as large as those from the production of conventional gas. They estimated life-cycle CH4 emissions of a well to be 3.6–7.9% of produced gas for shale compared to 1.7–6% for conventional gas. Moreover, they also calculated that on the 20-year horizon, emissions from shale gas are also at least 20% higher than those from coal production and possibly more than twice as high as for coal when expressed per unit of energy available during combustion and that there is little or no advantage over coal even on the 100-year horizon. According to the study, offshore gas had the lowest GWP (but still about 25% higher than coal), and Barnett shale the highest (about 2.5 times as much as coal).
