In the Earth System, carbon is stored in rocks (as carbonates), sediments, ocean and freshwaters, soils and terrestrial biomass, and the atmosphere. By far the larger reservoir of carbon is the deep water of the ocean, which is thought to contain about 80% of the Earth System’s carbon (excluding rock; see Figure 1.2). Oceanic sediments are thought to contain 4%. Ocean surface waters and the atmosphere each hold about 2% of the Earth system’s carbon reservoirs. Oil, gas, and coal reserves are thought to contribute another 3%. Soils and permafrost hold 5% and 4% of global carbon, respectively, while carbon stored in vegetation adds about 1%. The global carbon cycle includes the mechanical, chemical, and biological processes that transfer carbon among these reservoirs (see Figure 1.2). Reservoirs of carbon in the Earth system often are also referred to as “pools” or “stocks,” and transfers of carbon between reservoirs are known as “fluxes.” Some of these carbon fluxes are sensitive to climate, and their resulting responses to climate change are known as “carbon cycle–climate feedbacks.” A positive feedback can occur when carbon fluxes to the atmosphere increase as a result of, for example, increasing temperatures. More carbon in the atmosphere leads to further climate warming, possibly further increasing carbon fluxes to the atmosphere. Carbon cycle–climate feedbacks will be discussed further in Section 1.4.

1.2.1 Carbon Dioxide

The global carbon cycle comprises a fast carbon cycle, having relatively rapid exchanges among the ocean, terrestrial biosphere, and atmosphere, and a slow carbon cycle, involving exchanges with geological reservoirs such as deep soils, the deeper ocean, and rocks. Equilibration between the terrestrial biosphere and ocean occurs on millennial timescales, while redistribution of CO₂ among geological reservoirs requires tens to hundreds of thousands of years or longer. Figure 1.2 provides a pictorial representation of the exchanges of carbon among the main reservoirs, together with associated timescales.

Reservoirs for the fast components of the carbon cycle include the ocean, land vegetation and soils, freshwaters, shallow oceanic sediments, and the atmosphere. Based on estimates from the Intergovernmental Panel on Climate Change Fifth Assessment Report (IPCC AR5; IPCC 2013), about 830 petagrams of carbon (Pg C; 2000 to 2009 average) were present in the atmosphere, while 450 to 650 Pg C are stored in the terrestrial biosphere. Larger reservoirs of carbon exist in soils (1,500 to 2,400 Pg C; IPCC 2013), and soil organic carbon (SOC) densities are highest in moist boreal and tropical latitudes. Scharlemann et al. (2014) pointed out that these numbers are uncertain due to limited depth and sparse distribution of sampled or observed SOC profiles. The Arctic permafrost soils are estimated to contain 1,339 to 1,580 Pg C in the top 3 m of the soil column, with another 400 Pg C possible in deep soils (Schuur et al., 2015). Ocean waters and shallow sediments contain about 40,500 Pg C. The “fast-exchange” reservoirs of the ocean surface and marine biota hold only 900 Pg C and 3 Pg C, respectively. Turnover times for these fast- and ­slow-exchange reservoirs range from decades to millennia.

Exchange of carbon between the atmosphere and the terrestrial biosphere occurs via photosynthesis and respiration. Carbon is removed from the atmosphere by photosynthesis and fixed in leaves, roots, stems, and woody biomass. It is returned to the atmosphere through autotrophic (plant) respiration and heterotrophic (microbial) respiration of plant litter and soil carbon. Fire and other disturbances such as insect outbreaks and timber harvesting can be thought of as accelerated respiration processes, and the amount entering the atmosphere from these processes varies from year to year. Removal of CO₂ by photosynthesis is thought to have been slightly higher in the preindustrial atmosphere than emissions added from respiration and natural disturbances. Global total photosynthesis at that time is thought to have exceeded global respiration and emissions from natural disturbances so that net removal from the atmosphere by the land was about 1.7 Pg C per year. This removal is estimated to have been approximately in balance with outgassing from the ocean and freshwaters (Ciais et al., 2013; see Figure 1.2).

Gas exchange between the atmosphere and ocean depends on the difference between the partial pressure of CO₂ in surface water and that of CO₂ in the atmosphere (∆pCO₂). Carbon dioxide dissolves in ocean water to form carbonic acid (H₂CO₃), which then forms bicarbonate (HCO₃^–) and carbonate (CO₃^2–). These coupled reactions chemically buffer ocean water, thus regulating ocean pCO₂ and pH. Because pCO₂ can vary spatially, carbon outgasses from the ocean waters in some regions and is taken up in others. In regions where there is upwelling of nutrient-rich water and ocean waters are warm (e.g., in parts of the tropics), carbon is outgassed. In the North Atlantic, cold, sinking water removes carbon from the atmosphere. The Southern Ocean (latitudes south of 44°S) is another area where carbon is taken up. Carbon also is exchanged between land and ocean reservoirs via river transport to the coastal ocean.

Year-to-year variability of the global ocean CO₂ sink was thought to be small, at only about ±0.2 Pg C per year or 9% of annual ocean uptake (Wanninkhof et al., 2013); however, recent work by Landschutzer et al. (2016), based on comprehensive measurements of global ∆pCO₂ of ocean surface water, suggests that substantial decadal and interannual variability can exist. They found that during the 1990s, the global ocean sink was likely to have been significantly smaller than after year 2000 (–0.8 ± 0.5 Pg C per year and –2.0 ± 0.5 Pg C per year, respectively). They proposed 1) that these decadal variations are driven by extratropics and are linked with the atmospheric northern and southern annular modes and 2) that interannual variability is driven by the tropical ocean. The variability of the global land sink is larger, varying by 3 to 4 Pg C per year, and most of this variability likely occurs in the tropics (Baker et al., 2006). This global atmospheric CO₂ interannual variability arises primarily from land sink variability because of the strong anticorrelation between CO₂ and δ¹³C (e.g., Alden et al., 2010). Terrestrial net carbon exchange gives rise to significant δ¹³C variability, whereas air-sea gas exchange does not. The El Niño Southern Oscillation (ENSO) is thought to be a significant driver of tropical carbon flux variability for both the ocean and terrestrial ecosystems. During the warm phase of ENSO, the ocean takes up more carbon because of reduced upwelling and outgassing from the eastern Tropical Pacific. On land, ENSO is associated with outgassing from the terrestrial biosphere, a phenomenon likely associated with drought and warmer global temperatures. Indeed, the strong ENSO of 2016 pushed measured CO₂ concentrations at Mauna Loa to above 400 ppm, where they have remained (Betts et al., 2016).

The slow, or geological, carbon cycle operates on timescales of tens of millennia and longer. Fluxes to the atmosphere from volcanism, CO₂ removal from the atmosphere by chemical weathering, and ocean sediment formation together are a factor of 10 smaller than the fluxes of the fast carbon cycle. A vast amount of carbon is also stored in sedimentary rocks (100 × 10⁶ Pg C), with an estimated 4,000 Pg C stored as hydrocarbons (Ciais et al., 2013).

Ice core evidence suggests that during glacial periods atmospheric CO₂ was present at about 180 to 200 ppm. During interglacial periods, atmospheric CO₂ abundance was higher, between 270 to 290 ppm (Lüthi et al., 2008; Petit et al., 1999). The current atmospheric levels of 400 ppm are well outside the range that existed during the period resolved by ice cores; that is, 800,000 years before present. The most recent glacial period ended about 12,000 years ago, with the most recent glacial maximum occurring about 22,000 years ago. Even older evidence from Arctic lake sediments suggests that around 3.5 million years ago, Arctic summer temperatures were about 8°C warmer than today with atmospheric CO₂ levels around 400 ppm (Brigham-Grette et al., 2013). Contemporary CO₂ has surpassed 400 ppm, suggesting that the current Arctic is not yet in equilibrium with rapidly rising greenhouse gas concentrations and may become much warmer in the future.

Estimates for recent decades show significant trends and variability in the main components of the global carbon cycle (see Table 1.1). Only about half of human-driven emissions from fossil fuel burning, industry (e.g., cement manufacturing), and land-use change remains in the atmosphere, although the growth in atmospheric CO₂ is highly variable depending on emissions and the strength of uptake by land and ocean (see Table 1.1). Emissions have risen by about 70% from the 1980s to the most recent decade (2007 to 2016), while land and ocean have taken up 3.0 ± 0.8 and 2.4 ± 0.5 Pg C per year, respectively (Le Quéré et al., 2017). Of this amount, North America represents a rather substantial share of global carbon uptake (0.31 Pg C per year; see Ch. 2: The North American Carbon Budget). Figure 1.3a shows global average atmospheric CO₂ derived from in situ surface air samples. The steep rise in CO₂ reflects anthropogenic emissions, while the annual cycle reflects the seasonal uptake of vegetation, predominantly in the Northern Hemisphere.

1.2.2 Methane

Total global CH₄ emissions are approximately 550 teragrams (Tg) of CH₄ per year (1 Tg CH₄ per year = 10¹² grams of CH₄ per year; Saunois et al., 2016). Of this, roughly 40% comes from natural sources. The largest (and most uncertain) natural emissions of CH₄ are from wetlands, defined as regions that are permanently or seasonally waterlogged. Natural wetlands include high-latitude bogs and fens, tropical swamps, and temperate wetlands. Saturated soils in warm tropical environments tend to produce the most CH₄. However, warming Arctic temperatures raise concerns of increasing emissions from high-latitude wetlands and future decomposition of carbon currently stored in frozen Arctic soils (e.g., Schaefer et al., 2011; Schuur et al., 2015). Figure 1.4 provides a pictorial representation of the main components of the global methane cycle.

Estimates of global CH₄ emissions from wetlands range from 127 to 227 Tg CH₄ per year (Saunois et al., 2016), with most probable values between 167 and 185 Tg CH₄ per year. Most emissions occur in tropical regions (Matthews 1989; Melton et al., 2013; Saunois et al., 2016). Currently, only about 25 Tg CH₄ per year (i.e., 4% of global emissions) are thought to be emitted from high northern latitudes (AMAP 2015; Saunois et al., 2016). Because emissions are sensitive to temperature and precipitation, they exhibit significant seasonal cycles, especially at high latitudes, as well as interannual variability caused by moisture and temperature variability. Smaller amounts of CH₄ are emitted from fires, the ocean, and enteric fermentation in termites and wild animals (20 Tg CH₄ per year or less for each). In addition, up to 60 Tg CH₄ per year may be emitted from geological sources, such as seeps, clathrates, mud volcanoes, and geothermal systems (Etiope et al., 2008; Schwietzke et al., 2016).

Unlike CO₂, CH₄ has an atmospheric chemical sink that nearly balances total global emissions. Removal of atmospheric CH₄ by reaction with the hydroxyl radical (OH) results in a CH₄ atmospheric lifetime of about 9 to 10 years. Observationally constrained estimates of CH₄ lifetime suggest either small decreases of about 2% from 1980 to 2005 (Holmes et al., 2013) or stable CH₄ lifetimes with the possibility of interannual variability of about 2% (Montzka et al., 2011). CH₄ is a much more powerful greenhouse gas than CO₂ (on a per mass basis and over 100 years, CH₄ is about 25 times more effective at trapping heat than CO₂).

As shown in Figure 1.3b, atmospheric CH₄ increased rapidly during the 1980s and early 1990s before its growth leveled off between the mid-1990s and early 2000s. Methane has resumed its increase in the atmosphere since 2006, and observations show that this growth has even accelerated since 2014. The changing atmospheric CH₄ growth rate has been the subject of much debate, questioning why growth rate slowed for a decade starting in the mid-1990s. Several studies suggested that this slower rate was due to decreases in fugitive emissions from fossil fuel production (Aydin et al., 2011; Simpson et al., 2012) or to decreased emissions from anthropogenic microbial sources, such as rice agriculture (Kai et al., 2011). On the other hand, Dlugokencky et al. (1998, 2003) proposed that the slowing of CH₄ growth in the atmosphere was due to an approach to a quasi–steady state, reached when global sources and sinks are in balance. Consistent with this view, the study of Schwietzke et al. (2016) found that emissions from oil and gas production have remained stable over the past several decades, implying increasing efficiency in fossil fuel production industries while their production was increasing over time.

Dlugokencky et al. (2003) predicted that CH₄ would approach a steady state in the atmosphere of about 1,780 ppb by the 2010s if there were no major changes in its budget. The methane budget did change, however, because the atmospheric growth of CH₄ resumed its rise in 2006. The cause of the recent increase in CH₄ growth also has been much debated. Based on global observations of the CH₄ isotope, ¹³CH₄, the global growth in CH₄ appears likely to have been dominated by microbial sources in the tropics (wetlands or agriculture and waste), rather than fossil fuel production (Nisbet et al., 2016; Schaefer et al., 2016), as suggested by some studies (e.g., Rice et al., 2016). Other studies have argued that ¹³CH₄ may not be a very strong constraint on the global methane budget and that changes in the atmospheric CH₄ chemical sink are responsible for the global methane changes (Rigby et al., 2017; Turner et al., 2017). However, plausible chemical mechanisms that could explain the changes in the CH₄ sink have not been identified. Using space-based retrievals of carbon monoxide, Worden et al. (2017) argued that the isotopic data record also can be consistent with increased fossil fuel emissions if global biomass-burning emissions have decreased twice as much as estimates based on space-based observations of burned areas. If the recent rise of global atmospheric CH₄ is indeed due to increases in microbial emissions, then the question becomes whether anthropogenic or natural microbial sources are responsible. Some studies have suggested that anthropogenic microbial sources, such as livestock, are behind the increased atmospheric growth of CH₄ (Schaefer et al., 2016; Saunois et al., 2016). If the increase is due to emissions from wetlands, especially in the tropics, then this raises the possibility that changing climate could be changing natural emissions.

The carbon cycle undergoes perturbations caused by a variety of natural processes such as wildfires, droughts, insect infestations, and disease. These processes can themselves be affected by human activities, for example through GHG emissions that change climate, wildfire suppression, and land-use change. During longer periods, variations in the Earth’s orbit also drive significant perturbations to the global carbon cycle. Over the recent several centuries, human activity has resulted in perturbations to the carbon cycle that have no precedent in geological records. Anthropogenic emissions also can directly alter the chemistry of the atmosphere, possibly affecting its ability to remove pollutants. These human-caused carbon cycle perturbations are discussed in this section.

Since the dawn of the Industrial Age over 250 years ago, humans have significantly altered the global carbon cycle, chiefly by combustion of fossil fuels, but also by perturbing the natural carbon cycle. An example is the large-scale conversion of forests to agricultural land and rangeland. As a result, atmospheric concentrations of CO₂ and CH₄ have increased dramatically. Atmospheric CO₂ has increased from a preindustrial abundance of 280 ppm of dry air (MacFarling Meure et al., 2006) to more than 400 ppm in recent years (NOAA-­ESRL-GMD Trends 2017),² an increase of 43%. Methane has increased from a preindustrial abundance of about 700 ppb of dry air to current values of over 1,850 ppb, an increase of over 160%. Current understanding of the sources and sinks of atmospheric carbon supports the dominant role played by human activities, especially fossil fuel combustion, in the rapid rise of atmospheric carbon. For example, Tans (2009) demonstrated that accumulated carbon in the atmospheric and oceanic reservoirs since preindustrial times is approximately equivalent to the total amount emitted by fossil fuel combustion. If fossil fuel emissions were abruptly terminated, 20% to 40% of this carbon would remain airborne for millennia (Archer et al., 2009; Archer and Brovkin 2008; Solomon et al., 2009). Increases in atmospheric carbon, along with smaller contributions from other GHGs emitted by humans, have led to annual global mean temperatures that have risen by 0.85°C during 1880 to 2012 (IPCC 2013). If recent years are included, the global average temperature has increased by about 1.25°C since 1880 (Hansen et al., 2017).

1.3.1 Anthropogenic Emissions

By burning coal, oil, and gas, humans are accelerating the part of the geological carbon cycle that transfers carbon in rocks and sediments to the atmosphere. From 1870 to 2017, humans emitted 430 ± 20 Pg C as CO₂ to the atmosphere (Le Quéré et al., 2018). Global fossil fuel emissions of CO₂ increased at a rate of about 4% per year from 2000 to 2012, when emissions growth decreased to about 1% per year. In subsequent years, the growth of CO₂ emissions continued to decline, leveling off in 2015 (see Figure 1.4; Le Quéré et al., 2018), when global carbon emissions from fossil fuel use and cement production—an industry which releases CO₂ as a by-product of the chemical process that produces lime from limestone—was estimated to total 9.9 Pg C (about 100 times faster than natural geological fluxes; see Figure 1.2). This leveling off of emissions occurred even as the global economy was expanding (see Figure 1.5). In 2017, global CO₂ emissions rose again by an estimated 2%, likely due to faster economic growth and lower fossil fuel prices (Le Quéré et al., 2018).

Humans also can affect the global carbon cycle through land-use change, mainly by conversion of forests to agricultural land. Often deforestation is accomplished through use of fire. Emitted during the land-use conversion process from forest to other uses, CO₂ thereafter reduces carbon uptake. Reforestation of formerly agricultural land can cause increased carbon uptake over time. Cumulative emissions of carbon from land-use change (mainly clearing of land for agriculture) since 1750 are estimated at 225 ± 75 Pg C (Le Quéré et al., 2018).

Atmospheric CH₄ also is influenced by diverse human activities, ranging from food production (e.g., ruminants and rice) to waste (e.g., sewage and landfills) to fossil fuel production (e.g., coal, oil, and gas). Future increases in population likely will increase CH₄ emissions from agriculture and waste as demand rises for more food production. Furthermore, the current boom in shale oil and gas exploitation has focused attention on leakage from drilling, storage, and transport of fossil fuel (e.g., Peischl et al., 2015; Pétron et al., 2014). Chemical reaction with OH accounts for about 90% of the total CH₄ sink (Ehhalt 1974). These OH radicals, produced through the photolysis of ozone (O₃) in the presence of water vapor, are destroyed by reactions with CH₄ and other compounds. Uncertainty in the sink due to chemical loss by OH is 10% to 20%, because the OH distribution remains uncertain at regional to global scales (Saunois et al., 2016).

Relative to CO₂, CH₄ and other short-lived climate forcers such as black carbon have short atmospheric lifetimes; thus, estimates project that their mitigation potentially could reduce global mean warming by about 0.5°C by 2050, with air quality and agricultural productivity as co-benefits. Such mitigation, however, would not significantly limit maximum warming beyond 2050 (Shindell et al., 2012; Rogelj et al., 2014; National Academies of Sciences, Engineering, and Medicine 2018). Various strategies are possible for reducing emissions or enhancing the CH₄ sink. For example, some increases in agricultural and waste emissions possibly could be avoided through improved practices and changed dietary trends (Hall et al., 2009; see Ch. 5: Agriculture for more information on agricultural and food emissions). In addition, humans potentially can alter the chemical lifetime of CH₄ through emissions that affect the abundance of OH. Naik et al. (2013) found that OH might be about 10% lower than in preindustrial times, although with large uncertainty.

Current estimates reported by Saunois et al. (2016) for anthropogenic emissions average 328 Tg CH₄ per year (ranging from 259 to 370 Tg CH₄ per year). Extraction and processing of fossil fuels account for 32% to 34% of all anthropogenic emissions. Livestock, agriculture, landfills, and sewage together account for another 55% to 57%, with the remainder due to biomass and biofuel burning. A recent study using observations of the isotopic composition of CH₄ suggests that emissions from fossil fuel production and geological emissions may be 20% to 60% higher than previously thought. This increase would require a compensating reduction in microbial emissions from natural and anthropogenic sources (Schwietzke et al., 2016) for the atmosphere to be in balance with the observed global average CH₄ abundance.

Current CH₄ levels are unprecedented in over at least 800,000 years (Loulergue et al., 2008). Recent National Oceanic and Atmospheric Administration atmospheric network observations have shown that global CH₄ increased rapidly through the late 1990s, leveled off during the early 2000s, and began to increase again in 2007 (Dlugokencky et al., 2009; Rigby et al., 2008). These changes in global CH₄ are not well understood and are under debate. Although Dlugokencky et al. (1998, 2003) suggested that the plateau in CH₄ growth resulted from an approximate balance between global sources and sinks, some studies suggested that decreases in anthropogenic emissions (Aydin et al., 2011; Kai et al., 2011; Simpson et al., 2012) led to the period of slow CH₄ growth. Isotopic evidence points toward increased emissions from microbial sources as an explanation for the recent rise in global CH₄ (Nisbet et al., 2016; Schaefer et al., 2016; Schwietzke et al., 2016). However, increases in anthropogenic emissions also have been proposed (Rice et al., 2016), as well as decreases in the chemical loss (Rigby et al., 2017; Turner et al., 2017). Worden et al. (2017) have recently suggested a significant role for fossil fuel emissions in the recent growth of atmospheric CH₄ based on decreases in biomass burning that could change the interpretation of methane isotope observations. This result is based on space-based observations of atmospheric CO, which itself may be responding to changes in other sources besides biomass burning.

Figure 1.1 shows that CH₄ contributed just over 0.5 W/m² in 2017 to global total anthropogenic radiative forcing, an amount which is about one-fourth of that from CO₂. Although CH₄ is much more effective at absorbing infrared radiation (Hofmann et al., 2006; Myhre et al., 2013),³ it is about a hundred times less abundant in the atmosphere than CO₂.

1.3.2 North American Emissions in a Global Context

Historically, North America has been one of the world’s largest producers of human-caused CO₂ emissions. Between 1850 and 2011, the United States has added 27% of the cumulative emissions, compared with 25% from European Union (EU) countries and 11% from China, currently the world’s largest emitter (World Resources Institute et al., 2014).⁴ In 2015, North America emitted almost 15% (1.5 Pg C) of the 9.9 Pg C emitted globally (Olivier et al., 2016). Of North America’s annual total emissions, a majority (84%) came from the United States, while Canada and Mexico emitted 8.7% and 7.3%, respectively. Since the 2007 publication of the First State of the Carbon Cycle Report (SOCCR1), China has replaced the United States as the world’s top emitter of CO₂, adding 2.8 Pg C to the atmosphere in 2014, about twice U.S. emissions (Olivier et al., 2016). In terms of cumulative emissions, the United States is responsible for 100 Pg C out of a global total of 378 Pg C (UNFCCC 2013; World Resources Institute 2017). If land-use change and forestry are taken into account, U.S. contributions have totaled 134 Pg C out of a global total of 572 Pg C of net emissions. For comparison, historical emissions (including land-use change and forestry) of EU countries and China are 114 and 74 Pg C, respectively.

Both inventory (i.e., field measurements) and modeling techniques have been used to estimate land-based carbon sinks for North America (King et al., 2015). These estimates show that human-caused carbon emissions in North America are significantly higher than the land’s capacity to absorb and store them. For example, estimates suggest that between 2000 and 2009, only 15% to 49% (with a mean estimate of 26%) of North American fossil fuel emissions were absorbed by North American lands (King et al., 2015). As a result, North America is considered to be an overall net source of carbon to the atmosphere. However, the ability of North American land to take up and store carbon is significant. Globally, estimates suggest that over the past decade (2006 to 2015) 2.4 ± 0.5 Pg C per year were taken up by the ocean and 3.0 ± 0.8 Pg C per year were taken up by the terrestrial biosphere (Le Quéré et al., 2017). Of these totals, the amount taken up by the terrestrial biosphere in North America is estimated to be about 0.47 Pg C per year (King et al., 2015), or 15% of global terrestrial uptake.

Carbon uptake by North American lands is driven largely by the regrowth and recovery of forests from earlier human-driven changes in land cover and land use, such as forest clearing and harvesting (King et al., 2015), as well as increases in forest area from improved forest management practices (Melillo et al., 2014). Environmental influences on plant growth, such as the fertilizing effects of rising concentrations of atmospheric CO₂ and nitrogen, along with changes in climate including longer growing seasons in northern midlatitude regions also have contributed to increased carbon uptake in North America over the past two decades (King et al., 2015; Melillo et al., 2014; see Ch. 2: The North American Carbon Budget).

However, the emissions of other GHGs, primarily CH₄ and N₂O, partially offset the potential climate cooling induced by the uptake of CO₂ in North America (Tian et al., 2016). North America accounts for about 10% of natural (e.g., wetlands) and 12% of human-driven (e.g., agriculture and fossil fuels) global CH₄ emissions (Kirschke et al., 2013; see Ch. 2: The North American Carbon Budget).

²National Oceanic and Atmospheric Administration Global Monitoring Division, Trends in Atmospheric Carbon Dioxide
³Hofmann et al. (2006), updated at www.esrl.noaa.gov/gmd/aggi/
⁴World Resources Institute

Coupled carbon cycle–climate models forced with future “business as usual” emissions scenarios suggest that the changing carbon cycle will be a net positive feedback on climate, reinforcing warming, but the size of the projected feedback is highly uncertain (Friedlingstein et al., 2014). Besides the uncertain trajectories of human factors such as fossil fuel emissions, land use, or significant mitigation efforts, various natural processes can lead to the carbon cycle being a positive feedback. For example, a warming climate can lead to increased fires and droughts and less storage of carbon in the terrestrial biosphere. In particular, warming is expected to decrease carbon uptake in the tropics and midlatitudes. In the high latitudes, a warmer climate is expected to lead to a more productive biosphere and more uptake but also may result in increased respiration and release of stored CO₂ and CH₄ in soils and lakes. Negative feedbacks also are possible, such as increased atmospheric CO₂, leading to increased carbon storage in the terrestrial biosphere (e.g., Schimel et al., 2015), although the relative roles of this effect relative to land-use change, nitrogen deposition, and temperature increases on the cumulative land carbon sink over the last century are not fully understood (Huntzinger et al., 2017).

Human impacts on land use can directly impact climate. Deforestation and agriculture can affect carbon storage in soil and biomass. Fertilizer use also affects the global nitrogen budget and can increase carbon storage. Large-scale drainage of wetlands and conversion to agricultural land can reduce CH₄ emissions from anaerobic respiration while potentially increasing faster soil carbon loss through aerobic respiration.

The ocean carbon sink is driven primarily by the partial pressure difference of CO₂ between the atmosphere and the ocean surface (∆pCO₂). Although this mechanism would imply that increasing atmospheric CO₂ concentrations would, therefore, lead to increased uptake of CO₂ in the ocean, there actually is substantial uncertainty in future uptake due to uncertainty in future changes to ocean circulation, warming, and chemical changes, all of which would impact the ocean sink (Lovenduski et al., 2016; Randerson et al., 2015). In addition, the sequestration of CO₂ in ocean water also can lead to undesirable impacts as the ocean becomes more acidic. For example, ocean acidification disrupts the ability of organisms to build and maintain calcium carbonate (CaCO₃) shells, substantially perturbing ocean ecosystems.

Frozen Arctic soils compose another potential carbon cycle–climate feedback (see Ch. 11: Arctic and Boreal Carbon and Ch. 19: Future of the North American Carbon Cycle). An estimated 1,460 to 1,600 Pg C are frozen in Arctic soils, and warming has proceeded in the Arctic faster than in any other region. Current understanding suggests that approximately 146 to 160 Pg C, primarily as CO₂, could be vulnerable to thaw and release to the atmosphere over the next century (Schuur et al., 2015; see Ch. 11: Arctic and Boreal Carbon). This release of carbon from permafrost is likely to be gradual and occur on century timescales (Schuur et al., 2015). If the amount of carbon estimated to enter the atmosphere by Schuur et al. (2015) were released annually at a constant rate, emissions would be far lower than annual fossil fuel emissions (about 9 Pg C per year) but comparable to land-use change (0.9 Pg C per year).

Factors that will affect the carbon cycle are explored in much more depth in respective chapters of this report, and Ch. 19 describes future projections and the results of different IPCC scenarios on the North American carbon cycle in a global context.

Concern about the effects of climate change, on the one hand, and the difficulties of reducing emissions of carbon from fossil fuel use, on the other, have led to a target of limiting global average warming to no more than 2°C, with a more conservative target of 1.5°C to reduce the risks of the most serious effects of climate change (USGCRP 2017). The choice of 2°C reflects a balance between a realistic threshold and one that would result in a presumably tolerable amount of climate change. However, as Knutti et al. (2015) points out, no proof exists that this threshold maintains a “safe” level of warming, and the definition of “safe,” as well as the components of the Earth system that the term applies to, are themselves subjective. Several recent studies have suggested that the accumulated carbon in the atmosphere already may have committed the climate system to 2°C or more of global average temperature increase (Mauritsen and Pincus 2017; Raftery et al., 2017).

The relationship of cumulative carbon emissions to global temperature increase depends on the data constraints or model used to simulate the temperature response. Gillett et al. (2013) reports an observationally constrained range of 0.7 to 2.0°C per 1,000 Pg C (5% to 95% confidence interval) and a range of 0.8 to 2.4°C per 1,000 Pg C based on 15 models from the Coupled Model Intercomparison Project Phase 5 (CMIP5). Similarly, IPCC (2013) estimates that limiting the warming with a probability of >33%, >50%, and >67% to less than 2°C since the period 1861 to 1880 will require cumulative emissions from all anthropogenic sources to stay below about 1,570 Pg C, 1,210 Pg C, and 1,000 Pg C since that period, respectively. Cumulative emissions since 1850, including land-use change and forestry, are 572 Pg C (Global Carbon Project 2016; Peters et al., 2015; World Resources Institute 2017). However, this amount includes only the carbon from CO₂ emissions and does not include non-CO₂ emissions (i.e., primarily CH₄ and N₂O), which amount to an additional 210 Pg C equivalent from non-CO₂ sources, bringing the total to 779 Pg C equivalents (Peters et al., 2015). This amount implies that, to achieve a >33%, >50%, and >67% warming probability limited to below 2°C, amounts of no more than 791, 431, or 221 Pg C equivalent, respectively, can be emitted from 2017 forward. Current annual global emissions of CO₂ from fossil fuel combustion and cement production are 10.7 Pg C per year (Le Quéré et al., 2017), so this limit could be reached in less than 80, 40, or 20 years. Although technically achievable (Millar et al., 2017), the most conservative emissions reductions would require immediate and concerted action.

These simple estimates of cumulative emissions and their effect on future global temperature, however, have many uncertainties. Uncertainties in climate models include cloud, aerosol, and carbon cycle feedbacks. Carbon-climate feedbacks, such as the effect on carbon emissions from permafrost thaw, are highly uncertain and may significantly lower the cumulative amount of carbon that can be emitted before exceeding the 2°C global temperature increase.

Attempts to avoid the most severe impacts of climate change through management of the carbon cycle rely on reducing emissions and increasing storage in land and ocean reservoirs. Other means that focus on adaptation are not specifically addressed in this report. Evaluating and predicting the success of these strategies require an understanding of all the natural and anthropogenic components of the global carbon cycle because decreases in emissions or increases in sinks from mitigation activities may be offset partially or wholly by changes in other components. Globally, land and ocean sinks have averaged between 3.9 and 4.7 Pg C per year since 2000 (Le Quéré et al., 2016), growing over time in proportion to emissions (Ballantyne et al., 2012). The sink on land, accounting recently for about 25% of total emissions (Le Quéré et al., 2016), is consistent with the measured increase in carbon stocks of forests (Pan et al., 2011). In North America, the forest sink is currently about 223 Tg C per year (see Ch. 9: Forests), but increases in the frequency of wildfires and insect infestations in the western continent threaten to reduce that sink. The sink in Canadian forests, though much smaller than that in the United States, also is threatened by insects and wildfire and could become a significant source (Kurz et al., 2013), as has happened recently. Mexican forests also are thought to be a small sink based on estimates of regrowth of previously disturbed forests that exceed emissions from deforestation and forest degradation (see Ch. 9: Forests).

Options for managing emissions of carbon and other GHGs include 1) reduction or cessation of the use of fossil fuels, replacing them with renewable sources of energy (e.g., solar, wind, and water); 2) climate intervention via carbon dioxide removal (CDR), including carbon capture and storage (CCS), which involves absorption of emissions at point sources; and 3) negative emissions, using approaches to remove previously emitted CO₂ by increasing storage in terrestrial and ocean reservoirs. Climate intervention via albedo modification does not affect the carbon budget directly but is an attempt to counteract climate change by directly influencing the global radiation balance. For example, introducing aerosols into the stratosphere potentially could provide a global cooling effect but would not address other issues such as ocean acidification. Climate intervention will not be discussed here further; rather, the focus of this section is on actions that directly involve the carbon cycle.

The study of MacDonald et al. (2016) estimated that U.S. carbon emissions from the power sector could be reduced by as much as 80% relative to 1990 use without significantly increasing energy costs and using existing technology. Although some studies have argued that a complete transition to decarbonized energy systems is feasible (Jacobson et al., 2015), other authors have pointed out that a transition to a low-carbon energy system is likely to be difficult and expensive without using a range of options (Clack et al., 2017), including some contribution from fossil fuels. This issue is complex, and full discussion of it is beyond the scope of this report.

For the CCS option, there are many unknowns about its implementation and permanence. A special example of CCS involves renewable energy, in this case bioenergy CCS (BECCS), where energy is derived from burning biomass, capturing and storing the resulting CO₂, and then re-growing the biomass. Although BECCS is appealing because it replaces fossil fuels and removes carbon from the atmosphere, there is only one experimental biomass plant of this type and its technology suffers from the same uncertainty as other CCS types (Anderson and Peters 2016; Fuss et al., 2014).

Estimates of the potential for negative emissions are in the range of 1.6 to 4.4 Pg C per year or 34 to 105 Pg C by 2100 (Griscom et al., 2017; Houghton and Nassikas 2018). Achieving the potential of negative emissions, however, has other constraints involving competition for land area, water availability, albedo changes, and nutrient limitations (Smith et al., 2015). Most negative emissions activities on land are useful either as a bridge to a low–carbon emissions energy system for developing and implementing CCS or for assistance with future removals of previously emitted CO₂, but effects are limited in implementing long-term solutions because forests and soils cannot accumulate carbon at high rates indefinitely. The most rapid rates of carbon removal occur in the first 50 to 100 years of forest growth. Soils generally are slow to accumulate carbon, although that process in forests may last for centuries if the forests remain undisturbed (Luyssaert et al., 2008). Thus, negative emissions are a part of the portfolio of mitigation activities, but the timing of impacts needs to be considered. These negative emissions cannot compensate for future emissions that either continue at current rates or increase (Gasser et al., 2015). Furthermore, the effects of climate change on the carbon balance of terrestrial ecosystems are uncertain, as suggested by the increased mortality of U.S. forests from droughts, insects, and fires.

Another unknown is how much of an overshoot is possible—that is, by how much and for how long emissions could exceed the limit imposed by a 2°C ceiling and their effects still be reversible. Moreover, questions include: How would they be reversed with only limited, available negative emissions? What are the tipping points? For example, warming already is thawing permafrost and thereby exposing ­long-frozen organic carbon to oxidation. Estimates are that emissions of carbon from thawing permafrost could be 146 to 160 Pg C by 2100 (Schuur et al., 2015), enough to counter negative emissions. Similarly, disruption of tropical and subtropical ecosystems could lead to substantial releases of carbon into the atmosphere. Avoidance of tipping points is a paramount challenge to civilization. Only by continuing to seek a better understanding of the carbon cycle can the predictability of these events be improved.