Future of the North American Carbon Cycle

Key Finding 2

Land, ocean, coastal, and freshwater systems are currently net sinks of carbon from the atmosphere, taking up more carbon annually than they release. However, emerging understanding suggests that the future carbon uptake capacity of these systems may decline, depending on different emissions scenarios, with some reservoirs switching from a net sink to a net source of carbon to the atmosphere (high confidence).

Description of evidence base

Most work examining future carbon cycle changes and potential feedbacks with climate and rising atmospheric carbon dioxide (CO₂) has been conducted at the global scale as part of coupled carbon-climate model intercomparison efforts including the Coupled Model Intercomparison Project Phase 5 (CMIP5; Friedlingstein 2015; Friedlingstein et al., 2014). As a result, published estimates of projections specific to both the land carbon sink and coastal ocean carbon uptake in North America are lacking.

To provide an estimate of future land carbon sink evolution in North America, this chapter relied on the globally gridded net biome productivity simulated by nine CMIP5 models (Ciais et al., 2013; Friedlingstein 2015). With the exception of CESM1-BGC, which was not available on the CMIP5 data download page, the models and set of simulations used here (and in Figures 19.3 and 19.4) are the same as those used in Ch. 6 of the Intergovernmental Panel on Climate Change Fifth Assessment Report (IPCC; Table 6.11): CanESM2, GFDL-ESM2G, GFDL-ESM2M, HadGEM2–ES, IPSL–CM5A-LR, MIROC-ESM, MPI–ESM–LR, NorESM1–ME, and INMCM4. The simulation output was placed into a consistent 0.5° grid and trimmed to North America (10° to 70°N and 50° to 170°E). Projected land sink estimates were evaluated for all four of the Representative Concentration Pathways (RCPs; van Vuuren et al., 2011) used in the latest IPCC report:

1. RCP8.5 High Emissions Scenario. Projects increasing CO₂ and methane (CH₄) emissions over time due to increased energy intensity as a result of high population growth and lower rates of technology development leading to radiative forcing of 8.5 watts per square meter (W/m²) by 2100. This scenario assumes an increase in cropland and grassland area driven by the demands of population growth.

2. RCP6.0 Stabilization Scenario. Projects a range of technologies and strategies to reduce CO₂ emissions after the year 2080, coupled with fairly steady CH₄ emissions throughout the century to stabilize radiative forcing at 6 W/m² in 2100. This scenario assumes an increase in cropland area, but a decline in pasture area due to aggressive implementation of intensive animal husbandry.

3. RCP4.5 Stabilization Scenario. Projects a range of technologies and strategies to reduce CO₂ emissions after 2040, coupled with fairly steady CH₄ emissions throughout the century to stabilize radiative forcing at 4.5 W/m² in 2100. This scenario assumes a decrease in cropland and grassland area due to climate policies that value carbon in natural vegetation.

4. RCP2.6 Low Emissions Scenario. Projects an increased use of bioenergy and carbon capture and storage, which leads to substantial reduction in CO₂ emissions after 2020. This reduction coupled with declining CH₄ emissions from energy production, transportation, and livestock leads to a peak in radiative forcing of 3 W/m², followed by a decline to 2.6 W/m² by 2100. Cropland area increases, but largely as a result of bioenergy production. Grassland area remains relatively constant as the increase in animal production is offset by more intensive animal husbandry.

For the North American coastal ocean, this report used three CMIP5 models (GFDL-ESM2M [Dunne et al., 2013], HadGEM-ESM [Martin et al., 2011], and MIROC-ESM [Watanabe et al., 2011]) to estimate a range of historical (1870 to 1995) and future carbon uptake within the exclusive economic zones (EEZs) of North America (approximately 22.5 × 10⁶ km²). Since 1870, North American EEZs have taken up 2.6 to 3.4 petagrams of carbon (Pg C). These regions are projected to take up an additional 10 to 12 Pg C by 2050 and another 17 to 26 Pg C in the second half of this century (2050 to 2100). Global projections of ocean carbon uptake vary depending on emissions scenarios (Ciais et al., 2013). Under lower future emissions scenarios (e.g., RCP2.6 and RCP4.5), the strength of the ocean carbon sink starts to level off toward the end of the century. For the North American Pacific Coast, the combined effect of multiple factors (e.g., increasing atmospheric CO₂, surface warming, less vertical mixing with greater vertical stratification, and increases in horizontal temperature gradients) may lead to greater and more persistent CO₂ outgassing nearshore and lower productivity offshore (see Ch. 16: Coastal Ocean and Continental Shelves).

Major uncertainties

The balance between positive and negative influences of climate and atmospheric CO₂ on the global carbon cycle is not well constrained in models (see Figure 19.5; Ciais et al., 2013; Graven 2016). Although models tend to agree on the direction of the carbon uptake response to both climate warming and rising CO₂, they show low agreement on the magnitude (size) of this response (Ciais et al., 2013). In land carbon cycling, many current models do not consider nutrient cycle processes or the coupling of the nitrogen and carbon cycles (Ciais et al., 2013). In addition, model response to climate warming is highly uncertain. Climate warming could lead to an increase or decrease in carbon uptake, depending on a number of factors that will vary by region and the species present within a given ecosystem (Graven 2016). Major sources of uncertainty in models are projected changes in permafrost and soil carbon storage (see Section 19.7.2). Many models do not explicitly account for permafrost dynamics and include outdated representations of soil carbon turnover that are inconsistent with emerging scientific understanding (Bradford et al., 2016).

Assessment of confidence based on evidence and agreement, including short description of nature of evidence and level of agreement

Land, ocean, coastal, and freshwater systems are currently net sinks of carbon from the atmosphere. Although projections vary depending on future climate and carbon emissions scenarios, it is likely that under some future climate and CO₂ emissions scenarios these systems will turn from a net sink to a net source of carbon.

Summary sentence or paragraph that integrates the above information

It is the balance between the response of land and ocean systems to future climate and rising atmospheric CO₂ that will ultimately determine the strength and extent of carbon uptake by these systems and whether they continue to be net sink of carbon from the atmosphere or switch to being a net source.

Key Finding 3

Human-driven changes in land cover and land use will continue to be key contributors to carbon cycle changes into the future, both globally and in North America. Globally, land-use change is projected to contribute 10 to 100 Pg C to the atmosphere by 2050 and between 19 and 205 Pg C by 2100. Conversely, in the United States, land use and land-use change activities are projected to increase carbon stocks in terrestrial ecosystems by about 4 Pg C from 2015 to 2030. This projected increase is primarily driven by the growth of existing forests and management activities that promote ecosystem carbon uptake, often in response to changes in market, policy, and climate (high confidence).

Description of evidence base

Global estimates are based on Brovkin et al. (2013), who examined the difference in land carbon storage between the ensemble averages of simulations with and without land-use changes using RCP2.6 and RCP8.5. The RCP2.6 scenario assumes that climate change mitigation is partially achieved by increasing the use of bioenergy crops. Under this scenario, the global land area used for pastures is more or less constant over the simulation period, and increases in production (animal-based products) are achieved through changes in approaches to animal husbandry (Brovkin et al., 2013). In the RCP8.5 scenario, food demands and increasing population drive the expansion of croplands and pastures (and the loss of forested lands). The model ensemble includes six CMIP5 models for the projections: CanESM2, EC-Earth, HadGEM2-ES, IPSL-CM5A-LR, MIROC-ESM, and MPI-ESM-LR. Across all models, Brovkin et al. (2013) found a robust signal showing a loss of global land carbon storage because of projected land-use and land-cover change activities.

There is a lack of projections of emissions and sink trends for land use, land-use change, and forestry (LULUCF) activities specific to North America as a whole. U.S. estimates are based on the Second Biennial Report of the United States of America (U.S. Department of State 2016). That report presents a range in carbon sequestration estimates (689 to 1,118 teragrams [Tg] of CO₂ equivalent [CO₂e] per year by 2030) associated with U.S. land-use change and forestry activities. Also estimated is that emissions from forestry and land use will be 28 Tg CO₂e in 2030.

To project cumulative carbon uptake from 2015 to 2030, the emissions estimate associated with forestry and land use (28 Tg CO₂e) is subtracted from the low and high estimates of sequestration associated with forestry and land use (689 to 1,118 Tg CO₂e). These values are then combined and divided by 2 to arrive at an average projected net uptake per year in 2030 of 875.5 Tg CO₂e per year. This value is converted to teragrams of carbon (239 Tg C per year) and multiplied by 15 to arrive at a cumulative uptake of 3.6 Pg C from 2015 to 2030.

Major uncertainties

Uncertainties arise from how land use and land-use change information is implemented into the carbon cycle representation of ecosystem models (i.e., the inclusion or exclusion of specific land-use processes such as wood harvest; Brovkin et al., 2013). In global projections, uncertainty also arises from the lack of coupled carbon-nitrogen (and phosphorus) dynamics in models. The models in this study do not account for the effect of nitrogen or phosphorus limitation on land ecosystems or CO₂ fertilization.

For both the global and North American projections, there is also uncertainty in estimates of population growth and its potential impact on forest and agricultural land area. Moreover, there is general uncertainty in the potential future magnitude and timing of land-use change impacts on the land carbon cycle because of the difficulty in projecting the outcome of complex and interacting environmental, climate, and socioeconomic systems.

Assessment of confidence based on evidence and agreement, including short description of nature of evidence and level of agreement

Several studies generally agree with high confidence that direct human influence on land use and land-cover change is a large driver of future potential carbon cycle changes. Model projections for North America agree that U.S. LULUCF activities will continue to result in net carbon uptake (i.e., carbon sequestration) to 2030. However, uncertainty in population growth and its impact on forests and agricultural land leads to considerable uncertainty in carbon uptake projections beyond 2030 associated with land-use change and forestry activities.

Summary sentence or paragraph that integrates the above information

There is high confidence that land use, land-use change, and management play important roles in both the global and North American carbon cycles. However, the future magnitude and timing of carbon cycle changes emerging from land use and land-use change depend on a number of factors that are difficult to project, including population growth and environmental and economic policies, all of which will drive changes in land use.

Key Finding 4

The enhanced carbon uptake capacity of ocean and terrestrial systems in response to rising atmospheric CO₂ will likely diminish in the future. In the ocean, warmer and more CO₂-enriched waters are expected to take up less additional CO₂. On land, forest maturation, nutrient limitations, and decreased carbon residence time in soils will likely constrain terrestrial ecosystem response to rising CO₂ (high confidence).

Description of evidence base

Although models tend to agree on the direction of the carbon uptake response to rising CO₂, they show low agreement on the magnitude (i.e., size) of this response, particularly for terrestrial ecosystems (see Figure 19.5). However, some factors potentially important for limiting the CO₂ fertilization response of terrestrial ecosystems are not currently represented in models, including 1) the age distribution of forest trees, 2) nutrient limitation, and 3) soil carbon turnover rates.

Forest Age. Ecosystem CO₂ enrichment experiments in North American forests tend to show that, in the short term (e.g., up to 10 years), CO₂ fertilization increases forest production by 20% to 25% (McCarthy et al., 2010; Norby et al., 2010; Talhelm et al., 2014). However, most of these forest experiments were conducted in young forests that also were accumulating biomass under ambient CO₂ concentrations. The few experiments that have been conducted on individual trees in more mature forests tend to show little or no growth response (Bader et al., 2013; Klein et al., 2016).

Nutrient Limitation. Nutrients will likely constrain land carbon cycle response to rising CO₂ (e.g., Norby et al., 2010). Many current models do not consider nutrient cycle processes (Ciais et al., 2013; Hoffman et al., 2014), contributing substantial uncertainty to the overall accuracy of CO₂–carbon cycle feedback estimates. Even models that do consider nutrient cycling exhibit substantial uncertainty in responses of terrestrial ecosystems to increased atmospheric CO₂ (Walker et al., 2015; Zaehle and Dalmonech 2011).

Soil Carbon Turnover Rates. Results from some studies suggest that soil carbon storage may increase with rising atmospheric CO₂ (e.g., Iversen et al., 2012), even if the latter does not lead to increased carbon storage in forest biomass. However, soil carbon input may change microbial decomposition rates and the rate of soil carbon turnover, leading to less overall soil carbon storage (Hungate et al., 2013; van Groenigen et al., 2014).

In the ocean, warmer and more CO₂-enriched waters are expected to take up less additional CO₂ and be less resistant to changes in pH (Ciais et al., 2013). Several studies (Gattuso et al., 2015; Randerson et al., 2015; Bopp et al., 2013; Doney et al., 2009) have investigated in detail the impacts of contrasting emissions scenarios on ocean dynamics and marine and coastal ecosystems, including the goods and services that they provide. Alongside changes in ocean dynamics and a slowing of the ocean sink, these studies also highlight the fact that phytoplankton and zooplankton populations are likely to shift toward groups that favor higher temperature, greater physical stratification, and elevated CO₂ conditions, both in terms of trait diversity within groups (e.g., Dutkiewicz et al., 2013) and in some groups being favored over others (e.g., slow growing, CO₂-limited nitrogen fixers; Hutchins et al., 2007).

Major uncertainties

See previous section describing the evidence base.

Assessment of confidence based on evidence and agreement, including short description of nature of evidence and level of agreement

Models tend to agree on the direction of land and ocean carbon uptake response to rising CO₂, but they show less agreement on the magnitude of this response. However, multiple points of evidence suggest that the strength of net carbon uptake in response to rising CO₂ will decrease into the future.

Summary sentence or paragraph that integrates the above information

The recent increase in the carbon uptake capacity of ocean and terrestrial systems in response to rising atmospheric CO₂ from human-driven emissions will likely diminish in the future. Warmer and more CO₂-enriched ocean waters are expected to take up less CO₂ as climate warms due to a number of factors. Such factors, including forest maturation, nutrient limitations, and decreased carbon residence time in soils, will likely constrain terrestrial ecosystem response to rising CO₂.

Key Finding 5

Soil carbon losses in a warming climate will be a key determinant of the future North American carbon cycle. An important region of change will be the Arctic, where thawing permafrost and the release of previously frozen carbon will likely shift this region from a net sink to a net source of carbon to the atmosphere by the end of the century (very high confidence).

Description of evidence base

A meta-analysis of results from soil warming experiments indicates that soil carbon stock response to climate warming is variable but predictable and depends on the size of the soil carbon pool and the extent and duration of warming (Crowther et al., 2016). As a result, projected soil carbon losses are greatest at northern latitudes (e.g., Arctic and subarctic; see Figure 19.7, which have large soil carbon stocks and some of the most rapid rates of projected warming (Crowther et al., 2016; see also USGCRP 2017a and Section 19.3.3). With continued warming and large-scale losses of near-surface permafrost, almost all terrestrial carbon cycle models indicate that, by the end of this century, the Arctic could shift from a sink to a source of carbon (Cox et al., 2000; Fisher et al., 2014b).

Major uncertainties

Although there is considerable agreement that climate warming will lead to carbon loss from permafrost regions, the amplitude, timing, and form of carbon release remain topics of debate (e.g., McGuire et al., 2018; Lenton et al., 2008; Schuur et al., 2015; Slater and Lawrence 2013). This disagreement stems from a lack of understanding of three key factors that determine the potential climate feedback of the permafrost carbon pool: 1) the area and depth of permafrost vulnerable to release, 2) the speed with which carbon will be released from thawing soils, and 3) the form of carbon (e.g., CO₂ and CH₄) that will be released (Schuur et al., 2013, 2015).

Assessment of confidence based on evidence and agreement, including short description of nature of evidence and level of agreement

While some uncertainty remains about the timing, speed, and form of carbon release from permafrost thaw, there is strong agreement across multiple studies that climate warming will result in carbon loss from permafrost soils. Over time, under increased rates of warming in the Arctic, the carbon loss from permafrost thaw will likely cause high northern latitudes to switch from a net sink to a net source of carbon to the atmosphere.

Summary sentence or paragraph that integrates the above information

Although the amplitude, timing, and form of carbon released from thawing permafrost are still under study, there is very high confidence that warming will lead to soil carbon loss from permafrost regions.

Key Finding 6

Carbon storage in both terrestrial and aquatic systems is vulnerable to natural and human-driven disturbances. This vulnerability is likely to increase as disturbance regimes shift and disturbance severity increases with changing climatic conditions (high confidence).

Description of evidence base

Natural and human-driven disturbances will influence future vegetation carbon storage. Forest disturbance is a fundamental driver of terrestrial carbon cycle dynamics (Hicke et al., 2012). Harvesting, fire, wind throw, storms, pathogen and pest outbreaks, and drought collectively lead to the removal of 200 Tg C from U.S. forests annually (Williams et al., 2016). Initially, most disturbances shift an ecosystem to a carbon source, while recovery from disturbance is commonly associated with greater net ecosystem carbon storage (Magnani et al., 2007; Odum 1969). Hence, the effects of disturbance on carbon balance in forests are both immediate and lagged, and potentially long lasting. Given current management practices, climate change is likely to increase disturbance frequency and intensity across multiple spatial and temporal scales (Running 2008). Fire activity generally is expected to increase (Sommers et al., 2014; Westerling et al., 2006) in many regions, with fire seasons starting earlier and ending later compared to previous decades (Jolly et al., 2015). With climate warming, the range of insects (e.g., mountain pine beetle) is expected to expand into higher elevations and latitudes, putting previously unaffected forests at risk (Bentz et al., 2010; Kurz et al., 2008). Evidence suggests that the extent and severity of forest insect disturbances also are increasing with changing climate conditions (Kurz et al., 2008).

Freshwater ecosystems are particularly vulnerable to anthropogenic disturbances and are considered to be among the most threatened ecosystems on the planet (Vorosmarty et al., 2010). Human activities such as water management, river fragmentation by dams, alteration of natural flow, construction of water impoundments, and land-use changes have a major impact on freshwater ecology, biology, and carbon cycling. There is high confidence that direct human impacts—including increasing urbanization, expansion of irrigated agriculture, and growing demand for water resources—will continue to dominate the threats to most freshwater ecosystems globally over the next three decades (Settele et al., 2014).

Major uncertainties

Projections of future carbon cycle processes are highly sensitive to the ability of models to simulate external forcings. When projecting future carbon responses to natural and human-driven disturbances, there is a great deal of uncertainty (and intrinsic difficulty) in modeling disturbance events, particularly their timing, extent, and severity (Luo et al., 2015). Also, understanding and predicting the impacts of natural and human-driven disturbances on the carbon cycle require insights into and the ability to project management decisions, human use of land and aquatic systems, and the dynamic coupling and interconnectivity between natural and human-driven activities.

Assessment of confidence based on evidence and agreement, including short description of nature of evidence and level of agreement

While uncertainties remain in the ability to project the exact magnitude of carbon cycle impacts due to future disturbance events, the trajectory of land and aquatic carbon storage and loss is vulnerable to both natural and human-driven disturbances. As climate conditions change and the occurrence of extreme weather events increases, the impacts of disturbances on ecosystem carbon storage is likely to increase.

Summary sentence or paragraph that integrates the above information

Natural and human-driven disturbance will influence future vegetation carbon storage. Carbon storage in terrestrial and aquatic systems is vulnerable to disturbance events, and this vulnerability is likely to increase as disturbance regimes shift and disturbance severity increases with changing climatic conditions. However, the intrinsic predictability of disturbance events and their drivers is challenging.