Lead Authors:
Daniel J. Hayes, University of Maine
Rodrigo Vargas, University of Delaware
Contributing Authors:
Simone R. Alin, NOAA Pacific Marine Environmental Laboratory
Richard T. Conant, Colorado State University
Lucy R. Hutyra, Boston University
Andrew R. Jacobson, University of Colorado, Boulder, and NOAA Earth System Research Laboratory
Werner A. Kurz, Natural Resources Canada, Canadian Forest Service
Shuguang Liu, Central South University of Forestry and Technology
A. David McGuire, U.S. Geological Survey and University of Alaska, Fairbanks
Benjamin Poulter, NASA Goddard Space Flight Center
Christopher W. Woodall, USDA Forest Service
Science Lead:
Melanie A. Mayes, Oak Ridge National Laboratory
Review Editor:
Tara Hudiburg, University of Idaho
Federal Liaison:
Noel P. Gurwick, U.S. Agency for International Development

The North American Carbon Budget

SOCCR1 (CCSP 2007) concluded that North America was a net source of carbon to the atmosphere ca. 2003, with the magnitude of fossil fuel emissions outpacing the rate of carbon uptake by land sinks. The synthesis of carbon flux estimates in SOCCR2 suggests that North America has remained a carbon source in the decade since SOCCR1, continuing to contribute to the global rise in atmospheric CO2 and CH4 concentrations from 2004 to 2013. Synthesizing across the major continental-scale budget components, SOCCR2 assessments suggest that approximately 57% of the total fossil fuel emissions from Canada, the United States, and Mexico remains in the atmosphere after the offsetting portion is taken up by a net sink across North American land ecosystems, inland waters, and adjacent coastal ocean areas. This overall estimate of the “airborne fraction” of fossil fuel emissions is less than the estimated 70% reported in SOCCR1, a decrease that is a function of both a reduction in the total emissions estimate coupled with an increase in the net continental sink estimate for 2004 to 2013. The values in SOCCR2 also reflect additional information and improved understanding of components and sectors influencing the continental carbon budget, but large uncertainties in some components must be addressed to achieve a better understanding of the trends.

This report estimates that the total fossil fuel carbon source in North America from 2004 to 2013 was 1.8 Pg C per year, representing an approximately 5% reduction in annual emissions compared to the ca. 2003 estimate of 1.9 Pg C per year. The lower current emissions estimate is likely a result of changing technology, policy, and market factors (see Ch. 3: Energy Systems). Despite the modest reduction in emissions, the fossil fuel source still represents the largest single component in the continental-scale carbon budget. The relative contributions from each of the three countries have remained constant since SOCCR1, with the United States continuing to contribute the vast majority (85%) of total continental emissions. The total fossil fuel emissions from energy and transportation systems across North America likely will remain the dominant source category and continue to outpace the ability of the continental land ecosystems, inland waters, and adjacent coastal ocean areas to take up this carbon in the future.

North America’s natural and managed land ecosystems, inland waters, and adjacent coastal ocean areas likely will remain a net carbon sink, thereby partially constraining the airborne fraction of fossil fuel emissions and further mitigating climate impacts from rising atmospheric CO2. Bottom-up, ­inventory-based analyses have confirmed the existence of the continental carbon sink, but the uncertainty associated with these approaches provides less confidence in estimates of the sink’s magnitude than in the better-constrained estimates of fossil fuel emissions. The “best estimate” of the continental sink from 2004 to 2013 in SOCCR2 is 766 Tg C per year, compared to 505 Tg C per year estimated in SOCCR1. The difference in these two bottom-up estimates can be explained by the additional components considered in SOCCR2 that were not accounted for in SOCCR1. These components include Arctic and boreal ecosystems; estuaries; and updated information and accounting for grasslands, inland water fluxes, terrestrial and tidal wetlands, and the coastal ocean. Still, both the SOCCR1 and SOCCR2 estimates fall within the uncertainty bounds of the other and thus are not considered a trend nor significantly different from each other.

Given the large uncertainty in the bottom-up analysis, comparing it with top-down estimates is important to collectively provide an additional constraint on the overall continental sink estimate. Previous comparisons typically have shown mean estimates of the continental CO2 sink from top-down atmospheric models to be much greater than those from bottom-up inventory and biosphere models, although within the large range of uncertainty in these estimates (King et al., 2012; Pacala et al., 2001). In a progression of studies over time, mean land sink estimates based on atmospheric models have decreased from 1,700 ± 500 Tg C per year (Fan et al., 1998) to 890 ± 409 Tg C per year (King et al., 2015). Meanwhile, best estimates for the sum of sink components from inventory-based methods will increase as additional components are included in the calculation. For example, including estimates of highly uncertain components (e.g., woody encroachment, wetlands, and the net flux in inland waters) increased the sink estimate to 564 Tg C per year from the 325 Tg C per year that only considered reported inventory estimates for forests and agriculture (Hayes et al., 2012). In conclusion, the larger bottom-up sink estimates approach the lower end of the uncertainty in the atmospheric model estimates as these additional components are added, though they also greatly increase the uncertainty of the estimates (King et al., 2012).

SOCCR2 shows further convergence between the top-down, continental-scale carbon sink estimate from atmospheric modeling and the synthesis of estimates from bottom-up approaches across the major components of North America (see Figure 2.5). This convergence partly results from a series of operational, conceptual, and technological improvements. The analysis of a growing network of atmospheric measurements of CO2 and CH4 using inverse modeling techniques has increased significantly since SOCCR1. Several flux modeling systems produce regular continental-scale estimates on an operational basis, and regional inverse modeling studies are now focused on specific land areas and individual megacities (see Ch. 8: Observations of Atmospheric Carbon Dioxide and Methane). Furthermore, recent atmospheric inverse model analyses estimate the continental land sink to be 699 ± 82 Tg C per year, which includes all continental carbon fluxes from land and water but not the coastal ocean sink (see Ch. 8). These estimates are only slightly higher than the bottom-up estimate of 606 Tg C per year that is calculated by removing the coastal ocean sink from the continental total (see Table 2.2). Considering the uncertainty ranges of the two approaches, there is an apparent agreement in the magnitude of the continental carbon sink over the last decade between the top-down and bottom-up estimates in this report. The inverse model analysis of atmospheric CO2 data suggests that there is substantial variability in land-atmosphere carbon fluxes over North America from year to year, though a comparable analysis reported from bottom-up estimates is not possible here because of averaged stock change estimates over the longer time periods between inventories. Additionally, the atmospheric estimates show at least moderate evidence of an increasing rate of carbon uptake in the continental land sink from 2000 to 2014, but any such trend is difficult to ascertain from the bottom-up estimates between SOCCR1 and SOCCR2 because of differences in the components that are included and how they are calculated.


Figure 2.5: Estimates of the North American Carbon Sink in this Century

Figure 2.5: These estimates, in teragrams of carbon (Tg C) per year, are derived from inventory analysis, atmospheric inversion models (AIMs), and terrestrial biosphere models (TBMs). [Data sources: First State of the Carbon Cycle Report (SOCCR1; CCSP 2007), North American Carbon Program (NACP; Hayes et al., 2012), REgional Carbon Cycle Assessment and Processes (RECCAP) initative (King et al., 2015), and this report (SOCCR2). Publication year of each estimate is given in parenthesis.]


Given the general convergence with the current atmosphere-based estimates, the bottom-up estimates synthesized in this report are unlikely to be missing any major source or sink components in the budget (see Table 2.2). Similar to the continental sink estimates reported in SOCCR1, the forest sector is among the largest sinks (217 Tg C per year), along with smaller but persistent sinks in agricultural soils (15 Tg C per year) and terrestrial wetlands (58 Tg C per year) in SOCCR2. To reiterate, additional small-sink components for Arctic and boreal ecosystems (14 Tg C per year) and tidal wetlands and estuaries (17 Tg C per year) in this report were not considered in SOCCR1. The most significant components now included in SOCCR2 are the net uptakes by inland waters (260 Tg C per year) and by coastal ocean areas (160 Tg C per year). However, a large sink component associated with woody encroachment (120 Tg C per year) was included in SOCCR1 but is not explicitly separated in SOCCR2 because this potential sink mechanism is considered to be included within the forest and grassland estimates. The flux estimates from inland waters, the coastal ocean, and woody encroachment remain highly uncertain and should be prioritized for further study, given their potentially large contribution to the continental carbon budget.

Confidence in estimates of the overall, continental­scale carbon budget is expected to increase in the near future with more observations, improved data, and better understanding of the processes. More accurate, consistent, and highly resolved estimates among the various budget components likely will be helpful in informing management-scale decisions (see Ch. 18: Carbon Cycle Science in Support of Decision Making). Though atmospheric measurements provide an integrated constraint on the overall budget and can detect variability and trends over short time frames, they currently offer limited attribution capability with respect to the various individual components. Bottom-up measurements and inventory estimates are needed to make projections for specific sectors and at the finer spatial scales at which the sectors are managed. These inventories, however, are often expensive and difficult to undertake. Moreover, they do not always obtain all the required measurements with consistent precision and, in many cases, cannot resolve key trends in sources and sinks or attribute their causes. Results from terrestrial biosphere model simulations offer the potential for ­process-based attribution of ­regional-scale carbon cycle dynamics (Turner et al., 2016b), but variability in response across the ensemble of model results leads to uncertainty in the predictions (Huntzinger et al., 2012, 2017). The move toward more regional-scale and sector-targeted atmospheric analyses should offer substantial help with these efforts, but advancements in bottom-up biosphere modeling frameworks will be necessary to improve confidence in future projections of the North American carbon budget (see Ch. 19: Future of the North American Carbon Cycle). These estimates also will continue to benefit from the increasing availability of remote-sensing data provided by multiple platforms (Goetz and Dubayah 2014; Masek et al., 2015; Williams et al., 2014). Although there is value in retaining independence among the various top-down and bottom-up approaches for estimating and comparing carbon fluxes, the most significant progress likely will be made by increasing the formal integration of these approaches in future assessment and prediction frameworks that are more comprehensive and consistent.

See Full Chapter & References