Key Finding 1
Global concentrations of carbon dioxide (CO2) and methane (CH4) have increased almost linearly since the First State of the Carbon Cycle Report (CCSP 2007; see Figure 8.1). Over the period 2004 to 2013, global growth rates estimated from the National Oceanic and Atmospheric Administration’s (NOAA) marine boundary layer network average 2.0 ± 0.1 parts per million (ppm) per year for CO2 and 3.8 ± 0.5 parts per billion (ppb) per year for CH4. Global mean CO2 abundance as of 2013 was 395 ppm (compared to preindustrial levels of about 280 ppm), and CH4 stands at more than 1,810 ppb (compared to preindustrial levels of about 720 ppb) (very high confidence).
Description of evidence base
Global mean atmospheric growth rates and abundances of CO2 and CH4 are derived from publicly available tables on NOAA websites: 1) www.esrl.noaa.gov/gmd/ccgg/trends/global.html and 2) www.esrl.noaa.gov/gmd/ccgg/trends_ch4/.
Major uncertainties
The averages were calculated from the regularly updated marine boundary layer sites of NOAA’s Global Greenhouse Gas Reference Network. These averages are not associated with any recent literature. The methodology used to construct the global “surfaces” from which the global averages are computed is described in Masarie and Tans (1995). The uncertainties originate primarily from the incomplete sampling of the marine boundary layer by the NOAA network and the uncertainty associated with smoothing the raw data prior to creating the global surface. Measurement uncertainty of CO2 and CH4 is a minor component. Uncertainty calculations are described in detail at: www.esrl.noaa.gov/gmd/ccgg/mbl/mbl.html. While the atmospheric CO2 growth rate is relatively stable, there is strong decadal and interannual variability of CH4 emissions, making computation of an average inherently sensitive to the choice of time period. For instance, the CH4 growth rate averaged over 1997 to 2006 was 2.8 ppb per year, whereas over 2007 to 2015, it was instead 7.0 ppb per year.
Assessment of confidence based on evidence and agreement, including short description of nature of evidence and level of agreement
NOAA data are the gold standard for determining global growth rates and abundances because of extensive global coverage and high internal network compatibility, including high measurement precision. The trends and growth rates also agree well with estimates from other laboratories.
Summary sentence or paragraph that integrates the above information
NOAA CO2 and CH4 trends and abundances are publicly available, fully traceable, and represent the most comprehensive description of global CO2 and CH4.
Key Finding 2
Inverse model analyses of atmospheric CO2 data suggest substantial interannual variability in net carbon uptake over North America. Over the period 2004 to 2013, North American fossil fuel emissions from inventories average 1,774 ± 24 teragrams of carbon (Tg C) per year, partially offset by the land carbon sink of 699 ± 82 Tg C year. Additionally, inversion models suggest a trend toward an increasing sink during the period 2004 to 2013. These results contrast with the U.S. land sink estimates reported to the United Nations Framework Convention on Climate Change, which are smaller and show very little trend or interannual variability.
Description of evidence base
Fossil fuel emissions are from Carbon Dioxide Information Analysis Center (CDIAC) estimates (available from the U.S. Department of Energy’s Environmental Systems Science Data Infrastructure for a Virtual Ecosystem [ESS-DIVE] data archive, ess-dive.lbl.gov). The land carbon sink is based on the 10-year average of North American annual fluxes from four global inverse models, specified in the text. The error reported is twice the standard error of the mean of the 10 years and for the four models and mostly represents the amount of interannual variability. The evidence for a trend is based on a linear least-squares regression. The comparison of variability with the U.S. Environmental Protection Agency’s (EPA) estimate of the U.S. land sink is based on EPA data accessed at www.epa.gov/ghgemissions/inventory-us-greenhouse-gas-emissions-and-sinks-1990-2015.
Major uncertainties
Fossil fuel emissions uncertainty is very low (see Appendix E: Fossil Fuel Emissions Estimates for North America). Long-term means of CO2 sources and sinks derived from a given inverse model are highly uncertain. However, the interannual variability of fluxes from different models tends to agree well, suggesting lower uncertainty. EPA land flux estimates may not exhibit enough variability due to the U.S. Forest Service methodology, upon which EPA’s estimates are largely based.
Assessment of confidence based on evidence and agreement, including short description of nature of evidence and level of agreement
Fossil fuel uncertainty at the national, annual scale has the smallest uncertainty because it can be constrained by highly accurate information on imports and exports and internal usage. Inverse model-based estimates of CO2 sources and sinks contain numerous random and systematic errors including biases associated with wind fields and parameterization of vertical mixing. Because models exhibit different mean atmospheric transport, their long-term average fluxes can differ significantly. However, the interannual variability of fluxes among inverse models is much more similar, meaning that the difference between the inverse model and EPA flux variability is likely to be robust.
Estimated likelihood of impact or consequence, including short description of basis of estimate
The contrast between variability exhibited in the inverse model and the EPA estimates of land sink variability could cause EPA to re-examine its methodologies. Additionally, the emerging evidence that the North American CO2 sink is growing also could spur research in the “bottom-up” community and impact policy decisions.
Summary sentence or paragraph that integrates the above information
Regularly produced inverse modeling estimates of CO2 sources and sinks over North America are beginning to provide valuable information at least on interannual variability of terrestrial ecosystem fluxes.
Key Finding 3
During most of the study period covered by the Second State of the Carbon Cycle Report (2004 to 2012), inverse model analyses of atmospheric CH4 data show minimal interannual variability in emissions and no robust evidence of trends in either temperate or boreal regions. The absence of a trend in North American CH4 emissions contrasts starkly with global emissions, which show significant growth since 2007. Methane emissions for North America over the period 2004 to 2009 estimated from six inverse models average 66 ± 2 Tg CH4 per year. Over the same period, EPA-reported CH4 emissions equate to a climate impact of 13% of CO2 emissions, given a 100-year time horizon.
Description of evidence base
The conclusions of minimal interannual variability (standard deviation), trend (slope and its uncertainty), and mean flux are all based on fluxes from 14 inverse models used in the global CH4 budget analysis of the Global Carbon Project (Saunois et al., 2016). The 13% ratio of CH4 to CO2 warming impact is based on EPA CH4 and CO2 emission estimates using a 100-year global warming potential (GWP) value of 28.
Major uncertainties
Total CH4 emissions for North America include the inversely derived value of 60 Tg CH4 per year and the EPA anthropogenic emissions estimate for the United States, which would impact the 13% ratio. Inverse models are subject to poorly known uncertainties stemming from the use of biased priors, imperfect models of atmospheric transport, and the sparse network of in situ measurements.
Assessment of confidence based on evidence and agreement, including short description of nature of evidence and level of agreement
Total emissions have a high uncertainty (not reflected in the variability value stated in the Key Finding); note that EPA does not provide an uncertainty for its estimate. The absence of any trend has higher confidence, because numerous models with different methodologies contributed to this finding. However, the models used in the comparison did not uniformly cover the 2000 to 2013 period, making the conclusion less robust than that for CO2. On the other hand, the smaller variability relative to CO2 is consistent across models and is more robust. The 13% value is uncertain because of EPA’s CH4 emissions estimate and, to a lesser extent, the GWP uncertainty.
Estimated likelihood of impact or consequence, including short description of basis of estimate
The finding that CH4 is unlikely to have a temperate North American trend different from zero is significant, because there is great interest in the cumulative radiative forcing impact of CH4 emissions from the oil and gas sector. Moreover, while not a new finding, the simple calculation of CH4 having only 13% of the warming impact as CO2 should remind policymakers and scientists that CO2 emissions are substantially more important.
Summary sentence or paragraph that integrates the above information
The global and North American emissions were derived using atmospheric CH4 data assimilated in a wide variety of CH4 inverse models using both in situ and remote-sensing data. Although a consistent picture is emerging, the results are more uncertain than those for CO2, because estimates are not produced regularly over consistent timescales.
