- Second State of the Carbon Cycle Report (SOCCR2): A Sustained Assessment Report
- Featured Publications
- All Publications
finding 2.5 : key-message-2-5
The magnitude of the continental carbon sink over the last decade is estimated at 699 Tg C per year (±12%) using a top-down approach and 606 Tg C per year (±75%) using a bottom-up approach, indicating an apparent agreement between the two estimates considering their uncertainty ranges.*
*Note: Confidence level excluded due to Key Finding’s emphasis on methodological comparisons.
This finding is from chapter 2 of Second State of the Carbon Cycle Report (SOCCR2): A Sustained Assessment Report.
Description of evidence base: The integrated, continental-scale estimates of the overall carbon sink comprise compilations from 1) recent top-down, atmospheric approaches (see Ch. 8: Observations of Atmospheric Carbon Dioxide and Methane); 2) comparisons of bottom-up, inventory-, and model-based estimates from the various sector-focused chapters in this report; and 3) data and estimates synthesized in Table 2.2 and Figure 2.3 and discussed in the context of the results from previous continental carbon cycle synthesis efforts (e.g., CCSP 2007; Hayes et al., 2012; King et al., 2015).
New information and remaining uncertainties: The bottom-up estimate of the overall continental-scale carbon sink presented here is inferred from reconciling a number of estimates from different components, themselves often highly uncertain. Even components estimated in formal inventories (e.g., the forest sector) have pools and fluxes that are less well quantified (e.g., forest soils) and regional and temporal gaps in measurements. A large component of the uncertainty stems from limited information about the magnitude, spatial distribution, and temporal variability of carbon sources and sinks in inland, tidal, and coastal waters. Uncertainty in the top-down, atmospheric-based estimates is primarily from sparse observational networks and often poorly constrained models of atmospheric transport.
ProvenanceThis finding was derived from figure P.2: P.2. Likelihood and Confidence Evaluation
Alternatives : JSON YAML Turtle N-Triples JSON Triples RDF+XML RDF+JSON Graphviz SVG