Towards better-constrained assessments of the carbon balance of North America in the 21st Century: A comparison of recent model and inventory-based estimates

Background/Question/Methods

The North American carbon sink is generally considered to account for a large, but highly uncertain, portion of the northern, extra-tropical land-based sink, with estimates ranging from 15% to 100%. This uncertainty is owing to a number of sources, including the inherent differences in the approaches commonly used to develop estimates of land-atmosphere CO₂ exchange. Here, we examine estimates of carbon stocks and flux based on recent national forest and agricultural inventories for reporting zones in Canada, the U.S. and Mexico. We offer a novel perspective on using this measurement- and statistical- based information on carbon stocks to track both vertical (atmospheric sources and sinks) and lateral (transfers among reporting zones and sectors) fluxes of carbon at the continental scale. In addition, flux estimates derived from both bottom-up, terrestrial biosphere models (TBMs) and top-down, atmospheric inverse approaches have been collected for the Interim Synthesis activity under the North American Carbon Program, and these estimates have been processed to allow comparison at the spatial and temporal scales of the inventories. We then analyze the results of these different approaches in identifying sources, sinks and transfers of carbon across the North American continent.

Results/Conclusions

Analysis of the inventory-based data sets shows a net land-based sink of atmospheric CO₂ over the North American continent with a magnitude of 282 TgC yr^-1 between 2000 and 2006. In contrast, mean estimates from the ensembles of TBM and inverse models suggest a substantially larger continental sink over this time period (511 and 931 TgC yr^-1, respectively). The inventory-based data indicate that the overall sink is driven by carbon uptake through forest and crop productivity in the U.S., which is tempered by carbon releases from recent forest disturbances in Canada and land use change in Mexico. Although the magnitudes of the estimates are typically larger, the models generally agree with the inventory-based information on the relative patterns of the forest and crop land sinks across the reporting zones. The major source of disagreement between the approaches, however, relates to the fate of carbon in harvested forest and agricultural products. The inventory-based data indicate CO₂ sources from the respiration and decay of crop and forest products, which may not be accounted for in the modeling approaches.  We discuss other sources of uncertainty in each approach that should be prioritized for on-going assessment of the North American carbon balance.