Forests dominate the interannual variability of the North American carbon sink

Y.P. Shiga, A.M. Michalak, Y. Fang, K. SCHAEFER, A.E. ANDREWS, D.H. HUNTZINGER, C.R. SCHWALM, K. THONING and Y. WEI


SUMMARY

One of the most fundamental questions in climate science is the extent to which the world’s plants and soils will continue to take up a portion of the carbon emitted into the atmospheric through human activity. In this study, we use atmospheric observations of carbon dioxide to identify forests as the dominant region controlling how carbon uptake varies year-to-year for North America. Understanding how this carbon sinks responds to variability in temperature, precipitation, and other environmental factors helps us to understand how this sink will behave under future climate conditions.

Figure: Deciduous & mixed forests are the dominant contributor to North American NEE IAV based on atmospheric observations (diamonds), both in terms of the percent contribution (40%) and relative to the biome’s area (circles). TBMs (box-plots) show croplands (30%) as the largest contributor, although the spread among TBMs is too large to identify a dominant biome. For TBMs, boxes show interquartile range and median of the ten TBMs, whiskers represent TBMs within 1.5 times the interquartile range of either box edge, and plus signs represent outliers. For the atmospheric inversion results, the diamond represents the contribution estimate over the full time domain while the solid bar represents the range based on leave-one-year-out bootstrap estimates. The corresponding biome map is show in panel (b).


ABSTRACT

Understanding what drives the interannual variability (IAV) of the land carbon sink is crucial for improving future predictions of this important, yet uncertain, component of the climate system. While drivers of global and hemispheric-scale net ecosystem exchange (NEE) IAV have been investigated, our understanding of the drivers of NEE IAV at regional scales (e.g. sub-continental, biome-level) is quite poor. Here we explore the biome-level attribution and drivers of North American NEE using inverse estimates derived from a dense network of atmospheric CO2 observations. We find that deciduous broadleaf and mixed forests are the primary regions responsible for North American NEE IAV, which differs from the ecoregions identified for the globe and Northern Hemisphere. We also find that a suite of terrestrial biosphere models (TBMs) do not agree on the dominant biome contributing to NEE IAV, with TBMs falling along an apparent spectrum ranging between those with IAV dominated primarily by forested ecosystems to those with IAV dominated by non-forested ecosystems. Furthermore, this regional trade-off in TBM NEE IAV is found to be linked to differing regional responses to environmental drivers among TBMs. This work displays the importance of extra-tropical forests in driving continental NEE IAV and also highlights the challenges and limitations of using TBMs to inform regional-scale carbon flux dynamics.

Shiga, Y.P., A.M. Michalak, Y. Fang, K. Schaefer, A.E. Andrews, D.H. Huntzinger, C.R. Schwalm, K. Thoning, Y. Wei (2018) "Forests dominate the interannual variability of the North American carbon sink," Environmental Research Letters, 13 (8), 084015, doi:10.1088/1748-9326/aad505.