The Orbiting Carbon Observatory-2 (OCO-2) and in situ CO₂ data suggest a larger seasonal amplitude of the terrestrial carbon cycle compared to many dynamic global vegetation models

Existing, state-of-the-art vegetation models disagree by a factor of four on the seasonal amplitude of the global, terrestrial carbon cycle. This seasonal amplitude is likely increasing over time due to climate change, and disagreements among vegetation models therefore complicate efforts to quantify how climate change is impacting the carbon cycle. We evaluate the seasonal cycle of terrestrial CO₂ fluxes from an ensemble of vegetation models using CO₂ observations from the Orbiting Carbon Observatory-2 (OCO-2), in situ CO₂ observations, and inverse models. We find that vegetation models with a larger seasonal amplitude are also more sensitive to climate change, in that they exhibit a larger increase in amplitude during the past century. Furthermore, ten of the 17 models analyzed have a seasonal amplitude smaller than an ensemble of inverse CO₂ flux estimates based on OCO-2 observations; these discrepancies are largest across the Eastern US, boreal Asia, the Congo, and the Amazon. Vegetation models with larger seasonal amplitudes, when run through an atmospheric transport model (i.e. GEOS-Chem), typically exhibit a better fit compared to atmospheric CO₂ observations. We also find that vegetation models produce similar seasonal amplitudes of net CO₂ fluxes using very different combinations of gross primary production and respiration, making these model disagreements challenging to resolve.