New Phytol. http://doi.org/b9k3 (2017)

A major source of uncertainty in global gross primary production (GPP) modelling, and associated carbon-cycle dynamics, is the calculation of maximum photosynthetic carboxylation rate (Vcmax), one of two plant traits that closely determines photosynthetic rate. Various methods are used in terrestrial biosphere models to calculate these traits, each representing a different theory about how these traits scale, but the resultant errors have not previously been quantified.

Anthony P. Walker at Oak Ridge National Laboratory, USA, and co-workers investigate the impact of four trait-scaling hypotheses (plant functional type, nutrient limitation, environmental filtering, and plant plasticity) for Vcmax simulation and their impact on global patterns of GPP.

Modelled global GPP ranged between 108–128 PgC yr−1, representing around 65% of the uncertainty range found in GPP model intercomparison exercises. This uncertainty propagated through to a 27% coefficient of variation in net biome productivity. Encouragingly, all hypotheses produced global GPP estimates that were highly correlated with proxies of global GPP.