Abstract
Nine trait-based hypotheses (including static PFT values, empirical leaf nutrient relationships, empirical relationships to environment, and theoretical 'plant-centric' methods) are used to predict global Vcmax,25 distributions and we assess their impact on simulations of global gross primary productivity (GPP) in the Sheffield Dynamic Global Vegetation Model.
Global GPP varied from 104.8 PgC yr-1 to 136.9 PgC yr-1 with a coefficient of variation 10% across hypotheses, the variation in GPP propagated through to c. 35% variation in net biome productivity (NBP).
The most data-driven hypothesis (empirical relationships to environment) resulted in relative global GPP distributions that showed the strongest bias compared with relative distributions of solar induced fluorescence, suggesting a possible decoupling of GPP and Vcmax,25 at high latitudes.
These results demonstrate the uncertainty in both the geospatial distribution and long term trends in GPP that arises from various trait-based hypotheses of Vcmax,25. Plant-centric hypotheses resulted in shallower trends in global GPP 1901-2012 because the co-ordination hypothesis restricts CO2 responses of GPP to the CO2 response of light-triggered electron transport.
