Comparing global models of terrestrial net primary productivity (NPP): The importance of water availability to primary productivity

Given that neither absolute measures nor direct model validations of global terrestrial net primary productivity (NPP) are feasible, intercomparison of global NPP models provides an effective tool to check model consistency. For this study, we tested the assumption that water availability is the primary limiting factor of NPP in global terrestrial biospheric models. We compared a water balance coefficient (WBC), calculated as the difference of mean annual precipitation and potential evapotranspiration to NPP for each grid cell (0.5° × 0.5° longitude/latitude) in each of 14 models. We also evaluated different approaches used for introducing water budget limitations on NPP: (1) direct physiological control on evapotranspiration through canopy conductance; (2) climatological computation of constraints from supply/demand for ecosystem productivity; and (3) water limitation inferred from satellite data alone. Plots of NPP vs. WBC showed comparable patterns for the models using the same method for water balance limitation on NPP. While correlation plots revealed similar patterns for most global models, other environmental controls on NPP introduced substantial variability.