Abstract
AbstractDiffuse radiation can increase the light use efficiency for plant photosynthesis. However, the lack of observations limits the explorations of diffuse fertilization effects on the global scale. Here, we bias‐correct global hourly diffuse fraction (Kd) from a climate reanalysis using an artificial neural network (ANN) model in combination with site‐level observations. Evaluations at independent sites show that the updated Kd on average increases correlations by 9.7% and reduces root‐mean‐square errors (RMSEs) by 45.5% against measurements than the original reanalysis. The derived radiative fluxes are then used as input for a dynamic vegetation model to explore the impacts of diffuse radiation on global gross primary productivity (GPP) during 1981–2015. With the updated Kd, simulated GPP shows lower RMSEs against observations at 72 out of 76 FLUXNET sites, leading to a reduced RMSE of 9.9% on average. Moreover, simulations with updated Kd present higher global GPP (3.1%) than that with original Kd from Modern‐Era Retrospective Analysis for Research and Applications (MERRA) product. The simulations show that diffuse radiation, which accounts for 54% of the total shortwave radiation, contributes to long‐term global mean GPP by 1.49 g C m−2 day−1 (64.3%). During 1981–2015, the changes of direct radiation result in a global GPP trend of −0.49 g C m−2 yr−2, which is offset by 0.31 g C m−2 yr−2 (53.1%) following the enhancement of diffuse radiation.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call