Separating the effects of phenology and diffuse radiation on gross primary productivity in winter wheat

Gross primary productivity (GPP) has been reported to increase with the fraction of diffuse solar radiation, for a given total irradiance. We applied several approaches to observationally distinguish between diffuse radiation effects and other known influences on GPP, using 8 years of eddy-covariance measurements of winter wheat in the U.S. Southern Great Plains. The apparent enhancement of daily GPP due to diffuse radiation was reduced from 260% to at most 75%, either by subsampling over the peak growing season or by subtracting a 15-day moving average of GPP, suggesting a role of phenology. The diffuse radiation effect was further reduced to a statistically insignificant 22% after normalizing GPP by a spectral reflectance index to account for phenological variations in canopy photosynthetic capacity (Vcmax seasonality). Canopy photosynthetic capacity covaries with diffuse fraction per irradiance because both factors are dependent on the solar zenith angle on seasonal timescales. Using a two-leaf sun-shaded canopy radiative transfer model, we confirmed that the effects of phenological variations in photosynthetic capacity can appear qualitatively similar to the effects of diffuse radiation on GPP, and therefore can be difficult to distinguish using observations. The results suggest that two-leaf and multilayer canopy models may overestimate the effects of diffuse radiation on GPP if the leaf-to-canopy scaling of photosynthetic capacity is too far from optimal. The importance of controlling for phenology when inferring diffuse radiation effects on GPP raises new challenges and opportunities for using radiation measurements to improve carbon cycle models.