Do we (need to) care about canopy radiation schemes in DGVMs? Caveats and potential impacts

Dynamic global vegetation models (DGVMs) are an essential part of current state-of-the-art Earth system models. In recent years, the complexity of DGVMs has increased by incorporating new important processes like, e. g., nutrient cycling and land cover dynamics, while biogeophysical processes like surface radiation have not been developed much further. Canopy radiation models are however very important for the estimation of absorption and reflected fluxes and are essential for a proper estimation of surface carbon, energy and water fluxes. The present study provides an overview of current implementations of canopy radiation schemes in a couple of state-of-the-art DGVMs and assesses their accuracy in simulating canopy absorption and reflection for a variety of different surface conditions. Systematic deviations in surface albedo and fractions of absorbed photosynthetic active radiation (faPAR) are identified and potential impacts are assessed. The results show clear deviations for both, absorbed and reflected, surface solar radiation fluxes. FaPAR is typically underestimated, which results in an underestimation of gross primary productivity (GPP) for the investigated cases. The deviation can be as large as 25% in extreme cases. Deviations in surface albedo range between -0.15 <= Delta alpha <= 0.36, with a slight positive bias on the order of Delta alpha approximate to 0.04. Potential radiative forcing caused by albedo deviations is estimated at -1.25 <= RF <= -0.8 (Wm(-2)), caused by neglect of the diurnal cycle of surface albedo. The present study is the first one that provides an assessment of canopy RT schemes in different currently used DGVMs together with an assessment of the potential impact of the identified deviations. The paper illustrates that there is a general need to improve the canopy radiation schemes in DGVMs and provides different perspectives for their improvement.