The impact of forest canopy structure on modelled photosynthesis

Despite terrestrial vegetation being one of the largest carbon sinks, it’s representation within weather and climate models is simple due to computational and data constraints. Whilst the representation of many processes has been improved, the absorption of light by vegetation canopies is still described using the assumption of a plane-parallel turbid medium, which is not realistic. This approach assumes randomly distributed leaves, with no horizontal or vertical variability, and a flat canopy base. However, it is widely used as it permits the radiative transfer for vegetation to be described using a two-stream model (Sellers, 1985) that can be solved analytically. This work examines the importance of including realistic vegetation structure in photosynthesis calculations, testing the sensitivity of modelled Gross Primary Productivity (GPP) to the common turbid-media approximation by using detailed forest canopy information. We derive a methodology for calculating GPP from radiative transfer calculations from a high-resolution, computationally demanding, radiative transfer model, DART (Discrete Anisotropic Radiative Transfer). The GPP calculation  is based upon the photosynthesis scheme from the JULES (Joint UK Land Environment Simulator) land surface model. We explore the impacts of structure on GPP for six real forest canopies, from across the globe using 3D vegetation data collected using Terrestrial Lidar Scanning (TLS). The use of a 3D radiative transfer model allows us to investigate how much difference the two-stream approach introduces compared to detailed forest canopies with different levels of structure. We examine the profiles of both the absorbed radiation (fAPAR) and GPP. Across the six forest scenes used, there is generally a reduction in GPP as more structure is introduced. However, this is particularly the case in scenes where the horizontal variation of LAI is high. Additionally, in these scenes, we find that this increase in GPP is less pronounced at low sun angles. This work suggests that consideration should be taken in incorporating horizontal variability of the vegetation within weather and climate models, particularly when identifying the effects of forests on the global carbon budget.