Estimation of crown volume for a geometric radiation model from detailed measurements of tree structure

A stochastic statistical reconstruction model of gray alder (Alnus incana) was used to describe the architecture of a 5-year-old stand in Tõravere, Estonia. Using the modelled distribution of foliage, crown volume for alder trees was calculated. As the distribution of foliage in alder crowns was very inhomogeneous and leaf area density decreased rapidly with distance from the stem, the crown threshold α was defined as the fraction of foliage contained in the crown volume. Using different values of α, crowns of different shapes and sizes were generated. Crowns modelled as ellipsoids were used to calculate canopy transmittance; calculated transmittances were compared to those measured with a LAI-2000 canopy analyzer. Division of foliage into tree crowns increased accuracy of transmission predictions compared to those obtained with a modelled horizontally random canopy, but results depended on the chosen crown threshold. Best predictions for canopy transmittance were achieved with α = 0.95. Clumping index was also used to evaluate the performance of models; the best model with α = 0.95 produced clumping index values close to 1. Despite the clear influence of canopy geometric structure on transmission of shortwave radiation, the effect of leaf inclination on radiation penetration was found to be several times larger.