Sensitivity of simulated light interception and tree transpiration to the level of detail of 3D tree reconstructions

International audience

إنجليزي. The aim of this paper is to investigate the minimal level of detail of the tree geometry reconstruction required to enable an accurate estimate of the evaporative cooling effect of an individual tree. The Functional Structural Plant Modeling approach, which links the 3D tree structure to its functioning, is used to investigate the impact of the level of detail of a tree reconstruction on its simulated transpiration. Based on terrestrial laser scanning point cloud data of a nine-meter-high silver linden tree (Tilia tomentosa Moench), several methods of reconstruction of its crown were considered. They can be divided into three groups: (i) tree branching structure reconstructions where leafy shoots are reconstructed; (ii) envelope reconstructions such as 3D convex/concave envelopes; and (iii) voxel reconstructions where leaves are uniformly distributed within the given volume. Based on these methods, several mock-ups of resolved tree crowns from low to high level of details were created. The cooling performance of each tree mock-up was compared by simulating its transpiration rate with a validated 3D ecophysiological model based on the turbid medium approach. The resulting mock-ups differ in several characteristics related to tree metrics and light interception: volume, projected leaf area, leaf area index, leaf area density, leaf clumping and silhouette-to-total area ratio. An inter-comparison of the transpiration rates provided by these mock-ups shows that tree branching structure reconstruction methods perform better than envelope computation methods and that these differences are related to the light interception property of the reconstructed trees. This study provides guidelines to determine which structural characteristics of tree crowns must be measured and taken into account in order to carry out accurate estimates of the transpiration rates of individual trees.