Abstract
Modelling tree architectural development according to environmental factors is an emerging field of<br />research thanks to the development of adequate computational methods. However, modelling virtual<br />plants is greatly constrained by the difficulty in collecting the architectural and geometric data of woody<br />architecture and foliage. Moreover, partial understanding of functional processes that drive shoot growth<br />limits the analysis and the modelling of the processes involved. Cross-fertilization of insights into<br />architectural and morphological variability of Walnut tree on the one hand, and into growth physiology on<br />the other hand, enables to better understand crown development. Thanks to 3-dimensional measurements<br />of topology and geometry made on 7 to 9 years-old Walnut trees, we have analysed the impact on light<br />interception of variables required for virtual plant modelling that are tedious to measure and that render<br />extrapolation methods necessary. Secondly, we quantified the part of variance of growth variables which<br />can be explained by radiation intercepted by comparison with other factors such as topological position.<br />Our results point out the limitations of using allometric relationships and statistical laws for the<br />reconstruction of 3D canopies. Leaf clumping inside the crown, including the clumping of leaflets at leaf<br />scale, strongly determines the radiative properties of reconstructed crowns. Thanks to an approach based<br />on correlations and statistics, an effect of light intercepted by mother shoots on the growth of daughter<br />shoots was established. Strong correlations between shoot vigour (quantified by an index) and the length<br />of the pathway between the collar and shoots on the one hand and shoot growth on the other hand<br />suggest an effect of hydraulic architecture.