Modelling the boards properties distribution for a forest stand. Analysis of the propagation of errors in the simulation process. Application in Norway Spruce (Picea abies Karst.)

The need for management tools that integrate both growth and wood quality information led to the generation of detailed models which have mostly been developed for fast-growing coniferous plantations. These models that aim at assessing the timber quality of standing trees from typical inventory measurements (such as tree age, height and diameter at breast height of individual trees) are used to reconstruct the past growth of the trees and to predict their current internal structure, namely the ring width distribution. Both models are linked to aIlometric equations that estimate the characteristics of crown and branchiness, and to densitometric models that predict wood density from ring distribution. These models are integrated in a software called Win-EPIFN that simulates also the sawing and the grading of any board located in a stem for which morphology is known in detail. A previous study showed that this software was able to evaluate the wood properties values of an existing forest resource at one regional level. Our aim is now to improve the predictive qualities of these models and to be able to perform a direct comparison of the measured and simulated properties of boards for a stand. (Resume d'auteur)