Bayesian modeling of Douglas fir quality to predict the resource of extractable chemical

Abstract

Trees are hierarchical system including stem, branches and foliage. These compartments are linked together by their primary growth and secondary growth. Knottiness is the main factor at tree scale for wood quality. It has high extractive content and contribute to assess the stem economic value. However, the current modelling of this wood propriety faces several methodological difficulties. Using a homogenized database from 5 independent studies, a hierarchical Bayesian structural approach of the growth-branchiness- knottiness, was developed. Data from an automatic knottiness quantification algorithm developed during this study were added. The results validated the effectiveness of a structural quality growth approach within a bayesian hierarchical framework. They form a preliminary set of a complete model verifying the allometric equations obtained in the literature and allowing the simulation of quality according to silvicultural treatments.