Including variation in branch and tree properties improves timber grade estimates in Scots pine stands

• Published in: Canadian journal of forest research Vol. 48; no. 5; pp. 542 - 553
• Main Authors: Ojansuu, Risto, Mäkinen, Harri, Heinonen, Jaakko
• Format: Journal Article
• Language: English
• Published: Ottawa NRC Research Press 2018; Canadian Science Publishing NRC Research Press
• Subjects: branchaison; Branches; branchiness; Chain branching; Computer simulation; Forest management; Mathematical analysis; Mathematical models; Monte Carlo simulation; pin sylvestre; Pine trees; Pinus sylvestris; Quality; qualité du bois d’œuvre; random variation; Scots pine; simulation; Timber; timber quality; Variation; variation aléatoire
• ISSN: 0045-5067; 1208-6037
• DOI: 10.1139/cjfr-2017-0435

Deterministic modelling of roundwood quality with expected values of quality factors underestimates the variation in timber grade distribution and leads to sudden transitions of all of the trees in a size class between the quality classes. We constructed a recursive model chain that predicts the height of the lowest living and dead branches for a set of maximum branch diameters based on stem diameter, tree height, and height of the crown base. By using Monte-Carlo simulation, a timber grade distribution for a stand was integrated over the multidimensional random distribution of the recursive model chain. The results demonstrated that introducing random variation resulted in major changes in the calculated quality distributions. The timber grade distributions were wider compared with the expected value-based predictions. In particular, the proportions of the highest and lowest quality classes were under- or over-estimated in the deterministic predictions. Because of the random variation of tree properties, the timber grade distribution also became more continuous over time. The results also showed that the timber grade distributions can vary considerably between stands with similar stem diameter distributions. Understanding the effects of random quality variation will give guidance to forest managers to reach more profitable management regimes.