SUNConferences, New Frontiers in Forecasting Forests 2018

Font Size: 
THE LINK BETWEEN WOOD PROPERTY VARIATION AND LUMBER STIFFNESS: THE EFFECT OF INITIAL SPACING
Justin Erasmus, Brand Wessels, David Drew

##manager.scheduler.building##: Wallenburg Research Centre (STIAS)
##manager.scheduler.room##: Main auditorium
Date: 2018-09-27 01:55 PM – 02:20 PM
Last modified: 2018-05-11

Abstract


One of the most important strategic decisions forest managers make during the plantation establishment phase is the appropriate choice of spacing. The resulting growing space available to trees is one of the major factors limiting the full realization of the genetic potential of a tree. Initial spacing is also the most powerful management tool available to influence the raw material properties of wood and their within-tree variability, which in turn are major determinants of lumber performance. The management of fast growing plantations has generally focused on wood volume production with less concern for the nature and variability of solid wood properties, thereby inadvertently reducing the mechanical properties, particularly the stiffness, of recovered products. Pronounced changes in the stiffness, referred to as the modulus of elasticity (MOE), of juvenile wood,  can be largely accounted for by variations in basic wood properties: wood density, microfibril angle (the angle of microfibrils with respect to the vertical axis of the cell, MFA), ring width and knot properties. Therefore, the objective of this study is to examine the effects of changes in growth rate of Pinus patula, as mediated through initial spacing, on the within-tree variation of microfibril angle (MFA) and density. We also aimed to evaluate the effect of the resulting variation of density, MFA and knot properties on lumber performance – particularly MOE.

The study makes use of data on wood and lumber properties of 45 trees from an 20-year-old Pinus patula spacing trial located within the Mpumalanga escarpment (25.7667° S, 31.2392° E), which includes a wide range of tree spacing (1.83 x 1.83 m, 2.35 x 2.35 m, 3.02 x 3.02 m and 4.98 x 4.98 m). A 2.4 m saw log was removed above a breast height disc for all trees before being processed into 208 38x114 mm boards using a cant sawing pattern. The pith-to-bark properties of density, MFA and ring width of individual trees and year rings was measured at DBH level by the Silviscan 3 apparatus of CSIRO in Australia. Modified logistic and exponential non-linear mixed effects models, with trees as random effects, were used to evaluate the effect of initial spacing on the within-tree variation (pith value, radial rate of change and an outerwood value) of MFA and wood density as a function of cambial age. Ring width was subsequently also included in the predictions of ring-level properties in order to validate whether growth rate is sufficient in accounting for differences in tree spacing. Average MFA and density properties were also linked to lumber MOE through the corresponding annual rings present in the board of the same tree from which it was processed. Lumber MOE was then regressed against multiple wood properties, including knot characteristics, using a linear mixed effects model.

Preliminary results indicate that MOE improves significantly with reduced spacing. MFA, density and the number of knots per board proved to be the best predictors of lumber MOE (r2 = 0.73, and 0.82 with random effects) with MFA and density having the highest explanatory power as expected. The only significant difference in the radial variation of both MFA and density between spacing treatments was the radial rate of change. Radial gradients of MFA and density were significantly lower for the 4.9 x 4.9 m treatment.  Ring width contributed significantly to the MFA model only; however, there remained significant differences in the rate parameter. These results suggest that growth rate alone may have limited use as proxy for spacing in terms of its effect on MFA and is not a significant causal factor in density variation between spacing treatments. The annual Slenderness of trees were also made available and, in subsequent analysis, will be included in predictions and evaluated as a mechanical casual factor. Given the low stiffness of SA pine lumber associated with reduced harvesting ages and the high variability of basic wood properties in juvenile wood, the results of this study aids in understanding the key drivers of lumber stiffness and how changes in these properties caused by changes in growth, influences end product performance. An accurate model of MOE will enable wood processors to predict mechanical properties of trees from individual sites and plan their production accordingly. It will ultimately assist the sawmill industry to process acceptable structural products from SA pine, specifically Pinus patula – a species which dominates the structural lumber market. The final results will be presented after further analysis.


An account with this site is required in order to view papers. Click here to create an account.