Abstract: This paper examined the impact of commercial thinning on selected wood properties of jack pine (Pinus banksiana Lamb.). Wood properties evaluated include wood density (ring density, earlywood ring density, and latewood ring density), percentage of latewood in the ring, and wood bending properties. Nonlinear, mixed-effect models have been developed using data from three commercially thinned sites in eastern Canada. Ring density followed the same pattern as percentage of latewood, in which cambial age, relative height, and ring width were found to have important effects. Earlywood and latewood ring densities changed within the juvenile wood zone until a plateau was reached. Ring width affected earlywood and latewood ring densities mainly in narrow rings. Wood bending stiffness (measured by modulus of elasticity) and strength (measured by modulus of rupture) increased with cambial age and wood density; whereas, wood strength was also affected by ring width. Commercial thinning did not influence the developed models, but it had an indirect effect through increased ring width.
Resume : Cette etude a examine l'impact de l'eclaircie commerciale sur certaines proprietes du bois de pin gris (Pinus banksiana Lamb.). Ces proprietes incluaient la densite du bois (densite des cernes, densite du bois initial et densite du bois final), la proportion de bois final dans chaque cerne et les proprietes mecaniques en flexion. Des modeles non lineaires mixtes ont ete developpes a partir de donnees provenant de trois stations qui avaient subi une eclaircie commerciale dans fest du Canada. La densite des cernes s'est revelee etre fortement influencee par l'age cambial, la hauteur relative et la largeur des cernes, suivant ainsi un patron de variation similaire a celui de la proportion de bois final. La densite du bois initial et celle du bois final ont change a l'interieur de la zone de bois juvenile pour atteindre un plateau par la suite. La largeur des cernes a aussi affecte ces densites, en particulier dans le cas des cernes etroits. Les modulus d'elasticite (MOE) et de rupture (MOR) en flexion ont augmente avec l' age cambial et la densite du bois. De plus, le module de rupture a ete affecte par la largeur des cernes. L'eclaircie commerciale n'a pas change les parametres des modeles, mais elle a eu une influence indirecte a cause de son effet sur la largeur des cernes.
[Traduit par la Redaction]
Introduction
The Canadian forest industry has primarily been a volume-based and export-oriented commodity. In the last decade, wood products from fast-growing species have increased their share of the market, making it harder for producers such as Canada to compete on the global market. Moreover, the forest industry in Quebec is facing a 20% reduction of the annual allowable cut in public forests for boreal softwood lumber (Commission d'etude sur la gestion de la foret publique quebecoise 2004). In Ontario, the protected forest area (where forest harvesting activities are no longer allowed) has been increased to 12% in recent years under the Living Legacy Program. The reduced land base for fibre production coupled with increasing global competition has led the forest industry to enhance forest productivity and value through intensive forest management. The impacts of certain silvicultural practices have already been studied for jack pine (Pinus banksiana Lamb.) (Bell et al. 1990; Tong and Zhang 2005), which is the second most important commercial species in eastern Canada (OMNR 2004; Parent 2004). Jack pine was found to respond positively to intensive silviculture (Tong et al. 2004; Tong and Zhang 2005; Zhang et al. 2005). These studies showed that many of the wood quality attributes of jack pine are sensitive to spacing at an early age.
Very few studies have reported the effect of commercial thinning on the wood quality attributes of jack pine (Barbour et al. 1994). Thinning was found to reduce ring density (rd) by increasing the percentage of earlywood in the ring, but because of the limited range of data, the authors cautioned against generalizing the results. Scott et al. (1982) found that thinning slightly increased the wood density of jack pine, which they attributed to a site-specific effect. In loblolly pine (Pinus taeda L.), commercial thinning did not influence either the percentage of latewood (pctlw) in the ring or the rd (Tasissa and Burkhart 1998). The same results were observed for Scotts pine (Pinus sylvestris L.) in Sweden (Morling 2002). However, little research has investigated other properties, including wood mechanical properties such as strength and stiffness.
Recent work on wood density has focused on developing linear mixed models to incorporate inter- and intea-tree variations in wood density (Tasissa and Burkhart 1998; Guilley and Nepveu 2003; Guilley et al. 2003; Bouriaud et al. 2004) and wood strength and stiffness (Alteyrac et al. 2006). These models are limited to linear or linearizable functions, which often do not reflect the complex trends observed in the data. Advances in statistical methods enable the use of nonlinear models to fit hierarchal data that are used in wood quality research, such as rd (Pinheiro and Bates 2000; Mutz et al. 2004; Jordan et al. 2005; Fortin et al. 2006).
The influence of silvicultural practices on wood fibre quality and attributes will become more important to resource managers and policy makers as the Canadian forest products industry shifts towards intensive forest management and value-added manufacturing. This study provides information on the potential influence of commercial thinning on wood fibre attributes from long-term studies in unmanaged natural jack pine stands. Specifically, this paper quantifies the relationships of selected wood properties (rd, earlywood ring density (ewrd), latewood ring density (lwrd), and wood bending strength and stiffness) with ring width (rw), cambial age (hereinafter referred to as age), and relative height using a statistical modeling approach. The models developed are then used to investigate the effects of commercial thinning on the wood quality attributes studied (i.e., density and bending properties).
Material and methods
Site description, field work, and laboratory work
Stem analysis was carried out in three old, commercially thinned stands (Table 1). The Grand Lake Road site in New Brunswick had three thinning intensities, each with three plots: 0%, 25%, and 50% basal area removed. In the Quebec stands two thinning intensities were available, with one plot per thinning intensity: 0% and 20% or 22.6% basal area removed, for the Abitibi and Lac Saint-Jean sites, respectively. All thinnings were carried out from below, that is, small-diameter and (or) defective trees were removed.
The felled stems were chosen randomly in both the control and thinned plots (five stems per plot in the Grand Lake Road site (45 stems in total), 20 stems per plot in the Quebec sites (80 stems in total); Table 2). Three stems (two from the Grand Lake Road site, one from the Lac St-Jean site) were not analyzed because of manipulation errors in the laboratory. Disks were obtained at 0.2, 0.7, 1.3, and 2.0 m and every metre thereafter. The disks were then cut to yield 1.57 mm thick (longitudinal) samples in a north-south direction along the pith. The extraneous compounds were extracted from the strips in a Soxletwith a cyclohexane : ethanol (2:1) solution for 24 h then for 8 h with an ethanol solution. Following this step, the samples were placed in a hot-water bath for another 8 h, to ensure that all the extraneous compounds were removed. The samples were then air-dried under restraint to prevent warping. The density was measured with a direct-reading X-ray densitometer with a threshold density of 0.54 g/[cm.sup.3] for the earlywood to latewood transition. This yielded the density (oven-dry mass /green volume) and width of the earlywood, latewood, and whole ring. The pctlw in the ring is obtained by dividing the latewood width by the total rw. The height of the stem for every year was interpolated using the Carmean method (Carmean 1972).
A small log was taken between 1.0 and 1.3 m from each stem. These logs were then processed to 1 cm (radial) x 1 cm (tangential) x 15 cm (longitudinal) specimens from bark to pith along a northerly direction for bending tests. During the process, the exact location of the sample was noted, as well as the year of the rings that …
