ABSTRACT

The efficient utilization of woody biomass can be maximized by manufacturing engineered wood products instead of solid lumber due to substantial increases in yield from a log. For example, compared with a 40 percent yield in solid lumber from a log, parallel strand lumber, laminated strand lumber, and oriented strandboard can yield 65, 75, and about 80 to 90 percent, respectively (Schuler and Adair 2000, Balatinecz and Kretschmann 2001). Concurrently an increased demand for engineered wood products has led to increased utilization of small-diameter plantation trees that have received various management strategies to improve tree stem and wood qualties.

Many researchers have studied the effects of forest management on loblolly pine (Pinus taeda L.) wood anatomy (Bendtsen 1978; Taylor and Burton 1982; Megraw 1985; Bentdsen and Senft 1986; Cave and Walker 1994; Tasissa and Burkhart 1998; Groom et al. 2002; Jeong and Zink-Sharp 2012, 2013; among many others). Jeong and Zink-Sharp (2012, 2013) analyzed the effects of thinning and fertilization on the anatomical properties of young loblolly pine at different growth ring numbers and heights. Regardless of the effects of forest management, density increased as growth ring number increased, whereas density decreased as height increased. Different forest-managed trees had significantly different anatomical properties. Thinning and fertilization changed density, ring width, and latewood percentage. In addition, thinning significantly influenced earlywood and latewood tracheid length and cell dimension; fertilization significantly influenced cell dimension.

Likewise, there have been numerous studies on the effect of juvenile wood on the mechanical properties of loblolly pine. Tensile properties of the wood cells from juvenile wood were lower than those from mature wood because of higher microfibril angle (MFA; Mott et al. 1996, Groom et al. 2002). Jeong et al. (2009) reported that modulus of elasticity (MOE), ultimate tensile strength (UTS), and Poisson's ratio from earlywood and latewood generally increased with the increment in the growth ring positions except for specific gravity (SG) and UTS from latewood. Although the MOE for earlywood and latewood was not dependent on SG, the UTS for earlywood and latewood was highly dependent on SG. Jeong et al. (2009) also showed a positive relationship between MOE of earlywood, latewood, and tree height.

Cramer et al. (2005) proposed that SG and MFA were strong indicators of elastic properties of earlywood and latewood from loblolly pine. From the results of linear regression analysis, SG and MFA explained 75 percent of the accuracy of the elastic modulus value for latewood but less than 50 percent for earlywood. MOE increased with the increment in ring number and height, while shear modulus increased with ring number but decreased with height. Jeong et al. (2010) measured mechanical properties of differently oriented loblolly pine strands. The results showed that the mechanical properties of strands were strongly related with the growth ring orientation of strands.

Although many researchers have studied the effects of forest management on the anatomical properties of loblolly pine, there is a lack of information on the direct effects of anatomical structures on the mechanical properties of strand, particulate, and veneer components. The goal of this study was to begin the exploration of a link between the anatomical structures and mechanical properties of engineered wood strands from loblolly pine trees. The current study measured the MOE and UTS of loblolly pine wood strands extracted from the same stems used by Jeong and Zink-Sharp (2012, 2013) to investigate their anatomical structures.

Materials and Methods

Three young loblolly pine (Pinus taeda) trees were harvested from Reynolds Homestead Forest Resources Research Center in Critz, Virginia (36°40′N, 80°10′W). Each tree sampled was from a different forest management regime: the first tree was natural, the second tree was thinned, and the third tree was fertilized. Test specimens were prepared from blocks extracted from three 30-cm-long stem segments removed from each tree at uniformly spaced stem heights (0.5 to 1.5 m, 1.5 to 4.0 m, and 4.0 to 7.0 m). Anatomical descriptions of these stem segments at different growth ring numbers and heights were presented in the previous studies (Jeong and Zink-Sharp 2012, 2013). The blocks used for strand preparation were soaked in water and placed under a vacuum overnight to minimize surface damage during stranding. A laboratory flaker was used to process specimens to 127.0 by 25.4 by 0.5 mm (longitudinal by tangential by radial). Test specimen locations for each tension test are shown in Table 1. A minimum of 30 specimens were tested for each group. All strands were conditioned to 6 percent moisture content (MC) in an environmental chamber maintained at 30 percent relative humidity and 30°C before testing.

Table 1. Specimen arrangements of different heights and growth ring numbers from trees.

View Fullscreen

Tensile testing of the strands was conducted using an MTS universal testing machine equipped with an 889 N load cell and an extensometer. Gauge length for the extensometer and the distance between grips were 5.08 and 10.16 cm, respectively. Testing was conducted at a loading rate of 0.025 cm/min. Sample thickness and loading rate were adapted from Jeong et al. (2008). SG was measured according to ASTM D2395 (ASTM International 2005). Samples for measuring MC and SG were obtained from the same populations using untested strands. Hierarchical linear regression analysis was conducted to predict the MOE and UTS as a function of anatomical properties. The relationships between anatomical and mechanical properties (MOE and UTS) from the trees were evaluated and used to determine which anatomical properties were the most influential on mechanical properties using the R-square and stepwise selection methods.

Results and Discussion

Average SG of the strands at different heights and growth ring numbers is shown in Table 2. The lowest SG was found in the natural tree, whereas the highest SG was found from the fertilized tree. SG increased as the growth ring number increased, whereas SG decreased as the height increased. These trends agree with the anatomical results from Jeong and Zink-Sharp (2012, 2013). The variability of SG from the natural and fertilized trees decreased as growth ring number and height increased, but SG variability in the thinned tree was related to thinning year and height. With the increment in height, the impact from thinning on SG decreased. Tasissa and Burkhart (1997) reported that the impact of thinning gradually reached to a higher height of the tree after 4 years from the year of thinning.

Table 2. Moisture content (MC) and specific gravity (SG) of specimens from different trees.^a

View Fullscreen

MOE and UTS associated with growth ring number and height are shown in Figures 1 and 2. The highest MOE and UTS values with the lowest variation were found in the natural tree, whereas the lowest MOE and UTS values with the greatest variation were found in the fertilized tree. MOE and UTS from the different forest-managed trees increased as growth ring numbers increased except for 4.0 to 7.0 m from the thinned tree. With the increment in growth ring numbers, the variation of MOE from the natural tree increased, whereas the variation of MOE from the fertilized tree decreased. The variation of MOE from the thinned tree did not show any consistent trends with growth ring number. Compared with the variation of MOE, the variation of UTS from the different forest-managed trees showed a similar trend; however the magnitude of the variation was greater.

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

Increment of MOE and UTS from the different trees associated with increment of growth ring numbers can be explained by the growth of the anatomical structure. Although there were different magnitudes of anatomical change under different forest management regimes, with a corresponding increment in growth ring number, the density, tracheid length, latewood percentage, and the ratio of cell wall thickness to cell diameter from earlywood and latewood increased; ring width, however, decreased (Jeong and Zink-Sharp 2012, 2013). These anatomical changes associated with the increment of growth ring number had a positive influence on the mechanical properties of wood strands.

With the increment of height, MOE and UTS from different trees showed different behavior. As height increased from between 0.5 and 1.5 m to between 4.0 and 7.0 m, MOE and UTS from the natural tree increased 34.7 and 14.5 percent, respectively, which agreed with the results from previous studies (Cramer et al. 2005, Jeong et al. 2009). However, MOE and UTS from the fertilized tree decreased 36 and 40 percent, respectively, as height increased from between 0.5 and 1.5 m to between 4.0 and 7.0 m. Although MOE and UTS from the thinned tree fluctuated with increment of height, overall MOE and UTS from the thinned tree decreased from 0.5 to 7.0 m.

Inconsistent trends in MOE and UTS from the different trees with increments of height are difficult to explain by one specific anatomical property due to the complicated interaction between anatomical properties associated with height and forest management (Jeong and Zink-Sharp 2012, 2013). It can be postulated that interrelated anatomical properties collectively influenced the final mechanical properties.

Conventionally, SG is assumed to be strongly correlated with MOE. To some extent, it may be true when the differences in SG from different wood species are well above 0.1. However, SG within one tree may not differ more than 0.1. In other words, SG may not be a clear predictor of mechanical properties within one tree. Previous studies showed that mechanical properties were not dependent upon SG (McAlister et al. 1997, Cramer et al. 2005, Jeong 2008, Jeong et al. 2009). Likewise, SG may not be applicable in the prediction of mechanical properties of one wood species under different forest management.

It should be noted that MFA was known as one strong indicator to predict the mechanical properties of wood (Cave and Hutt 1969; Downes et al. 1993, 2002; Navi et al. 1995; Gindl and Teischinger 2002; Sedighi-Gilani and Navi 2007; Eder et al. 2009; Ivkovic et al. 2009; Salmen and Burgert 2009; Lachenbruch et al. 2010). Although the current study did not measure the MFA from the different forest-managed trees, it may help to provide more evidences in the prediction of the mechanical properties.

Figures 3 and 4 show the trends of MOE and SG and the trends of UTS and SG, respectively, according to increment of height. Different groups were selected for the comparisons with respect to three different heights in terms of consistent growth ring number and data availability. Given the increment of height, different trends between MOE and SG and between UTS and SG were observed from the different trees.

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

MOE from the natural and thinned trees did not follow the trends of SG. MOE from the natural tree increased with increment of height, whereas SG from the natural tree decreased at heights from between 0.5 and 1.5 m to between 1.5 and 4.0 m and increased from between 1.5 and 4.0 m to between 4.0 and 7.0 m. MOE from the thinned tree increased with the increment in height, whereas SG from the thinned tree decreased with increment of height. MOE from the fertilized tree, however, followed the trend of SG with increments of height.

UTS from the natural tree decreased slightly at heights from between 0.5 and 1.5 m to between 1.5 and 4.0 m but increased sharply from between 1.5 and 4.0 m to between 4.0 and 7.0 m, whereas SG from the natural tree decreased at heights from between 0.5 and 1.5 m to between 1.5 and 4.0 m but increased slightly from between 1.5 and 4.0 m to between 4.0 and 7.0 m. UTS from the thinned tree increased at stem heights from between 0.5 and 1.5 m to between1.5 and 4.0 m but decreased from between1.5 and 4.0 m to between 4.0 and 7.0 m; SG from the thinned tree decreased with the increment in height. Like the trend between MOE and SG from the fertilized tree, UTS followed the trend of SG from the fertilized tree with the increment in stem height.

It should be noted that mechanical properties can be represented as a vector quantity that has a certain direction and magnitude, whereas SG is a collective property interrelated with latewood percentage given ring width, and earlywood and latewood cell wall thickness given cell diameter. Therefore, higher SG may not always lead to higher mechanical properties. At the macro scale, wood material is assumed to be orthotropic material that has distinctive different mechanical properties in the longitudinal, radial, tangential directions (Jeong et al. 2010). These directional-dependent properties could not be explained by SG. However, it may be true for the isotropic material that SG is used to estimate physical properties. To predict mechanical properties of wood species that have a similar SG, more specific information on the relationship between anatomical properties and mechanical properties throughout different growth ring positions and heights may be required. Conventionally, most nondestructive devices for predicting mechanical properties of standing trees used SG as a key factor, which may lead to serious errors for MOE and UTS values of trees under different forest management.

Previous studies have shown that material properties were highly related to hierarchical structure of the material (Groom et al. 2002, Conrad et al. 2003, Cramer et al. 2005, Wittel et al. 2005, Burgert et al. 2007, Jeong et al. 2009, Lachenbruch et al. 2011, Kohan et al. 2012). Jeong and Zink-Sharp (2013) reported that thinning increased density, ring width, latewood percentage, earlywood and latewood tracheid length, and tangential latewood cell wall thickness. Jeong and Zink-Sharp (2012) also reported that fertilization increased ring width and earlywood cell diameter, but decreased density and latewood percentage. Different forest management practices influenced a complicated positive and negative interrelationship between the anatomical properties that existed in loblolly pine trees (Jeong and Zink-Sharp 2012, 2013). The effects of anatomical properties and interrelated anatomical properties from the different forest-managed trees on MOE and UTS of loblolly pine strands were analyzed using a hierarchical linear regression analysis.

Table 3 shows the results in different variable selection methods. R-square selection results showed that different trees had different anatomical properties that were correlated with MOE and UTS. The correlation coefficients between MOE and density, and between UTS and density, from the fertilized tree were found to be 0.71 and 0.73, respectively. It can be speculated that compared with the natural and thinned trees, a strong correlation between density and mechanical properties from the fertilized tree may be due to the uniform growth of cell dimension (Jeong and Zink-Sharp 2012). In spite of the fact that the highest SG was found from the fertilized tree, MOE and UTS from the fertilized tree were the lowest (Table 2), which indicated that SG could not be used as a predictor for the mechanical properties of loblolly pine.

Table 3. Hierarchical linear regression analysis of most influential wood characteristics on modulus of elasticity (MOE) and ultimate tensile strength (UTS) from trees.^a

View Fullscreen

The stepwise variable selection method chose the three most influential factors to predict MOE and UTS values by considering correlation between variables. With a combination of the three anatomical properties, MOE and UTS from the different forest-managed tree were predicted with R² ranging from 0.84 to 0.95. Except for the UTS from the fertilized tree, tangential latewood cell wall thickness (TLCWT) was included in the three most dominant factors to predict the MOE and UTS from different trees.

For the natural tree, a combination of TLCWT, radial latewood cell diameter (RLCD), and radial earlywood cell diameter (RECD) was selected as the most influential group of factors to predict MOE and a combination of TLCWT, latewood tracheid length (LTL), and RLCD was selected as the most influential to predict UTS. For the thinned tree, a combination of RW, LTL, and TLCWT was selected as the most influential group of factors to predict MOE and a combination of RW, TLCWT, and LTL was selected as the most influential to predict UTS. For the fertilized tree, a combination of density, tangential earlywood cell diameter (TECD), and TLCWT was selected as the most influential set of factors to predict MOE and a combination of density, earlywood tracheid length, and TECD was selected as the most influential to predict UTS.

From the R-square and stepwise selection results, it could be seen that mechanical properties of loblolly pine trees from different forest management practices were altered by anatomical properties and the interaction between anatomical properties. When mechanical properties of loblolly pine under different forest management are quantified nondestructively, specific anatomical information may increase in the prediction of specific mechanical properties.

Conclusions

MOE and UTS of loblolly pine increased as growth ring numbers increased. With increment of height, MOE and UTS from the natural tree increased, but MOE and UTS from the thinned and fertilized trees decreased. General trends of the mechanical properties of loblolly pine associated with growth ring number and height were found to be highly correlated to a certain combination of anatomical properties.

Several evidences were observed that SG was not a strong indicator to predict the mechanical properties of loblolly pine. Although the highest SG was found from the fertilized tree and the lowest SG was found from the natural tree, the highest MOE and UTS were found from the natural tree and the lowest MOE and UTS were found from the fertilized tree. With increment of height, SG decreased but MOE and UTS of loblolly pine from the natural and thinned trees increased. Altered anatomical properties and interrelated anatomical properties of loblolly pine were highly related with changes in MOE and UTS values associated with growth ring number and height.

Results from the R-square selection method showed that MOE and UTS values were correlated with different anatomical properties. However, prediction of the MOE and UTS with a specific anatomical property for different forest-managed trees seemed to be unreliable. Results from the stepwise selection method showed that MOE and UTS of loblolly pine under different forest management were better predicted with different combinations of anatomical properties, indicating R² values ranged from 0.84 to 0.95. It can be concluded that predicting MOE and UTS of loblolly pine under different forest management requires specific anatomical property information rather than SG.