text new page (beta)
English (pdf)
Article in xml format
Automatic translation
Cited by SciELO 
Cited by Google 
Similars in
SciELO 
Similars in Google 
Permalink
Introduction
Papua New Guinea (PNG) is an Oceanian country located in the south-western Pacific Ocean region. The country has a total land area of 46,3 million hectares of which 33,6 million ha is estimated to contain forest cover (FAO 2015). Primary and regenerated forests represent 52,4% and 47,6% of the forest area, respectively, accounting for a total growing stock of 5195 million m3. According to the PNG Forest Authority, there are more than 2000 tree species in PNG, of which 20% are utilized in one way or another for commercial use (PNGFA 2007). Commercial forest plantations started in the 1960s with sporadic progress and currently covering 62277 ha. The country aims to develop 240000 ha of commercially viable and sustainable commercial forest plantations by the year 2030 to sustain 3,6 million m3 of industrial timber currently harvested from natural forest annually.
Although the forest cover in PNG is large and diverse, knowledge of the current timber resource is scarce. In the 1970s, an exhaustive review of the mechanical properties of PNG timbers from native forests has been conducted by Bolza and Kloot (1976). However, the most valuable commercial timbers have since been harvested in large areas of the primary rainforest as a result of forest policies focussing on export logging. Secondary forests are recovering in terms of merchantable timber and carbon stocks but much of the potential timber resource is of lesser-known timber species.
Recent studies by Edwin and Ozarska (2015) and Kotlarewski et al. (2016) determined selected mechanical properties of some PNG timber species. However, there are still many research and structural challenges, constraints, and opportunities at different levels which need to be addressed to support the development of competitive value-added wood industries. The influence of environmental conditions and other growth factors on wood mechanical properties and the challenges associated with using this material has been the subject of numerous studies worldwide. Barrett and Kellogg (1991) noted a reduction in the mechanical properties of second growth Douglas-fir (Pseudotsuga menziesii) when compared to old-growth timber. Other studies also demonstrated that properties of wood produced from managed trees are different from wood produced in natural stands (Bendtsen 1978, Pearson and Gilmore 1980).
Currently, most local wood processing in PNG is focused on primary conversion of logs to low-grade building materials. Technical knowledge and capacity about efficient processing of different native timber species and produce a broader range of wood products are very low. The variability in properties of the wood material and of the finalized timber products indicates a potential for optimization too. Consequently, there is a need to provide research and technology development to develop and implement commercially sustainable log supply chains, knowledge, and capacity in wood science and processing technologies, as well as the processing structures which support successful domestic value-adding wood processing enterprises.
The aim of the project was to increase the contribution that utilization of forest resources makes to national and local economies, including landowners and processors, through the development of domestic value-added wood processing methods. The specific objectives were to 1) enhance the knowledge of wood properties of PNG timbers to facilitate greater value adding; 2) identify any relationship between log position in a tree and the mechanical properties of timber products to optimize the utilization of PNG timbers at various stages of the chain. Therefore, a testing program has been developed including the assessment of physical and mechanical properties of selected commercial and lesser-known species from secondary and plantation forests.
Materials and methods
Harvesting
The testing program has been divided into two groups of species to facilitate harvesting and specimens’ preparation logistic. Each group included 13 species from plantations and regenerated forests (also known as regrowth forests) located in the Morobe and West New Britain provinces, PNG (Table 1). Nine species were harvested from plantations and 17 from regenerated forests (3 softwoods, 23 hardwoods). A total of 130 trees, i.e. 5 trees per species, have been selected and harvested in accordance with ASTM D5536 (2010). The trees have been selected based on the following selection criteria: 1) the tree had to be more than 15 years old after regrowth or planting; 2) all trees for a specific species had to be from the same forest area; 3) the selected trees had to be representative of the population, with good form and merchantable height. Following harvesting, the total merchantable height of each tree has been further cut into 3 to 4 m-long logs, labeled as per height from ground (i.e. bottom, middle, and top section), and milled. The milled sawn boards were then kiln-dried to 12% moisture content.
Machining specimens and mechanical testing
Specimens have then been machined in accordance with ASTM D143 (2009) and conditioned to 23°C and 65% relative humidity until constant mass prior to testing. Whenever possible and applicable, a 50 x 50-mm specimen cross-section size has been selected. The selected mechanical properties for the present study were stiffness (MOE), flexural bending strength (MOR), compression strength parallel and perpendicular to the grain, shear strength parallel to the grain, and hardness (Janka). All the tests have been conducted using a universal testing machine (Instron model 5569, MA, USA). Prior to testing, each specimen has been measured to determine the volume and weighted to determine mass at the time of testing.
After every test, a section has been taken from the specimen near the point of failure, weighted, and placed at 103°C for 24 hours to determine the oven-dry mass and moisture content.
Static bending test (MOE and MOR): 20 specimens of dimensions 50 x 50 x 760 mm have been prepared per species and tested using a 3-point bending rig. The specimens have been tested using a loading speed of 2,5 mm/min until failure.
Compression strength parallel to the grain: 20 specimens of dimensions 25 x 25 x 100 mm have been prepared and tested per species. The speed of testing was 0,3 mm/min.
Compression strength perpendicular to the grain: 20 specimens of dimensions 50 x 50 x 200 mm have been prepared per species. The speed of testing was 0,6 mm/min.
Shear strength parallel to the grain: 20 specimens of dimensions 50 x 50 x 63 mm have been prepared and tested per species. The speed of testing was 0,6 mm/min. The portion of the piece that was sheared off has been used as a moisture content specimen.
Janka hardness: 20 specimens of dimensions 50 x 50 x 150 mm have been prepared and tested per species. The number of test penetrations included two (2) on a tangential surface, two (2) on a radial surface, and one (1) on each end. Only indentations on tangential and radial surfaces have been used to calculate the Janka hardness value. The speed of testing was 6 mm/min.
Statistical analysis
The data were analyzed in Minitab statistical software (Minitab 18, Version 18.1) to evaluate the effect within species of the position in a tree on the mechanical properties. A fit mixed effects model and a Fisher pairwise comparisons analysis (α = 0,05) were used. In the model, the tree identification number (tree #1 to #5) was used as a random factor where the species and position in the tree were the fixed factors.
Results and discussion
A total of 2,641 specimens from 26 species and 130 trees have been tested i.e. shear: 528 specimens; compression parallel: 526 specimens; compression perpendicular: 532 specimens; static bending: 525 specimens; hardness: 530 specimens. The total merchantable height of Magnolia tsiampacca trees was usually too short to obtain top sections. Therefore, only bottom and middle sections could generally be obtained for this species which has been excluded from the model.
Hopea iriana provided significantly higher mechanical testing results than any other species across all selected mechanical properties (Table 2). A group composed of Eucalyptus pellita, Homalium foetidum, and Intsia bijuga usually performed significantly better than the other tested species. Xanthophyllum papuanum, Anisoptera thurifera, and Castanospermum australe also typically performed above average. At the other end, a group formed of Octomeles sumatrana, Falcataria moluccana, Endospermum medullosum, Alstonia scholaris, Palaquium warbargianum, Elaeocarpus sphaericus, and Magnolia tsiampacca generally offered mechanical properties below the average.
Table 2: Summary of mechanical properties per species.
MOE: Modulus of elasticity or stiffness; MOR: Modulus of rupture or bending strength.
Effect of position in the tree on selected physical and mechanical properties
The physical and static bending properties per species based on the position in the tree are presented in Table 3. The statistical analysis demonstrated a very significant interaction (P-value <0,001) between species and position in the tree on the bending properties. Stiffness and bending strength tend to decrease or remain unchanged along the stem across all studied species. The average stiffness and bending strength values obtained from bottom and middle sections were significantly greater than those from top sections for 6 (Araucaria cunninghamii, Canarium oleosum, Hopea iriana, Intsia bijuga, Pangium edule, Xanthophyllum papuanum) and 5 (Canarium oleosum, Homalium foetidum, Intsia bijuga, Pangium edule, Xanthophyllum papuanum) species, respectively. Such a trend was not noticeable in species typically performing below the average of all tested species. This observation is in accordance with other studies reported in the literature (Harvald 1988, Shivnaraine 1989, Hojbo 1991, Kliger et al. 1995, Machado and Cruz 2005). In one case, i.e. Eucalyptus pellita, top sections provided significantly higher bending properties than bottom sections. Such a result might be explained by the fact that specimens were taken from logs without any consideration to the radial position within the tree. Consequently, some specimens obtained from bottom sections might have included a higher proportion of juvenile wood, usually not as strong as mature wood, if taken from the outer radial section of the tree and vice-versa. Kliger et al. (1995) also reported that strength and stiffness in butt logs usually tend to increase further away from the pith in the radial direction.
Table 3: Physical and bending properties per species and position in the tree (i.e. bottom, middle, or top section based on height from ground).
aAir dry density at 12% moisture content; bMean; cStandard deviation; d Number of specimens.
The position in the tree had a more limited impact on the compression strength properties. The species and position in tree interaction significantly influenced the compression perpendicular to grain in the case of two species, i.e. Eucalyptus deglupta and Pterocarpus indicus (Table 4). In both cases, top sections provided significantly higher average values than bottom sections (P-value <0,020). In the case of compression parallel to the grain, the crushing strength was not significantly influenced by the position in the tree for any of the tested species. The Fisher pairwise comparisons analysis identified two species, i.e. Castanospermum australe and Anisoptera thurifera, where the bottom sections provided significantly higher results than the top sections.
Table 4: Compression parallel and perpendicular to the grain per species based on position in the tree (based on height from ground).
aMean; bStandard deviation; cNumber of specimens *not applicable
Table 5: Shear parallel to the grain and hardness per species based on position in the tree (based on height from ground).
a Mean; b Standard deviation; cNumber of specimens.
The position in the tree also had a limited effect on shear and hardness testing results (Table 5). The model only identified species as a very significant factor (P-value <0,001) influencing shear strength and hardness. The Fisher pairwise comparisons analysis allowed identifying five species where hardness was significantly lower in top sections than bottom sections, i.e. Xanthophyllum papuanum, Homalium foetidum, Intsia bijuga, Syzygium spp., and Pometia pinnata. In the case of shear strength, four species could be identified as having significantly lower values in top sections versus bottom sections, i.e. Canarium oleosum, Homalium foetidum, Pometia pinnata, and Syzygium spp.. Like bending strength results, such a trend was usually noticeable in mid to high strength species.
Comparison between old-growth forests and regrowth forests or plantations
The mechanical properties of species obtained from plantations and regrowth forests were on average 24% lower than those from old-growth forests (Table 6, Bolza and Kloot 1976). However, it is not clear whether Bolza and Kloot (1976) used a primary (50 x 50 mm specimens) or secondary (25 x 25 mm or 20 x 20 mm) method for smaller size specimens. The authors also noted the extremely limited amount of test material to prepare specimens for many of the tested species, which suggested that smaller size specimens might have been considered here. Therefore, a length effect (Madsen 1990) where natural growth characteristics create cross sections with varying strengths along the length of a timber member might explain the lower mean values obtained in the present study. There was also no indication of tree age in the case of Bolza and Kloot (1976) which might have influenced the observed differences.
Octomeles sumatrana (-43%), Magnolia tsiampacca (-40%), and Syzygium spp. (-39%) were the most affected species when comparing mechanical testing results with their counterparts from old-growth forests. On the other end, Anisoptera thurifera (+4%), Elaeocarpus sphaericus (-10%), and Araucaria hunsteinii (-11%) were the least affected. The density variability observed between old-growth and regrowth or plantations timbers also support the differences observed between mechanical testing results. Octomeles sumatrana (-22%), Magnolia tsiampacca (-27%), and Syzygium spp. (-36%) were again some the species most affected a density reduction where Anisoptera thurifera (+7%), Araucaria hunsteinii (+5%), and Elaeocarpus sphaericus (0%) showed an increase or no difference. Overall, the density of plantations and regrowth forests timbers was 11% lower compared with old-growth forests. However, density alone does not seem to explain all the variability observed and other elements such as ring width might need to be considered (Kliger et al. 1995). Barr et al. (2015) also observed a weak correlation between the density and the bending strength for Intsia bijuga.
Table 6: Physical and mechanical properties of Papua New Guinea species from old-growth forests (Bolza and Kloot 1976).
ADD: Air dry density at 12% moisture content; MOE: Modulus of elasticity or stiffness; MOR: Modulus of rupture or bending strength. Araucaria cunninghamii, Falcataria moluccana, Castanospermum austral, Eucalyptus pellita, Palaquium warbargianum, Pangium edule, Pinus caribaea are not included in the list of studied species by Bolza and Kloot (1976).
Anisoptera thurifera was the only species from regrowth forests where the mechanical testing results were usually higher (+4%) across all selected mechanical properties than those from old-growth forests (Bolza and Kloot 1976). In the case of plantations species, Magnolia tsiampacca (-40%) and Eucalyptus deglupta (-35%) were the species showing the most important average decrease in terms of mechanical properties when compared with old-growth forests. Araucaria hunsteinii (-11%) and Terminalia brassii (-17%) were the two plantations species least affected when compared with their old-growth counterparts. Compression perpendicular to the grain (-31% on average across all species), hardness (-30%), and compression parallel to the grain (-29%) were the properties showing the highest drops when comparing with old-growth forests timbers. The bending stiffness (-23%) was the mechanical property least affected when compared with results from old-growth forests followed by bending strength and shear (-24% for both).
A study by Edwin and Ozarska (2015) support the observed decrease in mechanical properties of PNG timbers from regrowth forests and plantations when compared with old-growth forests. The physical and bending properties results of 6 hardwoods species from secondary forests in PNG are also in accordance with those obtained in the present study. A study by Haslett et al. (1991) on Anthocephalus chinensis from West Samoa also provided physical and bending properties fitting those of the present study with stiffness and bending strength of 6,8 GPa and 58 MPa, respectively, for an ADD of 340 kg/m3. A lower compression parallel to the grain (25 MPa) might be related to a site factor for growth traits as reported by Leksono et al. 2008 and Hung et al. (2015) for Eucalyptus pellita. Multiple studies on Eucalyptus pellita also demonstrated the site impact on timber mechanical properties (Bootle 2005, Kelin et al. 2006, Hung et al. 2015). Interlocked grain as an adaptive trait for tropical tree species in the rainforest has also been suggested by Cabrolier et al. (2009) after the authors noticed strong variations between and within trees. Where special attention was taken in the selection of specimens for the present study, the presence of interlocked grain could explain some of the differences observed between old-growth and secondary growth timbers. Barr et al. (2015) studied the impact of interlocked grain on bending properties of Intsia bijuga. The authors noted that the grain can vary significantly from straight to interlocked and influence significantly the bending strength. Interestingly, stiffnesses and densities found in the present study are similar to those found by Barr et al. 2015 i.e. 758 and 812 kg/m3 and 14,2 and 15 GPa, respectively. However, MOR values differed significantly (116,6 versus 152,5 MPa) potentially because of the presence of interlocked grain.
Conclusions
Six mechanical properties, namely bending strength, stiffness, compression parallel and perpendicular to the grain, shear parallel to the grain, and hardness, have been evaluated for 26 PNG species using 2641 small clear specimens from 130 trees.
The impact of the position in the tree on the selected mechanical properties has also been assessed. Stiffness and bending strength tend to decrease or remain unchanged along the stem across all studied species. Where shear and hardness testing results showed a similar trend to a lesser extent, the position in the tree had a much more limited impact on the compression strength properties. Further experiments where sampling would consider the radial position within the tree might accentuate observed trends. Therefore, segregating logs based on the position in the tree could be of interest where desired timber mechanical properties and costs associated with segregating is justifying optimum mechanical properties for the intended end use.
The mechanical properties of species obtained from plantations and regrowth forests were lower than those found in the literature from old-growth forests. Different factors including the size of specimens tested, the amount and provenance of tested material, and some adaptive traits for tropical tree species might explain some differences. However, comparisons of mechanical testing results with other recent studies tend to confirm a reduction of physical and mechanical properties when comparing with timbers from old-growth forests.
