Impact of location and forestry conditions on some physical and mechanical properties of northern Tunisian

Pinea pines have been used and cultivated for their edible pine nuts since prehistoric times. More recently, Pinus pinea L. has been introduced as an ornamental tree in Mediterranean regions, and is now often found in city parks and gardens. However, it has become naturalized outside North African cities to the point that it is now classified as an invasive species there. Its size (more than 15-20 m in height and 30- 40 cm in diameter) makes Pinus pinea a good candidate eco-resource for construc- tion materials in Tunisia. The focus of this study was to analyze the effects of geo- graphical location and forestry conditions on several physical and mechanical prop- erties of Northern Tunisian Pinea pine wood. The Pinus pinea wood samples studied were collected from four different geographical locations, divided into 87 plots according to tree population density and soil fertility class. Density, shrinkage and moisture content were measured and mechanical tests were performed on each wood sample. Preliminary results showed that Tunisian Pinea pine wood has very good dimensional stability in relation to its density. However, the modulus of rupture (MOR) in bending and compression strength is lower than in other Tunisian softwood species, whatever the geograph- ical situation and soil fertility. A linear regression analysis showed that only tree population density seems to have a signif- icant impact on the variability of Pinea wood properties, except for dimensional stability and volumetric shrinkage. Our findings suggest that the growing popula- tion density of P. pinea trees is the most important forestry parameter determining its mechanical and physical wood material properties. It could therefore be of interest to improve forestry conditions to obtain better wood quality.

Pinea pines have been used and cultivated for their edible pine nuts since prehistoric times.More recently, Pinus pinea L. has been introduced as an ornamental tree in Mediterranean regions, and is now often found in city parks and gardens.However, it has become naturalized outside North African cities to the point that it is now classified as an invasive species there.Its size (more than 15-20 m in height and 30-40 cm in diameter) makes Pinus pinea a good candidate eco-resource for construction materials in Tunisia.The focus of this study was to analyze the effects of geographical location and forestry conditions on several physical and mechanical properties of Northern Tunisian Pinea pine wood.The Pinus pinea wood samples studied were collected from four different geographical locations, divided into 87 plots according to tree population density and soil fertility class.Density, shrinkage and moisture content were measured and mechanical tests were performed on each wood sample.Preliminary results showed that Tunisian Pinea pine wood has very good dimensional stability in relation to its density.However, the modulus of rupture (MOR) in bending and compression strength is lower than in other Tunisian softwood species, whatever the geographical situation and soil fertility.A linear regression analysis showed that only tree population density seems to have a significant impact on the variability of Pinea wood properties, except for dimensional stability and volumetric shrinkage.Our findings suggest that the growing population density of P. pinea trees is the most important forestry parameter determining its mechanical and physical wood material properties.It could therefore be of interest to improve forestry conditions to obtain better wood quality.

Introduction
Pinea pine, Pinus pinea L., is one tree species which is found around the Mediterranean basin (Sbay, 2000;Khaldi, 2009).This wood species (figure 1) was successfully introduced in Tunisia at the beginning of the 20 th century along the Mediterranean coast line in order to stabilize the littoral dunes of Bizerte, situated in the north and along the north-eastern coast around the Cap Bon region (Hasnaoui, 2000).The success of these first plantations incited foresters to use this species to stabilize the littoral dunes of northwest Tunisia as well (Sghaier et al., 2006).Today, P. pinea accounts for over 21,000 ha of the Tunisian forest area (El Khorchani, 2010) and has become one of the most valuable species in Tunisian reforestation programs (Lesnino, 1997), not only for wood production (Polge, 1978) and paper pulp manufacturing (Hasnaoui, 2000), but also for its nuts which are widely enjoyed and used in many traditional Tunisian dishes (Khouja, 2003).
Natural regeneration is difficult to achieve in Tunisia.Cone yield and the amount of available seeds together with other limiting factors such as livestock and overgrazing often result in the failure of P. pinea stands to regenerate naturally.Pinus pinea is now listed as an invasive wood species (Agrimi and Ciancio, 1994).For these reasons, various appropriate ways to use P. pinea wood and its fruits should be found and the trees subsequently regenerated through sustainable management of P. pinea tree population and its fruit supply.Due to its tree dimensions (more than 15-20 m long and 30-40 cm diameter), P. pinea could also be a good potential green resource for structural material in Tunisia (Khaldi, 2009).In spite of the large interest in using this wood species in structure, there remains a great lack of knowledge about the properties of the wood material.This present study is part of a Tunisian forest development program, conducted by the National Institute of Agricultural Engineering Research, Water and Forestry (INRGREF) in Tunis.The aim of this work consists in analyzing the impacts of geographical location and forestry conditions on several physical and mechanical properties of Northern Tunisian Pinea pine' wood.Wood samples were collected from four different geographical locations divided into 87 plots according to Pinus pinea population density and the soil fertility class.Density, shrinkage and moisture content measurements as well as mechanical test (bending and compression strength) were performed on each wood sample.The results from these physical and mechanical tests were correlated to the tree growing conditions such as population density, soil fertility and geo-localization.These results could form a data base of Pinea pine' wood properties in order to have a preliminary idea of recoverable ways of this invasive Tunisian softwood species.
The climate of the North-West region covered by this study is wet and temperate.Annual rainfall varies, according to altitude from 640 mm/year on the coast to 1,572 mm/year in the region of Aïn Draham.Average annual temperatures range from 14.9 °C to 18.5 °C.The average of maximum temperature of the warmest month can reach more than 30 °C and the average maximum of the coldest month is around 4 °C.The soil is poorly developed in coastal dunes and leached brown forest in the mountains (Rejeb et al., 1996).Table I describes the topographic, climatic and soil characteristics from each selected area.
Each region was divided into several plots according to the P. pinea trees population density.These sections (6 ares of circular area) were classified into 6 density ranges from 500 to 3,000 trees/ha.Table II shows the plot repartition, according to area and tree population density.

Wood sample selection
The number of Pinea pine wood samples is also shown in table 2 & figure 2 according to its source region: 39 samples from Bizerte, 15 from Sejnane, 14 from Nefza-Tabarka and 20 from Aïn Draham.The selected trees have diameters (at 1.30 m from the ground) comprised between 25 and 35 cm.Relating to the method developed by Oger and Leclercq (1977), one tree without defects (knots, biotic and abiotic alterations) and with perfect rectitude was selected for each section (table II), providing one wood sample per section, making a total of 87 Pinea pine wood samples.To perform physical and mechanical tests, a wooden disk 50 cm in thickness was cut at 1.30 m from the ground for each selected tree.

Physical properties
To avoid errors during sampling, extreme cases such as excessively knotty trees or the presence of reaction wood or slope grain were taken into account (ISO 4471, 1982).From each disk, 12 samples of 3 cm width from bark to bark were cut.These samples were then cut into strip 2 cm thick.Moisture Content, basic density, Airdry density (after conditioning in a climatic room at 20 °C and 65% RH) and oven dry (after conditioning in an oven at 103 °C) density (D m12 , D m0 ) (ISO 13061-2, 2014), shrinkage (β) [tangential (βt), radial (βr), longitudinal (βl), volumetric (βv)] (ISO 4469, 1981) of the wood samples were determined according to International Organization for Standardization (ISO) Standards using wood specimens of 2 × 2 × 3 cm (along the grain).The shape factor (β t/ β r) was the ratio between tangential and radial shrinkage.Where m h is the humid mass of the initial sample and m 0 is the oven-dried mass of the wood sample.
Where D b is the basic density of wood (g.cm -3 ), D m0 is the oven-dried density of wood (g.cm -3 ) and D m12 is the Air-dried density of wood (g.cm -3 ).V h is the green volume of the specimen (cm 3 ), V 0 is the oven-dried volume of the sample and V 12 is the air-dried volume of wood sample.m 0 and m 12 are the oven-dried and air-dried weight of the sample (g), respectively.

Determining mechanical strength properties
In order to assess the effect of P. pinea growing conditions and locations on mechanical properties, three point bending (MOE, MOR) and compression tests were carried out for each of the selected wood tree samples, and results were compared.An INSTRON 4467 Universal Mechanical Test Machine was used for the measurements.Samples were conditioned in a climate-controlled room with 65% RH and at 22 °C for the time required to stabilize the samples weights.

Bending test
Three point static bending tests were carried out according to the EN 408 (2003).The sample size was 400 x 20 mm x 20 mm 3 (L x R x T).The moving head speed and span length were 0.09 mm.s -1 and 260 mm, respectively.The load deformation data obtained were analyzed to determine the modulus of elasticity (MOE) and the modulus of rupture (MOR).The tests were replicated on 20 samples from each selected P. pinea tree.

Compression strength parallel to grain
Compression tests were carried out according to the EN 408 (2003).Deviating from the norm, a reduced specimen size of 30 x 20 x 20 mm 3 (L x R x T) was used.The moving head speed was 0.09 mm.s -1 to ensure wood sample rupture within 1.5 to 2 minutes.The load deformation data obtained were analyzed to determine the MOR.20 specimens per selected tree were tested.

Statistical analysis
Physical and mechanical results were compared to Pinus pinea growth conditions as tree population density in the section (DEN), altitude (ALT), exposition (Expo) and region, using a linear regression statistical analysis.Such analyses attempted to model the relationship between two variables by fitting a linear equation to the observed data.One variable was considered an explanatory variable, while the other was considered to be a dependent variable.

Moisture content
Several studies have demonstrated that P. pinea growing has a high sensitivity to climatic conditions (Campelo et al., 2006;Mazza and Manetti, 2013) and more particularly to the geographical variability in rainfall which most influences tree ring formation (Mazza et al., 2014).Both rainfall and humidity have an important impact on the anatomical structure of the wood and on its physical and mechanical properties (Skaar, 1988).
Based on the Moisture Content measurements (figure 3), the climatic conditions were similar to each other whatever the tree growing regions.However, it appears that the MC (%) of P. pinea issued from Aïn Draham (69.1%) and Bizerte (64.7%) are slightly higher than those of trees from the Sejane region (60.2%) and the Nefza-Tabarka (60.9%).These results are in agreement with the ratio between rainfall quantity and dried periods of each harvesting site (table I).In addition, the average MC value of northern Tunisian woods tested in this study was 64.3%, which means that P. pinea' wood among those timbers with low initial humidity (Hasnaoui, 2000) thus make it possible to converting it to in material by reducing the duration of the drying process and limiting the risk of dimensional material deformations and cracks.

Mechanical properties
The mechanical test results of each wood sample issued from the four regions and conditioned at a temperature of 20 °C and 65% RH are shown in figure 4. Elasticity modulus (MOE) implies that deformations produced by low stress are recoverable after removing the load.Modulus of rupture (MOR) reflects the maximum capacity in bending or in compression and is proportional to the maximum moment borne by a sample.MOE is the ratio of stress to strain in compressing or bending deformation.
Based on the MOE in bending and in compression strength measurements (figure 4), tree growing area does not have a significant impact on P. pinea wood mechanical properties.The MOE average values in bending and in compression of northern Tunisian woods tested in this study were 11 980 MPa and 41 MPa, respectively.
Based on the MOR in bending and in compression strength measurements (figure 4), tree growing area does not have a significant impact on P. pinea wood mechanical properties.The MOR average values in bending and in compression of northern Tunisian woods tested in this study were 52.37 MPa and 37.26 MPa, respectively.These results show that the Northern Tunisian Pinus pinea' wood can be classified as having medium static bending strength (75 MPa < MOR < 110 MPa) and low axial compressive strength (30 MPa < MOR < 40 MPa) (Collardet and Besset, 1998).

Wood densities
Air-dried Density (D m12 ) is commonly used to compare different woods.Basic Density (Db), Oven-Dried Density (D m0 ) and Air-dried Density (D m12 ) have been measured on each wood sample.The results did not find any effect of the tree growth area on these densities.However, tree population density has an impact on wood material density (figure 5a).Table III presents the average values of the different densities.
According to Campredon (1967), Pinus pinea timbers which have been tested in this study can be classified as a light heavy weight wood (0.5 <D m12 < 0.6).

Discussion
To show the impact of forestry conditions on physical and mechanical properties, a linear regression analysis has been performed using the results cited above.The results from this statistical analysis are presented in table IV.Zobel and Van Buijtenen (1989) suggest that large structure variations are produced by changes in climate, site and management characteristics, due to the influence of these extrinsic factors on various activities.Our study indicated that there are no significant differences among wood coming from different growing areas in terms of wood mechanical properties, wood density, wood shrinkage variations and static bending properties.Table IV shows the coefficients of correlation between forestry conditions (Tree population density, Growing Area, Altitude), physical properties (Wood basic density -Db, volumetric shrinkage -βv) and mechanical strengths (Bending and compression strength -MOE and MOR) which were measured in this study.
Positive correlations were found between wood basic density, bending strength and compression strength.These results support previous studies (Zhang and Zhong, 1997;Betkas et al., 2002;Heräjärvi, 2004).More interesting correlations were found between tree population density, wood density and mechanical properties.These results (figure 5) show that, in this study, the most important parameter does not depend on climatic and geological conditions of tree growth, but more particularly on forestry conditions which were characterized by tree population density.
Generally, width of annual ring was inversely correlated with density of wood in softwood (Erten and Sözen, 1997).Because, while width of annual ring was getting increase, the latewood zone which is heavy portion of wood was getting  decrease (Berkel 1970).Indeed, some authors have reached similar findings by underscoring that an increase in tree plantation density results in a decrease in annual rings width (Nepveu, 1994;Pardé and Bouchon, 1988).According to Demirkir et al. (1997), mechanical and physical properties of softwood and hardwood species were affected from width of annual ring.Growth rate influences wood density due to the changes in the relative proportions of secondary cell walls and void volume (e.g.cell lumen) (Mäkinen et al., 2002;Saranpää, 2003) as well as in the relative amounts of chemical components of the cell wall (Saikku, 1975).These chemical changes and structural modifications impact wood mechanical properties.It is obvious that rapid growth rate due to a low population density results in low wood density with low mechanical strength properties (Saranpää, 2003).The trends observed in figure 5 show that the higher the tree population density, the greater the wood density and mechanical strength.
Finally, tree population density may have an effect on the concurrence between trees, assuming that the tree ages are comparable (trees age is comprised between 30 and 40 years old), and therefore a negative effect on ring width.Negative correlations were often reported between ring width and wood density for coniferous trees (Berkel, 1970); it would not be surprising to observe an indirect positive correlation between tree population density and wood density.However, it is very surprising that we obtain such a strong correlation between trees population and wood basic density.Results from this study should therefore not be considered as a generality.

Conclusion
The study focused on moisture content, densities, dimensional variations and mechanical properties of Northern Tunisian Stone pine woods issued from Bizerte, the Garaa Sejnane Watershed, Nefza-Tabarka and Aïn Draham areas.The physical and mechanical properties of these woods appear to depend more mainly on forestry population density than climatic conditions in all of the regions investigated.Indeed, results from the linear regression highlighted that the tree population density is the factor which has the greatest influence on Pinus pinea' wood basic density and mechanical properties.For most wood species, it appears that the denser the tree population, the higher the density and mechanical strength.In Tunisia, P. pinea' wood is mainly used for building boats, stilts, wood mines, lumber and poles.It is also used to produce of cellulose and mechanical pulp (Hasnaoui, 2000).Despite its physical and mechanical properties comparable or even superior to those of Spain, Tunisian Pinewood is mainly used in the crushing industry and to manufacture pallets.

Table IV.
Correlation matrix of all forestry conditions and physical/mechanical properties of the Northern Tunisian Pinus pinea wood sampled.

Tree population
Figure 1.Description of Pinus pinea L. tree.(a) Pinus pinea tree; (b) Recent plantation of stone pine in the region of Sejnane; (c) Cone from stone pine ; (d) Male inflorescence; (e) Female inflorescence.

Figure 2 .
Figure 2. Location and tree quantities of Pinus pinea L. wood samples.
Figure 3. Moisture Content (MC %) and shrinkages in different orientations of Pinus pinea woods according to their source regions.

Figure 4 .
Figure 4.The modulus of elasticity (MOE) and the modulus of rupture (MOR) in bending (a) and compression (b) of Pinus pinea woods related to their sourcing regions.
Figure 5. Correlation between Pinus pinea tree population density, respective wood Basic density (a) and wood mechanical strength [MOR in bending (b) and compression (c)].

Table II .
Plots repartition according to area and tree population density