Introduction 

Assessing the wood quality of individual trees and overall forest stands has become an important procedure in forest operations, as forestry and wood processing industries are under increasing economic pressure to maximize extracted value ([40]). For this reason, the estimation of timber species, quantity, and quality is critical for quantifying the productive value of a forest ([19], [25]). The worldwide shift in the wood supply from old-growth forests resources to intensively managed plantations increases the need of evaluating tree quality prior to harvest ([35]).

Significant efforts have been made to develop robust, non-destructive technologies (NDT) capable of predicting the intrinsic wood properties of individual living trees and assessing wood quality by stand and forest. In addition, non-destructive technologies including (mini) rhizotrons and ground-penetrating radar have been recently proposed to assess plant rooting distribution and growth ([26], [20]). The use of such technologies not only leads to greater profitability for the forest industry, but can also help foresters to make economic and environmental management decisions for treatment of individual trees and forest stands, improve thinning and harvesting operations, and efficiently allocate timber resources for optimal utilization ([40], [25]). Acoustic technologies have become well established as material evaluation tools, and their use has become widely accepted for quality control by the wood industry ([41]). With the development of portable and simple-to-use, time-of-flight and resonance-based tools, the use of acoustics in the forestry sector has increased, particularly in countries such as New Zealand ([34], [31], [3]), Australia, USA ([37], [39]), Hungary ([2], [6]) and the United Kingdom ([27]). In Japan, Nakamura ([21]) used ultrasonically-induced waves to assess larch trees and observed significant differences in acoustic velocities and modulus of elasticity (MOE) for trees in forest stands at different locations and with different density. The non-destructive technologies, equipment, and evaluation procedures arising from those efforts are now widespread, and their use is becoming more and more common also for European forest industries. This paper reports a case study using these tools in managed conifer forests in Italy.

The objective of the study was to determine if the effects of silvicultural practices on wood quality can be identified using acoustic measurement to assess the MOEd (dynamic modulus of elasticity) of standing trees with non-destructive method in Calabrian pine (Pinus nigra Arnold subsp. Calabrica) plantations. In Calabria, where the tests were conducted, forest cover is 40.6% compared to the national average of 34.7%. Each year, the average increase in wood volume in this region (equal to 6-8 m³ ha^-1) exceeds and sometimes doubles the estimated increase in other forests in and around southern Italy ([46], [24]). A secondary aim was to determine if the effects of silvicultural practices on wood quality can be identified using this technique.

  Materials and methods 

The time-of-flight method

The system used for measuring acoustic velocity in trees involves inserting two sensor probes (a transmit probe and a receiver probe) into the sapwood for few centimetres in a vertical plane, according to the manufacturer’s instructions, and introducing acoustic energy into the tree through a hammer impact. Wave propagation velocity in the tree is determined by measuring the time-of-flight of a single pulse wave. Time-of-flight (TOF) essentially measures the time for the stress wave to travel from the transmit probe to the receiver probe. Piezoelectric sensors inside the two probes are used to sense the wave passing.

The acoustic velocity is then calculated from the distance between the two sensor probes and the TOF data using the following equation (eqn. 1):

where CT is the acoustic velocity (m s^-1), S is the distance between the two probes (sensors) (m), and ΔT is the time-of-flight (s).

The FAKOPP TreeSonic™ (Fakopp Enterprise, Agfalva, Hungary) is an acoustic tool specifically designed to study standing trees ([1], [16], [10], [3], [2], [6], [42], [44]). The tool is comprised of a hand-held hammer and two probes, a transmitting accelerometer, and a receiving accelerometer. The vibration transducers (patented by the Weyerhaeuser Company) are equipped with a sliding hammer, resulting in quick operation ([11]). To measure the acoustic velocity wave, the start and stop sensors are driven through the bark at a 45° angle into the wood of the standing tree. Unlike other similar tools (Hitman Resonance Tool, IML Impulse Hammer and Microsecond FAKOPP) which are specifically modelled for radial measurement, the FAKOPP TreeSonic™ was developed to operate in the longitudinal direction along the stem (Fig. 1). During field acoustic measurements, the probes are aligned within a vertical plane on the same face.

Fig. 1 - Schematic representation of the acoustic wave system used for field-testing of standing trees.

  Enlarge/Shrink   Download   Full Width  Open in Viewer

For this study, a 1.00-meter testing span was roughly centered at breast height. The lower probe was placed about 40-60 cm above the ground. In conjunction with the acoustic measurement, the diameter at breast height of each tree was measured according to the common practice in forestry measurement ([24]). Four readings in the same position were recorded for each tree. Stress wave times were then converted to mean acoustic velocity for each tree.

The modulus of elasticity equation

The next step was to calculate the modulus of elasticity, which is the most important parameter in assessing wood quality. The dynamic modulus of elasticity (MOEd - [33]) can be estimated from the velocity of acoustic waves (acoustic velocity, CT) passing through the wood, according to the following formula (eqn. 2):

where WD is the wood density, expressed in kg m^-3, and AV is the velocity in km s^-1 ([23], [4], [15]). Wood density was determined by volumetric methods, which required measuring the volume and mass of dry wood in the laboratory. The density was computed starting from basic density measured on a small wood core (roughly 10 × 10 mm, after removing the bark) taken from the same section of the stem where the NDT measurement was carried out. To compute the basic density, the oven-dry mass (MC0%) was divided by its green volume, and measured based on the Archimedes’ law using a beaker with distilled water placed on a lab electronic balance (0.001 g of resolution) set to zero; a mechanical arm connected to the needle able to correctly immerse the wood specimen in the water, the top of the wood being just below the meniscus, thus avoiding a significant immersion of the needle. A small specimen from the core was devoted to get the moisture content of the wood via gravimetric test ([32]). Although the specimens were green, they had been stored in distilled water for 48 hours in order to saturate the wood. From basic density raw data, by taking into account the shrinkage percentage (Sh% = 100 · [decrease in size/ original size], applying a shrinkage rate of 0.14 and 0.25% for every % of mc, respectively for radial and tangential directions - [28]) from fibre saturation point down to 12% MC, the normal density (kg m^-3 12%) was computed. The original velocity recorded on green timber was considered, as it was not corrected according to the moisture content, following the manufacturer’s specification.

Study sites and tree samples

Four research plots were selected in 60-year-old Calabrian pine plantations located at Varco San Mauro in the municipality of Rose (Cosenza Tyrrhenian Presila). The four plots were previously subjected to different thinning treatments (Fig. 2): control (T - no thinning, approx. 1300 trees ha^-1); light thinning (A - 800 trees ha^-1); intermediate thinning (B - 600 trees ha^-1); and heavy thinning (C - 400 trees ha^-1). Thinning treatments have been carried out at intervals of 8-10 years each other. Plots A, B, and C were subsequently targeted for final harvest and regeneration. The historical information and the stand structural characteristics of the plots are reported in Tab. 1.

Fig. 2 - The four treatments in the study area.

  Enlarge/Shrink   Download   Full Width  Open in Viewer

Field tree measurements and acoustic wave tests were conducted from October 2014 to February 2015. A total of 450 trees were tested and tagged with a unique number. The sampled trees were stratified in four classes based on diameter at breast height (DBH): 30, 35, 40 and 45 cm. Such classes were chosen as considered representative of the typical size for commercial trees.

Statistical analysis

Analysis of tree stress wave time and tree diameter at breast height was performed at both individual tree and site level. The tree diameter at breast height (DBH, cm) and tree density (d, number of trees ha^-1, varying across treatments) were selected as independent variables (two groups in mixed model), while MOEd (calculated from eqn. 2) was the dependent variable of the regression analysis (eqn. 3):

Analysis of variance (ANOVA) was conducted to test for differences in MOEd between DBH classes and thinning regimes. The post-hoc Tukey’s test (α=0.05) was applied to detect significant differences among group means.

The software packages SPSS^® ver. 20.0 (IBM, Armonk, NY, USA) and MS Excel^® 2012 (Microsoft, Redmond, WA, USA) were used for the analysis.

  Results 

The wood density computed from basic density measured on the collected specimens is reported in Tab. 2.

Mean Fakopp velocity did not significantly differ among treatments T, A and B (4.36 ± 0.56 SD, 4.57 ± 0.29 and 4.46 ± 0.31 km s^-1, respectively), whereas it was significantly lower in treatment C (3.98 ± 0.52 km s^-1 - Fig. 3, Tab. S1 in Supplementary material). The average stress wave time of trees for plot A (light thinning) was about 5% higher than for the control (T); in plot C (heavy thinning), the value was lower by about 9% (Tab. 3).

Fig. 3 - Minimum, maximum and mean values of velocity (km s^-1) recorded at the four study sites. Error bars represent the standard deviation. (T): control; (A): light thinning; (B): intermediate thinning; (C): heavy thinning.

  Enlarge/Shrink   Download   Full Width  Open in Viewer

MOEd ranged from 6.537 to 14.328 MPa (mean = 10.901 MPa) in plot A, while in plot B it varied from 6.631 to 12.931 MPa, with a mean of 10.352 MPa (Tab. 3). Statistical analysis revealed significant differences in MOEd between the stands with moderate thinning (A and B) and those with heavy thinning (C - Tab. S1 in Supplementary material). Moreover, the MOEd values estimated for treatment C were lower than those for the control plot (Tab. 3).

Among the different diameter classes, the highest values of MOEd were registered for classes 30 and 35 cm (Fig. 4), though the differences were not statistically significant (p>0.05). This suggests that the thinning treatment affects the wood quality of standing trees. Indeed, it is expected that, for trees of the same age, larger trees with higher growth rate generally produce wood of lower density and stiffness ([45], [14]). Our results confirmed that light and moderate thinning produced significant benefits in terms of wood quality. The reduced stand density of plot C promoted a faster radial growth, which led to a lower wood density of the mature trees.

Fig. 4 - Mean MOEd values of standing trees estimated in the different diameter classes. Error bars represent the standard deviation. (T): control; (A): light thinning; (B): intermediate thinning; (C): heavy thinning.

  Enlarge/Shrink   Download   Full Width  Open in Viewer

The analysis of variance confirmed that the average MOEd of standing trees was significantly influenced by DBH and thinning regimes (Tab. 4). The results of the Tukey’s test confirmed the existence of significant differences (p < 0.05) in MOEd values among treatments (Tab. S1 in Supplementary material).

The results of the regression analysis carried out are reported in Tab. 5. Both parameters may be considered as fairly good predictors of MOEd, as they account for a significant portion of the total variance (R² = 0.34). The regression model obtained was as follows (eqn. 4):

where DBH is the diameter at breast height (cm), and d is the stand density (n ha^-1).

  Discussion 

This study demonstrates that the effect of silvicultural practices on wood properties can be successfully identified based on stress wave MOEd. Indeed, thinning caused a significant variation in terms of MOEd among DBH tree classes (Fig. 4). Similarly, Todaro & Macchioni ([29]) registered a significant variation of wood properties caused by thinnings in a Douglas-fir stand in Southern Italy. These results confirms that the reduction of stand density by thinning may cause a faster radial growth of the remnant trees, which involves a reduction of wood quality in terms of wood density and stiffness.

Acoustic waves based techniques aimed to assess the intrinsic wood properties of standing trees have been applied in many studies worldwide ([22], [36], [12], [13], [9], [37], [16]). Wang ([43]) found that a range of wood and fibre properties can be predicted through a simple acoustic measurement of standing trees. Moreover, many studies reported that, for softwood trees of the same age, stress wave velocity is higher for trees with slower growth rates and narrower rings. The species tested included Sitka spruce (Picea sitchensis), western hemlock (Tsuga heterophylla), jack pine (Pinus banksiana), ponderosa pine (Pinus ponderosa), and radiata pine (Pinus radiata - [38]).

In this study, the repeated intensive thinnings in the plot C have reduced the within-stand competition, allowing a faster tree growth, and this may explain the relationship between the acoustic velocity and DBH in this stand. Zhang ([45]) and Todaro & Macchioni ([29]) observed that the bigger trees with higher growth rate within a stand generally produced wood of lower stiffness. Chauhan & Walker (2006) have shown that the increased diameter of fast-growing trees, with its second moment of inertia, entails a smaller need of high material stiffness, while for trees of the same age, slow growing trees tend to have higher stiffness, enhancing tree resistance to various stresses like wind, crown weight, etc. Furthermore, younger wood has lower specific gravity and stiffness, while fast-growing trees has a larger diameter core of juvenile wood ([5]). All the above evidences may explain how wood quality and intrinsic wood properties are generally affected by silvicultural practices, especially those affecting stand density ([8], [18], [30], [7]).

The lower wave velocity with increased DBH observed in this study appears to be associated with a higher growth rate of trees, which adversely affects specific gravity as well as the strength and stiffness of the wood. The fairly high variability detected among trees of the same DBH class (age) across different sites may reflect the natural variation between individual trees as well as the thinning effect ([35]).

Wang ([36]) examined the effect of thinning treatments on both acoustic and static bending properties of young western hemlock and Sitka spruce trees from seven sites in southeast Alaska, finding that most trees with higher acoustic velocity and stiffness were growing in unthinned control stands and in stands subject to light and medium thinning, whereas the lowest values of acoustic velocity were found in heavily thinned stand.

The technology based on acoutic waves used in this study may also be applied to determine how environmental conditions and silvicultural treatments affect wood and fibre properties, thereby the most effective treatment can be selected for the desired fibre quality in future plantations ([43]). Since tree DBH is easy to measure, appropriate multivariate regression models using velocity and DBH as predictors can be established to estimate wood quality of standing trees.

  Conclusions 

A strong positive association was observed between MOEd of standing trees, their DBH and the stand thinning regime (in terms of stand density) in Calabrian pine plantations in southern Italy. Moreover, our results showed that tree diameter has a significant influence on acoustic wave measurements and a significant relationship with MOEd.

Acoustics is a practical, low-cost method to assess the wood quality of standing trees ([17]). Our results confirm that non-destructive stress wave techniques may be used to track wood property changes in standing trees, thus helping forest managers and industries in managing forest plantations to meet the desired wood and fibre qualities ([37]). Moreover, the information obtained by acoustic techniques may be relevant for tree breeding, pre-harvest assessment, and decision support at time of thinning. With continuous advances and refinements, this technology could assist in managing wood quality, assessing forest value, and improving the timber quality of future plantations ([41]).

  Acknowledgments 

This study is a part of the project “ALForLab” (PON03PE_00024_1) co-funded by the National Operational Programme for Research and Competitiveness (PON R&C) 2007-2013, through the European Regional Development Fund (ERDF) and national resource (Revolving Fund -Cohesion Action Plan [CAP] MIUR).

  References

  Supplementary Material

Tab. S1 - Results of the Tukey’ test on mean MOEd values registered in the different study plots.