An investigation of Eucalyptus nitens plantation wood quality to develop a segregation approach for diversifying resource utilisation

This thesis investigated wood properties and timber characteristics of plantation Eucalyptus nitens (E. nitens H. Deane and Maiden) grown primarily for fibre production to develop a segregation methodology and inform grading strategies for diversifying and improving the utilisation of this resource. Eucalyptus species are planted worldwide in monocultures, occupying 26% of the total planted area across the globe. Trees of this genus are grown under short rotation regimes with limited silvicultural interventions. Although mostly grown to produce fibre for the pulp and paper industry, timber from these fast-growing trees can be used for construction products, either as solid timber or in composite mass-laminated structures. The capacity to efficiently utilise fast-growing Eucalyptus wood for products other than pulp depends on the ability to locate and extract logs of the desired quality from the plantations, and to deliver them to the optimal processing stream. However, to achieve this, appropriate sorting, grading and processing of the raw wood resource and timber products is needed. The objective of this thesis was therefore to examine quality traits of fast-growing E. nitens wood at different scales, or production points, from the plantation setting to trees, logs, and timber products, to develop a segregation methodology and inform grading strategies which would improve and diversify the utilisation of the resource. The experimental studies conducted for this thesis were designed with a holistic view of the forest and wood products production chain. Hence, the investigation of wood quality traits of fast-growing E. nitens took place at the tree, log and sawn timber scale, and consideration was given to the relationships and impact of those traits on the recovery of high-quality products. The first experimental study, Chapter 2, describes a large-scale study characterising wood quality of the E. nitens estate planted across Tasmania, Australia. The study aimed to identify site and tree factors influencing density and stiffness of E. nitens wood to develop predictive models of these two important structural properties. The results indicate that density and stiffness vary considerably between and within plantations of E. nitens, depending on a sites' environmental and climatic conditions. Significant differences in tree density and stiffness were found between and within sites, with density and stiffness variation mostly due to differences between trees within a site (respectively of 70% and 60%). Site factors including extremes in temperature and precipitation, as well as site elevation, influenced both tree basic density and average basic density on the site, while stiffness at the tree and site level was negatively influenced by the site index, which describes site productivity. Tree factors associated with growth influenced these wood properties, with tree diameter positively influencing tree basic density and tree height, slenderness and basic density positively influencing tree stiffness. When employed in prediction models, these variables predicted density and stiffness of both individual trees and sites of a previously unseen dataset. Basic density was better predicted at the site level rather than at the tree level, explaining as much as 73% of site basic density variability, while stiffness was better predicted at the tree level, where it explained 59% of the total stiffness variability of the trees. These results can be used to inform the planning of tree plantings in suitable areas, and in the preparation of harvesting plans for required product types. Sites can be screened for their average density and chosen for sawlog production and within sites individual tree stiffness can be predicted and used as a way to segregate trees of higher stiffness. Selection of trees and segregation of logs for structural products require knowledge on wood stiffness, which can be obtained through the use of fast and reliable non-destructive testing techniques, such as acoustic wave velocity (AWV) screening. The second experimental study of this thesis, Chapter 3, examines the relationship among tree and log characteristics and sawn board structural grades to develop a segregation methodology for plantation E. nitens trees and logs. Results of this study show that modelling sawn timber stiffness with tree and log variables yielded modest coefficients of determination and hence do not support the determination of AWV segregation thresholds in a straightforward manner. Classification techniques, part of machine learning models, facilitated the identification of classification variables and could identify AWV thresholds on trees and logs for timber segregation. The use of such thresholds in the development of a segregation methodology allowed the allocation of trees or logs to categories suited for different product classes or types. In this study, different segregation scenarios were compared, utilising thresholds of AWV and log position for trees and logs and considering segregation only on trees, logs, or on both. It was found that incorporating the use of AWV thresholds to batch logs into different classes yields the best outcome in maximising production of high-quality board grades and logs for pulp. Operational constraints as well as the structure of the supply chain will likely determine which segregation methodology is best suited to a particular situation, which can also be directed a-priori by the requests of the forest products market. Although stiffness in logs and timber products is a highly sought-after quality, other log and timber characteristics determine the processing, recovery and quality of the final timber products. The third experimental study of this thesis, Chapter 4, aimed to investigate characteristics and quality traits of plantation E. nitens logs and explored how these traits impacted the recovery of sawn timber. Significant differences in log traits including volume, taper, log-end splits, density and stiffness were found among logs of different positions in the stem, although only log volume, taper, log splits and log stiffness impacted sawn timber recovery rates and characteristics. Log position influenced the number of knots in the sawn timber, as well as its density and stiffness; a larger number of knots was found in boards processed from upper logs, which were also denser and stiffer. Other end-use-related characteristics, such as dimensional distortions and sawn timber splits, were not significantly different among log positions. These results show that, when considering log traits for the development of grading standards, only those predominantly impacting sawn timber volume and value recovery should be used to batch logs in quality classes. Among them, log stiffness should be used as an indicator of timber quality and employed into grading standards to classify and batch logs destined for structural products. The rapid estimation of timber properties, and in particular stiffness, through the use of AWV screening can aid the identification of products of higher quality at the earliest stages of processing and be used for grading of the final timber material. The aim of the fourth experimental chapter of this thesis, Chapter 5, was to investigate the suitability of AWV screening to improve the segregation and grading of structural boards sourced from plantation E. nitens logs. Sawn board stiffness was influenced by log position in the stem and by processing stage, though no interaction was found between the two. Moreover, the stiffness tested after sawmilling was significantly correlated with that of the final dried and dressed timber (R\\(^2\\) of 60%), showing that early segregation based on log position and board stiffness tested through AWV screening after milling can aid in obtaining sawn timber of the required quality. This chapter also explored the ability of the traditional Visual-Stress Grading (VSG) and the novel AWV screening in facilitating the grading of sawn timber. It was found that grading via AWV leads to a much lower error in board grading (43.3%) in respect to VSG grading (82.5%), supporting the possibility to employ this technique to place structural boards of plantation E. nitens into stiffness classes, and rapidly screen the quality of the sawn material. Sawn timber products used for mass-laminated structures are required to have properties that satisfy end-user requirements. Fast-growing plantation timber may have low density and structural properties, especially if not segregated and screen for quality earlier in the production process. This would hinder the use of this resource for structural products requiring satisfactory density, stiffness and strength. The fifth experimental chapter of this thesis, Chapter 6, examined the potential to improve important wood properties through the application of Thermo-Hydro Mechanical (THM) treatments and compared the effects of such treatments in E. nitens and other two timbers. Results indicated that THM treatments can be used on plantation E. nitens timber, significantly increasing its basic density (improvement of 53% in density post-treatment). The treatment also led to an increase in other structural properties such as stiffness, strength and hardness. Another plantation species, the softwood P. radiata, did not perform as well as E. nitens, especially in terms of recovery of the original size and swelling after treatment, while THM-treated native Tasmanian Oak timber had similar increases in density, stiffness and hardness, indicating that wood modification treatments may be best used on hardwoods. The findings of this thesis show that the variability of important wood quality traits in E. nitens plantation trees can be used for the segregation of the resource at different scales and across specific processing stages. At the landscape scale environmental and climatic factors such as site elevation, rainfall, ...