Characterisation of Eucalyptus nitens plantations for veneer production

Eucalyptus nitens is widely planted in temperate-cold regions around the world, mainly for pulpwood production. However there is interest in its use for veneerbased engineered wood products to increase the value and market opportunities for the E. nitens plantation resource. This study examined such opportunities by studying factors which affect veneer recovery as well as those which influence wood properties important for veneer quality. This thesis had four main experiments. The first experiment examined log-end splitting with respect to log storage, steaming, plantation site and tree position. Log-end splitting is a major defect of Eucalyptus logs which has the potential to significantly reduce the recovery of veneer. Log-end splitting was assessed immediately post-felling, after transport and storage and immediately prior to veneer peeling, using upper and lower logs from each of 41 trees from three 20-22 year old plantations. The study found that log-end splitting varied across sites, was higher in upper than lower logs, and increased with time in storage particularly in the upper log. In the system studied, log steaming did not significantly affect the severity of log-end splitting. The three remaining studies dealt with the assessment of three important wood properties: density, microfibril angle and modulus of elasticity. SilviScan was the main approach used for assessing radial and site variation in these characteristics. However, in the second study an attempt was made to develop more rapid and cost effective approaches using near-infrared spectroscopy. Radial scans of breast-height samples from 86 trees and three sites were used to successfully develop nearinfrared calibrations of these wood properties against 1 mm resolution SilviScan measurements. However, the capacity of these models to predict wood properties when applied to independent sites was mixed and generally poor. SilviScan measurements were therefore used throughout the remainder of the thesis where comparisons across sites were the main focus. The third study examined radial change in density, microfibril angle and modulus of elasticity. While radial change in these veneer-critical wood properties is known to occur, the extent to which patterns vary, both between sites and trees within sites, is poorly understood in Eucalyptus. Radial models were developed to determine the trends in these wood properties from pith to cambium using samples from 2.5 m above ground level taken from forty-one 20-22 year old trees from three E. nitens plantations. Simple linear regression models were used to model density, while non-linear functions were used to model the radial variation of microfibril angle (asymptotic function) and modulus of elasticity (sigmoidal function), against both cambial age and percentage area from pith. The radial trends from pith to cambium in density, microfibril angle and modulus of elasticity found in this study were clear and matched the general trends in other studies of Eucalyptus. However, significant differences were obtained in the pattern of radial variation in these wood properties both among trees within sites as well as between sites. The radial patterns of change were similar regardless of whether they were assessed based on cambial age or percentage of the cross-sectional area and site rankings were the same. Lastly, the site variation in density, microfibril angle and modulus of elasticity across the Tasmanian E. nitens plantation estate was modelled. This study was based on area-weighted tree means obtained from radial SilviScan data. Three trees from each of 46 sites were used to develop the models and those from an additional 13 sites used as a validation data set. Site-level averages for the three wood properties were modelled using various forest, environmental and climatic variables as explanatory variables. Stepwise backward regression was used to select the most parsimonious linear model for each wood property. The final models included only plantation age and annual precipitation for all three wood properties as well as elevation for density. These models were well validated and used to map predicted spatial variation in these veneer-critical wood properties across the Tasmanian plantation estate to aid resource characterisation and forest management. A key result to emerge from this thesis was the importance of site-level factors on log traits and wood properties likely to impact the quantity and quality of veneer from Tasmanian E. nitens plantations. This thesis was able to characterise and predict variation in these wood properties, both within individual trees and across the Tasmanian plantation estate. These results have direct management implications, allowing quantification of the impact of different growing conditions on the wood properties of harvested logs and therefore on the potential veneer recovery. Coupled with growth and economic models, these results will assist in optimisation of available forestry resources and future planning