Between-tree variation in stem volume, wood density, fibre length and Kraft pulping properties of Eucalyptus globulus and the utility of field- portable NIR Spectroscopy and wood cores in evaluating pulpwood quality properties standing trees

摘要

Pulpwood quality characteristics of 75 individual trees of 10 to 11-year-old Eucalyptus globulus were evaluated to assess their suitability for inclusion in a clonal tree improvement program. The trees were phenotypically selected for good growth and form from three commercial plantations in contrasting rainfall zones in south-western Western Australia. All sampled trees were assessed for merchantable tree volume, heartwood proportion, bark proportion, basic density and Kraft pulping properties. Simulated wood core samples extracted at breast height from each tree were evaluated for basic density and near infrared spectroscopy (NIRS) predicted pulp yield. The application of a field-portable Polychromix Phazir NIRS system as a tool for rapid determination of pulp yield in standing trees was tested as was the reliability of Pilodyn penetration in predicting basic density at breast height.rnThe selected trees showed a considerable degree of variation within and between sites for all the evaluated properties, indicating a potential for tree improvement through cloning of individual trees matched to specific sites. Across the three sites, individual tree values ranged from 451 to 658 kg nr3 for chip basic density; 0.78 to 1.12 mm for fibre length; 47.8 to 58.8% for screened pulp yield at Kapparnnumber 18, with an active alkali from 10.5 to 15.8% and 264 to 371 kg m-3 for pulp-wood productivity.rnCore samples were better predictors of whole tree chip basic density compared to Pilodyn penetration, with respective coefficients of determination of 0.83 and 0.61. The NIRS spectra taken from bark free surfaces on standing trees using the field-portable Polychromix Phazir produced moderately high predictions of total pulp yield (R2 = 0.70 and SEP = 0.85%) but poorer predictive capacity for screened pulp yield (R2 = 0.61 and SEP = 1.73%). In contrast, the laboratory-based NIRS method using wood core samples was found to be a superior predictor with R2 values of 0.85 and 0.88 respectively for screened and total pulp yield. The respective standard errors of prediction were 1.35 and 1.00%.