Application of non-destructive evaluation techniques to the prediction of solid-wood suitability of plantation-grown Eucalyptus nitens logs

Non-destructive evaluation measurements of longitudinal growth strain (LGS) and acoustic wave velocity (A WV) were collected on 22-year-old Eucalyptus nitens trees from a plantation thinning trial in Tasmania. A total of 81 trees from five residual stocking treatments including an unthinned control were assessed. LGS was measured at breast height at four aspects per tree. Standing tree-A WV was determined at two opposite aspects per tree. Log-A WV was collected on the 5.7 m-long bush-logs and on 162 lower and upper 2.7 m-long sawlogs. Sawlogs from 41 trees with diameter over bark (DBHOB) under 43 cm were back-sawn, and those from 40 trees with DBHOB over 43 cm were quarter-sawn. Wood shrinkage and basic density were measured on 25x25x40 mm wood blocks cut from wood disks immediately above the first and second sawlogs of the back-sawn trees at heights of approximately 3m and 6 m above ground. Three blocks were cut at each height, at radial distances 25%, 50% and 75% from pith to cambium. Linear mixed models and multiple linear regressions (MLR) were applied to examine: a) effects of stocking, DBHOB and other factors on LGS, AWV and wood block traits; and b) the ability to predict sawing traits (log end-splitting; sawn-board distortion, shrinkage and checking propensity), product recoveries and value; and wood mechanical properties (timber stiffness, strength and hardness). Stocking was significantly positively related only to log-A WV. DBHOB was negatively related to standing tree-A WV and log-A WV, and positively related to log end-splitting, which was also higher in the upper sawlogs and positively related to LGS and log-A WV. Measurement aspect affected both LGS and AWV with higher levels on the north-west and sout-west aspects, facing the prevailing wind direction. Basic density increased with stem height and from pith to cambium,'whereas tangential and radial shrinkage and collapse were lower at the upper sampling point, while they also increased from pith to cambium. LGS was positively associated with sawn-board distortion and losses of board volume from end-docking. Standing tree-A WV, log-A WV and basic density were positively related to wood mechanical properties. Log-A WV was also positively related to board bow, shrinkage, select recovery and product value. Levels of shrinkage and collapse in wood-blocks were related to board shrinkage (positive), internal checking (positive), select recovery (negative) and product value (negative). Overall LGS, AWV and block traits had a modest capability to predict sawlog performance, explaining respectively up to 20%, 17% and 6% of additional variance in log end-splitting, timber stiffness, and product value, in those MLR that already incorporated significant explanatory factors such as sawing method, DBHOB and log position.