Evaluation of non-destructive methods of measuring growth stress in Eucalyptus globulus: relationships between strain, wood properties and stress

High levels of growth stress are implicated in causing end splitting of logs, deflection during sawing and deformation of boards as stresses are released during sawing operations. Level of stress is a function of strain and the elastic modulus of the wood (MOE). Levels of peripheral strain can be measured on standing trees and, if the MOE is known, stresses can be estimated. The validity of using peripheral strain measurements relies on underlying theoretical models that relate strain to expected patterns of stress distribution and levels of board deflection. This study evaluates these theoretical relationships by determining relationships of stress and strain with board deflection, end splitting and a range of wood properties. Peripheral strain levels were extremely variable within the bottom log and little evidence was found for consistent patterns of variation, although measurements generally increased with increasing height above ground. Sampling on two sides of the standing tree at breast height appeared to be a suitable strategy, with the mean for these strain readings having a correlation (r) of 0.86 with the average strain in the bottom log. Growth strain was not a reliable predictor of board deflection and cannot be recommended as a non-destructive sampling method. Overall there was a poor relationship between growth strain and board deflection. No consistent relationships were found between a range of wood properties and growth strain or board deflection across both sites. Stress levels were calculated for each tree as the product of growth strain and modulus of elasticity and the relationship between calculated stress and mean board deflection determined. No relationship was found at either site with correlations being very close to zero. The underlying theoretical relationships between stress and strain were examined and several questions raised about the validity of such models.