Unexplained site variance in machine strength grading of Norway spruce structural timber

The aim of the study was to assess site effects in terms of unexplained site variance in machine strength grading of Norway spruce structural timber. The site effects were estimated for grading based on axial resonance frequency and timber length, and for grading based on dynamic modulus of elasticity. Timber was collected from 14 sites in Norway, and linear mixed models were developed based on 1188 boards. The study showed that strength grading based on axial resonance frequency and timber length leaves out effects of site that are related to altitude, latitude and site index. The variance could be reduced when the grading was based on dynamic modulus of elasticity. For both grading methods, the site effects were smaller for bending strength than for modulus of elasticity and density. Major parts of the site effects were explained by mass density, and simulations showed that it is possible to fulfil the requirements of the strength classes with a higher yield when the sorting is based on a combination of exclusion by mass density and exclusion by the frequency-based indicating property.