Growth stresses in eucalypts: evaluation and development of measurement techniques

• Chinese Academy of Forestry, China
• Queensland Forest Research Institute, Australia

Strain forces generated in the release of growth stress in eucalypt species can result in serious distortion and quality degradation in sawn timber. They are a major constraint to use of plantation-grown eucalypts for high-value products such as construction, joinery and furniture.
Many developing countries have now established large areas of eucalypt plantations that are in use for fuelwood, pulp and poles. These plantations could yield higher-value products if wood quality problems including growth stresses could be solved. The plantation industry needs a reliable, efficient and non-destructive method of predicting growth stress in standing trees and in logs prior to processing, in place of existing methods that are slow and cumbersome and produce less than satisfactory results.

This project aimed to develop methods whereby research and industry could rapidly and accurately screen large numbers of trees or logs and reliably classify them for severity of growth stress.

Evidence from earlier research suggests that growth stress problems can be successfully managed through a combination of tree breeding and silviculture, and technological improvements on log processing through sawing equipment design and sawing strategies. In this project the research team sought to establish an efficient and reliable measurement system to quantify growth strains in standing trees and logs. A literature review was undertaken to evaluate the utility characteristics of existing methods developed in Australia and in other countries. It compared the Nicholson method, the CIRAD method and the strain gauge method.
In the laboratory four groups of back-sawn regrowth Eucalyptus delegatensis specimens (three specimens per group) were subjected to a range of target tensile loads. The load was maintained after reaching the target level and a 20 mm hole was drilled. Grooves were cut near the measurement points after the specimens were unloaded. Measurements were taken at various stages of the tests to compare the three methods.
The team carried out field comparison of the three methods on 119 E. globulus trees (7-9 years old), 28 Corymbia citriodora subsp. variegata trees (3.5 years old) and 35 E. cloeziana trees (6.5 years old) to determine the relative suitability of existing methods for grouping trees into broad classes of growth strain levels. Three growth strain measurements were made on each tree within a 500 mm vertical space, using the three methods.
The team also explored faster and less destructive alternatives to measure or predict growth strains in trees or logs, e.g. near infra-red (NIR) analysis.

Based on the evaluation of the utility characteristics, the CIRAD-fort method was recommended for routine field growth strain assessment and the strain gauge method for scientific research. Strain correlated significantly between the three methods tested. The strength of the relationships indicates that strain by one method could be estimated from another with reasonable confidence. This in turn indicates that the three methods are compatible in segregating trees into similar broad strain classes.

In the investigation of faster and less destructive alternatives to measure or predict growth strains in trees or logs the team collected high resolution SilviScan data, NIR spectrum and digital images from E. globulus wood specimens that were closely matched with the locations where growth strain was measured. Drilling resistance data were also obtained using a digital microprobe (DmP) at zones beside and above growth strain measurement locations in 22 E. globulus trees that represented a wide range of the CIRAD-method displacement measurements.

Based on the results, SilviScan was deemed capable of prediction of growth strain near tree surface with a moderate level of accuracy. As the use of SilviScan technology is becoming more popular, people may use the methods in this report to predict surface growth strain of E. globulus from microfibril angle (MFA) when accuracy is not critical or growth strain measurements tools are unavailable.

Results indicated the potential of NIR analysis in predicting the longitudinal strain. Further research is highly recommended to improve the calibration, since NIR analysis has the advantage that its specimen collection causes little injury to tree stems and also scan spectra may be used to predict a number of other wood properties.

The digital image correlation method was not found reliable in this study because of poor quality images, largely due to inconsistent photograph settings when outside the laboratory environment. The difficulty of obtaining quality digital photos means that the method does not appear suitable for routine growth strain assessment in the field. It could be useful in fundamental research on growth stress, but much more research is needed before this method will be ready for growth strain measurement.

The DmP showed potential for an approximate, quick and low-stem-injury prediction method of growth strain in standing trees. It is portable, user-friendly, low-effort, and cheap to operate. Although it is not designed for accurate predictions of growth strain, it is still worth further investigation as a means of rapid and coarse segregation of lowly stressed trees for sawlogs and for developing procedures for field measurement.