RPBC's breeding objective (or breeding 'goal') is to breed trees that have good growth rates, tree form and wood properties at harvest. The breeding objective assumes that a stand at harvest age is valued based on the volume and the quality of the wood produced. RPBC's breeding objective includes volume, branching index, wood density and stiffness at harvest. While a breeding objective can be specific to only one end use, RPBC recognizes that shareholders have a variety of end uses in mind. Developing an objective in a breeding program that is all things to all foresters is not possible. As such, RPBC has elected to focus on a general-purpose objective but has put an emphasis on volume growth and structural timber.

TopTree ranks trees on their ability to meet this objective using selection-age traits that are combined into a RPBC selection index. In TopTree, this is referred to as the ‘Baseline’ index and values are in $NPV. Selection indices can be used in breeding to improve multiple traits simultaneously. There are currently five traits included in the RPBC index measured at selection age (5 to 8 years old) that have been identified as having a direct economic value to the Australasian forestry sector at final harvest age (25 to 35 years old). These traits can be categorised as either 'Production' traits (including tree form) or ‘Wood Property’ traits. Estimated breeding values of these five traits are multiplied by their respective economic weights to produce the selection index.

The traits are weighted to maximise response in the RPBC breeding objective. The index considers the economic importance, heritability and correlations among traits for the purpose of maximising the correlation between the selection index and breeding objective.

The effective emphasis of the five individual traits within the RPBC index are divided between 61% production efficiency traits and 38% wood property traits. TopTree allows users to alter the importance of these traits in the index either at harvest age (the economic value of the breeding objective trait) or the economic weight on the selection index. For example, if users want to put more emphasis on growth, then more weight can be applied on volume at harvest age (the breeding objective) or at selection age (selection index). Note that because relationships between traits are accounted for in the calculations, e.g., the negative correlation between growth and density, the customisation of one trait will likely have an impact on the other traits also.  Modifications in the future may include Dothistroma resistance as a ‘Production’ trait, but estimated breeding values for Dothistroma are provided separately and can be used to help with the selection of germplasm if required.

Accuracy of an EBV in the correlation between true and estimated breeding value. The accuracy is between 0 and 1 (or 0% and 100%). Accuracy is related to the precision of the predicted breeding value and is higher (more precise) when more information is used, e.g. on own performance, or on relatives and progeny. In the extreme case of an ortet having no information the accuracy of a breeding value is 0, and with a very large amount of information the accuracy will approach 1. The accuracy is higher for traits with a higher heritability, but as more information is used the effect of heritability becomes less. Also, the accuracy of a parent average (e.g. the average EBV from the cross between two ortets) depends on the parent EBV accuracy and not on heritability (but with low heritability it will be more difficult for a parent to achieve a certain accuracy). Reliability is sometimes used instead of accuracy and is simply [accuracy]².

In the example shown below, ortet ID 313003 has been tested as a clone so it has a good information available on its own performance (as it has been replicated across several test sites) as well as information sourced from relatives. Depending on the amount of information available, a trait such as corewood density tends to have a higher accuracy than DBH because its heritability is higher, as is the case below.

For deployment, we are interested in multiplying either the selected genotype (clonal deployment) or its progeny (via seed production). For the individual genotype, if we have low accuracy, the performance of the individual tree could be a long way from its predicted value (+ve or -ve). In contrast, only small changes from predicted values are expected at high accuracies. In deploying germplasm with low accuracy, the risk of commercial plantations not performing as expected increases. However, younger ortets tend to have lower accuracies. In a typical breeding programme, the average estimated breeding value should increase over time and is accordingly higher for younger ortets. Germplasm choice can therefore come down to the forester’s appetite for risk and approaches and capacity to spread that risk when deploying germplasm to the production forest. In practice, production clones deployed for clonal forestry are generally well tested and therefore confidence in their performance should be reasonably high. The clonal production company should have comprehensive information available on the performance of their production clones. In commercial seed production, parents tend not to be introduced into a control pollinated orchard until seed orchardists have confidence in the EBVs of the individual. Further, control- or open-pollinated seed mixes usually involve several parents so the changes from predicted values in any individual genotype should not noticeably alter the overall average merit of the seedlot.

Selection Traits:

• Diameter at Breast Height (DBH): Included as the principal production trait. Measured in mm at 1.4m above ground level.
• Branch Frequency: A production trait, branching is based on a 1 (uni-nodal) - 6 (multi-nodal) score of branch cluster frequency.
• Straightness: A production trait, stem straightness is based on a 1 (crooked) - 6 (very straight) score.
• Corewood Density: Included as a wood property trait, Corewood density at selection age in kg/m3 is measured from wood cores or using an IML Resistograph tool on standing trees.
• MoE: The focal wood property trait, Modulus of Elasticity is predicted by acoustic velocity (GigaPascals, Gpa) using either a Tree Tap, ST300 or IML Hammer tool on standing trees.

Harvest Traits

• Volume: Economic value ($NPV) of harvest volume (kg m^-3)represents the additional economic return per marginal unit improvement in harvest volume.  Volume is 44% of the total emphasis of the RPBC (baseline) breeding objective.
• Branching index (BIX): Economic value ($NPV) of BIX at harvest (BIX is the average size of four branches per 5.5 m log, using the largest branch in each quadrant) represents the additional economic return per marginal unit improvement in BIX. BIX is 17% of the total emphasis of the RPBC (baseline) breeding objective.
• Density: Economic value ($NPV) of wood density (kg m^-3) at harvest represents the additional economic return per marginal unit improvement in wood density. Density is 12% of the total emphasis of the RPBC (baseline) breeding objective.
• Stiffness: Economic value ($NPV) of wood stiffness (Gpa) at harvest represents the additional economic return per marginal unit improvement in wood stiffness. Stiffness is 26% of the total emphasis of the RPBC (baseline) breeding objective.