Mike Carson, the Business Development Manager at Gemnetics, is writing a short article series about breeding topics arising from the perspective that he and Sue Carson have developed during their (combined) 79 years involvement in forest tree breeding.

Mike has used Gemview on a daily basis to track his trial breeding data and view trends and relationships in order to increase the productivity of the Forest Genetics breeding and clonal deployment program.

Mike's articles will be published on our Gem news page. If you would like to contact Mike to comment or share your own insights, please contact him through our contact us page, or directly through Mike's Linkedin profile.

The Acid Test of tree breeding programs is to demonstrate genetic gains in forest plantations

Thanks to the successful efforts of past and present tree breeders, most forest owners and managers believe that breeding programs are delivering genetic gains in productivity and profitability to their forest estates. In fact, most of the countries with a long experience of growing exotic tree species on a plantation scale, including for example Brazil, South Africa, Chile, New Zealand, and Australia, have recorded the evidence for genetic gains from harvests of full-rotation stands of improved trees.

However, such historic evidence is not in itself proof of continuing gains from tree breeding. Forest owners and managers require breeders to provide hard evidence that the progenies and clones being deployed currently are delivering the expected gains. This can be more easily said than done, and too often breeders are tempted to duck the question by falling back on gain projections based only on progeny trial results, with sometimes mixed outcomes.

Tree breeders typically utilise a range of trial designs to evaluate candidate selections, and to provide a limited attempt at validation of productivity and other gains. This figure illustrates a typical approach, in which single-tree-plot trials are used for initial evaluation of progenies and/or clones, and a combination of row-plot and large-block designs are employed to validate the performance of a selected set of genotypes.

This approach has proven very effective in selection of candidate genotypes for breeding and for deployment, but it often falls short of providing growers with the level of confidence they need to support increased use of high-gain genotypes on their own estates, particularly when additional costs of breeding and propagation enter the picture. Tree breeding trials are seldom replicated widely and numerously across forest estates, so cannot hope to be fully representative of all the forest site types and management regimes present. Also, foresters have a natural suspicion that trials may receive special treatment in the hands of breeders and researchers, and wish to see the evidence of gain under operational conditions.

This is the ‘easier said than done’ challenge, since stand-level comparisons offer none of the experimental design benefits provided by well-replicated, randomised trials. Trial entries are not all grown under controlled conditions or with appropriate genetic controls, so that statistically-significant differences and rankings for performance traits can not always be reliably estimated. Inevitably, any stand-level comparisons among different genetic treestocks will be confounded with various operational effects, including:

• Differential nursery effects on growth.

• Differential site preparation and establishment effects, including different planters.

• Differences in site slope and aspect, etc.

Thankfully, meaningful stand-level comparisons can sometimes be achieved, as for example in a recent B.For.Sci. (Hons) dissertation at the University of Canterbury, School of Forestry, by Acacia Farmery. Acacia’s project was aimed at comparing the performance of radiata pine clonal varieties developed by Forest Genetics Ltd with ‘best-available’ control-pollinated orchard seedlots in the N.Z. north island east coast forests managed by Pan Pac Forests Ltd. A key objective of Acacia’s project was to compare trial-based ranking information provided by Forest Genetics with her stand-level measurements in the Pan Pac plantations, as illustrated in Figure 2, below.

Acacia’s conclusions were that “Overall, the clones performed well…. and performed the same in Pan Pac Forest Products’ production forests as they did in Forest Genetics’ trials”, and “The production plantings of clones in Pan Pac’s forests have been found to perform very similarly to their performance in Forest Genetics’ trials. This is very beneficial to Pan Pac as it indicates that they can rely on the ratings derived by Forest Genetics as a measure of the performance of clones on their forest estate.”

Table 3: Each clones performance in the study compared to trial performance for five traits.

NB: Comments in each field of the table refer to the level of agreement between the performance estimates provided by Forest Genetics from BLUP analyses across their country-wide trial network, and the stand-level measurements in Pan Pac Forests’ estate obtained in the project by Acacia Farmery.