Presentation at Subtropical Forest Research Institute, Chinese Academy of Forestry, August 19, 2010
Topics to be presented
Variation in fertility Seed orchard turnover Optimal number of clones Prediction of fertility prior to orchard establishment Deployment of clones to a seed orchard Thinning of seed orchards using the linear deployment
algorithm Biological seed production Impact of growth and seed related characters on seed
procurement
Material and Methods
Pinus sylvestris main tree species Two types of material
Seed orchard data, collected by the authors and/or extracted and analysed from published works
Models developed using real data Calculation of variance components using
ANOVA ASReml
Microsoft Excel spreadsheets used for formulation of models and Excel-tool Solver for optimizing
Mathematical framework
N
iipN
1
2
The sibling coefficient, , expresses the probability that successful gametes (sibs) will originate from the same parent compared to the case with no differences in parental fertility
N is the census number of the parents
pi is the probability that a gene in the offspringoriginates from parent i
Mathematical framework, cont’d
1)1(2
N
NCV
The relationship between sibling coefficient and coefficient of variation (CV)
≥1; if =1, all individuals have the same fertility; if =2, it means that the probability that tworandomly drawn successful gametes share thesame parent is twice that where fertilities are equal across the population
Variation in fertility
Why focus on female fertility? Seeds are the source of income from a
seed orchard Cones and seeds are used to audit the
operation About 2/3 of the tree improvement effect
comes from the seed parents as half of the pollen parents are outside the orchard
Seeds from a known tree can be harvested and counted, which can be considered as the exact number of successful female gametes of the parent
Seed orchard (cf Table 2)
Character
Variance component among clones a single year
Variance component among clones over years
Variance component of clone-year interaction (years)
Variance component among ramets within clone and year
Heritability(broad sense for individual year)
Askerud Seeds/ ramet 3813c 1.38 2318 1.23 1495 (3) 1856 0.673
Långtora Seeds/ ramet 0b 1.00 * * 4775 0.000
Lustnäset Seeds/ ramet 5830c 1.58 4751 1.48 1079 (2) 3302 0.638
RobertsforsCones/ ramet
2455 1.25 * * 2029 0.548
Skaholma Seeds/ ramet 5466 1.55 * * 1946 0.737
SävarCones/ rameta
1258 1.13 * * 2039 0.382
GnievkovoCones/ ramet
313c 1.03 149 1.01 164 (2) 976 0.243
NebraskaCones/ ramet
10271c 2.03 3242 1.32 7029 (2) 2896 0.780
ViitaselkiCones/ ramet
* 2824 1.28 * 940
VilhelminmäkiFemale strobili/ ramet
2541c 1.25 1970 1.20 571 (2) 5607 0.312
VilhelminmäkiCones/ ramet
* 2722 1.27 * 5281
Average all SOs 3550 1.35 2568 1.26 2068 2722 0.479
Average for SO with clone-year interaction
43861.437
2500 1.250 2068 2893 0.521
Seed orchard turnover
A model was developed to evaluate the benefit of various options. Options and variables included in the model are: Seed orchard size Planting density Type of orchard material (grafts, cuttings) Establishment and management costs Cone harvest and seed processing costs Development over time of the seed orchard crop Rate of genetic progress in long-term-breeding
Seed orchard turnover, cont’d
Genetic penalty, representing the gain differential between the seed orchard and a hypothetical new orchard incorporating the latest genetic progress in the breeding population
Impact of pollen contamination How the genetic quality influences the value of the
seeds The orchards productive lifespan
Seed orchard turnover, cont’d
Optimal rotation age for Pinus sylvestris orchards is suggested to be around 30 years Cone harvest starts at age 8
For Picea abies 40 years is optimal Cone harvest starts at
age 15
Optimal number of clones
Maximize a goodness criterion (“benefit”) for orchards. It’s a function of: # of tested genotypes available for selection and
planted in seed orchard The contribution of pollen from:
The ramet itself The closest neighbors The rest of the orchard and contamination
Variation among genotypes for fertility Frequency of selfing
Optimal number of clones, cont’d
Production of selfed genotypes Gene diversity (=status number) Influence of contamination Genetic variation among candidates Correlation between selection criterion (e.g. height
in progeny test) and value for forestry (e.g. production in forests from the orchard)
# of clones harvested
Optimal number of clones, cont’d
Optimum # of clones in Pinus sylvestris is suggested to be 16, assuming ψ=2
If ψ=1.24 as mentioned before, then optimum # is 11
Prediction of fertility prior to orchard establishment
Fertility varies over years Cumulative cone-yield data would
provide greater reliability Correlations between female fertility
in clone archives and performance of the same clones in seed orchards was close to zero
Thus it’s not worthwhile to collect data as a selection criterion when designing new seed orchards
Deployment of clones to a seed orchard
“Linear deployment” means that clones are deployed proportional to their breeding values
A higher proportion of pollen from a clone constitutes a higher probability of self-fertilization, but seldom leads to fertile seed
Outcrossing pollen is more efficient than pollen that is delivered to ramets of the same clone
Deployment of clones to a seed orchard, cont’d
“Outcrossing effective number” is coined to describe the balance between # of ramets and the effective # of the realised seed crop
Comparison between optimal and linear deployment of clones, under same outcrossing effective number, produced similar results.
At low effective numbers, impact of selfing will be greater
Deployment of clones to a seed orchard, cont’d
Deployment of clones to a seed orchard, cont’d
Pinus sylvestris seed orchards are suggested to be established with 20-25 tested clones, linearly deployed according to their breeding values, with an effective number of 15-18 clones based on number of ramets planted
Thinning of s.o. using the linear deployment algorithm
Thinning by linear deployment results simultaneously in greater genetic gain, higher effective clone #, lower thinning intensity
Gives more flexibility for future thinning, mass production of controlled crosses, selective harvest etc
Thinning of s.o. using the linear deployment algorithm
ParameterBefore thinning
Max. Ne Max. gain at given Ne Truncation selection1
After thinning
Clones 32 32 32 31 29 26 24.35 32
Ramets 5351 3644 3644 3644 3644 3644 3644 3644
Gain 105.96 106.89 108.18 108.52 108.76 108.92 108.97 108.52
Ne 19.99 26.8 24 22 20 18 16.82 22.00
g0 - -667.71 94.86 97.84 100.37 101.15 101.82
b 0 0.20102 16.17 24.87 49.05 79.89
Biological seed production
Seed orchards are often located on abandoned farm land, with favorable climate and soil conditions, which normally increases the seed production
120-150 seeds/m2 or 10 kg/hectare would be possible Seed orchards are generally young compared to seed
stands, and thus seed production potential is underestimated
The studied seed orchards indicate a average biological seed production of 9 kg/hectare
Impact of growth and seed related characters on seed procurement
Growth, morphology and number of strobili have limited genetic variation, thus should not be considered when selecting clones for orchards = take care of it with cultural management
Seed-related characters have impact on seed procurement
Around ¼ of cones are situated in the top level of the crown, ½ in the middle and ¼ in bottom level
Impact of growth and seed related characters on seed procurement, cont’d
Cost for harvesting cones is dependent on tree height, thus pruning is recommended
Variation in fresh weight of cones – related to ripening- can have impact on seed procurement
Summary
When establishing a new seed orchard, little emphasis should be put on selecting clones with high fertility
A seed orchard with tested clones should contain 20-25 clones, linearly deployed, resulting in an effective number of 15-18 clones
Harvest of cones can often be started as soon as the first cones are available, but contaminating pollen can change the adaptability of the seed
For Pinus sylvestris, the optimal active life time of seed orchards seems to be 30 years, for Picea abies 40 years
Highs costs of cone harvest can be reduced by pruning the seed orchard trees
Thank you for your attention