Etablering av genetiska gallringsförsök i björk

The knowledge about the volume production over whole rotation in improved birch stands is needed to give robust information and recommendations for operational forestry. This might increase an interest in planting birch and in a long term increase the diversity of raw material available for the forestry sector. The long-term birch experiments that can provide such information are nowadays quite scare in Sweden.

In the following project, a new thinnings experiment in 12 years old birch stand was established. Throughout 12 years the experiment was the genetic progeny trial where progenies from 23 full-sib crossings between selected genotypes were tested. The breeding value of each tree was known before thinning experiment was established.

Six experimental plots were thinned with three treatments: thinning for high breeding value after thinning (H_AV), thinning for low breeding value after thinning (L_AV) and usual selective thinning (V_GL). There were two replications of each treatment. One plot remained unthinned (OGL). For each plot around 30% of basal area was removed.

The selection in the treatments which considered breeding values was done with the help of the developed program that took spatial distribution and/or diameter into consideration. It was assumed that trees with high or low breeding values might tend to group in the experimental area as it has been observed for diameter. The selection just based on the breeding value or diameter could lead to the clustering of the trees after thinning.

In all thinned plots the Clark-Evans coefficient index (R, 1954) was higher than 1 and indicated a regular distribution in the stand. Such value of the R coefficient secures the high production possibilities of the growing stand.

In the plots where the usual selective thinning was done (V_GL), the effect of two other treatments (H_AV, L_AV) and thinning based just on diameter were simulated by adjusting parameters of the developed program i.e. consideration of diameter in selection and type of spatial consideration. The thinning effect on Clark-Evans aggregation index and average breeding value was studied. Neither consideration of diameter nor the different spatial consideration affected aggregation index (R) and the average breeding value of remaining and removed trees. In the plots where thinning for high breeding value was applied aggregation index was on average 1.50 while it was 1.70 for plots where trees with low breeding values were promoted. It indicated that trees in the L_AV plots had higher tendency to regular distribution over the plot as the index deviated more from one which indicates a random distribution. However, it is hard to estimate a practical importance of the difference as both were in the limit that secured high productivity of the stand.

The selection in the field resulted in lower average breeding value of the remaining trees (0.89) than in the simulated H_AV thinning (1.08). Thinning based solely on diameter gave similar average breeding value (1.03) as the H_AV thinning considering breeding value and/or spatial distribution. The difference might indicate higher growth potential when breeding value of the tree is carefully considered as trees with high breeding value are expected to have higher growth rate over whole rotation. Moreover, diameter showed to be a quite reliable indicator of the breeding value. The average breeding value of remaining trees when just diameter was used as a selection criterion was slightly lower than in the treatments considering both breeding value and spatial distribution while spatial distribution coefficient was only slightly lower.

The V_GL selective thinning which is usually done in operational forestry might reduce the average breeding value in the remaining stand. It might reduce a growth potential of the stand and cause suboptimal realization of the expected genetic gain. The lower breeding value might be compensated by better quality of the trees after selective thinning. It is important to indicate that thinnings done via simulations did not consider the quality traits that were considered when selecting trees in the treatment V_GL i.e. double tops, spike knots.

The simulations in the StandWise HEUREKA system did not reveal the differences in the stand development which could favoured the stands where selection of superior genotypes was done. The system is favouring the high stem density instead i.e. L_AV with high number of stems after first thinning stands were most productive when no more thinnings were done in the stand. Two additional thinnings in the H_AV and L_AV treatments with the goal to equalise the number of stems per hectare after the last thinning resulted in almost the same volume production over whole rotation age. It might indicate that the differences in the average breeding value between plots were not high enough to have a long-term effect on stand productivity. On the other hand, it might also indicate that the Heureka can not handle the breeding value in a proper way. The follow-up measurements in the stands would give more evidence to draw robust conclusions.

In the long-time perspective the experiment will allow studying both the development of the whole stands with different average breeding value but also long-term development of trees with specific breeding value to see how decrease of the competition influence their growth performance. The results of such studies might be used for estimations and calibrations of genetic gains obtained in the breeding program, calibration of Heureka predictions and for the practical recommendation for selection in thinnings.