ISSN:
1573-5095
Keywords:
height variation
;
repeatability
;
sampling height
;
tree breeding
Source:
Springer Online Journal Archives 1860-2000
Topics:
Agriculture, Forestry, Horticulture, Fishery, Domestic Science, Nutrition
Notes:
Abstract Longitudinal patterns of within tree variation for basic density were determined for plantation Eucalyptus globulus (ages 5 and 10 years) and E. nitens (ages 5, 10 and 15 years) growing in three geographic areas in Tasmania. Each tree was sampled by taking discs from a combination of percentage heights (0, 10, 20 ___ 70%) and fixed height samples (0.5 m, 0.7 m ___ 1.5 m). At each of the fixed heights, a single pilodyn reading was taken from each of 4 aspects. Weighted whole tree density was calculated from the percentage height samples and used for correlation and regression analysis with the fixed height samples. Both species showed an initial drop in density between the felling cut (zero height) and 0.5 m, followed by a linear increase in density between 10% and 70% of tree height. Slope of fitted regressions were parallel within species but differed significantly between the species (1.14 kg/m3 and 1.62 kg/m3 per 1% increase in height above 10% for E. globulus and E. nitens, respectively). Density at all fixed heights was highly correlated with whole tree values for E. globulus, but results were variable across sites for E. nitens. Both species were found to contain trees which produced aberrant pilodyn readings. For E. globulus, the optimal sampling height was 1.3 m above ground and the mean pilodyn reading was found to predict whole tree density with an accuracy of ±21 kg/m3. For E. nitens, optimal sampling height was 1.5 m above ground. However, pilodyn readings around the stem were not very repeatable and correlations with whole tree density were lower, resulting in the accuracy of prediction of whole tree density being ±26 kg/m3.
Type of Medium:
Electronic Resource
URL:
http://dx.doi.org/10.1023/A:1006544918632
Permalink