Some considerations on optimal plot size in field experiments with forest trees
Stern K. (1968). Some considerations on optimal plot size in field experiments with forest trees. Silva Fennica vol. 2 no. 4 article id 4776. https://doi.org/10.14214/sf.a14561
18 permanent sample subplots of the Swedish and the Hassian Forestry Institute, each measured in equal intervals for several decades, were divided into subplots of different size. An analysis of variance was calculated for every set of subplot size. The development of intraclass-correlations over years and over different sizes of subplots could be explained if three different correlations were assumed: soil-correlation, correlations from irregular distribution of the trees, and correlation resulting from competition. Intraclass-correlations were positive or negative depending on dominance of one or two of these correlations.
An explanatory simultation study of competitional variance showed the effect of the degree of competitional correlations on the variance of means of subplots of different sizes. If this coefficient was small, all variances of subplots means within the range investigated became larger than expected in experiments without competition, with larger coefficients the variances of means of the smaller subplots became smaller, those of larger subplots larger than expected.
Plots of medium or large size are probably optimal for long term experiments with forest trees, if all sources of costs in such experiments are taken into account.
The PDF includes a summary in English and Finnish.
metsänarviointi; menetelmät; koealat; inventointi; koealakoko; metsänmittaus
forest inventory; sampling; forest mensuration; methods; sample plots; sample plot size
Published in 1968
Available at https://doi.org/10.14214/sf.a14561 | Download PDF