Statistical Aspects of Sampling for Height in Tree Species Trials

SYNOPSIS The paper, based on a re-analysis of existing species trials in Eucalyptus plantations, discusses some basic aspects of sampling. The first part of the study deals with the selection of response variables in such trials: mean height, the median and upper quartile of the height distribution, periodic height increment, a trend coefficient expressing the least-squares estimate of height growth and the total length of the trees within the plot. The second part deals with subsampling and examines the effect of a reduced sample size on the expected variance of the means. A practical test was carried out to evaluate the results of subsampling.

Hali tarjima qilinmagan