Sample size for foliar analyses of coastal Douglas-fir

Abstract

Coefficients of variation were calculated for 17 foliar elements or element ratios in 10 stands of Pseudotsugamenziesii var. menziesii (Mirb.) Franco in the Vancouver Forest Region of British Columbia. These were found to vary widely between elements and, in some cases, between stands. Sample sizes necessary to produce estimates of mean foliar element concentrations within specified error limits under various α and β significance levels and estimates of coefficient of variation were calculated. Concentration of the least variable foliar elements (N, P, S, and K) could be estimated to within 10% with a sample of 21 trees and within 5% with a sample of 68 trees, assuming average coefficients of variation for each nutrient and α and β significance levels of 0.95. The variability reported in this study can be used to determine sample sizes for operational diagnoses of foliar element concentration if pilot samples are prohibitively expensive. A pilot sample should be used in conjunction with explicit β significance levels when determining foliar sample size for research purposes, especially if the samples are to be composited prior to chemical analysis. Analysis of the interrelationships between final sample size, pilot sample size, maximum desirable sampling error, and significance levels can provide the sample designer with information useful for the design of an efficient sample.