Lateral variability of forest floor properties under second-growth Douglas-fir stands and the usefulness of composite sampling techniques

Abstract

Lateral variability of forest floor physical and chemical properties is examined in LF and H horizons under six naturally regenerated, second-growth Douglas-fir (Pseudotsugamenziesii (Mirb.) Franco) stands in coastal southwestern British Columbia. The number of samples required to predict a mean value at two confidence levels (P = 0.01 and 0.05) and two allowable errors (10 and 20%) are given for each variable. Total C, N, P, S, Zn, pH, and lipids were the least variable, requiring 2–13 samples to estimate a plot mean with a 10% allowable error at the 95% confidence level in LF horizons and 3–51 samples per lot in H horizons. Total K, Cu, and Mn were found to have moderately high lateral variability, while total Ca, Mn, Al, and Fe all required large numbers of samples to estimate the plot mean. In the second part of the paper, composite samples weighted by field depth and bulk density are compared with the depth and bulk density–weighted arithmetic mean of subsamples analyzed individually. Values from analysis of composite samples were within one standard deviation of the mean, with the exception of P and Cu in the LF horizons and lipids in both horizons. Composite values and mean values were significantly correlated across the six sites for all variables except lipids in LF horizons and total C and Mn in both horizons. Composite samples are suggested to provide an adequate estimate of the mean value of subsamples analyzed individually for most purposes and, for some variables (i.e., Ca, Fe, Al, and Mn), the only feasible method of obtaining an estimate of the mean.