Effect of plot size on estimates of top height in douglas-fir

Abstract

Top height is frequently used to estimate site quality and growth potential of forest stands. Top height estimates are quantile estimators and thus sensitive to scaling. If top height is defined as the height of the largest dbh tree per 0.01 ha of area, then it follows that any estimate based on ntop trees in a ntop·0.01 plot will be a function of ntop, the distribution of dbh values, and the height-diameter relationship of the stand. This study quantifies for second-growth Douglas-fir (Pseudotsuga menziesii) stands along the west coast of British Columbia top height estimates derived from plots of size 0.01, 0.03, 0.05, and 1.0 ha. A pool of 635 permanent and remeasured sample plots allowed the estimation of expected stand-level diameter distributions (Weibull) and height-diameter relationships. Simulated sampling from these distributions quantified the impact of plot size. Top height, derived from plots of 0.01 ha, 0.03 ha and 0.05 ha were 5.6%, 3.0%, and 2.5% lower, respectively, than the estimates based on 1.0 ha plots. Yield predictions based on top heights from 0.01 ha plots were, on average, 8% lower than predictions from 1.0 haplots. Three of four top height predictions based on 0.01 ha plots were significantly (P< 0.05) lower than predictions from a 1.0 ha plot.