A comparison of N-tree distance sampling with fixed-radius plot and variable-radius point sampling methods

N-tree distance sampling is examined analytically and empirically as an alternative to the traditional forest inventory methods, fixed-area plot and variable-radius point sampling. With n-tree distance sampling, the number of trees sampled at every location is the same (n). The radius of the circular plot is the distance from the sample location to the center of the n{dollar}rmsp{lcub}th{rcub}{dollar} tree. The method is easy to conduct in field applications.; The estimators of trees per acre and basal area per acre are unbiased for all three methods in random or clustered spatially patterned forests modeled by the Poisson or negative binomial distributions, respectively. A formula was derived to quantify the precision relationship of n-tree sampling and plot sampling when estimating the number of trees per acre in these forest types. The number of trees needed for n-tree distance sampling to be as precise as plot sampling depends on the fixed-radius plot size, the forest density, and the amount of clustering.; N-tree sampling of stands produced estimates which were slightly biased. The direction of this bias depended on both the spatial pattern of the forest and the form of the sampling estimator. Although the coefficients of variation for n-tree sampling tended to be larger than those of traditional methods it was compared with, increasing the number of samples for the n-tree method made the precision of the methods equivalent. A time study demonstrated that when few trees are included in an n-tree sample, increasing the number of samples may be a cost effective alternative.