The walk through and fro estimator for edge bias avoidance

Abstract

Edge bias can result from the failure of sampling estimators to account for the true sampling probabilities of trees near the edges of a stand. Numerous estimators for basal area, each with an associated sampling method, have been proposed so as to avoid or reduce edge bias. The "walkthrough" estimator by Ducey et al. (2004) is easier to use than most competitors. A modification of that estimator is proposed here and referred to as the walk through and fro estimator. Along the line that connects the sample point and a particular sample tree, distances from the tree to any intersected stand boundaries are recorded if those distances are shorter than the radius of the conceptual tree-centered plot. The recorded distances are used to calculate sample weights that allow for unbiased estimation of basal area under all circumstances. The mean squared error for this estimator is generally less than that of the walk through estimator but greater than that of the tree-concentric estimator. Several variations on the proposed estimator are described. © 2013 by the Society of American Foresters.