Relative Efficiency of Point Sampling Change Estimators

Abstract

Concerns about the efficiency and the reliability of point sampling to estimate change in forest growth variables have been expressed ever since point sampling appeared in the literature more than 60 years ago. Change estimators for point samples based on point-to-tree distance in variable-radius plots were introduced about 30 years ago but are rarely implemented despite easy access to point-to-tree distance. The statistical efficiency and bias of these newer estimators were compared to traditional fixed-area plot estimators using stem-mapped permanent sample plots. Methods using variable-radius plots and point-to-tree distance were more efficient to estimate volume and basal area while fixed-area plots were more efficient to estimate trees/ha. Compatible and time-additive estimators are examined for estimating survivor, mortality, and ingrowth change using point samples. These estimators are unbiased under unrestrictive conditions.