Tree-ring analysis can be used to estimate growth in forest inventory, and to supplement or replace permanent sample plot data in the development of growth models. Costs can be reduced by measuring growth rings in a carefully selected subset of the trees in a plot, making use of the information on current tree size. Efficient sampling strategies for estimating past basal area were investigated by simulation with permanent sample plot data.
Combinations of simple random, stratified, and variable probability sampling with minimum variance unbiased, ratio, and regression estimators were tested, as well as a method based on order statistics from the Weibull distribution. Best results were obtained with a variable probability sampling procedure related to systematic sampling.