Efficient Sampling Techniques for Timber Sale Surveys and Inventory Updates
Three different regression estimators, predicting tree or plot volume from tree diameter squared times total height or previous plot volume 5 or 10 years earlier for specific populations, were studied with respect to minimum relative bias, minimum variance, percentage of confidence intervals containing true volume, and consistency in estimating total volume using six sampling rules. The sampling rules were: restricted random sampling (RSRS), sampling from the range of x (psR), pps sampling (standard against which all others are compared), stratified pps sampling (spps), pscx, and spscx procedures. The latter two sampling rules yielded the most efficient volume estimators in all three populations. Stratified pps sampling is less reliable than the pscx procedures but better than pps sampling. Restricted simple random sampling and sampling from the range of x are less reliable than pps sampling. Forest Sci. 31:857-866.
No Supplementary Data
No Article Media