Simulation of design-unbiased point-to-particle sampling compared to alternatives on plantation rows
This article is part of a larger document. View the larger document here.Abstract
Total quantities of tree attributes can be estimated in plantations by sampling on plantation rows using several methods. At random sample points on a row, either fixed row lengths or variable row lengths with a fixed number of sample trees can be assessed. Ratio of means or mean of ratios estimators can be developed for the fixed number of trees option but are not designunbiased. Ducey’s estimator samples a fixed number of trees and is design-unbiased, as is sampling fixed row length. Simulations indicted that some of the ratio estimators could have substantial bias on rows containing large gaps, but for other ratio estimators bias was minimal in practical terms. Ducey’s method is unbiased but had a root mean square error slightly larger than some of the ratio estimators.