Under a point sampling system with terrestrial sample plots the design-variances of the estimators depends on the spatial distribution of the tree locations in the forest. An universal framework for forest inventory optimisation needs therefore a spatial population model. In recent research work carried out at the Swiss Federal Institute of Technology in Zürich, Switzerland, it has been shown, that a simple population model assuming a stratified Poisson distribution of the tree locations, leads to simple expressions of the anticipated variance (mean design-variance under the model) under a wide variety of sample designs and to intuitively understandable optimisation rules.
Some of the findings of these studies are presented and detailed examples for the optimisation of the tree selection probabilities and plot sizes are given. Furthermore, strategies for measuring the efficiency of sample design alternatives with respect to the most important target variables of an inventory are given, and further developments are outlined.