Comparing two approaches to populations of sample plots in forest inventories.

摘要

In this paper, we compare two approaches for a definition of populations of sample plots for forest inventories. The first is intuitively more rapidly accessible where the forest area is subdivided into mutually exclusive plots of identical size and shape in such a way that the entire area is covered without gaps. This builds a discrete population of N countable plots from which a sample is then selected. Obviously, this approach is only feasible for a limited number of plot types (rectangular, triangular and hexagonal fixed area plots). The other approach considers all (dimensionless) points in the forest area as possible reference- (center-) points for a sample plot. Under that approach, the population is infinite and the set of plot values is continuous. With some methodological considerations, it is illustrated, that with this second approach, some times also referred to as Monte Carlo approach, it is much more straightforward to describe and explain sampling and response design elements such as tree selection probabilities and the mirage technique for border plot correction.