CONTRIBUTIONS TO A DISCUSSION ABOUT SYSTEMATIC-SAMPLING IN FOREST INVENTORY

This paper provides 2 examples to demonstrate the problem of variance estimation in the case of systematic sampling. Low sampling errors are closely related to a remarkable overestimation of the variance of estimators of population means and sums. Therefore high precision as a result of systematic sampling is neutralized, because we must judge the results of an inventory by overestimated sampling errors. Stratified random sampling is propagated as an alternative, which allows unbiased variance estimation and in many cases comparable or even lower sampling errors. Obviously the additional efforts for the planning of stratified samples in forest inventory will be reduced substantially by aid of forest information systems. But systematic sampling will keep importance for small scale inventories because of its clear advantages in locating sample plots.