Sampling design to study vegetation at the regional level

Abstract

Stratified sampling provides quality needed in the study of relations between species and environmental conditions. A sampling is then optimal in Neyman sense (or N-optimal) when all strata have the same number of relevés. In order to reach it an algorithm as well as the relations and thresholds that help evaluate the quality of the sampling are proposed. A real case example from Burkina Faso is discussed.