Composite samples in environmental studies

Abstract

In ecological studies, bulk material such as soil, water, plant material etc. is commonly investigated. An important feature of bulk sampling is the possibility of physical mixing of initial increments of material into a composite sample. In ecological studies, chemical analysis cost far outweighs all other expenses. Compositing samples into composite samples reduces the number of laboratory samples for chemical analysis and thus diminishes the total cost.

Due to economic constraints some researchers acquire only one composite sample in the study. In this case an estimate of the mean can be obtained, however no estimate of sample precision is possible. Hence, this approach is not recommendable for long-term ecological studies as statistical tests are not applicable.

In this paper we present the composite sample alternative of the two-stage sampling plan. When instead of the two-stage design its composite sample alternative is used, the total expenses is reduced considerably. We present the mean and the variance of the mean for the composite sample alternative. In situation commonly encountered in ecological studies the impact on compositing on the precision of the mean is small. Efficiency of the composite sample design is greater than efficiency of the corresponding two-stage design, thus we recommend the composite sample design for ecological studies.