Sampling strategy for estimating bacterial density in large lakes

Abstract

The question of how to use past information about the spatial variability of bacteriological data to develop a strategy for future data collection is investigated when the lake can be divided into a number of «homogeneous » zones. A new method is given for forming the zones and determining the number of zones. Under the assumption that the number of bacteria found in successive, independant, and equal samples from each zone are identically distributed and that the objective of sampling is to estimate the mean bacterial density of the lake, a criterion is given which allows the optimal allocation of stations to zones. In particular, detailed analysis is made when the distribution of bacteria within each zone can be represented by a Poisson distribution. Further, the number of stations that need to be sampled in order to estimate the mean bacterial density of the lake with a pre-specified precision is given. The usefulness of the methods of this paper is illustrated using coliform data from Lake Erie.