The estimation of biological population size at large scale by incomplete area surveys and replicated counts

A two-stage estimation of biological population abundance is considered at large scale. In the first stage some area units are selected using without-replacement sampling, while in the second stage the estimation of abundance in each selected unit is performed through a suitable counting strategy. In order to account for the variability due to varying sizes. the use of unit size as an auxiliary variable is proposed at the design level by handling the inclusion probabilities of the units. at the estimation level by combining, the single abundance estimates as in the ratio and regression criteria, or at both levels. The results of an artificial comparison suggest the joint use of simple random sampling and ratio criterion. The asymptotic properties of the resulting estimator are derived when the estimation of abundances at the second stage is performed through the replicated use of the counting strategy. Copyright (C) 2002 John Wiley Sons, Ltd.