6 Environmental adaptive sampling

Abstract Adaptive sampling designs are designs in which the procedure for selecting sites or units at which to observe the variable of interest — such as the level of a pollutant or the abundance of a species — depends on values observed during the survey. For example, additional observations may be made in the vicinity of sites at which high levels of the pollutant or high abundance of the species are encountered. In adaptive cluster sampling whenever the variable of interest satisfies a specified condition units in the neighborhood of that unit are added to the sample. With adaptive allocation designs sample sizes in strate or primary units depend on observed values. When detectability is imperfect in adaptive sampling, estimation methods taking the imperfect detectability into account are available. For rare, clustered or spatially uneven populations adaptive designs can produce substantial gains in precision compared to conventional designs of equivalent sample size. Additional advantages of the adaptive sampling include increasing the yield of the sample — so that, for example, more individuals of a rare species are observed — and helping to find local maxima — the ‘hot spots’ of environmental pollution surveys.