Applying sequential adaptive strategies for sampling animal populations: An empirical study

Traditional sampling methods may prove inadequate when dealing with spatially clustered populations or when studying rare events or traits that are not easily detectable across the target population. When both scenarios occur simultaneously, adaptive sampling strategies can represent a viable option to enhance the detectability of cases of interest. This paper delves into the application of a novel class of sequential adaptive sampling strategies to animal surveys. These strategies, originally proposed for human population tuberculosis prevalence surveys, allow oversampling of the rare interest variables while managing on-field constraints. This ensures that the unfixed sample size, typical of adaptive sampling, does not compromise overall cost-effectiveness. We explore a strategy within this class that integrates an adaptive component into a Poisson sequential selection. The aim is twofold: to intensify the detection of cases by exploiting the spatial clustering and to provide a flexible framework for managing logistics and budget constraints. To illustrate the strengths and weaknesses of this Poisson-based sequential adaptive sampling strategy compared to traditional sampling methods, a simulation study was conducted on a blue-winged teal population in Florida, USA. The results showcase the benefits of the proposed strategy and open avenues for future methodological and practical improvements.