Adaptive Sampling for Spatial Capture-Recapture: An efficient sampling scheme for rare or patchily distributed species

Abstract

Rare species present challenges to data collection, particularly when the species is spatially clustered over large areas, such that the encounter frequency of the organism is low. Sampling where the organism is absent consumes resources, and offers relatively low-quality information which are often difficult to model using standard statistical methods. In adaptive sampling, a probabilistic sampling method is employed first, and additional effort is allocated in the vicinity of sites where some measured index variable - assumed to be proportional to local population size - exceeds an a priori threshold. We applied this principle to the spatial capture-recapture (SCR) analytical framework in a Bayesian hierarchical model incorporating capture-recapture (CR) and index information from unsampled sites to estimate density. We assessed the adaptively sampled SCR model (AS-SCR) by simulating CR data and compared performance with a standard SCR baseline (F-SCR), adaptive SCR discarding index information (AS-SCR-), and standard SCR applied at a simple random sample of sites. Under AS-SCR, we observed minimal bias and comparable variance with respect to parameter estimates provided by the standard F-SCR model and sampling implementation, but with substantially reduced effort and significant cost saving potential. This represents the first application of adaptive sampling to SCR.