2012/11/24 by J. Andrew Royle, Royle, J. Andrew, Sarah J. Converse +3 · 9 citations
Computer Science · Environmental Science · Mathematics · #Abundance (ecology) #Bayes' theorem #Bayesian Methods and Mixture Models #Bayesian hierarchical modeling #Bayesian probability #Biology #Categorical variable #Census and Population Estimation #Computer science #Data mining #Demography #Ecology #Econometrics #FOS: Computer and information sciences #Hierarchical database model #Latent variable #Latent variable model #Mark and recapture #Markov chain Monte Carlo #Mathematics #Methodology (stat.ME) #Multinomial distribution #Population #Population model #Population size #Statistics #Variation (astronomy) #Wildlife Ecology and Conservation #stat.ME
paper · pdf · doi:10.48550/arxiv.1211.5706
published in arXiv (Cornell University) (Cornell University) · 2 figures
arxiv created 2012/11/24 · openalex publication_date 2012/11/24 · arxiv updated 2012/11/27 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
Capture-recapture studies are widely used to obtain information about abundance (population size or density) of animal populations. A common design is that in which multiple distinct populations are sampled, and the research objective is modeling variation in population size Ns; s=1,2,...,S among the populations such as estimating a treatment effect or some other source of variation related to landscape structure. The problem is naturally resolved using hierarchical models. We provide a Bayesian formulation of such models using data augmentation which preserves the individual encounter histories in the model and, as such, is amenable to modeling individual effects. We formulate the model by conditioning on the total population size among all populations. In this case, the abundance model can be formulated as a multinomial model that allocates individuals among sites. MCMC is easily carried out by the introduction of a categorical individual effect, gi, which partitions the total population size. The prior distribution for the latent variable g is derived from the model assumed for the population sizes Ns.