Title: Bayesian capture-recapture models with temporary emigration and heterogeneity using mixtures of changepoint processes
Authors: Eleni Matechou - University of Kent (United Kingdom) [presenting]
Raffaele Argiento - University of Torino (Italy)
Abstract: In populations where individuals are detected with probability lower than one, capture-recapture methodology is often employed for estimating population size and, in open populations, arrival and departure times. However, all the existing capture-recapture models either assume that emigration is permanent and each individual performs a single visit to the site of interest, or require the use of specialised sampling schemes that assume population closure for certain times of the sampling period. We propose a novel approach for modelling capture-recapture data on open populations that exhibit temporary emigration, whilst also accounting for individual heterogeneity. Our modelling approach combines changepoint processes, fitted using an adaptive approach, for inferring the number and timing of individual visits with Bayesian mixture modelling, fitted using a nonparametric approach, for identifying clusters of individuals with similar visit patterns and capture probabilities. The proposed method is extremely flexible as it can be applied to any capture-recapture data set and is not reliant upon specialised sampling schemes. We demonstrate our model and algorithm using a motivating data set on anglers fishing in the Gaula river in Norway and the results provide us with the first ever estimate of the size of the population of anglers fishing during the season as well as new insights on the visit patterns of anglers and their probabilities of catching salmon whilst at the river.