Estimating dispersal among numerous sites using capture-recapture data

Dispersal affects processes as diverse as habitat selection, population growth, and gene flow. Inference about dispersal and its variation is thus crucial for assessing population and evolutionary dynamics. Two approaches are generally used to estimate dispersal in free-ranging animals. First, multisite capture recapture models estimate movement rates among sites while accounting for survival and detection probabilities. This approach, however, is limited in the number of sites that can be considered. Second, diffusion models estimate movements within discrete habitat using a diffusion coefficient, resulting in a continuous processing of space. However, this approach has been rarely used because of its mathematical and implementation complexity. Here, we develop a multi-event capture recapture approach that circumvents the issue of too many sites while being relatively simple to be implemented in existing software. Moreover, this new approach allows the quantifying of memory effects, whereby the decision of dispersing or not on a given year impacts the survival or dispersal likelihood of the following year. We illustrate our approach using a long-term data set on the breeding ecology of a declining passerine in southern Quebec, Canada, the Tree Swallow (Tachycineta bicolor).