Chapter 9 - Closed-Population Mark-Recapture Models

Abstract

Mark-recapture, occupancy, and distance sampling all involve models for missing data. In each the mechanism by which data are missing is presumed to depend on the values of unknown parameters, such as the population size or the catchability of the animals. For problems of this nature, the mechanism by which the data went missing must play a dominant role in any model if reasonable inference is to result. This chapter describes mark-recapture models in the context of missing data, and highlights the usefulness of the complete data likelihood (CDL) for mark-recapture modeling, and demonstrates the use of data augmentation for abundance and hierarchical modeling for multimodel inference. In both cases, the required models is implemented in BUGS without the need for customized code based on reversible jump Markov chain Monte Carlo, despite the potential difficulties due to varying model dimensions.