Abstract
This chapter shows how hidden Markov models (HMMs) can be used to develop capture–recapture and occupancy models, traditionally used to study the dynamics of populations and the distribution of species in a context of imperfect detection. It shows how the HMM formulation permits the estimation of hidden variables in two different case studies. These case studies include: estimating the prevalence of dog–wolf hybrids with uncertain individual identification; and estimating the distribution of a wolf population with species identification errors and heterogeneous detection. The hidden variables encountered in the study of animal populations are living/dead; developmental states, which are generally discrete, such as sexual maturity; epidemiological states; or social states. HMM will be used to model species distribution in a case featuring identification errors and heterogeneous detection. The main advantage of the HMM approach lies in the ability to infer the ecological states of individuals and species which are partially observable: these are hidden variables.