A Bayesian multi-state model with data augmentation for estimating population size and effect of inbreeding on survival

Abstract

A joint model framework for estimating population sizes over time and survival probabilities while considering inbreeding and age as covariates in the survival function is elaborated. This methods is tested with data simulated over two small (N≈150) close to extinction open populations, that aims to imitate wild individuals under decline and bottleneck dynamics. A Hidden Markov Model (HMM) perspective with a multi-state formulation that considers young and adult individuals is applied with data augmentation to account for non-seen individuals. The transition probabilities are estimated, and different treatments of the covariates and levels of data seen are compared. Our results suggest that the model framework correctly estimates different population size trends but the parameter estimation is challenging in some cases. The proposed model presents a new way to use the inbreeding coefficient, and further implementations on a real conservation data of wild species can help the decision-making process for the management of small populations.