Posterior Predictive Checks for Model Assessment in Occupancy Modeling

摘要: Though methodology and motivation for the use of posterior predictive checks has been introduced in some textbooks (eg Gelman et al. 2004), the technique is not commonplace in, for example, applied ecological studies using Bayesian techniques. In many studies using Bayesian techniques, the only semblance of model checking is the use of DIC for model selection, but this does not provide practically meaningful information about the quality of the fit. In this paper, we demonstrate the use of posterior predictive checks for model assessment under an occupancy study of small land birds in the Helena National Forest. Occupancy of these birds was recorded at many different sites for several years before the onset of a massive pine beetle epidemic as well as for several years after the onset of the epidemic. We present two rather simple models to describe the occupancy of the bird species at the different sites and demonstrate the use of posterior predictive checks for model assessment and comparison of these two models.