A Bayesian Approach to Estimating Mallard Duckling Survival: Understanding the Relationships with Wetland Density and Hatch Date

Mallard (Anas platyrhynchos) duckling survival is an important, but not well understood, component of reproductive success for the species. To further investigate how mallard duckling survival is related to hatch date and wetland density, we analyze duckling counts for broods in southwestern Manitoba during the late 1980’s. Various models are explored to explain the patterns in the observed data and these approaches are compared to the original analysis (Rotella and Ratti 1992). We demonstrate fitting models using a Bayesian approach and discuss the advantages of this framework in comparison to maximum likelihood estimation. Different Bayesian models are compared to one another and evaluated using posterior predictive checks. Zeroinflated binomial models appear to provide an ecologically realistic description of these data and allow a more complete understanding of how duckling survival is related to wetland density and hatch date. Estimating Mallard Duckling Survival CONTENTS