Estimating spatiotemporal variation in mortality improves conservation insights

Wildlife population dynamics change across time and space, but few studies use formal spatiotemporal statistical models to understand these patterns. The population distribution of mallards (Anas platyrhyncos) in the Prairie Pothole Region (PPR) has shifted dramatically over the past several decades as agricultural and climate patterns change. We examined spatiotemporal variation in survival, natural mortality, and harvest mortality for mallards in the PPR from 1974–2023 using intrinsic conditional autoregressive models in a Bayesian framework. The results showed substantial variation over time and space for all demographic rates that was related to habitat variables and varied by age and sex classes. Specifically, environmental relationships indicate that habitats with less agricultural activity and more wetlands in breeding grounds may benefit dabbling duck populations but decrease survival via increased natural mortality driven by reproductive risks. We also found a worrying trend: females exhibited declining survival probabilities across the time series.