Abstracts

Presented by Todd W. Arnold - Email: arnol065@umn.edu

Duckling survival is the most influential factor affecting population growth of mallards (Anas platyrhynchos) in New Zealand. Understanding how habitat choices affect duckling survival can provide valuable insights for managing landscapes to increase productivity. During 20142015, we conducted 2,252 observations of 190 brood-rearing female mallards on two study sites in New Zealand (NZ). We used ArcGIS Pro to identify used and available habitats around brood travel routes and we used these data to predict duckling survival from interval-specific observation matrices of offspring counts and other covariates. We analyzed data using a Bayesian hierarchical generalized linear model that simultaneously estimated daily duckling and brood survival, and individual duckling and brood detection. Habitat selection was strongest when drains, ponds, effluent ponds or sedge habitat constituted greater than 11%, 7%, 5%, or 13% of the brood buffer, respectively. Mallards that selected brood-sites with greater proportions of effluent ponds experienced higher duckling survival, indicating adaptive selection. However, mallards also selected drains and ponds despite lower duckling survival associated with greater amounts of these habitat types, suggesting these habitats may be ecological traps. Although we found no evidence of selection or avoidance of streams and hedgerows, both habitats were associated with higher duckling survival, implying mallards may not always recognize beneficial habitats. Mallards in NZ exhibited both adaptive and maladaptive patterns of habitat selection of brood-rearing areas, which may result from their non-native status, anthropogenic land use, or intense habitat fragmentation.