Multi-scale habitat selection affects offspring survival in a precocial species.

In theory, habitat preferences should be adaptive. Accordingly, fitness is often assumed to be greater in preferred habitats; however, this assumption is rarely tested and, when it is, the results are often equivocal. Habitat preferences may not directly convey fitness advantages if animals are constrained by tradeoffs with other selective pressures like predation or food availability. We address unresolved questions about the survival consequences of habitat choices made during brood-rearing in a precocial species with exclusive maternal care (mallard Anas platyrhynchos, n = 582 radio-marked females on 27 sites over 8 years). We directly linked duckling survival with habitat selection patterns at two spatial scales using logistic regression and model selection techniques. At the landscape scale (55-80 km(2)), females that demonstrated stronger selection of areas with more cover type 4 wetlands and greater total cover type 3 wetland area (wetlands with large expanses of open water surrounded by either a narrow or wide peripheral band of vegetation, respectively) had lower duckling survival rates than did females that demonstrated weaker selection of these habitats. At finer scales (0.32-7.16 km(2)), females selected brood-rearing areas with a greater proportion of wetland habitat with no consequences for duckling survival. However, females that avoided woody perennial habitats composed of trees and shrubs fledged more ducklings. The relationship between habitat selection and survival depended on both spatial scale and habitats considered. Females did not consistently select brood-rearing habitats that conferred the greatest benefits, an unexpected finding, although one that has also been reported in other recent studies of breeding birds.