Breeding-site infidelity in greater snow geese: a consequence of constraints on laying date?

We studied annual variation in the distribution of greater snow goose (Chen caerulescens atlantica) nests and colonies at a High Arctic breeding site (Bylot Island, Northwest Territories, Canada) from 1988 to 1994. Annual distribution and size of colonies within the study area were highly variable. The distance between colony centres in successive years ranged from 2.9 to 6.9 km ((x) over tilde = 3.9 km), with no overlap of the areas occupied. Timing of snowmelt and nest-site availability were also very variable. In some years, the area was snow-free at the time the geese arrived, whereas in others, snow persisted for up to 3 weeks after their arrival. However, even in the 3 years of earliest snowmelt (1988, 1993, and 1994), geese used markedly different nesting areas, in both lowland and upland areas. Geese nested in association with snowy owls (Nyctea scandiaca) in the 2 years that owls nested and their nesting success was markedly improved, probably because of the deterrent effect of owls on arctic foxes (Alopex lagopus), the main nest predator. This suggests that protection from nest predators is an important determinant of nest-site selection. The lack of fidelity to a specific nesting location is unusual in geese. We suggest that large annual variability in timing of snowmelt, predation pressure, and availability of feeding areas during the prenesting period, combined with the need to lay at on optimal date because of the short arctic nesting season, probably explain the low nest-site fidelity in this population.