Survival and habitat utilisation: a case for compositional analysis.

Abstract

Avian survival rates can vary not only by age and year, but also in relation to individual habitat use. Radiotelemetry makes it possible to determine repeatedly the location of a tagged individual. Its habitat use is then given by the proportion of radiolocations in each habitat type. Radiotelemetry is also a direct method of establishing whether a marked bird is still alive. The importance of different habitat types for survival may be assessed by compositional analysis. This technique recognizes that the sample size is the number of tagged animals, and that radiolocations serve to subsample each individual's habitat use. It allows for the analytical complication that, for a given individual, the proportions describing habitat use sum to one over all habitat types. The principles of compositional analysis are explained and demonstrated by reference to survival and habitat data collected for radiotagged pheasants on the Swedish island of Gotland in 1984 and 1985. There was no effect of age or year on survival, but pheasants that survived made significantly greater use of vegetable crops and less use of pasture than those that died.