Analysing habitat use from radio-tracking data.

Abstract

In order to develop an effective conservation strategy for protecting a wild gamebird species it is often essential to know its detailed habitat requirements. A general idea of the broad biotype that a species occupies (e.g. forest, agricultural land, marshes etc.) is not sufficient. We need to find how it uses habitat types within these areas and which of these are most important for its continued survival. In order to acquire this type of information, detailed studies of the birds' locations are required, often using techniques such as radio telemetry.
Radio telemetry enables the biologist to follow the daily movements and activities of an animal without disturbing it. The technique has been used with great success on pheasants (Davison 1981, Hill and Robertson 1988, Young et al. 1991). For each marked animal, the distribution of its radio-locations provides a measure of its habitat use. The use of these data in the analysis of the habitat utilization of marked animals presents a number of statistical and conceptual problems involving sample size, non-independence of data and assessment of habitat availability. Several techniques commonly in use or recommended for habitat analysis (Neu et al. 1974, Alldredge and Ratti 1986, White and Garrott 1990) often ignore one or more of these problems, thereby casting doubt on the technique's validity.
This paper illustrates an alternative method of analysing habitat utilization based on compositional analysis (Aitchison 1986). The problems and pitfalls associated with analyses of habitat utilization are discussed, followed by an explanation of how compositional analysis copes with these problems. Further details are provided in Aebischer and Robertson (1992) and in Aebischer, Robertson and Kenward (in press). In the second part of the paper, a worked example is given which uses data collected from Grey Partridges Perdix perdix, released in southern England as part of an attempted restocking exercise.