Estimating temporal independence of radio-telemetry data on animal activity

Radio-telemetry is an excellent tool for gathering data on the biology of animals and their interactions with the environment they inhabit. Many methods have been developed for analyses of spatial information, on home range size and utilization density. Activity patterns are often described using radio-tracking data, but no generally accepted method is currently available specifically for determining the temporal independence of this type of data for statistical inference. Activity rhythms have generally been analysed by ecologists with the assumption that data are temporally independent, or by subjectively fixing an independence interval, based on attributes of their ranging behaviour. Although some good approximations of activity patterns can be obtained in these ways, we underline the need for a functionally correct method of estimating independence interval. Here we use semi-variograms to estimate the minimum interval required for the readings to be sequentially independent. This geostatistical tool is applied to the analysis of data on activity of Chilean foxes (Pseudalopex culpaeus) and Chacoan peccaries (Catagonus wagneri). Data were collected in the field by radio-tracking over 24-hr periods, with readings on activity state taken every 15 min. The spatial dimension in which the theory of geostatistics lies has been transferred into the time dimension, so that the correlation interval is expressed in time units (min). Time of independence as estimated by the variogram was 110 min for foxes, while data on peccaries indicated that they have long periods of activity, more suitable for time-series analysis. Copyright 1999 Academic Press.