Development of playback census methods for Storm Petrels Hydrobates pelagicus

Abstract

The factors affecting the probability of nest site attendance and response to playback were examined on Mousa, Shetland, during 1996 with the aim of developing a standard method of censusing Storm Petrels using diurnal playback. Nest site attendance by breeding birds peaked at 90% during July. Site attendance by non-breeding birds was much lower at 15% and this increased as the season progressed. Studies of breeding phenology at colonies throughout Britain and Ireland demonstrated that the average laying dates at most colonies were in late June, but those at Mousa were on average two weeks earlier and at Inish Glora one week later. This analysis also suggested that site attendance is highest in mid-July and so this represents the best time of year to conduct playback censuses at colonies in Britain and Ireland. Rates of response to playback of the male purr call were highly variable, with evidence for effects of the environment, colony location, playback equipment and year. This suggests that the application of a single correction factor to playback survey data collected at different colonies is likely to result in serious errors in estimation of population status. We recommend that the probability of detecting a nest is examined during each survey so that a colony-specific correction factor can be applied to the census data.