Observer error in exit counts of flying-foxes (Pteropus spp.)

Abstract

Population estimation and monitoring is a fundamental component in the conservation management of any species. For species such as flying-foxes (Pteropus spp.) population estimation is complicated by the large number of animals involved, their mobility and the conditions under which counts are conducted. Because count results are used in the determination of management requirements, they are the focus of much critical attention. Despite this, while measures of the precision of fly-out counts of Pteropus spp. have been published, measures of their accuracy have not. In this paper we present an assessment of the accuracy and precision of observer counts of dusk fly-outs by comparing recordings of observers' counts with a video of the same fly-out. Observer's counts were significantly related to the video-count (rs = 0.69, P < 0.0001), with the average observer's count underestimating the video-count by 14.7% (± 25, s.d.) of the video-count. Observers' errors increased with the rate at which flying-foxes left the camp and with the width of the fly-out stream. These results suggest that while observers' errors are inherent in dusk fly-out counts, these errors are manageable and relatively predictable. Other sources of error are likely to have a greater impact on the final population estimate at both camp and regional scales.