Effectiveness and efficiency of avian species detection: a comparison between field observers and automatic recording devices

The monitoring of animal populations is essential for reporting on the state of the environment, with birds often used as indicators of ecosystem health. Traditionally, bird monitoring has been done by field observers; however, there has been recent interest in use of automatic recording devices (ARDs) as an alternative. A monitoring programme managed by the New Zealand Department of Conservation (DOC), used observers and ARDs concurrently for three survey seasons, providing the opportunity to compare results in terms of effectiveness and efficiency. The difference in species-richness estimates from the two methods was small, with the observer method detecting slightly higher numbers of species in all habitat types. Detection probabilities for individual species, derived from occupancy analysis, were similar between methods, with a few exceptions: bellbird (Anthornis melanura), brown creeper (Mohoua novaeseelandiae), tui (Prosthemadera novaeseelandiae), North/South Island robin (Petroica longipes/australis), and rifleman (Acanthisitta chloris). Bellbird and rifleman had a higher probability of being detected by ARDs, whilst the remainder were more likely to be detected by observers. Differences in detection probability may be due to identification confusion in the case of bellbird and tui, and observer ability to detect and identify birds visually for brown creeper and North/South Island robin. The relationship between indices of abundance from the observer and ARD methods varied between species and habitat types. These inconsistencies suggested that the ARD results did not correlate closely with observed abundance, which may limit the ARD method to provision of confirmed presence data. Observer counts proved to be more time-efficient given present levels of processing technology, mainly due to the longer processing time required for ARD recordings. However higher numbers of people were required for observer counts, which may be problematic when there is a shortage of appropriately skilled observers at the required time of year.