|
Résumé :
|
Reliable species distribution data are important for valid scientific conclusions and effective conservation planning. Mismatch between survey timing and animal behaviors that influence detection may result in false absences that can lead to poorly informed management decisions. Birds exhibit a diversity of diel vocalization patterns, but many large-scale multispecies surveys are based on the songbird dawn chorus, indicating the potential for bias to detect birds with other diel vocalization patterns. In this study, we quantified bias in point counts and morning acoustic recordings to measure the number of occupied sites detected for a set of dawn chorusing birds (songbirds) and irregularly vocalizing wetland birds (waterfowl) relative to estimates obtained from 10-min acoustic recordings conducted hourly throughout 24-h periods for three consecutive days. Furthermore, we investigated which revisitation schedule—same day or different day sampling, as well as increased sampling effort—best minimized false-negative detections for songbirds and waterfowl. Morning surveys significantly underestimated the number of occupied sites for 10 of 13 species. No differences were found between same-day and between-day revisitation schedules with identical sampling effort, regardless of whether birds exhibited a dawn chorus or irregular vocalization patterns. Detection improved with increased sampling effort. Subsampled recordings captured the majority of occupied sites for songbirds (up to 87% of occupied sites detected), but less so for waterfowl (up to 60% of occupied sites detected). Accurate detection for irregularly vocalizing species such as waterfowl will require more intensive sampling effort (likely throughout 24-h periods) when using acoustic recordings.
|