Constellation, le dépôt institutionnel de l'Université du Québec à Chicoutimi

Résumé

Presence-absence data is often used to determine the preferred habitat of a given organism. However, with presence-absence datasets there is a problem associated with the comparison between habitats when there is an inter-habitat variation in the proportion of false zeros. Using conspecific playbacks and time of reaction of black-backed woodpeckers in burned and unburned forests, the present study determined whether detection probabilities were similar. The period of time required to detect this species was shorter in recently burned sites than in mature forest stands (F2,235 = 22.1, df = 2, P<0.0001). To accurately compare these habitats it is important to assure the same proportion of presence and false zeros in each habitat during the census. To achieve this, we propose a time corrected method. Because inter-habitat differences in detectability exist for black-backed woodpeckers, and probably many other organisms, caution is needed when interpreting presence-absence data in the context of habitat comparisons or when monitoring biodiversity in different habitats.