A habitat use model for the barred owl (strix varia) in Cheshire County, New Hampshire

Wansleben, Thomas W.
Jon Atwood, PhD
Department of Environmental Studies
The Barred Owl (Strix varia) is considered to be a common raptor in southwestern New Hampshire, yet little quantitative information exists concerning their distribution and habitat requirements. Current avian monitoring programs such as the Breeding Bird Survey and the Audubon Christmas Bird Count do not adequately represent the distribution or popular trends for the Barred Owl as they do not occur during their breeding season when the species is easily detected. In order for conservation efforts to effectively consider the barred Owl we must understand how habitat characteristics influence their distribution. The purpose of this research was to characterize the habitat surrounding sites occupied by Barred Owls in order to develop a model that could predict the probability of Barred owl distribution within a landscape. From Feb-Apr 2006, broadcast surveys detected Barred Owls at 47 of 130 sites throughout Cheshire County, New Hampshire. Habitat characteristics were generated using satellite-derived landcover data managed through Geographic Information Systems (GIS) software. Characteristics of sites where Barred Owls were detected were compared to sites where they were not detected, and a multiple logistic regression model based on these results was used to predict the probability of Barred Owl occurrence throughout the non-surveyed portion of the study region. Barred Owls were detected in landscapes that contained higher percentages of forested and undeveloped land cover when compared to areas where there were not detected, thus supporting their preference for forest-dominated landscapes with minimal human alterations. Owls were also found in areas that contained lower densities of roads and rivers. Small landscape fragments <40.5 ha (100 acres) were also more prevalent in sites where barred Owls were not detected, reflecting the species' preference for largely unfragmented landscapes. The best logistic regression model correctly classified only 50% of sites where barred Owls were detected, suggesting that the species' distribution may be influenced by localized habitat characteristics or spatial features different from those that were measured. Nonetheless, the results of this study indicate that efforts to conserve viable populations of Barred owls should focus on protection of large unfragmented forested landscapes, where development and conversion of forest to non-forest uses is kept to a minimum. The results also indicate the need for more detailed studies into the habitat requirements of the Barred Owl.

Read Full Text