Breeding bird density in woodlots: Effects of depth and buildings at the edges

A model is presented that describes the breeding bird density of a species in a woodlot, given that this density may change from edge to interior. The primary variable is a measure of depth in the woodlot that depends on dist ances to four woodlot edges. The model applies to forest-interior species a nd to edge species, and it also accounts for different degrees of threat or attraction at different edges. During 1993-1996, we collected bird abundan ce data using point counts at 387 sample plots in 176 woodlots, including t wo big woods (>5000 ha), in eastern Pennsylvania, USA. Models were develope d for 36 species by using nonlinear regression guided by tests of residuals against other variables measured at each sample plot. Densities of 20 of t he 36 species were affected by the depth in the woodlot. The density of one forest-interior species, Ovenbird (Seiurus aurocapillus), decreased farthe r away from the big woods, which can be explained qualitatively by dispersa l dynamics. Densities of 21 of the 36 species were affected by buildings ne ar the woodlots: 10 species increased, and 11 decreased. One preliminary co nservation conclusion is that it is important to preserve big woods (>5000 ha) because at least one common forest species (Ovenbird) might disappear f rom nearby woodlots if the big woods were to become 'fragmented. A second c onservation conclusion is that certain species can persist only in the abse nce of nearby buildings. Perhaps these species would persist near buildings if the negative biotic interactions associated with buildings were removed , but more information is needed before this possibility can be evaluated. Our model-building procedure and statistical approach might be useful for m odeling a number of variables concerning birds or other animals in various habitats where edge effects are important.