Bird responses to forest loss are influence by habitat specialization

Abstract

Aim Due to intrinsic differences in the sensitivity to habitat grain among species, studies performed at different extent are necessary to understand the consequences of forest loss and fragmentation. Using a large database, we explored the responses of birds to changes in forest cover and the role of habitat specialization in the strength of this response.

Location Southern Atlantic forest of Argentina, Brazil and Paraguay.

Methods We used data on bird occurrences recorded in 1,384 point counts (2004–2011), estimated forest cover and number of forest fragments in two radii (60 and 960 m), centred at each point count. For each bird species, we extracted the geographical and altitudinal range as two indirect measures of habitat specialization. We used general linear model and Akaike information criterion to explore the influence of the type of habitat, the amount of habitat and fragmentation pattern on the probability of species occurrence and the influence of habitat specialization on the strength of response.

Results Of the 28 species analysed, 15 (55%) responded either to forest loss or the number of fragments, either positively or negatively. In these 15 species, the probability of occurrence of 67% was better explained by a specific extent (either 60 or 960 m). The strength of the response to forest loss decreased with both the geographical and the altitudinal range of species.

Main conclusions Our study shows that a large proportion of species responded to forest loss at a specific extent and that the magnitude of the response is related to species specialization. A single-extent approach to multispecies studies may not be enough to preserve the whole community due to differences in sensitivity to habitat grain. Maintaining forest cover at multiple extents and managing anthropogenic habitats to increase their suitability for native species are essential to preserve communities in highly fragmented landscapes such as the Atlantic forest.