Temporal variation of bird assemblages in moderately fragmented and less-fragmented boreal forest landscapes: A multi-scale approach

Abstract

To test whether landscape structure affects the temporal variability of bird populations and species turnover, we estimated breeding bird abundances in a boreal forest area in northern Finland during the years 1990-1995. The analyses were conducted at three spatial scales: point-count station (4 ha), block (12.5 km2), and landscape unit levels. At the census point scale, temporal variability (CV) of generalist species correlated negatively with amount of unforested habitat. At the block scale, temporal variability of forest edge and managed forest species correlated negatively with the amount of forested area in the landscape. No response was observed among sedentary, migratory, and virgin forest species at these lower scales. At the landscape unit scale, forest-edge and managed-forest species such as chaffinch (Fringilla coelebs L.), and redpoll (Carduelis flammea L.) showed higher temporal variability, whereas sedentary and virgin forest species such as cuckoo (Cuculus canorus L.), raven (Corvus corax L.), and hooded crow (Corvus corone cornix L.) showed lower temporal variability in the moderately fragmented landscape than in the less-fragmented landscape. Of the virgin-forest species, crossbills (Loxia spp.) and titmice (Parus spp.) showed more variability in the less-fragmented landscape, whereas the rest of the virgin-forest species (hawks, grouses, woodpeckers, etc.) reflected more temporal variability in the moderately fragmented landscape unit. The results were only partially in accordance with expected trends, and there were differences between scales in variability patterns of species and species groups. Large-scale landscape characteristics mainly affected the temporal variability of sedentary and virgin-forest species, whereas species not restricted to special habitat structures were more affected by smaller-scale habitat structure.