Isolation, patch size and matrix effects on bird assemblages in forest reserves

Habitat fragmentation is one of the most studied topics in ecology but our knowledge is still limited particularly concerning matrix effects on species distribution in a human-dominated landscape. We tested the ability of random sampling hypothesis, colonization–extinction dynamics and matrix-related concepts to explain the variation in species richness, total bird density and community composition in old-forest bird assemblages in two contrasting landscapes. We collected data on breeding bird abundances from 66 old-growth forest reserves in NE Finland and six larger areas in adjacent Russian Karelia using the line transect method. In Finland, protected old-forest patches are embedded in a matrix dominated by young regeneration stands. In Russia, study areas were situated in continuous, old forest dominated landscapes. Bird assemblages in old-forest patches embedded in human-modified matrix in Finland were not random samples from Russian bird assemblages. In the Finnish assemblages, species richness was lower and total bird density higher. Species richness declined with increasing distance (isolation) to Russia. Bird assemblages in large forest reserves in Finland close to Russia were structurally more similar to assemblages in the continuous reference landscape than those in small and more distant reserves. The results support the idea that several mechanisms related to colonisation–extinction dynamics and to matrix resource availability influence species distribution in fragmented landscapes but in species-specific ways. We conclude that even though small and isolated protected areas may foster high relative bird species density their ecological integrity is compromised, and therefore, improving matrix quality around reserves may lead to considerable conservation benefits.