Bird community variation across Polylepis woodland fragments and matrix habitats: implications for biodiversity conservation within a high Andean landscape

摘要

The scattered and dwindling Polylepis woodlands of the high Andean global hotspot have been identified as being of particular importance to biodiversity conservation, and yet little is known of the make-up of their faunal communities, how these vary across landscapes, and how well species might tolerate matrix/edge habitats. We examined the bird communities and vegetation characteristics of Polylepis woodlands and the surrounding matrix habitats at three sites in the Cordillera Vilcanota, southern Perú (3,400–4,500 m). The vegetation structure of woodlands varied significantly across the three sites but all were dominated by two Polylepis tree species, with mossy ground cover. Matrix habitats were treeless and dominated by ground-level puna grass-steppe or boulder scree vegetation. Bird species richness and diversity, encounter rates and the number of globally-threatened and restricted-range bird species were consistently higher in the Polylepis forests, than in matrix habitat. We used canonical correspondence analysis (CCA) to identify habitat gradients across the landscape, and to classify bird species according to their association with Polylepis, the matrix or Polylepis-matrix interface. There were few matrix-restricted bird species, but around half the bird community, including fourteen threatened or restricted-range species, were Polylepis-dependant. Many of these species had very narrow niches. The Polylepis-matrix interface was dominated by species traditionally considered invasive ecological generalists. Our study illustrates the overriding importance of Polylepis interior habitats, indicating that conservation strategies for high Andean birds must focus on patch size maintenance/enlargement, enhancement of within-patch habitat quality, and efforts to safeguard connectivity of suitable habitat across what is essentially an inhospitable puna/scree matrix.