Forest patch size and isolation as drivers of bird species richness in Maputaland, Mozambique
Article first published online: 29 JUN 2010
© 2010 Blackwell Publishing Ltd
Journal of Biogeography
Volume 37, Issue 10, pages 1884–1893, October 2010
How to Cite
Guldemond, R. A. R. and van Aarde, R. J. (2010), Forest patch size and isolation as drivers of bird species richness in Maputaland, Mozambique. Journal of Biogeography, 37: 1884–1893. doi: 10.1111/j.1365-2699.2010.02338.x
- Issue published online: 29 JUN 2010
- Article first published online: 29 JUN 2010
- Conservation biogeography;
- feeding guilds;
- forest birds;
- fragment size;
- island biogeography;
- number of species
Aim To examine the response of forest-dependent and generalist bird assemblages and feeding guilds to patch characteristics of forest fragments in the context of island biogeography and metapopulation theories in an area that supports high levels of species diversity and endemism.
Location Dune, riverine, sand and swamp forest fragments in southern Mozambique’s Maputaland.
Methods We recorded 20 forest-dependent and 69 generalist bird species at 220 survey points in 30 forest fragments that ranged in size from 5 to 7432 ha and that were 100 to 5100 m apart. We used linear regressions to relate the number of species to forest fragment size, isolation, perimeter, fractal dimension, shape and core area for frugivores, insectivores, granivores, nectarivores, carnivores and omnivores separately for forest-dependent and generalist bird assemblages. We tested for modality in occupancy frequency distributions of species in forest fragments and used the binary matrix temperature calculator to determine whether the assemblages had a nested structure separately for forest-dependent and generalist bird species.
Results The number of forest-dependent and generalist bird species categorized into the feeding guilds varied independently from forest patch parameter values. Occupancy models showed a random distribution for forest-dependent birds, and generalists had a unimodal distribution. Both the forest-dependent and generalist bird assemblages had a non-nested structure. The presence of eight rare forest-dependent and four endemic bird species in 22 of the 30 fragments contributed to the non-nested structure of these bird assemblages.
Main conclusions Fragmented forests did not induce the expected responses of the forest bird assemblages in that the number of species was not explained by patch characteristics, nor could we find evidence for metapopulation dynamics. Non-nestedness may be caused by the few rare and endemic species occupying forest fragments with a wide range of patch characteristics. Knowledge of the response of species to forest fragmentation can benefit conservation planning catering for rare and endemic species. Our study suggests that all forest fragments in the region may have to be included in a conservation network to complement existing landscape-based plans to protect the region.