The biogeography of forest birds in the Limpopo Province, South Africa.

Le résumé fourni par la source

Forest assemblage composition is determined by local ecological (e.g. patch area, species interactions), landscape (e.g. patch connectivity) and regional (e.g. historical change in forest distribution) processes. I investigated the relative effect of these processes on bird and frog assemblage composition in two isolated archipelagos of Afrotemperate forest in the Limpopo Province. The linear relationship between local and regional species diversity suggests that forest bird assemblages in the Limpopo Province are unsaturated. In addition, 66% of bird species and 42% of frog species in southern African forests are generalist species (i.e., forest associated as opposed to forest dependent), suggesting that matrix species have invaded forest assemblages. I thus argue that forest bird and frog assemblage composition is primarily determined by regional (historical) processes and that local ecological processes play a relatively minor role. Forests in the Limpopo Province were eliminated by major climatic changes during the Quaternary with major forest expansion only in the last 6000 years. Limpopo Province forest assemblages have thus established fairly recently. No forest dependent frogs and one forest dependent bird have established in the Limpopo Province forests from the relatively proximate forests in eastern Zimbabwe. This suggests that the Limpopo River catchment has acted as a significant barrier to the dispersal of forest vertebrate faunas. Cluster analyses showed that the forest bird and frog assemblages are essentially Afrotemperate and South African in origin with all forest dependent frogs and 97% of forest dependent birds occurring in the KwaZulu-Natal scarp forests. In addition the most important environmental gradient of change in the southern African forest bird faunas was the geographical distance from northern KwaZulu-Natal. This gradient is congruent with a major northward radiation of faunas from the KwaZulu-Natal scarp into the Limpopo Province. As a result the Limpopo Province forests have low biodiversity values compared to the KwaZulu-Natal scarp because forest frog and bird faunas are largely derived from the latter region. However, the importance of the Limpopo Province forests lies in their protection of threatened vertebrates as well as in providing landscape heterogeneity and ecological services to the surrounding matrix. Soutpansberg forest bird assemblages appear to be more robust and resilient and comprise a significantly greater proportion of forest associated species than those of the Limpopo Province Drakensberg. This is likely to be a consequence of more severe climatic extinction filtering of these faunas caused primarily by the proximity of the Soutpansberg forests to the arid Limpopo valley during the development of these forests. Consequently, regional and historical processes have played a relatively greater role in determining forest bird assemblages in the Soutpansberg than in the Limpopo Province Drakensberg and species richness in the former region was not significantly affected by local ecological processes (including forest area, isolation and habitat heterogeneity). Forest area and habitat heterogeneity did, however, affect forest bird species richness and abundance in the Limpopo Province Drakensberg where the relatively lower importance of regional processes (compared to the Soutpansberg) has combined with anthropogenic disturbance of smaller forests to increase the influence of local ecological processes. However, the role of local processes in determining local species richness is likely to increase in both archipelagos if the current rates of anthropogenic change and disturbance to forests are sustained. Forests greater than 138 ha (minimum critical patch size) are needed to avoid an island effect on bird species richness in the Limpopo Province Drakensberg. However, the long-term conservation of vertebrate assemblages in Limpopo province forests depends upon the successful conservation of evolutionary and landscape processes. This can best be achieved by maximising forest connectivity and landscape heterogeneity through the protection of both riparian corridors and forests of all sizes. The maintenance of historical dispersal routes, in particular connectivity along the escarpment with the scarp forests of KwaZulu-Natal, is important. This would require the protection of forests on the KwaZulu-Natal scarp and along the entire northern Drakensberg escarpment.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.