Aims (1) To define the physical correlates of indigenous forest in KwaZulu-Natal province and develop a model, based on climatic parameters, to predict the potential distribution of forest subtypes in the province. (2) To explore the impact of palaeoclimatic change on forest distribution, providing an insight into the regional-scale/historical forces shaping the pattern and composition of present-day forest communities. (3) To investigate potential future shifts in forest distribution associated with projected climate change.Location KwaZulu-Natal province, South Africa.Methods A BIOCLIM-type approach is adopted. Bioclimatic 'profiles' for eight different forest subtypes are defined from a series of grid overlays of current forest distribution against nineteen climatic and geographical variables, using ArcInfo GIS grid-based processing. A principal components analysis is performed on a selection of individual forests to identify those variables most significant in distinguishing different forest subtypes. Five models are developed to predict the distribution of forest subtypes from their bioclimatic profiles. Maps of the potential distribution of forest subtypes predicted by these models under current climatic conditions are produced, and model accuracy assessed. One model is applied to two palaeoclimatic scenarios, the Last Glacial Maximum (LGM) (≈18,000 BP) and the Holocene altithermal (≈7000 BP), and to projected future climate under a doubling in global atmospheric carbon dioxide.Results Seven variables; altitude, mean annual temperature, annual rainfall range, potential evaporation, annual temperature range, mean annual precipitation and mean winter rainfall, are most important in distinguishing different forest subtypes. Under the most accurate model, the potential present-day distribution of all forest subtypes is more extensive than is actually observed, but is supported by recent historical evidence. During the LGM, Afromontane forest occupied a much reduced and highly fragmented area in the midaltitude region currently occupied by scarp forest. During the Holocene altithermal, forest expanded in area, with a mixing of Afromontane and Indian Ocean coastal belt forest elements along the present-day scarp forest belt. Under projected climatic conditions, forest shifts in altitude and latitude and occupies an area similar to its current potential and more extensive than its actual current distribution.Main conclusions Biogeographical history and present physical diversity play a major role in the evolution and persistence of the diversity of forest in KwaZulu-Natal. It is important to adopt a long-term and regional perspective to forest ecology, biogeography, conservation and management. The area and altitudinal and latitudinal distribution of forest subtypes show considerable sensitivity to climate change. The isolation of forest by anthropogenic landscape change has limited its radiation potential and ability to track environmental change. Long-term forest preservation requires reserves in climatically st...

Aim To examine the influence of climatic extinction filtering during the last glacial maximum (LGM; c. 18,000 yr bp) and of the subsequent recolonization of forest faunas on contemporary assemblage composition in southern African forests.Location South Africa, Mozambique, Swaziland, Zimbabwe.Methods Data comprised presence/absence by quarter-degree grid cell for forest-dependent and forest-associated birds, non-volant mammals and frogs. Twenty-one forest subregions were assigned to one of three previously identified forest types: Afrotemperate, scarp, and Indian Ocean coastal belt. Differences among forest types were examined through patterns and gradients of species richness and endemism, assemblage similarity, species turnover, and coefficients of species dispersal direction. The influence of contemporary environment on assemblage composition was investigated using partial canonical correspondence analysis. Several alternative biogeographical hypotheses for the recolonization of forest faunas were tested.Results Afrotemperate faunas are relatively species-poor, have low species turnover, and are unsaturated and infiltrated by generalist species. In northern and central regions, communities are supplemented by recolonization from scarp forest refugia, and among frogs by autochthanous speciation in localized refugia. Scarp faunas are relatively species-rich, contain many forest-dependent species, have high species turnover, and overlap with coastal and Afrotemperate faunas. Coastal forests are relatively species-rich with high species turnover.Main conclusions Afrotemperate communities were affected most by climatic extinction filtering events. Scarp forests were Afrotemperate refugia during the LGM and are a contemporary overlap zone between Afrotemperate and coastal forest. Coastal faunas derive from post-LGM colonization along the eastern seaboard from tropical East African refugia. The greatest diversity is achieved in scarp and coastal forest faunas in northern KwaZulu-Natal province. This historical centre of diversity has influenced the faunal diversity of nearly all other forests in South Africa. The response of vertebrate taxa to large-scale, historical processes is dependent on their relative mobility: forest birds best illustrate patterns resulting from post-glacial faunal dispersal, while among mammals and frogs the legacy of climatic extinction filtering remains stronger.

Available empirical evidence suggests that many primate populations are increasingly threatened by anthropogenic actions and we present evidence to indicate that Africa is a continent of particular concern in terms of global primate conservation. We review the causes and consequences of decline in primate diversity in Africa and argue that the major causes of decline fall into four interrelated categories: deforestation, bushmeat harvest, disease and climate change. We go on to evaluate the rarity and distribution of species to identify those species that may be particularly vulnerable to threats and examine whether these species share any characteristic traits. Two factors are identified that suggest that our current evaluation of extinction risk may be overly optimistic; evidence suggests that the value of existing forest fragments may have been credited with greater conservation value in supporting primate populations than they actually have and it is clear that the extinction debt from historical deforestation has not being adequately considered. We use this evaluation to suggest what future actions will be advantageous to advance primate conservation in Africa and evaluate some very positive conservation gains that are currently occurring.

No abstract

No abstract