Did forests survive the cold in a hotspot?

Article metrics

Tropical forests harbour a large proportion of the world's biodiversity, they act as a major reservoir for carbon in the global cycling of that element, and they are under threat. Yet, surprisingly, little is known about their detailed history — particularly in Africa. A group of European palaeoecologists1 has now sought to close this gap in our knowledge with an intensive study, published in the Journal of Biogeography, of the highland region of east-central Africa, running south from the Virunga volcanoes in Uganda, through Rwanda into Burundi (Fig. 1). They conclude that current biodiverse ‘hotspots’ were not always so species rich.

Figure 1: Map of east Africa, showing the swamp locations studied by Jolly et al.1.

Their studies of these five locations indicate that the tropical forest in this region came either from a core refugium in the west which survived the climatic shocks of the Pleistocene, or one that was derived from dispersed patches of forest. (Figure adapted from ref. 1.)

A number of questions need to be answered regarding the history of the forests in tropical Africa. To what extent were they affected (and possibly fragmented) by the colder climates of the high-latitude glacial episodes? Have they expanded, contracted or moved in the form of intact communities? Or have they, like their temperate counterparts, changed their composition as individual taxa have responded to climate change according to their own requirements? We also need to know whether the current biodiversity hotspots are located in sites where relict fragments of the forest survived the cold, dry climate of the last glacial events. And to what extent has human activity — even early in prehistory — modified the pattern of forest spread in the Holocene (the past 10,000 years)?

Figure 2

Montane vegetation, to the east of the sites studied by Jolly et al.1.

Assembling all of the available information, in 1982, Hamilton2 suggested that certain core areas, probably resulting from refuges that had survived the climatic shocks of the Pleistocene (roughly the past two million years), acted as centres for forest spread and recolonization of central Africa during the Holocene. Among these cores were eastern Zaire (including the western fringes of Uganda, Rwanda and Burundi, bordering the Western Rift Valley), the coastal area of eastern Tanzania, and the west coast of Africa in Cameroon/Gabon. Possible minor core areas were also proposed, in northern Mozambique and in Ethiopia. According to Hamilton, the plant and animal species spread from these postulated refugia as they moved into the intervening areas and reassembled the interglacial forest.

To put this hypothesis to the test, palaeoecological information is needed to document patterns of vegetation change in these regions. Moreover, a number of sites must be investigated, to detect differential movements and to avoid the problems that arise from the possible use of an atypical site. In the latest study by Jolly et al.1, five suitable sites were located along the highlands on the eastern side of the Western Rift Valley, running in a north-south transect of about 300 km. All sites were peat-accumulating swamps within the broad-leaved forest zone. Their depths ranged from 7 to 23 metres, and all contained well-preserved pollen. Some of the cores contained material dating from as far back as 18,000 years before present (BP), thus taking the vegetation history record back to the coldest stage of the last glacial episode.

The oldest sediments contained pollen typical of dry ericaceous scrub and grassland. This indicates that the area supported little forest during the height of the cold episode. There is evidence for the elements of montane forest, but the pollen could have been transported over some distance in this generally tree-less landscape. Alternatively, there could have been patches of montane forest in particularly favourable locations. Expansion of the montane forest varies in its timing between the sites but, unexpectedly, it seems to occur at the higher-altitude locations before the lower ones (about 11,000 and 10,000 yr BP, respectively). But precise radiocarbon dating is hampered by a radiocarbon plateau, making resolution difficult3.

It is apparent from the pollen diagrams that the composition of the richer forest — which follows after about 10,000 yr BPat all sites — varies between the locations. The genus Alchornea (Euphorbiaceae), for example, expands more rapidly and to higher levels at the northern end of the transect than in the south. By contrast, Macaranga (also Euphorbiaceae) is more abundant and persistent in the south. Increasing levels of scrub and forest-edge indicators in recent times (from about 2,300 yr BPonwards) suggest forest disturbance by human activity, but the time at which these clearances began varies between sites.

In palaeoecological research, especially in east Africa, single-site studies are the norm. But the regional, multi-site study by Jolly et al.1 provides at least some of the answers to the historical biogeographical questions of the Old World tropics. If there was a core refugium of forest in this region, it certainly did not extend east of the Western Rift Valley, where the uplands were covered by ericaceous scrub during the height of the cold period. Either the refugium was further west in Zaire or, perhaps, we should be considering the survival of much smaller, dispersed patches of forest rather than a major refugial forest block.

When the climate became warmer and wetter in the early Holocene, equatorial forests did not march up the mountains in a phalanx. They seem, instead, to have dropped behind the lines in guerrilla fashion. They then expanded from the favoured locations in which they established themselves, sometimes moving downwards from higher altitudes. Palaeoecological studies of temperate forest4 indicate that the forest composition varied from site to site, depending both on local environmental conditions and the expansion patterns of the different taxa. The specific constitution of the forest has been dynamic in both time and space, most recently because of human disturbance and clearance for agriculture. The same seems to be true of tropical forest.

Biodiversity hotspots in the tropics are unlikely to be explained simply on the basis of refugium theory. But survival history has undoubtedly played an important part in their current richness. The range of opportunities presented in the Holocene development of ecosystems — both in terms of microhabitat diversity and disturbance features — must also have contributed to the variety of species combinations that can be achieved and maintained in these areas. Such is the essence of a biodiverse hotspot.


  1. 1

    Jolly, al. J. Biogeogr. 24, 495–512 (1997).

  2. 2

    Hamilton, A. Environmental History of East Africa (Academic, London, 1982).

  3. 3

    Becker, B., Kromer, B. & Trimborn, P. Nature 353, 647–649 (1991).

  4. 4

    Jacobson, G. L., Webb, T. & Grimm, E. C. in The Geology of North America Vol. K-3 (eds Ruddiman, W. F. & Wright, H. E.) 277-288 (Geol. Soc. Am., New York, 1987).

Download references

Author information

Rights and permissions

Reprints and Permissions

About this article

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.