Assessment of the Neotropical dry forest refugia hypothesis in the light of palaeoecological data and vegetation model simulations

Abstract

Abstract Seasonally dry tropical forests exhibit a markedly discontinuous, disjunct distribution pattern across South America. Previous botanical inventories have revealed that these dry forest areas exhibit close floristic links with one another, which has led to the hypothesis that they constitute modern‐day refugia, resulting from vicariance (i.e. fragmentation) of a formerly more widespread, continuous distribution during more arid climatic conditions associated with the Last Glacial Maximum (LGM). This paper examines the validity of this ‘dry forest refugia’ hypothesis in the light of previously published fossil pollen data and vegetation model simulations. The model simulations suggest that major dry forest areas, such as the Bolivian Chiquitano Dry Forest, are not Pleistocene refugia, but instead developed during the Holocene by population expansions from parts of southern Amazonia presently covered by evergreen rainforest. The limited pollen data available indicate that glacial‐age seasonally dry tropical forests, irrespective of their distribution, were floristically quite different from those of today, with key dry forest species, such as Anadenanthera , only appearing in the Holocene. I suggest that a more parsimonious explanation for the current biogeographic pattern of these South American dry forests is population migration (perhaps by long‐distance dispersal) since the LGM, rather than vicariance. Copyright © 2004 John Wiley & Sons, Ltd.