Are modern pollen data representative of west African vegetation?

Vegetation reconstructions are proposed in central Atlantic and north-west Africa from modern pollen data derived from an original data set of 452 samples and 1075 pollen taxa along a north-south transect from Congo to Egypt. We use multivariate statistical methods to link pollen samples to vegetation types, climate parameters (both annual and seasonal) and biomes. Analysis is based on a precise relationship between pollen assemblages and vegetation types and a re-assessment of Plant Functional Types and biomes at a sub-continental scale. The results show that the 13 main vegetation types described in western Africa are correctly defined by pollen data (both main pollen types and percentages) and both annual and seasonal climate parameters can be assigned to each of them. This includes distinction between the northern and southern types of dry rain forest from Cameroon and Congo and between tropical and non-tropical desert areas. Biomes are correctly reconstructed for all the vegetation types except for the tropical dry forest (TDFO) in the Sudanian vegetation zone (probably due to the strong anthropogenic disturbance of the natural landscape) and the tropical rain forest (TRFO) near the Equator (due to the highly complex biology of the rain forest, still poorly studied).