H. Elenga et al., Use of plots to define pollen-vegetation relationships in densely forestedecosystems of Tropical Africa, REV PALAE P, 112(1-3), 2000, pp. 79-96
Modern soil samples from South Congo were analyzed for pollen content and c
ompared to forest inventories to define modern pollen-vegetation relationsh
ips. A correspondence analysis (CA) was applied independently to botanical
and pollen data and a hierarchical cluster analysis to pollen data only. Su
bsequently, a CA using a presence-absence approach has been made to directl
y compare the two types of data. Results show that the pollen rain and flor
istic composition of the sampled sites are not directly linked to altitudin
al or precipitation gradients, but clear evidence of variation in relation
to hygromorphy and soil type is detected. The forests occurring in swampy e
nvironments are well differentiated from the forests developed on well-drai
ned soils by pollen and floristic data. Among forests on well-drained soils
, a good distinction can be made between those growing on sandy soils and t
hose growing on ferralitic soils. The comparison between pollen spectra and
vegetation shows site-to-site variations in pollen assemblages in relation
to the floristic heterogeneity of forests, and it appears that few taxa sh
ow a good correlation between plant cover and pollen abundance. (C) 2000 El
sevier Science B.V. All rights reserved.