Changes in the vegetation and trade winds in equatorial Northwest Africa 140,000-70,000 yr B.P., as deduced from two marine pollen records

The deep-sea cores M 16415-2 and M 16416-2 at about 9°N off Sierra Leone were analysed palynologically for the time interval 140,000-70,000 yr B.P. Results were presented in absolute (pollen concentration and pollen influx) and relative diagrams (pollen percentage). In a previous study it was evidenced that in northwest Africa pollen is mainly transported to the Atlantic by wind, so that the efficiency of aeolian pollen transport (pollen flux) could be used to evaluate changes in the intensity of the northeast trade winds. The glacial episodes (represented by the oxygen isotope stages 6 and 4) are characterized by strong northeast trade winds, whereas the last interglacial (stage 5) is characterized by weak trade winds. The pollen influx diagram shows that the intensity of the trade winds increased slightly during the relatively cool intervals of stage 5 (viz. 5.4 and 5.2). Tropical forest had maximally expanded around 124,000 yr B.P. (stage 5.5), around 98,000 yr B.P. (transition of stage 5.3 to 5.2), and around 70,000 yr B.P. (first part of stage 4): an increasing delay of the response of tropical forest to global intervals with maximum temperature is apparent during the last interglacial. As tropical forests need continuous humidity, the record of tropical forest monitors changes in climatic humidity south of the Sahara. During the last interglacial, the southern boundary of the Sahara shifted only little: expansions and contractions of the tropical forest area are correlated with contra-oscillations of the grass-dominated savanna zone. Great latitudinal shifts of the desert savanna boundary, on the contrary, occurred during the penultimate glacial interglacial transition (around 128,000 yr B.P.) to the north, and during the last interglacial-glacial transition (around 65,000 yr B.P.) to the south.

Geochemical composition of snow samples from Antarctica, Greenland and northeast Canada

In this paper, we present new detailed data on the trace metal content of more than 200 shallow polar snow samples collected at various depths in numerous locations mainly in Antarctica and Greenland. The samples were collected in ultraclean plexiglass or teflon tubes from the walls of hand dug pits, using stringent contamination free techniques controlled by severe blank tests. They were then analysed for Na, Mg, K, Ca, Fe, Al, Mn, Pb, Cd, Cu, Zn and Ag in clean room conditions by flameless atomic absorption, after a preconcentration step (by non boiling evaporation in teflon bulbs) which includes dissolving any solid particles by concentrated nitric and hydrofluoric acids. The overall precision on the measured concentrations is of the order of 10 % for all the metals except Pb (20 %) and Cd (35 %), using 95 % confidence limits. The data obtained are compared with those published previously in the literature. Part of these previous data are shown to be erroneously too high, probably because of con-tamination problems both during field collection and analysis.