In order to reconstruct intermediate water circulation of the equatorial Atlantic and paleoproductivity of the equatorial divergence, silt grain-size analysis were performed on sediments of the north-western African continental slope (ODP Site 959) and on pelagic sediments of the Guinea Basin (ODP Site 663) for the last 1 Ma. Silt grain-size records of ODP Site 959 monitor distinct fluctuations of the intermediate water circulation of the equatorial Atlantic especially during terminations. Based on the terrigenous "sortable silt" fraction, grain-size spectra of the terrigenous silt and their statistic grain-size parameters, variations of the UNADW/GNAIW circulation could be deduced due to glacial/interglacial changes. The observed gradients should have been strengthened by the exposed setting of ODP Site 959. During the period from 1 Ma-today a low but continous sedimentation was determined at Site 959. The investigated sediments were divided into three facies types due to their compositional, structural and textural characteristics. Special focus was set on the distribution of the terrigenous sortable silt fraction and their statistic grain-size parameters. Bioturbated sediments of Facies Type A reflect an increasing influence of hemipelagic sedimentation whereas laminated sediments (Facies Type C) could be classified as deposited by contour currents. The third facies type (Type B) of rest sediments is characterised by erosive contacts and by periods of almost pure clayey and sandy deposits with reduced silt fraction controlled by strong or growing influence of contour currents. Obviously Facies Type B mainly occured on terminations where pulses of grain-size finings shifted to coarse silt mean values. These silt-sized minima display characteristic peaks on termination IX-V and monitor increased contour current intensities of the intermediate watermasses. In comparision with silt grain-size records of Site 14 K (Manighetti et al., 1995) in a similar water depth on the Rockall-Plateau the following interpretations could be inferred: The fluctuations of the bottom water intensities image variable production rates and depth ranges of intermediate water masses. During glacial times LNADW production was strongly reduced due to increased sea-ice coverage. According to this reduction of the LNADW formation in the GIN-Seas the source area of LNADW has been shifted to the southern Labrador Sea. There one separate intermediate water mass, the Glacial North Atlantic Intermediate Water (GNAIW), which is predominantly attributed to LSW has been formed. In the North Atlantic the GNAIW is located at a water depth domain between 1000 and 2000 m whereas it extends till 2400 - 2600 m water depth in the South Atlantic (e.g. deMenocal etal., 1992). Hence the reduced formation of LSW during glacial times caused the glacial enhanced current intensities observed. The total volume of GNAIW was increased significantly by this high production in contrast to the interglacial UNADW, LSW respectively which led to intense glacial current intensity. Hence during interglacial periods of strong thermohaline circulation current velocities were relatively low. The characteristic minima of the silt mean data on terminations could be attributed to highest GNAIW production at the end of glacial times; the following shift to higher silt mean values displayed lower production of GNAIW.
The total phyto- and zooplancton of the equatorial divergence is distinctly dominated by carbonate producing organisms. Consequently based on the grain-size signature of the biogenous carbonate at ODP Site 663 the paleoproductivity of the equatorial divergence could be carried out in high resolution for the last 1 Ma. At this Site the data of the silt analysis identified the production of both, carbonate phyto- and zooplancton quantitatively and enable balancing of the processes due to their contribution to pC02• To calculate the carbonate loss according to water mass induced or metabolic carbonate dissolution a multi parameter approach of several carbonate dissolution indicators has been performed. Therefore especially the fractionation index of planctic foraminifera is figured out to react most sensitive. During the MBE and the MPT enhanced carbonate dissolution occurred due to stronger entering and mixing of AABW, which caused a rise of the lysocline level. The contribution of metabolic, supralysoclinal carbonate dissolution is relative low in comparison with coastal upwelling areas. Except of IS 2 values of the Carbon Rain Ratio didn't exceed 0.02 (Bickert et al., 1996). During the remaining periods the system of the equatorial divergence is considered to be dominantly controlled by productivity. To reconstruct the paleoproductivity on Site 663 the total amount of foraminifera and coccolith carbonate has been determined based on LECO data and component analysis of the sand fraction of the last 1 Ma. Compared to calculations of Samthein et al. (1992) and of Rühlemann (1996) significant correspondence and some differences could be inferred. According to Berger & Herguera (1992) paleoproductivity calculations based on Corg represent the primary production. The (%) coccolith carbonate calculations of core 663 display the carbonate primary production. The correspondence of both parameters is high in general and points out the usefulness of the methods applied. Deviations could be caused by maintenance effects of Corg, an imprint by terrigenous organic material or changes in the contribution of opaline primary producers. The comparison of the foraminifera and coccolith carbonate with the percentages of coarse and fine silt pointed out a strong correspondence of (%) coarse silt and foraminifera carbonate and of (%) fine silt and coccolith carbonate. On one hand percentages of coarse silt and foraminifera carbonate are generally enhanced during interglacials and during the warmer periods of the glacials. On the other hand coccolith carbonate and (%) fine silt reach highest values during glacial times and during the cold stages of the interglacials. These distribution patterns were caused by glacial enhanced upwelling intensity of the equatorial divergence due to glacial increased trade wind zonality and/or increased advection of the Benguela Current in the Equatorial Atlantic. SEM-examinations carried out a precoincidance of the coccolith species C. leptoporus and Gephyrocapsa sp. in the silt fraction of core 663, whereas the clay fraction is represented mainly by the species Gephyrocapsa sp. and E. huxleyi (since IS 5/4).
This pattern could be attributed to the different grain-size ranges of the single species. In the data of the fine silt mean a tendency to glacial enhanced and interglacial reduced phi-values is observed, which displays a predominance of the bigger species C. leptoporus at interglacials and of the smaller species Gephyrocapsa sp. during glacial times. The carbonate content of the clay fraction is increased during interglacial times, which could reflect enhanced occurrence of the species E. huxleyi (since IS 5/4). Based on the approach performed in this study the production of carbonate primary and secondary producers has been displayed and could be used to balance the processes of the "biological pump" and the "carbonate pump" against each other. Therefore the relationship between coccolith carbonate/carbonate bulk has been worked out as a useful parameter to compare the processes of photosynthesis and carbonate production. The distribution of this parameter characterises the glacials as time periods of enhanced appearance of autotrophic primary producers. This pattern can document the stronger glacial activity of the "biological pump" in high production areas in comparison with the "carbonate pump" and could be an explaination of the interglacial pC02 rise observed. In this scenario the behaviour of coccoliths as a source or a sink of C02 still remains a question.
