Upwelling off southwest Africa is at present centered on the inner shelf and at the shelf edge. The Benguela Current flows roughly parallel to the coast and within ~180 km of it south of 25°S, and then turns to the west over the Walvis Ridge between 23° and 20°S (Fig. 8). At about 20°S, warm, tropical-water masses from the north meet the cold Benguela Current water. Eddies of cold, upwelled water contain radiolarian and diatom skeletons, which are transported from the upwelling area to the northern part of the Walvis Ridge, where they have been sampled at Deep Sea Drilling Project (DSDP) Sites 532 (Hay, Sibuet, et al., 1984) and 362 (Bolli, Ryan, et al., 1978).
According to previous studies, during the last glacial maximum (LGM) eddies formed farther north and the Benguela Current flowed parallel to the coast and over the Walvis Ridge to reach the Angola Basin, finally bearing to the west at about 17°S. Sediments deposited at Site 532 during the LGM apparently confirm the absence of upwelling eddies by containing zero to very few opal skeletons (Hay, Sibuet, et al., 1984; Diester-Haass, 1985). Upwelling may have continued to occur on the African shelf, but the Benguela Current then did not transport that upwelling signal to the Walvis Ridge. However, from the distribution of foraminiferal assemblages at Site 532, it appears that the northeastern Walvis Ridge was in fact characterized by intensified upwelling and a westward expansion of coastal upwelling cells at glacial periods during the last 500,000 yr (Oberhänsli, 1991). The issue of contrasting models of glacial/interglacial upwelling dynamics in this region is unresolved. It hinges on the question of why opaline fossils show contrary abundance variations, with respect to the productivity record from other proxy indicators.
The results from Sites 362 and 532 can be used to reconstruct, tentatively, the evolution of the Benguela Current during the past 10 m.y. This evolution is characterized, on the whole, by increasing rates of accumulation of organic carbon. In addition, there are indications from changing correlations between percent carbonate, percent Corg, and diatom abundance that the dynamics of the system undergo stepwise modification. In this connection, as well, a distinct opal maximum in the early Quaternary is of great interest (Fig. 9). The nature of this transition is not clear; perhaps it is a response to the migration of the polar front to its modern position.
The evolution of the climate of the northern hemisphere, and particularly that of northern Europe, is linked to the exchange of heat between the South Atlantic and the North Atlantic Oceans (Fig. 10). This energy transport, operating over large distances, is involved in the formation and magnitude of polar ice caps. In today's world, a net heat transfer from the South Atlantic to the North Atlantic exists in currents above the thermocline (Fig. 11). A part of the heat contribution from the South Atlantic is believed to originate from the Indian Ocean via the Agulhas Current. The Benguela Current is a connection between the waters north of the polar front in the South Atlantic and the Equatorial Currents of the Atlantic. Northward and southward shifts of the Southern Ocean polar front constrict or expand, respectively, the interchange of heat from the Indian Ocean to the South Atlantic (McIntyre et al., 1989). This interchange presumably has a drastic impact on the heat budget of the Benguela Current and, consequently, that of the entire Atlantic Ocean. Such variations in heat transfers should appear as changes in the course and intensity of currents and productivity regimes and should be recorded in the sediments accumulating along the southwest African margin.
An important element of the heat transfer dynamics is the deep-circulation pattern. Traditionally, the focus in reconstructing this pattern has been on the properties and boundaries of NADW related water masses, as seen in the delta13C of benthic foraminifers. The emphasis has been on glacial-to-interglacial contrast (Fig. 12). This contrast shows that NADW production was greatly reduced during glacial periods (as also reflected in the pattern of carbonate preservation). More recent studies have added much detail to this story (summarized in Bickert and Wefer, 1996) (Fig. 13). It appears that the strength of the NADW is reflected in the differences between eastern and western basins and in gradients within the eastern basin. Information on associated changes at depths above the NADW has been sparse. It must be assumed that the strength of the nutrient maximum underlying the Benguela upwelling regions (Fig. 14) is somehow coupled to the evolution of NADW, which in turn influences dynamics of intermediate water-mass formation to the south. At this point, we do not know how the different cycles are related, so little or nothing can be said about causal relationships.
Paleoceanographic interpretations regarding the history of the Benguela Current are derived mainly from a single location off southwest Africa (Site 532) and must be considered preliminary. Given the indications that the axis and the intensity of the Benguela Current have changed over the past 15 m.y. and that productivity has fluctuated with glacial/interglacial cycles, confirmation and refinement of these ideas is needed. Although DSDP Legs 40, 74, and 75 occupied sites in the Cape and Angola Basins and on the Walvis Ridge, these sites are situated too far offshore to provide the needed information. The Benguela Current and its associated upwelling are not recorded well in the sediments at these sites. Even Sites 362 and 532 on the Walvis Ridge are too far offshore to contain a direct record of upwelling. They receive an indirect record of near-coastal upwelling from material transported to their location by the Benguela Current. Furthermore, modern coring technology (advanced hydraulic piston corer [APC], extended core barrel [XCB]) allows for high-resolution studies by avoiding much of the drilling disturbance present in the Leg 40 cores. Such high-resolution work is crucial if the dynamics of upwelling are to be captured back to the Miocene on a scale of glacial/interglacial cycles. Information from an array of sites situated in the southern and central Cape Basin, on the Walvis Ridge, and in the southern Angola Basin would allow the construction of a coherent picture.