Inasmuch as upwelling along the southwestern coast of Africa is driven by winds and is closely tied to the strength of the offshore currents, it reflects processes important in the heat transfer from the South Atlantic to the North Atlantic Oceans. However, no simple relationship between productivity and heat transfer can be assumed.
At least two factors introduce important complications. The first is the back pressure developed from a glaciated Northern Hemisphere, which moves the Intertropical Convergence Zone from its northern position toward the equator (Flohn, 1985). The back-pressure problem demands that productivity patterns be reconstructed equally well for the systems off Iberia and northwest Africa for comparison with the southern African regions.
The second factor concerns changes in the nutrient content of subsurface waters feeding the upwelling system (Hay and Brock, 1992; Lange and Berger, 1993; Herguera and Berger, 1994). Changes in nutrient content of subsurface waters ("thermocline fertility") are readily appreciated when comparing depositional rates of opal and organic matter (Berger et al., 1994, 1997; Berger and Wefer, 1996; Schneider et al., 1996, 1997). Thus, great care must be taken to capture both the influence of physical mixing and upwelling (which depend on the wind field) and that of thermocline fertility (which depends on the quality of upper intermediate waters, among other things).
The MOM off southwestern Africa is of special interest in the context of evolution of the Benguela Current and the concomitant development of thermocline fertility. It is centered near 2.2 Ma and follows a rapid increase of diatom productivity near 3 Ma. In its early stage, it is marked by a strong Antarctic component, as well as by an admixture of warm-water forms. After MOM time, diatomaceous remains typical of strong coastal upwelling dominate. This sequence suggests a greater availability of silicate in subsurface waters during the MOM than either before or after. We hypothesize that the MOM marks a time when the Benguela Current (and all other currents interacting with it at its point of origin) were flowing less vigorously than today (Fig. 26). The silicate front (Fig. 27A) was less well developed than in the Quaternary, and subsurface waters were richer in silicate north of 40°S than they are today (e.g., Westall and Fenner, 1990, p. 780). Because the Agulhas Retroflection was less active to the west of the Cape of Good Hope (it depends on the inertia of a strong Agulhas Current), waters of Antarctic origin were entrained sporadically into the region off southwest Africa up to Namibia ("AAE" in Fig. 26B). Poleward undercurrent flow ("PUC" in Fig. 26) was stronger during MOM time, being less opposed than today by the Benguela Current and northward-flowing Antarctic Intermediate Water.
Our scenario (Fig. 26B) suggests a number of tests:
Our hypothesis states that the system goes through an optimum with respect to diatom deposition. It implies that warming moves us closer to the optimum; hence, interglacials should have greater opal deposition on and south of the Walvis Ridge, and indeed they do (Diester-Haass et al., 1990, 1992; Hay and Brock, 1992). Also, north of the ridge, a more normal pattern should be found, with glacial sediment having the greater opal content (as described by Schneider et al., 1996, 1997). Furthermore, before reaching the MOM, more opal should accumulate on and south of the ridge during cold periods because cooling moves the system closer to the optimum. There are indications that this may be so (Diester-Haass et al., 1992; Hay and Brock, 1992).
These distinct patterns on various time scales suggest to us that upwelling is not the only factor controlling opal deposition; the quality of the upwelled water is equally important (cf. Barron and Baldauf, 1989; Baldauf and Barron, 1990). Increased flow of the Benguela Current presumably parallels increased upwelling; how-ever, it also seems to be associated with processes that decrease the silicate content of waters within the thermocline off southwestern Africa. Analogous observations have been made elsewhere (Berger et al., 1997). Clues to the mechanisms responsible for these patterns will be found in the phase relationships between different productivity proxies and between productivity and the ice-age cycles of temperature and sea level.
The Agulhas Current Retroflection ("AgR" in Fig. 26A) is thought to play a major role in providing warm water to the Benguela Current near its point of origin (Lutjeharms, 1996; Shannon and Nelson, 1996). The Agulhas Current overshoots beyond the Cape of Good Hope and spawns large eddies, which enter the Benguela Current as the Agulhas water turns back toward the Indian Ocean (Fig. 28). The history of the import of Indian Ocean water (some 6 sv at present; 1sv = 106 m3/s) clearly is of great importance to the question of heat transport in the Benguela Current system. We suspect that there is an optimum for warm-water acquisition when the inertia of the Agulhas Current is great enough to inject waters well into the South Atlantic Ocean, but the countervailing currents (South Atlantic Current and Antarctic Circumpolar Current) were not as strong as they are today. It is possible that such optimum injection was contemporaneous with the MOM, in which case our hypothesis would have to be modified to account for sporadic warm-water incursions from the Indian Ocean into the Cape Basin.