Much of what we know or guess about long-term changes in productivity is derived from studying glacial/interglacial contrasts and glacial/interglacial cycles. In the area surveyed during Leg 175, variations in productivity are generated in different ways, within different geographic settings (off the Congo, near the Angola Dome, at the Walvis Ridge, and in the upwelling cells south of the Ridge; see articles in Wefer et al., 1996a). Much detailed work will be necessary to document fluctuations in productivity in these various settings and to tie the fluctuations in oceanic conditions to the corresponding changes in climate on the adjacent continent.
Work completed in preparation for Leg 175 on piston cores has shown that productivity systems are extremely sensitive to climatic change on several time scales. The influence of precessionally modulated seasonal insolation is especially strong in the tropical regions (Schneider et al., 1994, 1996, 1997; Wefer et al., 1996b), as has been known for some time (McIntyre et al., 1989; Molfino and McIntyre, 1990; McIntyre and Molfino, 1996). The effect is still strong at the latitude of the Walvis Ridge (Wefer et al., 1996b). Three factors are sufficient for obtaining a reasonable statistical model of organic carbon accumulation on the Walvis Ridge: (1) the general climatic state of the world, as reflected in a standard oxygen-isotope curve; (2) the precessional effect, as measured by insolation in July at 15°N; and (3) the decay of organic matter during diagenesis.
The example given in Figure 22 is based on total organic carbon (TOC) measurements on Core GeoB 1028 (Geosciences Bremen), located close to DSDP Site 532. The time scale is based on oxygen-isotope stratigraphy (data in Wefer et al., 1996b). The TOC model is of the form
where SL stands for sea level and is expressed as an arbitrary index, a linear transformation of the G. sacculifer isotope record in ODP 806 (806sox; Berger et al., 1995). INS is the insolation in July at 15°N, expressed as an arbitrary index ranging between 0.5 and 1.5, a linear transformation of the actual irradiation values given in Berger and Loutre (1991). The exponent a is taken as 2/3; this defines the weighting of the two explanatory variables as 2 to 1. The resulting calculated curve is shown as TOCcalc and is compared with TOCmeas (model and target; Fig. 22). The difference between model and target, the residual, shows decreasing values with depth in core, indicating the effects of diagenesis. The slope is decreasing with depth, reflecting diminishing rates of decay, with the most active destruction occurring in the uppermost meter of the sediment column. From the nature of the residual, there is a hint that the system is more sensitive toward precession during cold climate states than during warm ones on a 100-k.y. scale.
The appearance of color reflectance data at Site 1075 and subsequently occupied sites invited a search for precessional signals within color cycles. It would be of great interest to document a diminishing influence of precessional power going from the tropics to the Cape of Good Hope.
Unfortunately, the origin of color differences is not entirely clear; varying contents of carbonate and organic matter are involved, as well as other components such as terrigenous supply and authigenic minerals. Also, comparison between different holes at the same site shows that amplitudes of color variation are subject to artifact, such as exposure to air. For example, although Hole 1075A shows increasing amplitudes with depth, Hole 1075C shows the exact opposite in the long-term trend (see "Lithostratigraphy" section, "Site 1075" chapter, this volume). Short-term changes in amplitude presumably are less subject to alteration and may preserve the precessional beats.
To test this idea, we first establish a detailed time scale using magnetic susceptibility (measured continuously on unopened cores). To avoid circular reasoning, we do not use reflectance to establish a scale. Inspection of raw magnetic susceptibility data, combined with previous knowledge about sedimentation rates in the region (Schneider et al., 1994, 1997), suggested a direct correlation with the global oxygen-isotope stratigraphy. The fit is in fact excellent (Fig. 23), especially if susceptibility data are partially integrated going upward in the sequence (labeled "partially integrated susceptibility" in Fig. 23; for details, see Berger et al., Chap. 22, this volume). The integration assumes that susceptibility is spiked on deglacial transitions and has the effect of smoothing the curve and moving it toward slightly younger ages.
The time scale derived from correlation, in the fashion shown in Fig. 23, places the Brunhes/Matuyama boundary at the correct depth (i.e., just before the susceptibility maximum corresponding to Stage 19 in the correlation). With the time scale established, reflectance data were automatically dated and could be analyzed for precessional information. Results show strong precessional cycles in the red/blue ratio, especially in the last 600 k.y. (Milankovitch Chron) when the 100-k.y. cycle dominates climatic change (Fig. 24). Total reflectance, on the whole, has less prominent precessional information, but more so in the middle Pleistocene (Croll Chron, ~600–1200 k.y.). Although a correspondence of forcing and response is evident for both reflectance and color, the match is not as good as one would hope (perhaps because of the above-mentioned artifact).
Precessional information enters the record through productivity variations because of cyclic fluctuations in the wind field (McIntyre et al., 1989; Molfino and McIntyre, 1990; Schneider et al., 1994, 1996; Jansen et al., 1996; McIntyre and Molfino, 1996; Mix and Morey, 1996; Wefer et al., 1996b). Simulations with a climate model suggest that sea-surface temperature and wind fields respond to changes in seasonal insolation (Kutzbach and Liu, 1997). In particular, modeling indicates that 6000 yr ago, when seasonal contrast was greater than today because of late summer perihelion, sea-surface temperature (SST) was greater in late summer, and winds had a stronger landward component toward North Africa (Fig. 25). A landward component is also suggested for South Africa. The effect of such a change in wind field would be to weaken the trades and hence decrease upwelling in the eastern equatorial Atlantic. Also, the offshore component of flow along the southwestern margin would be weakened, reducing upwelling in the entire Benguela Current system. Whether heat transfer from the South Atlantic to the North Atlantic would increase or decrease is not clear; presumably, there would be a shift from transport by currents to transport by moist winds.
The experiment by Kutzbach and Liu (1997) illustrates how the climate record in Africa will have to be consulted to fully understand the history of upwelling in the Angola-Benguela systems. It is clear from their results that the influence of tropical dynamics will be strong well beyond the tropics. Although we expect that precessional information will tend to dominate in the tropical portion of the Leg 175 megatransect (based on the earlier findings cited), we anticipate a strong precessional signal even off the Cape of Good Hope. As mentioned, the relationships to climatic change on the continent will be of special interest (Jansen et al., 1984).