INTRODUCTION

During Ocean Drilling Program (ODP) Leg 175, 13 sites were drilled off the western coast of Africa from 5° to 32°S latitude (Shipboard Scientific Party, 1998a) (Fig. F1), the tip of southern Africa to the Congo River. This area is the location of the Angola-Benguela Current (ABC) system, a major component of South Atlantic Ocean circulation and one of the greatest upwelling areas of the world (Berger et al., 1998a). The aim of this study is to reconstruct the climatic history along the coast of Africa over the past 1.5 m.y. using a multiproxy approach and, therefore, employing several methods and techniques to obtain records of different proxy indicators. In addition, by comparing records from the two different areas drilled during ODP Leg 175, Sites 1076 and 1077, located in the Congo Basin at a latitude of 5°S, and Site 1081, situated farther south at the Walvis Ridge at 17°S, local vs. global influences can be established.

In particular, records from these areas can be used to consider factors such as the source and supply of terrestrially derived material to the South Atlantic, the compositions and volumes of which may provide evidence for climatic change on the African continent, and the variability and evolution of upwelling and productivity, often assumed to represent a response to changes in the marine environment (e.g., Siesser, 1980; Hay and Brock, 1992). In addition, these two factors may interact; for example, there is evidence for influence of continentally derived iron on marine productivity (Martin, 1992). As a consequence, these records can also be used to consider the coupling and decoupling of marine and continental influences in the area.

Furthermore, by investigating the past 1.5 m.y., it is possible to establish the influence of both long- and short-term climatic variability on the marine and continental environments in addition to determining the influence of the mid-Pleistocene revolution (MPR) and, therefore, to consider the response of the oceans and continents to such climatic fluctuations and events.

The South Atlantic Ocean plays a crucial role in global climate by transferring a huge amount of heat to the North Atlantic (Reid, 1996), and because of this, the South Atlantic is quite different in character from the North Atlantic. The North Atlantic pays for its heat with nutrient deficiency because heat arrives in nutrient-depleted waters (Berger and Wefer, 1996), and the surface waters of the North Atlantic are warm and nutrient poor in comparison to the cold, upwelled, nutrient-rich waters of the South Atlantic. However, evidence suggests that this was not always the case and that heat transfer may have been reduced during glacial periods (Raymo et al., 1990; Williams et al., 1998).

A major component of the heat transfer system from the South Atlantic is the ABC system, which consists of the Angola Current and the Benguela Current (Fig. F2) (Oberhänsli, 1991) and is one of the major upwelling systems in the world. Currently, the cold, nutrient-rich waters of the Benguela Current flow northward, parallel to and within ~200 miles of the coast of southwest Africa. At ~20°S latitude, these waters meet the southward-flowing Angola Current and develop the Angola-Benguela Front (ABF) (Meeuwis and Lutjeharms, 1990). At this front, the Benguela Current is deflected west and merges with the South Equatorial Current.

The present-day latitude of the ABF coincides approximately with the location of the Walvis Ridge. The Walvis Ridge forms a barrier to the northward and southward flow of deep water below 3000 m (Shannon and Nelson, 1996) and effectively separates two areas of intense upwelling. North of the ridge, upwelling and productivity are dominated by Congo River outflow and the Angola Dome, an area of offshore upwelling generated by a sluggish cyclonic gyre formed by the meeting and mixing of the Benguela Current and the Angola Current (Berger et al., 1998a). To the south, wind-driven upwelling interacts with the northward flow of the Benguela Current to create upwelling mainly on the landward side of the current (Berger et al., 1998a).

Studies into the evolution of the Benguela Current demonstrated that it has gradually strengthened since its initiation during the late Miocene (Siesser, 1980), causing it to flow farther north before turning westward and, therefore, shifting its associated upwelling cells northward (Siesser, 1980; Diester-Haass et al., 1992). Hypotheses on the glacial-interglacial variability of the ABC system, however, have proved contentious. Many studies assume that the ABF and the westward deflection of the current shifted north during glacial periods in connection with the global increase in ocean circulation (McIntyre et al., 1976; Jansen et al., 1996), and reconstructions of the last glacial maximum suggested that the current extended into the Gulf of Guinea, or even as far north as the equator (McIntyre et al., 1989). Yet, other studies indicate only a slight intensification of the Benguela Current and a limited shift of the ABF by a few degrees (Jansen et al., 1984; Diester-Haass, 1985), or no significant shift at all (Schneider et al., 1995).

Records from the African continent suggest that the continent has suffered a gradual increase in cooling and aridity over at least the past 2 m.y. (Dupont et al., 1989; Tiedemann et al., 1989) in addition to short-term variations in response to glacial-interglacial cycles. The issue of the relative influences of high-latitude global climate change vs. low-latitude local fluctuations in climate is, however, unresolved (deMenocal et al., 1993). The effects of these forcing factors can often be considered in terms of the dominant cyclicities in proxy climate records because high-latitude forcing is attributed to changes in global ice volumes and sea levels and the 100- and 41-k.y. eccentricity and obliquity cycles, whereas low-latitude changes represent changes in precipitation and the monsoons, which have been shown to respond to the 23-k.y. precessional cycle.

Various climatic modeling attempts suggested that tropical climate change over Africa was driven by either one or the other of these factors. For example, pelagic sediments from the eastern equatorial Atlantic showed increased quartz fluxes during glacial periods (Tiedemann et al., 1989), indicating that the African continental climate was more arid during glacial maxima and, therefore, that tropical climate change responded passively to high-latitude climatic change. In contrast, evidence from eastern equatorial sea-surface temperatures (SSTs) showed 23-k.y. periodicities, suggesting the association of upwelling with wind-driven changes in precession (McIntyre et al., 1989) and, therefore, that the South Atlantic and adjacent African continent was driven directly by low-latitude changes. However, deMenocal et al. (1993) ascertained that both these forcing factors have a role to play. Eolian dust and phytolith records from ODP Site 663, located off the northwest coast of Africa, are dominated by 100- and 41-k.y. cycles, yet freshwater diatom records, which provide evidence of lake levels and precipitation, manifest the 23-k.y. cycle.

Sites 1076 and 1077 are located in the Congo Basin. Hole 1076A is at a shallow-water site, located in 1402 m water depth. The sediments have generally high organic carbon values and consist of olive-gray and greenish gray clays. Sedimentation rates vary between 50 and 210 m/m.y., and there is much evidence of reworked material (Shipboard Scientific Party, 1998c). Hole 1077A sediments are from an intermediate water depth of 2394 m and are composed of greenish gray diatom- and nannofossil-rich clays (Shipboard Scientific Party, 1998d).

A complex environment exists in the Congo Basin, dominated by river input, seasonal coastal upwelling, and incursions from the South Atlantic Ocean (Shipboard Scientific Party, 1998a). The Congo River supplies freshwater, nutrients, and terrigenous material, including freshwater diatoms, phytoliths, and clay minerals, all of which contain evidence of climate change in the drainage basin of the river to the ocean and generated the Congo Fan, which extends for >1000 km into the Angola Basin (Jansen et al., 1984). The fan consists of fairly fine-grained, muddy material, as much of the sandy sediment is trapped in or close to the mouth of the river.

The Walvis Ridge is a basaltic abutment formed from hotspot activity during the Early Cretaceous period (Dean and Gardner, 1985), located at a latitude of 20°S off the coast of Namibia. It extends southwestward from the continental margin for >2500 km toward the Mid-Atlantic Ridge (Shannon and Nelson, 1996). Hole 1081A is at a shallow-water site, located at a 794-m water depth on the ridge. The sediments consist of gray clays with varying amounts of diatoms, nannofossils, foraminifers, radiolarians, and authigenic minerals such as glauconite, pyrite, and dolomite. Sedimentation rates are high and vary from 70 to 150 m/m.y. (Shipboard Scientific Party, 1998e). This site is located directly below the upwelling center on the only topographic high in the area. It provides a comparison with records from ODP Site 532 and Deep Sea Drilling Project (DSDP) Site 362, which are located on the eastern side of the ridge seaward of the upwelling center but contain an upwelling signal that has been transported by the Benguela Current (Shipboard Scientific Party, 1998a).

The MPR refers to the transition observed in proxy climatic records from symmetrical low-amplitude, high-frequency (41 k.y.) ice volume variations to high-amplitude, low-frequency (100 k.y.) asymmetrical sawtoothed ice volume variations indicating gradual ice buildup terminated by rapid deglaciation events (Broecker and van Donk, 1970). The MPR resulted in a change in the mean state of the global climate system, including lower global temperatures, increased global ice volume, and lower sea-surface temperatures (Shackleton et al., 1990).

The timing, duration, and cause of the MPR, however, are something of a mystery. The first 100-k.y. cycle in ¹⁸O records is generally observed at marine isotope Stage 22/23 (0.9 Ma) (Shackleton and Opdyke, 1977). Yet, the classical sawtoothed large-amplitude fluctuations characteristic of the late Pleistocene do not appear until isotope Stage 17 (0.7 Ma) (Shackleton et al., 1990). Studies regarding the timing of the MPR both in terms of its midpoint and the rate of change frequently disagree, and there is evidence to indicate that the MPR may have occurred as early as 1.2 Ma or as late as 0.4 Ma and that its duration was as little as 50 k.y. or as great as 500 k.y. (Prell, 1982; Ruddiman et al., 1989). Yet, it is known that the MPR was a global event, and its occurrence has been documented in both marine and continental records worldwide.

Most importantly, however, the MPR demonstrates that the causal link between insolation and ice volume suggested by Milankovitch (1930; Imbrie et al., 1984) is, in fact, more complex than it first might appear. During the Pliocene and early Pleistocene, it appears that a linear relationship between orbital forcing and ice volume and climatic variations existed (Imbrie et al., 1992). Yet, reconstructions of insolation values (Berger and Loutre, 1991) suggest that there was no significant change in the pattern of insolation at the time of the MPR to account for the transition observed in climatic variations from 41- to 100-k.y. cycles.

Similarly, Imbrie et al. (1993) demonstrated that the amplitude of the 100-k.y. forcing on insolation is at least one order of magnitude smaller than the same insolation signal in the 23- and 41-k.y. bands. Yet, the response of the climate system in these two bands combined is less than one-half the amplitude observed in the 100-k.y. period. Therefore, although the proxy climatic and ice volume records are dominated by the 100-k.y. cycle, the power of the 100-k.y. eccentricity signal on insolation is essentially zero (Shackleton and Opdyke, 1977).