The Paleogene represents a climatically dynamic period in Earth history. Stable isotope and other temperature proxy records reveal a complex history of warming and cooling, characterized by periods of both gradual and rapid change (Miller et al., 1987; Miller and Katz, 1987; Stott et al., 1990; Zachos et al., 2001, 1994). Major events include a 1-m.y.-long global warming trend that began in the late Paleocene and climaxed in the early Eocene in a 1- to 2-m.y.-long climatic optimum (early Eocene Climatic Optimum [EECO]) and a 12-m.y.-long stepped cooling trend that began in the early middle Eocene and culminated in the earliest Oligocene with the appearance of continental-scale ice sheets (Hambrey et al., 1991; Zachos et al., 1992). One of the more prominent events is a transient but extreme greenhouse interval known as the Paleocene–Eocene Thermal Maximum (PETM). This event occurred at ~55.0 Ma. Major changes in ocean chemistry as inferred from shifts in carbon isotope patterns and distribution and preservation patterns of terrigenous and biogenic sediments on the seafloor (e.g., Bralower et al., 1995; Kennett and Stott, 1991; Robert and Kennett, 1997) characterize the PETM. In addition, distinct shifts in the distribution of key groups of fauna and flora occurred in the oceans and on land (e.g., Kelly et al., 1998; Koch et al., 1995, 1992; Thomas and Shackleton, 1996; Thomas, 1998; Wing, 1998). Another notable event is the earliest Oligocene Glacial Maximum (EOGM, or Oi-1), a brief but extreme glacial interval that occurred at ~33.4 Ma and marks the transition to permanent glacial conditions on Antarctica (e.g., Miller et al., 1987, 1991; Zachos et al., 1996). This event, like the PETM, has caused large-scale perturbations in ocean chemistry and paleoecology (Barrera and Huber, 1993, 1991; Salamy and Zachos, 1999; Thomas and Gooday, 1996; Thunell and Corliss, 1986).

Multiple hypotheses exist to explain the large-scale, long-term changes in Paleogene climate, although none have yet gained universal acceptance. In general, among many factors, the role of ocean gateways (continental geography) and greenhouse gas levels are considered as key variables. Theoretical models have invoked either the absence of a circum-Antarctic current or higher greenhouse levels or some combination of both to account for the EECO (Barron, 1985; Bice et al., 2000; Sloan and Barron, 1992; Sloan and Rea, 1996; Sloan et al., 1995, 1992). Similarly, the Oligocene glaciation has been attributed to both the initiation of the Antarctic Circumpolar Current and a reduction in greenhouse levels and (e.g., Kennett and Shackleton, 1976; Mikolajewicz et al., 1993; Oglesby, 1991; Raymo et al., 1990; Rind and Chandler, 1991). Some of the more abrupt transient excursions are more likely to have been forced by rapid changes in greenhouse gas levels because they occur over short timescales (e.g., 10³–10⁴ yr) and, most importantly, are accompanied by geochemical and isotopic anomalies suggestive of major perturbations in the carbon and sulfur cycles (Dickens et al., 1997, 1995; Paytan et al., 1998; Pearson and Palmer, 2000; Schmitz et al., 1997; Stott et al., 1990; Zachos et al., 1993).

Progress in characterizing Paleogene oceanography and climate history, particularly the transient events, has been hampered by the lack of high-quality, high-resolution multicored sequences. Most sites cored prior to Leg 198 suffer from poor recovery and drilling disturbance, and few were multicored or drilled as part of depth transects. The few exceptions are sites recovered during recent Ocean Drilling Program (ODP) legs including Sites 865, 999, 1001, 1051, and Bass River (Bralower et al., 1997, 1995; Miller et al., 1998; Norris and Röhl, 1999; Röhl et al., 2000, 2001, 2003). High-resolution records produced from these sites yield a wealth of exciting, important evidence of climate change to be more fully explored with additional data.

The ODP extreme climate advisory panel (Program Planning Group [PPG]) recognized the dearth of high-resolution records across climate transients, and the panel formulated new questions concerning extreme climates (Kroon et al., 2000) and potential drilling targets among which was the Walvis Ridge area. The ODP Leg 208 proposal, which resulted from the PPG discussions, described as its major goal the acquisition of sediment archives necessary to characterize such short-term changes in ocean chemistry and circulation that theoretically should have accompanied these climatic extremes, as well as their effects on the oceanic biota. Acquisition of more complete records of these events at a higher temporal resolution is required for both formulating and testing hypotheses on the origin of these and similar events.

Walvis Ridge, located in the eastern South Atlantic Ocean (Fig. F1), is one of the few known locations where it is possible to recover Paleogene sediments, including the PETM and EOGM, over a broad range of depths, including depths exceeding 4.5 km. The ridge was the target of drilling by Deep Sea Drilling Project (DSDP) Leg 74, which occupied Sites 525–529 on the northern flank of the ridge between depths of 2.5 and 4.2 km (Moore, Rabinowitz, et al., 1984). Paleogene pelagic sediments characterized by moderate sedimentation rates (~6–15 m/m.y.) and good magnetic stratigraphy were recovered at each site. However, because of the poor recovery (~50%–75%) and coring disturbances, especially with the rotary core barrel in unlithified sediments, only short segments of the sequences were recovered fully intact, and none of the sequences were double cored. Technical problems, combined with the lack of high-resolution shipboard core logs, limited high-resolution cyclostratigraphic investigations to a few short segments of the Cretaceous (Herbert and D'Hondt, 1990). Nevertheless, subsequent shore-based studies of low-resolution samples collected from these cores were instrumental in adding to our understanding of long-term Maastrichtian and Paleogene paleoceanography of the South Atlantic Ocean (e.g., calcite compensation depth [CCD], carbon isotope stratigraphy, deep-sea temperature/ice volume) (e.g., Moore, Rabinowitz, et al., 1984; Hsü and Weissert, 1985, and papers within; Shackleton, 1987). Nearly complete PETM intervals were recovered at the shallowest and deepest Sites 525 and 527. Stable isotope analysis of foraminifers recovered from these sites have helped to constrain the magnitude of the deep Atlantic biogeochemical and environmental changes during this event (Thomas et al., 1999; Thomas and Shackleton, 1996). At the remaining sites, the PETM was not recovered because of core gaps (poor recovery).

During the winter of 2000, a seismic survey of the southeastern Walvis Ridge was undertaken by the Meteor (Cruise M49/1). The survey extended coverage of the Leg 74 sites to the north and northeast, where more continuous and slightly thicker sediment sequences were discovered (Fig. F2). The higher fidelity multichannel seismic data generated during the survey allowed identification of several areas where the PETM and other critical intervals could be recovered by advanced hydraulic piston corer/extended core barrel (APC/XCB) drilling. The Leg 208 sites cover a paleodepth range of 2.3 km, sufficient to constrain large and subtle changes in the chemistry of deep and intermediate waters. Advances in coring technology and drilling strategies (i.e., multiple hole composite sections) allowed for 100% recovery of sequences that were only partially recovered during Leg 74. In addition, recent advances in data acquisition and cyclostratigraphy enabled high-precision correlation and dating of these sediments.

Walvis Ridge is a northeast-southwest-trending aseismic ridge that divides the eastern South Atlantic Ocean into two basins, the Angola Basin to the north and the Cape Basin to the south (Figs. F1, F3). The ridge consists of a series of interconnected crustal blocks that slope gradually toward the northwest and more steeply toward the southeast. Magnetic and gravity anomalies indicate that Walvis Ridge was formed by hotspot volcanism near the spreading ridge as the Atlantic basin gradually widened (Rabinowitz and Simpson, 1984) and is subsequently assumed to have followed a simple thermal subsidence model with ~1.1 km of subsidence since the Maastrichtian (Moore, Rabinowitz, et al., 1984).

Pelagic sediments drape most of the ridge and generally increase in thickness toward the continental margin (Moore, Rabinowitz, et al., 1984). In the vicinity of the primary Leg 208 target area on the southeastern portion of the ridge, sediment thickness varies from ~300 m on the deep (>4.5 km) seafloor adjacent to the ridge to ~600 m near the summit (~2.5 km), a pattern that is clearly expressed by seismic sequences in the multichannel seismic profiles (Fig. F4). The sediments are primarily calcareous oozes and chalks that range in age from Campanian to Holocene. The Neogene sequences consist primarily of nannofossil and foraminifer nannofossil oozes with relatively high carbonate contents, often in excess of 90%, although much lower values were recorded in lower through middle Miocene sections. Turbidites and slumps are present in some intervals. The underlying Paleogene sediments are dominated by nannofossil and foraminifer-bearing nannofossil oozes to chalks. Carbonate contents are also high, generally in excess of 80% through most of the Paleocene, Eocene, and Oligocene with the exception of several short, carbonate-poor intervals at the deeper sites (e.g., Paleocene and upper Eocene of Site 527) that represent episodes of CCD shoaling. A few thin chert layers are present below the upper Paleocene of the shallowest Sites 525 and 528 and in the lower Eocene of Site 529. Slump deposits of various scales are present in the upper Paleocene at Site 529. In most sections, calcareous microfossil preservation varies from good to excellent. The natural remnant magnetism of the Walvis Ridge sediments appears to be strong and stable. As a result, the quality of the magnetic polarity records is excellent, particularly in the Upper Cretaceous and lower Paleocene intervals (Chave, 1984).

Sediment accumulation rates on Walvis Ridge have varied considerably with time. On average, the highest rates (~8–13 m/m.y.) occur in the Pleistocene, Paleocene, and Maastrichtian. Much lower sedimentation rates (1–5 m/m.y.) as well as unconformities are common at most sites in the Neogene. For example, most of the Pleistocene is absent at the shallowest sites (Sites 525 and 526) whereas the lower and middle Miocene is absent at several of the deeper sites. The Neogene unconformities appear to be predominantly erosional in nature although dissolution clearly affected deposition downslope of the ridge (i.e., lower–middle Miocene at Site 527). Unconformities are also present in the upper Eocene and Oligocene. In contrast, at middepths (Site 529), most of the Paleogene is present but the middle–upper Miocene is either condensed or absent. In the deepest section (Site 527), the middle–upper Eocene is relatively condensed and the Oligocene and Miocene are highly condensed and/or absent. Nevertheless, it appears that for most of the early Paleogene, deposition was more or less continuous over much of the ridge. This continuity in sediment accumulation is reflected to some extent in the similar patterns in the low-resolution carbon isotope stratigraphies for four of the Leg 74 sites (Shackleton and Hall, 1984; Shackleton, 1987).

Cyclical variations in sedimentation are evident in various lithologic and physical property indexes, particularly in the more expanded Maastrichtian and Paleocene sequences. Spectral analysis of sediment color banding reveals the presence of a strong precessional beat in upper Maastrichtian and lower Paleocene sediments at Sites 525 and 528 (Herbert and D'Hondt, 1990). Similar cycles, although more subtle, may be present in the upper Paleocene and lower Eocene.