IODP Proceedings    Volume contents     Search

doi:10.2204/iodp.proc.342.101.2014

Background and objectives: Newfoundland sediment drifts

Earth’s history holds a rich record of abrupt climate and ecological changes. However, the dynamics of these state changes in Earth’s ecosystems are often not well resolved, particularly in past warm climates that have analogs to future global change. The complexity of Earth’s biological and climatological systems hinders the use of models to confidently predict future ecosystem dynamics, yet the planet is rapidly approaching a highly altered climate forcing that has no historical analog for the past 30 m.y. Our approach is to seek historical analogs to future climates in Earth’s past. However, to understand the dynamics of rapidly shifting environments, it is critical that we target records that have adequate fidelity to study the processes we seek to understand.

A long-standing approach to understanding Paleogene and Cretaceous oceanography has been to drill sediments that record the long sweep of Earth’s history. This methodology has allowed reconstruction of ocean climate and ecosystem changes over the past ~180 m.y. and has allowed many “extreme climate” events to be placed into context both with respect to broad climate trends and increasingly well resolved chronologies. Such deep time records, however, necessarily involve a trade-off between outlining long histories and understanding ocean dynamics on timescales similar to (or shorter than) the ~1500 y mixing time of the ocean. The vast majority of work on Pliocene–Pleistocene climate dynamics long ago began to target archives with finely resolved chronologies (millennial to century scale) and, in the case of ocean sediments, high deposition rates. This focus on highly resolved records has the potential to reveal dynamics that operate on timescales that have relevance to human society. Expedition 342 is founded on the objective of extending this dynamics-focused approach, for the first time, into the archives of the past greenhouse Earth.

Our program requires a high sedimentation–rate record of the Eocene and Oligocene in which we can resolve both rates and magnitudes of past extreme climate events as well as the background variability in climate and ecosystems from these times. Because we target sediment drifts that accumulate faster than typical deep-sea sediments, we should be able to reconstruct the history of a warm Earth with unusual fidelity. A related objective is to acquire these records along a depth transect of drill cores between ~5 and 3 km water depth. Because the ocean is layered, with different water masses formed in various regions around the world arranged above one another, our depth transect of drill sites will permit a detailed reconstruction of the chemistry, circulation, and history of greenhouse Earth’s oceans. These two things—an unusually detailed climate history and a detailed assessment of the structure and circulation of the warm-world ocean—will help us test models of Earth’s climate and ecosystem evolution that have been difficult or impossible to resolve with typical deep-sea or land-based records of the warm Earth and its transition into the icehouse world.

Geological record of abrupt climate dynamics as future-Earth analogs

The primary objective of Expedition 342 is to answer a series of pressing questions about the rate and magnitude of past ecosystem changes:

  • How similar are past abrupt climate changes to model expectations for the environmental changes triggered by human modification of Earth’s environments?

  • What are the consequences for Earth’s biota of variations in magnitude or rate of change during abrupt events?

  • Are there any ancient analogs to what we are doing to the planet now or has humanity launched a carbon cycle perturbation without geological analog?

  • How long will perturbations similar to those of the “Anthropocene” last?

  • Are we missing fundamental feedbacks in our bid to model the Earth’s future under higher greenhouse gas forcing?

  • What is the climate sensitivity to different rates or magnitudes of greenhouse gas forcing?

The Eocene is not a perfect analog to the near future (e.g., Haywood et al., 2011) but understanding Eocene climate dynamics will provide information on what to expect from a warmer planet.

Rapid ecosystem change on a warm Earth

A particular goal of Expedition 342 is to reconstruct the carbonate lysocline to provide constraints for carbon cycle models. An extremely successful result of Ocean Drilling Program (ODP) Leg 208 (Zachos, Kroon, Blum, et al., 2004) was the acquisition of a ~2000 m depth transect that permitted estimation of the magnitude of carbonate compensation depth (CCD) fluctuations in the Paleogene, particularly the Paleocene/Eocene Thermal Maximum (PETM) (Zachos et al., 2005). This depth transect showed that the CCD excursion associated with the PETM was >2000 m and suggested that the amount of CO2 added to the biosphere during the PETM must have been much greater than previously documented (Panchuk et al., 2008). In turn, these data were interpreted to suggest that the source of greenhouse gases could not have been primarily gas hydrates because the combined magnitude of the temperature excursion, the CCD fluctuation, and the size of the δ13C anomaly all suggest a carbon source more similar to marine organic carbon than bacterial methane (Panchuk et al., 2008). This example emphasizes how important it is to capture the magnitude of these carbonate preservation excursions because they provide a direct estimate of the change in ocean undersaturation and hence the transient change in CO2 storage during extreme climate events. When combined with δ13C data sets and a range of numerical modeling approaches, the size of the CCD excursion can be used to study the source of isotopically light carbon transferred during such events and, ultimately, climate sensitivity to CO2 forcing (e.g., Pagani et al., 2006a; Goodwin et al., 2009; Dunkley Jones et al., 2010; Dickens, 2011; DeConto et al., 2012; Pälike et al., 2012).

Understanding the spatial and depth dimensions of the CCD change is clearly key to modeling the likely sources and amounts of carbon added to the biosphere during the PETM, but our current observational data do not yet provide a clear picture of either variable. If the methane hydrate hypothesis is untenable for the PETM and other extreme warm transients—so called “hyperthermals”—then it becomes even more important to identify a viable alternative. This task would be made much easier if we had a clear idea of the total amount of CO2 put into the ocean and atmosphere during the PETM and other Paleogene hyperthermal events. A well-resolved depth transect that could be used to identify the full depth range of the CCD excursion is critical to resolving the amount and type of carbon introduced into the biosphere during Paleogene hyperthermals. Recent evidence that the PETM and Paleogene global warming events may be orbitally forced or at least conditioned (Lourens et al., 2005; Galeotti et al., 2010; Sexton et al., 2011) implies that the source of carbon was not an extraterrestrial impact or one-time volcanic eruption (Kent et al., 2003; Svensen et al., 2004). Thus, long astronomically calibrated records of CCD changes through the warm Eocene may be critical to unraveling the origins of the PETM and other Paleogene events (e.g., the Middle Eocene Climatic Optimum [MECO], a prominent reversal in the long-term global cooling trend from peak Cenozoic warmth in the early Eocene) (Bohaty et al., 2009; Bilj et al., 2010).

The carbonate content of deep-sea sediments is a sensitive measure of the productivity, weathering, and fluxes of carbon within the biosphere. The CCD reflects changes in carbon erosion, deposition, and silicate weathering and, in association with δ13C of marine and terrestrial proxies (benthic foraminifers, terrestrial biomarkers, and planktonic–benthic gradients), can be used to gauge changes in the size and distribution of carbon reservoirs (Higgins and Schrag, 2006; Leon-Rodriguez and Dickens, 2010). However, although the general form of CCD changes has long been moderately well known for the various ocean basins (van Andel, 1975; Peterson et al., 1992), we are only just starting to gather well-resolved records at the secular scale and to lay the foundations to reveal the rates and amplitude of change in the CCD on short timescales. This information is very important for understanding transient episodes such as the various extinction events, hyperthermals, and glacial events of the Paleogene (e.g., Pälike et al., 2012). Expedition 342 is well placed to provide records of CCD change from the North Atlantic Ocean at unprecedented temporal resolution across the largest Paleogene water depth transect ever attempted.

In addition to reconstructing the sources and fate of carbon that produces extreme climate events, we also need to know the rate and full magnitude of these events and the response of marine biota. In most deep-sea records, slow sedimentation rates mean that the full magnitude of the events may have been bioturbated or dissolved away (Zeebe and Zachos, 2007). The first solution to the problem of resolving transient events is to target sites where sedimentation rates are higher than the ocean average. Expedition 342 is aimed first and foremost at obtaining highly expanded records suitable for studies of Paleogene climate by targeting drift sediments of hypothesized Oligocene–middle Eocene age. The depth transect approach and acquisition of expanded carbonate-rich sections near the limits of their northern extent will allow us to study faunal and floral processes such as population turnover and latitudinal and stratigraphic range expansion during intervals of ocean acidification and global warming (e.g., Gibbs et al., 2006; Edgar et al., 2010, 2013).

The record on the Newfoundland ridges should allow a companion study to Walvis Ridge drilling (Leg 208) to be conducted under the flow path of northern-source deep water. We have not previously obtained a good record of the history of the carbonate chemistry of the Deep Western Boundary Current in the North Atlantic. The Leg 208 transect on Walvis Ridge is likely influenced mainly by southern source waters by virtue of its location on the eastern side of the South Atlantic Basin. Furthermore, we can see in the Newfoundland seismic data that the toe of J-Anomaly Ridge (at ~5 km water depth) likely lies close to the average depth of the CCD because the drift package thins markedly at this point (Fig. F1). Therefore, drilling on J-Anomaly Ridge, combined with the shallower sites on Southeast Newfoundland Ridge, positioned us to recover a depth transect spanning >2.5 km (Fig. F2). It is of course possible that the amplitude of CCD shift during the most extreme Paleogene transient events could still exceed our transect, but we are better placed to obtain a record of the full magnitude of CCD excursions than previous studies. The transect is particularly suited to resolving fluctuations in the carbonate chemistry of the truly deep ocean, a large change from most previous ocean drilling sites that recovered Paleogene records, which rarely targeted paleodepths below ~2500 m.

Rapid ecosystem changes on a cold Earth

In many ways the EOT represents the obverse of the PETM. Across the EOT, the Pacific Ocean appears to undergo pronounced deacidification associated with the onset of major Cenozoic polar ice sheets (Fig. F3) (Coxall et al., 2005; Merico et al., 2008). Ultimately, ice sheet initiation appears to have been triggered by an orbitally forced interval inhibiting warm summers, but some other factor, probably long-term drawdown in atmospheric carbon dioxide levels, must have conditioned the climate system (DeConto and Pollard, 2003; Coxall et al., 2005; DeConto et al., 2008). Records from the tropical Pacific (ODP Leg 199; Lyle, Wilson, Janecek, et al., 2002) and IODP Expedition 320/321 demonstrate that the CCD deepened permanently (by ~1 km) in the earliest Oligocene in two 40 k.y. “jumps” and in lockstep with the stepwise onset of Antarctic glaciation, as recorded in benthic foraminifer δ18O (Coxall and Wilson, 2011; Coxall et al., 2005). However, it is unclear to what extent this lockstep behavior represents the global picture because sufficiently complete EOT depth transects from other ocean basins and latitudes are not available. Comparatively low resolution records across a highly condensed and reworked Eocene/Oligocene boundary from Walvis Ridge (Leg 208; Zachos, Kroon, Blum, et al., 2004) have been interpreted in terms of a much smaller (~200 m) and nonpermanent CCD deepening. Furthermore, a new reconstruction of the carbonate compensation system based upon Leg 199 and Expedition 320/321 results suggests that the CCD in the equatorial Pacific underwent temporary large-amplitude changes in the middle and late Eocene (Pälike et al., 2012). In the end, it is not clear whether CCD behavior in the equatorial Pacific represents the global picture, driven by wholesale change in deep-sea CO3 ion concentration, or is to some extent a regional signal driven by changes in export production with important implications for our understanding of contemporaneous carbon cycling (Dunkley Jones et al., 2008; Coxall and Wilson, 2011; Pälike et al., 2012).

The large but temporary CCD events of the middle Eocene equatorial Pacific, or “calcium carbonate accumulation events” (Lyle et al., 2005), have been interpreted (by analogy with the EOT and on the basis of discontinuous stable isotope records) to be a lithologic expression of large-scale bipolar continental glaciation (Tripati et al., 2005, 2008), but this interpretation is contested (Edgar et al., 2007; Eldrett et al., 2007, 2009; DeConto et al., 2008). The middle Eocene to lower Oligocene depth transect on the Newfoundland ridges presents an ideal opportunity to generate high-resolution records of changes in mass accumulation of CaCO3 needed to test the hypothesis that the CCD changes recently documented in the equatorial Pacific are globally representative and intimately associated with the onset of continental glaciation. Together with the floral, faunal, and geochemical proxy records, such as compound-specific δ13Corg and δ11B in foraminiferal calcite (e.g., Pagani et al., 2005; Pearson et al., 2009) that could be generated, it would be possible to test for biotic turnover and pCO2 drawdown across these intervals, with important implications for our understanding of the causes and consequences of Cenozoic cooling.

History of the Deep Western Boundary Current

Numerous models suggest that thermohaline circulation can exist in multiple equilibrium states (e.g., Marotzke and Willebrand, 1991), with the potential for catastrophic disruption or collapse. For example, in the recent past the warmth and stability of interglacial climates is notably distinct from the rapid and abrupt cycles of warming and cooling so striking in the late Pleistocene glacials (e.g., Dansgaard et al., 1993). It has been proposed that enhanced climate instability during cold periods may originate from salinity feedbacks, which set up multiple thermohaline modes (Keeling and Stephens, 2001; Rahmstorf, 1995; Stommel and Arons, 1960), or from instabilities that lead to periodic purges and the collapse of large ice sheets (Alley and Clark, 1999). Analogs to Pleistocene suborbitally paced ice-rafting cycles are documented in the latest Pliocene and earliest Pleistocene North Atlantic associated with the intensification of northern hemisphere glaciation and development of a Laurentide ice sheet (Bailey et al., 2010, 2012). Yet the pre-Pliocene history of continental ice sheet variability in the northern hemisphere is controversial (e.g., Tripati et al., 2005, 2008; DeConto et al., 2008; Edgar et al., 2007), and we know nothing of Paleogene suborbital variability in climate and ocean structure in the North Atlantic.

Deep ocean circulation is crucial to defining the global climatic regime and its stability for two reasons. First, the sinking of surface water distributes thermal energy around the globe. Second, the chemistry and nutrient composition of the dominant deepwater masses influence the partitioning of CO2 between the deep ocean and the atmosphere. The pattern, strength, and stability of thermohaline circulation arise from a combination of wind forcing (Toggweiler and Samuels, 1995) and thermal or salinity density contrasts.

Something akin to modern North Atlantic Deep Water appears to be absent until the late Miocene or early Pliocene, when a combination of tectonic subsidence of the Greenland-Scotland Ridge and northern hemisphere refrigeration began to form cold, dense overflow waters in the Nordic seas and the Labrador Sea (Oppo et al., 1995; Wright and Miller, 1993; Wright et al., 1992). However, a northern component of deep water clearly formed in the North Atlantic throughout the Neogene and Oligocene, judging from geochemical differences between Atlantic and Indo-Pacific waters (Wright et al., 1992). Intensification of deepwater formation in the North Atlantic is proposed to account for acceleration of the Deep Western Boundary Current during the Oligocene, leading to widespread erosion along continental margins and formation of the seismic reflection, Horizon Au, in the western North Atlantic. Subsequent current-controlled sedimentation formed major sediment drifts throughout the North Atlantic (Arthur et al., 1989; Miller and Fairbanks, 1985; Miller et al., 1987; Tucholke and Mountain, 1986; Tucholke, 1979; Tucholke and Vogt, 1979).

Determining the flow history of the Atlantic Deep Western Boundary Current

Direct evidence for significant flow in a deep boundary current before the Oligocene is sparse (Davies et al., 2001; Hohbein et al., 2012). Tucholke and Mountain (1986) suggested that the Erik and Gloria Drifts south of Greenland may have begun to grow in the middle Eocene, based on interpreted ages of deep reflections in the drifts. The presence of onlapping reflectors and depositional structures on Blake Nose along the mid-Atlantic margin suggest erosion by an intermediate water mass centered above 2000 meters below sea level (mbsl) in the early Eocene (Katz et al., 1999; Norris et al., 2001a, 2001b). The seismic record from Blake Nose also shows evidence of condensed sections and slumps on the tip of Blake Nose (~2600 mbsl) that could indicate deeper erosional flow along the Blake Escarpment in the early Eocene (Norris, Kroon, Klaus, et al., 1998). Erosion on Blake Nose may have been caused by shallow parts of the Deep Western Boundary Current rather than by the northward-flowing Gulf Stream. The area of erosion is >100 km east of the main flow of the Gulf Stream, which is constrained by the location of the Florida Straight and Suwanee Channel. In the deep western North Atlantic basin, Mountain and Miller (1992) presented seismic evidence for late Paleocene bottom currents over the Bermuda Rise that could have a source analogous to Antarctic Bottom Water. Although limited, all of these data suggest that both a southern-source water mass and a northern-source water mass may have been present in the deep North Atlantic and circulated strongly enough to control seafloor deposition and erosion during the relatively warm climates of the early Paleogene. This conclusion raises the possibility that our drilling operations can expand upon some of the extraordinary paleoclimate results yielded during IODP coring in the Arctic (e.g., Brinkhuis et al., 2006; Moran et al., 2006; Pagani et al., 2006b; Sluijs et al., 2006).

The Expedition 342 drilling transect lies directly under the flow path of the present Atlantic Deep Western Boundary Current and is therefore ideally placed to monitor the chemistry and temperature of water exiting the Nordic basins in the Paleogene (Fig. F4). Therefore, the J-Anomaly Ridge and Southeast Newfoundland Ridge sites are perfectly placed to act as end-members for studying the contribution of northern-source water to the rest of the global ocean. We already have a South Atlantic/Southern Ocean end-member site (Weddell Sea, ODP Site 690) and Pacific end-member sites (Shatsky Rise, ODP Leg 198), but we have nothing equivalent in the North Atlantic except for the subtropical Blake Nose (ODP Leg 171B) sites, many of which are located in water depths that are too shallow to truly record Deep Western Boundary Current properties.

Why do we need end-members? If all we had were equatorial sites (far from the high-latitude sites of deepwater formation), we could still determine whether Atlantic water was being exported into the Pacific and vice versa, but we would not know where this water originated. Barring the unexpected revival of models of warm, saline deep water (Brass et al., 1982; Friedrich et al., 2008), it seems most likely that deep water dominantly forms at high latitudes (Bice and Marotzke, 2002; Bice et al., 1997), as is the case today and as is predicted in ocean circulation models for the Paleogene. Hence, in order to offer any concrete observational data to test ocean modeling for past warm climates, we must know where deep water is sourced and how deepwater production varies during different climate states. Drilling Newfoundland Ridge sediment drifts offers a first-rate opportunity to obtain a true North Atlantic end-member for these studies. The large depth range covered by the depth transect will also allow studies of the full range of ocean properties from the abyss to intermediate waters for the first time anywhere in the Paleogene.

In order to capture rapid shifts in circulation patterns, it is critical to obtain records that allow us to reconstruct climate at suborbital temporal resolution. The Newfoundland sediment drifts offer an unprecedented opportunity to achieve this goal from sites close to the source of northern overflow water in the Paleogene.

These drifts also allow us to resolve the changes in circulation that accompany various Paleogene extreme climate events. A case in point is the inferred reversal in deepwater overturning during the PETM (Nuñes and Norris, 2006; Tripati and Elderfield, 2005). A compilation of δ13C data from benthic foraminifers suggests that the “aging gradient” in δ13C between Southern Ocean and northern hemisphere sites abruptly reverses during this interval of global warming (Fig. F5) (Nuñes and Norris, 2006). Southern Ocean overturning in the late Paleocene is interpreted to give way to northern overturning during the PETM and then gradually revert to Southern Ocean deepwater formation over the next ~100–150 k.y. An alternative explanation is that the reconstructed gradient reversal is a stratigraphic artifact created by the effects of dissolution on sediment accumulation in northern Atlantic sites (Zeebe and Zachos, 2007). A reversal in deep ocean circulation is seen in global climate model experiments of the PETM (e.g., Bice and Marotzke, 2001) and may have occurred during other Paleogene climate events.

Drilling on the Newfoundland ridges can be used to look for such reversals in deepwater formation during a host of extreme climate events using three techniques. First, we can use depth transects to produce highly resolved records of the vertical structure of deep water at critical sites such as the Newfoundland ridges. Second, we can examine geographic gradients in benthic foraminifer geochemistry (e.g., δ13C) or neodynium isotopes on fish teeth using the Newfoundland sites as a northern end-member for comparison with sites elsewhere in the world. Third, we can use the history of drift formation itself to assess the strength of the Deep Western Boundary Current and estimate the rate of formation of deep water in the North Atlantic.

Depth transects have been shown to produce highly resolved records of changes in the source of deep water in the Pleistocene (e.g., ODP Leg 154; Curry, Shackleton, Richter, et al., 1995) and the Paleogene (ODP Legs 198, 207, and 208). For example, the hypothesized change in deepwater circulation during the PETM (Fig. F5) may have similarities to changes in deepwater formation during Pleistocene interstadials, in which outflow from the Nordic seas shift from deepwater formation to intermediate water formation as North Atlantic overturning wanes. In the Pleistocene, reduced North Atlantic overturning is associated with the intrusion of Antarctic Bottom Water well into the northern North Atlantic beneath Glacial North Atlantic Intermediate Water (Oppo and Lehman, 1993). We can look for similar changes in deepwater source areas by geochemical analysis of benthic foraminiferal calcite along depth transects. Appropriately well-preserved material can also be used to generate proxy records for the total CO32– content of the deep and intermediate waters using planktonic foraminifer weight and percent fragmentation.

Expedition 342’s deepest sites lie within the carbonate lysocline in the Paleogene and may have very little carbonate given the condensed nature of the transparent (Eocene) horizon on J-Anomaly Ridge (Fig. F1). Foraminifer preservation will clearly suffer in sediments that have been subjected to extensive calcite dissolution and may well preclude foraminifer-based geochemical studies, particularly in planktonic foraminifers. However, if the sites can be correlated (by physical property records) to well-dated upslope sites (as was done during Leg 154), we can use the deepwater site to pin the depth of the CCD and therefore evaluate the full magnitude of CCD variations throughout the Paleogene. The deepwater sites should also preserve a record of more refractory substrates, such as fish teeth, that can be used in Nd isotope studies of deep ocean flow paths (e.g., Via and Thomas, 2006; Scher et al., 2011; Thomas et al., 2003).

An additional approach for identifying patterns of deepwater formation relies on dating drift sedimentary packages and analyzing their thicknesses and distributions in seismic profiles. The dates we currently have for the sediment drifts on the Newfoundland ridges strongly suggest that major drift formation began with the deposition of thick acoustically transparent sequences and then changed character in the early Oligocene to a more mudwave-dominated system (Fig. F6). Verification of Eocene drift formation would require a reassessment of our inferences of the relationship between the rates of deepwater formation and perhaps polar glaciation because it has been generally held that drift formation began at the major onset of Antarctic glaciation in the earliest Oligocene (Davies et al., 2001) (although recent results from the Greenland-Scotland Ridge indicate drift formation began at the early to middle Eocene transition [Hohbein et al., 2012]).

Studies of deposition rates, sediment sources, and grain size would help us evaluate the origins and current regimes under which these drifts formed, as well as provide precise dates for changes in drift formation that could be compared with geochemical evidence for changes in deepwater sources. To achieve these primary goals we propose to combine proxies from sedimentology, clay mineralogy, and organic/isotope geochemistry to be applied specifically to the current-sensitive silt fraction. Continuous granulometric records will provide direct information on the variability of current strength and thus the history of the Deep Western Boundary Current. Clay mineral assemblages are excellent tracers for provenance studies, and organic matter composition will support the identification of source areas of fine-grained material and the presence of allochthonous organic matter from higher northern latitudes. As a working hypothesis we consider that intervals of enhanced strength of the Deep Western Boundary Current should have resulted in stronger advection of a higher latitude component and generally coarser grain sizes in the silt fraction, whereas intervals of reduced current activity should primarily record more local sediment sources and surface-ocean signatures as well as finer grained sediments.

High-resolution records of the onset and development of Cenozoic glaciation

The canonical view of the onset of Cenozoic glaciation is that it took place in two main steps: (east) Antarctic ice sheets were established around the time of the Eocene/Oligocene boundary (~34 Ma), whereas northern hemisphere glaciation was not triggered until ~3–7 Ma (e.g., Miller et al., 1987; Zachos et al., 2001). However, on the basis of sediment recovered in the last phase of ODP and the initial phase of IODP, it has been suggested that this view of Earth’s climate history may need revision. Dropstones have been reported from the Arctic (IODP Expedition 303) in sediments of ~45 Ma age (Backmann et al., 2005). Discontinuous δ18O records in bulk and benthic foraminiferal calcite from the equatorial Pacific (Leg 199) were interpreted in terms of extensive ice sheet development in both hemispheres, together with a huge (>150 m) eustatic sea level fall around 42 Ma (Tripati et al., 2005, 2008). This interpretation is controversial (Lear et al., 2004; Edgar et al., 2007; DeConto et al., 2008). Records from the Norwegian-Greenland Sea and the Arctic Lomonosov Ridge suggest that winter sea-ice formation was initiated around the start of the middle Eocene and that isolated alpine outlet glaciers existed on Greenland by the late Eocene (Eldrett et al., 2007; Stickley et al., 2009). In contrast with early interpretations (Lear et al., 2000), it is now also clear that the magnitude of the δ18O increase across the EOT (~33.5 Ma) is too large to reflect ice growth on Antarctica alone (Coxall et al., 2005), and there is growing evidence for contemporaneous global cooling (Lear et al., 2004, 2008; Liu et al., 2009; Eldrett et al., 2009).

Expedition 342 provided an opportunity to shed new light on these aspects of Paleogene climate in the high northern latitudes in unprecedented stratigraphic detail. The middle to upper Eocene and lower Oligocene sections targeted here will allow us to generate the high-resolution records of changes in sedimentation rate, clay mineral assemblages, and occurrence and provenance of ice-rafted debris (IRD) that are needed to achieve these goals. A potential problem exists with differentiating between putative Paleogene IRD and sediment eroded from the continental margin. In Pleistocene strata, IRD is often identified because of its coarse grain size (sand sized and larger) and its distinctive provenance (e.g., “red-coated grains” from Labrador and basalt from Iceland) (Bond and Lotti, 1995; Hemming et al., 1998). IRD in the silt fraction has been identified by a number of techniques, including modeling of end-member components using a wide spectrum of grain-size analyses (Weltje, 1997; Weltje and Prins, 2003; Prins et al., 2002). IRD might also be identified by combining grain size (especially sand content) and mineralogical information during “cold” events inferred from light stable isotope studies and geochemical assessment of provenance.

Clay minerals are excellent tracers of provenance in the modern North Atlantic and should have broadly similar distributions in the Paleogene. Continental-sourced clay should dominate the west Greenland and Canadian margins, whereas volcanic-sourced clay should be more typical of the Paleogene flood basalt province in eastern Greenland. In the modern North Atlantic, smectite, illite, chlorite, and kaolinite constitute the major proportion of the clay fraction. Smectite has been regarded as a major tracer for Iceland-Faeroe–derived material (Fagel et al., 1996) and could be expected from the Greenland volcanic province in the Paleogene. Where smectites are lacking, chlorite is the typical high-latitude clay from old metamorphic sources. A lack of volcanogenic and weathered basalt–derived clay (nontronite and amorphous minerals) suggests a rather restricted contribution from eastern Greenland and mid-ocean-ridge sources. Clay mineral type (e.g., montmorillonite versus beidellite) and percentage data and Sm/Nd ratios in the clay fraction of Reykjanes Ridge sediment indicates a dominant terrigenous contribution from young continental crust that may derive from Europe and/or the Arctic (Fagel et al., 2001; Innocent et al., 2000). Today, the clay fraction may also contain terrigenous calcite and dolomite derived from glacier milk in the Hudson Bay area. Such a signal could be transported for much longer distances than sand-sized IRD.

Astronomical calibration of chronostratigraphic markers and the geological timescale

A detailed geological timescale that can be applied widely was a primary goal of Expedition 342 drilling. Indeed, a highly resolved timescale is becoming increasingly critical in studies of Paleogene paleoceanography as we begin to extend work on millennial timescales into deep time and require increasingly high resolution correlation to evaluate the evolution of interbasin geochemical gradients. Determination of the rates at which Earth processes take place and how these rates change are key not only to developing an understanding of Earth history but also to accurately describing the nature and rate of the processes themselves.

The discovery of orbitally driven variations in Earth’s climate and their preservation in the marine sedimentary record has been a major advance in chronostratigraphy. The application of this technique is not without assumptions, including the predictability of the exact beat of the climatic metronome (Pälike and Shackleton, 2000; Pälike et al., 2004). However, if constrained by a paleomagnetic reversal stratigraphy, and if the same ages are obtained for geologic boundaries in different regions with different sensitivities to each of the different orbital forcings, we can develop a strong confidence in the timescale. A complete orbitally tuned Cenozoic timescale can provide estimates of process rates that have comparable precision throughout the Cenozoic.

We are working toward an astronomically tuned Cenozoic timescale. There is a semianchored astronomical timescale back to 30 Ma (mostly from Leg 154). No magnetostratigraphy was recovered during Leg 154 (Curry, Shackleton, Richter, et al., 1995), but investigations from ODP Legs 177 (Gersonde, Hodell, Blum, et al., 1999) and 199 have resulted in verification and refinement of results from Leg 154 (e.g., Billups et al., 2004; Wade and Pälike, 2004), as well as an extension of the tuning across the EOT (Coxall et al., 2005). Several other ODP legs and IODP expeditions have provided suitable data for work of this sort, most recently from the Pacific Equatorial Age Transect (Expedition 320/321; Pälike et al., 2010). Additional records are available for the time intervals of 35–42 Ma (Legs 171B and 177), ~45–56 Ma (Leg 207), 53–57 Ma (Legs 171B, 198, and 208), and 62–65 Ma (Legs 165, 171B, 198, and 208) (Pälike et al., 2001; Westerhold and Röhl, 2009; Norris and Röhl, 1999; Lourens et al., 2005; Sigurdsson, Leckie, Acton, et al., 1997). From this perspective, acquisition of an astronomically tuned record of the late Eocene and the early middle Eocene would be extremely important to span existing gaps in our tuning efforts and to provide single-site, high-resolution data that allow splicing, verification, and extension of previous efforts.

For the Paleocene and Eocene, a shallow CCD has hampered efforts to obtain long, uninterrupted carbonate sequences that allow high-resolution paleoclimatic studies with traditional geochemical studies (e.g., stable isotope analysis). The tectonic setting of the sites targeted by Expedition 342 is likely to overcome this problem because the sites track above the CCD for the critical time slices needed, with potentially very expanded Paleocene, Eocene, and Oligocene sedimentary deposits. In addition, for the time intervals that have already been astronomically tuned, a comparison between low-latitude sites from the Pacific and Atlantic will be complemented by the North Atlantic setting of Expedition 342, offering the chance to decipher the processes controlling the amplification of, for example, ~41 k.y. obliquity versus ~100 and 405 k.y. eccentricity cycles and to test the hypothesis that there are different dominant astronomical forcing mechanisms between Earth’s warm and cool periods.

Drilling strategy

A primary challenge in recovering highly detailed records of past greenhouse conditions is the typically low sedimentation rates found throughout much of the deep ocean. Most ocean drill cores that recover Paleogene and Cretaceous records, for example, have sedimentation rates of 1 cm/k.y. or less—rates that are too low to resolve events with durations <1000 y in the best of cases. Indeed, burrowing and other mixing processes distort these slowly accumulating records still further, distributing even instantaneous events such as meteorite impact ejecta or volcanic ash beds over tens of centimeters of the sediment record (Hull et al., 2011). Mixing therefore both reduces the apparent magnitude of past abrupt events or transitions and makes these events appear to occur much more slowly than was actually the case. It is possible, of course, to unmix the record, either by modeling the mixing process or by inverse modeling (Heinze et al., 1999; Panchuk et al., 2008; Ridgwell, 2007), but these methods will always produce a hypothesis about rates or magnitudes of events that is difficult to test empirically.

Another challenge in Paleogene paleoceanography is that old sediment is typically buried beneath younger sediment, with consequent impacts for the preservation of the substrates used in proxy-based paleoenvironmental reconstruction and our ability to obtain full records of sedimentary sequences. Low-temperature diagenesis degrades the quality of most microfossil groups as well as chemical tracers of ocean conditions. Burial diagenesis also impedes coring, decreasing the probability that sequences can be recovered in continuous sections and producing gaps in the resulting record. Because hard sediments cannot be obtained by piston coring, we also lose the ability to obtain oriented cores for paleomagnetic work. Finally, burial beneath a thick sedimentary pile increases drilling time and reduces the time available to obtain extensive records of deeply buried, older sediments.

Our strategy, therefore, was to focus on drift sediments with high deposition rates as opposed to open-ocean pelagic records. We targeted drifts that are not buried significantly and are likely to preserve sediments that have not undergone major burial diagenesis. The lack of burial improves the chance that the upper sediment column is available through piston coring and the acquisition of oriented cores. Our approach also samples drift packages that occur over a wide variety of water depths with the hope of obtaining cores over the widest possible range of water depths.

Expedition 342 is focused on the Paleogene record on the Newfoundland ridges. Although there is an extensive Cretaceous record of both drifts and fossil reefs in the seismic record, we did not have time to do justice to Cretaceous objectives without sacrificing our studies of the Paleogene system. Furthermore, our particular area of focus was the middle Eocene to Oligocene interval where thick sediment drift deposits preserve unusually expanded records of the transition from the greenhouse world of the Eocene climatic optimum to the glaciated world of the Oligocene.

Drilling during Expedition 342 completed the North Atlantic objectives laid out in a 1997 Marine Aspects of Earth System History workshop on warm-period dynamics and the ODP Extreme Climates program planning group (PPG) (see www.odplegacy.org/program_admin/sas/ppg.html). Expedition 342 had its genesis in the Extreme Climates PPG, which backed proposals that later became Legs 198, 199, 207, and 208. It is somewhat fitting that Expedition 342 is the last of this powerful lineup of drilling programs because it is the only one to focus on drilling high deposition–rate sequences rather than long but low temporal resolution records. Drilling also addressed initiatives of the IODP Initial Science Plan in the areas of extreme climates and rapid climate change. Finally, our expedition takes up proposals of the recent National Research Council report Understanding Earth’s Deep Past: Lessons for Our Climate Future (National Research Council, 2011), which advocates focused efforts to resolve the timescale and use mechanisms of past hyperthermal events as possible analogs for future global change.

The Newfoundland ridges

On 15 April 1912, the RMS Titanic, en route westward from Southampton, England, to New York City, USA, hit an iceberg off the Grand Banks of Newfoundland and sank, killing more than 1500 people. The two halves of the wreck lie between the volcanic seamounts of the Southeast Newfoundland Ridge because there the southward-flowing surface water of the cold Labrador Sea carries icebergs to their intersection with the warm tongue of the Gulf Stream. Today the Titanic is bathed by the Deep Western Boundary Current because this new abyssal water passes at depth under the Gulf Stream on its circuit throughout the deep basins of the world oceans (Fig. F4). The shape of the North Atlantic margin suggests that a similar current configuration occurred in the past, with any deep water formed in the North Atlantic constrained to flow over the Newfoundland ridges. Therefore, Expedition 342 drill sites are particularly useful to monitor the overturning history of the North Atlantic Ocean.

The Newfoundland ridges are mantled with some of the oldest sediment drifts known in the deep sea and range in age from the Late Cretaceous to Paleogene. Pliocene–Pleistocene drifts in the northeastern Atlantic commonly have sedimentation rates of 4–20 cm/k.y. and therefore can be used to study rates of abrupt climate change (Channell et al., 2010). Previous drilling of drifts on Blake Nose (off the southeastern United States) revealed sedimentation rates in the middle Eocene of ~5–6 cm/k.y., far higher than the ~1 cm/k.y. rates typical of previous Paleogene-focused drilling targets (Norris et al., 2001b). The Newfoundland sediment drifts also have high accumulation rates, which will allow us to obtain records of warm-period climates and evolution with unusual fidelity, and these will be particularly useful for assessing rates of change in the Earth system during both transient episodes of extreme warming (analogous to the near future) and transitions from warm climates into the glaciated world.

Structure and stratigraphy of J-Anomaly Ridge and Southeast Newfoundland Ridge

J-Anomaly Ridge and Southeast Newfoundland Ridge (Fig. F7) formed along the axis of the mid-Cretaceous Mid-Atlantic Ridge in a fashion analogous to the modern Reykjanes Ridge and Iceland (Tucholke and Ludwig, 1982). The tops of both ridges were above sea level in the Aptian and subsequently subsided to abyssal depths by the Late Cretaceous (Tucholke and Vogt, 1979). Aptian rudist platform carbonates were drilled at Deep Sea Drilling Project (DSDP) Site 384 (now at 3900 mbsl) (Tucholke, Vogt, et al., 1979), and buried reef-like seismic features are present on the flank of J-Anomaly Ridge and Southeast Newfoundland Ridge. Pelagic carbonates began to blanket the tops of the ridges by ~75–80 Ma (Tucholke, Vogt, et al., 1979) and accumulated to thicknesses of nearly 1.5 km by the late Neogene.

Five principal sedimentary sequences are evident on the Newfoundland ridges, bounded by reflection Horizons A–D (Fig. F6). The uppermost sequence displays well-defined internal reflections and mantles the northern side of Southeast Newfoundland Ridge. Nearly all piston cores of this sequence collected from the Newfoundland ridges (Fig. F8) are of Pleistocene age, including “long piston” Core MD95-2027. The presence of thick Pleistocene sections on the north side of the Newfoundland ridges may reflect the accumulation of iceberg-transported sediment derived from the Hudson Bay and Greenland. To the south, on Southeast Newfoundland Ridge and J-Anomaly Ridge, the Pleistocene cover is nearly absent and is represented by only a thin (3–5 m thick) layer of Pleistocene foraminifer sand mixed with glacially transported sediment based on gravity cores collected during the Knorr 179-1 site survey. The thin Pleistocene sediment that covers our target sediment likely reflects the barrier imposed by the warm Gulf Stream to icebergs drifting southward along the Newfoundland margin. Pleistocene glacial sand protects the Paleogene sediment drifts from extensive erosion and preserves them in unconsolidated condition, as seen in piston Cores KNR179-1-13PC and KNR179-1-15PC from Southeast Newfoundland Ridge.

The second sequence, bounded by reflection Horizons A and B, displays poorly defined, discontinuous reflections (Fig. F6) and is probably of Oligocene and Neogene age on the basis of its similar acoustic character to other drifts in the North Atlantic (Davies et al., 2001). In some areas, the discontinuous reflections can be resolved as fields of sediment waves, suggesting that much of the unit was deposited under strong directional bottom currents. Large parts of the southern flank of Southeast Newfoundland Ridge and the northern end of J-Anomaly Ridge are covered by this sequence, with thicknesses of >700 m. We have no cores that firmly date this sequence because of failure of the hydraulic winch during the site survey cruise.

The third sequence, bounded by reflection Horizons B and C, is seismically transparent and has a poorly defined contact with the overlying sequence of discontinuous reflections (Fig. F6). The absence of a strong reflector between these seismic units suggests that seismic Unit 3 has a conformable relationship with overlying Unit 2. Piston cores and seismic ties to Site 384 show that this sequence is of early Eocene age (nannofossil Zones NP14 and NP15) and younger. Its great thickness (as thick as 800 m; Figs. F9, F10) suggests an unusually expanded sequence of lower and middle Eocene sediment on J-Anomaly Ridge. Furthermore, it is possible that there is a complete Eocene–Oligocene sequence in the drilling transect on Southeast Newfoundland Ridge (Fig. F10). Piston core samples show the main lithology in the lower to middle Eocene section is a clay-rich white to yellow nannofossil ooze. The absence of strong internal reflections suggests that the sequence is not punctuated by major hiatuses but was deposited steadily like many modern Pliocene–Pleistocene drifts in the North Atlantic. This sequence thins below ~4.5 km present water depth (~4 km in the Eocene), apparently reflecting reduced sedimentation rates in the lysocline and below the CCD. Thinning of the Eocene package at ~4 km paleodepth is broadly consistent with the position of the Eocene CCD estimated from prior North Atlantic drilling (Fig. F2) (Tucholke and Vogt, 1979).

The fourth sequence, bounded by reflection Horizons C and D (Fig. F6), is of Cretaceous–Paleogene age, is >500 m thick, and crops out on the base of J-Anomaly Ridge (e.g., Site U1403; Fig. F1) and in numerous places on the crest of Southeast Newfoundland Ridge, including outcrops in moats around several seamounts. These strata also display drift-like morphology, albeit of smaller size than the Eocene drifts, and are characterized by mostly well defined parallel reflections like the Pleistocene cover section. This sequence was drilled at Site 384 (Figs. F6, F10), which recovered Campanian to lowermost Eocene beige calcareous ooze and soft chalk with excellent magnetic stratigraphy and well-preserved foraminifers and calcareous nannofossils, as well as radiolarians in the upper Paleocene (Berggren et al., 2000). The unconformities at the K/Pg boundary and the Paleocene/Eocene boundary at Site 384 are expected because the site is located in a condensed section on top of a ridge. All recent Paleogene–Cretaceous drilling expeditions (Legs 171B, 198, 199, 207, and 208) (Norris, Kroon, Klaus, et al., 1998; Bralower, Premoli Silva, Malone, et al., 2002; Lyle, Wilson, Janecek, et al., 2002; Erbacher, Mosher, Malone, et al., 2004; Zachos, Kroon, Blum, et al., 2004) recovered one or more of these boundary sections despite their absence in older DSDP holes upon which the new drilling legs were based.

The fifth sequence, underlying reflection Horizon D, displays dense but parallel reflections (Fig. F6) and crops out on the northwest slope of J-Anomaly Ridge, the crest and flanks of Southeast Newfoundland Ridge, and apparently in the pelagic caps of several seamounts. This sequence consists of several discrete seismic sequences separated by possible unconformities, indicated by truncations of reflectors. This entire seismic package is likely of mid-Cretaceous to early Late Cretaceous age on the basis of seismic ties to Site 384. Some of these sequences lap up against seismically identified buried reefs and are as thick as 450 m. The reefs and surrounding sediment are probably Barremian–Albian in age on the basis of results from Site 384 (Tucholke and Vogt, 1979) and Sr isotope stratigraphy (P. Wilson, unpubl. data), in keeping with the estimated ages of buried reefs off Florida (Hutchinson et al., 1995; Norris et al., 2001a). Representative summaries of our interpretations of the sequence of sedimentary packages on J-Anomaly and Southeast Newfoundland Ridges are shown in Figure F10.

The modern Deep Western Boundary Current

The area east of the Grand Banks is a region critical to understanding the history of deepwater circulation in the North Atlantic because it is the gateway between bottom water sources in the Norwegian-Greenland and Labrador Seas to the north and the main basins of the North Atlantic to the south. Denmark Straight Overflow Water is the main deepwater mass, centered at ~3500 mbsl and overlain by Labrador Sea Water at ~1500 mbsl (Pickart et al., 1999). Southeast Newfoundland Ridge is a major barrier to deep southward flow and diverts the Deep Western Boundary Current offshore into the path of the northeasterly flow of the Gulf Stream. The deepest part of the bottom current follows submarine contours around the southeastern end of the ridge and continues west around J-Anomaly Ridge and along the Nova Scotian continental rise (Fig. F4). Shallower portions of the current follow contours around the crest of Southeast Newfoundland Ridge and also interact with seamounts on the ridge, forming local moats and drifts (Fig. F9). A particularly good example of these current-formed drifts are the sinuous deposits of calcareous clay and nannofossil ooze that form drifts over J-Anomaly Ridge (Fig. F11).

The Gulf Stream actually reaches the seafloor over Southeast Newfoundland Ridge and may contribute to bottom scouring. East of the ridge, Meinen and Watts (2000) found that the mean North Atlantic Current clearly extends to the bottom. The measured bottom currents are strong enough to suspend sediments but probably not strong enough to cause extensive erosion. Still, we must bear in mind the possibility that erosion on J-Anomaly and Southeast Newfoundland Ridges is related to a southward-flowing deep boundary current and/or a northward-flowing surface current that regionally extends to the bottom.

Newfoundland sediment drifts

One of the main advantages of drilling the Newfoundland sediment drift complex is the near-absence of Neogene sedimentary cover. Most areas were swept by sufficiently strong currents during the later Cenozoic to prevent extensive deposition of younger strata on the southern side of the ridges or in patches around the seamounts. Although we do not have firm dates on when these strong currents were initiated, they are probably a post-Oligocene feature (on the basis of sedimentation rate changes at ODP Site 1276 in the Newfoundland Basin) and may reflect the full development of North Atlantic Deep Water, possibly in combination with a strengthened Gulf Stream. Before this time, the remarkable thickness, absence of internal reflections, and drift morphology suggest that the Paleogene section is likely to be hugely expanded, with sedimentation rates much higher than the 0.5–1 cm/k.y. typically encountered in the deep sea. A similar transition from drift deposition to nondeposition in the latest Eocene was observed on Blake Plateau off the Florida-South Carolina margin during Leg 171B (Norris, Kroon, Klaus, et al., 1998). On Blake Nose, the Eocene and older sections are unusually expanded, with deposition rates as high as 6 cm/k.y. throughout the later middle and late Eocene (Norris et al., 2001a, 2001b). We viewed the Leg 171B results as a favorable prognosis for the outcome of drilling the Newfoundland ridges.

The primary drilling targets for Expedition 342 are in plastered drifts that exist in two places: (1) the southern toe and eastern flank of J-Anomaly Ridge (Fig. F11) and (2) the north-facing slopes of seamounts on Southeast Newfoundland Ridge (Fig. F12). In total, we targeted four different drifts during Expedition 342 to compare the sedimentation histories of drifts formed under different hydrodynamic conditions.

Drift sedimentation clearly has a complex history on the Newfoundland ridges, with an initial phase of drift formation on the eastern flank of J-Anomaly Ridge in the Late Cretaceous (probably starting in the Campanian or early Maastrichtian, judging from Site 384 drilling results) (Berggren et al., 2000; Tucholke and Ludwig, 1982) and continuing through most or all of the Paleocene. Site 384 was spot cored in a highly condensed section, and the Paleocene/Eocene boundary was not recovered. However, acoustic character does change at some point in the upper Paleocene or lower to middle Eocene with the deposition of an acoustically transparent layer (seismic Unit 3; Fig. F6). This transparent seismic unit is our primary drilling target because the absence of internal reflectors suggests that it is a conformable sequence with a good likelihood of being correlative across the depth transect.

Drift morphology suggests that the primary drift deposits formed mostly under a southward-flowing bottom current in the Eocene. This current formed plastered drifts on the northeast and southwest faces of seamounts on Southeast Newfoundland Ridge, as well as a long episodically growing and accreting ridge system between J-Anomaly Ridge and Southeast Newfoundland Ridge (Fig. F11). The southeastern flank of Southeast Newfoundland Ridge has a very thick seismically transparent drift (part of seismic Unit 3) that is overlain by a younger drift deposit that displays complex internal reflectors consistent with sediment waves (seismic Unit 2). We provisionally assigned the transition between these drift packages to the EOT.

The considerable thickness of the middle Eocene to ?upper Eocene seismically transparent interval (seismic Unit 3) requires that traditional methods of coring Paleogene targets will have to give way to more focused drilling objectives. Most previous Paleogene and Cretaceous drilling was designed to obtain records spanning tens of millions of years. In contrast, previous drilling legs that targeted Pleistocene drifts typically cored the upper parts of the sediment packages in order to obtain highly expanded records of the late Pleistocene. Our goal was to follow the “Pleistocene strategy,” in which we cored expanded drift records at the expense of long time series.