The continuous Paleogene sedimentary sections from Leg 199 provide new opportunities to document and understand the Cenozoic evolution of the equatorial Pacific. These studies include understanding how productivity has changed, the timing of glaciations, and the involvement of long period-changes in orbitally driven insolation, timing and extent of rapid changes in climate state, and involvement of the carbon cycle in climate change.
Moore et al. (2004) used the newly updated ages of radiolarian zone boundaries to explore changes in equatorial Pacific productivity from the Paleocene to the late Miocene. Furthermore, Parés and Moore (2005) tracked northward Pacific plate movement by assuming that the linear sedimentation maximum found in each time slice from Moore et al. (2004) marks the position of the paleo-equator.
The modern Pacific equatorial zone has significantly higher productivity and higher biogenic sedimentation than other pelagic biomes (Murray et al., 1993; Honjo et al., 1995) (Fig. F6). Biogenic mass accumulation rates (MAR) between 0° and 1° north or south of the equator are ~50% higher than biogenic MAR between 1° and 2° north or south. Because the productivity is caused by Ekman divergence, a physical process that is driven by the change in sign of coriolis force at the equator, there should always be upwelling at the equator as long as the trade winds blow from east to west (Wyrtki, 1981; Chavez et al., 1996) (Fig. F7). The nutrients added by upwelling deeper waters into the euphotic zone fertilize the equatorial region.
More than 90% of the sediments in the equatorial Pacific are biogenic in origin, so sedimentation rates are a semiquantitative proxy of paleoproductivity in the region. Moore et al. (2004) used these ideas to study the Cenozoic evolution of Pacific equatorial zone. The most striking discovery is of a secondary productivity zone in the northern tropical Pacific in the Eocene and early Oligocene. The secondary zone is probably associated with divergence at the boundary of the Eocene North Equatorial Current with the North Equatorial Countercurrent (Fig. F8) (Moore et al., 2004). Moore et al. (2004) also note that deposition of biogenic sediments in the equatorial sediment tongue varied by a factor of 2 to 5 in different time slices representing 5- to 6-m.y. averages. Such large variability, even when averaged over such long periods, makes it clear that large-scale variations of equatorial Pacific winds and upwelling have occurred as part of the evolution of the Cenozoic.
Moore et al. (2004) noted that the paleoposition of the equatorial sediment tongue did not always backtrack to an equatorial position. Parés and Moore (2005) used the misalignment of the equator backtracked via a stationary hotspot model of Pacific plate motion to explore the motion of the Hawaiian hotspot in the Cenozoic. They found a heartening agreement between "sedimentary" Pacific plate polar wander and Cenozoic poles estimated by other means (Tarduno et al., 2003) Through this method Parés and Moore (2005) were able to discern new detail of Cenozoic Pacific plate motion. The results support other data that indicate southward motion of the Hawaiian hotspot since the Cretaceous at a rate of ~32 km/m.y. and the general northward motion of the Pacific plate since 53 Ma of 12°, or ~25 km/m.y.
Arguably, the single most important paleoclimate rationale for ocean drilling is that deep-sea sediments provide the most stratigraphically complete and globally representative proxy records of global change. One of the outstanding highlights of Leg 199 was our recovery of stratigraphically complete sequences (the sections drilled are virtually free of hiatuses at the biostratigraphic zone and magnetochron level), representative of substantial tracts of the central Pacific Ocean. Undoubtedly, the most elegant demonstration of this important highlight is the detailed stratigraphic correlation of Oligocene sedimentary sequences, including the Eocene–Oligocene and Oligocene–Miocene transitions, recovered at Sites 1218 and 1219 (located nearly 750 km apart) to the decimeter level (Pälike et al., this volume).
In many ways the Oligocene (~33.7–23.03 Ma) represents something of a "neglected middle child" of Cenozoic paleoceanography—caught between the early Paleogene Greenhouse and the well-developed Neogene Icehouse. This situation is at least partly attributable to the perception that the Oligocene marks a prolonged interval of relative stasis in paleoclimate and biotic turnover, as indicated in deep-sea micropaleontological communities by conservative body plans, confusing taxonomies, and low biostratigraphic resolution. However, in some respects the Oligocene represents the most interesting piece of the Cenozoic paleoceanographic puzzle because it offers an opportunity to unravel the processes that lie behind the transition from a world free of large-scale continental icecaps and rapid eustatic sea level oscillations to one dominated by these climatic changes. Recent progress in our understanding of Oligocene paleoclimates has largely been driven by seismic stratigraphy and scientific drilling on continental margins (e.g., Browning et al., 1996; Miller et al., 1996; Pekar et al., 2002). In contrast, benthic stable isotope compilations (e.g., Miller et al., 1987; Zachos et al., 2001) show that paleoclimate records for the deep oceans through the Oligocene rely heavily on old DSDP sites, largely from the Atlantic Ocean (Miller et al., 1987, 1988, 1990, 1991, 1993), with a relatively low sample spacing. Leg 199 presented an excellent opportunity to generate low-latitude deep-sea records throughout the Oligocene in order to test models developed from continental margin sequences for the pattern and timing of changes in global temperature and continental ice volume.
Previous studies designated "Oi events" in the Oligocene to represent transient glaciations based on intervals where particularly heavy oxygen isotope ratios were observed in relatively low resolution data sets from several sites (e.g., Miller et al., 1991). High-resolution (up to 2 k.y.) Oligocene stable isotope stratigraphies now exist for bulk sediment (top Subchron C6An.2n through base Subchron C6Cn.3n; most of Chron C9n; top Chron C12n through base Chron C13n), planktonic (top Chron C9n through base Subchron C11n.1n), and benthic foraminiferal calcite (the entire Oligocene, including the E–O and O–M transitions) from Site 1218 (Wade and Pälike, 2004; Coxall et al., 2005; Tripati et al., this volume; Pälike et al., submitted [N1], this volume). These new data sets from Site 1218 demonstrate pronounced cyclicity in the orbital bands and shed new light on the Oi events. It is now clear that such events occur more regularly than previously supposed and are characterized by higher amplitude 18O change at orbital timescales as well as regular cyclic behavior in 13C records, predominantly with a 405-k.y.-long eccentricity spacing. In fact, the new data sets have made it possible to develop a new naming scheme for Oligocene glacial maxima that is based on the 405-k.y. eccentricity cycle count (Wade and Pälike, 2004), avoiding some of the confusion caused by previous conflicting naming schemes. The occurrence of these glacial maxima is striking because intervals with the highest 18O values consistently correspond to obliquity minima modulated with a period of 1.2 m.y. (and triggered to minima in the ~405 and ~100 k.y.) eccentricity cycles (Wade and Pälike, 2004; Coxall et al., 2005; Pälike et al., submitted [N1]). This important result supports the view that seasonality plays a key role in determining the development of glacial maxima. Specifically, the key factor in the development of the major Oligocene glaciations is not the occurrence of particularly cold winters or the occurrence of particularly cold summers but rather the prolonged absence of particularly warm (Southern Hemisphere) summers.
Two methods have been employed to help quantify the magnitude of Antarctic ice volume and thus sea level change during the Oligocene using Site 1218 sediments. Where both benthic and planktonic 18O records are available (top Chron C9n through base Subchron C11n.1n), sea level changes of up to 60 m have been estimated from benthic 18O data on the basis that ice volume fluctuations cannot exceed the variation recorded in paired planktonic data (Wade and Pälike, 2004). The other method is to use paired Mg/Ca and 18O data in benthic foraminifers to calculate seawater 18O (sw) (Lear et al., 2000). Application of this method to the Site 1218 section suggests that a large Antarctic ice sheet formed during Oi-1 and subsequently fluctuated throughout the Oligocene on both short (<0.5 m.y.) and long (2–3 m.y.) timescales, between ~50% and 100% of its maximum earliest Oligocene size (consistent with estimates of sea level derived from sequence stratigraphy) (Lear et al., 2004).
Site 1218 13C stratigraphies for the Oligocene also demonstrate pronounced cyclicity in the orbital bands with a particularly strong imprint of the ~405-k.y. eccentricity cycle, indicative of changes in global carbon burial. Where both benthic and planktonic records are available, we see that alterations of high-latitude temperatures and Antarctic ice volume had a significant impact on equatorial productivity (Wade and Pälike, 2004). Cross-spectral analysis of the Site 1218 Oligocene stable isotope data sets indicates a complex pattern of lead-lag relationships. Benthic 18O lags 13C in the early Oligocene, corresponding to ~8 k.y. in the 40-k.y. band. The lag suggests that the response of the global carbon cycle to Earth's obliquity helped to force changes in ice volume and high-latitude temperature (Coxall et al., 2005). For the entire Oligocene, an ~20-k.y. lag of 13C compared to 18O is found for the long eccentricity (405 k.y.) cycles suggesting that, in turn, changes in Antarctic ice volume and high-latitude temperature had a significant impact on global carbon cycling, and ocean buffering of the global carbon cycle (Pälike et al., submitted [N1]).
The Eocene–Oligocene transition represents a pivotal point in the transition from the Greenhouse world of the Cretaceous and early Paleogene into the late Paleogene–Neogene Icehouse. Yet, the rarity of complete deep-sea sections across this interval has limited our understanding of the dynamics of this important step leading to the modern Icehouse world.
The Eocene–Oligocene transition is marked by a large rapid increase in the benthic foraminiferal calcite 18O record in earliest Oligocene time (Oi-1). This excursion was first ascribed to a 5°C temperature drop, but more recently Oi-1 has been associated with the onset of continental ice accumulation on Antarctica (Shackleton and Kennett, 1975; Shackleton, 1986; Miller et al., 1987, 1991; Zachos et al., 1992). Such confusion reflects the long-standing difficulty of separating the effects of temperature and ice on benthic 18O (Lear et al., 2000) and thus the relative importance of cooling vs. higher Antarctic snowfall. A benthic Mg/Ca record for the E/O transition from DSDP Site 522 (South Atlantic) shows no significant change corresponding to "Oi-1," suggesting that all of the 18O increase associated with Oi-1 can be ascribed to ice growth with no concomitant decrease in polar temperatures (Lear et al., 2000). Support for this suggestion exists from modeling experiments and sedimentary records (Kohn et al., 2004; Grimes et al., 2005). According to these data, the trigger for continental glaciation lay in the hydrological cycle rather than the carbon cycle. Specifically, it has been proposed that the opening of the Australian-Antarctic Seaway in earliest Oligocene time might have enhanced the supply of moisture as snow to the Antarctic interior (Lear et al., 2000). On the other hand, paleoproductivity studies suggest that the near-contemporaneous increase in seawater 13C was driven by increased rates of Corg burial in marine sediments, and this factor may be implicated in global cooling and ice sheet growth (Diester-Haass and Zahn, 2001).
Even if global cooling did not trigger Cenozoic Antarctic glaciation, some of the feedback processes associated with the development of a large ice sheet might be expected to cause cooling. Either way, long-standing global lithologic compilations indicate a pronounced deepening of the CCD near the Eocene–Oligocene transition (van Andel et al., 1975; Delaney and Boyle, 1988) that has not been studied in detail.
Until recently, our most complete records of the E/O boundary came from two mid-latitude sites, both in the Southern Hemisphere (DSDP Site 522 and ODP Site 744). Leg 199 radically improved our archive of deep-sea sections across this important climate transition by recovering E/O boundary sections from five Northern Hemisphere sites (1217, 1218, 1219, 1220, and 1221). Taken together, these sites provide a valuable opportunity to study the chain of events across the E/O boundary within the framework of a depth and latitudinal transect. Throughout this transect, the transition from the Eocene to the Oligocene is instantly recognizable by a sharp upsection shift from SiO2-rich and carbonate-poor to carbonate-rich and SiO2-poor sediments. Shipboard observations indicated that the CCD deepened substantially, rapidly, and permanently across the Eocene–Oligocene transition (Fig. F9, from Shipboard Party, 2002). One objective of postcruise research was to quantify this relationship.
Coxall et al. (2005) generated high-resolution (up to 2 k.y.) records of percent (and mass accumulation rate) CaCO3 and benthic 18O and 13C for the E–O transition at Site 1218. These data show that the CCD deepening across the E–O transition occurred much faster than previously documented, occurring in two jumps of about 40-k.y. each, separated by a ~200-k.y. "plateau." Remarkably, this deepening occurred synchronously with the stepwise onset of Antarctic ice sheet growth, as indicated by the parallel increase in the 18O series (Coxall et al., 2005). Ultimately, the trigger for these events is interpreted to have been an orbital configuration favoring the prolonged absence of warm Antarctic summers (low eccentricity and low-amplitude obliquity change) (Coxall et al., 2005). Yet, some additional conditioning factor must have been at work because there is no evidence that the low-eccentricity obliquity "node" conditions at 34 Ma were any more extreme than those that occurred regularly during preceding Eocene time (Coxall et al., 2005). A likely candidate for this climate conditioning factor is long-term decline in atmospheric carbon dioxide (e.g., DeConto and Pollard, 2003). The amplitude of 18O increase in the Site 1218 data is three times larger than that modeled by DeConto and Pollard (2003) and impossibly large to be explained by ice growth on Antarctica alone (Coxall et al., 2005). This important observation raises the intriguing possibility that Northern Hemisphere glaciation was initiated much earlier than widely believed. Alternatively, the data can be explained by concomitant global cooling (by ~4°C). This alternative calls into question a diverse range of data sets, including benthic Mg/Ca records, that show little evidence for cooling across the E/O boundary. Lear et al. (2004) present a Mg/Ca record for the E–O transition at Site 1218 and compare it to the one generated at DSDP Site 522 by Lear et al. (2000). This analysis, together with work on the sensitivity of Mg partitioning between seawater and benthic foraminiferal calcite, suggests that some other factor (e.g., pH or [CO32–]) acts to partially mask the cooling signal in the Mg/Ca records for the E–O transition.
The "lock-step" behavior of the 18O and CCD records for the E–O transition at Site 1218 suggests that ocean acidity and continental glaciation are in some way intimately related and points to a teleconnection between Antarctica and the equatorial Pacific. The association between a decrease in ocean acidity and glaciation can be understood in terms of declining atmospheric carbon dioxide and low-resolution proxy records for Cenozoic pCO2 change indicate significantly lower levels of pCO2 for the Oligocene than for the Eocene (Pagani et al., 2005). However, no proxy pCO2 data are yet available for the E–O transition itself.
Modeling studies suggest that an increase in deep-sea [CO32–] associated with a 1-km deepening of the CCD today yields a drawdown in atmospheric CO2 of <25 ppmv (Sigman and Boyle, 2000; Zeebe and Westbroek, 2003). Nevertheless, the 100-ppm drawdown of atmospheric CO2 at the Last Glacial Maximum (LGM) in the Pleistocene is associated with only a 100-m change in CCD, not a 4000-m change (Farrell and Prell, 1989), suggesting strong negative feedback and much less sensitive CCD response to atmospheric CO2 change than simple models predict.
Antarctic glaciation may have triggered CCD deepening (Coxall et al., 2005) by a change in global ratio of CaCO3 to Corg burial. In fact, the pronounced 13C increase seen in the E–O transition records suggest that the CCD deepened to compensate for increased Corg burial ratio relative to CaCO3 (Coxall et al., 2005). However, the mechanism responsible for bringing about this change is a matter of debate. Coxall et al. (2005) reason that a simple mechanism to explain these events is shelf-to-basin CaCO3 fractionation caused by glacioeustatic sea level fall associated with Antarctic glaciation. Rea and Lyle (2005) argue that the sea level fall associated with the glaciation (50 m, based on New Jersey margin records, e.g., Miller et al., 2005) is too small to have been the sole factor and invoke an increase in the delivery to the ocean of weathered Ca and HCO3– ions. Olivarez Lyle and Lyle (this volume, 2006) suggest that a temperature drop associated with the E/O may have increased Corg burial significantly by reducing the metabolic decomposition of organic matter before burial. Such a reduction in the dissolved inorganic carbon pool relative to weathered cations would also deepen the CCD. These (and other) competing hypotheses are now being tested using a range of modeling techniques.
A chert-rich interval at roughly the early/middle Eocene boundary at all sites except Site 1215 (see Fig. F2) operationally divides Leg 199 Eocene studies into those of the Paleocene–early Eocene and those from the middle Eocene to the Eocene/Oligocene boundary. At Site 1215, the northernmost site of Leg 199, sedimentation switched from carbonates to red clay at ~52 Ma, in the early Eocene (Raffi et al., 2005), so studies of this site also fall within the Paleocene–early Eocene category. We discuss all the Paleocene–early Eocene work in the section on the Paleocene/Eocene boundary interval.
The Eocene is relatively poorly studied but Earth's climate is thought to have evolved from a transient warm burst (the PETM, at 55 Ma) to maximum Cenozoic warmth over a period of ~5 m.y. From that point onward, Earth was on a cooling path, punctuated by significant rapid changes, such as that of the Eocene/Oligocene boundary (see Zachos et al., 2001, for a good overview). The paucity of high-resolution records in the Eocene has caused the impression that the Eocene was marked by only gradual change. More detailed records reveal significant numbers of transient warm and cold events lasting from a few hundred thousand years up to 2 m.y. (e.g., Bohaty and Zachos, 2003; Lourens et al., 2005; Lyle et al., this volume).
The early Eocene began with one of the most important Cenozoic transient events, the Paleocene/Eocene boundary event, or PETM (Zachos et al., 2001). The upper? Paleocene and lower Eocene sediments recovered during Leg 199 are marked by a very distinctive P/E boundary section (see Lyle, Wilson, Janecek, et al., 2002; Nomura et al., 2002; Faul and Paytan, this volume; Knoop, this volume; Murphy et al., this volume; Nomura and Takata, this volume; Nuñes and Norris, this volume), with significant geochemical variations as well as the disappearance of carbonate that marks the PETM worldwide. The Paleocene and lower Eocene sediments are also distinguished by low numbers of radiolarians, unlike later Eocene intervals.
Within the Leg 199 transect, the lower Eocene has high carbonate, primarily reflecting the drilling strategy of the leg to sample shallow early Eocene sediment sections. The Eocene had a very shallow CCD, roughly 3200–3300 meters below sea level (mbsl), compared to a late Pleistocene equatorial Pacific CCD of ~4600 mbsl (Heath et al., 1977; Delaney and Boyle, 1988; Lyle, Wilson, Janecek, et al., 2002; Lyle, 2003; Rea and Lyle, 2005; Lyle et al., this volume). The early Eocene transect was therefore drilled on ocean crust only slightly older than the P/E boundary to be near the shallow topography of the early Eocene rise crest (~2800 m). One interesting feature of the early Eocene is that carbonate was better preserved off the equator and not at the equator, opposite to all Neogene trends.
The early–middle Eocene transition, the warmest part of the Cenozoic, was poorly recovered at all Leg 199 sites because a significant numbers of chert horizons were found in this interval (Lyle, Wilson, Janecek, et al., 2002). During the leg there was significant speculation as to why the early/middle Eocene boundary was consistently chertified, but there has not yet been further study. Hypotheses that would explain extensive chert formation at the early/middle Eocene boundary range from the chemical to the kinetic, and the true explanation probably contains elements of many of the hypotheses. For example, biogenic opal sedimentation increased between the early and middle Eocene. Such a change in sediment composition may have initiated chert formation and the initial chert formation may have autocatalyzed further chertification. Because amorphous biogenic silica is converted to chert more rapidly in the presence of carbonate (Kastner et al., 1977), it may be significant that the chert interval lies just above the lower carbonates. Also, the conversion rate of biogenic opal to chert increases with temperature. It is probably more than coincidence that the warmest part of the Cenozoic, when bottom water temperatures exceeded 12°C, also has the highest development of cherts.
Oxygen isotope values trend toward heavier values all through the middle and late Eocene, implying a long-term trend toward cooling conditions (Zachos et al., 2001). However, other paleoceanographic proxies show significantly more variability. Average rates of biogenic deposition in the equatorial Pacific are high between 50 and 40 Ma, or during the entire middle Eocene (Moore et al., 2004). Sedimentation rates dropped consistently throughout the late Eocene, with the latest Eocene occasionally being represented by a hiatus in off-equator sites along the Leg 199 transect (Lyle, Wilson, Janecek, et al., 2002). Carbonate accumulation events occurred roughly every 2.5 m.y. in the middle Eocene but switched to ~1-m.y. intervals after 37.5 Ma (Lyle et al., this volume). The largest of the events (CAE-3, Lyle et al., this volume) occurred between 42 and 40 Ma and was also a productivity event (Olivarez Lyle and Lyle, this volume). The CAE-3 interval is the only one in the middle and late Eocene to have significant numbers of diatoms in the radiolarian oozes, up to 50% of the total biogenic silica fraction near the end of CAE-3. Steiger (this volume) performed a preliminary study on the environmental changes in the CAE-3 interval from a radiolarian faunal analysis. He interprets the changes in radiolarian fauna to indicate cooling during the event. There clearly is additional information preserved in the fauna, but much of this information depends upon further studies to better determine the environments of specific radiolarian species.
The Eocene/Oligocene boundary has long been recognized as the largest change of the CCD in the Cenozoic (van Andel et al., 1975; Heath et al., 1977; Delaney and Boyle, 1988; Rea and Lyle, 2005). However, until Leg 199 it was not recognized that the Eocene/Oligocene boundary also marked the change from a regime of high-amplitude CCD changes to one of relatively constant CCD. The equatorial Pacific CCD oscillated by several hundred meters a number of times during the Eocene (Rea and Lyle, 2005; Lyle et al., this volume) but has only changed by 250 m or so throughout the Neogene (Lyle, 2003) and only by ~100 m in the Pleistocene (Farrell and Prell, 1989).
The major change of CCD at the Eocene/Oligocene boundary is attributed to the sudden lowering of sea level associated with the first continent-wide glaciation of Antarctica (Coxall et al., 2005). Yet, Rea and Lyle (2005) show that, assuming a sea level fall of ~50 m (Miller et al., 2005), the expansion of abyssal seafloor area available for carbonate deposition is significantly larger than the loss of shelf and shallow sea area by sea level lowering. The area of abyssal carbonate deposition expanded by ~25% of global Earth area, while lost shelf area occupied only 4% of global Earth area. Lyle et al. (unpubl. data) argue that the weathering flux of Ca and alkalinity must have increased rapidly by ~20% at the E/O boundary or that the dissolved inorganic carbon pool must have dropped dramatically.
A series of transient CCD events occurred in the Eocene (Lyle, Wilson, Janecek, et al., 2002; Lyle et al., this volume). Lyle et al. (this volume) identified seven CCD events in the middle and late Eocene, demonstrated that high carbonate is associated with cooling, and showed that the largest event (CAE-3) ended because of carbon cycle changes. Such a change implies a large increase in atmospheric CO2 associated with the shallowing of the CCD by more than 800 m and the warming conditions. The large gyrations of the CCD in the Eocene imply that many stabilizing feedbacks of the Pleistocene and Holocene carbon cycle are missing in the Eocene.
Tripati et al. (2005a, 2005b) suggested that the CAE-3 cooling event was associated with 120 m of sea level change based upon comparison of oxygen isotopes (a combined ice volume and temperature signal) and Mg/Ca (controlled by temperature) in benthic foraminifers. This ice volume estimate is high in comparison to all available other lines of geological evidence, raising the possibility that the Mg/Ca temperature estimates are compromised. The CAE-3 transition is marked by major changes in carbonate dissolution, and strong dissolution artifacts are known to exist for the Mg/Ca paleothermometer (Lear et al., 2004). In addition, the basal section of sediments, just below the interval in question, contains euhedral diagenetic dolomite rendering these sediments of questionable reliability for paleothermometric study (Lyle, Wilson, Janecek, et al., 2002).
Independent evidence of sea level rise at the end of CAE-3 does not support the existence of a large ice cap. Seismic stratigraphy along the New Jersey margin suggests that the sea level rise at the end of CAE-3 was only ~30–50 m (Browning et al., 1996; Miller et al., 1998, 2005). The sheer volume of ice needed to cause a 120 m sea level drop is another important concern. Such an icecap is as large as the combined Northern and Southern Hemisphere ice caps at the LGM (21 ka), but evidence for extensive middle Eocene ice sheet development outside Antarctica does not exist. If the ice cap were confined to Antarctica, it would have to be double the thickness of the modern Antarctic ice cap under warmer conditions—a physical impossibility given the limits imposed on area by the Antarctic continent and strong dependence of ice sheet flow on stress (Coxall et al., 2005).
Clearly, the biogeochemical cycles of the Eocene equatorial Pacific clearly were organized differently than the Holocene equatorial Pacific. One of the most impressive features of the middle and late Eocene are the presence of sugary radiolarian oozes that contain ~70% biogenic silica. Despite the high biogenic opal contents, there are almost no diatom tests (only a few percent, based on smear slides). In contrast, Holocene equatorial Pacific sediments contain 2 to 10 times as many diatoms as radiolarians (Lyle, Wilson, Janecek, et al., 2002). The Eocene thus has a short supply of larger phytoplankton characteristic of the modern equatorial Pacific. New estimates of opal deposition rate (Moore, 2005) suggest that the Eocene actually had relatively low MAR of opal compared to the Neogene.
The Eocene equatorial Pacific also has extremely low Corg contents and Corg MAR that are an order of magnitude below those of the Holocene or Pleistocene equatorial Pacific. Olivarez Lyle and Lyle (this volume, 2006) explored the implications of these low Corg contents and propose that the low Corg burial is caused by higher metabolic demand by marine ecosystems under high-temperature conditions. Such a metabolic demand in both marine and terrestrial ecosystems is an important feedback maintaining warm climate conditions.
Olivarez Lyle and Lyle (this volume, 2006) first compared Corg MAR to biogenic Ba MAR (a proxy for productivity) (Dymond et al., 1992) and determined that productivity in the Eocene equatorial Pacific was not radically lower than the Holocene. Instead, low Corg was caused by much higher organic matter degradation before ultimate burial. After eliminating the possibility of much higher oxygen exposure times, the likely cause of such an increase in degradation is the temperature-dependence of heterotrophic metabolism. Metabolic respiration rates roughly double for a 10°C change in environmental temperature (Gillooly et al., 2001; Brown et al., 2004). Even though the amount of pelagic burial is small, changes in the net burial of Corg could significantly affect levels of atmospheric CO2 on timescales of 104 years. In addition, the increase in organic matter degradation would affect the other shallow organic carbon reservoirs in soils and deltaic sediments, so that the feedback is probably stronger than a prediction based on pelagic sediments alone.