Earth and Planetary Science Letters 228 (2004) 391 – 405 www.elsevier.com/locate/epsl

Circulation in the Southern Ocean during the Paleogene inferred from neodymium isotopes Howie D. Scher*, Ellen E. Martin Department of Geological Sciences, University of Florida, Gainesville, FL, USA Received 14 April 2004; received in revised form 10 October 2004; accepted 10 October 2004 Editor: E. Boyle

Abstract Long-term records of neodymium (Nd) isotopes from sedimentary archives can be influenced by both changes in water mass mixing and continental weathering. Results of Nd isotopic analyses of fossil fish teeth from ODP Site 689 (Maud Rise, Southern Ocean) provide a long, continuous, high-resolution marine sediment Nd isotope record (expressed in e Nd units). Correlation of down core secular variations between the e Nd record, d 13C values from benthic foraminifera, and clay mineral assemblages demonstrates that long-term variability of Nd isotope ratios reflect changes in ocean circulation, and that only minor fluctuations in e Nd values are associated with changes in continental weathering on Antarctica. Nonradiogenic e Nd values at Site 689 during the middle Eocene require the contribution of an end member with e Ndb 9.5. Southern Ocean deep water may have been too radiogenic in the middle Eocene (e Nd= 8.5), though this end member may not be fully characterized. A possible source of deep water outside of the Southern Ocean in the middle Eocene is the Tethys Sea (e Nd= 9.3 to 9.8). The presence of Warm Saline Deep Water (WSDW) on Maud Rise is consistent with the Nd isotope results. The onset of more radiogenic e Nd values at ~40.8 Ma coincides with other changes at Site 689 which are consistent with a switch from a warm bottom water mass in the middle Eocene to a colder bottom water mass in the late middle Eocene. A rapid shift to radiogenic e Nd values beginning at 37 Ma is best explained by the opening of Drake Passage. The shift coincides with increases in phytoplankton production throughout the Atlantic sector of the Southern Ocean that document the development of upwelling cells presumably related to more effective latitudinal circulation. After the Eocene/Oligocene boundary when large-scale ice sheets developed on Antarctica, Southern Ocean sourced water masses, such as Antarctic Intermediate Water (AAIW) and Antarctic Bottom Water (AABW), had a greater influence on the hydrography of the study area. An early Oligocene trend to nonradiogenic compositions resulted in similar values to the modern e Nd values of these water masses. The modern e Nd values of AAIW and AABW reflect a significant contribution of North Atlantic Deep Water (NADW), thus decreasing e Nd values in the early Oligocene may have resulted from the export of Northern Component Water (NCW, similar to modern NADW). During the late Oligocene and early Miocene, the long-term trends of the record follow benthic d 13C values. Variability in the Nd isotope record most likely reflects fluctuations in ocean circulation arising from changes in the relative contributions of

* Corresponding author. Tel.: +1 352 392 2231; fax: +1 352 392 9294. E-mail address: [email protected] (H.D. Scher). 0012-821X/$ - see front matter D 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.epsl.2004.10.016

392

H.D. Scher, E.E. Martin / Earth and Planetary Science Letters 228 (2004) 391–405

different end member water masses to the Southern Ocean. An interval where e Nd values and d 13C values are not correlated may reflect the influence of a short-lived weathering event on the e Nd record. Early Miocene e Nd values resemble those of modern Southern Ocean water masses, indicating a shift toward present-day patterns of ocean circulation. D 2004 Elsevier B.V. All rights reserved. Keywords: fossil fish teeth; neodymium isotope ratios; ocean circulation; Warm Saline Deep Water; Drake Passage

1. Introduction The Nd isotope ratio of seawater (143Nd/144Nd) demonstrates a strong correlation with water mass. Dissolved Nd has a short seawater residence time (600–2000 years; [1–3]) and is sourced predominantly from the continents, thus the Nd isotope ratio of a water mass tends to reflect the geology of its source area. Sedimentary archives that contain measurable levels of Nd are gaining recognition in paleoceanography because recent studies suggest that water mass mixing, i.e., ocean circulation, influences the Nd isotopic ratio of seawater. Therefore, Nd isotopes have the potential to yield patterns of past ocean circulation, for which there are few reliable proxies.

Proxies for ocean circulation are critical for examining links between ocean circulation and global climate change. However, applying Nd isotopes as a proxy for ocean circulation has been complicated by the competing influence of another paleoclimate signal, changes in continental weathering. The composition and/or provenance of material entering the source area of a water mass, can also bear upon the Nd isotopic ratio of seawater. Distinguishing between ocean circulation and continental weathering as sources of variability in long-term Nd isotope records is crucial for the application of Nd isotopes to Cenozoic paleoceanography. This paper reports the findings of a multi-proxy approach designed to deconvolve the contributing signals to the Nd isotope ratio of seawater in the

Fig. 1. Paleogeographic reconstruction of the late Eocene showing the location of ODP Site 689, the locations of other DSDP and ODP sites discussed in the text, and relevant tectonic gateways. Arrows illustrate possible pathways for deep water exchange between ocean basins. From the Ocean Drilling Stratigraphic Network (OSDN).

H.D. Scher, E.E. Martin / Earth and Planetary Science Letters 228 (2004) 391–405

Atlantic sector of the Southern Ocean. Nd isotope variability is compared to proxy records from the same location that record a strong response to changes in either continental weathering, as documented by clay mineralogy, or ocean circulation, as defined by d 13C values from benthic foraminifera. In this study, fossil fish teeth from ODP Site 689 (64.318S, 3.068E, 2080 m) (Fig. 1) were used to generate a high-resolution (average 270 ky) Nd isotope record over a 20-My interval from the middle Eocene to early Miocene. The high yield of fossil fish teeth permitted the construction of this record at a resolution that is unparalleled during this interval in the Cenozoic. ODP Site 689 has been extensively studied, and many proxy records have been generated that span the relevant interval. d 18O and d 13C values of benthic foraminifera were measured by Kennett and Stott [4], Mackensen and Ehrmann [5], and DiesterHaass and Zahn [6]. The relative abundance of clay minerals, which reflect continental weathering conditions on Antarctica, was determined by Ehrmann and Mackensen [7] and Robert et al. [8]. Paleoproductivity in the surface waters overlying Maud Rise have been derived from benthic foraminiferal accumulation rates [6]. Nd isotope variability is sufficiently resolved and can be directly compared to variability in other proxy records. The length of the record coupled with close spacing of samples provides an excellent opportunity to better understand the nature and causes of long-term secular variability of Nd isotopes in seawater.

2. Background 2.1. Seawater e Nd values Nd isotope investigations of seawater samples demonstrate that modern North Atlantic Deep Water (NADW) has an e Nd value of 13.5 [9], reflecting the weathering input of Archean age rocks into the Labrador Sea. e Nd units represent the difference of the 143Nd/144Nd ratio in parts per 104 from the chondritic uniform reservoir (CHUR) [10]. The e Nd value of modern Pacific seawater is very distinct (e Nd= 4) [11], which is due to the contribution of young, circum-Pacific volcanogenic sources. Water

393

masses sourced in the Southern Ocean such as Antarctic Bottom Water (AABW) and Antarctic Intermediate Water (AAIW) have intermediate e Nd values ~ 8 [1,12,13], which reflect mixing between Pacific and Atlantic seawater. The input of weathered material in the dissolved and suspended load of rivers can modify e Nd values of seawater proximal to such sources. For example, the Orange River in southern Africa drains terrains of Proterozoic age such as the Orange River Group (e Nd= 13.5 to 24) [14]. Seawater around South Africa has a nonradiogenic signature [15] presumably resulting from input from the Orange River. 2.2. Constraining sources of variability in Southern Ocean Nd records Long-term Nd isotope records from ferromanganese (Fe–Mn) crusts have demonstrated that presentday provinciality between the Pacific and Atlantic oceans has been maintained for much of the Cenozoic [16–20]. It follows that variability of Nd isotope ratios in Southern Ocean sedimentary records during the Cenozoic should reflect changes in the proportion of end member water masses mixing in the Southern Ocean, as well as changes in continental weathering. Although the e Nd values of water masses that are likely to influence the interpretation of Paleogene e Nd records from the Southern Ocean have been loosely constrained, the Nd isotopic contribution from continental weathering during the relevant interval must also be examined. The portion of the Nd isotope signal in Southern Ocean records that is attributable to continental weathering is likely to reflect the growth of ice sheets on Antarctica. However, it is difficult to estimate the e Nd value from weathering of Antarctic rocks because much of the continent is ice covered. From exposures above the ice, the spatial limit of Precambrian basement rocks have been traced and amount to a significant portion of the continent [21]. Geochemical investigations of basement exposures yield nonradiogenic Nd isotope signatures. Gabbros from northern Victoria Land demonstrate e Nd values that range from 14 to 19 [22] and granulites from the Wilson Terrain have e Nd values of 16 [23]. e Nd values from Grenville-age gneisses in the Maud Province are 10.5 [24]. Enderby Land outcrop

394

H.D. Scher, E.E. Martin / Earth and Planetary Science Letters 228 (2004) 391–405

samples have very low Nd isotope ratios (typically e Nd ~ 30 to 50), consistent with its Archean age [25–27]. Borg et al. [28–31] measured a large range of Nd isotope ratios (e Nd=0 to 35) on outcrop samples from the Transantarctic Mountains near the western Ross Sea. Thus, it is likely that material entering the Southern Ocean from weathering on Antarctica has an e Nd value that is less radiogenic than the modern seawater value of 8. 2.3. Archives of seawater Nd The application of Nd isotopes from Fe–Mn sediment archives to questions in paleoceanography have provided insight into gateway events [16,32], Northern Hemisphere glaciation [33], and the history of NADW export to the Southern Ocean [20,34]; however, these studies have been limited to the Neogene. Robust Nd isotope records from the Paleogene have been more elusive. Fe–Mn crusts do not adequately resolve the natural variability of Nd isotopes on Paleogene time scales due to very slow growth rates (1–15 mm/My) and poor age control beyond 10 Ma (see Frank [35] for a recent review). The resolution and age control limitations imposed on Fe–Mn Nd isotope records have not permitted precise correlation to other paleoceanographic proxy records, and have limited the ability to deconvolve contributing signals from ocean circulation and continental weathering, although it is clear that some long-term records do reflect weathering inputs [36]. In recent years, fossil fish teeth have been used to construct relatively high resolution Nd isotope records [37–40]. In a post-mortem mineralogical transformation of hydroxyfluorapatite to fluorapatite that occurs at the sediment–water interface, fish teeth acquire Nd concentrations that average 300 ppm. The post-mortem addition of Nd to fish teeth overwhelm very low levels of Nd that were incorporated in vivo, thus passing the e Nd signal of bottom water into fish teeth [41]. The e Nd signal carried by fossil fish teeth is resistant to alteration during burial and diagenesis. When exposed to similar bottom waters, Nd isotope data from fossil fish teeth corroborate those from Fe–Mn crusts [37] and authigenic Fe–Mn coatings [39]. The occurrence of high Nd concentrations in fossil fish teeth within precisely dated ODP sections has provided a means

to more accurately examine seawater Nd isotope variability during the Paleogene.

3. Methods 3.1. Analytical methods Fossil fish teeth were hand picked out of the N125-Am fraction of samples that had were washed and sieved with deionized water. Most samples consisted of three to seven teeth and were cleaned using the oxidative/reductive procedure after Boyle [42], Boyle and Keigwin [43], and Boyle (personal communication, 1993) to chemically remove Fe–Mn coatings. Samples were dissolved in aqua regia and the solution was transferred to clean Teflonk beakers. All samples were spiked for Nd concentration measurements. Selected samples were also spiked for Sm concentration measurements. Samples were then evaporated to dryness in preparation for cation exchange chemistry. Samples were redissolved in 0.75 N HCl and the solution was passed through a quartz glass column packed with Mitsubishik cation exchange resin. The column was washed with 1.7 N HCl to remove coexisting elements such as Ca and Mg. Sr was then eluted with 1.7 N HCl. The Sr isotope data for these samples is discussed in Martin and Scher [39]. Ba, which negatively impacts the analysis of Nd isotopes, was removed by washing the column with 2 N HNO3. The rare earth elements (REE) were then eluted with 4.5 N HCl. The solution containing the REE was evaporated to dryness then redissolved in 0.75 N HCl and passed through a separate though identical column packed with Mitsubishik cation exchange resin treated with NH4OH. The column was washed with distilled 0.2 M Alpha hydroxyisobuteric acid (AlphaHIBA) buffered to pH ~4.6 with NH4OH, to isolate Sm and subsequently to isolate Nd. The solutions containing purified Sm and Nd were evaporated to dryness, redissolved in ~20 Al of aqua regia to remove AlphaHIBA, and evaporated to dryness. The total Nd blank for this procedure is 6 pg. Isotopic ratios were analyzed on a Micromass Sector 54 thermal ionization mass spectrometer (TIMS) in dynamic mode at the University of Florida. Samples for Nd analysis were redissolved in 8 N

H.D. Scher, E.E. Martin / Earth and Planetary Science Letters 228 (2004) 391–405

395

Fig. 2. Nd isotope ratios from fossil fish teeth and polarity reversal stratigraphy vs. depth for ODP Site 689. Nd isotope ratios are plotted as e Nd(T) values, which are calculated from the average 147Sm/144Nd values of selected (samples Table 1 in the Appendix). Error bars are the 2r external reproducibility of the JNdi-1 and Ames Nd standards. The polarity reversal stratigraphy is from Spieh [47] and is shown with reference to geologic epochs. The diagonal lines in the polarity reversal stratigraphy at 66.86 and 65.5 mbsf represent the unconformities discussed in Section 4.2.

nitric acid and loaded onto zone-refined Re filaments with silica oxide gel, and analyzed as NdO+. Using 142 Nd16O as the monitor peak, beams of 0.5 V were measured for 200 ratios. Mass fractionation was corrected to 146Nd16O/144Nd16O=0.722254. Samples for Sm analysis were redissolved in 8 N nitric acid and loaded onto Tantalum filaments. Replicate analyses of an internal standard (AMES Nd) during the 6 months in which samples were analyzed yielded 0.512138 (F0.000012, 2r external reproducibility, n=40). Replicate analyses of the international Nd standard JNdi-1 from September, 2003 to February, 2004 yielded 0.512102 (F0.000012, 2j external reproducibility, n=65). No correction has been applied to the Nd isotope data. Internal measurement errors of samples are listed in Table 1 in the Appendix. Ten samples from this study were spiked and analyzed for Sm in order to determine the 147Sm/144Nd ratios preserved by the teeth at various levels in the core. The range of 147Sm/144Nd values from teeth at ODP Site 689 is 0.1212–0.1303, in agreement with 147 Sm/144Nd values from fossil fish teeth in other marine cores [37,38,40]. An average 147Sm/144Nd value of 0.1248 was applied to all samples to calculate

e Nd(T) values. This correction ranges from 0.4 to 0.2 e Nd units between the oldest and youngest samples (see Table 1). 3.2. Age model The age model used in this study is from Mead and Hodell [44] and was modified to the time scale of Cande and Kent [45]. Two unconformities are present in the upper part of the section. At 66.86 mbsf, early Miocene sediments lie unconformably over late Oligocene sediments (~5 My hiatus). The second, at 65.5 mbsf, lies within a normally magnetized interval and was recognized by biostratigraphy [46] and Sr isotopes [44]. The sediments overlying this unconformity (b1 My hiatus) have been assigned to Chron C5En by means of Sr isotope chemostratigraphy [44].

4. Results Sm and Nd isotope data are reported in Table 1 in the Appendix. Nd concentrations of these samples are discussed in Martin and Scher [39]. Nd isotope results

396

H.D. Scher, E.E. Martin / Earth and Planetary Science Letters 228 (2004) 391–405

are presented as e Nd(T) values and are plotted as a function of depth (Fig. 2). To effectively compare other proxy records from Site 689, the Nd isotope data has been plotted against depth, and the magnetostratigraphy [47] is shown so the results can be discussed with reference to relevant geologic epochs. In addition the Nd isotope data are plotted against age to show the ages of paleoceanographic events inferred from the data. Clearly, there is variability in the Nd isotope data that exceeds analytical precision (Fig. 2). The e Nd record is dominated by a pattern of secular variability that begins with very nonradiogenic values in the middle Eocene, then shifts stepwise to the most radiogenic values observed in the record by the latest Eocene. Through the Oligocene and Miocene there are variations of ~1 e Nd unit. First-order fluctuations are generally smooth, well-resolved shifts in e Nd values with amplitudes exceeding reported external reproducibility. The Nd isotopic results are discussed with

reference to four time slices that display unique Nd isotopic patterns. 4.1. Middle Eocene (183–140 mbsf, 46–37 Ma) This interval is distinguished by a prominent step in e Nd values that occurs at 162 mbsf. Below the step, e Nd values average 9.25 and display little variability. At 162 mbsf, e Nd values step up 0.75 e Nd units to an average value of 8.5 for the remainder of the middle Eocene. 4.2. Late Eocene (140–121 mbsf, 37–33.7 Ma) The late Eocene interval is dominated by a very pronounced increase of 1.15 e Nd units. In a welldefined shift beginning at 139.50 mbsf, values increase from 8.5 in the late Eocene to 7.35 at 125 mbsf, in the latest Eocene. The values that culminate this increasing trend are the most radiogenic

Fig. 3. e Nd record (top panel), clay mineral assemblages (middle panels), and benthic d 18O (bottom panel) vs. depth for ODP Site 689. The data for the clay mineral assemblages are from Ehrmann and Mackensen [7]. Benthic d 18O data are from Diester-Haass and Zahn [6]. The gray bars originating at the bottom of the diagram call attention to dramatic changes in the clay mineral assemblage and d 18O record. The gray bars emanating from the top of the diagram highlight the dramatic shifts in Nd isotope ratios.

H.D. Scher, E.E. Martin / Earth and Planetary Science Letters 228 (2004) 391–405

values measured and are more radiogenic than present-day e Nd values at this location [1]. e Nd values rapidly decrease to 8.35 in the latest Eocene, but radiogenic values are restored by the lowermost Oligocene. 4.3. Early Oligocene (121–93 mbsf, 33.7–28.5 Ma) In the early Oligocene interval, e Nd values decrease towards nonradiogenic compositions averaging 8.5 by the late early Oligocene. 4.4. Late Oligocene to early Miocene (93–60 mbsf, 28.5–16 Ma) During the late Oligocene, e Nd values again increase, reaching radiogenic compositions around 7.9; however, this trend is interrupted by a rapid excursion to nonradiogenic values of 9.1 beginning at 88 mbsf. The departure to nonradiogenic compositions is brief and by 80 mbsf e Nd values have recovered to radiogenic compositions of 7.65. The late Oligocene interval ends abruptly at the hiatus at 66.86 mbsf and is overlain by middle early Miocene sediments with e Nd values averaging 8.7.

5. Sources of long-term Nd isotope variability 5.1. Comparison to records of continental weathering The group of primary clay minerals consisting of smectite, illite, kaolinite and chlorite present in deep sea sediment are initially formed on nearby continents. The relative abundances of these clay minerals are indicative of various weathering processes, which are ultimately controlled by climate [48]. Illite and chlorite are chemically immature and dominate clay mineral assemblages in regions characterized by physical weathering [49–51]. The occurrence of smectite in marine sedimentary sequences can be indicative of chemical weathering under warm, humid conditions [48]. Kaolinite is generally indicative of intense chemical weathering conditions, though it can occur in polar sedimentary sequences from the mechanical weathering of kaolinite deposits [48,52,53]. Changes in the style of continental weathering on Antarctica that

397

resulted from the growth of ice sheets are recorded in the clay mineral assemblage from Site 689 [7,54,55]. Dramatic changes in continental weathering style are indicated by two intervals of pronounced change in the middle Eocene and early Oligocene (Fig. 3). The pre-Oligocene history of Antarctic glaciation has been inferred from direct evidence in the form of glaciomarine sediments in cores from the Antarctic margin [53,56–58]. Deposits of waterlain glacial tills, sands and diamictites indicate brief and localized episodes of glaciation on Antarctica in the middle and late Eocene. At Site 689, the appearance of chlorite in detectable quantities at 154 mbsf (38.6 Ma) reflects an increase in physical weathering associated with these early glaciations (Fig. 3). The simultaneous appearance of kaolinite in this assemblage is likely due to the physical weathering of older kaolinite bearing sediments on Antarctica, such as those found in the Beacon Supergroup [7]. There is no shift in Nd isotopes at Site 689 that corresponds with the change in clay minerals at 154 mbsf; instead e Nd values between 154.44 and 153.33 mbsf are virtually unchanged. There is a 0.5 e Nd unit increase in slightly younger material between 153.33 and 151.40 mbsf (Fig. 3), however, it is unlikely that the proximity of the increase in e Nd units to the shift in clay minerals is significant. This is in part because an increase in weathered material derived from Antarctica should result in decreasing e Nd values. Moreover, shifts in e Nd values of this magnitude are observed in other parts of the record, including earlier intervals when there is no documented evidence of glaciation on Antarctica. At approximately 122 mbsf, coinciding with the oxygen isotope shift at the Eocene/Oligocene boundary, a major change occurs in the character of the clay mineral assemblage (Fig. 3). During the Eocene, illite amounted to less than 20% of the total clay accumulating at this site. At ~122 mbsf, illite increased to about 60%, and remained high throughout the remainder of the record. This shift reflects the switch from predominant chemical weathering conditions, which prevailed on Antarctica before the early Oligocene glaciation, to a physical weathering regime following the build up of ice sheets. A rapid excursion to nonradiogenic e Nd values occurs from 124.24 to 120.95 mbsf that may be related to the early Oligocene glaciation as recorded by d 18O values (Fig. 3). The shift may reflect a brief interval when a very large amount of nonradiogenic Nd

398

H.D. Scher, E.E. Martin / Earth and Planetary Science Letters 228 (2004) 391–405

was delivered to the Southern Ocean, released by mechanical weathering of Precambrian basement provinces on Antarctica during rapid ice sheet growth. Radiogenic values are restored by 118.7 mbsf, thus if e Nd values did respond to such a weathering event, the effects were transitory. The changes observed in the clay mineral assemblage described above are associated with short term fluctuations in e Nd values. However, long-term variability of e Nd values is independent from changes in the style of continental weathering on Antarctica (Fig. 3). There is no change in the clay mineral assemblage surrounding the large step in e Nd values during the middle Eocene. Likewise, the sharp increase in e Nd values observed during the late Eocene is not associated with large changes in the relative abundance of clay minerals. The difference in the timing of the shifts between Nd isotope ratios and clay minerals suggests that changes in continental weathering are not responsible for the first order secular variability observed in the Nd isotope record from Site 689. It is likely, then, that the Nd isotope record reflects changes in ocean circulation. 5.2. Comparison to benthic d 13C Similarities between the Nd isotope record and the record of d 13C values of benthic foraminifera Cibicidoides [6] also support the idea the Nd isotope record

reflects changes in ocean circulation. After the late Eocene, variations in e Nd values display a close coherence to the long-term trend of the benthic d 13C record (Fig. 4). While benthic foraminiferal d 13C is often used as a nutrient proxy to reconstruct ocean circulation, the signal can be overprinted by productivity changes in surface waters and changes in the size of the oceanic carbon reservoir. Mackensen and Ehrmann [5] demonstrated that benthic d 13C trends at Site 689 do not reflect ocean circulation during the Eocene on the basis that similar trends are observed in sites at a range of depths on the Kerguelan Plateau. However, it was concluded that benthic d 13C variability in the Oligocene do reflect ocean circulation because the trend to lighter d 13C values at Site 689 is not observed at Sites 738 and 744 (Fig. 1). Moreover, an increase in local productivity over Maud Rise that could have led to lighter benthic d 13C values during the Oligocene is not supported by paleoproductivity proxies (Fig. 5). Despite differences in the geochemical cycling of Nd and carbon in seawater, and different host phases for these elements in marine sediments, the similarity between long-term trends of the e Nd and d 13C records during the Oligocene suggests that both tracers respond to the same paleoceanographic signal. The advantage of the e Nd record is that it provides more information regarding the source area of water mass end members.

Fig. 4. Nd isotope ratios from fossil fish teeth and y13C from benthic foraminifer Cibicidoides vs. depth for ODP Site 689. The benthic d 13C record (dashed line) is from Diester-Haass and Zahn [6].

H.D. Scher, E.E. Martin / Earth and Planetary Science Letters 228 (2004) 391–405

399

Fig. 5. Nd isotope ratios from fossil fish teeth and paleoproductivity vs. age for ODP Site 689. The age model used for Site 689 is from Mead and Hodell [44] and has been recalibrated to the ages of Cande and Kent [45]. The paleoproductivity record is from Diester-Haass and Zahn [6] and is derived from the abundance of benthic foraminifera.

6. Southern Ocean paleoceanography from Nd isotopes

mixing; though some water mass end member compositions are poorly constrained.

A rudimentary understanding of global seawater Nd isotope patterns during the Paleogene is provided by previous investigations of marine glauconite deposits [59], Fe–Mn crusts [16–20] and fossil fish teeth [38–40]. With this limited Nd isotope database for the Paleogene ocean, the Nd isotope record from Site 689 can be interpreted in terms of water mass

6.1. Middle Eocene deep water sources Reconstructions of deep water circulation during the Paleogene indicate that the Southern Ocean was the predominant source of deep water [60–63]. However, e Nd values at Site 689 during the middle Eocene (e Nd= 9.1 to 9.5) are slightly less radio-

Fig. 6. Nd isotope ratios vs. age for the early Miocene through middle Eocene. The data for ODP Site 689 are from this study. The Nd isotope records from the Atlantic, Pacific and Indian were measured from ferromanganese crusts from abyssal depths [16–18]. The shaded rectangle denotes the range of qNd values estimated of the Tethys Sea from middle Eocene age authigenic glauconite deposits in the Helvetic belt of the Alps [59].

400

H.D. Scher, E.E. Martin / Earth and Planetary Science Letters 228 (2004) 391–405

genic than the only published estimate of Southern Ocean deep water (e Nd= 9.1 in the early Eocene [38]). Thomas et al. [38] did observe e Nd values as low as 10.6 on Maud Rise briefly in the early Eocene, however e Nd values from ODP Site 1090 (Fig. 1) are much less radiogenic (e Nd= 8.5 [64]) during the middle Eocene. There are two possible explanations for the Nd isotope data at Site 689. First, if Southern Ocean deep water during the middle Eocene is not fully constrained by available data, then Site 689 e Nd values may simply reflect a middle Eocene analogue to the deep water source described in Thomas et al. [38]. The other possibility is that middle Eocene e Nd values at Site 1090 are characteristic of Southern Ocean deep water [64] and the observation that Site 689 is less radiogenic than this end member requires that the nonradiogenic signal was propagated to Site 689 from a deep water source outside of the Southern Ocean. The only known sources of seawater outside of the Southern Ocean with such a nonradiogenic signal (e Ndb 9.5) in the middle Eocene were the North Atlantic (e Nd = 11) [16,18,19] and the Tethys Sea (e Nd= 9.3 to 9.8) [59] (Fig. 6). The North Atlantic end member e Nd value is based on Fe–Mn crusts, while the Tethys seawater estimate is based on e Nd values of Rb–Sr and K–Ar dated glauconite deposits from the Helvetic belt of the Alps, interpreted as the northern continental shelf of the Tethys Sea. Downwelling in the North Atlantic may have been possible during the middle Eocene on the basis of seasurface temperature (SST) constraints from d 18O values of planktonic foraminifera, which indicate a temperature of ~13 8C in the northeastern Atlantic [65]. However, export of deep water from the North Atlantic to the Southern Ocean in the middle Eocene is unlikely based on reconstructions of deep ocean circulation using d 13C values of benthic foraminifera [61,63]. In these reconstructions, values in the Southern Ocean remain high relative to the North Atlantic and Pacific indicating the dominance of Southern Ocean deep water. The hypothesis that a subtropical source of deep water, known as Warm Saline Deep Water (WSDW), formed in regions of net evaporation during warm climate intervals has been contentious since it was first proposed by Chamberlin [66]. Many geochem-

ical, faunal, and sedimentological records have been interpreted as reflecting a shift from high latitude deep water production to an inferred source of low latitude deep water production [4,60,61,62,67–72]. However, much of the evidence for a low latitude deep water source is equivocal, no direct evidence exists for WSDW, and general circulation models (GCMs) with early Paleogene boundary conditions fail to produce a stable mode of salinity-induced downwelling in the Tethys Sea, which was a large, low latitude seaway at the time [73]. Yet the presence of Tethys-derived seawater in the Southern Ocean appears to be the most likely explanation for the Nd isotope data from Site 689, and thus supports the production of WSDW in the low latitude Tethys Sea and subsequent export to the Southern Ocean. Further, Nd isotope investigations may strengthen or weaken the WSDW interpretation, however it does corroborate previously published geological evidence for a warm water mass on Maud Rise. Kennett and Stott [4] attributed reversed depth gradients of oxygen isotope data from Sites 689 and 690 to differences in bottom water temperatures, concluding that incursions of a warm deep water mass beneath colder Southern Ocean seawater led to the 0.5x difference in d 18O values. Other changes at Site 689 that coincide with the onset of radiogenic values at 40.8 Ma (~162 mbsf), such as increasing d 18O values (Fig. 3) and decreased carbonate preservation [74], are also consistent with a change from a warm bottom water mass during the middle Eocene to colder bottom water in the late middle Eocene. 6.2. Constraints on Drake Passage The timing of the opening of Drake Passage has been a long-standing debate driven by the hypothesis of Kennett [75] linking the opening of Drake Passage to initiation of the Antarctic Circumpolar Current (ACC) and development of ice sheets on Antarctica. Estimates for the opening of Drake Passage to deep water flow range from around the Oligocene/Miocene boundary [76–78] to the early Oligocene [79, 80]. Despite numerous attempts to constrain the opening by dating the onset of the ACC with other proxies (see recent review by Barker and Thomas [81]) the debate has endured.

H.D. Scher, E.E. Martin / Earth and Planetary Science Letters 228 (2004) 391–405

Nd isotopes offer an interesting approach to this problem in that the Nd isotopic ratio of Pacific and Atlantic deep waters comprise, respectively, the most and least radiogenic seawater in the ocean, a distinction that has persisted for much of the Cenozoic [16–19] (Fig. 6). Assuming that the radiogenic signature of Pacific seawater was effectively absent from the Atlantic sector of the Southern Ocean when Drake Passage was closed, the opening of Drake Passage should introduce the radiogenic fingerprint of Pacific seawater to the study area. In the late Eocene, a dramatic shift in e Nd values leads to the most radiogenic values observed in this study. The resulting value (e Nd= 7.3) is more radiogenic than any intermediate or deep water mass in the present-day Southern Ocean [1,12,13,82]. The only obvious explanation for such radiogenic e Nd values is the influx of Pacific deep water into the study area. The calculated paleodepth curve for Site 689, using the thermal subsidence model of Parsons and Sclater [83] indicates a depth of 1600 m in the late Eocene [72]. Thus, the data supports that Drake Passage was open to shallow and possibly intermediate depths by the late Eocene (~37 Ma). This estimate is in excellent agreement with Diester-Haass and Zahn [6], who reached a similar conclusion based on an increase in proxy measurements of local surface productivity during the late Eocene (Fig. 5). Increased phytoplankton production also occurred in other parts of the Atlantic sector of the Southern Ocean during the late Eocene [84], indicating a change in the nutrient profile of the surface layer, presumably resulting from the development of upwelling cells. Based on the Nd isotope data, widespread changes in productivity in the Atlantic sector of the Southern Ocean can be linked to the opening of Drake Passage to intermediate depths and the association of accelerated ocean currents with more effective latitudinal circulation. Variability of benthic d 13C values which accompany the shift in e Nd values likely reflect more pronounced changes in ventilation of the water column [6]. An alternative pathway for Pacific seawater to the vicinity of Site 689 in the late Eocene was through the Central American Seaway and subsequent export into the southern high latitudes (Fig. 1). It is intriguing that the Nd isotope record from DSDP Site 357 on Rio Grande Rise shows a shift to radiogenic e Nd values in

401

the middle Eocene [85], which supports a pathway for Pacific water into the tropical South Atlantic. However, the preferred interpretation for the Nd isotope data from Site 689 is the pathway through Drake Passage, as it is a more direct route to the study area and is consistent with other observations. 6.3. Oligocene to early Miocene ocean circulation During the Oligocene and Miocene, long-term variability of e Nd values follows the benthic d 13C record (Fig. 4), and most likely reflects fluctuations in ocean circulation arising from changes in the relative contributions of different end member water masses to the Southern Ocean. Following the rapid shift to radiogenic values in the late Eocene, e Nd values decrease slightly in the early Oligocene to a lower mean value of ~ 8.5, close to values observed in present-day Southern Ocean sourced water masses, AAIW and AABW [1,12,13,82]. The increasing influence of these water masses following the major glaciation of Antarctica likely played a more significant role in the hydrography of the study area [65]. The modern e Nd values of Southern Ocean sourced water masses reflect a significant contribution of NADW, so decreasing e Nd values at this time may have resulted from the export of Northern Component Water (NCW, similar to modern NADW) [86]. Southwesterly dipping downlap reflections of early Oligocene sediments within the Southeast Faeroes drift provide evidence of a southerly flow regime in the North Atlantic [87]. During the late Oligocene, there is a long-term increase in e Nd values, followed by a decrease in e Nd values through 24.8 Ma (66.90 mbsf), just below the hiatus at 66.86 mbsf where the Oligocene section ends. Strengthening of the ACC during the late Oligocene as suggested by Barker [78] may account for some of the positive fluctuations in e Nd values. A short-lived excursion is superimposed upon the longterm trend, from 28.2 to 27.13 Ma (87.96 to 80.23 mbsf ) when e Nd values fall to very nonradiogenic values (e Nd= 9.1), approaching values observed during the middle Eocene. It is not immediately clear what this excursion represents. Based on the similarity of the values during the excursion to middle Eocene values, it is possible that WSDW was present briefly at the study area in the late Oligocene, though there is

402

H.D. Scher, E.E. Martin / Earth and Planetary Science Letters 228 (2004) 391–405

no other evidence to support this supposition. An alternative explanation is that the excursion represents a pulse of NCW. There is stable isotopic evidence for a similar pulse of NCW during the early Oligocene [86]. A final possibility is that a climatically induced weathering event introduced nonradiogenic Nd into the Southern Ocean. The long-term decrease in e Nd values at the top of the Oligocene section follows the trend of the benthic d 13C record (Fig. 4). Above the hiatus at 66.86 mbsf, early Miocene e Nd values average 8.5. This value is similar to modern e Nd values of deep and intermediate waters in the Southern Ocean and perhaps reflects the beginning of modern deep water circulation patterns as suggested by Woodruff and Savin [88].

7. Conclusions Nd isotope ratios have been measured from middle Eocene to early Miocene age fossil fish teeth from ODP Site 689. The record represents the longest highresolution record of Paleogene Nd isotopes. Using multiple paleoceanographic proxy records, the contributing signals from ocean circulation and continental weathering were deconvolved from the Nd isotope record enabling an examination of the nature and causes of secular variability of Nd isotopes at Site 689 through the Paleogene. Long-term secular variations of Nd isotopes are not associated with major changes in continental weathering on Antarctica as revealed by changes in clay mineral abundances. Instead, a close correspondence with d 13C values is observed, suggesting the Nd isotope variability reflects changes in deep water circulation. Though the e Nd and d 13C records demonstrate similar trends, the e Nd record provides more information about the circulation of water mass end members. Therefore, the Nd isotope data can be used to examine the evolution of ocean circulation in the Southern Ocean during the Paleogene. The e Nd values of the water mass overlying Site 689 during the middle Eocene are less radiogenic than estimates for Southern Ocean deep water at the time and require the contribution of a water mass with e Nd b 9.5. The contribution from a Tethys Sea end

member, known as WSDW, provides an intriguing explanation for the data. This interpretation is consistent with oxygen isotope data and sedimentological evidence for WSDW on Maud Rise during the Paleogene. A dramatic shift toward radiogenic e Nd values in the late Eocene is best explained by an influx of Pacific seawater into the Atlantic Ocean, signifying the opening of Drake Passage by 37 Ma. Increases in phytoplankton production throughout the Atlantic sector of the Southern Ocean also occur in the late Eocene, and indicate the development of upwelling cells associated with more effective latitudinal circulation. The Nd isotope data places important constraints on the timing of the opening of Drake Passage and indicates that flow through Drake Passage was established to shallow, and possibly intermediate depths, prior to large-scale development of ice sheets on Antarctica. Long-term variability of e Nd values during the Oligocene and early Miocene closely follows the trend of benthic d 13C values. The only exception occurs in the late Oligocene during a short-lived excursion to nonradiogenic e Nd values similar to those observed in the middle Eocene. It is unclear whether the excursion reflects the fingerprint of a pulse of nonradiogenic seawater to the study area, such as NCW or WSDW, or a climatically induced weathering event that introduced nonradiogenic material from Antarctica into the Southern Ocean. During the early Miocene, e Nd values average 8.5 which is similar to modern values for deep and intermediate water in the Southern Ocean and perhaps reflects the emergence of circulation patterns similar to today.

Acknowledgments We appreciate reviews by Tim Bralower and Martin Frank whose comments have led to improvements in the manuscript. HDS is grateful to J. Lyons for providing valuable assistance in the lab. A debt of gratitude is extended to R. Thomas for his technical expertise with the TIMS at UF. Samples were provided by the Ocean Drilling Program. This research was supported by an NSF Career award to EEM (OCE-962970).

H.D. Scher, E.E. Martin / Earth and Planetary Science Letters 228 (2004) 391–405

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/ j.epsl.2004.10.016.

References [1] C. Jeandel, Concentration and isotopic composition of Nd in the South Atlantic Ocean, Earth Planet. Sci. Lett. 117 (1993) 581 – 591. [2] C. Jeandel, J.K. Bishop, A. Zindler, Exchange of neodymium and its isotopes between seawater and small and large particles in the Sargasso Sea, Geochim. Cosmochim. Acta 59 (1995) 535 – 547. [3] K. Tachikawa, C. Jeandel, M. Roy-Barman, A new approach to the Nd residence time in the ocean; the role of atmospheric inputs, Earth Planet. Sci. Lett. 170 (1999) 433 – 446. [4] J.P. Kennett, L.D. Stott, Proteus and Proto-Oceanus; ancestral Paleogene oceans as revealed from Antarctic stable isotopic results; ODP Leg 113, Proc. Ocean Drill. Program Sci. Results 113 (1990) 865 – 880. [5] A. Mackensen, W.U. Ehrmann, Middle Eocene through early Oligocene climate history and paleoceanography in the Southern Ocean; stable oxygen and carbon isotopes from ODP Sites on Maud Rise and Kerguelen Plateau, Mar. Geol. 108 (1992) 1 – 27. [6] L. Diester-Haass, R. Zahn, Eocene–Oligocene transition in the Southern Ocean; history of water mass circulation and biological productivity, Geology 24 (1996) 163 – 166. [7] W.U. Ehrmann, A. Mackensen, Sedimentological evidence for the formation of an East Antarctic ice sheet in Eocene/ Oligocene time, Palaeogeogr. Palaeoclimatol. Palaeoecol. 93 (1992) 85 – 112. [8] C. Robert, L. Diester-Haass, H. Chamley, Late Eocene– Oligocene oceanographic development at southern high latitudes, from terrigenous and biogenic particles; a comparison of Kerguelen Plateau and Maud Rise, ODP Sites 744 and 689, Mar. Geol. 191 (2002) 37 – 54. [9] D.J. Piepgras, G.J. Wasserburg, Rare earth element transport in the western North Atlantic inferred from Nd isotopic observations, Geochim. Cosmochim. Acta 51 (1987) 1257 – 1271. [10] D.J. DePaolo, G.J. Wasserburg, Nd isotopic variations and petrogenetic models, Geophys. Res. Lett. 3 (1976) 249 – 252. [11] D.J. Piepgras, S.B. Jacobsen, The isotopic composition of neodymium in the North Pacific, Geochim. Cosmochim. Acta 52 (1988) 1373 – 1381. [12] D.J. Piepgras, G.J. Wasserburg, Isotopic composition of neodymium in waters from the Drake Passage, Science 217 (1982) 207 – 214. [13] C.J. Bertram, H. Elderfield, The geochemical balance of the rare earth elements and neodymium isotopes in the oceans, Geochim. Cosmochim. Acta 57 (1993) 1957 – 1986.

403

[14] D.L. Reid, H.J. Welke, A.J. Erlank, A. Moyes, The Orange River Group; a major Proterozoic calcalkaline volcanic belt in the western Namaqua Province, Southern Africa; geochemistry and mineralization of Proterozoic volcanic suites, Geol. Soc. London Spec. Publ. 33 (1987) 327 – 346. [15] F. Albare`de, S.L. Goldstein, World map of Nd isotopes in seafloor ferromanganese deposits, Geology 20 (1992) 761 – 763. [16] K.W. Burton, H. Ling, R.K. O’Nions, Closure of the Central American Isthmus and its effect on deep-water formation in the North Atlantic, Nature 386 (1997) 382 – 385. [17] H.F. Ling, K.W. Burton, R.K. O’Nions, B.S. Kamber, F. von Blanckenburg, A.J. Gibb, J.R. Hein, Evolution of Nd and Pb isotopes in Central Pacific seawater from ferromanganese crusts, Earth Planet. Sci. Lett. 146 (1997) 1 – 12. [18] R.K. O’Nions, M. Frank, F. von Blanckenburg, H.F. Ling, Secular variation of Nd and Pb isotopes in ferromanganese crusts from the Atlantic, Indian and Pacific oceans, Earth Planet. Sci. Lett. 155 (1998) 15 – 28. [19] K.W. Burton, D. Lee, J.N. Christensen, A.N. Halliday, J.R. Hein, Actual timing of neodymium isotopic variations recorded by Fe–Mn crusts in the western North Atlantic, Earth Planet. Sci. Lett. 171 (1999) 149 – 156. [20] M. Frank, N. Whiteley, S. Kasten, J.R. Hein, K. O’Nions, North Atlantic Deep Water export to the Southern Ocean over the past 14 Myr; evidence from Nd and Pb isotopes in ferromanganese crusts, Paleoceanography 17 (2002) 12.1 – 12.10. [21] S.L. Harley, Archaean–Cambrian crustal development of East Antarctica; metamorphic characteristics and tectonic implications; Proterozoic East Gondwana; supercontinent assembly and breakup, Geol. Soc. London Spec. Publ. 206 (2003) 203 – 230. [22] L. Dallai, C. Ghezzo, Z.D. Sharp, Oxygen isotope evidence for crustal assimilation and magma mixing in the Granite Harbour Intrusives, northern Victoria Land, Antarctica, Lithos 67 (2003) 135 – 151. [23] F. Talarico, L. Borsi, B. Lombardo, Relict granulites in the Ross Orogen of northern Victoria Land (Antarctica): II. Geochemistry and palaeo-tectonic implications; tectonics of East Antarctica, Precambrian Res. 75 (1995) 157 – 174. [24] C.D. Wareham, R.J. Pankhurst, R.J. Thomas, B.C. Storey, G.H. Grantham, J. Jacobs, B.M. Eglington, Pb, Nd, and Sr isotope mapping of Grenville-age crustal provinces in Rodinia, J. Geol. 106 (1998) 647 – 659. [25] D.J. DePaolo, W.I. Manton, E.S. Grew, M. Halpern, Sm–Nd, Rb–Sr and U–Th–Pb systematics of granulite facies rocks from Fyfe Hills, Enderby Land, Antarctica, Nature 298 (1982) 614 – 618. [26] L.P. Black, J.W. Sheraton, P.R. James, Late Archaean granites of the Napier Complex, Enderby Land, Antarctica; a comparison of Rb–Sr, Sm–Nd and U–Pb isotopic systematics in a complex terrain, Precambrian Res. 32 (1986) 343 – 368. [27] L.P. Black, M.T. McCulloch, Evidence for isotopic equilibration of Sm–Nd wholerock systems in early Archaean crust of Enderby Land, Antarctica, Earth Planet. Sci. Lett. 82 (1987) 15 – 24. [28] S.G. Borg, E. Stump, B.W. Chappell, M.T. McCulloch, D. Wyborn, R.L. Armstrong, J.R. Holloway, Granitoids of

404

[29]

[30]

[31]

[32]

[33]

[34]

[35]

[36]

[37]

[38]

[39]

[40]

[41]

[42] [43]

[44]

H.D. Scher, E.E. Martin / Earth and Planetary Science Letters 228 (2004) 391–405 northern Victoria Land, Antarctica; implications of chemical and isotopic variations to regional crustal structure and tectonics, Am. J. Sci. 287 (1987) 127 – 169. S.G. Borg, D.J. DePaolo, B.M. Smith, Isotopic structure and tectonics of the central Transantarctic Mountains, J. Geophys. Res., B, Solid Earth Planets 95 (1990) 6647 – 6667. S.G. Borg, D.J. DePaolo, A tectonic model of the Antarctic Gondwana margin with implications for southeastern Australia; isotopic and geochemical evidence; accretionary tectonics and composite continents, Tectonophysics 196 (1991) 339 – 358. S.G. Borg, D.J. DePaolo, Laurentia, Australia, and Antarctica as a late Proterozoic supercontinent; constraints from isotopic mapping, Geology 22 (1994) 307 – 310. M. Frank, B.C. Reynolds, R.K. O’Nions, Nd and Pb isotopes in Atlantic and Pacific water masses before and after closure of the Panama Gateway, Geology 27 (1999) 1147 – 1150. F. von Blanckenburg, T.F. Naegler, Weathering versus circulation-controlled changes in radiogenic isotope tracer composition of the Labrador Sea and North Atlantic deep water, Paleoceanography 16 (2001) 424 – 434. R.L. Rutberg, S.R. Hemming, S.L. Goldstein, Reduced North Atlantic deep water flux to the glacial Southern Ocean inferred from neodymium isotope ratios, Nature 405 (2000) 935 – 938. M. Frank, Radiogenic isotopes; tracers of past ocean circulation and erosional input, Rev. Geophys. 40 (2002) DOI:10.1029/2000RG000094. M. Frank, van de Flierdt, Tina, A.N. Halliday, P.W. Kubik, B. Hattendorf, D. Gunther, Evolution of deep water and weathering inputs in the central Atlantic Ocean over the past 33 Myr, Paleoceanography 18 (2003) DOI:10.1029/2003PA000919. E.E. Martin, B.A. Haley, Fossil fish teeth as proxies for seawater Sr and Nd isotopes, Geochim. Cosmochim. Acta 64 (2000) 835 – 847. D.J. Thomas, T.J. Bralower, C.E. Jones, Neodymium isotopic reconstruction of late Paleocene–early Eocene thermohaline circulation, Earth Planet. Sci. Lett. 209 (2003) 309 – 322. E. Martin, H. Scher, D., Preservation of seawater Sr and Nd isotopes in fossil fish teeth: bad news and good news, Earth Planet. Sci. Lett. 220 (2004) 25 – 39. D.J. Thomas, Evidence for deep-water production in the North Pacific Ocean during the early Cenozoic warm interval, Nature 430 (2004) 65 – 68. H. Elderfield, R. Pagett, Rare earth elements in ichthyoliths: variations with redox conditions and depositional environments, Sci. Total Environ. 49 (1986) 175 – 197. E.A. Boyle, Cadmium, zinc, copper, and barium in foraminifera tests, Earth Planet. Sci. Lett. 53 (1981) 11 – 35. E.A. Boyle, L.D. Keigwin, Comparison of Atlantic and Pacific paleochemical records for the last 215,000 years; changes in deep ocean circulation and chemical inventories, Earth Planet. Sci. Lett. 76 (1985) 135 – 150. G.A. Mead, D.A. Hodell, Controls on the 87Sr/86Sr composition of seawater from the middle Eocene to Oligocene; Hole 689B, Maud Rise, Antarctica, Paleoceanography 10 (1995) 327 – 346.

[45] S.C. Cande, D.V. Kent, Revised calibration of the geomagnetic polarity timescale for the Late Cretaceous and Cenozoic, J. Geophys. Res., B, Solid Earth Planets 100 (1995) 6093 – 6095. [46] R.E. Gersonde, L.H. Burckle, Neogene diatom biostratigraphy of ODP Leg, 113, Weddell Sea (Antarctic Ocean), Proc. Ocean Drill. Program Sci. Results 113 (1990) 761 – 789. [47] V. Spiess, Cenozoic magnetostratigraphy of Leg 113 drill sites, Maud Rise, Weddell Sea Antarctica, Proc. Ocean Drill. Program Sci. Results 113 (1990) 261 – 315. [48] H. Chamley, Clay Sedimentology, Springer-Verlag, Berlin, 1989, 623 pp. [49] P.E. Biscaye, Mineralogy and sedimentation of Recent deepsea clay in the Atlantic Ocean and adjacent seas and oceans, Geol. Soc. Am. Bull. 76 (1965) 803 – 831. [50] J.J. Griffin, H. Windom, E.D. Goldberg, The distribution of clay minerals in the world ocean, Deep-Sea Res. Oceanogr. Abstr. 15 (1968) 433 – 459. [51] K.C. Moriarty, Clay minerals in Southeast Indian Ocean sediments, transport mechanisms and depositional environments, Mar. Geol. 25 (1977) 149 – 174. [52] D.A. Darby, Kaolinite and other clay minerals in Arctic Ocean sediments, J. Sediment. Petrol. 45 (1975) 272 – 279. [53] M.J. Hambrey, W.U. Ehrmann, B. Larsen, Cenozoic glacial record of the Prydz Bay continental shelf, East Antarctica, Proc. Ocean Drill. Program Sci. Results 119 (1991) 77 – 132. [54] J.P. Kennett, P.F. Barker, Latest Cretaceous to Cenozoic climate and oceanographic developments in the Weddell Sea, Antarctica; an ocean-drilling perspective, Proc. Ocean Drill. Program Sci. Results 113 (1990) 937 – 960. [55] L. Diester-Haass, C. Robert, H. Chamley, Paleoceanographic and paleoclimatic evolution in the Weddell Sea (Antarctica) during the middle Eocene–late Oligocene, from a coarse sediment fraction and clay mineral data (ODP Site 689), Mar. Geol. 114 (1993) 233 – 250. [56] P.J. Barrett, M.J. Hambrey, D.M. Harwood, A.R. Pyne, P.N. Webb, Synthesis, in: P.J. Barrett (Ed.), Antarctic Cenozoic history from the CIROS-1 drillhole, McMurdo Sound, Department of Scientific and Industrial Research (DSIR), Wellington, New Zealand, 1989, pp. 241 – 251. [57] S.W. Wise Jr, J.R. Breza, D.M. Harwood, W. Wei, Paleogene glacial history of Antarctica, in: D.W. Mueller, J.A. McKenzie, H. Weissert (Eds.), Controversies in modern geology, Acad. Press, London, United Kingdom, 1991, pp. 133 – 171. [58] J.R. Breza, S.W. Wise Jr., Lower Oligocene ice-rafted debris on the Kerguelen Plateau; evidence for East Antarctic continental glaciation, Proc. Ocean Drill. Program Sci. Results 120 (1992) 161 – 178. [59] P. Stille, H. Fischer, Secular variation in the isotopic composition of Nd in Tethys seawater, Geochim. Cosmochim. Acta 54 (1990) 3139 – 3145. [60] G.S. Mountain, K.G. Miller, Seismic and geologic evidence for early Paleogene deepwater circulation in the western North Atlantic, Paleoceanography 7 (1992) 423 – 439. [61] D.K. Pak, K.G. Miller, Paleocene to Eocene benthic foraminiferal isotopes and assemblages; implications for deepwater circulation, Paleoceanography 7 (1992) 405 – 422.

H.D. Scher, E.E. Martin / Earth and Planetary Science Letters 228 (2004) 391–405 [62] J.C. Zachos, D.K. Rea, K. Seto, R. Nomura, N. Niitsuma, Paleogene and Early Neogene deep water paleoceanography of the Indian ocean as determined from benthic foraminifer stable carbon and oxygen isotope records, in: R.A. Duncan, D.K. Rea, R.B. Kidd, U. von Rad, J.K. Weissel (Eds.), The Indian Ocean: A Synthesis of Results from the Ocean Drilling Program, AGU Monograph, vol. 70, 1992, pp. 351 – 386. [63] J.D. Wright, K.G. Miller, Southern Ocean influences on late Eocene to Miocene deepwater circulation, in: J.P. Kennett, D.A. Warnke (Eds.), The Antarctic paleoenvironment; a perspective on global change: Part two, American Geophysical Union, Washington, DC, United States (USA), 1993, pp. 1 – 25. [64] H.D. Scher, E.E. Martin, Eocene To Miocene Southern Ocean deep water circulation revealed from fossil fish teeth Nd isotopes, Eos Trans. AGU 82 (2001) Fall Meet. Suppl., Abstract OS31C-0468. [65] J.C. Zachos, L.D. Stott, K.C. Lohmann, Evolution of early Cenozoic marine temperatures, Paleoceanography 9 (1994) 353 – 387. [66] T.C. Chamberlin, On a possible reversal of deep-sea circulation and its influence on geologic climates, Proc. Am. Philos. Soc. (1906) 33 – 43. [67] N.J. Shackleton, E.B. Kraus, H.P. Hanson, The deep-sea sediment record of climate variability; climatic variability in the oceans; papers from an international symposium held at the Cooperative Institute for Marine and Atmospheric Studies, Prog. Oceanogr. 11 (1982) 199 – 218. [68] M.L. Prentice, R.K. Matthews, Cenozoic ice-volume history; development of a composite oxygen isotope record; with Suppl. Data 88-26, Geology 16 (1988) 963 – 966. [69] J.P. Kennett, L.D. Stott, Abrupt deep-sea warming, palaeoceanographic changes and benthic extinctions at the end of the Palaeocene, Nature 353 (1991) 225 – 229. [70] D.K. Rea, J.C. Zachos, R.M. Owen, P.D. Gingerich, Global change at the Paleocene–Eocene boundary; climatic and evolutionary consequences of tectonic events, Palaeogeogr. Palaeoclimatol. Palaeoecol. 79 (1990) 117 – 128. [71] J.C. Zachos, K.C. Lohmann, J.C.G. Walker, S.W. Wise, Abrupt climate changes and transient climates during the Paleogene; a marine perspective; Centennial special issue, J. Geol. 101 (1993) 191 – 213. [72] G.A. Mead, D.A. Hodell, P.F. Ciesielski, Late Eocene to Oligocene vertical oxygen isotopic gradients in the South Atlantic; implications for warm saline deep water, in: J.P. Kennett, D.A. Warnke (Eds.), The Antarctic paleoenvironment; a perspective on global change: Part two, American Geophysical Union, Washington, DC, 1993, pp. 27 – 48. [73] K.L. Bice, J. Marotzke, Numerical evidence against reversed thermohaline circulation in the warm Paleocene/Eocene ocean, J. Geophys. Res., C, Oceans 106 (2001) 11 – 11542. [74] L. Diester-Haass, Middle Eocene to early Oligocene paleoceanography of the Antarctic Ocean (Maud Rise ODP Leg 113, Site 689): change from a low to a high productivity

[75]

[76] [77]

[78]

[79]

[80]

[81]

[82]

[83]

[84]

[85]

[86]

[87]

[88]

405

ocean., Palaeogeogr. Palaeoclimatol. Palaeoecol. 113 (1995) 311 – 334. J.P. Kennett, Cenozoic evolution of Antarctic glaciation, the Circum-Antarctic Ocean, and their impact on global paleoceanography, J. Geophys. Res. 82 (1977) 3843 – 3860. P.F. Barker, J. Burrell, The opening of Drake Passage, Mar. Geol. 25 (1977) 15 – 34. P.F. Barker, J. Burrell, The influence upon Southern Ocean circulation, sedimentation, and climate of the opening of Drake Passage; Antarctic geoscience, in: C. Craddock (Ed.), Antarctic Geoscience, University of Wisconsin Press, Madison, WI, 1982, pp. 377 – 385. P.F. Barker, Scotia Sea regional tectonic evolution; implications for mantle flow and palaeocirculation, Earth-Sci. Rev. 55 (2001) 1 – 39. L.A. Lawver, L.M. Gahagan, Opening of Drake Passage and its impact on Cenozoic ocean circulation, in: T.J. Crowley, K.C. Burke (Eds.), Tectonic Boundary Conditions for Climate Reconstructions, Oxford University Press, Oxford, 1998, pp. 212 – 223. L.A. Lawver, L.M. Gahagan, Evolution of Cenozoic seaways in the Circum-Antarctic region; Antarctic Cenozoic palaeoenvironments; geologic record and models, Palaeogeogr. Palaeoclimatol. Palaeoecol. 198 (2003) 11 – 37. P.F. Barker, E. Thomas, Origin, signature and palaeoclimatic influence of the Antarctic Circumpolar Current, Earth-Sci. Rev. 66 (2004) 143 – 162. F. Albare`de, S.L. Goldstein, D. Dautel, The neodymium isotopic composition of manganese nodules from the Southern and Indian oceans, the global oceanic neodymium budget, and their bearing on deep ocean circulation, Geochim. Cosmochim. Acta 61 (1997) 1277 – 1291. B. Parsons, J.G. Sclater, An analysis of the variation of ocean floor bathymetry and heat flow with age, J. Geophys. Res. 82 (1977) 803 – 827. B. Diekmann, G. Kuhn, R. Gersonde, A. Mackensen, Middle Eocene to early Miocene environmental changes in the subAntarctic Southern Ocean; evidence from biogenic and terrigenous depositional patterns at ODP Site 1090, Glob. Planet. Change 40 (2004) 295 – 313. M.R. Palmer, H. Elderfield, Rare earth elements and neodymium isotopes in ferromanganese oxide coatings of Cenozoic foraminifera from the Atlantic Ocean, Geochim. Cosmochim. Acta 50 (1986) 409 – 417. K. Miller, G., Middle Eocene to Oligocene stable isotopes, climate, and deep water history: the terminal Eocene event? in: D. Prothero, W.A. Berggren (Eds.), Eocene–Oligocene Climatic and Biotic Evolution, Princeton University, Princeton, NJ, 1992, pp. 160 – 177. R. Davies, J. Cartwright, J. Pike, C. Line, Early Oligocene initiation of North Atlantic Deep Water formation, Nature 410 (2001) 917 – 920. F. Woodruff, S.M. Savin, Miocene deepwater oceanography, Paleoceanography 4 (1989) 87 – 140.