Article | Open | Published:

Major intensification of Atlantic overturning circulation at the onset of Paleogene greenhouse warmth

Nature Communicationsvolume 9, Article number: 4954 (2018) | Download Citation


During the Late Cretaceous and early Cenozoic the Earth experienced prolonged climatic cooling most likely caused by decreasing volcanic activity and atmospheric CO2 levels. However, the causes and mechanisms of subsequent major global warming culminating in the late Paleocene to Eocene greenhouse climate remain enigmatic. We present deep and intermediate water Nd-isotope records from the North and South Atlantic to decipher the control of the opening Atlantic Ocean on ocean circulation and its linkages to the evolution of global climate. The marked convergence of Nd-isotope signatures 59 million years ago indicates a major intensification of deep-water exchange between the North and South Atlantic, which coincided with the turning point of deep-water temperatures towards early Paleogene warming. We propose that this intensification of Atlantic overturning circulation in concert with increased atmospheric CO2 from continental rifting marked a climatic tipping point contributing to a more efficient distribution of heat over the planet.


The Earth underwent long-term climatic cooling between the peak-greenhouse intervals of the mid-Cretaceous and the Eocene1,2,3,4,5. Globally averaged deep-water temperatures gradually declined by almost 10 °C from 72 to 59 Ma, as estimated from benthic foraminiferal oxygen-isotope data3,6. This cooling has been ascribed to decreasing atmospheric CO2 levels7,8,9 through global reduction of volcanism and sea-floor spreading rates10 combined with changes in ocean circulation patterns3. In contrast, there is no comprehensive model explaining how the greenhouse conditions of the Eocene were established and what the roles of atmospheric CO2 and ocean circulation were in promoting global warming. Mechanisms proposed so far have solely focussed on increased atmospheric CO2 levels either induced by carbon cycle changes6, rates of continental rifting11, or by enhanced volcanism of the North Atlantic igneous province12,13. The role of changes in overturning circulation caused by the opening of the Atlantic Ocean and related changes in oceanic heat transport has, however, not been addressed yet.

While circum-equatorial flow, which had dominated circulation in the proto-North Atlantic earlier in the Cretaceous, gradually declined14, the ongoing opening and deepening of the Atlantic basin15,16 led to increased North-South connectivity, although the timing of the establishment of a deep-water connection remains debated17,18,19,20,21,22. Enhanced latitudinal water-mass exchange likely promoted the distribution of heat across the planet via the thermohaline conveyor and resulted in reduced temperature contrasts between the equator and the poles. To distinguish tectonic constraints on circulation from climatically driven changes, the role of subsiding submarine barriers has to be assessed. We determine the timing of the establishment of a persistent deep-water connection between the North and South Atlantic by combining deep-water neodymium (Nd) isotope and temperature records.

Assessing the role of ocean circulation on Earth’s climate in the latest Cretaceous and early Paleogene requires tight constraints on the modes and locations of deep-water formation and the extent of mixing of different deep-water masses. Information on past water mass mixing and exchange can be derived from Nd-isotope signatures (143Nd/144Nd, expressed as εNd(t)) of authigenic, seawater-derived sedimentary archives such as ferromanganese coatings of sediment particles or fish debris, which have been demonstrated to incorporate the Nd-isotope composition of ambient deep waters23. Deep-water masses mainly acquire their Nd-isotope signatures from continental contributions via rivers and dust inputs in their source areas23, as well as through exchange processes with ocean margin sediments24. These characteristic Nd-isotope compositions of deep-water masses are then conservatively advected and mixed over large distances in the open ocean given that the average Nd residence time of 400–2000 years is similar to the global ocean mixing time23,25. Analysis of the Nd-isotope composition of authigenic sedimentary archives thus allows the reconstruction of changes in deep-water mixing over time.

Existing Late Cretaceous Atlantic seawater εNd(t) signatures display a large spread in values (Fig. 1) that led to the suggestion that different mechanisms and locations of deep-water formation operated simultaneously21,26,27,28,29,30. There are indications that intermediate and deep-water exchange commenced as early as at 90 Ma in the Late Turonian20, although the deep Atlantic Ocean potentially operated as a number of sub-basins with limited connectivity until the Maastrichtian22. The large variability in Cretaceous εNd(t) values has so far been interpreted to reflect different modes and locations of deep-water formation either in the southern high latitudes21,22,29,30 or in the North Atlantic26,27,28, or local deep-water formation in relatively shallow sub-basins separated at depth22. The spread in εNd(t) values was likely further enhanced by local boundary exchange and weathering inputs into the relatively small Atlantic basins. Since the Cretaceous Atlantic Ocean was limited in depth and width, its contact area with the margins was large compared to its volume. Regional processes such as boundary exchange thus had a profound effect on water-column chemistry, as is the case for modern near-shore settings24,31 or restricted sub-basins32. The Cretaceous Nd-isotope records from Demerara Rise and the Cape Verde Basin exemplify this effect by local weathering inputs of highly unradiogenic Nd from the old cratons of South America and Africa27,28 (Fig. 1). Despite the potential influence of continental inputs near ocean margins, several open-ocean sites in the Late Cretaceous North and South Atlantic show parallel Nd-isotope trends. These parallel trends have been interpreted to reflect the formation and northward flow of a southern-sourced deep-water mass, “Southern Component Water”30, although the behaviour of individual εNd(t) records is highly variable on a time scale of millions of years and patterns of change are dissimilar between localities.

Fig. 1
Fig. 1

Nd-isotope data across the latest Cretaceous – early Paleogene. In this compilation of published Atlantic εNd(t) records and new Nd-isotope data from Site 369 (this study) only records with four or more data points over the time-interval 72–50 Ma are included. The εNd(t) data of Demerara Rise (Sites 1258, 1260, 1261), Maud Rise, Blake Nose and Site 525 have been smoothed by LOWESS regression (f = 0.3) with the dashed lines representing the 97.5% confidence interval. The εNd(t) data are flanked by the geomagnetic polarity time scale with chrons C31-C24. PETM: Paleocene/Eocene thermal maximum, K/Pg: Cretaceous/Paleogene Boundary

From the Paleocene–Eocene Thermal Maximum (PETM) at 56 Ma onwards, most open-ocean εNd(t) signatures from the North and South Atlantic were within a narrow range of -8 to -1028,33,34, indicating common water masses at bathyal and abyssal depths. There is, however, a lack of data for the Paleocene, which limits our understanding of when and to what extent deep waters exchanged and when the Atlantic started to play a key role in hemispheric oceanic heat exchange. A compilation of existing εNd(t) records for the period of time from 72 to 50 Ma (Fig. 1) shows that Nd-isotope data for the Paleocene are only available from a limited number of sites and, with the exception of Demerara Rise26,28, are of limited resolution (less than one sample per two million years). The Paleocene, however, marks the time when the Atlantic significantly widened and deepened, which potentially paved the way for similar-to-modern ocean overturning processes17. Here we fill this gap and present new Paleocene intermediate- and deep-water Nd-isotope records from the North and South Atlantic Ocean. Five ocean drilling sites were selected from paleo-water depths between 500 and 4500 m (Supplementary Table 1/Fig. 2) to obtain seawater Nd-isotope records covering the critical time span from the end-Cretaceous to the early Paleogene.

Fig. 2
Fig. 2

Paleogeographic setting. a Paleogeographic reconstruction at 60 Ma. Plate tectonic configuration after ref. 63 and global bathymetry from ref. 39. Circles indicate Atlantic Ocean drilling sites for which Nd-isotope data are available (at least four points in the time-interval 72–50 Ma), DR: Demerara Rise. Diamonds indicate locations for which new data are presented here. Numbers indicate DSDP/ODP/IODP Sites. b Three-dimensional plot of the reconstructed bathymetry of the South Atlantic with the Rio Grande Rise – Walvis Ridge barrier at 60 Ma from ref. 40. c Contour plot of the bathymetry in panel B, with black numbers indicating depths in meters


Seawater origin of Nd-isotope signatures

Seawater Nd-isotope signatures were obtained by leaching ferromanganese coatings of bulk sediments that are considered a reliable archive if sufficiently weak leaching procedures are applied35. The εNd(t) variability of the detrital material was also determined for selected samples in this study (details in “methods” section), to evaluate the potential influence of local weathering inputs. The εNd(t) signatures of the detrital fractions and the leached ferromanganese oxide coatings show similar long-term trends at Sites 516 and U1403 and parts of the records at Sites 1267 and 525. Despite following parallel trends, most detrital εNd(t) values are significantly offset from those of the coatings supporting the validity of the seawater εNd(t) signatures extracted from the coatings at the offshore locations of our studied sites as faithful recorders of past water mass mixing (Fig. 3). The Nd-isotope composition of the water-masses themselves may have been influenced to some extent by local factors such as boundary exchange processes that mainly occur when deep-water circulation is slow and/or the sites were located in small or partly isolated basins with high detrital input31,32,36. In addition, the dissolved seawater Nd-isotope signature may have been incorporated into the hydrogenous component of pelagic clays20,37, which may partly explain the similarity in the long-term evolution of the detrital and leached εNd(t) values.

Fig. 3
Fig. 3

New Nd-isotope results. Neodymium isotope values of ferromanganese coatings (coloured symbols) and detrital fractions (open symbols) measured in this study and previously for Site 52522. Error bars indicate 2 s.d. external reproducibility

Parallel trends and convergence of Nd-isotope values

Our new seawater Nd-isotope records from the North and South Atlantic (Fig. 3 and Supplementary Tables 2 to 6) display a wide range of values (−2 to −11) in the Maastrichtian interval (72.1–66 Ma) with parallel trends that converge to a common value of -8 to -9 at 59 Ma (Fig. 4). Our North Atlantic record from Site U1403 ends at 58 Ma, but εNd(t) values between −9.2 and 8 around 57 Ma at northern Site 54938 corroborate our findings (Fig. 1).

Fig. 4
Fig. 4

The convergence of Nd-isotope values. Maastrichtian to early Eocene Nd-isotope records compared to atmospheric CO2 levels and deep-sea temperature evolution. Open symbols depict new εNd(t) data for North Atlantic Site U1403 (Newfoundland margin, dark blue) and South Atlantic Sites 516 (Rio Grande Rise, red), 1267 (northern flank of the Walvis Ridge, green) and 525 (top of the Walvis Ridge, yellow). Filled symbols depict selected high-resolution εNd(t) data from Demerara Rise in the North Atlantic26,28 (purple), Sites 52733 and 126234 on the flank of the Walvis Ridge (green), Site 126334 and Site 52522 on the Walvis Ridge (yellow), and Maud Rise in the Southern Ocean20 (dark blue). The εNd(t) data have been smoothed by LOWESS regression (f = 0.3) with the shaded areas representing the 97.5% confidence interval. The εNd(t) data are flanked by the geomagnetic polarity time scale on the left and on the right by reconstructed atmospheric CO2 levels11, benthic oxygen isotope data from the Pacific41 and the Atlantic13, and the suggested evolution of the Atlantic Ocean. The Paleogene age model of the Sites rests on a combination of magnetic and nannofossil stratigraphy plotted against the GTS 201264. C31-C24: paleomagnetic chrons, δ18Ob: oxygen isotope ratio from benthic foraminifera, VPDB: Vienna Pee Dee Belemnite standard, PETM: Paleocene/Eocene Thermal Maximum, DR: Demerara Rise, K/Pg: Cretaceous/Paleogene boundary, MR: Maud Rise

Sites 525, 1267 and 516 in the South Atlantic, and Site U1403 in the North Atlantic show a trend of decreasing εNd(t) from approximately 70 to 63 Ma, with lowest values reached in the first half of the Paleocene. This decrease may reflect the reduction in active volcanism and exposed volcanic terrains in and around the Atlantic Ocean20. Nd-isotope values at Site 525 were positively offset from εNd(t) signatures at comparably shallow Site 516 on the Rio Grande Rise and nearby deeper Site 1267 at the base of the north-western slope of the Walvis Ridge until the end of the Cretaceous. This positive offset was most likely caused by the weathering influx of volcanic material from the partially subaerially exposed Walvis Ridge in the latest Cretaceous15,39. The offset decreased as the ridge and Site 525 subsided.

From approximately 64 Ma onwards, average εNd(t) values display an increasing trend until 60–59 Ma. We assign this trend to the enhanced volumetric flow of deep and intermediate water masses in the opening South Atlantic Basin which likely led to a decrease of the influence of local inputs and boundary change effects. In addition, the observed trend coincides with a first phase of magmatic activity of the North Atlantic Igneous Province from 62 to 61 Ma13, which may have supplied radiogenic Nd, and ongoing deepening of the study sites that may have reduced unradiogenic weathering inputs from nearby continents.

From 59 Ma onwards, the Nd-isotope signatures at all newly studied sites, as well as Demerara Rise26, decrease together and our εNd(t) results fall within a narrow range of -7 to -9.5 for the period 58.5–56.5 Ma. This convergence may reflect increasing admixture of southern-sourced deep water, which would have carried a εNd(t) signature similar to that at Maud Rise of approximately -9 in the Paleocene (Fig. 4)20.

Opening of the Atlantic Ocean and climatic implications

Recent paleobathymetric reconstructions show that deep oceanic basins in the Atlantic Ocean, like the Cape and the Angola basins, were constricted until the end of the Cretaceous15. Deeper structures, such as the Vema and Hunter channels flanking Rio Grande Rise only allowed intermediate-water exchange at depths shallower than 2500 m. In the Paleocene, the South Atlantic deepened and widened, with the western portion of the Rio Grande Rise having subsided below 2500 m water depth at 60 Ma and the Argentine and Brazil basins reaching depths of over 5500 m in the early Eocene15,40 (Fig. 2). The close correspondence in εNd(t) signatures at 59 Ma suggests a common deep-water signature (εNd(t) −9 to −8) in the South and North Atlantic (Fig. 4). We interpret the converging trend of Nd-isotope signatures to reflect an increasingly efficient deep-ocean circulation in the Atlantic Ocean with the dominant deep-water masses most likely originating in the high southern latitudes. Such a southern origin of deep water is consistent with recent modelling results17 suggesting locations of major deep-water formation in the Southern Ocean, potentially supplemented by a minor source of deep water formed offshore North America. At the same time, our data show that the sub-basins of the deep Atlantic became fully connected by subsidence of the Rio Grande Rise near 59 Ma, accompanied by the widening and deepening of the equatorial gateway17,20. The improved connectivity and the increased volumetric exchange of water masses in the Atlantic Ocean at 59 Ma allowed modern-like open-ocean processes and water-mass mixing to be established, which decreased the sensitivity of the Nd-isotope composition of seawater to local effects such as terrigenous and coastal sedimentary inputs. The convergence of Nd-isotope signatures across the entire Atlantic Ocean spanning paleo-waterdepths of 500 to 4500 m, further suggests that between 62 and 59 Ma, both local tectonic restrictions as well as the vertical stratification of the deep Atlantic Ocean decreased and a global mode of thermohaline circulation was initiated.

The close correspondence in Nd-isotope values among sites at 59 Ma coincided with the onset of the mid-Paleocene global climate warming as evident from benthic foraminiferal oxygen isotopes13,41 (Fig. 4). Based on a recent reconstruction of continental rift length histories11 in comparison to the long-term evolution of atmospheric pCO28, the underlying cause of this warming may lie in the increased cumulative length of incipient continental rifts. Despite a reconstructed gradual increase in pCO2 levels during the end of the Cretaceous and earliest Paleocene8,11 (Fig. 4), as well as an initial magmatic phase of the North Atlantic Igneous Province from 62 to 61 Ma13, the long-term increasing trend in bottom-water temperatures did not start until 59 Ma13,41, when pCO2 started to increase at a higher rate8,11 (Fig. 4).


We hypothesize that the strengthened Atlantic overturning circulation suggested by our data enhanced oceanic poleward heat transport thereby contributing to global climate warming culminating in the peak greenhouse conditions of the Eocene. Global warming may itself have enhanced vertical mixing through increased occurrence of storms and cyclones42 that enabled more efficient overturning circulation in the Atlantic Ocean. Both the deepening of the Rio Grande Rise and enhanced mixing associated with global warming would have increased the capacity of the overturning circulation in the Atlantic Ocean to transport heat. These interpretations of our new Nd-isotope data are consistent with observed changes in Late Cretaceous to early Paleogene Nd-isotope records from the Pacific Ocean42 and Earth system modelling results, which indicate that vigorous ocean circulation and strong vertical mixing resulted in increased oceanic heat transport and reduced equator–pole temperature gradients42,43. Higher oceanic heat transport efficiency likely also set the stage for the occurrence of brief hyperthermals which were frequently superimposed on the overall temperature rise of the Eocene hothouse41. Together with increasing atmospheric CO2 levels8,11, the changing paleogeography of the Atlantic Ocean may have contributed to the boundary conditions that pushed the Earth’s climate into a greenhouse state.


Extraction of Nd isotopes

For Nd-isotope analyses of past seawater extracted from ferromanganese oxide coatings, bulk sediment samples consisting mainly of nannofossil oozes and chalks were dried and homogenised in an agate mortar. To extract the authigenic, seawater-derived Nd-isotope signature, ~2.5 g of powder was treated following the procedure described in ref. 44, omitting the carbonate removal step45. Powdered samples were rinsed three times with de-ionized (MQ) water, after which 10 ml of MQ was added and 10 ml of a 0.05 M hydroxylamine hydrochloride/15% acetic acid solution, buffered with NaOH to a pH of 4. Samples were placed on a shaker for 1 h and centrifuged. The supernatant containing the seawater Nd-isotope signature of the ferromanganese oxide coatings was pipetted off and dried down. For determining the detrital εNd signature, selected samples underwent an additional 12 h leaching step with 20 ml of the hydroxylamine solution (above), after which samples were rinsed with MQ three times and ~50 mg of dried sample was dissolved in a mixture of aqua regia and HF. As preparatory steps for column chemistry, all samples were refluxed in concentrated HNO3 at 80 °C overnight, centrifuged, and 80% of the supernatant was dried down. Twice, 0.5 ml of 1 M HCl was added and the sample was dried down, after which the samples were redissolved in 0.5 ml 1 M HCl. Samples were passed through cation-exchange columns with 0.8 ml AG50W-X12 resin (mesh size 200‒400 μm), using standard procedures, to separate Sr and the Rare Earth Elements (REEs), as well as removing most of the Ba46. A second set of columns with 2 ml Ln-Spec resin (mesh size 50‒100 μm) was used to separate Nd from the other REEs and remaining Ba47.

Analytical procedure

Neodymium isotope ratios were measured on a Nu Instruments Multiple Collector Inductively Coupled Plasma Mass Spectrometer (MC-ICPMS). The majority of samples were measured at GEOMAR Kiel, Germany, and a subset of samples at the department of Earth Sciences of Oxford University, UK (Supplementary Tables 2, 5 and 6). Measured 143Nd/144Nd results were mass-bias corrected to a 146Nd/144Nd ratio of 0.7219 and were normalized to the accepted 143Nd/144Nd value of 0.512115 for the JNdi-1 standard48, which was measured after every third sample.

The results were decay-corrected for the time of deposition by (143Nd/144Nd)sample(t) = (143Nd/144Nd)sample(0) – [(147Sm/144Nd)sample(0) * (℮ʎt – 1)] where t is time, the decay constant ʎ is 6.54 × 10−12, and using an average 147Sm/144Nd ratio of 0.12422. Nd-isotope ratios are reported as εNd(t) values with respect to the Chondritic Uniform Reservoir (CHUR), which are calculated as εNd(t) = [(143Nd/144Nd)sample(t) / (143Nd/144Nd)CHUR(t) − 1] × 104 using a (143Nd/144Nd)CHUR(0) value of 0.512638, and a (147Sm/144Nd)CHUR(0) of 0.196649. External reproducibility (2σ) of the measurements was between 0.15 and 0.54 εNd units and procedural Nd blanks were ≤ 30 pg Nd and thus negligible.

Age models

Age models for the individual sites were generated by an integrated approach of magneto- and biostratigraphy and if available astrochronology. All datum levels are tied to the Geological Timescale GTS2012. Ages of polarity chrons are from ref. 50, and of calcareous nannofossils (NP zonation) from ref. 51 as compiled in ref. 52. In detail the following data are used and summarized in Supplementary Table 7. Tie points for Site 516 are defined by magneto- and calcareous nannofossil stratigraphy given in ref. 53. Tie points for Site 525 are defined by polarity chrons in the Maastrichtian54 and by calcareous nannofossils in the Paleocene55. Tie points for Site 1267 are derived from precession cycle counting for the upper Paleocene (until 58.2 Ma ago)56 and polarity chrons for the lower to middle Paleocene and the Maastrichtian57. Ages of neodymium isotope data from Site 126234 and Site 52733 were converted to GTS 2012. Tie points for Site 369 follow the age model of the Shipboard Scientific Party58. Tie points for Site U1403 are defined by first occurrences (FO) of calcareous nannofossils for the Paleocene59 with an adjustment for the FO of Lithoptychius spp. at 227 m depth rCCSF (corresponding to the first radiation of fasciculithids according to refs. 60,]61) and by astronomical tuning of 405 kyr cycles and carbon isotope stratigraphy62.

Data availability

The authors declare that all the data generated during this study are available within the manuscript and its supplementary information file.

Additional information

Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.


  1. 1.

    Clarke, L. J. & Jenkyns, H. C. New oxygen isotope evidence for long-term Cretaceous climatic change in the Southern Hemisphere. Geology 27, 699–702 (1999).

  2. 2.

    Cramer, B. S., Toggweiler, J. R., Wright, J. D., Katz, M. E. & Miller, K. G. Ocean overturning since the Late Cretaceous: Inferences from a new benthic foraminiferal isotope compilation. Paleoceanography 24, PA4216 (2009).

  3. 3.

    Friedrich, O., Norris, R. D. & Erbacher, J. Evolution of middle to Late Cretaceous oceans—A 55 m.y. record of Earth’s temperature and carbon cycle. Geology 40, 107–110 (2012).

  4. 4.

    Huber, B. T., MacLeod, K. G., Watkins, D. K. & Coffin, M. F. The rise and fall of the Cretaceous hot greenhouse climate. Glob. Planet. Change 167, 1–23 (2018).

  5. 5.

    Huber, B. T., Hodell, D. A. & Hamilton, C. P. Middle–Late Cretaceous climate of the southern high latitudes: Stable isotopic evidence for minimal equator-to-pole thermal gradients. Geol. Soc. Am. Bull. 107, 1164–1191 (1995).

  6. 6.

    Zachos, J. C., Dickens, G. R. & Zeebe, R. E. An early Cenozoic perspective on greenhouse warming and carbon-cycle dynamics. Nature 451, 279–283 (2008).

  7. 7.

    Anagnostou, E. et al. Changing atmospheric CO2concentration was the primary driver of early Cenozoic climate. Nature 533, 380–384 (2016).

  8. 8.

    Foster, G. L., Royer, D. L. & Lunt, D. J. Future climate forcing potentially without precedent in the last 420 million years. Nat. Commun. 8, 14845 (2017).

  9. 9.

    Pucéat, E. et al.Fish tooth δ18O revising Late Cretaceous meridional upper ocean water temperature gradients. Geology 35, 107–110 (2007).

  10. 10.

    Larson, R. L. Latest pulse of Earth: evidence for a mid-Cretaceous superplume. Geology 19, 547–550 (1991).

  11. 11.

    Brune, S., Williams, S. E. & Müller, R. D. Potential links between continental rifting, CO2 degassing and climate change through time. Nat. Geosci. 10, 941–946 (2017).

  12. 12.

    Gutjahr, M. et al. Very large release of mostly volcanic carbon during the Palaeocene–Eocene thermal maximum. Nature 548, 573–577 (2017).

  13. 13.

    Westerhold, T., Röhl, U., Donner, B., McCarren, H. K. & Zachos, J. C. A complete high-resolution Paleocene benthic stable isotope record for the central Pacific (ODP Site 1209). Paleoceanography 26, PA2216 (2011).

  14. 14.

    Pucéat, E., Lecuyer, C. & Reisberg, L. Neodymium isotope evolution of NW Tethyan upper ocean waters throughout the Cretaceous. Earth Planet. Sci. Lett. 236, 705–720 (2005).

  15. 15.

    Pérez-Díaz, L. & Eagles, G. South Atlantic paleobathymetry since early Cretaceous. Sci. Rep. 7, 11819 (2017).

  16. 16.

    Sewall, J. O. et al. Climate model boundary conditions for four Cretaceous time slices. Clim. Past. 3, 647–657 (2007).

  17. 17.

    Donnadieu, Y., Pucéat, E., Moiroud, M., Guillocheau, F. & Deconinck, J.-F. A better-ventilated ocean triggered by Late Cretaceous changes in continental configuration. Nat. Commun. 7, 10316 (2016).

  18. 18.

    Frank, T. D. & Arthur, M. A. Tectonic forcings of Maastrichtian ocean‐climate evolution. Paleoceanography 14, 103–117 (1999).

  19. 19.

    Friedrich, O. & Erbacher, J. Benthic foraminiferal assemblages from Demerara Rise (ODP Leg 207, western tropical Atlantic): possible evidence for a progressive opening of the Equatorial Atlantic Gateway. Cretac. Res. 27, 377–397 (2006).

  20. 20.

    Moiroud, M. et al. Evolution of neodymium isotopic signature of seawater during the Late Cretaceous: implications for intermediate and deep circulation. Gondwana Res. 36, 503–522 (2015).

  21. 21.

    Robinson, S. A., Murphy, D. P., Vance, D. & Thomas, D. J. Formation of “Southern Component Water” in the Late Cretaceous: evidence from Nd-isotopes. Geology 38, 871–874 (2010).

  22. 22.

    Voigt, S. et al.Tectonically restricted deep-ocean circulation at the end of the Cretaceous greenhouse. Earth Planet. Sci. Lett. 369–370, 169–177 (2013).

  23. 23.

    Frank, M. Radiogenic isotopes: tracers of past ocean circulation and erosional input. Rev. Geophys. 40, (2002).

  24. 24.

    Lacan, F. & Jeandel, C. Neodymium isotopes as a new tool for quantifying exchange fluxes at the continent–ocean interface. Earth Planet. Sci. Lett. 232, 245–257 (2005).

  25. 25.

    Rempfer, J., Stocker, T. F., Joos, F., Dutay, J.-C. & Siddall, M. Modelling Nd-isotopes with a coarse resolution ocean circulation model: Sensitivities to model parameters and source/sink distributions. Geochim. Cosmochim. Acta 75, 5927–5950 (2011).

  26. 26.

    MacLeod, K. G., Isaza Londoño, C., Martin, E. E., Jiménez Berrocoso, Á. & Basak, C. Changes in North Atlantic circulation at the end of the Cretaceous greenhouse interval. Nat. Geosci. 4, 779–782 (2011).

  27. 27.

    MacLeod, K. G., Martin, E. E. & Blair, S. W. Nd isotopic excursion across Cretaceous ocean anoxic event 2 (Cenomanian-Turonian) in the tropical North Atlantic. Geology 36, 811–814 (2008).

  28. 28.

    Martin, E. E., MacLeod, K. G., Jiménez Berrocoso, A. & Bourbon, E. Water mass circulation on Demerara Rise during the Late Cretaceous based on Nd isotopes. Earth Planet. Sci. Lett. 327328, 111–120 (2012).

  29. 29.

    Murphy, D. P. & Thomas, D. J. The evolution of Late Cretaceous deep-ocean circulation in the Atlantic basins: Neodymium isotope evidence from South Atlantic drill sites for tectonic controls. Geochem. Geophys. Geosystems 14, 5323–5340 (2013).

  30. 30.

    Robinson, S. A. & Vance, D. Widespread and synchronous change in deep-ocean circulation in the North and South Atlantic during the Late Cretaceous. Paleoceanography 27, PA1102 (2012).

  31. 31.

    Grasse, P. et al. Short-term variability of dissolved rare earth elements and neodymium isotopes in the entire water column of the Panama Basin. Earth Planet. Sci. Lett. 475, 242–253 (2017).

  32. 32.

    Osborne, A. H., Haley, B. A., Hathorne, E. C., Flögel, S. & Frank, M. Neodymium isotopes and concentrations in Caribbean seawater: Tracing water mass mixing and continental input in a semi-enclosed ocean basin. Earth Planet. Sci. Lett. 406, 174–186 (2014).

  33. 33.

    Abbott, A. N., Haley, B. A., Tripati, A. K. & Frank, M. Constraints on ocean circulation at the Paleocene–Eocene Thermal Maximum from neodymium isotopes. Clim. Past Discuss. 11, 2557–2583 (2015).

  34. 34.

    Via, R. K. & Thomas, D. J. Evolution of Atlantic thermohaline circulation: Early Oligocene onset of deep-water production in the North Atlantic. Geology 34, 441–444 (2006).

  35. 35.

    Blaser, P. et al. Extracting foraminiferal seawater Nd isotope signatures from bulk deep sea sediment by chemical leaching. Chem. Geol. 439, 189–204 (2016).

  36. 36.

    Yu, Z. et al. Seasonal variations in dissolved neodymium isotope composition in the Bay of Bengal. Earth Planet. Sci. Lett. 479, 310–321 (2017).

  37. 37.

    Goldstein, S. L. & O’Nions, R. K. Nd and Sr isotopic relationships in pelagic clays and ferromanganese deposits. Nature 292, 324–327 (1981).

  38. 38.

    Thomas, D. J., Bralower, T. J. & Jones, C. E. Neodymium isotopic reconstruction of late Paleocene–early Eocene thermohaline circulation. Earth Planet. Sci. Lett. 209, 309–322 (2003).

  39. 39.

    Müller, R. D., Sdrolias, M., Gaina, C., Steinberger, B. & Heine, C. Long-term sea-level fluctuations driven by ocean basin dynamics. Science 319, 1357–1362 (2008).

  40. 40.

    Pérez-Díaz, L. & Eagles, G. Estimating palaeobathymetry with quantified uncertainties: a workflow illustrated with South Atlantic data. Geol. Soc. Lond. Spec. Publ. SP476, 1 (2018).

  41. 41.

    Littler, K., Röhl, U., Westerhold, T. & Zachos, J. C. A high-resolution benthic stable-isotope record for the South Atlantic: Implications for orbital-scale changes in Late Paleocene–Early Eocene climate and carbon cycling. Earth Planet. Sci. Lett. 401, 18–30 (2014).

  42. 42.

    Thomas, D. J., Korty, R., Huber, M., Schubert, J. A. & Haines, B. Nd isotopic structure of the Pacific Ocean 70-30 Ma and numerical evidence for vigorous ocean circulation and ocean heat transport in a greenhouse world. Paleoceanography 29, 454–469 (2014).

  43. 43.

    Sijp, W. P. & England, M. H. The effect of low ancient greenhouse climate temperature gradients on the ocean’s overturning circulation. Clim. Past. 12, 543–552 (2016).

  44. 44.

    Gutjahr, M. et al. Reliable extraction of a deepwater trace metal isotope signal from Fe–Mn oxyhydroxide coatings of marine sediments. Chem. Geol. 242, 351–370 (2007).

  45. 45.

    Stumpf, R., Frank, M., Schönfeld, J. & Haley, B. A. Late Quaternary variability of Mediterranean outflow water from radiogenic Nd and Pb isotopes. Quat. Sci. Rev. 29, 2462–2472 (2010).

  46. 46.

    Barrat, J. et al. Determination of rare earth elements in sixteen silicate reference samples by ICP_MS after Tm addition and ion exchange separation. Geostand. Newsl. 20, 133–139 (1996).

  47. 47.

    Le Fèvre, B. & Pin, C. A straightforward separation scheme for concomitant Lu–Hf and Sm–Nd isotope ratio and isotope dilution analysis. Anal. Chim. Acta 543, 209–221 (2005).

  48. 48.

    Tanaka, T. et al. JNdi-1: a neodymium isotopic reference in consistency with LaJolla neodymium. Chem. Geol. 168, 279–281 (2000).

  49. 49.

    Jacobsen, S. B. & Wasserburg, G. J. Sm-Nd isotopic evolution of chondrites. Earth Planet. Sci. Lett. 50, 139–155 (1980).

  50. 50.

    Ogg, J. G. Geomagnetic Polarity Time Scale. in: The Geological Time Scale Vol. 1 (eds Gradstein, F. M., Ogg, J. G., Schmitz, M. D. & Ogg, G.) 85–113 (Elsevier, Amsterdam, 2012).

  51. 51.

    Berggren, W. A., Kent, D. V., Swisher, C. C. & Aubry, M.-P. A Cenozoic geochronology and chronostratigraphy. Soc. Sed. Geol. Spec. Publ. 54, 129–212 (1995).

  52. 52.

    Anthonissen, D. E. & Ogg, J. G. Cenozoic and Cretaceous biochronology of planktonic foraminifera and calcareous nannofossils. in: The Geological Time Scale Vol. 2 (eds Gradstein, F. M., Ogg, J. G., Schmitz, M. D. & Ogg, G.) 1083–1027 (Elsevier, Amsterdam, 2012).

  53. 53.

    Berggren, W. A. et al. Magnetobiostratigraphy of Deep Sea Drilling Project leg 72, Sites 515-518, Rio Grande Rise (South Atlantic). In: Barker, P.F. et al., Init. Rep. Deep Sea Drill. Proj. 72, 938–948 (1983).

  54. 54.

    Chave, A. D. Lower Paleocene-Upper Cretaceous magnetostratigraphy, Sites 525, 527, 528, and 529, Deep Sea Drilling Project Leg 74. In: Moore, T.C. et al. Init. Rep. Deep Sea Drill. Proj. 74, 525–531 (1984).

  55. 55.

    Manivit, H. Paleogene and Upper Cretaceous calcareous nannofossils from Deep Sea Drilling Project Leg 74. In: Moore, T.C. et al. Init. Rep. Deep Sea Drill. Proj. 74, 475–499 (1984).

  56. 56.

    Westerhold, T. et al. On the duration of magnetochrons C24r and C25n and the timing of early Eocene global warming events: Implications from the Ocean Drilling Program Leg 208 Walvis Ridge depth transect. Paleoceanography 22, PA2201 (2007).

  57. 57.

    Zachos, J. C. et al. In: Leg 208 summary. Proc. Ocean Drill. Prog Init. Rep. 208, 1–112 (2004).

  58. 58.

    Lancelot, Y. Seibold, E., et al. Volume 41. Init. Rep. Deep Sea Drill. Proj. 41, (U.S. Government Printing Office, Washington, 1977).

  59. 59.

    Norris, R. D., Wilson, P. A., Blum, P. & Expedition 342 Scientists. Site 1403. In: Paleogene Newfoundland Sediment Drifts and MDHDS Test. Proc. Integrat. Ocean Drill. Prog. 342, (2014).

  60. 60.

    Monechi, S., Reale, V., Bernaola, G. & Balestra, B. The Danian/Selandian boundary at Site 1262 (South Atlantic) and in the Tethyan region: biomagnetostratigraphy, evolutionary trends in fasciculiths and environmental effects of the Latest Danian Event. Mar. Micro. 98, 28–40 (2013).

  61. 61.

    Dinarès-Turell, J., Westerhold, T., Pujalte, V., Röhl, U. & Kroon, D. Astronomical calibration of the Danian stage (Early Paleocene) revisited: Settling chronologies of sedimentary records across the Atlantic and Pacific Oceans. Earth Planet. Sci. Lett. 405, 119–131 (2014).

  62. 62.

    Batenburg, S. J. et al. Late Maastrichtian carbon isotope stratigraphy and cyclostratigraphy of the Newfoundland Margin (Site U1403, IODP Leg 342). Newsl. Stratigr. 51, 245–260 (2018).

  63. 63.

    Hay, W. W. et al. Alternative global Cretaceous paleogeography. In: Special Paper 332: Evolution of the Cretaceous Ocean-Climate System 1–47 (Geological Society of America, Boulder, Colorado, 1999).

  64. 64.

    Gradstein, F., Ogg, J., Schmitz, M. & Ogg, G. (Eds) The Geologic Time Scale 2012 (Elsevier, Amsterdam, 2012).

Download references


We would like to thank the International Ocean Discovery Program (IODP) for providing samples. Authors were funded by the German Research Foundation (DFG) under grant numbers DFG VO 687/14, FR2544/8, and FR1198/11, BO2505/8 and EH89/20. We would like to thank Jutta Heinze and Chris Siebert at GEOMAR, Kiel and Alan Hsieh at Oxford University, UK for smooth operation of the laboratory and the mass spectrometers.

Author information


  1. Institut für Geowissenschaften, Goethe-Universität Frankfurt, Altenhöferallee 1, Frankfurt am Main, 60438, Germany

    • S. J. Batenburg
    • , S. Voigt
    •  & T. Klein
  2. Department of Earth Sciences, University of Oxford, South Parks Road, Oxford, OX1 3AN, UK

    • S. J. Batenburg
  3. Institut für Geowissenschaften, Ruprecht-Karls-Universität Heidelberg, Im Neuenheimer Feld 234-236, 69120, Heidelberg, Germany

    • O. Friedrich
  4. GEOMAR Helmholtz-Zentrum für Ozeanforschung Kiel, Wischhofstr. 1-3, Kiel, 24148, Germany

    • A. H. Osborne
    •  & M. Frank
  5. Bundesanstalt für Geowissenschaften und Rohstoffe, Stilleweg 2, 30655, Hannover, Germany

    • A. Bornemann
  6. Department of Earth Sciences, Royal Holloway, University of London, Egham, TW20 0EX, UK

    • L. Pérez-Díaz


  1. Search for S. J. Batenburg in:

  2. Search for S. Voigt in:

  3. Search for O. Friedrich in:

  4. Search for A. H. Osborne in:

  5. Search for A. Bornemann in:

  6. Search for T. Klein in:

  7. Search for L. Pérez-Díaz in:

  8. Search for M. Frank in:


S.J.B., S.V., O.F. and M.F. developed the project. S.J.B., A.H.O. and T.K. performed Nd-isotope analyses. L.P-D. provided bathymetric information. All authors contributed to the writing of the manuscript.

Competing interests

The authors declare no competing interests.

Corresponding author

Correspondence to S. J. Batenburg.

Electronic supplementary material

About this article

Publication history






By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.