Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Atlantic Deep-water Response to the Early Pliocene Shoaling of the Central American Seaway


The early Pliocene shoaling of the Central American Seaway (CAS), ~4.7–4.2 million years ago (mega annum-Ma), is thought to have strengthened Atlantic Meridional Overturning Circulation (AMOC). The associated increase in northward flux of heat and moisture may have significantly influenced the evolution of Pliocene climate. While some evidence for the predicted increase in North Atlantic Deep Water (NADW) formation exists in the Caribbean and Western Atlantic, similar evidence is missing in the wider Atlantic. Here, we present stable carbon (δ13C) and oxygen (δ18O) isotope records from the Southeast Atlantic-a key region for monitoring the southern extent of NADW. Using these data, together with other δ13C and δ18O records from the Atlantic, we assess the impact of the early Pliocene CAS shoaling phase on deep-water circulation. We find that NADW formation was vigorous prior to 4.7 Ma and showed limited subsequent change. Hence, the overall structure of the deep Atlantic was largely unaffected by the early Pliocene CAS shoaling, corroborating other evidence that indicates larger changes in NADW resulted from earlier and deeper shoaling phases. This finding implies that the early Pliocene shoaling of the CAS had no profound impact on the evolution of climate.


Changes in major ocean-gateways during the late Neogene are considered to have had far-reaching impacts on the evolution of climate, mediated through large-scale changes in ocean circulation. Notably, it has been proposed that the progressive closure of the CAS, which occurred during the Pliocene1, was a necessary precondition for establishing enhanced (modern-like) Atlantic Meridional Overturning Circulation (AMOC) conditions2,3. The so-called “Panama Closure Hypothesis” posits that as the CAS closed, diminishing upper-water exchange between the Pacific Ocean and Caribbean Sea led to enhanced flow of warm and salty surface water into the North Atlantic via the Gulf Stream2,3. The additional salt carried northward would have increased density in the surface waters of deep-water formation regions, such as the Nordic and Labrador Seas. In turn, this is expected to have led to stronger NADW formation, driving enhanced AMOC along with a marked reorganization of the deep-water flow in the Atlantic Ocean3,4. Such closure induced changes have been implicated in enhancing ice sheet growth and promoting the onset of Northern Hemisphere Glaciation through an increase in the North Atlantic moisture budget5,6,7, as well as in triggering an overall shoaling of the tropical thermocline and promoting the development of the Eastern Equatorial Pacific cold tongue8. In contrast, it has also been argued that the resulting increase in northward heat flux hindered ice sheet growth and delayed Northern Hemisphere Glaciation9,10.

Elucidating the timing and nature of the CAS closure is crucial for evaluating its impact on AMOC and climate. However, geological records reveal it as a complex process, with many details remaining unclear and contested11,12,13,14,15,16. The timeline of closure events is estimated to range from the emergence of land in the CAS as early as ~15–20 Ma17,18, to closure of the deep-water connection by ~7–10 Ma19,20 and complete restriction of Pacific-Caribbean surface-water exchange between ~3.5–2.5 Ma12. In addition, while there is strong general agreement between modeling studies that AMOC strengthens in response to CAS closure, there is a wide range in the magnitude of AMOC response, especially for the final few hundred meters of shoaling (for an overview, see20 and references therein). Hence, both the anticipated timing and the magnitude of CAS induced changes in AMOC are uncertain.

Nevertheless, arguably the clearest and best-constrained evidence for the final stages of CAS closure, particularly with regards to its impact on regional surface-water oceanography, is in the development of a permanent inter-basin surface-water δ18O-gradient between the Pacific and Caribbean Sea (Fig. 1). An initial δ18O-gradient occurred as early as ~4.7 Ma, followed by a strengthening between 4.4–4.2 Ma, implying restricted upper-water exchange across the CAS2,21,22. In line with the δ18O data, Mg/Ca analyses on planktonic foraminifera indicate a build-up of heat and salinity in the Caribbean8,22,23. Haug and Tiedemann3 suggested that this interval reflects a major phase in the final closure of the seaway, with shoaling to within 100 m water depth by 4.2 Ma. More recently, however, Mestas-Nuñez and Molnar13 have suggested that these changes in the Caribbean may be unrelated to CAS shoaling: Instead, they propose it may be a consequence of a shifting dominance of El Niño- to La Niña-like conditions in the eastern tropical Pacific. Regardless of the cause, the build-up of heat and salinity in Caribbean surface waters during the period ~4.7–4.2 Ma is uncontested and, as a consequence, the expectation of an increase in AMOC remains. Hence, in this study, the early Pliocene shoaling of the CAS refers to these well-documented changes in Caribbean surface-water oceanography.

Figure 1

Surface water (planktic) δ18O records (‰ vs Vienna Pee Dee Belemnite) from the Caribbean (Site 999)21 and Pacific (Site 849)62 showing the emergence of a permanent gradient between ~4.7–4.2 Ma, interpreted to reflect a major shoaling phase of the CAS3.

Bold lines represent 50 kyr running averages.

Based on the original interpretation of surface water changes in the Caribbean, many studies have marked the early Pliocene as a key interval in the CAS closure history and linked it to proxy records that indicate an enhancement in NADW3,4,5,8,19,21,24,25. However, the evidence for NADW changes during the early Pliocene is mainly limited to the Caribbean Sea and deep equatorial Western Atlantic (Fig. 2). As both bathymetric constraints from major ocean ridges and differences between water mass advection within the Deep Western Boundary Current (DWBC) and interior pathways are markedly influential in setting up gradients and pathways within the Atlantic26,27,28 (Fig. 3), it is straightforward to envisage scenarios whereby regional changes in water mass prevalence are not representative of large-scale changes in the wider deep Atlantic. Indeed, important regional differences in Pleistocene deep-water circulation have been shown previously29,30,31,32,33, reinforcing the need to consider previously documented changes in a wider spatial context.

Figure 2

The main evidence for NADW changes in response to the early Pliocene shoaling of the CAS.

(a) δ13C (‰ vs Vienna Pee Dee Belemnite) time series from Caribbean Site 999 and North Atlantic Site 982. The increasing δ13C values at Site 999 are interpreted to reflect a progressive increase in AMOC3; (b) δ18O timeseries (‰ vs Vienna Pee Dee Belemnite) from western equatorial Atlantic Sites 925 and 929. The reversal of δ18O gradients after ~4.7 Ma are interpreted to reflect an increase in temperature and salinity due to enhanced northward heat and salt transport25; (c) % sand content timeseries for Sites 999 and 9293. The increases in average and variability of % sand are interpreted to reflect enhanced carbonate preservation in response to an increase in dominance of low-corrosive NADW at both sites3. Bold lines represent 50 kyr running averages. References for isotopic data are given in Table 1.

Figure 3

Dissolved oxygen concentration maps of the Atlantic at different depths (modified from the WOCE Atlantic Ocean Atlas 26 to show features of deep-water circulation and the locations of all sites referred to in the text), reflecting aspects of the chemical structure of the deep-Atlantic.

Arrows depict general NADW circulation patterns and blue ovals indicate areas of NADW formation. Sites are displayed on maps according to their relevant depth (Table 1). Sites with isotopic data presented in Figs 4, 5, 6 plotted with a red circle, while other sites mentioned in the text are plotted with a red square. Site 999 is situated at 2828 m in the Caribbean, but is expected to record deep-waters entering across the Atlantic-Caribbean sills at ~1600–1900 m depth. High oxygen concentrations result from recent and prolonged contact with the atmosphere, while low oxygen concentrations result from microbial respiration of organic matter over time. NADW, with high oxygen concentrations, is a well “ventilated” water mass, in contrast with deep-waters entering the Atlantic from the south. These maps outline the modern dominance of NADW and highlight the importance of bathymetric constraints and deep-water pathways (e.g. the Deep Western Boundary Current versus interior pathways) in setting gradients in water-mass properties within the Atlantic.

In light of these considerations, we find that evidence for large-scale changes in AMOC during the early Pliocene is hitherto unclear. In particular, changes in the southern extent of NADW—an important indicator of AMOC—are so far not well sampled. In the Atlantic sector of the Southern Ocean, low resolution and poorly time-constrained benthic stable isotope records exist for Ocean Drilling Program (ODP) Sites 108834,35 (2080 m depth, 41°S, 14°E) and 70436 (2530 m depth, 47°S, 7°E), but these lie within a region of complex ocean mixing, with time-varying contributions from Circumpolar Deep Water and Antarctic Intermediate Water. Indeed, questions have arisen over the origin of the signal recorded in this region during the Pliocene37,38, with a recent proposition that, during the mid-Pliocene, it may be mostly responding to changes in ventilation from the south and so does not reflect changes in AMOC39.

In order to better assess early Pliocene changes in AMOC, we present benthic stable oxygen and carbon isotope data from ODP Site 1264 (2505 m depth, 29°S, 3°E), recently published by Bell et al.33, situated on the northern flank of Walvis Ridge in the Angola Basin of the Southeast Atlantic. Walvis Ridge forms an almost impassible barrier in most locations below ~3500 m and restricts water flow up to depths of ~2500 m, and, together with the Mid-Atlantic Ridge, acts to isolate the Angola Basin from deep-waters entering from the south40,41. As a result of this bathymetry, together with its position along a major export pathway of NADW41, below ~1500 m NADW dominates all depths in the Angola Basin. Furthermore, at ~18°S to the south of Walvis Ridge, Site 704 (also at ~2500 m depth) provides an important complementary “downstream” record for Site 1264 and allows an assessment of the relative strength of NADW presence in the Southeast Atlantic. Hence, Site 1264 is uniquely positioned to monitor fluctuations in NADW.

Using these new δ18O and δ13C data, together with published data from a number of other sites in the wider Atlantic (Fig. 3, Table 1), we assess the impact of the early Pliocene (4.7–4.2 Ma) shoaling phase of the CAS on AMOC.

Table 1 All site locations, water depth, average age resolution for the period 5.3–3.6 Ma and sources of isotopic data used in this study.

Benthic foraminiferal δ13C records are a well-established tool for reconstructing circulation changes in the deep-ocean42,43, reflecting ambient nutrient concentration and the air-sea gas exchange history of source waters44. Meanwhile, δ18O records serve as a useful additional water-mass tracer32,33,45 as they document changes in the temperature and δ18O of ambient seawater, the latter of which co-varies with salinity changes. Hence, both δ13C and δ18O measurements are suitable tracers of chemical and physical changes in water-mass properties. An increase in NADW production is expected to increase δ13C values and/or reduce horizontal and vertical δ13C-gradients in the deep (>1 km depth) Atlantice.g.46,47 as NADW has a higher δ13C signature than deep-waters originating from the south48. Concomitant changes in δ18O values are less straightforward to predict: it is possible that an increase in AMOC will be accompanied by an increase in δ18O, reflecting higher salinities and a low-latitude δ18Oseawater signature33, but this may be offset by the effects of warmer waters on δ18O during the equilibrium precipitation of calcite and/or local meltwater enriched in 16O from the north.


Timeseries of benthic δ13C and δ18O data from all Atlantic sites are presented in Figs 4 and 5, respectively, while site locations and data references are given in Table 1. The benthic δ18O stack of Lisiecki and Raymo49 (hereafter LR04) is also shown in Fig. 5 next to each record so that stratigraphic alignment can be seen. In order to directly assess the impact of the early Pliocene shoaling phase, we also compare average Atlantic deep-water conditions for time slices on either side of the period 4.7–4.2 Ma (Fig. 6).

Figure 4

Time series of δ13C data (‰ vs Vienna Pee Dee Belemnite) from all sites shown in Fig. 1.

Data has been resampled at common 10 kyr time steps (linear integration). Horizontal black lines show average δ13C values for the time slice intervals 5.0–4.7 Ma (Pre CAS shoaling) and 4.2–3.6 Ma (Post CAS shoaling). Arrows indicate approximate modern δ13C values for each site location48. Note that data for Site 607 has been resampled at 10 kyr intervals, as is available from32, while data from Site 704 prior to 4.2 Ma is displayed as symbols in order to highlight the low resolution and stratigraphic uncertainty. Data references are given in Table 1.

Figure 5

Time series δ18O data (‰ vs Vienna Pee Dee Belemnite) from all sites shown in Fig. 1 (thick colored lines).

Data has been resampled at common 10 kyr time steps (linear integration). Horizontal thick black lines show average δ18O values for the time slice intervals 5.0–4.7 Ma (Pre CAS shoaling) and 4.2–3.6 Ma (Post CAS shoaling). The LR04 stack38 is shown along side each record (thin gray lines), on the same scale, to show stratigraphic alignment (see Methods). Note that data from Site 1264 has been resampled at 3 kyr intervals so that stratigraphic alignment can be clearly seen, while Site 704 prior to 4.2 Ma is displayed as symbols in order to highlight the low resolution and stratigraphic uncertainty. References for isotopic data are given in Table 1.

Figure 6

Average δ13C and δ18O values (both plotted in ‰ vs Vienna Pee Dee Belemnite) for the time slice intervals 5.0–4.7 Ma (Pre CAS shoaling) and 4.2–3.6 Ma (Post CAS shoaling) plotted as anomalies with respect to changes at Pacific Site 849 (see text).

Square symbols indicate sites recording ocean water masses at intermediate depths (1000–2000 m) while diamond symbols indicate sites recording water masses in the deep ocean (>2000 m). Error bars are 2σ combined error on the difference, Due to poor time resolution for Site 704 and lack of original time series data for Site 607, no error bars are shown for these sites. Note that errors serve only as an indication of the confidence in mean values calculated. Data references are given in Table 1.

Early Pliocene Atlantic δ13C records generally record high values, averaging between 0.5–1.0‰, directly comparable to values in the modern ocean (Fig. 4) and show little long-term trend. Interestingly, Site 1264 often records higher δ13C values than today, similar to modern core NADW values (~1‰). On the other hand, Sites 704 and 999 record low average values (~0–0.5‰) in early Pliocene, with Site 999 displaying a long-term increase from values averaging ~0.2‰ at the start of the Pliocene, rising to an average of ~0.65‰ by 3.7–3.6 Ma.

Atlantic δ18O records show a spread in average values similar in range to δ13C records. Sites 1264 and 704 have consistently higher δ18O values than the LR04 stack, while Sites 929 and 982 have consistently lower values. Long-term trends are generally small, with the notable exception of Site 925, which shows a large increase in δ18O at ~4.7 Ma.

Figure 6 shows average paired δ18O-δ13C anomalies for each site calculated by subtracting the mean ocean δ18O-δ13C changes, as recorded by ODP Site 849 in the deep Pacific37, for the intervals 5.0–4.7 Ma (Pre CAS shoaling) and 4.2–3.6 Ma (Post CAS shoaling). After mean ocean changes are removed, a significant increase (p = < 0.05 for student t-tests) in δ13C is only seen at Site 999 (Δδ13C = +0.18‰). δ13C changes in the remaining sites are small and not significant. Meanwhile, a significant and prominent shift in δ18O of +0.28‰ is observed for Site 925.


The δ13C increase at Site 999, together with a concomitant increase in carbonate preservation3 (Fig. 2), is consistent with a southward shift in the boundary of Atlantic intermediate waters due to enhanced upper NADW formation in the region of the Labrador Sea, coeval with the progressive CAS shoaling phase. This interpretation is supported by modeling predictions50 and by the early Pliocene onset of drift deposits in sediments from the Labrador Sea (Site 646), indicating enhanced bottom water currents51. Rather than implying an overall restructuring of the deep and intermediate water circulation, however, the lack of substantial change elsewhere in the Atlantic Ocean suggests relatively local changes in water mass prevalence are being recorded in the Caribbean.

Data from Site 1264 provide a reference for Site 704, forming a latitudinal transect (~28–47°S) of the Southeast Atlantic at ~2500 m depth. Since this region and depth is situated along a major export pathway for NADW41 and encompasses the modern boundary between NADW and Southern Source Water (SSW) (as reflected in Fig. 3), variations in Site 1264–704 δ13C-gradients gradients are expected to be sensitive to latitudinal movements in the extent of NADW and hence reflect the strength of NADW formation. Throughout the interval studied and with no apparent change in response to CAS shoaling, Site 1264–704 δ13C-gradients varied in the range of ~0.2–0.6‰, although sparse data from Site 704 indicate short periods of lower gradients. This is similar to the modern δ13C-gradient for this region, which is spatially variable and in the range of ~0.2–0.4‰48. δ13C–gradients between Site 1264 and sites in the North Atlantic are also similar to the modern situation, with the exception of Site 982–1264 gradients, which are even lower. Therefore, these observations provide a convincing basis for inferring that NADW formation during the early Pliocene was generally strong, with periods possibly stronger than today and was seemingly unaffected by the CAS shoaling phase.

Previous studies have utilized δ13C data from Site 704 in order to reconstruct the relative strength of AMOC for intervals that encompass the early Pliocene35,52,53. Based on comparisons with available North Atlantic and Pacific records and assuming a northern signal for δ13C changes, these studies speculated that NADW formation was already strong during the late Miocene, from ~6–7 Ma onwards and that the early Pliocene progressive shoaling of the CAS (~4.7–4.2 Ma) had little impact. Similar conclusions were made with data from Site 108834, which is at a slightly shallower depth in the same region (2080 m depth, 41°S, 14°E). Importantly, data from Site 126433, which is uniquely located to record northern sourced deep-waters, substantiate these interpretations by demonstrating a positive N-S δ13C-gradient within the South Atlantic; a finding which is independent of remaining uncertainties in Southern Ocean ventilation e.g.39.

Thus, given the confirmation of a northern δ13C-signal penetrating into the South Atlantic, we are able to quantitatively estimate the relative contribution of NADW versus SSW to waters bathing Site 1264. For a given site (x), this may be calculated using the equation of Oppo and Fairbanks54:

The δ13CSSW is the end-member carbon isotopic value for SSW, δ13CNADW is the end-member carbon isotopic value for NADW and δ13Cx is the carbon isotopic value at site x in the Atlantic. For Site 1264, the calculation yields values of ~70%, both before and after the CAS shoaling phase (4.7–4.2 Ma), using NADW and SSW end-member values of Sites 607 and 704, respectively. By comparison, based on phosphate measurements26, this calculation indicates values of ~60–70% – depending on end-member values used – for the modern situation at Site 1264. This indicates that early Pliocene NADW formation was similar to the present, despite an open CAS. It should be noted, however, that this calculation relies on several assumptions, including the mixing of only two deep-water masses between all three sites and that the end-member values of these water masses are properly captured by Sites 607 and 704.

The δ18O-gradient between Sites 925 and 607 (Fig. 6) records changes in mid-depth (3000–3500 m) waters between the western equatorial Atlantic and Northwest Atlantic, respectively (Fig. 3). Prior to the early Pliocene CAS shoaling phase, the large (~0.3‰) δ18O-gradient between these two sites points to differences in temperature and δ18Oseawater of the deep-waters flooding both locations, suggesting separate source regions in the North Atlantic (Fig. 6). Maintaining such a large δ18O-gradient at similar depths within the same basin would require a deep-water flow regime with little mixing between the core of the Deep Western Boundary Current (DWBC) (recorded by Site 92525) and the basin interior (recorded by Site 60735). Such a scenario may have existed due to weaker flow speeds, which promote more stable, laminar boundary flow28. Following from this, the absence of high-δ13C NADW with a low δ18O-signature being recorded elsewhere in the Atlantic could be explained if interior pathways were more important than the DWBC for deep-water export in the North Atlantic, as may be the case in the modern ocean27. Alternatively, it is also possible that the δ18O-shift at Site 925 is an artifact of measurement offsets between laboratories, which can be as large as 0.3‰55. However, the reversal of the Site 929–925 δ18O-gradient should remain a robust result, as both sets of data were measured at the same laboratory25.

Billups et al.25 recognized the early Pliocene δ18O increase at Site 925 and the resulting reversed δ18O-gradient with underlying waters at Site 929 (Fig. 2). Based on the requirement of a stable vertical density structure, they concluded that Site 925 deep-waters must have become warmer and saltier between ~4.2–3.7 Ma. Figures 4 and 5 indicate that deep-water conditions at Site 925 were unusual prior to 4.7 Ma and that, subsequently, the prevalence of deep-waters sourcing other mid-depth (2500–3500 m) sites (i.e. Sites 607, 659 & 1264) increased. The heavier δ18O-signature of deep-waters recorded at Site 925 may also reflect a strengthened low-latitude surface-water signal33, in line with inferences of an increase in temperature and salinity.

While stable isotope evidence suggests that deep-water change in response to the progressive CAS shoaling phase (4.7–4.2 Ma) was restricted to a strengthening of upper NADW formation and an altered flow regime of the DWBC, sedimentary evidence of increased carbonate preservation suggests a further increase in the prevalence of low-corrosive NADW. An increase in the sedimentary carbonate fraction from ~0% to ~40% is seen during the early Pliocene in the eastern North Atlantic (Site 66556, 4755 m). Given the highly nonlinear relationship between %CaCO3 and dissolution, however, this may only reflect small changes in preservation. Meanwhile, in the western equatorial Atlantic (Site 929, 4350 m), an abrupt increase in % sand fraction variability is seen after ~4.6 Ma3 (Fig. 2), likely relating to phases of enhanced carbonate preservation24. While this is not reflected in average values (Fig. 7), it is consistent with an increase in peak δ13C values at Site 92925,57. These observations may be explained by a deepening of the lower boundary of NADW, such that preservation conditions at Site 929 became sensitive to variations in AMOC after CAS shoaling. In conjunction with evidence for enhanced deep-water formation in the Labrador Sea, this is remarkably consistent with modeling predictions by Mikolajewicz & Crowley50.

Figure 7

A schematic of Atlantic deep-water changes in response to the early Pliocene CAS shoaling phase (~4.7–4.2 Ma).

The dashed line in b) shows the boundary between Southern Sourced Water (SSW) and North Atlantic Deep Water (NADW) prior to CAS shoaling, for reference. The red arrows on the map sections indicate the northward surface flow of warm water, which was possibly slightly enhanced after CAS shoaling. Green circles indicate areas of deep-water formation in the North Atlantic. Maps generated using Adobe Illustrator.


Figure 7 provides an illustrative summary of deep-water changes inferred from the stable isotope and carbonate preservation evidence described above. Prior to the early Pliocene CAS shoaling phase (Fig. 7a), we find strong evidence for a well-ventilated deep Atlantic as far a ~28oS, suggesting that NADW formation was active despite the open gateway. Low δ13C-gradients between various sites across the Atlantic indicate that deep-water circulation was rigorous, comparable to and frequently stronger than today. In particular, our new data confirm the dominance of NADW in the Southeast Atlantic. However, upper-NADW formation in the Labrador Sea was reduced and, taking changes in δ18O data at face value, deep-waters within the DWBC were distinct from those filling the wider Atlantic, implying limited mixing with deep-waters in the basin interior, possibly due to low flow speeds.

The early Pliocene progressive CAS shoaling phase had a limited effect on the overall geometry of Atlantic deep-waters (Fig. 7b). This is indicated by the negligible changes to the isotopic signature of deep-waters at most Atlantic sites. Instead, changes were restricted to a few key regions: Upper-NADW formation in the Labrador Sea began a long-term increase3, while deep-waters sourcing the DWBC became more similar to those filling the basin interior, possibly reflecting an increase in temperature and salinity25 in response to enhancements in the Gulf Stream. Meanwhile, the depth of the lower boundary of NADW deepened slightly, increasing carbonate preservation in the deep North Atlantic (>4000 m).

We conclude that the lack of change in large parts of the Atlantic Ocean demonstrates that the early Pliocene shoaling phase of the CAS had little impact on AMOC and, therefore, climate. Our data corroborate data and modeling studies that propose significant NADW flow existed with an open CASe.g.34,35,46,50,52,58 and suggest that large scale changes in Atlantic deep-water circulation, if any, would have occurred in response to earlier and deeper shoaling phases of the CAS, possibly during the late Miocene19.


10 cm3 samples were taken from Site 1264 cores at 1.5 cm intervals, using the shipboard splice to create a meters composite depth (mcd) scale. For the large majority of Site 1264 samples (30.01–57.45 mcd = 3.715–5.300 Ma), processing and measurements were carried out at the University of Florida. Benthic species, Cibicidoides wuellerstorfi, were picked from the >150 μm size fraction. The foraminifera were then cleaned in an ultrasonic bath to remove fine-grained particles and soaked in 15% H2O2 to remove surface organic contaminants prior to analysis. The number of specimens of C. wuellerstorfi varied from 1 to 4 and foraminiferal tests were crushed before analysis. The foraminiferal calcite was reacted in 70 oC orthophosphoric acid using a Kiel III carbonate preparation device. Evolved CO2 gas was measured online with a Finnigan-MAT 252 mass spectrometer.

For samples between 27.80–30.01 mcd (=3.600–3.715 Ma), samples were processed and measured at Vrije Universiteit, Amsterdam. Measurements were performed on 1–3 C. wuellerstorfi, selected from the >200 μm size fraction, using a Finnigan 251 Gas Source mass spectrometer equipped with Kiel type automated carbonate extraction line.

All stable isotope results are reported relative to Vienna Pee Dee Belemnite and calibrated using in house standards that are correlated to the international standard, NBS19. Analytical precision for both δ18O and δ13C is better than ±0.1‰. For all sites assessed, δ18O values are adjusted by +0.64‰ to account for disequilibrium with surrounding seawater59.

Systematic inter-laboratory offsets are a possible artifact in δ18O records, generated by pooling data from different labs. Indeed, our experience from a small number of duplicate measurements taken between different laboratories indicates potential offsets of up to 0.1–0.2‰, in line with previous findings of up to 0.3‰55. However, given practical limitations on the number of duplicate measurements possible, including limited availability of C. wuellerstorfi specimens, such offsets are difficult to correct for. Hence, we are aware of potential offsets of <0.3‰.

All age models used in this study were derived from the alignment of δ18O time series to the global benthic δ18O stack of Lisiecki and Raymo49 (the LR04 stack), as described in Bell et al.33 and shown in Fig. 5. The low resolution and short hiatuses present in early Pliocene data from Site 704, however, prevented reliable alignment, so the original age model36 was used. Averages calculated from these data may thus be affected by age model uncertainties and missing information.

Additional Information

How to cite this article: Bell, D. B. et al. Atlantic Deep-water Response to the Early Pliocene Shoaling of the Central American Seaway. Sci. Rep. 5, 12252; doi: 10.1038/srep12252 (2015).


  1. Schmidt, D. in Deep. Perspect. Clim. Chang. Marrying Signal from Comput. Model. Biol. Proxies ( Williams, M., Haywood, A. M., Gregory, F. J. & Schmidt, D. N. ) 429–444 (The Micropalaeontological Society, 2007). doi: 10.1144/TMS2.21.

  2. Keigwin, L. D. Isotopic Paleoceanography of the Caribbean and East Pacific: Role of Panama Uplift in Late Neogene Time. Science (80-. ). 217, 350–353 (1982).

    ADS  CAS  Google Scholar 

  3. Haug, G. H. & Tiedemann, R. Effect of the formation of the Isthmus of Panama on Atlantic Ocean circulation. Nature 393, 673–767 (1998).

    ADS  CAS  Google Scholar 

  4. Burton, K. W., Ling, H.-F. & O’Nions, K. Closure of the Central American Isthmus and its effect on deep-water formation in the North Atlantic. Nature 386, 382–385 (1997).

    ADS  CAS  Google Scholar 

  5. Driscoll, N. W. & Haug, G. H. A Short Circuit in Thermohaline Circulation: A Cause for Northern Hemisphere Glaciation? Science (80-. ). 282, 436–438 (1998).

    ADS  CAS  Google Scholar 

  6. Bartoli, G. et al. Final closure of Panama and the onset of northern hemisphere glaciation. Earth Planet. Sci. Lett. 237, 33–44 (2005).

    ADS  CAS  Google Scholar 

  7. Lunt, D. J., Valdes, P. J., Haywood, A. & Rutt, I. C. Closure of the Panama Seaway during the Pliocene: implications for climate and Northern Hemisphere glaciation. Clim. Dyn. 30, 1–18 (2007).

    Google Scholar 

  8. Steph, S. et al. Early Pliocene increase in thermohaline overturning: A precondition for the development of the modern equatorial Pacific cold tongue. Paleoceanography 25, (2010).

  9. Berger, W. H. & Wefer, G. in South Atl. Present Past Circ. ( Wefer, G. ) 363–410 (Springer, New York, 1996).

  10. Klocker, A., Prange, M. & Schulz, M. Testing the influence of the Central American Seaway on orbitally forced Northern Hemisphere glaciation. Geophys. Res. Lett. 32, 1–4 (2005).

    Google Scholar 

  11. Stone, R. Battle for the Americas. Science 341, 230–3 (2013).

    ADS  CAS  PubMed  Google Scholar 

  12. Molnar, P. Closing of the Central American Seaway and the Ice Age: A critical review. Paleoceanography 23, 1–15 (2008).

    Google Scholar 

  13. Mestas‐Nuñez, A. & Molnar, P. A mechanism for freshening the Caribbean Sea in pre‐Ice Age time. Paleoceanography 508–517 (2014). doi: 10.1002/2013PA002515.Received

  14. Kirby, M. X., Jones, D. S. & MacFadden, B. J. Lower Miocene stratigraphy along the Panama Canal and its bearing on the Central American Peninsula. PLoS One 3, e2791 (2008).

    ADS  PubMed  PubMed Central  Google Scholar 

  15. Coates, A. G. et al. Geological Society of America Bulletin Closure of the Isthmus of Panama: The near-shore marine record of Costa Rica and western Panama. (1992). doi: 10.1130/0016-7606(1992)104&lt;0814

  16. Coates, A. G., Collins, L. S., Aubry, M.-P. & Berggren, W. A. The Geology of the Darien, Panama and the late Miocene-Pliocene collision of the Panama arc with northwestern South America. Geol. Soc. Am. Bull. 116, 1327 (2004).

    ADS  Google Scholar 

  17. Montes, C. et al. Arc-continent collision and orocline formation: Closing of the Central American seaway. J. Geophys. Res. 117, 1–25 (2012).

    Google Scholar 

  18. Montes, C. et al. Evidence for middle Eocene and younger land emergence in central Panama: Implications for Isthmus closure. Geol. Soc. Am. Bull. 124, 780–799 (2012).

    ADS  CAS  Google Scholar 

  19. Osborne, A. et al. The seawater neodymium and lead isotope record of the final stages of Central American Seaway closure. Paleoceanography 29, 715–729 (2014).

    ADS  Google Scholar 

  20. Sepulchre, P. et al. Consequences of shoaling of the Central American Seaway determined from modeling Nd isotopes. Paleoceanography (2013). doi: 10.1002/2013PA002501.Received.

  21. Haug, G., Tiedemann, R., Zahn, R. & Ravelo, A. Role of Panama uplift on oceanic freshwater balance. Geology 29, 207–210 (2001).

    ADS  CAS  Google Scholar 

  22. Steph, S. et al. Changes in Caribbean surface hydrography during the Pliocene shoaling of the Central American Seaway. Paleoceanography 21, 1–25 (2006).

    Google Scholar 

  23. Groeneveld, J. et al. Foraminiferal Mg/Ca increase in the Caribbean during the Pliocene: Western Atlantic Warm Pool formation, salinity influence, or diagenetic overprint? Geochemistry Geophys. Geosystems 9, 1–21 (2008).

    Google Scholar 

  24. Frenz, M., Henrich, R. & Zychla, B. Carbonate preservation patterns at the Ceará Rise – Evidence for the Pliocene super conveyor. Mar. Geol. 232, 173–180 (2006).

    ADS  Google Scholar 

  25. Billups, K., Ravelo, A. & Zachos, J. C. Early Pliocene deep water circulation in the western equatorial Atlantic: Implications for high-latitude climate change. Paleoceanography 13, 84–95 (1998).

    ADS  Google Scholar 

  26. Koltermann, K. P., Gouretski, V. V. & Jancke, K. Hydrographic Atlas of the World Ocean Circulation Experiment (WOCE). Volume 3, Atlantic Ocean (eds. M. Sparrow, P. Chapman & J. Gould ). (International WOCE project office, Southampton, UK, ISBN 090417557X 2011).

  27. Gary, S. F., Susan Lozier, M., Böning, C. W. & Biastoch, A. Deciphering the pathways for the deep limb of the Meridional Overturning Circulation. Deep Sea Res. Part II Top. Stud. Oceanogr. 58, 1781–1797 (2011).

    ADS  Google Scholar 

  28. Dengler, M. et al. Break-up of the Atlantic deep western boundary current into eddies at 8 degrees S. Nature 432, 1018–20 (2004).

    ADS  CAS  PubMed  Google Scholar 

  29. Bickert, T., Curry, W. & Wefer, G. Late Pliocene to Holocene (2.6-0 Ma) western equatorial Atlantic deep-water circulation: Inferences from benthic stable isotopes. Proc. Ocean Drill. Program, Sci. Results 154 (1997).

  30. Venz, K. A. & Hodell, D. A. New evidence for changes in Plio-Pleistocene deep water circulation from Southern Ocean ODP Leg 177 Site 1090. Paleoceanography 182, 197–220 (2002).

    Google Scholar 

  31. Raymo, M. E., Oppo, D. W. & Curry, W. The mid-Pleistocene climate transition: A deep sea carbon isotopic perspective. Paleoceanography 12, 546–559 (1997).

    ADS  Google Scholar 

  32. Raymo, M. E. et al. Stability of North Atlantic water masses in face of pronounced climate variability during the Pleistocene. Paleoceanography 19, 1–13 (2004).

    Google Scholar 

  33. Bell, D. B., Jung, S. J. A., Kroon, D., Lourens, L. J. & Hodell, D. A. Local and regional trends in Plio-Pleistocene δ18O records from benthic foraminifera. Geochemistry, Geophys. Geosystems (2014). doi: 10.1002/2014GC005297

  34. Billups, K. Late Miocene through early Pliocene deep water circulation and climate change viewed from the sub-Antarctic South Atlantic. Palaeogeogr. Palaeoclimatol. Palaeoecol. 185, 287–307 (2002).

    Google Scholar 

  35. Hodell, D. A. & Venz-Curtis, K. A. Late Neogene history of deepwater ventilation in the Southern Ocean. Geochemistry Geophys. Geosystems 7, 1–16 (2006).

    Google Scholar 

  36. Hodell, D. A. & Venz, K. A. Toward a high-resolution stable isotopic record of the Southern Ocean during the Pliocene-Pleistocene (4.8 to 0.8 Ma). Antarct. Res. Ser. 56, 265–310 (1992).

    Google Scholar 

  37. Mix, A. C. et al. 17. Benthic foraminifer stable isotope record from Site 849 (0-5 Ma): local and global climate changes. Proc. Ocean Drill. Program, Sci. Results 138, 371–412 (1995).

    Google Scholar 

  38. Andersson, C., Warnke, D. A., Channell, J. E. T., Stoner, J. & Jansen, E. The mid-Pliocene (4.3-2.6 Ma) benthic stable isotope record of the Southern Ocean: ODP Sites 1092 and 704, Meteor Rise. Palaeogeogr. Palaeoclimatol. Palaeoecol. 182, 165–181 (2002).

    Google Scholar 

  39. Zhang, Z.-S. et al. Mid-pliocene Atlantic Meridional Overturning Circulation not unlike modern. Clim. Past 9, 1495–1504 (2013).

    Google Scholar 

  40. Connary, S. D. & Ewing, M. Penetration of Antarctic bottom water from the Cape Basin into the Angola Basin. J. Geophys. Res. 79, 463–469 (1974).

    ADS  Google Scholar 

  41. Arhan, M., Mercier, H. & Park, Y.-H. On the deep water circulation of the eastern South Atlantic Ocean. Deep Sea Res. Part I Oceanogr. Res. Pap. 50, 889–916 (2003).

    ADS  Google Scholar 

  42. Curry, W. B., Duplessy, J. C., Labeyrie, L. D. & Shackleton, N. J. Changes in the distribution of δ13C of deep water ΣCO2 between the last glaciation and the Holocene. Paleoceanography 3, 317–341 (1988).

    ADS  Google Scholar 

  43. Sarnthein, M. et al. Changes in east Atlantic deepwater circulation over the last 30,000 years: Eight time slice reconstructions. Paleoceanography 9, 209–267 (1994).

    ADS  Google Scholar 

  44. Lynch-Stieglitz, J., Stocker, T. F., Broecker, W. S. & Fairbanks, R. G. The influence of air-sea exchange on the isotopic composition of oceanic carbon: observations and modeling. Paleoceanography 9, 653–665 (1995).

    CAS  Google Scholar 

  45. Curry, W. & Oppo, D. Glacial water mass geometry and the distribution of δ13C of ΣCO 2 in the western Atlantic Ocean. Paleoceanography 20, 1–13 (2005).

    Google Scholar 

  46. Schneider, B. & Schmittner, A. Simulating the impact of the Panamanian seaway closure on ocean circulation, marine productivity and nutrient cycling. Earth Planet. Sci. Lett. 246, 367–380 (2006).

    ADS  CAS  Google Scholar 

  47. Butzin, M., Lohmann, G. & Bickert, T. Miocene ocean circulation inferred from marine carbon cycle modeling combined with benthic isotope records. Paleoceanography 26, PA1203doi: 10.1029/2009PA001901 (2011).

    Article  ADS  Google Scholar 

  48. Kroopnick, P. The distribution of 13C in Atlantic Ocean. Earth Planet. Sci. Lett. 49, (1980).

  49. Lisiecki, L. E. & Raymo, M. E. A Pliocene-Pleistocene stack of 57 globally distributed benthic δ18O records. Paleoceanography 20, 1–17 (2005).

    Google Scholar 

  50. Mikolajewicz, U. & Crowley, T. J. Response of a coupled ocean/energy balance model to restricted flow through the Central American Isthmus. Paleoceanography 12, 429 (1997).

    ADS  Google Scholar 

  51. Kaminski, M. A., Gradstein, F. M., Scott, D. B. & Mackinnon, K. D. 37. Neogene benthic foraminifera biostratigraphy and deep-water history of Sites 645, 646 and 647, Baffin Bay and Labrador Sea. Proc. Ocean Drill. Program, Sci. Results 105, 731–756 (1989).

    Google Scholar 

  52. Poore, H. R., Samworth, R., White, N. J., Jones, S. M. & McCave, I. N. Neogene overflow of Northern Component Water at the Greenland-Scotland Ridge. Geochemistry Geophys. Geosystems 7, Q06010doi: 10.1029/2005GC001085 (2006).

    Article  ADS  CAS  Google Scholar 

  53. Ravelo, A. C. & Andreasen, D. H. Enhanced circulation during a warm period. Geophys. Res. Lett. 27, 1001–1004 (2000).

    ADS  Google Scholar 

  54. Oppo, D. W. & Fairbanks, R. G. Variability in the deep and intermediate water circulation of the Atlantic Ocean during the past 25,000 years. Northern Hemisphere modulation of the Southern Ocean. 86, 1–15 (1987).

    CAS  Google Scholar 

  55. Ostermann, D. R. & Curry, W. B. Calibration of stable isotopic data: An enriched δ180 standard used for source gas mixing detection and correction. Paleoceanography 15, 353–360 (2000).

    ADS  Google Scholar 

  56. Ruddiman, W. F., Sarnthein, M. & Baldauf, J. et al. Chapter 10. Site 665. Proc. Ocean Drill. Program, Initial Reports 108, 741–789 (1988).

    Google Scholar 

  57. Billups, K., Ravelo, A. C. & Zachos, J. C. 21. Early Pliocene deep-water circulation: stable isotope evidence for enhanced Northern Component Deep Water. Proc. Ocean Drill. Program, Sci. Results 154, 319–330 (1997).

    CAS  Google Scholar 

  58. Nisancioglu, K. H., Raymo, M. E. & Stone, P. H. Reorganization of Miocene deep water circulation in response to the shoaling of the Central American Seaway. Paleoceanography 18, 1–12 (2003).

    Google Scholar 

  59. Shackleton, N. J. & Hall, M. A. Chapter 16. Oxygen and carbon isotope stratigraphy of Deep Sea Drilling Project hole 552a: Plio-Pleistocene glacial history. Initial Reports Deep Sea Drill. Proj. 81, 599–609 (1984).

    CAS  Google Scholar 

  60. Tiedemann, R., Sarnthein, M. & Shackleton, N. J. Astronomical timescale for the Pliocene Atlantic d18O and dust flux records of Ocean Drilling Program site 659. Paleoceanography 9, 619–638 (1994).

    ADS  Google Scholar 

  61. Tiedemann, R. & Franz, S. O. 20. Deep-water circulation, chemistry and terrigenous sediment supply in the equatorial Atlantic during the Pliocene, 3.3-2.6 Ma and 5-4.5 Ma. Proc. Ocean Drill. Program, Sci. Results 154, 299–318 (1997).

    CAS  Google Scholar 

  62. Cannariato, K. G. & Ravelo, A. C. Pliocene-Pleistocene evolution of eastern tropical surface water circulation and thermocline depth and hemisphere gradient. Paleoceanography 12, 805–820 (1997).

    ADS  Google Scholar 

Download references


This work was supported by a NERC Ph.D. studentship to D.Bell and NERC grant NE/G01602X/1. Colin Chilcott is thanked for assistance with the mass spectrometer.

Author information




D.B., S.J. and D.K. wrote the main text, while D.H., L.L. and M.R. contributed to study design and data interpretation. D.B. prepared all figures. All authors reviewed the manuscript.

Ethics declarations

Competing interests

The authors declare no competing financial interests.

Rights and permissions

This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Bell, D., Jung, S., Kroon, D. et al. Atlantic Deep-water Response to the Early Pliocene Shoaling of the Central American Seaway. Sci Rep 5, 12252 (2015).

Download citation

Further reading


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.


Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing