GLACIAL CYCLES

The residence time of SouthernOcean surface waters and the100,000-year ice age cycleAdam P. Hasenfratz1,2*, Samuel L. Jaccard2*, Alfredo Martínez-García3,Daniel M. Sigman4, David A. Hodell5, Derek Vance6, Stefano M. Bernasconi1,Helga (Kikki) F. Kleiven7, F. Alexander Haumann8,9, Gerald H. Haug1,3

From 1.25 million to 700,000 years ago, the ice age cycle deepened and lengthenedfrom 41,000- to 100,000-year periodicity, a transition that remains unexplained.Using surface- and bottom-dwelling foraminifera from the Antarctic Zone of theSouthern Ocean to reconstruct the deep-to-surface supply of water during the iceages of the past 1.5 million years, we found that a reduction in deep watersupply and a concomitant freshening of the surface ocean coincided with theemergence of the high-amplitude 100,000-year glacial cycle. We propose that thisslowing of deep-to-surface circulation (i.e., a longer residence time for Antarcticsurface waters) prolonged ice ages by allowing the Antarctic halocline to strengthen,which increased the resistance of the Antarctic upper water column to orbitallypaced drivers of carbon dioxide release.

The mid-Pleistocene transition (MPT) oc-curred in the absence of any discerniblechanges in the orbital parameters that con-trol the seasonality and meridional dis-tribution of incoming solar radiation (1),

driving decades of research into the possiblecontributors to the change (2–7). Evidence isgrowing that the Southern Ocean was criticalto the global cooling of the late Pleistocene iceages, enhancing ocean carbon storage, lower-ing atmospheric CO2 (pCO2) concentrations, andthus weakening the global greenhouse effect(8–10). The available data are consistent with anincrease in the amplitude of glacial/interglacialchanges in pCO2 since the MPT, largely througha deepening of the CO2 minima reached duringice ages (10–12). Thus, the evolution of glacial/interglacial cycles since 1.25 million years agomay have reflected the progress of glacial-stage conditions of the Southern Ocean towardthose observed in the glacial maxima of thelate Pleistocene.In the modern Southern Ocean, nutrient-

and CO2-rich deep waters ascend to the sur-face, where the scarcity of light and iron im-

pedes the consumption of major nutrientsby phytoplankton, allowing for the evasion ofpreviously sequestered carbon. In the AntarcticZone (AZ) south of the Antarctic Polar Front,the evasion of CO2 could have been inhibitedduring ice ages by increased sea ice cover (13)or by a reduction in the exchange of waterbetween the Antarctic surface and the deepocean (14–16). The relative importance ofthese two alternative causes of CO2 declinehave not been resolved for the late Pleistoceneice ages, let alone the MPT. Data on the evo-lution of AZ conditions across the MPT mayoffer new insight.In this study, through measurements on

the sediments from Ocean Drilling Project(ODP) Site 1094 (53.2°S, 05.1°E, 2807 m waterdepth) (fig. S1), we sought to determine thehistory of the relationship between surfaceand deep waters across the ice ages of theMPT. ODP 1094 is currently bathed by LowerCircumpolar Deep Water (LCDW). Above, Up-per Circumpolar Deep Water (UCDW) upwellsto the surface, and vertical mixing betweenUCDW and surface waters also occurs. How-ever, the surface waters are lower in salinity,and the resulting gradient in salinity (thehalocline), and thus density, moderates ver-tical exchange (17). Biogenic opal and bariumaccumulation rates indicate that export pro-duction at the site reached minima in eachof the glacial stages back to pre-MPT (9, 10),indicating that ODP 1094 remained withinthe AZ throughout the ice ages of the past1.5 million years. Consistent with this no-tion, sea surface temperatures (SSTs) recon-structed for the glacials throughout the past1.5 million years were rarely >1.5°C higherthan that of the Holocene and the other in-

terglacials (see below), when the AntarcticPolar Front was north of the study site—a con-clusion further supported by diatom assemb-lages (18).In benthic and planktonic foraminiferal

tests from ODP 1094, we have measured thecalcite d18Ο (d18ΟC) and the Mg/Ca ratio, thelatter serving as a proxy for seawater temper-ature (5, 19). Seawater d18Ο (d18ΟW) was de-termined by removing the temperature effecton d18ΟC using Mg/Ca-derived temperatureestimates (5). Although d18Ο is a traditionaltool for reconstructing regional temperatureand salinity as well as global ice volume, wefocus here on its utility as a passive tracer inthe upper water column of the AZ in orderto infer circulation rates. The AZ upper wa-ter column d18ΟW is lowered by the excess ofprecipitation relative to evaporation, withan additional reduction due to melting ofAntarctic land ice, ice shelves, and/or ice-bergs. The amplitude of this d18Ο loweringdepends not only on the rates of these pro-cesses but also on the residence time of wa-ter in the upper water column of the AZ.Enhanced input of deep water into the AZsurface, in the absence of other changes, wouldlower the residence time and thus raise thed18ΟW of the AZ surface, which would bereflected in the d18Ο of planktonic fora-minifera. Conversely, reduced input of deepwaters would increase the residence time,allowing the AZ surface and deep d18ΟW todiverge. Following this logic, in the absenceof major changes in precipitation or glacialmelt rate, coupled planktonic and benthic d18ΟC

measurements in the AZ, corrected to yieldchanges in d18ΟW, can record past changesin the residence time of water in the AZ surfaceand thus the rate of deep-to-surface circula-tion in the region. As a result of the multi-yearresidence time of water in the surface AZ andthe lack of major effects from the formationand melting of sea ice (20), modern d18ΟW

varies little across the open AZ [by <0.3 per mil(‰)] (21, 22). This mitigates concerns regardingthe use of a single site to infer changes in theAZ as a whole.The planktonic [Neogloboquadrina pachyderma

(sinistral)] and benthic (Cibicidoides spp.,Melonispompilioides) d18ΟC records are highly corre-lated (Fig. 1A), which is consistent with a sub-stantial contribution from whole-ocean d18ΟW

change and/or the possibility that deep watersare responsive to climate evolution at high south-ern latitudes (5). To constrain the temperaturecomponent in the foraminiferal d18ΟC records,we systematically corrected the Mg/Ca recordsof both N. pachyderma (s.) and M. pompilioidesfor Mn contamination (23) and converted theseto temperatures (figs. S2 to S4). [Cibicidoideswas not used for Mg/Ca paleothermometrybecause of a carbonate saturation effect onthis taxon (24).] Our new bottom-water records,although discontinuous in the late Pleisto-cene (due to a scarcity of shells), provide aconstraint on the Pleistocene evolution of

RESEARCH

Hasenfratz et al., Science 363, 1080–1084 (2019) 8 March 2019 1 of 5

1Geological Institute, Department of Earth Sciences, ETHZürich, Zürich, Switzerland. 2Institute of GeologicalSciences and Oeschger Centre for Climate ChangeResearch, University of Bern, Bern, Switzerland. 3MaxPlanck Institute for Chemistry, Mainz, Germany.4Department of Geosciences, Guyot Hall, PrincetonUniversity, Princeton, NJ, USA. 5Department of EarthSciences, University of Cambridge, Cambridge, UK.6Institute of Geochemistry and Petrology, ETH Zürich,Zürich, Switzerland. 7Department of Earth Science,University of Bergen and Bjerknes Centre for ClimateResearch, Bergen, Norway. 8British Antarctic Survey,Cambridge, UK. 9Atmospheric and Oceanic Sciences,Princeton University, Princeton, NJ, USA.*Corresponding author. Email: adam.hasenfratz@geo.unibe.ch(A.P.H.); samuel.jaccard@geo.unibe.ch (S.L.J.)

on October 17, 2020

http://science.sciencem

ag.org/D

ownloaded from

Hasenfratz et al., Science 363, 1080–1084 (2019) 8 March 2019 2 of 5

Fig. 1. Antarctic Southern Ocean surface andbottom water records. Surface- and bottom-waterdata are shown in red and black, respectively.(A) Foraminiferal d18OC of planktonic N. pachyderma(s.) [(45) and this study] and of benthic Cibicidoidesspp. and M. pompilioides (composite record) fromODP Site 1094. (B) ODP 1094 SST derived fromN. pachyderma (s.) Mg/Ca compared to Antarctictemperature (25) (blue). (C) ODP 1094 BWTderivedfrom M. pompilioides Mg/Ca compared to ODP Site1123 (5). (D) d18OW of surface and bottom watersat ODP Site 1094. The uncertainty envelopes includethe uncertainties from Mn correction and calibration(1s SD). The stars represent the modern summertemperature (46) and the d18OW at ODP 1094(47, 48) (fig. S10) and ODP 1123 (21) for thesurface and deep water, respectively. VPDB, ViennaPeeDee belemnite; SMOW, standard meanocean water.

Fig. 2. Glacial Antarctic Southern Oceansurface and bottom water records from ODPSite 1094. (A) Planktonic and benthic foraminiferald18O. (B) Surface- and bottom-water temperature.(C) Surface and bottom water d18O. Bottom-waterd18O for the LGM (barren of benthic foraminifera)was estimated using the d18O of Cibicidoides spp.and assuming BWT of –1.5° ± 0.5°C. The greensymbol represents the pore water d18O at theLGM at ODP Site 1093 (49) (fig. S1). Error barsare 1s SD.

RESEARCH | REPORTon O

ctober 17, 2020

http://science.sciencemag.org/

Dow

nloaded from

temperature and d18ΟW of the deep SouthernOcean south of the Antarctic Polar Front.The SST record compares well with ice core–

derived Antarctic temperatures (25) during therecent glacial cycles (Fig. 1B). Prior to ~600,000years ago, however, the two records displaydistinct patterns, with the glacial SSTs gradu-ally decreasing from ~2°C to near-freezing whileAntarctic temperatures remained relatively sta-ble. For almost all glacials after theMPT as wellas some that preceded the MPT, reconstructedbottom-water temperatures (BWTs) are closeto the freezing point of seawater (Fig. 1C). Deep-water temperatures in the Southwest Pacific(5) are similar but somewhat warmer duringthe past few hundred thousand years, consistentwith the modern BWT difference between thetwo sites.In contrast to the glacial maxima, the Mg/Ca

records within the deglacials and interglacialsbear larger uncertainties owing to correctionsrequired for Mg associated with Mn phasescoated on foraminifera (fig. S2). Hence, this studyfocuses on the peak glacials. To aid comparisonbetween the glacial evolution of surface and deepwaters, we averaged all the peak glacials on thebasis of the LR04 benthic d18ΟC stack (23, 26)(Fig. 2 and figs. S5 to S8). Starting at MIS 16(~700,000 years ago) or perhaps slightly earlier,the glacial temperature and d18ΟW data indicatecontraction of the vertical temperature gradientand an expansion of the vertical d18ΟW gradient(Fig. 3, C and D).Under globally colder conditions,most climate

models simulate a decline in precipitation inthe AZ (27). Similarly, with reduced snowfallon Antarctica under a colder climate (28), coastaldischarge of meltwater and icebergs into the AZmay have declined. Because these reductions inthe freshwater input would act to increase thed18ΟW of the AZ surface, the observed decline ind18ΟW in the surface relative to the deep watermust result from an increase in the residencetime of surface waters during the ice ages sincetheMPT. Given the dominance of the underlyingdeep ocean as a source of new water to the AZsurface, the d18ΟW decline thus indicates a re-duction in the rate of deep-water inflow into theAZ surface mixed layer. This reduction mighthave been caused by a reduction in wind-drivenupwelling and/or turbulent or convective mix-ing across the base of the AZ surface mixedlayer. Assuming an ice age decrease of 10% inthe precipitation excess and meltwater dis-charge to the AZ surface ocean south of 50°S (27)and accounting for the expected temperature-related d18Ο decline of precipitation over the AZand the meltwater discharged from Antarcticaduring the glacials, the reconstructed verticald18ΟW gradient since the MPT indicates morethan a halving of the supply of the water fromthe subsurface (200 to 500 m) relative to mod-ern rates (Fig. 4). Thus, our results provideproxy evidence for a substantial reduction inthe supply of subsurface water to the AZ surfaceduring the late Pleistocene ice ages, and alsoindicate that this condition arose only at the

Hasenfratz et al., Science 363, 1080–1084 (2019) 8 March 2019 3 of 5

Fig. 3. Paleoceanographic changes in the Southern Ocean across the MPT. (A) AtmosphericCO2 records (10–12, 41). The dashed line around the CO2 data from (12) indicates the ageuncertainty. (B) Benthic d18OC stack (26) and normalized power spectral density of the 100,000-yearcycle in the data calculated using 500,000-year windows and ½ lags. (C and D) Surface-to-deepvertical gradients in glacial temperature and seawater d18O at ODP Site 1094 (see Fig. 2, B and C).(E) Iron flux to Subantarctic ODP Site 1090 (8) (fig. S1).

RESEARCH | REPORTon O

ctober 17, 2020

http://science.sciencemag.org/

Dow

nloaded from

end of the MPT. Today, the input of subsurfacewater warms the AZ surface. Thus, the cooling ofglacial-age AZ surface waters over theMPT (Figs.1B and 2B) may have been a consequence of thereconstructed decline in this input (29).In glacial stages prior to MIS 16, no d18ΟW

lowering of the surface relative to the subsurfaceis discernible, and the opposite is observed forseveral glacials, with a slightly lower d18ΟW re-constructed for deep water (Fig. 3D). This is incontrast to the expected signal from glacial-stage reductions in surface/deep water exchangeinferred from export production records (9, 10).During glacials prior to MIS 16, warmer AZSST may have quickly melted calving icebergs

near the coast, lowering the d18ΟW of coastalsurface waters and the deep water that theyform. This would have lowered the d18Οw ofdeep water forming near the Antarctic marginmore than it lowered the d18Οw of open AZsurface waters. In the Holocene, melting is alsoconcentrated along the coasts, and it does in-fluence the d18ΟW of newly formed deep water(30). However, southern-sourced deep waterhas dominated the deep Antarctic only duringthe glacials of the past 900,000 years (7), suchthat the effects of these surface AZ processes onbottom-water d18ΟW at ODP 1094 would havebeen stronger during the glacials than duringthe interglacials.

d18ΟW is mechanistically related to salinity(21): As with meltwater from the Antarctic icesheet, an excess in precipitation relative to evap-oration over the AZ lowers both d18ΟW andsalinity, while the upwelling subsurface waterincreases both quantities. Therefore, the ob-served surface d18ΟWdecline at ODP 1094 duringthe past 700,000 years likely also signals a sa-linity decline. Although Antarctic ice has a verylow d18ΟW, the rate of glacial meltwater additionis currently only 10 to 20% that of excess pre-cipitation (23, 31) and was likely even lower dur-ing glacial intervals (28). Neglecting it in ourcalculation, the decreased d18ΟW in the surfaceocean during the ice ages of the late Pleistocenewould have corresponded to a salinity decreaseof around 0.6 practical salinity units (psu) relativeto the underlying subsurface ocean, suggesting adoubling of the halocline strength compared totoday (23) (Fig. 4).Sea ice formation, the loss of brine from the

surface to the ocean interior, and the subsequentmelting of the sea ice at the AZ surface alsolower the salinity of the AZ surface mixed layer,especially in the open ocean (32). Because itseffects on d18ΟW are negligible, the salinity of theAZ surface during glacials would have been evenlower than calculated from d18ΟW. Thus, althoughthe multiple processes at work in the AZ cur-rently preclude a quantitative reconstructionof salinity based on d18ΟW, the data reportedhere provide robust evidence for a strengthen-ing of the AZ halocline during the post-MPT iceages. A strengthening of the AZ halocline wouldhave been promoted by the reconstructed in-crease in surface water residence time, whichitself would have resulted from a decrease inthe subsurface-to-surface water supply rate (33).Alternatively, or in addition, thehalocline strength-ening may have caused the decrease in sub-surface water supply through an increase in thevertical density gradient and thus the resistance tosurface/subsurface water exchange (34). Indeed,our reconstructions of temperature and salinitytogether indicate stronger density stratificationduring the peak glacials of the past 700,000 years(Fig. 3). Thus, a positive feedback between re-duced flow of deep water into the AZ surface andits freshening may have been a central dynamicin the glacial ocean.For at least the two most recent ice ages,

diatom- and coral-bound nitrogen isotope dataindicate that low subsurface-to-surface supplyof nitrate (and thus also water) was associatedwith more complete nitrate consumption inthe surface AZ waters (35–37), signaling an in-crease in the efficiency of the biological car-bon pump in the region and thus reduced AZCO2 leakage (14). Moreover, the reduction insubsurface-to-surface water supply since MIS16 may have involved a complementary decreasein the flow of AZ surface waters into the deepocean, which represents an additional mecha-nism for diminishing AZ CO2 leakage (14, 38, 39).The initial drivers of the reconstructed reduc-tion in subsurface-to-surface flow in the AZ didnot necessarily involve a complementary decline

Hasenfratz et al., Science 363, 1080–1084 (2019) 8 March 2019 4 of 5

Fig. 4. Reconstructeddecreases in the subsurfacewater supply to, and thesalinity of, the AntarcticZone surface during the iceages since the mid-Pleistocene transition.(A) Rates of net precipitation(Fp), meltwater discharge fromthe continent (Fm), and thesupply of subsurface (>200 mdepth) water to the AZ surface(Fsubsurf.), as well as theird18O and salinity (S), for themodern ice age (MIS 6).The subscripts p and m referto precipitation and meltwater.(B) Same as (A) for thepost-MPT ice age (MIS 16).The lateral and downwardfluxes out of the AZ surfaceare indicated by dashedarrows and must togethermatch the supply ofsubsurface water, meltwaterdischarge, and precipitationto the surface. Regulartypeface indicates observa-tions, either for the modern orfor glacials based onour data; bold typeface indi-cates calculated values using a simple mass balance approach (23). This approach neglectschanges in meridional transport and mixing, and does not take into account changes incontinental ice volume because they affect planktonic and benthic d18O equally (23). d18Op

was estimated using modern and glacial air temperature estimates at our study region (23).The asterisks in (B) indicate that d18OW,subsurf. and Ssubsurf. are artificially held at theirmodern values. d18OW,surface and Ssurface are calculated from these modern values and theobserved surface-to-deep differences. In this way, the only properties indicated to changebetween (A) and (B) are those involving the surface AZ. The d18OW,surface value in (B) is basedon the average d18OW surface-to-bottom water gradient observed for the post-MPT glacials(–0.48‰), which is then corrected for the modern d18OW difference between the shallowsubsurface ocean (200 to 500 m depth) and the d18OW of the bottom water at the 2807-m-deepcore site (a d18OW difference of 0.17‰). The subsurface-to-deep d18OW gradient is assumedto be constant, which is judged to be the greatest weakness in this calculation (50). Themodern d18OW values are from measurements obtained close to ODP 1094 (47, 48) (fig. S10),the modern d18Om value is based on d18O measurements of surface snow close to the coastof the Antarctic continent (23, 48), the modern salinity values are from (51), and Fp andFm are from (31). For the glacials, the net precipitation and the discharge of meltwater aredecreased by 10%, reflecting the results of model simulations (27).

RESEARCH | REPORTon O

ctober 17, 2020

http://science.sciencemag.org/

Dow

nloaded from

in surface-to-subsurface water flow (i.e., in deepocean ventilation). Nonetheless, the increase inthe residence time of AZ surface waters wouldhave contributed to the strengthening of theAZ halocline, which may have had the eventualconsequence of slowing deep ocean ventilation(33, 40).Changes in AZ overturning and their CO2

effects have been proposed to contribute to theorbitally paced glacial cycles throughout thePleistocene (9, 10). By 800,000 years ago,atmospheric pCO2 had reached an ice age min-imum value of ~180 parts per million (ppm)that has been roughly maintained over the latePleistocene (41) (Fig. 3A). In contrast, prior tothe MPT, ice age pCO2 minima are estimatedto be consistently greater than 200 ppm (10).An increase in Fe-bearing dust supply to theSubantarctic Ocean represents one possible ex-planation for this downward shift (8–10). Muchof the intensification of the glacial temperatureand d18ΟW gradients reconstructed here occurredfrom ~700,000 years ago onward, late in theMPT(Fig. 3, C and D). Thus, at least this portion of theincrease in AZ surface residence time occurredwithout concomitant declines in the ice age mi-nima in pCO2.However, the increase in AZ surface resi-

dence time at MIS 16 coincides with an abruptincrease in benthic d18ΟC, which has been sug-gested to reflect an increase in global con-tinental ice volume (3, 26, 42) (Fig. 3B). Giventhat deep water temperatures were close to thefreezing point during the ice ages within thepast 1 million years (5, 43) (Fig. 3B), the d18ΟC

increase observed in our record (Fig. 1A) pointsto a substantial increase in ice volume (3). Thelatter is corroborated by the first occurrenceof Heinrich stadials in the North Atlantic,640,000 years ago, indicating that the volumeof the Laurentide Ice Sheet possibly exceededa critical threshold (42, 44). This contrastswith the previously proposed notion of rela-tively stable glacial ice volume after the so-called900,000-year event (5). The lack of any d18ΟW

increase in the Southwest Pacific record at MIS16 may have been due to local hydrographicchanges (fig. S9).The ice volume increase at MIS 16 coincides

with the establishment of the dominant high-amplitude quasi–100,000-year glacial cycles asrevealed by spectral analysis of the benthic d18ΟC

stack (3, 26) (Fig. 3B). Thus, the rise in ice vol-ume may have been the result of the lengthen-ing of the glacial intervals, which allowed forlonger periods of uninterrupted ice sheet growth.We propose that the lengthening of ice ages, inturn, was promoted by the increased residencetime of Antarctic surface waters reconstructedhere. With a low glacial baseline rate of sub-surface water supply to the AZ surface, whichallowed the halocline to strengthen, and thepositive feedback between the two, it may havebeen more difficult than previously for orbitallypaced changes to enhance AZ surface-to-deepwater exchange to a rate that would have causedCO2 venting in the AZ surface. Consistent with

this view, model experiments indicate that, rel-ative to interglacial conditions, as increased nu-trient consumption and/or reduced deep oceanventilation in the AZ together become more ex-treme, the pCO2 declines become progressivelysmaller, with the system leveling off to a pCO2

of ~175 ppm (39). Thus, AZ changes at the endof the MPT may have been secondary in loweringpCO2 but central in preventing subsequentpCO2 rises, ensuring that glacial conditionspersisted despite orbital changes to the con-trary. In this view, the reconstructed change inglacial AZ conditions at the end of the MPT mayhave allowed Northern Hemisphere ice sheetsto survive periods of orbitally paced summer in-solation maxima on a more regular basis, therebyincreasing the longevity of ice ages.

REFERENCES AND NOTES

1. N. Pisias, T. Moore Jr., Earth Planet. Sci. Lett. 52, 450–458(1981).

2. K. A. Venz, D. A. Hodell, Palaeogeogr. Palaeoclimatol.Palaeoecol. 182, 197–220 (2002).

3. P. U. Clark et al., Quat. Sci. Rev. 25, 3150–3184 (2006).4. M. E. Raymo, P. Huybers, Nature 451, 284–285 (2008).5. H. Elderfield et al., Science 337, 704–709 (2012).6. A. Abe-Ouchi et al., Nature 500, 190–193 (2013).7. L. D. Pena, S. L. Goldstein, Science 345, 318–322 (2014).8. A. Martínez-Garcia et al., Nature 476, 312–315 (2011).9. S. L. Jaccard et al., Science 339, 1419–1423 (2013).10. T. B. Chalk et al., Proc. Natl. Acad. Sci. U.S.A. 114, 13114–13119

(2017).11. B. Hönisch, N. G. Hemming, D. Archer, M. Siddall,

J. F. McManus, Science 324, 1551–1554 (2009).12. J. A. Higgins et al., Proc. Natl. Acad. Sci. U.S.A. 112, 6887–6891

(2015).13. B. B. Stephens, R. F. Keeling, Nature 404, 171–174

(2000).14. J. L. Sarmiento, J. R. Toggweiler, Nature 308, 621–624

(1984).15. R. François et al., Nature 389, 929–935 (1997).16. D. M. Sigman, E. A. Boyle, Nature 407, 859–869 (2000).17. Shipboard Scientific Party, in Proc. ODP, Init. Reports, Leg 177,

R. Gersonde et al., Eds. (Ocean Drilling Program, 1999),vol. 177, chap. 9.

18. A. E. S. Kemp, I. Grigorov, R. B. Pearce, A. C. Naveira Garabato,Quat. Sci. Rev. 29, 1993–2009 (2010).

19. D. Nürnberg, J. Bijma, C. Hemleben, Geochim. Cosmochim. Acta60, 803–814 (1996).

20. J. R. O’Neil, J. Phys. Chem. 72, 3683–3684 (1968).21. A. N. LeGrande, G. A. Schmidt, Geophys. Res. Lett. 33, L12604

(2006).22. R. Srivastava, R. Ramesh, S. Prakash, R. Anilkumar,

M. Sudhakar, Geophys. Res. Lett. 34, L24603 (2007).23. See supplementary materials.24. H. Elderfield, J. Yu, P. Anand, T. Kiefer, B. Nyland, Earth Planet.

Sci. Lett. 250, 633–649 (2006).25. J. Jouzel et al., Science 317, 793–796 (2007).26. L. E. Lisiecki, M. E. Raymo, Paleoceanography 20, PA1003

(2005).27. W. R. Boos, J. Clim. 25, 992–1006 (2012).28. C. Lorius et al., Nature 316, 591–596 (1985).29. L. Menviel, A. Timmermann, O. Elison Timm, A. Mouchet,

Paleoceanography 25, PA4231 (2010).30. J. R. Toggweiler, B. Samuels, J. Phys. Oceanogr. 25, 1980–1997

(1995).31. R. P. Abernathey et al., Nat. Geosci. 9, 596–601 (2016).32. F. A. Haumann, N. Gruber, M. Münnich, I. Frenger, S. Kern,

Nature 537, 89–92 (2016).33. A. M. de Boer, J. R. Toggweiler, D. M. Sigman, J. Phys. Oceanogr.

38, 435–450 (2008).34. G. H. Haug, D. M. Sigman, R. Tiedemann, T. F. Pedersen,

M. Sarnthein, Nature 401, 779–782 (1999).35. R. S. Robinson, D. M. Sigman, Quat. Sci. Rev. 27, 1076–1090

(2008).36. A. Studer et al., Paleoceanography 30, 845–862 (2015).37. X. T. Wang et al., Proc. Natl. Acad. Sci. U.S.A. 114, 3352–3357

(2017).

38. J. R. Toggweiler, Paleoceanography 14, 571–588 (1999).39. M. P. Hain, D. M. Sigman, G. H. Haug, Global Biogeochem.

Cycles 24, GB4023 (2010).40. J. R. Toggweiler, J. L. Russell, S. R. Carson, Paleoceanography

21, PA2005 (2006).41. D. Lüthi et al., Nature 453, 379–382 (2008).42. D. A. Hodell, J. E. T. Channell, Clim. Past 12, 1805–1828

(2016).43. S. Bates, M. Siddall, C. Waelbroeck, Quat. Sci. Rev. 88, 147–158

(2014).44. D. A. Hodell, J. E. T. Channell, J. H. Curtis, O. E. Romero,

U. Röhl, Paleoceanography 23, PA4218 (2008).45. D. A. Hodell, C. D. Charles, J. H. Curtis, P. G. Mortyn,

U. S. Ninnemann, K. A. Venz, in Proc. ODP, Sci. Results, Leg177, R. Gersonde et al., Eds. (Ocean Drilling Program, 2003),vol. 177, chap. 9.

46. R. A. Locarnini, A. V. Mishonov, J. I. Antonov, T. P. Boyer,H. E. Garcia, O. K. Baranova, M. M. Zweng, C. R. Paver,J. R. Reagan, D. R. Johnson, M. Hamilton, D. Seidov, in NOAAAtlas NESDIS 73, S. Levitus, Ed. (NOAA, 2013).

47. A. Mackensen, H.-W. Hubberten, N. Scheele, R. Schlitzer,Paleoceanography 11, 203–215 (1996).

48. F. A. Haumann, K. Leonard, M. P. Meredith, C. Arrowsmith,I. V. Gorodetskaya, J. Hutchings, M. Lehning, M. J. Leng,S. Stammerjohn, M. Tsukernik, Y. Weber, Seawater stableisotope sample measurements from the AntarcticCircumnavigation Expedition (ACE) (2019).

49. J. F. Adkins, K. McIntyre, D. P. Schrag, Science 298, 1769–1773(2002).

50. D. C. Lund, J. F. Adkins, R. Ferrari, Paleoceanography 26,PA1213 (2011).

51. M. M. Zweng, J. R. Reagan, J. I. Antonov, R. A. Locarnini,V. Mishonov, T. P. Boyer, H. E. Garcia, O. K. Baranova,D. R. Johnson, D. Seidov, M. M. Biddle, in NOAA Atlas NESDIS74, S. Levitus, Ed. (NOAA, 2013).

ACKNOWLEDGMENTS

We thank U. Röhl, V. Lukies, and S. Steffer for technical supportduring the x-ray fluorescence analysis at MARUM; S. Bishop andM. Jaggi (ETH Zürich) and I. Mather (Cambridge) for technicalassistance with the isotopic analysis; C. Archer, S. Westermann (ETHZürich), and M. Greaves (Cambridge) for technical support with thetrace metal analysis; T. Gordijn for helping with the samplepreparation; and J. Gottschalk and J. Roberts for helpful discussions.Funding: Supported by ETH Research Grant ETH-04 11-1 (A.P.H.);Swiss National Science Foundation grants PP00P2_144811 andPP00P2_172915 (S.L.J.), PZ00P2_141424 (A.M.-G.), and 175162(F.A.H.); the Max Planck Society (A.M.-G. and G.H.H.); and NSF grantOPP-1401489 and ExxonMobil through Princeton University’sAndlinger Center for Energy and the Environment (D.M.S.). Modernoxygen isotope sample collection by F.A.H. was supported by theBNP Paribas Foundation and ACE, which was a scientific expeditioncarried out under the auspices of the Swiss Polar Institute,supported by funding from the ACE Foundation and FerringPharmaceuticals. This research used sediment samples provided bythe ODP, which is supported by NSF and participating countriesunder management of Joint Oceanographic Institutions (JOI) Inc.Author contributions: A.P.H., S.L.J., A.M.-G., D.A.H., and G.H.H.designed the study; D.A.H. and H.F.K. provided washed and sievedsediment samples from 0 to 121 meter composite depth andproduced planktonic foraminiferal stable isotope data within thisdepth range; A.P.H. washed and sieved the sediments, preparedthe samples, performed most of the analysis, and wrote the firstversion of the manuscript with assistance from S.L.J. andA.M.-G. throughout the entire study and from D.M.S. during thewriting; F.A.H. compiled modern seawater d18O data for theSouthern Ocean and helped with their interpretation in context ofthis study; S.M.B., D.A.H., and D.V. were involved in the isotopicand elemental analysis, respectively; and all authors providedsubstantial input to the final manuscript. Data and materialsavailability: Data not previously reported are archived at theBern Open Repository and Information System (BORIS)(https://boris.unibe.ch/id/eprint/127004).

SUPPLEMENTARY MATERIALS

www.sciencemag.org/content/363/6431/1080/suppl/DC1Materials and MethodsFigs. S1 to S15References (52–82)

26 March 2018; accepted 23 January 201910.1126/science.aat7067

Hasenfratz et al., Science 363, 1080–1084 (2019) 8 March 2019 5 of 5

RESEARCH | REPORTon O

ctober 17, 2020

http://science.sciencemag.org/

Dow

nloaded from

The residence time of Southern Ocean surface waters and the 100,000-year ice age cycle

M. Bernasconi, Helga (Kikki) F. Kleiven, F. Alexander Haumann and Gerald H. HaugAdam P. Hasenfratz, Samuel L. Jaccard, Alfredo Martínez-García, Daniel M. Sigman, David A. Hodell, Derek Vance, Stefano

DOI: 10.1126/science.aat7067 (6431), 1080-1084.363Science 

, this issue p. 1080; see also p. 1040Sciencedioxide release.may have caused more prolonged ice ages by making the Antarctic less responsive to orbitally paced drivers of carbon100,000-year cycle coincided with a reduction in deep-water supply and a freshening of the surface ocean. This slowing depths to the surface over the past 1.5 million years (see the Perspective by Menviel). The emergence of theplanktonic foraminifera from the Antarctic in order to reconstruct changes in the rate of transfer of ocean water from the

measured the oxygen isotope composition and magnesium/calcium ratio in benthic andet al.epoch. Why? Hasenfratz The periodicity of glacial cycles changed from 100,000 to 41,000 years during the middle of the Pleistocene

Resetting the glacial timer

ARTICLE TOOLS http://science.sciencemag.org/content/363/6431/1080

MATERIALSSUPPLEMENTARY http://science.sciencemag.org/content/suppl/2019/03/06/363.6431.1080.DC1

CONTENTRELATED http://science.sciencemag.org/content/sci/363/6431/1040.full

REFERENCES

http://science.sciencemag.org/content/363/6431/1080#BIBLThis article cites 80 articles, 9 of which you can access for free

PERMISSIONS http://www.sciencemag.org/help/reprints-and-permissions

Terms of ServiceUse of this article is subject to the

is a registered trademark of AAAS.ScienceScience, 1200 New York Avenue NW, Washington, DC 20005. The title (print ISSN 0036-8075; online ISSN 1095-9203) is published by the American Association for the Advancement ofScience

Science. No claim to original U.S. Government WorksCopyright © 2019 The Authors, some rights reserved; exclusive licensee American Association for the Advancement of

on October 17, 2020

http://science.sciencem

ag.org/D

ownloaded from