6
DOI: 10.1126/science.1203580 , 347 (2011); 334 Science , et al. Stephen Barker 800,000 Years of Abrupt Climate Variability This copy is for your personal, non-commercial use only. clicking here. colleagues, clients, or customers by , you can order high-quality copies for your If you wish to distribute this article to others here. following the guidelines can be obtained by Permission to republish or repurpose articles or portions of articles ): October 24, 2011 www.sciencemag.org (this infomation is current as of The following resources related to this article are available online at http://www.sciencemag.org/content/334/6054/347.full.html version of this article at: including high-resolution figures, can be found in the online Updated information and services, http://www.sciencemag.org/content/suppl/2011/09/08/science.1203580.DC1.html can be found at: Supporting Online Material http://www.sciencemag.org/content/334/6054/347.full.html#related found at: can be related to this article A list of selected additional articles on the Science Web sites http://www.sciencemag.org/content/334/6054/347.full.html#ref-list-1 , 7 of which can be accessed free: cites 33 articles This article registered trademark of AAAS. is a Science 2011 by the American Association for the Advancement of Science; all rights reserved. The title Copyright American Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005. (print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by the Science on October 24, 2011 www.sciencemag.org Downloaded from

800,000 Years of Abrupt Climate Variability Stephen …web-static-aws.seas.harvard.edu/climate/seminars/pdfs/...800,000 Years of Abrupt Climate Variability Stephen Barker,1* Gregor

  • Upload
    others

  • View
    8

  • Download
    0

Embed Size (px)

Citation preview

DOI: 10.1126/science.1203580, 347 (2011);334 Science

, et al.Stephen Barker800,000 Years of Abrupt Climate Variability

This copy is for your personal, non-commercial use only.

clicking here.colleagues, clients, or customers by , you can order high-quality copies for yourIf you wish to distribute this article to others

  here.following the guidelines

can be obtained byPermission to republish or repurpose articles or portions of articles

  ): October 24, 2011 www.sciencemag.org (this infomation is current as of

The following resources related to this article are available online at

http://www.sciencemag.org/content/334/6054/347.full.htmlversion of this article at:

including high-resolution figures, can be found in the onlineUpdated information and services,

http://www.sciencemag.org/content/suppl/2011/09/08/science.1203580.DC1.html can be found at: Supporting Online Material

http://www.sciencemag.org/content/334/6054/347.full.html#relatedfound at:

can berelated to this article A list of selected additional articles on the Science Web sites

http://www.sciencemag.org/content/334/6054/347.full.html#ref-list-1, 7 of which can be accessed free:cites 33 articlesThis article

registered trademark of AAAS. is aScience2011 by the American Association for the Advancement of Science; all rights reserved. The title

CopyrightAmerican Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005. (print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by theScience

on

Oct

ober

24,

201

1w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

800,000 Years of AbruptClimate VariabilityStephen Barker,1* Gregor Knorr,2 R. Lawrence Edwards,3 Frédéric Parrenin,4,5

Aaron E. Putnam,6 Luke C. Skinner,7 Eric Wolff,8 Martin Ziegler1

We constructed an 800,000-year synthetic record of Greenland climate variability based on thethermal bipolar seesaw model. Our Greenland analog reproduces much of the variability seen inthe Greenland ice cores over the past 100,000 years. The synthetic record shows strong similaritywith the absolutely dated speleothem record from China, allowing us to place ice core recordswithin an absolute timeframe for the past 400,000 years. Hence, it provides both a stratigraphicreference and a conceptual basis for assessing the long-term evolution of millennial-scale variabilityand its potential role in climate change at longer time scales. Indeed, we provide evidence for aubiquitous association between bipolar seesaw oscillations and glacial terminations throughout theMiddle to Late Pleistocene.

Ice core records from Greenland first demon-strated the existence of repeated, large, abruptshifts in Northern Hemisphere climate dur-

ing the last ice age (1, 2). These shifts are oneexpression of a global system that is capable ofdriving major changes in climate componentssuch as ocean temperatures (3, 4) and monsoonrainfall (5). The Greenland records provide anarchetypal view of abrupt climate variability (6)over the last glacial cycle, which was character-ized by rapid alternations between cold (stadial)and warmer (interstadial) conditions [known asDansgaard-Oeschger (D-O) oscillations]. But iron-ically, the very high temporal resolution of theserecords makes it difficult to look farther back intime; the high accumulation rates on the Green-land ice sheet mean that more than 3000 m ofice may represent just 100,000 years of climatehistory. Fortunately, climate records preserved inAntarctic ice (7) enable us to address this funda-mental problem.

The thermal bipolar seesaw model (8, 9) at-tempts to explain the observed relationship betweenmillennial-scale temperature variability observedin Greenland and Antarctica by calling on var-iations in the strength of the Atlantic meridionaloverturning circulation (AMOC). The north-ward heat transport associatedwith this circulation(10) implies that changes in the strength of over-turning should lead to opposing temperature re-sponses in either hemisphere. According to theseesaw model, a transition from weak to strongAMOCwould cause an abrupt warming across the

North Atlantic region (a D-O warming event)while temperatures across Antarctica would (ingeneral) shift fromwarming to cooling. The ocean-atmosphere climate system is an integrated andsynergistic system, and it is important to note thatthe overall concept of the bipolar seesaw we in-voke here is not restricted to oceanic processes butalso includes atmospheric shifts that may be re-lated to the variationswe are interested in (9,11–14).

According to the thermal bipolar seesawmodel, we should observe an antiphase relation-ship between the Greenland temperature anom-aly and the rate of change of Antarctic temperature(9). This can be illustrated by a lead/lag analysisof the methane-tuned temperature records afterremoval of their orbital time scale variability (6)(Greenland, GLT_hi, and Antarctica, AAT_hi)and the first time derivative of Antarctic temper-ature, AAT_hi´ (Fig. 1). Comparison of the undif-ferentiated records illustrates the historical debateas to whether the two signals are positively cor-related, with a southern lead of 1000 to 1600years, or negatively correlated, with the north lead-ing by 400 to 800 years (15, 16). However, asimplied in (9) and illustrated in Fig. 1, a nearzero-phase anticorrelation is observed betweenGLT_hi and AAT_hi´. Uncertainties in the iceage–gas age offset (Dage), which may be up tohundreds of years (17), mean that an exact anti-phase relationship is unlikely to be observed (6).As described by (18) using a similar approach,the process of differentiating amplifies noise inthe original temperature record. Smoothing therecord of AAT_hi before differentiating reducesthis noise but will compromise the ability ofAAT_hi´ to replicate the abrupt nature of D-Owarming events and reduce the predicted ampli-tude of smaller events. The choice of smoothingwindow is therefore a trade-off between these ef-fects (6) (Fig. 1).

The empirical relations illustrated in Fig. 1 of-fer the possibility of producing a synthetic recordof Greenland climate using the Antarctic record,with the purpose of reconstructing the natureof northern variability beyond the present limitof the Greenland records. The record of Green-

land temperature (GLT) is broken down into itsorbital and millennial time scale components,GLT_lo and GLT_hi, respectively (Fig. 2 and fig.S5), where GLT_lo is a 7000-year smoothing ofGLT (6) and GLT_hi is the difference betweenGLT and GLT_lo. We consider GLT_hi as the north-ern temperature anomaly with respect to meanbackground conditions. Building on Fig. 1, weassume that the rate of Antarctic temperaturechange is inversely proportional to the northerntemperature anomaly. We therefore scale the am-plitude of AAT_hi´ to match that of GLT_hi toproduce a synthetic record of northern millennial-scale temperature variability, GLT_syn_hi (Fig. 2C)(6). It can be seen that a synthetic reconstructionof GLT could be made by combining GLT_syn_hiwith an estimate for GLT_lo. The orbital timescale components of the Greenland and Ant-arctic temperature records (GLT_lo and AAT_lo,respectively) are highly correlated, with the south-ern record leading the north by ~2000 years(6). We therefore incorporate longer time scalevariations into our reconstruction by substitut-ing GLT_lo with a scaled version of AAT_lo,shifted by 2000 years (which we call GLT_syn_lo)(fig. S5) (6). Our full reconstruction, GLT_syn(Fig. 2D), is then the sum of GLT_syn_lo andGLT_syn_hi.

Our formulation of the thermal bipolar see-saw concept is qualitatively analogous to that of(9) in that it implies the existence of a heat reser-voir that convolves the northern signal, produc-ing a southern signal with a longer characteristictime scale. Our approach is slightly different inthat we relate the rate of Antarctic temperaturechange directly to the northern temperature anom-aly. Indeed, we note that for some long stadialevents, particularly those associated with gla-cial terminations, Antarctic temperatures appear torise unabated until an abrupt warming event oc-curs in the north (19). On the other hand, ourformulation does not imply that Antarctic temper-atures must continue to rise indefinitely when-ever Greenland is cold, only while it is cold withrespect to background conditions (defined bythe orbital time scale component). We also notethat northern temperature (regardless of back-ground conditions) is not always constant through-out stadial events. For example, Greenland warmedsignificantly during cold stadial 21 (Fig. 2D).By our formulation, the rate of Antarctic tem-perature rise during this event would decreasecorrespondingly, in line with observations (Fig.2A) (20).

We used a thresholding approach for predict-ing the occurrence of abrupt Greenland warmingevents based on minima in the second time dif-ferential of AAT (AAT´´) (Fig. 2, E and F). Thishas an advantage over use of the first differential(decreasing through zero) because it is capable ofdistinguishing between events of varying mag-nitude and incorporates information about con-ditions before and after an abrupt event (6). Using arelatively insensitive threshold (blue dashed line inFig. 2F), we are able to identify the largest D-O

1School of Earth and Ocean Sciences, Cardiff University, CardiffCF10 3AT, UK. 2Alfred Wegener Institute, 27570 Bremerhaven,Germany. 3Department of Earth Sciences, University of Min-nesota, Minneapolis, MN 55455, USA. 4Laboratoire de Glaciologie,CNRS and Joseph Fourier University, 38400 Grenoble, France.5Laboratoire Chrono-Environnement, 25000 Besançon, France.6Department of Earth Sciences and Climate Change Institute,University of Maine, Orono, ME 04469, USA. 7Godwin Lab-oratory for Palaeoclimate Research, Department of EarthSciences, University of Cambridge, Cambridge CB2 3EQ, UK.8British Antarctic Survey, Madingley Road, High Cross,Cambridge CB3 0ET, UK.

*To whom correspondence should be addressed. E-mail:[email protected]

www.sciencemag.org SCIENCE VOL 334 21 OCTOBER 2011 347

REPORTS

on

Oct

ober

24,

201

1w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

temperature shifts recorded in Greenland, whereas“smaller” events, such as D-O 2, require a moresensitive threshold (red line). Accordingly, we areable to identify almost all of the canonical D-Oevents over the past 90,000 years without intro-ducing “spurious events.” We incorporate a cor-rection, based on the time integrated per 55-cmsample of ice (7), to account for loss of temporalresolution in the deeper parts of the ice core (6).

Our synthetic reconstruction of Greenland tem-perature closely resembles the observed recordin terms of both its timing and the structure of in-dividual events (Fig. 2). This is despite probablevariability in the relationships between d18O [forthe Greenland Ice Sheet Project 2 ice core (GISP2)]or dD [for the European Project for Ice Coringin Antarctica Dome C ice core (EDC)] versuslocal temperature through time (21) and othermillennial-scale variability that might not be re-lated to the bipolar seesaw. We find similar resultsusing alternative Antarctic ice core records (6).On the basis of the predictive ability of GLT_synover the past ~100,000 years, we can extend ourreconstruction back to ~800,000 years ago (Fig.3 and Fig. 4). In doing so we implicitly assumethat the empirical relationships observed over

the last glacial cycle held during earlier periods.Given the inherent uncertainty in this assump-tion, we do not claim particular skill at pre-dicting the absolute amplitude of earlier events;however, we do suggest that to the extent thatthe underlying physical mechanisms did persistthroughout the past 800,000 years, the timingand overall structure of events will be relativelyrobust.

To test this hypothesis, we compared our syn-thetic reconstructions with real climate records.The record of Asian monsoon variability derivedfrom cave deposits (speleothems) in China (5, 22)is one of the best candidates for this task (Fig. 3).The Chinese speleothem d18O record is thoughtto represent changes in the proportion of lowd18O (summer monsoon) rainfall within annualtotals (22) and can be considered a measure ofthe amount of summer monsoon rainfall or mon-soon intensity. The combined record from severaldeposits taken from a number of caves providesa continuous, absolutely dated record over thepast ~400,000 years (22). The record is domi-nated by orbital time scale changes, possibly re-lated to the influence of boreal summer insolationon the strength of the Asian monsoon (6). Re-

moval of this variability by normalizing to theinsolation curve (6) reveals distinctive millennialtime scale activity that has been shown to cor-respond with D-O variability over Greenlandduring the last glacial cycle (5, 22). This corre-spondence is thought to be caused by latitudinalshifts in the position of the Intertropical Con-vergence Zone (ITCZ) and related atmospher-ic phenomena in response to variations in theAMOC and related changes in North Atlantictemperature (11).

There is a strong one-to-one correspondencebetween inferred weak-monsoon events and ourreconstructed cold events in Greenland (Fig. 3).Moreover, there are pronounced similarities inthe structure of abrupt events, particularly duringdeglacial episodes (terminations); the multipleweak-monsoon events associated with glacialterminations of the Late Pleistocene (22) are re-flected by multiple cold events in our records.Our reconstruction suggests the occurrence oflarge-amplitude “D-O–type” oscillations between160,000 and 180,000 years ago [during marineisotope stage 6 (MIS 6)] (Fig. 3). These may becompared with similar events in the records ofplanktonic d18O and tree pollen from a marine

3 2 1 0 1 2 3−0.2

0

0.2

0.4

0.6

0.8

1

Lag (kyr)

North leads South leadsB

Methane

Cor

rela

tion

Cor

rela

tion

North leads South leads

3 2 1 0 1 2 3−0.8

−0.4

0.4

0.8

Lag (kyr)

0

1000yr700yr200yr

C

Vs AAT_hi

-10

0

10

30 40 50 60 70 80AA

T_h

i (‰

)

Age (kyr) (EDC3)

-0.05

0

0.05

-0.01

0

0.01

-0.01

0

0.01

-2

0

2

GL

T_h

i (‰

)

-100

0

100

ED

C C

H4

(ppb

v)

-100

0

100

GIS

P C

H4

(ppb

v)

A

Fig. 1. Ice core records from Greenland (GISP2) and Antarctica (EDC) (all records have orbital timescale variability removed; kyr, thousands of years). (A) Methane (17, 25) with tuning points (crosses)used to place all records on the EDC3 age model (26), Greenland temperature (GLT_hi with 200-yearsmoothing) derived from d18O of ice (2), and Antarctic temperature (AAT_hi, 200-year smoothing)from dD of ice (7). First derivatives of AAT_hi (AAT_hi´) are shown for various smoothing lengths (in brackets) of the undifferentiated record. (B) Lead/lagcorrelation of the methane records suggests successful tuning of the gas records. (C) Lead/lag correlations between GLT_hi, AAT_hi, and AAT_hi´ reveal thewell-known relation between northern and southern temperature records and the antiphase relationship between GLT_hi and AAT_hi´.

21 OCTOBER 2011 VOL 334 SCIENCE www.sciencemag.org348

REPORTS

on

Oct

ober

24,

201

1w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

Fig. 2. Reconstructing millennial-scale climatevariability over Greenland using the Antarctic tem-perature record. (A) The record of dD from the EDCice core (AAT) (7) with a 7000-year smoothing ofthe same record (AAT_lo). (B) Removal of orbitaltime scale variability and application of a 700-year smoothing (6) produces AAT_hi. (C) AAT_hi isdifferentiated and then scaled to GLT_hi (greencurve) to produce GLT_syn_hi (orange curve). (D)A synthetic reconstruction of Greenland temper-ature variability (GLT_syn; red curve) constructedby adding GLT_syn_hi to GLT_syn_lo (6). Greencurve is GISP2 d18O placed on EDC3 via methanetuning (Fig. 1). (E and F) Minima in AAT´´ below athreshold (dashed lines) are used to predict theoccurrence of major warming events in Greenland(F), identified by the corresponding colored dots in(E). A threshold value of –1.2 × 10−5 (red dashedline and dots) has good success at picking canon-ical D-O events [green numbers in (D) and (E)]without introducing spurious events.

1

2

34 5 6 7

8

910

1112

13

141516

17

18

19 20

21

22

0

1 2 3 4 5 6 7 8 91011 1213 14151617 18 19 20 21 220

10 20 30 40 50 60 70 80 90Age kyr (EDC3)

Age kyr (EDC3)10 20 30 40 50 60 70 80 90

GLT

E-44

-40

-36

GL

T (

δ 18

O, ‰

)

D

GLT_syn

-2x10-5

0

2x10-5

F

-10

0

10B

AA

T_h

i

-450

-430

-410

-390

ED

C δ

D (

‰)

AAAT

AAT_lo

-4

-2

0

2

4

C

GLT_hiGLT_syn_hi

GL

T _

hi

Fig. 3. (A to H) Compari-son of reconstructed Green-land climate variability withother records. The normal-ized record of monsoonvariability from China (A)(5, 22), marine records fromthe Iberian Margin [(C)to (E)] (23, 32), and therecord of atmospheric CH4(G) (25) share many fea-tures in common with ourrecords derived from theAntarctic temperaturerecord[(B) and (F)]. Colored dotsin (H) represent the oc-currence of D-O eventspredicted from AAT´´ usinga fixed (red) or variable(blue) threshold (6). Allrecords are on the EDC3time scale (26) except themonsoon record which ison its own absolute timescale (22). Thepollen recordfrom MD95-2042 (32) (C)was placed on EDC3 by tun-ing the corresponding plank-tonic d18O record (24) toGLT_recon. Gray bars indi-cate cold conditions andperiods of weak monsoon.Glacial terminations are indicated by Roman numerals.

400

500

600

700

CH

4 (p

pbv)G-40

-35

-30

GL

T_s

yn (

‰) F

120 140 160 180 200 220 240 260Age (kyr)

0

20

40

Tre

e po

llen

(%)

E

20

40

60

Med

iterr

anea

npo

llen

(%)

D

1

2

3

plan

kton

ic δ

18O

(‰

)

C-4

0

4

GL T

_syn

_hi (

‰)

B

Spe

leot

hem

δ18

O(n

orm

aliz

ed) A

H

T-II T-III

www.sciencemag.org SCIENCE VOL 334 21 OCTOBER 2011 349

REPORTS

on

Oct

ober

24,

201

1w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

sediment core taken from the Iberian Margin (23).On the basis of the findings of Shackleton et al.(24), the Iberian Margin records were tuned tothe EDC dD record via the record of benthicd18O from the same core (23). The tuning exer-cise did not involve the surface records, whichtherefore provide a quasi-objective “target” forcomparing with our reconstruction [which shouldbe aligned with the surface-ocean records accord-ing to (24)], and there is good agreement in thetiming and structure of the abrupt events during

MIS 6. We also note good agreement betweenour reconstructions and the record of atmosphericCH4 (25). Our predicted D-O warming events aregenerally aligned with sharp increases in CH4

(similar to the observed relationship during MIS3). This relationship holds for the entire 800,000-year record (Fig. 4) and provides critical ground-truthing for our reconstruction.

Building on previous studies (22), we usedthe precise and absolutely dated Chinese speleo-them record to place our reconstruction on an

absolute time scale for the past 400,000 years.We did this by aligning the cold events in ourreconstruction with the inferred weak-monsoonevents in the speleothem record (Fig. 4) (6). TheEDC3 age scale (which remains the fundamen-tal basis for our model) was derived through acombination of ice flow modeling and variousage markers, including orbital tuning constraints(26). By tuning the millennial-scale features ofGLT_syn to the speleothem record, we providea refinement of the age scale that provides an

Fig. 4. (A and B) 800,000years of abrupt climate varia-bility. Records of North AtlanticIRD (4), monsoon rainfall (5, 22)(normalized) and SST fromthe Iberian Margin (27) allshow strong similarities withour reconstruction of Greenlandclimate variability (GLT_syn_hiand GLT_syn). Glacial termi-nations (identified by Romannumerals) are characterized bycold conditions across Green-land and the North Atlanticand weakened monsoon rain-fall, with a corresponding risein atmospheric CO2 (33), fol-lowed by an abrupt warmingover Greenland, strengtheningof the monsoon, and sharp risein atmospheric CH4 (25). Pinkboxes indicate terminal North-ern Hemisphere cold periods.Red and blue dots are predictedD-O warming events using afixed or variable threshold, re-spectively. Lowermost curves ineach panel are moving win-dows of the standard devia-tion of AAT_hi´, our ”bipolarseesaw activity index“ (notethat orbital time scale varia-tions have been removed; blueis 5000-year window; green is10,000-year window). Increasedmillennial-scale activity is gen-erally observed during transi-tions between climate states,with minimal activity duringinterglacials and glacial max-ima. All records are on the newSpeleo-Age (A) or the EDC3(B) time scale except the ben-thic d18O stack (LR04) of (34),which is on its own time scale.

TI TII TIIIA TIII TIV

TV TVI TVII

400

600

800

CH

4 (p

pbv)

-45

-40

-35

-30

GL T

_syn

(‰

)

-4

0

4

GL T

_syn

_hi

(‰)

AA

T_h

i’ S

D (

x10-3

)

0

5

10

3

4

5 δ18O

(LR

04)

400 450 500 550 600 650 700 750 800Age (kyr) (EDC3)

B

200

300

CO

2 (p

pmv)

400

600

800

CH

4 (p

pbv)

5

10

15

20

Iber

ian

Mar

gin

SS

T (

ºC)-45

-40

-35

-30

GL T

_syn

(‰

)

-4

0

4

GL T

_syn

_hi

(‰)

Spe

leot

hem

δ18

O(n

orm

aliz

ed)

0

50

100

Nor

th A

tlant

icIR

D (

%)

Age (kyr) ‘Speleo-Age’0 50 100 150 200 250 300 350 400

A

3

4

5δ18O

(LR

04)

AA

T_h

i’ S

D (

x10-3

)

0

5

10

200

300

CO

2 (p

pmv)

21 OCTOBER 2011 VOL 334 SCIENCE www.sciencemag.org350

REPORTS

on

Oct

ober

24,

201

1w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

alternative to the ultimate dependence on orbitaltuning. In addition to providing an absolute timescale for the ice and gas records from Antarctica,we can also use our absolutely dated Green-land reconstruction as a tuning target for otherhigh-resolution paleo-records, such as recordsof ice-rafted debris (IRD) from a North Atlanticsediment core (4) and a record of sea surfacetemperature (SST) from a core off the IberianMargin (27) (Fig. 4). Each of these records hasbeen tuned to our reconstruction on its absolutetime scale (6).

Our synthetic records confirm that millennialtime scale variability and abrupt climate oscilla-tions occurred in Greenland throughout the past800,000 years, and more specifically they sug-gest that the underlying physical mechanisms rep-resented by the conceptual thermal bipolar seesawwere relatively invariant throughout this period.In line with observations for the last glacial pe-riod (28), our reconstructions suggest that higher-amplitude variability and more frequent D-O–likewarming events occurred when climate was inan intermediate state or during the transitionsbetween states (Fig. 4). Extending the observa-tions of (22), we find that glacial terminationsof the Middle to Late Pleistocene in generalwere characterized by oscillations of the bipolarseesaw.

This apparently ubiquitous association ofmillennial-scale climate variability with glacialterminations raises an important question: Is thismode of variability a necessary component of de-glacial climate change, or merely a complicatingfactor? Previous studies (28, 29) have suggestedthat D-O–type variability might represent an in-herent resonance of the climate system, attaininga high amplitude only within certain windowsof opportunity (i.e., intermediate climate states).Given that global climate must pass through such

a window during deglaciation, one could arguethat terminal oscillations of the bipolar seesaw aremerely a symptom of deglacial climate change(29). However, the precise correspondence observedbetween bipolar seesaw oscillations and changesin atmospheric CO2 during glacial terminations(Fig. 4) suggests that the bipolar seesaw may playmore than just a passive role in the mechanismof deglaciation (i.e., through the positive feedbacksassociated with increasing CO2) (14, 19, 22). Withthe supercritical size of continental ice sheets asa possible precondition (30), and in combinationwith the right insolation forcing (31) and ice al-bedo feedbacks, the CO2 rise associated with anoscillation of the bipolar seesaw could providethe necessary additional forcing to promote de-glaciation. In this sense, the overall mechanismof glacial termination during the Middle to LatePleistocene might be viewed as the timely andnecessary interaction between millennial and or-bital time scale variations.

References and Notes1. W. Dansgaard et al., Science 218, 1273 (1982).2. M. Stuiver, P. M. Grootes, Quat. Res. 53, 277

(2000).3. G. Bond et al., Nature 365, 143 (1993).4. J. F. McManus, D. W. Oppo, J. L. Cullen, Science 283, 971

(1999).5. Y. J. Wang et al., Nature 451, 1090 (2008).6. See supporting material on Science Online.7. J. Jouzel et al., Science 317, 793 (2007).8. W. S. Broecker, Paleoceanography 13, 119 (1998).9. T. F. Stocker, S. J. Johnsen, Paleoceanography 18, 1087

(2003).10. K. E. Trenberth, J. M. Caron, J. Clim. 14, 3433

(2001).11. M. Vellinga, R. A. Wood, Clim. Change 54, 251

(2002).12. J. C. H. Chiang, M. Biasutti, D. S. Battisti,

Paleoceanography 18, 1094 (2003).13. R. F. Anderson et al., Science 323, 1443 (2009).14. S. Barker et al., Nature 457, 1097 (2009).

15. A. Schmittner, O. A. Saenko, A. J. Weaver, Quat. Sci. Rev.22, 659 (2003).

16. E. J. Steig, R. B. Alley, Ann. Glaciol. 35, 451(2002).

17. T. Blunier, E. J. Brook, Science 291, 109 (2001).18. M. Siddall et al., Quat. Sci. Rev. 25, 3185 (2006).19. E. W. Wolff, H. Fischer, R. Rothlisberger, Nat. Geosci. 2,

206 (2009).20. E. Capron et al., Quat. Sci. Rev. 29, 222 (2010).21. E. Capron et al., Clim. Past 6, 345 (2010).22. H. Cheng et al., Science 326, 248 (2009).23. V. Margari et al., Nat. Geosci. 3, 127 (2010).24. N. J. Shackleton, M. A. Hall, E. Vincent,

Paleoceanography 15, 565 (2000).25. L. Loulergue et al., Nature 453, 383 (2008).26. F. Parrenin et al., Clim. Past 3, 485 (2007).27. B. Martrat et al., Science 317, 502 (2007).28. M. Schulz, W. H. Berger, M. Sarnthein, P. M. Grootes,

Geophys. Res. Lett. 26, 3385 (1999).29. A. Sima, A. Paul, M. Schulz, Earth Planet. Sci. Lett. 222,

741 (2004).30. M. E. Raymo, Paleoceanography 12, 577 (1997).31. J. D. Hays, J. Imbrie, N. J. Shackleton, Science 194, 1121

(1976).32. M. F. Sánchez Goñi, F. Eynaud, J. L. Turon, N. J. Shackleton,

Earth Planet. Sci. Lett. 171, 123 (1999).33. D. Lüthi et al., Nature 453, 379 (2008).34. L. E. Lisiecki, M. E. Raymo, Paleoceanography 20,

PA1003 (2005).Acknowledgments: We thank the authors of all of the

studies cited here for making their results available forthis work. Supported by a Philip Leverhulme Prize (S.B.),Natural Environment Research Council (UK) awardsNE/F002734/1 and NE/G004021/1 (S.B.), and NSF grants0502535 and 1103403 (R.L.E.). This study is also partof the British Antarctic Survey Polar Science for PlanetEarth Programme, funded by the Natural EnvironmentResearch Council (UK).

Supporting Online Materialwww.sciencemag.org/cgi/content/full/science.1203580/DC1Materials and MethodsFigs. S1 to S14Tables S1 to S3References

31 January 2011; accepted 26 August 2011Published online 8 September 2011;10.1126/science.1203580

Pre-Clovis Mastodon Hunting 13,800Years Ago at the Manis Site, WashingtonMichael R. Waters,1* Thomas W. Stafford Jr.,2,5 H. Gregory McDonald,3 Carl Gustafson,4

Morten Rasmussen,5 Enrico Cappellini,5 Jesper V. Olsen,6 Damian Szklarczyk,6 Lars Juhl Jensen,6

M. Thomas P. Gilbert,5 Eske Willerslev5

The tip of a projectile point made of mastodon bone is embedded in a rib of a single disarticulatedmastodon at the Manis site in the state of Washington. Radiocarbon dating and DNA analysisshow that the rib is associated with the other remains and dates to 13,800 years ago. Thus, osseousprojectile points, common to the Beringian Upper Paleolithic and Clovis, were made and usedduring pre-Clovis times in North America. The Manis site, combined with evidence of mammothhunting at sites in Wisconsin, provides evidence that people were hunting proboscideans atleast two millennia before Clovis.

Recent studies have strengthened the casethat the makers of Clovis projectile pointswere not the first people to occupy the

Americas (1–5). If hunting by humans was re-sponsible for the megafauna extinction at the

end of the Pleistocene (6), hunting pressuresmust have begun millennia before Clovis (7).Here we reexamine the evidence from the Manissite in the state of Washington (8), an early mas-todon kill that dates to 800 years before Clovis.

Between 1977 and 1979, a single male mas-todon (Mammut americanum) was excavated fromsediments at the base of a kettle pond at theManis site (figs. S1 to S3) (8–10). Some boneswere spirally fractured, multiple flakes were re-moved from one long bone fragment, and otherbones showed cut marks (8, 11, 12). The onlydocumented artifact associated with the masto-don was a foreign osseous fragment, interpretedas the tip of a bone or antler projectile point,

1Center for the Study of the First Americans, Departments ofAnthropology and Geography, Texas A&M University, 4352TAMU, College Station, TX 77843–4352, USA. 2Stafford Re-search, 200 Acadia Avenue, Lafayette, CO 80026–1845, USA.3Park Museum Management Program, National Park Service,1201 Oakridge Drive, Suite 150, Fort Collins, CO 80525, USA.4245 Southeast Derby Street, Pullman, WA 99163–2217, USA.5Centre for GeoGenetics, University of Copenhagen, ØsterVoldgade 5-7, 1350 Copenhagen, Denmark, 6Novo NordiskFoundation Center for Protein Research, Faculty of Health Sci-ences, University of Copenhagen, Blegdamsvej 3b, 2200 Co-penhagen, Denmark.

*To whom correspondence should be addressed. E-mail:[email protected]

www.sciencemag.org SCIENCE VOL 334 21 OCTOBER 2011 351

REPORTS

on

Oct

ober

24,

201

1w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from