northern hemisphere glaciation during the globally warm ...the early late pliocene (3.6 to ~3.0...

OPEN 3 ACCESS Freely available online © PLOSI o -

Northern Hemisphere Glaciation during the Globally Warm Early Late PlioceneStijn De Schepper1'2*, Jeroen Groeneveld3, B. David A Naafs40, Cédéric Van Renterghem5, Jan Hennissen6, Martin J. Head6'7, Stephen Louwye5, Karl Fabian81 Department of Earth Science, University of Bergen, Bergen, Norway, 2 Geosciences Department, University of Bremen, Bremen, Germany, 3 MARUM - Center for Marine Environmental Sciences, University of Bremen, Bremen, Germany, 4 Alfred Wegener Institute for Polar and Marine Research, Bremerhaven, Germany, 5 Research Unit Palaeontology, Ghent University, Ghent, Belgium, 6 Department of Earth Sciences, University of Toronto, Toronto, Ontario, Canada, 7 Department of Earth Sciences, Brock University, St. Catharines, Ontario, Canada, 8 Norwegian Geological Survey, Trondheim, Norway

AbstractThe early Late Pliocene (3.6 to ~3.0 m illion years ago) Is the last extended Interval In Earth's history when atmospheric C02 concentrations were comparable to today's and global climate was warmer. Yet a severe global glaciation during marine Isotope stage (MIS) M2 Interrupted this phase o f global warmth ~3.30 m illion years ago, and Is seen as a premature attem pt o f the climate system to establish an Ice-age world. Here we propose a conceptual model for the glaciation and déglaciation o f MIS M2 based on geochemical and palynologlcal records from five marine sediment cores along a Caribbean to eastern North Atlantic transect. Our records show that Increased Paclfic-to-Atlantlc flow via the Central American Seaway weakened the North Atlantic Current and attendant northward heat transport prior to MIS M2. The consequent cooling o f the northern high latitude oceans perm itted expansion o f the continental lee sheets during MIS M2, despite near-modern atmospheric C02 concentrations. Sea level drop during this glaciation halted the Inflow o f Pacific water to the Atlantic via the Central American Seaway, allowing the build-up o f a Caribbean Warm Pool. Once this warm pool was large enough, the Gulf Stream-North Atlantic Current system was relnvlgorated, leading to significant northward heat transport that terminated the glaciation. Before and after MIS M2, heat transport via the North Atlantic Current was crucial In maintaining warm climates comparable to those predicted for the end o f this century.

Citation: De Schepper S, Groeneveld J, Naafs BDA, Van Renterghem C, Hennissen J, et al. (2013) Northern Hemisphere Glaciation during the Globally Warm Early Late Pliocene. PLoS ONE 8(12): e81508. doi:10.1371/journal.pone.0081508

Editor: Victoria C Smith, University of Oxford, United Kingdom

Received May 7, 2013; Accepted October 14, 2013; Published December 12, 2013

Copyright: © 2013 De Schepper et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Funding: This research was supported by Deutsche Forschungsgemeinschaft grants SCHE 1665/2-1 and SCHE 1665/2-2 (SDS) and NA973/1-1 (BDAN), the University of Bergen (SDS), a MARUM Student Summer Fellowship (CVR), and a NSERC Canada Discovery Grant (MJH). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Com peting Interests: The authors have declared that no competing interests exist.

* E-mail: [email protected]

a Current address: Organic Geochemistry Unit, School of Chemistry, University of Bristol, Bristol, United Kingdom

IntroductionT h e early Late Pliocene (early Piacenzian) from 3.6 to

~ 3 .0 m illion years ago (Ma) is the last sustained interval in E arth ’s history w hen global clim ate was w arm er than today. T he ~ 3 .3—3.0 M a tim e slab known as the m id-P iacenzian W arm Period (mPW P, Figure 1) has been studied intensively as a potential analogue for our future global clim ate [1]. T h e m PW P is characterised by ~ 3 "C w arm er global tem peratures [2], 10M 0 m higher sea-level [3], reduced continental ice sheets [4], and an A tlantic m eridional overturning circulation (AM OC) com parable to [5] or stronger than [6] preindustrial levels. A tm ospheric C 0 2 concentrations w ere h igher th an preindustrial values, and likely as high as the m odern anthropogenic values o f ~ 4 0 0 ppm [7-9] (Figure 1C). T h e m PW P clim ate is a good approxim ation for the w arm clim atic conditions o f the entire early Late Pliocene. This w arm stable clim ate was nonetheless in terrup ted by a short-lived, intense global glaciation (3.305-3.285 Ma) during m arine isotope stage (MIS) M 2 [10,11] (Figure 1). In the L R 04 Plio-Pleistocene benthic Sl80 stack [10], M IS M 2 starts as a low-am plitude glaciation typical o f the Pliocene, bu t deepens steeply betw een 3.305 and 3.285 M a to reach values characteristic o f early

Q u a ternary glaciations. W e distinguish this b rie f interval o f intense glaciation (3.305-3.285 Ma) w ithin the longer interval o f M IS M 2 (3.312-3.264 Ma) as defined in LR 04 [10], T h e associated glacio- eustatic sea level drop is reflected in a m ajor depositional sequence boundary [12] with sea level estim ated at 10 m ± 10-15 m, 40 m ± 1 0 m, o r indeed up to 65 m ± 15-25 m below present [13-15] (Figure IB). G iven this large uncertain ty in reconstructed sea level for M IS M 2, it is difficult to quantify how the volum e of the no rthern and southern hem isphere ice sheets changed. LTsing the Holocene-like, relatively cool and dry Arctic clim ate at Lake El’gygytgyn (northeast Arctic Russia) as an approxim ation o f the b ro ad er Arctic clim ate, ice advance during M IS M 2 is thought to have occurred in Alaska, G reenland, Svalbard and A ntarctica, w hereas substantial expansion in N orth A m erica was less likely [16], Estim ates for ice volum e increase in A ntarctica correspond to a sea level drop o f ~ 8 m [17] or even ~ 1 8 m [13], bu t cannot not fully explain the ~0.5%o benthic foram inifera! S180 shift at this tim e [10], D irect and indirect evidence o f glaciation support expansion of the A ntarctic ice sheet [18,19], a considerable ice advance o f the G reen land an d Svalbard /B aren ts Sea ice sheets

PLOS ONE I www.plosone.org 1 December 2013 | Volume 8 | Issue 12 | e81508

mailto:[email protected]

http://www.plosone.org

Northern Hemisphere Glaciation in a Warm Climate

Age (Ma)2.40 2.50 2.60 2.70 2.80 2.90 3.00 3.10 3.20 3.30 3.40 3.50 3.60 3.70 3.80 3.90 4.00

Series/StagePMAG

Pleistocene Early Pliocene (Zanclean)Matuyama SjemmothJ ' ’■ Gilbert

/ LR04M2

MG6G ¡2 modern value

\refs. [13], [14], [15], [79]

eg /—o £O ra o. n

500400300200100403020100

pCO^anno 2012]

Preindustrial pCO

refs. [8] alkenone, [80] s to m ata , [8] and [9] boron

AMOC+

T T T T T~ T T T "2.40 2.50 2.60 2.70 2.80 2.90 3.00 3.10 3.20 3.30

Age (Ma)3.40 3.50 3.60

— I—

3.70

AMOC-~ r T

3.80 3.90 4.00

Figure 1. M arine isotope stage M2 in the long-term climate evolution of the Pliocene. (A) Time scale, including palaeom agnetic reversals (PMAG) and th e LR04 benthic iso tope stack [10], o range shading show s mid-Piacenzian Warm Period (= mid-Pliocene Warm Period), grey shading show s m arine isotope stage MIS M2; (B) sea level estim ates for th e Pliocene to Pleistocene [13-15,79]; (C) Late Pliocene atm ospheric carbon dioxide concentrations based on boron, alkenones and leaf stom ata [8,9,80]; (D) long-term carbonate-sand record a t ODP Site 999 as an indicator for Pacific w ater flow through th e Central American Seaway into th e Atlantic and AMOC [27], doi:10.1371/journal.pone.0081508.g001

[20-23], ice cap expansion in Iceland [24], and possibly in Alaska and the C anad ian Rocky M ountains [25] (Figure 2).

In terrup ting an interval o f global w arm th, M IS M 2 has been proposed as an early, failed a ttem pt by the E a rth ’s clim ate to establish a p a tte rn o f intense and frequent N orthern H em isphere glaciations [26,27]. It was no t until ~ 50 0 ,0 0 0 years later th a t this p a tte rn em erged, likely due to decreasing atm ospheric carbon dioxide concentrations during the Late Pliocene [8,28], T he decline in atm ospheric carbon dioxide concentrations [7-9], increasing global ice volum e [10,11], cooling o f ocean surface waters [29-31], and tectonic closure o f ocean gateways [27,32] since the Late M iocene m ay well have ultim ately facilitated glaciation in the late Late Pliocene, bu t these long-term processes are an unlikely cause o f the short-lived M IS M 2 glaciation. Similarly, variations in astronom ical forcing alone cannot explain the intense glaciation o f M IS M 2 because intervals w ith similar astronom ical forcing occurred th roughout the Late Pliocene w ithout leading to intense glaciation. T h e isolated na ture o f the M IS M 2 glaciation in the otherwise w arm clim ate o f the early Late Pliocene m ust be the result o f a specific forcing, unique w ithin this tim e period.

W e established high-resolution palynological and geochemical records from five ocean drilling sites along a southw est-northeast transect in the N orth A tlantic covering the C aribbean W arm Pool, G ulf Stream , subtropical gyre an d N orth A tlantic C urren t (NACI) over the interval 3 .400-3 .180 M a to determ ine the role o f ocean circulation in causing the extensive glaciation o f M IS M2 (Figure 2). O u r surface w ater mass, sea surface tem perature

(SST), relative salinity reconstructions, an d carbonate-sand records provide direct evidence that the unique conditions responsible for glaciation during M IS M 2 relate to an increased Pacific-to- A tlantic flow via the C entral A m erican Seaway (CAS) prio r to M IS M 2. This w eakened northw ard heat transport due to a shift o f the NACI. T h e conceptual m odel proposed here links an open CIAS with glaciation in the N orthern H em isphere and contrasts with hypotheses that propose the closure o f the CIAS as a cause for the intensification o f N orthern H em isphere glaciation a round 2.6 M a [33], o r as a delaying factor [34] or a p recondition for ice sheet expansion in the N orthern H em isphere [27].

Materials and MethodsSam ples were collected a t the IO D P B rem en Clore R epository

(Germany) an d G ulf Cloast R epository (College Station, Texas, LTSA) from five sites constituting a transect betw een the C aribbean Sea (O D P Site 999), w estern N orth Atlantic (DSDP Site 603), and the eastern N o rth Atlantic (DSDP Site 610, IO D P Sites FT 1308 and U1313). T h e foram inifera! geochem istry da ta and palyno- m orph assemblages were acquired from the same samples at each o f the five ocean drilling sites, and samples for biom arker (alkenone) analysis were taken from the same sample depths at three sites. All generated dinoflagellate cyst and geochem ical proxy d a ta are accessible th rough the database PA N G A EA at h t tp : / / do i.pangaea.de/10 .1594/P A N G A E A .804677 . Previously p u b lished M g/C ia an d dinoflagellate cyst da ta [26,35] are also available a t h ttp ://d o i.p an g aea .d e /1 0 .1 5 9 4 /P A N G A E A .7 5 8 7 1 0 and h ttp ://d o i.p an g aea .d e /1 0 .1 5 9 4 /P A N G A E A .7 5 8 7 1 1 .


http://doi.pangaea.de/10.1594/PANGAEA.758710

http://doi.pangaea.de/10.1594/PANGAEA.758711



90°W 60°W 30°W 0“

U1307subpolar gyre

8CTN

60°N

40°N

20°N

NAD

y r U 1 3 0 8

subtropical gyre

sed im en to log ica l ev id e n c e for g laciation a n d IRD during MIS M2

te rrestria l ev id e n c e for g laciation

Waterdepth

28 2 8 m 46 3 3 m 3 413 m 3 872 m 2 4 1 7 m

Coordinates

O D P 999 12 4 4 ’N, 7 8 4 4 WD S D P 6 0 3 3 5 a3 0 ’N, 70°2 ’WIODP U 1313 4 1 ‘0 ’N, 3 2 '5 7 ’W IODP U 1308 49°53’N, 24°14’W D S D P 6 1 0 5 3 '1 3 ’N, 18°53’W

G S = Gulf S treamNAC = North A tlantic C urren tNAD = North A tlantic Drift

S S T('C ) 0 5 10 15 20 25 30

Figure 2. M odem North Atlantic surface circulation with modern sea surface temperatures (World Ocean Atlas 2005 [81]). Each studied site is indicated by the same colour in subsequent figures, and other sites discussed in the text are shown in white, lee caps on Greenland are schematic representations of Pliocene reconstructions [4].doi:10.1371/journal.pone.0081508.g002

Dinoflagellate cyst preparation technique and assemblage interpretation

O u r laboratory technique allows dinoflagellate cysts and foram inifera to be extracted from the same samples (full details in [26]). E ach sample was first w et sieved at 125 pm to concentrate the foram inifera an d ensure that the palynom orphs pass through the sieve for further processing. T h e fraction re ta ined on the sieve (> 125 pm) was dried and weighed before being picked for foram inifera. T h e sedim ent filtrate (< 125 pm) was dried and weighed, and Lycopodium clavatum tablets were added before applying standard palynological p repara tion techniques involving cold HC1 and H F acids [36], No oxidation, alkali or ultrasonic treatm ents were used. O rganic residues were sieved th rough a 10- pm nylon m esh and strew m ounted onto m icroscope slides using glycerine jelly. D inoflagellate cysts were counted under 400x m agnification w ith counts varying betw een 44 an d 527 (average 267) specimens pe r sample. In addition, acritarchs and terrestrial palynom orphs were also enum erated during the dinoflagellate cyst counts. Palynom orph concentrations an d erro r estimates were then calculated based on the pa lynom orph an d Lycopodium clavatum counts and the dry weight o f the < 1 2 5 pm fraction [37]. O nly

relative abundance variations th a t are statistically significant according to the procedure described in ref. [38] have been used for interpretation . D a ta presented in refs. [26] an d [35] were used alongside our newly generated da ta (Figure 3).

T h e assemblage com position o f dinoflagellate cysts in core-top samples is largely related to the present-day overlying w ater masses [39^-2], and reflects the interplay betw een tem perature, salinity, nutrients, sea ice cover an d light availability. Present- day, last interglacial [43] and Pliocene [26] dinoflagellate cyst assemblages recovered from the eastern N orth A tlantic consisting o f high abundances o f Operculodinium centrocarpum sensu W all & Dale (1966) (herein 0. centrocarpum) all reflect the presence of the N orth Atlantic C urren t (NAC). A t D S D P Site 610 and IO D P Site U l 308, 0. centrocarpum concentrations are highest also when the relative abundances are high, independently corroborating the value o f 0 . centrocarpum as a NAC indicator species.

Full authorial citations o f species discussed in the text are given in T ab le 1.




full g lac ia l, c o n d itio n s

3.PMAG

O o

b ëCi) ^w 7c

® 2 2 -

18-14-

P « “ „ e I 2 2 :y o -2 -ioI a? % 18

H o ^ i4w d 14 cn

180

Age (Ma)3.340 3.360 3.380 3.400

LR04I 603 IIU1308II 610 IIU1313]

MG4

|U1308|

coring gap

o £ 18-16-

ï p 14-

§ g « 12-1 0 -

3.0-.

. 2 .0 -

' 1 . 0 -

0 -NAC

Warmwatertaxa

] Operculodinium centrocarpum I Invertocysta lacrymosa + tabulata ] Impagidinium solidum I Spiniferites mirabilis I Impagidinium patulum ] Impagidinium paradoxum ] Impagidinium aculeatum I Polysphaeridium zoharyi ] Pentapharsodinium dalei I Impagidinium pallidum ] Filisphaera filifera I Bitectatodinium tepikiense ] Nematosphaeropsis labyrinthus I Spiniferites spp.I Others

water

180 3.3203.200 3.220 3.240 3.260 3.340 3.360 3.380 3.400Age (Ma)

Figure 3. North Atlantic palaeoceanographic proxy records from DSDP Sites 603 and 610, and IODP Sites U1308 and U1313 between 3.400 and 3 .180 Ma. Circles with w hite fill are data poin ts from this study, circles with colour fill are from [30,55] (C) and [26,35] (D). Vertical grey bars represen t glacials, white bars are interglacials. (A) palaeom agnetic reversals; (B) benth ic isotope age m odels for each site tuned to the LR04 stack [10] (black line), thin coloured lines are data, thick coloured line is 4-point running m ean; (C) alkenone SSTs (0-10 m w ater depth) including calibration related error (shading), horizontal lines represent m odern average annual tem pera tu re at 0 -50 m w ater d ep th for each site; (D) SSTs a t 0 -60 m w ater d ep th based on Mg/Ca of G. bulloides, including a 1 C error bar (shading), horizontal lines represent m odern average spring (March-April-May) tem pera tu re a t 0 -50 m w ater d ep th for each site; (E) SWTMg/ca of G. inflata a t 300-400 m w ater dep th ; (F) calculated S18Osw.¡ce as estim ate of salinity; thick coloured lines are a 4-point running m ean; (G-J) dinoflagellate cyst assem blage com position. High abundances of 0. centrocarpum (yellow) indicate an active NAC. Bluish colours = cool-w ater species, reddish colours = w arm -w ater species; (G) and (H) contain data p resen ted in [26] and [35], (J) B. tepikiense and F. filifera are g rouped to g e th er a t DSDP Site 603 and are represented by th e colour for B. tepikiense (purple); Impagidinium cf. pallidum is represented as Impagidinium pallidum. doi:10.1371/journal.pone.0081508.g003




T a b le 1 . Dinoflagellate cyst species mentioned in the text and figures: abbreviation, full authorial citation and grouping.

Abbreviation Full species name

B. tepikiense Bitectatodinium tepikiense Wilson 1973

F. filifera Filisphaera filifera Bujak 1984 emend. Head 1994

1. aculeatum Impagidinium aculeatum (Wall 1967) Lentin & Williams 1981

1. paradoxum Impagidinium paradoxum (Wall 1967) Stover & Evitt 1978

1. pallidum Impagidinium pallidum Bujak 1984

1. patulum Impagidinium patulum (Wall 1967) Stover & Evitt 1978

1. solidum Impagidinium solidum Versteegh & Zevenboom in Versteegh 1995

1. lacrymosa Invertocysta lacrymosa Edwards 1984

1. tabulata Invertocysta tabulata Edwards 1984

N. labyrinthus Nematosphaeropsis labyrinthus (Ostenfeld 1903) Reid 1974

0. centrocarpum Operculodinium centrocarpum sensu Wall & Dale 1966

0. israelianum Operculodinium israelianum (Rossignol 1962) Wall 1967

P. dalei Cyst of Pentapharsodinium dalei Indelicato & Loeblich III 1986

P. zoharyi Polysphaeridium zoharyi (Rossignol 1962) Bujak et al. 1980

RBC Round brown cysts

S. mirabilis Spiniferites mirabilis (Rossignol 1964) Sarjeant 1970, and

Spiniferites hyperacanthus (Deflandre & Cookson 1955) Cookson & Eisenack 1974

Spiniferites/ Achomosphaera sp p. Spiniferites spp. Mantel I 1850, and

Achomosphaera spp. Evitt 1963

Others Contains all other dinoflagellate cyst taxa counted.

doi:10.1371 /journal.pone.0081508.t001

Geochemistry: 81S0 and Mg/Ca o f foraminifera, calculating and interpreting sea surface temperature and relative salinity (818Osw_ice)

Foram inifera were picked from the > 1 2 5 pm dry fraction of each sample. Planktonic foram inifera isotope da ta were m easured using a Finnigan M A T 251 mass spectrom eter at the Isotope L aboratory, Geosciences D epartm ent, LTniversity o f B rem en using

five specimens per sample o f Globigerina bulloides (250-315 pm) for Sites 603, 610, FT 1308 and LU 313 and five specimens pe r sample o f Globigerinoides sacculifer (250-355 pm) for Site 999. Benthic foram inifera! isotope da ta a re based on a t least one > 2 5 0 pm specim en of Cibicidoides wuellerstorfi or Uvigerina perigrina per sample. Cibicidoides wuellerstorfi S180 values have been corrected by adding 0.64 %o [44]. T h e standard deviation o f the analyses is based on an

SiteCore,

Hole type Section HalfTop(cm)

Bottom(cm)

Depth(mbsf)

Samplename

Inclinationn

Inc(low)

Inc(high) Comments

603 C 17 X 1 W 91 93 136.51 C17S1091 -8 .4 -20.8 4.8 Questionablereliability

T a b le 2 . Palaeomagnetic data for the reversal at the base o f the Mammoth Subchron in DSDP Hole 603C (depth Indicated In bold).

603 C 17 X 2 W 47 49 137.57 C17S2047 -19.3 -43.1 13.2 Questionablereliability

603 C 17 X 3 W 11 13 138.71 C17S3011 -32.9 -35.0 -30.7

603 C 17 X 3 W 42 44 139.02 C17S3042 -15.4 -18.1 -12.6

603 C 17 X 3 W 102 104 139.62 C17S3102 -50.0 -51.7 -48.3

603 C 17 X 3 w 107 109 139.67 C17S3107 -24.7 -29.3 -19.6

603 C 17 X 4 w 119 121 141.29 C17S4119 -19.7 -34.4 -1 .7 Questionablereliability

603 C 17 X 5 w 41 43 142.01 C17S5041 16.0 12.0 19.9

603 C 17 X 6 w 42 44 143.52 C17S6042 14.8 11.1 18.3

603 C 18 X 1 w 11 13 145.31 C18S1011 -22.9 -33.0 -11.0

603 C 18 X 1 w 99 101 146.19 C18S1099 28.6 27.1 30.1

603 C 18 X 1 w 117 119 146.37 C18S1117 39.6 36.5 42.5

The reversal at the top of the Mammoth Subchron could not be assigned.




in-house Solnhofen carbonate standard with a value o f 0.07%o. Values are reported relative to th a t o f the V ienna Pee Dee Belem nite (VPDB) calibrated using N ational B ureau o f S tandards (NBS) 18, 19, and 20 standards.

For M g /C a m easurem ents, we used 20-25 specimens per sample o f G. bulloides (250-315 pm) or G. sacculifer (250-355 pm) and 20 specimens pe r sample o f Globorotalia inflata (250-400 pm) from Sites 610, U 1308 and U 1313. T h e cleaning procedure for M g /C a m easurem ents is described elsewhere [45]. After dissolution in 0.5 m L 0.075 M Q D H N 0 3, the samples were centrifuged an d diluted for analysis on an IC P -O E S (Perkin E lm er O p tim a 3300R) a t the Geosciences D epartm ent, U niversity o f Brem en. T h e analytical precision o f the M g /C a analyses for G. bulloides, G. sacculifer, and G. inflata com bined was 0.17% (n = 459). R eproducibility based on replicate samples (n = 32) o f bo th G. bulloides and G. sacculifer was ±0 .11 m m o l/m o l (~3.3% ). T h e validity o f analyses was checked by analysing an artificial in- house standard to m onitor drift o f the IC P -O E S (M g /C a = 2.93 m m ol/m ol) and the lim estone standard E C R M 752-1 (M g/ C a = 3.75 m m ol/m ol) to allow inter-laboratory com parison [46]. A l /Ca, F e /C a , and M n /C a were sim ultaneously analysed with M g /C a to p revent con tam inated samples from being included in the interpretation . W e used the following calibration, established from core-top sedim ent samples in the N orth A dantic, to transform the foram inifera! M g /C a ratios o f G. bulloides into SS T Mg/ Ca: M g /C a = 0 .5 2 exp 0.10 T [47]. W e in terp ret the SSTMg/Ca value of G. bulloides as spring to sum m er SSTs of the upper 60 m o f the w ater colum n [26,48,49] because the oxygen isotope com position o f G. bulloides reflects the northw ard- m igrating phytoplankton spring b loom in the N orth A tlantic[47,50]. T h e SSTMg/Ca value of G. sacculifer represents the annual m ixed-layer tem perature o f the upper 75 m of the w ater colum n for C aribbean Site 999 [51]. M g /C a values were transform ed into palaeo-seaw ater tem peratures using the following equation: M g /C a = 0 .491 exp 0.033 T [52]. A lthough G. inflata calcifies th roughout the w ater colum n, the SW T Mg/ 0a based on mostly non-encrusted G. inflata represents the tem peratu re o f the perm an en t therm ocline [53]. W e used the following calibration to calculate tem peratures: M g /C a = 0 .7 2 exp 0.076 T . C om bining analytical an d calibration errors, we estim ate the erro r on M g /C a palaeo tem perature reconstruction for shallow- dwelling foram inifera as ± 1 .0 -1 .5 °C [47], w hereas for the deeper-dwelling G. inflata the erro r is estim ated to be ± 2 - 2 .5°C[53].

T h e oxygen isotope com position of seawater (8 O sw) was calculated via a standard form ula [54], Since M g /C a an d S180 were m easured on the same planktonic foram iniferal species, the possible effects o f seasonality and h ab ita t differences are m inim ised. W e used the LR 04 global benthic foram iniferal S180 stack [10] as an approxim ation for changes in ice volum e over the studied interval. After norm alizing the L R 04 record, we subtracted it from S18O sw, resulting in a 8 18O sw_;ce record th a t approxim ates local variations in salinity.

AlkenonesAll alkenone da ta from Sites 610 an d 1308 are new, whereas

da ta from IO D P Site U 1313 have been published earlier [30,55] (Figure 3C). T h e m odified alkenone unsaturation index U \-¡ [56,57] was m easured using a G C /T O F -M S system [58] on separate samples, taken from the same depths as those used for foram iniferal M g /C a and dinoflagellate cyst analyses. U \-¡ in com bination with a global core-top calibration was used to calculate annual m ean SST (top 10 m) [59]. T h e analytical technique, calibration and reliability o f alkenone-based SSTs for

the Pliocene is detailed elsewhere [30,55]. T h e calibration erro r on the alkenone SSTs is ~ 1 .5 °C [59].

T h e global core-top calibration gives the highest correlation with annual m ean SSTs, b u t locally alkenone-based SSTs could reflect the tem peratu re o f the growing season (spring in the N orth Atlantic) [60]. A lthough this affects the absolute SST estimates, it does no t influence the relative trends in our records. T he exception w ould be if the alkenone producers shifted their p roduction season on a glacial/interglacial basis, b u t there is no evidence for such behaviour. How ever, if such shifts did occur during glacials the alkenone producers w ould have delayed their p roduction towards sum m er to avoid the colder spring surface conditions. This implies th a t the cooling observed in the alkenone records during M IS M 2 w ould actually underestim ate the true cooling.

Carbonate sand fractionW e generated high-resolution carbonate sand fraction data

from O D P H ole 999A over the study interval. In addition, we used the available low resolution, long-term carbonate sand fraction record of the same site [27]. T h e sand conten t (> 63 pm) o f deep- sea carbonates is considered [27] a sensitive indicator o f changes in carbonate dissolution: sand con ten t (foraminifer tests) decreases as dissolution progresses. A low -carbonate sand fraction was in terp reted to reflect a poorly-ventilated deep C aribbean w ater mass. C arbonate dissolution a t Site 999, caused by entry o f A ntarctic In term ediate W ater (AAIW) into the C aribbean Basin in place o f N orth Atlantic D eep W ater (NADW), implies an open C entral A m erican Seaway and a weak overturning circulation [27].

Palaeomagnetic measurementsT h e positions o f m agnetic reversals for the M am m oth Subchron

in D S D P Holes 603C and 610A [61], and IO D P H ole 1308C [62] were re-m easured in this study to increase precision by analysing discrete, oriented samples a t 4 -2 3 cm resolution (Tables 2—41). O rien ted cubic polystyrene boxes (7.2 cm 3) w ere taken, avoiding visible m ineral concretions an d areas influenced by the coring process, from the working halves o f Sections 603C-17X 1 to 603C- 18x1 (12 samples betw een 136.51 and 146.37 mbsf), Sections 610A-17H 3 to 610A -17H 4 (17 samples betw een 156.74 and 158.75 mbsf), Sections 610A-17H 6 to 610A-18H1 (19 samples betw een 161.20 an d 163.85 mbsf), and Sections U 1308C -26H 2 to U 1308C -26H 5 (20 samples betw een 230.48 and 239.45 mbsf). T h e discrete samples were m easured a t the Geosciences D epartm ent, U niversity o f B rem en on a cryogenic m agnetom eter (model 2G Enterprises 755 H R). T h e natural rem anen t m agnetization (NRM ) was dem agnetized in nine steps (10-100 m T), and inclination and relative declination, an d their confidence intervals were determ ined from line fits o f straight-line segments in a Zijderveld diagram . N ote th a t absolute declination depends on section and core orientations, w hereas inclination values rely on the fact th a t the drill hole is to a very good approxim ation perpendicular. All ages o f the m agnetic reversals are according to the A T N T S 2004 [63],

In D S D P H ole 610A, the upper boundary o f the M am m oth Subchron was found betw een 158.35 m bsf (positive inclination) and 158.75 m bsf (negative inclination). T h e reversal a t the base o f the M am m oth Subchron is m ore difficult to identify due to a coring gap betw een Cores 610A -17H an d 610A-18H , and disturbed sedim ent in the upper 25 cm of Section 610A-18H1 [64], b u t m ust be located betw een 161.85 m bsf (negative inclination) an d 163.11 m bsf (positive inclination). T he reversals bounding the M am m oth Subchron in IO D P Site U l 308 are betw een 254.56 an d 255.46 m cd a t the top, and betw een 262.91




Table 3. Palaeomagnetic data for the reversals at the base and top o f the Mammoth Subchron in DSDP Hole 61OA.

Site HoleCore,type Section Half

Top(cm)

Bottom(cm)

Depth(mbsf)

Samplename

Inclinationn

Inc(low)

Inc(high) Comments

610 A 17 H 3 W 73 75 156.74 Mam-N1821 41.2 10.6 57.4

610 A 17 H 3 W 88 90 156.89 Mam-N1822 48.9 36.4 57.2

610 A 17 H 3 w 98 100 156.99 Mam-N1823 69.6 44.4 77.2

610 A 17 H 3 w 108 110 157.09 Mam-N1824 62.9 53.9 68.5

610 A 17 H 3 w 115 117 157.16 Mam-N1825 53 45.9 58.3

610 A 17 H 3 w 127 129 157.28 Mam-N1826 56.6 50.7 61.1

610 A 17 H 3 w 139 141 157.40 Mam-N1827 55.4 30.4 66.6

610 A 17 H 4 w 6 8 157.57 Mam-N1828 51.9 39.2 60

610 A 17 H 4 w 21 23 157.72 Mam-N1829 42.1 -42.5 69.8

610 A 17 H 4 w 26 28 157.77 Mam-N1830 45.4 29 55.8

610 A 17 H 4 w 36 38 157.87 Mam-N1831 49.8 34.3 59.4

610 A 17 H 4 w 52 54 158.03 Mam-N1832 78.2 71.2 81.5

610 A 17 H 4 w 69 71 158.20 Mam-N1833 48 40.1 54.1

610 A 17 H 4 w 84 86 158.35 Mam-N1834 45.3 25 57.3

610 A 17 H 4 w 92 94 158.43 Mam-N1835 22.7 2.5 38.4

610 A 17 H 4 w 114 116 158.65 Mam-N1836 -17.2 -4 0 12.4

610 A 17 H 4 w 124 126 158.75 Mam-N1837 -30.7 -41.8 -16.5

610 A 17 H 6 w 69 71 161.20 Mam-N1838 -66.5 -7 0 -61.6

610 A 17 H 6 w 88 90 161.39 Mam-N1839 -58.7 -61.6 -55.3

610 A 17 H 6 w 93 95 161.44 Mam-N1840 -50.2 -53.7 -46.1

610 A 17 H 6 w 101 103 161.52 Mam-N1841 -56.8 -58.9 -54.5

610 A 17 H 6 w 117 119 161.68 Mam-N1842 64.1 62.3 65.7

610 A 17 H 6 w 134 136 161.85 Mam-N1843 -39.2 -41 -37.3

610 A 17 H CC w 11 13 161.97 Mam-N1844 20.8 10.9 29.6

610 A 18 H 1 w 2 4 162.63 Mam-N1845 44.1 43 45.3 Disturbed?

610 A 18 H 1 w 8 10 162.69 Mam-N1846 -3 .7 -6.5 -1 Disturbed?

610 A 18 H 1 w 14 16 162.75 Mam-N1847 43.4 41.8 45 Disturbed?

610 A 18 H 1 w 23 25 162.84 Mam-N1848 7.1 6.1 8.1 Disturbed?

610 A 18 H 1 w 27 29 162.88 Mam-N1849 -20.6 -23.9 -17.1 Disturbed?

610 A 18 H 1 w 31 33 162.92 Mam-N1850 -71.3 -74.9 -65.6

610 A 18 H 1 w 41 43 163.02 Mam-N1851 -57.1 -59.4 -54.4

610 A 18 H 1 w 50 52 163.11 Mam-N1852 34.2 31 37.3

610 A 18 H 1 w 68 70 163.29 Mam-N1853 9.7 -8 .9 26.6

610 A 18 H 1 w 86 88 163.47 Mam-N1854 57.3 34.5 67.6

610 A 18 H 1 w 101 103 163.62 Mam-N1855 37.4 -12.4 60.2

610 A 18 H 1 w 124 126 163.85 Mam-N1856 17.5 4.7 28.7

Depths in bold demonstrate the position of the reversals.Note: The reversal at the base of the Mammoth Subchron is difficult to identify due to a core gap and potentially disturbed sediments in the upper 0-30 cm of Section 610A-18H1.

and 264.41 m cd at the bo ttom [62]. O u r re-assessment places the top o f the subchron betw een 254.59 an d 254.67 m cd, and the bottom betw een 263.21 and 263.46 m cd. T he offset betw een predicted an d m easured position in IO D P Site U l 308 is small and considered to be w ithin the m argin o f accuracy of the m ethods. At IO D P Site U l 313, the reported palaeom agnetic reversal o f the base o f the M am m oth Subchron lies a t 153.68 m cd ±0 .1 m [65], w ithin the glacial m axim um o f M IS M 2. T h e offset m ay be the result o f the field geom etry and distance betw een the sites, variation in the m agnetic lock-in tim e and dep th in the sediment,

and the sites used to determ ine the M am m oth Subchron in the G eom agnetic Polarity T im e Scale.

Age modelsA n age m odel was established for each hole (Figures S1-S4) by

tuning its benthic foram iniferal stable oxygen isotope record to the LR 04 benthic foram iniferal isotope stack [10], w ith the palaeo- m agnetic reversals as guidelines only, using the software p rogram AnalySeries 2.0.4.2 [66]. T he accuracy o f each age m odel depends on the accuracy of the LR 04 benthic stack w hich is estim ated a t 15




Table 4. Palaeomagnetic data for the reversals at the base and top o f the Mammoth Subchron in IODP Hole U1308C.

Core, Top Bottom Depth InclinationSite Hole type Section H alf (cm) (cm) (mbsf) Sam ple nam e (°) Inc (low) Inc (high) Comments

1308 C 25 H 6 W 7 9 230.48 Mam-N1857 44.7 35.1 51.9

1308 C 25 H 6 W 15 17 230.56 Mam-N1858 - 38.1 -42.7 - 32.8

1308 C 25 H 6 W 25 27 230.66 Mam-N1859 -76.6 -79.9 -70.5

1308 C 25 H 6 W 35 37 230.76 Mam-N1860 -6 2 -64.5 -59.1

1308 C 25 H 6 W 42 44 230.83 Mam-N1861 -5 7 -58.6 -55.2

1308 C 25 H 6 W 48 50 230.89 Mam-N1862 -51.8 -53.6 -49.8

1308 C 25 H 6 W 56 58 230.97 Mam-N1863 -60.6 -61.5 -59.6

1308 C 25 H 6 W 65 67 231.06 Mam-N1864 -6 9 -70.5 -67.3

1308 C 25 H 6 W 77 79 231.18 Mam-N1865 -54.8 -56.4 -53.1

1308 C 25 H 6 W 87 89 231.28 Mam-N1866 -72.4 -73.1 -71.6

1308 C 25 H 6 W 94 96 231.35 Mam-N1867 -6 7 -67.7 -66.3

1308 C 26 H 5 W 33 35 238.74 Mam-N1868 -51.7 -52.4 -51

1308 C 26 H 5 W 40 42 238.81 Mam-N1869 -75.4 -76.1 -74.6

1308 C 26 H 5 W 48 50 238.89 Mam-N1870 -54.8 -56.7 -52.8

1308 C 26 H 5 W 56 58 238.97 Mam-N1871 -57.9 -61 -54.2

1308 C 26 H 5 W 63 65 239.04 Mam-N1872 -66.5 -6 9 -63.5

1308 C 26 H 5 W 72 74 239.13 Mam-N1873 13 -31.1 46.8

1308 C 26 H 5 W 83 85 239.24 Mam-N1874 -42.9 -54.4 -24.8

1308 C 26 H 5 W 94 96 239.35 Mam-N1875 61 0.6 74.4

1308 C 26 H 5 W 104 106 239.45 Mam-N1876 65.6 62.9 67.9

Depths in bold demonstrate the position of the reversals.

kyr betw een 3 an d 4 M a [10] and the accuracy of the graphic correlation o f the benthic 8 1 8 0 records w ith the L R 04 tuning target. T h e tie-points used for the age m odels o f each hole and their correlation coefficients are p resented in Tables 5 -9 and Figures S1-S5. T h e ages in the Results an d Discussion section are reported w ith high precision (3 decimals) to dem onstrate (1) the relative age difference betw een samples w ithin one site, an d (2) the relative age of events in relation to the onset (~3 .315 Ma), full glaciation (~ 3 .305-3 .285 Ma), and term ination o f M IS M2

T a b le 5. Tie-points for the age model of DSDP Hole 610A.

Depth (mbsf) Age (Ma)

158.60 3.207

159.00 3.233

159.32 3.237

159.90 3.253

160.20 3.263

161.27 3.285

161.56 3.290

161.75 3.301

163.05 3.315

163.89 3.326

164.09 3.332

167.08 3.596

Note: mbsf = metres below sea floor. doi:10.1371 /journal.pone.0081508.t005

(~3 .285 Ma). These ages should neither be considered as absolute ages, nor as evidence for suborbital age control.

T h e original age m odel o f O D P Site 999 [27] is based on correlating the benthic S180 record to the astronom ically dated benthic S180 records from equatorial East Pacific O D P Site 846 [67] and equatorial East Atlantic O D P Site 659 [68] for the time interval 5 -2 M a. T h e existing age m odel was updated [51] to the newly generated, orbitally-tuned age m odel o f East Pacific IO D P Site 1241 [69] (Figure S5). For this study, we used the LR 04 benthic S180 stack [10] to fine-tune the glacial-interglacial transitions a round M IS M 2.

T a b le 6 . Tie-points for the age model of IODP Hole U1308C

Depth (mcd) Age (Ma)

254.66 3.207

259.27 3.265

261.11 3.284

262.35 3.302

263.25 3.320

263.38 3.327

264.70 3.340

Note: mcd = metres composite depth. doi:10.1371 /journal.pone.0081508.t006




T a b le 7 . Tie-points for the age model of IODP Site U1313(primary splice).

Depth (mcd) Age (Ma)

149.92 3.224

150.83 3.237

153.07 3.285

154.39 3.311

157.23 3.372

159.72 3.419

Note: mcd = metres composite depth.doi:10.1371 /journal.pone.0081508.t007

Results and Discussion

Events during interglacial MIS MG1 leading to glaciationT h e early Late Pliocene was w arm er th an today, and prio r to

~ 3.315 M a our geochem ical proxies and dinoflagellate cyst assemblages dem onstrate a surface circulation com parable to today’s bu t with elevated tem peratures in the high-latitude N orth Atlantic. W e record an active G ulf S tream over Site 603 as illustrated by the high SSTs (ca. 19.5"C) and the presence of 0. centrocarpum and such w arm w ater dinoflagellate cyst taxa as Impagidinium aculeatum, I. paradoxum, I. patulum, I. solidum, and Polysphaeridium, zoharyi w hich are also present there today [41], W arm (ca. 20"C) an d oligotrophic surface waters a t the subtropical gyre Site U 1313 are reflected in the dom inance of I. aculeatum, I. paradoxum, I. patulum, and Invertocysta spp. An active NAC brought w arm waters (15.2—18.6"C in the upperm ost 60 m) northw ard over Sites LT308 and 610 (Figure 3C, 3D, 3G , 3H), expressed in the dinoflagellate cyst assemblages by the dom inance o f 0. centrocarpum an d the persistent presence o f the w arm w ater species Spiniferites mirabilis. T h e less steep m eridional SST gradient com pared to present, especially visible in the SS T aik and to lesser extent in the SS T M?/Ca (Figure 3C, 3D, 4), indicates generally w arm er conditions in the h igher latitudes com pared to today.

A lthough deeper-w ater exchange via the CAS had been restricted since ~ 4 .6 M a [27], shallow Pacific-to-Atlantic exchange occurred well into the Late Pliocene [51]. This implies that A tlantic m eridional overturning circulation (AM OC) [6] was able to function even w hen the CAS was partially open. Following a m axim um in A M O C due to m inim al Pacific-to-Atlantic through-flow a round 3.6 M a, a gradual increase in through-flow via an open CAS culm inated im m ediately p rio r to M IS M2[27,51]. At C aribbean Site 999, we record betw een ~ 3 .3 2 0 and ~ 3.315 M a a drop in SST and salinity (818O sw_;cc) (Figure 5D, 5E), a low carbonate sand-fraction (Figure 5F), and high productivity evidenced by high dinoflagellate cyst concentrations dom inated by heterotrophic species (round b row n cysts; Figure 5G, 5F1). T h e low carbonate sand-fraction indicates a poorly ventilated deep C aribbean w ater mass and carbonate dissolution caused by entry o f A ntarctic In term ediate W ater (AAIW) into the C aribbean Basin in favour o f N orth Atlantic D eep W ater (NADW) - in terpreted as evidence o f a weak overturning circulation [2 7]. T h e drop in SST and salinity po in t to an increased inflow o f cooler, less saline Pacific waters to the C aribbean. Furtherm ore, the inferred high productivity is fully consistent w ith nutrient-rich waters from the Pacific entering the C aribbean [70], C onsidered altogether, this evidence shows that Pacific-to-Atlantic through-flow via the CAS during interglacial M IS M G1 exceeded a critical threshold, thereby reducing the A M O C . This was likely aided by the high

T a b le 8 . Tie-points for the age model o f DSDP Hole 603C.

Depth (mbsf) Age (Ma)

136.70 3.194

138.25 3.214

139.79 3.237

140.27 3.252

142.85 3.286

145.71 3.330

Note: mbsf = metres below sea floor. doi:10.1371 /journal.pone.0081508.t008

sea levels at that tim e [12,15] an d a longer-term gradual weakening o f the therm ohaline circulation since 3.6 M a [27] that b rought the clim ate system closer to a tipping point. P rior to M IS M G1 (Figure ID , 5C) high sea levels also occurred, bu t Pacific-to- A tlantic through-flow appears not to have w eakened the NAC, and glaciation in the N orthern H em isphere rem ained restricted.

D uring m axim al Pacific inflow via the open CAS during interglacial M IS M G1 (~ 3 .315-3 .320 Ma), contem poraneous changes occurred in the N o rth Atlantic surface circulation. At ~ 3.315 M a, a m ajor reduction in northw ard flow o f w arm NAC waters is reflected at Sites 610 and LU 308 by a m ajor tu rnover o f the dinoflagellate cyst assemblages w ithin 1-2 kyrs, and an initial cooling o f the surface waters is registered (Figure 3C, 3D, 3G , 3H). This corroborates m odelling studies showing that an open CAS results in a w eakened A M O C an d hence northw ard heat transport [71,72]. Between ~ 3 .3 1 5 an d 3.305 M a, a persistent NAC influence is recorded at Site LU 308 w hen SSTs decreased further and 0. centrocarpum (our NA C tracer) rem ained present in low abundance. At the same tim e we find a peak abundance o f 0. centrocarpum (Figure 31) and surface-water cooling (Figure 3D) at subtropical gyre Site LU 313. W e in terpret this sequence of events as an initial reduction in northw ard flowing w arm w ater o f the NAC at ~ 3.315 M a, followed by a gradual southw ard deflection o f the NAC betw een ~ 3.315 an d 3.305 M a. It is im portan t to note that the initial reduction in northw ard transport o f w arm NAC w ater occurred at ~ 3.315 M a w ithin the interglacial M IS M G 1, well before the M IS M 2 glacial m axim um at 3.295 M a.

T h e southw ard shift o f the NAC prio r to M IS M 2 led to the cessation of northw ard heat transport during the full glacial conditions o f M IS M 2 (3.305-3.285 Ma). D uring the glacial conditions, subtropical gyre circulation persisted as attested by a

T a b le 9 . Tie-points for the age model o f ODP Site 999.

Old age (Ma) New age (Ma)

3.205 3.205

3.239 3.245

3.276 3.280

3.296 3.295

3.319 3.320

3.342 3.340

3.355 3.365

3.371 3.375

Note: old age from ref. [78], new age from this study. doi:10.1371 /journal.pone.0081508.t009




610 U1308 U1313

U1308 U131313.2 16.3

Full glacial conditions 3.305-3.285 Ma

3.332-3.305 Ma

610 U1308 U1313J 18.6 ,^~ '1 î4 l ' ...... ]16||20.1|2ÖÖ 08

Temperature C 10 12 14 16 18 20

Figure 4. Reconstructed sea-surface temperatures for the studied intervals prior to, during and after the full glacial conditions of MIS M2 along an eastern North Atlantic transect through DSDP Site 610, and IODP Sites U1308 and U1313 from 6 0 N , 10 W to 3 0 N , 30 W. Background represents presen t day sea-w ater tem pera tu res o f the upper 500 m (from WOA2005, [81]). The insets below each site represen t th e surface w ater tem pera tu re based on alkenones (alk; surface w ater 0 -10 m), Mg/Ca ratios o f Globigerina bulloides (mgca; mixed layer 0 - 60 m) and sea-w ater tem pera tu re a t 3 0 0 ^ 0 0 m based on Mg/Ca ratios o f Globorotalia inflata (inf). The surface w aters (SSTa!k, SSTMg/Ca G. bulloides) show im portant cooling during MIS M2 with a s te ep N-S tem pera tu re grad ien t established. Low resolution sea-w ater tem pera tu re (SWTMg/Ca) reconstructions o f Globorotalia inflata show th a t w ater a t th e perm anen t therm ocline (300-400 m depth) rem ained stable and warm th roughou t the entire studied period. The glacial had no major effect on th e d eeper surface w aters, except possibly at Site U1308 w here one sam ple recorded tem pera tu res as low as 11.8 C during MIS M2. The average values o f 13.5, 14.2 and 15.1 C at Site 610, U1308 and U1313 respectively over th e entire period illustrate th a t th e entire upper w ater colum n during th e Pliocene was w arm er than today in th e North Atlantic. doi:10.1371/journal.pone.0081508.g004

continuance o f the G ulf S tream at Site 603 w here no m ajor changes in the dinoflagellate cyst assemblages were recorded (Figure 3J). T h e contrasting proxy evidence of cool surface SST alk [30], w arm er m ixed-layer SST M?/Ca and largely unchanged dinoflagellate cyst assemblages at subtropical gyre Site U l 313 is difficult to in terpret (Figure 3G, 3D, 31). D uring the earlier glacials M IS M G 4 and M G 2, bo th geochem ical proxies record a cooling, suggesting a fundam entally different oceanography for M IS M 2. It is no t know n w hether the divergence in SS T M?/Ca values during M IS M 2 was caused by different a genotype o f G. bulloides becom ing dom inan t in a changed oceanographic setting [73]. Irrespective o f the ultim ate cause, we consider the contrasting SST proxy records in com bination with the palynological da ta as evidence o f a southw ard shift o f the NAC that affected Site U 1313 during M IS M 2 bu t no t during prio r glacials.

Glaciation in the Northern Hemisphere during MIS M2A t the two n o rth ern sites, dinoflagellate cyst assemblages

indicate subpolar conditions (Bitectatodinium tepikiense, Filisphaera filifera, I. pallidum, Nematosphaeropsis labyrinthus, Pentapharsodinium dalei) at Site 610 an d oligotrophic conditions (I. aculeatum, I. paradoxum) at Site U l 308 (Figure 3G, 3H), while surface waters a t bo th sites cooled by 3M "C to tem peratures only ju st h igher th an today (Figure 3C, 3D). This cooling a t the northern sites established a steep latitudinal SST gradient in the N orth Atlantic (Figure 4), causing the therm al isolation o f G reen land from northw ard heat transport. As a com parison, a 3 -4 "C cooling o f the N ordic Seas was necessary for the last glacial inception (~ 115,000 years ago) in

Scandinavia [74], T h e increased m eridional SST gradient will have reduced air tem peratu re and increased snowfall over m ost o f N orth Am erica, bo th factors favourable to ice sheet inception [29]. W e dem onstrate th a t sufficiently cool surface waters were present in the n o rth ern high latitude oceans, and propose th a t these were crucial for the glaciation in the N orthern H em isphere during M IS M 2. T h e m oisture requ ired to build a large ice sheet in the N orthern H em isphere was presum ably already present in the atm osphere, because Pliocene climates were generally w etter than today [75]. It is nevertheless likely that after the southw ard shift o f the NA C and cooling o f the n o rth ern high-latitude surface waters, carbon cycle (vegetation, C 0 2) [28,71] an d perhaps sea ice (albedo) feedbacks also contributed to the m ajor glaciation during M IS M 2. Nevertheless, the extent o f N orthern H em isphere glaciation during M IS M 2 rem ained smaller than a typical Q u a ternary glaciation, bu t m ay have been larger than at present. T h e h igher 8 18O bcnthic values during M IS M 2 com pared to today (Figure 1A; 3.74%o vs. 3.23%o, [10]) indeed imply that M IS M 2 ice sheets were larger than today. Indirect evidence o f expanded ice sheets in the N orthern H em isphere is found in several sediment and ice-rafted debris records from the Arctic O cean, N ordic Seas and n o rthern N orth A tlantic [20-23] w hich indicate that the G reen land and Svalbard /B aren ts Sea ice sheets reached the coastline. Glacial deposits on Iceland [24] and possibly also in the C anad ian Rocky M ountains and Alaska [25] dem onstrate the presence of ice caps there. W ith SSTs approach ing present day values (Figure 3C, 3D) and a Holocene-like Arctic clim ate prevailing during M IS M 2 [16], the developm ent o f a significant




Age (Ma)3.

PMAG

3.0-O 3.2-ca> ~ 3.4-

o 1COCO

%)

£o 3 .8 -4 .0 -

20-

M am m oth3.285

fu ll glacial conditions

Site 999

— E - 2 0 - Wanganui Basin [12] Based on benthic isotope stack [13]

Mg/Ca of ostracoda ¡14]

Cooling starts to MIS M2

I - « 23Warming starts early in MIS M2

999 8180 modern

NADW

8 -5 20£ 15i - 03 13

O 1 1 0 -

NADW

q -Q oo

800

£ 4 0 0

Round brown cysts[ Impagidinium aculeatum Impagidinium paradoxum Impagidinium patulum

Polysphaeridium zoharyi Operculodinium israelianum Invertocysta lacrymosa Operculodinium centrocarpum Nematosphaeropsis labyrinthus Spiniferites spp.Others

k” wi k

Productivity

3.280i r

3.180 3.200n r-

3.2201 r-

3.2401 r

3.260 3.300 Age (Ma)

1 r-3.320

1 r-3.340

1 r-3.360

n 1------ 13.380 3.400

Figure 5. Caribbean Sea palaeoceanographic proxy records from ODP Site 999 between 3.400 and 3.180 Ma. Vertical grey bars represen t glacials, vertical white bars are interglacials. All circles with w hite fill are da ta from this study. (A) Palaeom agnetic reversals; (B) benthic isotope age m odel for Site 999 tuned to th e LR04 stack [10] (black line); (C) sea level estim ates; (D) SST-Mg/ca of G. sacculifer, thick black line represents 4-point running m ean, shading represents calibration error; (E) 818Osw.ice estim ate of salinity; (F) carbonate-sand fraction, filled black circles are data from [27], thick line represents 4-point running m ean. Lower (higher) values reflect decreased (in-creased) North Atlantic Deep W ater (NADW) influence a t the site. (G) dinoflagellate cyst assem blage com position; presence o f round brow n cysts indicate high productivity and inflow of Pacific w ater; (H) dinoflagellate cyst concentration, including error bar (light blue shading). doi:10.1371/journal.pone.0081508.g005

N orth A m erican ice sheet is unlikely. Therefore, to explain the observed ~0.5%o benthic isotope shift [10], a considerable expansion o f the A ntarctic ice sheet m ust have occurred also [16], Nevertheless, the possibility o f an ice cap in N orth Am erica

during M IS M 2 should not be excluded given the evidence o f an ice cap in the N orth A m erican interior that did no t reach the N orth Atlantic coastline at ~ 3 .5 M a [76], w hen glacials (e.g. M IS M G6) w ere less severe th an during M IS M 2 (Figure 1).




(1) Northward heat transport via NAC

AMOC AND NORTHWARD HEAT ! TRANSPORT VIA THE NAC

re-established after expansion and warming of CWP

Î -3 .2 8 5 Ma

GLACIALLY CLOSED PANAMA Sea-level drop halted inflow of cool

and fresh Pacific water into the Caribbean. From glacial maximum, the Caribbean Warm Pool builds up.

Pacific C aribbean

higher SST

P anam aIsthmus

Northward heat transport via active NAC. SSTs in mid- to high-latitude N. Atlantic

are 2 -3 'C higher than today. Mountain glaciation on Greenland. Central American Seaway is open.

Interglacial high sea level

PLIOCENE WARM CLIMATE

MIS M2 GLACIATION

Pacific C aribbean

cool, lower salinity water* high

W 1 productivity

P anam a \

/

MAXIMAL PACIFIC-TO- ATLANTIC TROUGH-FLOW via open Panam a Seaway

during interglacial MG1

-3 .3 1 5 Ma1REDUCTION IN NORTHWARD

FLOWING WARM WATER VIA NAC Southward shift of the NAC

No NAC, no northward heat transport.Mid- to high latitude Atlantic SSTs 3-4 "C cooler.

Glaciation of Greenland, Iceland, Svalbard- Barents Sea and questionably N. America.

Central American Seaw ay is closed.

A

Glacial lower se a level Cool high-latitude oceans

Figure 6. Conceptual model of glaciation and deglaciation of the Northern Hemisphere during MIS M2 in the otherwise globally warm early Late Pliocene. Numbers show sequence o f events. doi:10.1371/journal.pone.0081508.g006

Glacial closure o f the Central American Seaway led to deglaciation

O u r results further dem onstrate that the sea level drop a t the full glaciation o f M IS M 2 [12,15] closed the CAS an d effectively halted the inflow o f Pacific w ater into the Atlantic realm . T he oligotrophic conditions at Site 999 shown in the absence of heterotrophic dinoflagellate species an d low cyst concentrations (Figure 5G, 5H) during M IS M 2 suggest no inflow of nutrient- rich Pacific waters [70], T h e increasing SST and salinity (S18O sw_;cc) at Site 999 from the glacial m axim um at ~ 3 .2 9 5 M a onw ards show that C aribbean surface waters becam e w arm er an d m ore saline while rem aining oligotrophic (Figure 5D, 5E, 5G, 5H ). T his is a reflection o f the build-up of the C aribbean W arm Pool already from the glacial m axim um onwards. T h e expansion an d w arm ing o f the C aribbean W arm Pool are essential for re-establishing A M O C and northw ard heat transport, as observed for the last déglaciation [77]. At a round 3.285 M a, Site 999 is characterised by high SSTs and salinity, a re tu rn to biologically productive conditions, an d increased carbonate preservation (high carbonate sand-fraction) (Figure 5F). This indicates a C aribbean W arm Pool sufficiently large an d w arm to re-invigorate the A M O C . T h e re-established northw ard heat transport an d active NAC flowing along its m odern pathw ay a round 3.285 M a is reflected by the rapid turnover w ithin 1-2 kyrs o f the dinoflagellate cyst assemblages in the eastern N orth A tlantic Sites 610 and U l 308 w here 0. centrocarpum becom es dom inant again (Figure 3G , 3F1). By then, the w arm climates o f the m PW P [1] were established, with N orth A tlantic SSTs ~ 3 " C above present values (Figure 3C!, 3D), a m odern-like A M O C [5] bu t w ith a reduced m eridional

sea-surface tem perature gradient (Figure 4), an d a G reen land ice sheet that was reduced to isolated m ountain glaciers [4], As such, the glaciation during M IS M 2 appears responsible for its own demise.

ConclusionsO u r study identifies links betw een CAS through-flow , NAC

variability, high latitude sea-surface tem peratures, an d N orthern Flem isphere glaciation during Late Pliocene M IS M 2 (~ 3 .30 Ma). W e provide a conceptual m odel based on palynological and geochem ical records in the N orth A tlantic an d C aribbean for glacial expansion and consequent déglaciation during an otherwise globally w arm er w orld (Figure 6).

A long-term global cooling tren d (reflected in SST, C 0 2, and ice volum e records; Figure 1) p reconditioned the N orthern Flem isphere for glaciation during the early Late Pliocene. Flowever, the ultim ate tipping po in t for intense glaciation during M IS M 2 was the through-flow o f Pacific w ater via an open CAS into the A tlantic, ultim ately resulting in a steep SST gradient in the N orth Atlantic and therm al isolation o f the high latitudes. An open CAS as the trigger for N orthern Flem isphere glaciation contrasts w ith the usually invoked CAS closure as either the cause, precondition, o r delaying factor for the intensification of N orthern Flem isphere glaciation w hich occurred 500,000 years later [27,33,34], R ecent m odelling experim ents indicate that the closure o f the CAS actually had no effect on the Late Pliocene G reen land ice sheet, an d dem onstrate that declining atm ospheric carbon dioxide concentrations were the driving factor beh ind the intensification o f N orthern Flem isphere glaciation a t ~ 2 .7 5 M a [28,71],




O u r records in fact dem onstrate th a t the glacio-eustatic closure o f the CAS during M IS M 2 eventually re-established northw ard h eat transport in the N orth A dantic. Following the expansion of the A ntarctic an d N orthern H em isphere ice sheets (including G reenland, Iceland, Svalbard /B aren ts region an d questionably the interior o f N o rth America) to a volum e seemingly larger than present, sea level fell to m ore th an 10 m an d possibly as m uch as 65 m below present (Figure IB). T his closed the CAS and halted the flow o f Pacific w ater into the N orth Atlantic, allowing the C aribbean W arm Pool to accum ulate. In tim e, this re-invigorated the G ulf S tream /N o rth A tlantic C u rren t system and provided northw ard h eat transport, leading to high-latitude N orth A tlantic surface waters th a t were 3°C w arm er th an present and consequent re treat o f the G reen land ice sheet to m ountainous areas in the east and southeast during the m PW P.

T h e transition from M IS M 2 to the m PW P can be seen as the evolution of a w orld with com parable global tem peratures to p resent and slightly larger ice sheets, to a w orld with global tem peratures ~ 3 °C higher than today an d glaciation strongly dim inished and localised in the N orthern H em isphere. A lthough operating on a longer tim e scale, this clim ate transition can provide valuable insights into the p resent anthropogenically-forced clim ate transition towards a globally w arm er planet, being com parable to projections for the end o f this century. In view of this projected clim ate w arm ing, our results from the Late Pliocene show th a t high-latitude N orth A tlantic surface circulation and SSTs are a crucial factor in the expansion an d contraction of N orthern H em isphere ice sheets.

Supporting InformationFigure SI Age m o d e l for D SD P H ole 610A b a sed on the correla tion o f oxygen iso to p e reco rd s from the stu d ied in terva ls w ith the LR04 b en th ic oxygen iso to p e global stack [10]. Left panel: core-sections, polarity subchrons, including uncertain ty interval for the exact position o f each reversal o f the M am m oth Subchron, and benthic isotope record against depth (mbsf). Middle panel: correlation o f the benthic record (thin red line, raw data; thick red line, 4-point runn ing m ean) to the L R 04 global stack of benth ic isotope records [10] plo tted against time. G rey shading represents the m arine isotope stage boundaries from [10]: m arine isotope stage M 2 was defined betw een 3.264 and 3.312 M a. W e consider the full glaciation to occur betw een 3.305 and 3.385 M a (light grey). T h in black lines betw een left and m iddle panel show the tie points used (listed in inset). Right panel: sedim entation rate based on our age m odel. Inset gives the tie

References1. D o w se tt H J, R o b in so n M M , H a y w o o d A M , H ill D J, D o la n A M , e t al. (2012)

A ssessing co n fid en ce in P liocene sea surface tem p e ra tu re s to eva lu a te p red ic tiv e m odels. N a t G lim C h a n g 2: 1-7 . d o i:1 0 .1 0 3 8 /n c lim a te l4 5 5 .

2. H a y w o o d A M , V aldes PJ (2004) M o d e llin g P liocene w a rm th : co n tr ib u tio n o f a tm o sp h e re , o cean s a n d cryosphere . E a r th P la n e t Sei L e tt 218: 363—377. d o i:1 0 .1 0 1 6 /S 0 0 12 -8 2 1X{03)00685-X .

3. R ay m o M E , M itro v ica J X , O ’L eary M J, D eco n to R M , H e a r ty PJ (2011) D e p a r tu re s f ro m eustasy in P liocene sea-level records. N a t G eosci 4: 328—332. d o i:1 0 .1 0 3 8 /n g e o l 118.

4. D o la n A M , H a y w o o d A M , H ill D J, D o w se tt H J, H u n te r SJ, e t al. (2011) S ensitivity o f P liocene ice sheets to o rb ita l forcing. P a laeo g eo g r P a laeo c lim ato l P a laeo eco l 309: 9 8 -1 1 0 . d o i:1 0 .1 0 1 6 /j.p a lae o .2 0 1 1 .0 3 .0 3 0 .

5. Z h a n g Z S , N isanciog lu K H , C h a n d le r M A , H a y w o o d A M , O tto -B lie sn e r BL, e t al. (2013) M id -P lio cen e A d a n tic m erid io n a l o v e rtu rn in g c ircu la tio n n o t unlike m o d ern . G lim P a s t 9: 1 4 9 5 -1 5 0 4 . doi: 1 0 .5 1 9 4 /c p -9 -1495-2013.

6 . R ay m o M E , G ra n t B, H o ro w itz M , R a u G H (1996) M id -P lio cen e w arm th : s tro n g er g reen h o u se a n d s tro n g er conveyor. M a r M ic ro p a leo n to l 27: 313—326. doi: 1 0 .1 0 16/0377-8398{95)00048-8 .

7. P a g a n i M , L iu Z , L aR iv ie re J , R av e lo A C (2009) H ig h E arth -sy stem c lim ate sensitivity d e te rm in e d f ro m P liocene c a rb o n d iox ide c o n cen tra tio n s. N a t G eosci 3: 2 7 -3 0 . d o i:1 0 .1 0 3 8 /n g e o 7 2 4 .

points used, an d correlation values o f the benthic record running m ean and raw da ta with the LR 04 global stack. Note: H ole 610A shows a coring gap betw een Cores 610A -17H and 610A-18H , and sedim ent d isturbance in the upper 25 cm o f Section 610A-18H 1. (TIF)

Figure S2 Age m o d e l for IO D P Site U 1308 b a sed on the correla tion o f oxygen iso to p e reco rd s from the stu d ied in terva ls w ith th e LR04 b en th ic oxygen iso to p e g lobal stack [10]. Left, middle and right panel and inset as for Figure S I. (TIF)

Figure S3 Age m o d e l for IO D P Site U 1313 b a sed on the correla tion o f oxygen iso to p e reco rd s from the stu d ied in terva ls w ith th e LR04 b en th ic oxygen iso to p e g lobal stack [10]. Left, middle and right panel and inset as for Figure S I. (TIF)

Figure S4 Age m o d e l for D SD P Site 603 b a sed on the correla tion o f oxygen iso to p e reco rd s from the stu d ied in terva ls and p a la eo m a g n etic rev ersa ls w ith the LR04 b en th ic oxygen iso to p e g lobal stack [10]. Left, middle and right panel and inset as for Figure S 1.(TIF)

Figure S5 Show n on the le f t are the b en th ic 5ls O g lobal LR04 stack [10] com p ared to th e b en th ic 5ls O record o f IO D P Site 1241 [69]. O n the right, the LR 04 global stack is com pared to the old [51] and new (this study) benthic S180 curve o f O D P Site 999. T h e latter is a fine-tuning o f the [51] record to the L R 04 stack.(TIF)

AcknowledgmentsM. Segl is thanked for isotope measurements, S. Pape for assistance with the ICP-O ES, and M. Hoins, S. Forke and J. Engelke for assistance in the lab (MARUM and University of Bremen). Samples were supplied by the Integrated Ocean Drilling Program. W e thank Alan M. Haywood and one anonymous reviewer for the constructive comments.

Author ContributionsConceived and designed the experiments: SDS JG BDAN. Performed the experiments: SDS JG BDAN C V R JH KF. Analyzed the data: SDS JG BDAN C V R J H M JH SL KF. Contributed reagents/m aterials/analysis tools: SDS JG BDAN M JH SL KF. W rote the paper: SDS JG BDAN C V R J H M JH SL KF.

8 . Seki O , F oster G L , S ch m id t D N , M ack en sen A , K a w a m u ra K , e t al. (2010) A lkenone a n d b o ro n -b a se d P liocene pC C V records. E a r th P la n e t Sei L e tt 292: 2 0 1 -2 1 1 . doi: 1 0 .1 0 1 6 /j.e p s l.2 0 1 0.01 .037 .

9. B arto li G , H ö n isc h B, Z eeb e R E (2011) A tm o sp h eric C O 2 dec line d u r in g the P liocene in tensifica tion o f N o r th e rn H e m isp h e re g laciations. P a leo cean o g rap h y 26: PA 4213 . d o i:1 0 .1 0 2 9 /2 0 1 0 P A 0 0 2 0 5 5 .

10. Lisiecki L E , R a y m o M E (2005) A P liocene-P le istocene stack o f 57 globally d is tr ib u ted b e n th ic S 180 records. P a leo cean o g rap h y 20: P A 1003. d o i:1 0 .1 0 2 9 / 2004PA 001071 .

11. M udelsee M , R a y m o M E (2005) Slow d ynam ics o f th e N o r th e rn H em isp h ere g lac ia tion . P a leo c e a n o g ra p h y 20: P A 4022. d o i:1 0 .1 0 2 9 /2 0 0 5 P A 0 0 1 1 5 3 .

12. M ille r K G , K o m in z M A , B ro w n in g JV , W rig h tJ D , M o u n ta in G S , e t al. (2005) T h e P h a n e ro z o ic rec o rd o f g lobal sea-level change. Science 310: 1293—1298. doi: 1 0 .112 6 /sc ie n ce . 1116412.

13. N a ish T R , W ilson G S (2009) C o n s tra in ts o n the a m p litu d e o f M id-P liocene {3.6—2.4 M a) eu sta tic sea-level flu c tu atio n s f ro m th e N e w Z e a la n d shallow - m arin e sed im en t reco rd . P h il T ra n s A M a th Phys E n g Sei 367: 169—187. d o i:1 0 .1 0 9 8 /rs ta . 2008.0223.

14. M ille r K G , W rig h t J D , B ro w n in g JV , K u lp ecz A, K o m in z M , e t al. (2012) H ig h tid e o f th e w a rm P liocene: Im p lica tio n s o f g lobal sea level for A n tarc tic d eg lac iad o n . G eo logy 40: 4 0 7 M 1 0 . d o i:1 0 .1 1 3 0 /G 3 2 8 6 9 .1 .




15. D w y er G S , C h a n d le r M A (2009) M id -P lio cen e sea level a n d co n tin e n ta l ice vo lu m e b a sed o n co u p led b e n th ic M g /C a p a la e o tem p e ra tu re s a n d oxygen isotopes. P h il T ra n s A M a th Phys E n g Sei 367: 1 5 7 -1 6 8 . doi: 1 0 .1 0 9 8 / rsta. 2008.0222.

16. B r ig h am -G re tte J , M elles M , M in y u k P, A n d reev A , T a raso v P , e t al. (2013) P liocene w a rm th , p o la r am plification , a n d s tep p ed P leistocene coo ling rec o rd e d in N E A rc tic R ussia. Science 340, 1 4 2 1 -1 4 2 7 . doi: 1 0 .1 1 2 6 /sc ien ce . 1233137.

17. P o lla rd D , D e C o n to R M (2009) M o d e llin g W est A n ta rc tic ice sh eet g ro w th a n d collapse th ro u g h th e p a s t five m illion years. N a tu re 458: 3 2 9 -3 3 3 . d o i:1 0 .1 0 3 8 / n a tu re0 7 8 0 9 .

18. N a ish T , Pow ell R , L evy R , W ilson G , S ch e re r R , e t al. (2009) O b liq u ity -p a c e d P lio c e n e W e st A n ta rc t ic ice s h e e t o sc illa tio n s . N a tu re 45 8 : 3 2 2 —328. doi: 1 0 .1 0 3 8 /n a tu re0 7 8 6 7 .

19. Passch ier S (2011) L inkages b e tw e en E ast A n ta rc tic le e S h ee t e x te n t a n d S o u th e rn O c e a n tem p e ra tu re s b a sed o n a P liocene h ig h -re so lu tio n rec o rd o f ice- ra fted deb ris o ff P ry d z Bay, E ast A n ta rc tica . P a leo c e a n o g ra p h y 26: PA 4204. doi: 1 0 .1 0 2 9 /2 0 10PA 002061.

20. K le iv en H F , J a n se n E , F ro n v a l T , S m ith T M (2002) In ten sif icad o n o f N o r th e rn H e m isp h e re g laciations in th e c ircu m A tlan tic reg io n {3.5—2.4 M a) — ice-ra fted d e tr itu s ev idence. P a laeo g eo g r P a laeo c lim a to l P a laeo eco l 184: 213—223. doi: 1 0 .1016 /S 0031-0182{01)00407-2

21. K n ies J , M a tth ie s s e n J , V o g t C , L a b e rg J S (2009) T h e P lio-P leistocene g laciation o f th e B a re n ts S ea—S v a lb a rd reg io n : a n e w m o d e l b a s e d o n rev ised c h ro n o s tra tig rap h y . Q u a t Sei R ev 28: 8 1 2 -8 2 9 . d o i:1 0 .1 0 1 6 /j.q u asc irev .2 0 0 8 . 1 2 .0 0 2 .

22. M o ra n K , B ack m an J , B rinkhu is H , C lem en s S C , C ro n in T , e t al. (2006) T h e C en o zo ic p a la e o e n v iro n m en t o f th e A rc tic O c e a n . N a tu re 441: 60 1 —605. doi: 1 0 .1 0 3 8 /n a tu re0 4 8 0 0 .

23. S a rn th e in M , B arto li G , P ra n g e M , S c h m ittn e r A , S ch n e id e r B, e t al. (2009) M id -P lio c e n e shifts in o c e a n o v e r tu rn in g c irc u la tio n a n d th e o n se t o f Q u a te rn a ry -s ty le clim ates. C lim P a s t 5: 2 6 9 -2 8 3 . doi: 1 0 .5 1 9 4 /cp-5 -269-2009 .

24. G e irsd ó ttir A (2011) P liocene a n d P leistocene g laciations o f Iceland : a b r ie f o verv iew o f th e g lacial h istory. In : E h lers J , G ib b a rd P L , H u g h e s P D , ed itors. D ev e lo p m en ts in Q u a te rn a ry S cience, V ol. 15. A m ste rd am : Elsevier. 199—210. doi: 1 0 .1 0 1 6 /B 9 7 8 -0 -4 4 4 -5 3 4 4 7 -7 .0 0 0 1 6 -7 .

25. B a re n d re g t R W , D u k -R o d k in A {2011) C h ro n o lo g y a n d e x te n t o f L a te C en o zo ic ice sheets in N o r th A m erica : a m ag n e to s tra tig ra p h ic a l assessm ent. In : E h lers J , G ib b a rd P L , H u g h es P D , ed ito rs. D ev elo p m en ts in Q u a te rn a ry S cience, V ol.15. A m ste rd am : E lsevier. 4 1 9 M 2 6 . d o i:1 0 .1 0 1 6 /B 9 7 8 -0 -4 4 4 -5 3 4 4 7 -7 .0 0 0 3 2 -5 .

26. D e S c h e p p e r S, H e a d M J, G ro en ev e ld J {2009) N o r th A d a n tic C u rre n t variab ility th ro u g h m arin e iso tope s tage M 2 {circa 3.3 M a) d u r in g th e m id- P liocene. P a leo cean o g rap h y 24: P A 4206. d o i:1 0 .1 0 2 9 /2 0 0 8 P A 0 0 1 7 2 5 .

27. H a u g G H , T ie d e m a n n R {1998) E ffect o f th e fo rm a tio n o f th e Isth m u s o f P a n a m a o n A d a n tic O c e a n th erm o h alin e c ircu lad o n . N a tu re 393: 67 3 —676. d o i:1 0 .1 0 3 8 /3 1 4 4 7 .

28. L u n t D J, F oster G , H a y w o o d A M , S to n e E {2008) L ate P liocene G re e n la n d g lac ia tio n co n tro lled b y a decline in a tm o sp h eric C O 2 levels. N a tu re 454: 1102— 1106. doi: 1 0 .1 0 3 8 /n a tu re0 7 2 2 3 .

29. B rierley C M , F ed o ro v A V {2010) R ela tive im p o rta n c e o f m erid io n a l a n d zonal sea surface te m p e ra tu re g rad ien ts for th e o n se t o f th e ice ages a n d P lio cen es P le istocene c lim ate evolu tion . P a leo c e a n o g ra p h y 25: P A 2214. d o i:1 0 .1 0 2 9 / 2009PA 001809 .

30. N aafs B D A , S tein R , H e fte r J , K hélifi N , D e S ch ep p er S, e t al. {2010) L ate P liocene c h an g es in th e N o r th A d a n tic C u rre n t. E a r th P la n e t Sei L e tt 298: 434— 442. d o i:1 0 .1 0 1 6 /j.ep s l.2 0 1 0 .0 8 .0 2 3 .

31. L aw ren ce K T , S osd ian S, W h ite H E , R o se n th a l Y {2010) N o r th A d a n d c c lim ate ev o lu d o n th ro u g h th e P lio -P leistocene c lim ate transitions. E a r th P la n e t Sei L e tt 300: 3 2 9 -3 4 2 . d o i:1 0 .1 0 1 6 /j.e p s l.2 0 1 0 .10.013.

32. C a n e M A , M o ln a r P {2001) C lo sin g o f th e In d o n es ia n seaw ay as a p rec u rso r to east A frican a rid ificad o n a ro u n d 3—f | m illion y ea rs ago. N a tu re 411: 157—162. doi: 1 0 .1 0 3 8 /3 5 0 7 5 5 0 0 .

33. D risco ll N W , H a u g G H {1998) A sh o rt c ircu it in th e rm o h a lin e c ircu la tion : a cause fo r N o r th e rn H e m isp h e re g lac iadon? Science 282: 4 3 6 —f3 8 . doi: 1 0 .1 1 2 6 / sc ience .282 .5388 .436 .

34. B erger W , W efe r G {1996) E x p ed id o n s in to th e Past: P a leo c e a n o g ra p h ic studies in the S o u th A d an tic . In W efer G , B erg er W H , S ied ler G , W e b b D J, ed itors. T h e S o u th A dan tic : P re se n t a n d P ast C ircu la tio n . B erlin , H eid e lb erg : Springer- V erlag . 363—410. doi: 1 0 .1 0 0 7 /9 7 8 -3 -6 4 2 -8 0 3 5 3 -6 _ 2 1 .

35. D e S ch ep p er S, F ischer E l , G ro en ev e ld J , H e a d M J, M a t th ie s s e n J {2011) D e c ip h e r in g th e p a la e o e c o lo g y o f L a te P lio c e n e a n d E a r ly P le is to cen e d in o fla g e lla te cysts. P a la e o g e o g r P a la e o c lim a to l P a lae o e c o l 309: 17—32. d o i:1 0 .1 0 1 6 /j .p a la e o .2 0 1 1.04.020.

36. D e S c h e p p e r S, H e a d M J {2008) N e w d inoflagellate cyst a n d a c rita rc h tax a fro m th e P liocene a n d P le istocene o f the ea ste rn N o r th A d a n d c {DSD P S ite 610).

J Syst P a leo n t 6 : 10 1 -1 1 7 . d o i:1 0 .1 0 1 7 /S 1 4 7 7 2 0 1 9 0 7 0 0 2 1 6 7 .37. S to ck m arr J {1971) T ab le ts w ith spores used in ab so lu te p o llen analysis. P o llen e t

Spores 13: 6 1 5 -6 2 1 .38. H eslo p D , D e S ch ep p er S, P roske U {2011) D iag n o sin g th e u n c e rta in ty o f tax a

re lad v e ab u n d a n ce s d e riv ed f ro m c o u n t d a ta . M a r M ic ro p a leo n to l 79: 114—120. doi: 1 0 .1 0 1 6 /j.m a rm ic ro . 2011 .01 .007 .

39. H a r la n d R {1983) D is tr ib u tio n m ap s o f R e c e n t d inoflagellate cysts in b o tto m sed im en ts f ro m th e N o r th A d a n d c O c e a n a n d a d ja c e n t seas. P a laeo n to lo g y 26: 3 2 1 -3 8 7 .

40. R o c h o n A , d e V e rn a l A , T u r o n J -L , M a tth ie s se n J , H e a d M J {1999) D is tr ib u tio n o f rec e n t d inoflagellate cysts in surface sed im en ts f ro m th e N o r th A d a n tic O c e a n a n d a d ja c e n t seas in re la d o n to sea-surface p a ra m e te rs . A A S P C o n tr ib u d o n s Series 35. D allas, T exas: A A SP F o u n d a d o n . 146.

41. Z o n n e v e ld K A F , M a rre t F, V ers teeg h G JM , Bogus K , B o n n e t S, e t al. {2012) A d as o f m o d e rn d inoflagellate cyst d istr ib u tio n b a sed o n 2405 d a tap o in ts . R ev P a lae b o t Palynol. d o i:1 0 .1 0 1 6 /j.re v p a lb o .2 0 1 2 .0 8 .0 0 3 .

42. R a d i T , d e V e rn a l A {2008) D inocysts as p ro x y o f p r im a ry p ro d u ctiv ity in m id - h ig h la titu d es o f th e N o r th e rn H em isp h e re . M a r M ic ro p a leo n to l 6 8 : 84—114. doi: 1 0 .1 0 1 6 /j.m a rm ic ro . 2008 .01 .012 .

43. E y n a u d F, T u r o n J -L , D u p r a tJ {2004) C o m p a riso n o f the H o lo cen e a n d E em ian p a la e o e n v iro n m en ts in th e S o u th Ice lan d ic Basin: d ino flagellate cysts as p rox ies for th e N o r th A d a n d c surface c ircu lad o n . R ev P a lae b o t P a lyno l 128: 55—79. doi: 10 .1016 /S 0034-6667{03)00112-X .

44. S h ack le ton N J, H a ii M A {1984) O x y g en a n d c a rb o n iso tope s tra tig rap h y o f D e e p Sea D rillin g P ro je c t H o le 552A : P lio -P leistocene g lacial h istory. D S D P In it R e p 81: 5 9 9 -6 0 9 . d o i:1 0 .2 9 7 3 /d sd p .p ro c .8 1 .1 16.1984.

45. B ark er S, G reav es M , E lderfie ld H {2003) A stu d y o f c lean in g p ro ce d u re s used for fo ram in ife ra l M g /C a p a le o th e rm o m e try . G eo ch em G eophys G eosyst 4: 8407 , d o i:1 0 .1 0 2 9 /2 0 0 3 G C 0 0 0 5 5 9 .

46. G reaves M , C a illo n N , R e b a u b ie r H , B arto li G , B o h arty S, e t al. {2008) In te r la b o ra to ry c o m p ariso n stu d y o f ca lib ra tio n s ta n d a rd s fo r fo ram in ife ra l M g / C a th e rm o m e try . G e o c h e m G e o p h y s G eosyst 9: Q 0 8 0 1 0 . d o i:1 0 .1 0 2 9 / 2008 G G 0 0 1 9 7 4 .

47. E lderfie ld H , G an ssen G {2000) P ast te m p e ra tu re a n d S180 o f su rface o cean w aters in fe rred fro m fo ram in ife ra l M g /C a ratio s. N a tu re 405: 442—445. doi: 1 0 .1 0 3 8 /3 5 0 1 3 0 3 3 .

48. S ch iebe l R , B ijm a J , H em le b e n C {1997) P o p u la tio n d y n am ics o f th e p lan k d c fo ram in ife r Globigerina bulloides f ro m th e ea ste rn N o r th A d an d c . D eep Sea R es I 44{9-10): 1 7 0 1 -1 7 1 3 . do i:10 .1016 /S 0 9 6 7 -0 6 3 7 {9 7 )0 0 0 3 6 -8 .

49. C h a p m a n M R {2010) S easonal p ro d u c tio n p a tte rn s o f p lan k to n ic fo ram in ife ra in th e N E A d a n tic O c e a n : im plica tions for p a le o te m p e ra tu re a n d h y d ro g rap h ic reco n stru c tio n s . P a leo cean o g rap h y 25: P A I 101. d o i:1 0 .1 0 2 9 /2 0 0 8 P A 0 0 1 7 0 8 .

50. G an ssen G M , K ro o n D {2000) T h e iso top ic s ig n a tu re o f p lan k to n ic fo ram in ife ra f ro m N E A d a n tic su rface sedim ents: im plica tions fo r th e rec o n s tru c tio n o f p a s t ocean ic cond itions. J G eo l S oc L o n d 157: 693—699. doi: 1 0 .1 1 4 4 /jg s .l5 7 .3 .6 9 3 .

51. S tep h S, T ie d e m a n n R , P ra n g e M , G ro en ev e ld J , N ü rn b e rg D , e t al. {2006) C h a n g e s in C a r ib b e a n su rface h y d ro g ra p h y d u r in g the P liocene shoaling o f the C e n tra l A m erican S eaw ay. P a leo cean o g rap h y 21: P A 4221. doi: 1 0 .1 0 2 9 / 2008P A 001645

52. N ü rn b e rg D , M ü lle r A , S ch n e id e r R R {2010) P a leo -sea surface tem p e ra tu re c a lcu lations in th e eq u a to r ia l ea st A d an tic f ro m M g /C a ratio s in p lan k tic fo ram in ife ra : A co m p ariso n to sea surface te m p e ra tu re estim ates f ro m U 3 7 ,

oxygen isotopes, a n d fo ram in ife ra l tran sfe r fu n ctio n . P a leo c e a n o g ra p h y 15: 124— 134. doi: 1 0 .1 0 2 9 / 1999PA 000370.

53. G ro en ev e ld J , C hiessi C M {2011) M g /C a o f Globorotalia inflata as a rec o rd e r o f p e rm a n e n t th erm o clin e tem p e ra tu re s in th e S o u th A d an tic . P a leo cean o g rap h y 26: PA 2203 . d o i:1 0 .1 0 2 9 /2 0 1 0 P A 0 0 1 9 4 0 .

54. S h ack le ton N J {1974) A tta in m e n t o f iso topic e q u ilib riu m b e tw e en o cean w ater a n d b e n th o n ic fo ram in ife ra g en u s Uvigerina: Iso top ic c h an g es in th e o cean d u r in g th e la s t glacial. C o llo q I n t G N R S 219: 2 0 3 -2 1 0 .

55. N aafs B D A , H e fte r J , A c to n G , H a u g G H , M a rtin e z -G a rc ia A , e t al. {2012) S tre n g th e n in g o f N o r th A m erican d u s t sources d u r in g th e la te P liocene {2.7Ma). E a r th P la n e t Sei L e tt 3 1 7 -3 1 8 : 8 -1 9 . doi: 1 0 .1 0 1 6 /j.e p s l.2 0 1 1.11.026.

56. B rasseli SG , E g lin g to n G , M arlo w e IT , P fla u m a n n U , S a rn th e in M {1986) M o lecu lar stra tig rap h y : a n ew tool for c lim atic assessm ent. N a tu re 320: 129— 133. doi: 1 0 .1 0 3 8 /320129a0.

57. P ra h l F G , W a k eh a m S G {1987) C a lib ra tio n o f u n sa tu ra tio n p a tte rn s in long- c h a in k e to n e co m positions for p a le o te m p e ra tu re assessm ent. N a tu re 330: 367— 369. doi: 1 0 .1 0 3 8 /330367a0.

58. H e fte r J {2008) A nalysis o f a lk en o n e u n sa tu ra tio n ind ices w ith fas t gas c h ro m a to g ra p h y /tim e -o f-f lig h t m ass sp ec tro m etry . A n a l C h e m 80: 2161— 2170. d o i:1 0 .1 0 2 1 /a c 7 0 2 1 9 4 m .

59. M ü lle r I J , K irs t G , R u h la n d G , vo n S to rch I, R osell-M elé A {1998) C a lib ra tio n o f th e a lk en o n e p a le o te m p e ra tu re in d ex U 37 b a sed o n co re-tops f ro m th e easte rn S o u th A d a n tic a n d th e g lobal o c e an {60°N—60°S). G eo ch im C o sm o ch em A cta 62: 1 7 5 7 -1 7 7 2 . do i:10 .1016 /S 0 0 1 6 -7 0 3 7 {9 8 )0 0 0 9 7 -0 .

60. S ch iebe l R , B ru p b a c h e r U , S ch m id tk o S, N a u sc h G , W a n ie k J J , e t al. {2011) S p rin g co cco lith o p h o re p ro d u c tio n a n d d ispersion in th e te m p e ra te easte rn N o r th A t la n t ic O c e a n . J G e o p h y s R e s 116: C 0 8 0 3 0 . d o i :1 0 .1 0 2 9 / 2 0 1 0 JC 006841 .

61. C le m e n t B M , R o b in so n F {1987) T h e m ag n e to s tra tig ra p h y o f L eg 9 4 sed im ents. D S D P I n k R e p 94: 6 3 5 -6 5 0 . d o i:1 0 .2 9 7 3 /d sd p .p ro c .9 4 .1 1 2 .1 9 8 7 .

62. E x p ed itio n 303 S cientists {2006) Site U l 308. P ro c IO D P 3 0 3 /3 0 6 : 1-98 . d o i:1 0 .2 2 0 4 /io d p .p ro c . 3 0 3306 .108 .2006 .

63. L o u ren s L, H ilg en F , Sh ack le to n N J, L ask ar J , W ilson D {2005) T h e N eo g en e . In : G ra d s te in F M , O g g J G , S m ith A G , ed itors. A G eo log ica l T im e S cale 2004. C a m b r id g e : C a m b r id g e U n iv e r s i ty P re s s . 4 0 9 - 4 3 0 . d o i: 1 0 .1 0 1 7 / C B O 9 7 8 0 5 11536045

64. S h ip b o a rd S cientific P a r ty {1987) S ite 610. D S D P In it R e p 94: 351—470. d o i :1 0 .2 9 7 3 /d sd p .p ro c .9 4 .106.1987.

65. E x p ed itio n 306 Scientists {2006) S ite U 13 1 3 . P ro c IO D P 3 0 3 /3 0 6 : 1 -1 2 4 . d o i:1 0 .2 2 0 4 /io d p .p ro c . 3 0 3306 .112 .2006 .




6 6 . P a illa rd D , L ab ey rie L , Y io u P (1996) M acin to sh P ro g ra m pe rfo rm s tim e-series analysis. E os T ra n s A G U 77: 379. d o i:1 0 .1 0 2 9 /9 6 E 0 0 0 2 5 9 .

67. S hack le to n N J, H a ii M A , P a te D , Pisias N , M a y er L, e t al. (1995) P liocene stable iso tope s tra tig rap h y o f O D P S ite 846. P ro c O D P Sei R es 138: 337—356. doi: 10 .2 9 7 3 /o d p .p ro c .s r . 138 .117.1995.

6 8 . T ie d e m a n n R , S a rn th e in M , S hack le to n N J (1994) A s tro n o m ic tim escale fo r the P liocene A d a n tic S 180 a n d d u s t flux reco rd s o f O c e a n D rilling P ro g ra m Site 659. P a leo cean o g rap h y 9: 6 1 9 -6 3 8 . doi: 1 0 .1 0 2 9 /9 4 P A 0 0 2 0 8 .

69 . T ie d e m a n n R , S tu rm A , S te p h S, L u n d SP , S to n e r J S , e t al. (2006) A stronom ica lly c a lib ra te d tim escales f ro m 6 to 2.5 M a a n d b e n th ic isotope s tra tig rap h ies , Sites 1236, 1237, 1239, a n d 1241. P ro c O D P Sei R es 202: 1-69 . doi: 1 0 .2 9 7 3 /o d p .p ro c .sr . 202 .210 .2007 .

70. S ch n e id e r B, S ch m ittn e r A (2006) S im u la tin g th e im p a c t o f th e P a n a m a n ia n seaw ay closure o n o cean c ircu la tio n , m arin e p ro d u ctiv ity a n d n u tr ie n t cycling. E a r th P la n e t Sei L e tt 246: 3 6 7 -3 8 0 . d o i:1 0 .1 0 1 6 /j.ep s l.2 0 0 6 .0 4 .0 2 8 .

71. L u n t D J, V aldes PJ, H a y w o o d A M , R u t t IG (2008) C lo su re o f th e P a n a m a Seaw ay d u r in g th e P liocene: im plica tions fo r c lim ate a n d N o r th e rn H em isp h ere g lac ia tion . G lim D y n 30: 1 -1 8 . d o i:1 0 .1 0 0 7 /s0 0 3 8 2 -0 0 7 -0 2 6 5 -6 .

72. K lo ck e r A, P ran g e M , Schulz M (2005) T e s tin g th e in fluence o f th e C e n tra l A m e ric a n Seaw ay o n o rb ita lly fo rced N o r th e rn H e m isp h e re g lacia tion . G eophys R es L e tt 32: L 03 7 0 3 . d o i:1 0 .1 0 2 9 /2 0 0 4 g l0 2 1 5 6 4 .

73. D ar lin g K F , K u c e ra M , W a d e G M , v o n L an g e n P, P ak D (2003) Seasonal d istr ib u tio n o f gene tic types o f p lan k to n ic fo ram in ife r m o rp h o sp ec ie s in the S an ta B a rb a ra C h a n n e l a n d its p a le o c e a n o g ra p h ic im plications. P a leo cean o g rap h y 18(2): 1032. d o i:1 0 .1 0 2 9 /2 0 0 1 P A 0 0 0 7 2 3 .

74. B o rn A , K a g e y a m a M , N isanciog lu K H (2010) W a rm N o rd ic Seas d e lay ed g lacial in ce p tio n in S can d in av ia . G lim P ast 6 : 8 1 7 -8 2 6 . d o i:1 0 .5 1 9 4 /c p -6 -8 1 7 - 2 0 1 0 .

75. S a lzm an n U , H a y w o o d A M , L u n t D J, V ald es I J , H ill D J (2008) A n ew global b io m e re c o n s tru c tio n a n d d a ta -m o d e l c o m p ariso n for th e M id d le Pliocene. G lob E co l B iogeogr 17: 4 3 2 -4 4 7 . doi: 10.111 l / j . l4 6 6 -8 2 3 8 .2 0 0 8 .0 0 3 8 1 .x .

76. G ao G , M cA n d rew s J H , W a n g X , M enzies J , T u r to n C L , e t al. {2012) G lac ia tio n o f N o r th A m erica in th e J a m e s Bay L o w lan d , C a n a d a , 3.5 M a. G eo logy 40: 9 7 5 -9 7 8 . doi: 1 0 .1 1 3 0 /G 3 3092.1.

77. C a rlso n A E , O p p o D W , G am e R E , L eg ran d e A N , K eig w in L D , e t al. (2008) S u b tro p ica l A d a n tic salin ity v ariab ility a n d A d a n tic m erid io n a l c ircu la tio n d u r in g th e las t d eg lac iation . G eology 36: 9 9 1 -9 9 4 . d o i:1 0 .1 1 3 0 /G 2 5 0 8 0 A .l.

78. S tep h S, T ie d e m a n n R , G ro en ev e ld J , S tu rm A , N ü rn b e rg D (2006) P liocene chan g es in tro p ica l E a s t Pacific u p p e r o cean s trad ficad o n : response to tro p ical gatew ays? P ro c O D P Sei R es 202: 1 -5 1 . d o i:1 0 .2 9 7 3 /o d p .p ro c .s r .2 0 2 .211 .2006 .

79. B in tan ja R , v a n d e W a l R S W (2008) N o r th A m e ric a n ice-shee t d y n am ics a n d th e o n se t o f 100 ,000-year glacial cycles. N a tu re 454: 8 6 9 -8 7 2 . d o i:1 0 .1 0 3 8 / n a tu re0 7 1 5 8 .

80. K ü rsc h n e r W , v a n d e r B u rg h J , V isscher H , D ilch e r D (1996) O a k leaves as b iosensors o f la te N e o g e n e a n d early P le istocene p a le o a tm o sp h e ric C O 2

c o n cen trad o n s . M a r M ic ro p a le o n to l 27: 299—312.81. L o c a rn in i R A , M ish o n o v A V , A n to n o v J I , B oyer T P , G a rc ia H E (2006) W o rld

O c e a n A d as 2005, V o lu m e 1: T e m p era tu re . In : L evitus S, ed ito r. N O A A A das N E S D IS 61. W ash in g to n , D .G .: U .S . G ov. P rin tin g O ffice, p . 182. Available: h t t p : / /w w w .n o d c .n o a a .g o v /O C 5 /in d p ro d .h tm l .


http://www.nodc.noaa.gov/OC5/indprod.html


northern hemisphere glaciation during the globally warm ...the early late pliocene (3.6 to ~3.0...

Documents