Benthic isotope evidence for changes of the ... - Wiley Online Library

19 downloads 0 Views 1MB Size Report
und Museum, Christian-Albrechts Universitaet, ... [Wust, 1936; Worthington, 1976; Reid, 1981;. Lacombe ... formation in the Norwegian Sea [Reid, 1979], that.
PALEOCEANOGRAPHY, VOL. 2, NO. 6, PAGES543-559, DECEMBER 1987

BENTHIC ISOTOPE EVIDENCE FOR CHANGES OF THE MEDITERRANEAN OUTFLOW

DURING THE LATE QUATERNARY

RainerZahn1 andMichael Sarnthein Geologisch-Palaeontologisches Institut und Museum, Christian-AlbrechtsUniversitaet, Olshausenstrasse 40-60, Kiel,

FederalRepublicof Germany Helmut

Erlenkeuser

Institutfuer Reineund Angewandte Kernphysik,Christian-AlbrechtsUniversitaet, Olshausenstrasse 40-60, Kiel,

FederalRepublicof Germany

Abstract. A suit of sediment cores close to and

southof theStraitof Gibraltar(12ø-36øN,5002800 m waterdepth)wereanalyzedfor stable isotopesin epibenthicforaminifersCibicidoides

climaticchangeof the carbonisotopesignalsof bothwatermasses indicates,thatthesupplyof salineMOW to thenorthAtlanticmay be less

wuellerstorfi andPlanulin•a ariminensi•s. During

importantfor the formation of NADW than previouslyassumed.

valuesof up to 1.6"/ooat intermediate depths

INTRODUCTION

butalsoto thenorthasrevealedby dataof intermediate depthcoresnorthof 38øN(in

pronounced highsalinity(>38"/00) component to the modernhydrographyof the AtlanticOcean [Wust, 1936;Worthington,1976; Reid, 1981;

peakglacial times,..the dataexhibit higher •5•3C neartheStraitof Gibraltar(36øN).Thevalues decrease to thesouthasevidenced by ourdata,

Duplessyet al. [ 1987]). Thus, the distribution

pattern of •5•3C provides crucial evidence foran increased influenceof nutrientdepleted MediterraneanOutflow Water (MOW) on the glacialnortheast Atlantichydrography. During oxygenisotopeTerminationsI and II, the

meridionalcarbonisotopegradientindicatesa significantlydecreasedbut still activeMOW. As

deduced fromthe•5•80fluctuations, the temperaturesof the MOW in the Atlantic were

lowerduringglacialtimesby asmuchas5øC. Duringglaci. aj timesandduringTerminationI the maximum •St•C values of the MOW correlate with

minimumvaluesof theNorthAtlanticDeepWater (NADW) andviceversa.Thisinverseresponse to

•NowatCollege of Oceanography, Oregon State University, Corvallis, OR 97331.

Copyright1987 by theAmericanGeophysical Union. Papernumber7PO834. 0883- 8305/87/007P-0834510.00

Mediterranean OutflowWater .. (MOW) provides a

Lacombeand Richez, 1985; Washburnand Kaese, 1987]. Upon leavingthe Straitof Gibraltar,it sinksto water depthsbetween800 and 1500 m [Zenk, 1971; Armi and Farmer, 1985; Gascard and

Richez, 1985] and spreadsto the west and southwest[Kawase and Sarmiento, 1986; Kaese and

Zenk, 1987]. Along the northwestAfrican continentalmargin, the MOW can be tracedsouth-

wardsto CapeBlancat21øN[e.g.,Mittelstaedt et al., 1975; Huber et al., 1977]. To the south of the

Greenland-Scottland Ridge, a northernbranchof the MOW

mixes with overflow

waters from the

Norwegian-Greenlandseasandwith deepwaters from the LabradorSea [Dooley andMeincke, 1981; Talley and McCartney, 1982; Kaeseet al., 1986; Kawase and Sarmiento, 1986]. As a result, a

relatively_warm (7.7< TøC< 11.6)andsaline

(35.0 < S•'/oo < 36.1) water massforms, which buildsthe upperlayer of the North Atlantic Deep Water (NADW) and can be easily traced throughoutthe Atlantic and even into the Pacific Ocean [Wust, 1936]. Thus the advection of MOW

into the Atlantic is likely to have oceanwide implications.It has alsobeen suggestedthat the

544

MOW may directlycontributeto thedeepwater formationin the NorwegianSea[Reid, 1979], that is, to the formation of the NADW itself, although this model is still under debate.

The advectionof the modemMOW dependson the deepwaterformationin the Mediterranean, whichis closelytied to the climaticconditionsof this area [Bethoux, 1979a; Lacombeet al., 1981].

In thisadjacentsea,deepandintermediate water formationresultsfrom excessevaporationof surfacewatersandwinterly deepconvection, particularlyin the northernBalearenandthe Levantineseas[Medoc Group, 1970; Lacombeet al., 1981;BrydenandStommel,1982;Millot, 1987]. A numberof studiespresentedgeochemical, micropaleontological, andstableisotopeevidence for markedchangesof bothclimaticand hydrographic conditions in theMediterranean duringthelate Quaternary[e.g.,Thunellet al., 1977, 1984; Williams et al., 1978; Thiede, 1980; Nesteroffet al., 1981; Cita andGrignani, 1983; Rossignol-Strick et al., 1982;Rossignol-Strick, 1985;VergnaudGrazziniet al., 1986;Ganssen and Troelstra, 1987; ten Haven et al., 1987]. These studiesrevealedthat the glacialwinterly sea

surfacetemperatures (SST) of the Mediterranean

Seawereconsiderably colder byabout 7øCand

that the seasonalSST variationswere higher than

today.Furthermore, theytestifyto a high freshwaterinputinto theeasternMediterranean duringthelastdeglacial.Thissituation ledto a highlystratifiedwatercolumnand,in turn,to intermittent periodsof deepwaterstagnation as evidenced by theformationof sapropels (details in the work of Thunell et al. [ 1984]). These

changes of theMediterranean hydrography canbe expectedto alsoaffectthe intermediate water andpossiblyevendeepwaterhydrography of the Atlantic Ocean.It will closelyrespondto possible variationsin both the advectionrate and density of the MOW, as well as to its nutrient com-

position.The mostextrememodelis a totalshutoff of MOW advectionduringglacialand/or deglacialtimesasa resultof theenhanced freshwatersupplyinto the MediterraneanSea [e.g.,Huanget al., 1972;Diester-Haass, 1973; Bethoux, 1979a].

In thisstudywe presentstableisotoperecords of epibenthic foraminifersfrom 13 sediment cores, which have been obtained from the northeast

Atlantic continentalmargin off and to the south of the Straitof Gibraltar(Figure 1, Table 1) (the

isotopedataareavailableuponrequestfromthe seniorauthor).The coresform threebathymetric

transects between 20ø and35øNranging from500 to 2800 rn waterdepth,thatis, crossingthe modemflow pathof the MOW (Figure1). For comparative purposes, additional isotoperecords

Zahnet al.: Late QuaternaryMediterranean Outflow convincingevidencethattheintermediate depth levels of the northeast Atlantic Ocean were well

oxygenated duringthepastglacial-interglacial cycle.By analogywith themodemcirculation pattern,we concludethatthedataindicatea continuousadvectionof nutrient-depletedMOW to the Atlanticduringthe last 140,000yearsB.P. METHODS

StableIsotopeAnalysis

Stableisotopemeasurements wereperformedat theC 14 Laboratoryof Kiel Universityon botha MicromassVG 602D anda FinniganMAT 251 massspectrometer, thelatterwith theCARBOKIEL automatedpreparationsystem.Both instruments were calibrated to the Pee Dee

Belemnite(PDB) scalethroughNationalBureauof Standards(NBS) 19 andNBS 20 carbonate standards. Additionally,intercalibrationwas checkedwith a laboratorystandard(Solnhofen limestone),thusallowingfor comparison of values measuredwith bothsystems.Externalreproducibilities(includingcarbonate r•reparation and

spec[ometric ana12/sis) for•51•O and•513C were

0.10•'/oo and 0.05"/00 for the VG 602D, and

0.070/00and0.040/00for theMAT 251, respectively. Measurements weremadeon monospecificbenthicforaminiferalsamples of 1-10 specimens, whichwerepickedfromthe250-to 400-gm-sizefraction.Sampletreatment and preparation followedstandard procedures [e.g., Ganssen,1983; Zahn, 1986].

Mostisotopicdatahavebeenobtainedfrom Cibicidoides wuellerstorfi. The carbonate of this

species exhibits •5•Cvalues whichlieveryclose to those of the 2;CO.. of the ambient bottom

water [Belanger eta•.,1981; Graham etal.,1981; Duplessyet al., 1984;Zahnet al., 1986]because of its habitat,which is well elevatedabovethe sedimentsurface[Lutze and Thiel, 1987]. This

species, however,livesalongthecontinental marginof northwestAfrica mainly at water depthsbelow3000 rn anddoesnotoccurat

depthsshallower thanabout1000rn [Lutze,1980]. To determinethepastcirculationpatternsalsoin theupperdepthrangeof themodernMOW, that is, above 1000 m, isotopemeasurements were carriedout on Planulinaariminensisd'Orbigny

[ 1826].This species livesoff northwestAfrica at waterdepthsbetween300 and 1000rn [Lutze, 1980].Comparison of its oxygenandcarbon isotopedatawith thoseof C. wuellerstorfi show highcorrelations of 0.93 and0.91,respectively (Figure2). Slightshiftsin theC. wuellerstorfi

versus P. ariminensis isotot•e valuesof-0.3ø/00

and+0.20/00occur for•5•8'O and•5•3C. Larger

of six northeastAtlantic deep seacoreswere

offsets of some1ø/ooaremainlyrestricted to

used(Figure1, Table1). Ourisotopedataprovide

TerminationsI and II, wherethey obviouslyresult

Zahnet al.- Late QuaternaryMediterraneanOutflow degrees 40

"'•( •S•

'LY-II-13A•

545

northern

latitude

30

20

I

I

10

I

15666

I

-- / ++•5669 ,•1 0 .•

lO

o

.s.

• • •'

20 -

+ + 603013289 +12347

12379 •



-

12328

20

/ degrees 40

northern

latitude

30

20

I

I

lO

b

+15663 +1600•

+ 15666

LY-II-13A

MOW

+

+ 15670

11944

+ 15627 •

16017 +



12345

+ 16004

-1

+ 16030

+

-2

+ 15669

NADW +

+

12379 13289

15672

+ 12392

+ 12347

3'

1230F

12328

-3

Fig. 1. Positionsof sedimentcores.(a) Locationmap. Isohalinesat 1000 m water depth indicatemajor flow path of the MOW (adoptedfrom Kaeseand Zenk [ 1987]). (b) Depth distributionof the coresand of major modernwater masses. from bioturbationalmixing of oxygenisotopically heavyshells,that is, of glacialshellsof C. wuellerstorfiinto interglacialsectionsof the records(Figure 2a, core 15666). Thus testsof P. ariminensisappearto providea valuablebasis

Since most of our sediment cores lack the

modern sediment surface, benthos versus water

isotopicadjustments were basedon a separateset of stableisotopedataof surfacesedimentsamples

[Ganssen, 1983].The/5•80 equilibrium values for

for/5•3C-based studies onthepaleohydrography of

calcite were calculated after O'Neil et al. [ 1969]

upperintermediatewater massessuchas the

from a te_rnperature and salinityprofile at

of the bottom water ZCO• and thus, also those of

profile from the samelocationwas availablefrom Duplessy[ 1972]. Largely oxygenisotopedata from Uvigerina were usedto deducethe oxygenisotope stratigraphyand to infer possibletemperature/ salinity(TS-) fluctuationsin the MOW. Oxygen isotopevaluesof this genusare consideredas

MOW.Based onKroopnick [1971],the/5•3C values ep•benth•cforarmmferacanbe depletedby as

muchas0.350/00wherethefluxof organic carbonreachesa maximum near the seafloor(see alsoMcCorkle [ 1987]), which, however, is not the case at our core sites.

36øN/09øW [Stramma, 1981].A hydrographic •13C

546

Zahn et al.: Late QuaternaryMediterraneanOutflow TABLE 1. Locationsand Depthsof SedimentCores Average

Core Number

Northern Latitude

Western Longitude

Depth, m

Sedimentation Rates

Ambient Water Mass

Profile A

MOW a

LY-II-13A

35o58.0'

07o49.0'

1201

22.8

11944

35ø39.1'

08o03.7

1765

5.7

15663

34055.2'

06050.7'

500

16.6

NACW

15666

34057.6'

07ø07.1'

803

2.3

MOW

15669

34053.5'

07048.9'

2022

5.1

NADW

15670

34054.5'

07034.6'

1482

4.1

NADW

15672

34051.6'

08007.6'

2455

4.4

NADW

Profile

MOW/NADW

B

15627

29ø10.0'

12ø05.2'

1021

2.9

MOW

16004

29o58.7'

10o38.8'

1512

4.0

NADW

16006

29ø14.8'

11029.8'

796

4.0

MOW/NADW

Profile C

12379

23o08.1'

17ø44.7'

2136

11.1

NADW b

16017

21ø14.7'

17ø48.2'

812

17.4

EUC/NADW

16030

21ø14.1'

18ø03.3'

1500

6.0

NADW

Reference Cores

12309

26o50.3'

15o06.6'

2820

10.4

NADW

12328

21o08.7'

18o34.4'

2778

14.0

NADW

12345

15028.8'

17021.6'

945

29.8

SACW

12347

15049.5'

17051.7'

2576

22.1

NADW

12392

25010.3'

16050.7'

2575

8.5

NADWC,d

13289

18004.4'

18000.6'

2490

6.6

NADW

SU 81-21

38015.0'

09033.0'

1260

MOW e

SU 81-44

44015.0'

02042.0'

1173

MOW e

SU 81-45

44006.0'

02029.0'

994

MOW e

EUC Equatorward UnderCurrent;MOW Mediterranean OutflowWater;NACW North AtlanticCentralWater;NADW NorthAtlanticDeepWater;SACW SouthAtlanticCentralWater. a Samplesprovidedby L. Diester-Haass.

bStratigraphy rearranged afterKoopmann [1979]. c Data from Shackleton [ 1977].

dDatafromZahnet al. [ 1986]. e Data from Duplessyet al. [ 1987] beingcloseto equilibriumwith the ambient seawatervalues [Shackleton,1974]. For the rest

equivalents by applyinga conversion factorof

of the cores the data of C. wuellerstorfi/

closelymatchesthevalueof 0.69"/00, whichwas

P. ariminensiswere normalizedto Uvigerina

obtained from 423 double measurements on the

0.640/00 [Shackleton etal.,1984]c. Thisfactor

Zahn et al.: Late QuaternaryMediterraneanOutflow

5q?

16006

6180 vs

PDB

ß

i

ß

i

ß

ß

i

ß

ß

ß

-

-a- P.arlmlnensls -e- C.wuellerstorfi 15666 1

6180 vs

PDB 2

i

0

:20

i

i

i

40

60

80

i

i

100

ß

120

i

140

ago ( 1,000 yr)

b

4-

•180 PDB

R=0.93/

vs

v S PDB



I

ß

,

2

ß

!

3

o

4

I

1



2

C. wuellorstorfl

Fig. 2. Comparisonof stableisotopedataobtainedfrom C. wuellerstorfiandP. ariminensis.(a) down-coreoxygenisotopeprofilesand (b) correlationgraphs,withoutvaluesfrom oxygenisotopeTerminationsI andII (for discussion seetext).

respectivespeciesalong 19 sedimentcoresfrom

fraction> 125 gm. The calculationof the radio-

the northeastAtlantic [Zahn, 1986].

carbondata,(Table2) is basedon 95% of theNBS

Radio CarbonAges

oxalic acid •'•C standardacitvity and on the conventional(Libby) half-life of 5568 yearsfor

•4C.Corrections forisotope fractionation were notapplied since the6•3Cvalues ofthedated

For a more precisechronostratigraphical control of the sedimentsyoungerthan30,000 years, radiocarbonageswere measuredon six cores

effectofisotope fractionation onthe•4Cages

for cores12379 and 12392 (in Koopmann[ 1979], andunpublished data).Samplescontainedbetween 10 and30 g of carbonateandwere from the size

will closelybalancethe reservoirage of about400 yearsof the ocean'ssurfacelayer. Time controlof the oxygenisotoperecords older than stage2 was adoptedfrom the

(Table2). A fewfurther•4Cdatawereavailable

carbonates werecloseto 0ø/ooPDB,thatis, the

5•48

Zahn et al.: Late QuaternaryMediterraneanOutflow PROFILE A (35øN) oxygen isotope stages i

ß

i

20

&]80 2

vs PDB

ß

!

,

i

oxygen isotope stages

,

.

ß

,;o

&180

15663

3

.!,

ß

40 60 8; 1;0 1:•0 140kyr B.P. vs

11944

PDB

1,765 m

500m

4

11 2

15666

3

803m

15669 2,022 m

4

3

2

15672

3i!iii: ..• LY-II-13A 4

2,455m

1,201 m

5

4

1

6

3 ' '•

15670

4

1,482 m

,

0

20

40

60

80

1 0

, 120! ! 140!

age (1,000 yr B.P.)

,

0

40 age (1,000 yr B.P.)

Fig. 3. Oxygenisotopestratigraphies of coresat profilesA-C. Valuesof C. wuellerstorfiand P. ariminensis arenormalized to Uvi[erinavalues.Curveof core15666basedonUvigerina

values. Arrows markposition of ]4C•tges listed inTable2. Forchronostratigraphy seetext. CARTUNE time scale(modified from CARPOR and STUNE time scales;Herterich and Sarnthein [1984]).

thoseexpectedfrom the respectiveoxygen isotopiccurve structure.In core 16017, maximum

offsets between expected andmeasured 14Cages reach14,000 years.This benthicisotoperecord

Stratigraphy

exhibits an unusual scatter,which is also observed

in threeplanktonoxygenisotopeprofilesfrom the Most of our oxygenisotoperecordsfrom the upperand middle northeastAtlantic continental margin(Figure3) closelyresemblethe established oxygenisotopehistoryof theworld ocean[e.g., Pisias et al., 1984]. Most of them contain the climatic transitionof Termination I; somealso

containTerminationII (Figure3). Only core 15627 lackssediments youngerthan15,000yearsB.P., possiblydueto a cotingloss.Core 15666might lack the sediments of the oxygenisotopesubstages5c and5d asrevealedby the anomalously smoothcourseof the up.per stage5 curve,aswell

asbytherather sharp/51•O peakofsubstage 5e.

Despitethe generallywell-definedstructureof

ourO-isotope curves, the14Cages of cores

15666, 16006, and 16017 from shallower water

depthsappearto be too old whencomparedwith

samecore[Zahn,1986].Furthermore, the

data which have been obtained from the smaller

sizefraction(125-200 gm) reveala significantly older age than thosefrom the coarsersize fractions(>200 gm; sample211-222 cm; Table 2).

Hence thediscrepancies between 14Cages and expected 5180curveagesandthe14Coffsets

betweensamplesof differentsizefractionsfrom the samesamplemay be relatedto downslope redeposition of older,fine-grainedcarbonate

particles. Thusthe14Cdataof these shallow site coreswere rejectedand stratigraphic controlwas setby analogywith the oxygenisotopeprofilesof coresfrom the deepsea. Averagesedimentation ratesof the coresrange from 3 crn/1000yearsto >20 cm/1000yearsalong the northwestAfrican continentalmargin(Table

Zahn et al.' Late QuaternaryMediterraneanOutflow

5119

PROFILEB (29øN)

PROFILE C (22øN) oxygen isotope stages

oxygen Isotope stages

s

', '

'

'

'

I

I

',;o

vs PD8

16006

5180 2

ß

i

ß

ß

i

ß

ß

i

ß

I

,

•180

vs PDB 3

16017

3

796

i

20 40 60 8; 100 1•0 140kyr B.P. 812m

m 4

4

3

3 •

16030

15627 1,021 rn

1,500 m

4

5

11 3

16004 4

12379

1,512 rn

4 0

20

40

60

80

100

120

140

2,136 rn

5

age (1,000 yr 6

ß

0

2O

age (1,000 yr B.P.)

Fig. 3. (continued) 1). Thus the time resolutionof our stratigraphic recordsreaches250-1200 yearsfor sampling intervals of 5-10 cm.

RESULTS

AND

DISCUSSION

watermass ofheavy•5•3C values of about +1.30ø/oo(Figure 5b,> recent recent recent l) tlldop aele•

i

t •)

(re>l) qldep aoIet•

55•

Zahn et al.: Late QuaternaryMediterraneanOutflow age

(1,000 yr B.P.)

10

I

20

i >1.2



och-6 carbonshift:A changein 13 I2 the oceans C/ C ratio 6.2 million yearsago, Mar. Micropaleontol.,5, 185-203, 1980. Washburn, L., and R. H. Kaese, Double diffusion

and the distributionof the densityratio in the Mediterranean

water front southeast of the

Zahn et al.: Late QuaternaryMediterraneanOutflow Acores,J. Phys.Oceanogr.,17, 12-25, 1987. Williams, D. F., R. C. Thunell, and J.P. Kennett,

Periodicfreshwaterflooding and stagnationof the easternMediterraneanSea duringthe late Quaternary,Science,201,252-254, 1978. Worthington,L. V., On the North Atlantic circulation,JohnHopkinsOceanogr.Stud.,6, 110 pp., 1976. Wust, G., Schichtungund Zirkulationdes AtlantischenOzeans,Wiss, Ergeb.Dtsch.Aft. Exped.Forsch,Schiff "Meteor" 1.925-1927,6 (1/2), 1936. Zahn, R., Spatquartare Entwicklungyon Kustenauftrieb und Tiefenwasserzirkulation

im

Nordost-Atlantik, Rekonstruktion anhand stabiler

Isotopekalkschaliger Foraminiferen,Ph.D. thesis,111 pp., Kiel University,Kiel, Federal Republicof Germany, 1986. Zahn, R., K. Winn. and M. Sarnthein, Benthic

foraminiferal•13C andaccumulation ratesof

559

organiccarbon(UvigerinaperegrinaGroupand Cibicidoideswuellerstorfi),Paleoceanography, _1, 27-42, 1986.

Zenk, W., Zur SchichtungdesMittelmeerwassers westlichyon Gibraltar,"Meteor"-Forschungsergeb.,Reihe A, 9, 1-30, 1971. H. Erlenkeuser, Institut fuer Reine und

AngewandteKemphysik,C-14 Labor, ChristianAlbrechts Universitaet, Olshausenstrasse40-60,

D-2300 Kiel, FederalRepublicof Germany. M. SarntheinandR. Zahn, GeologischPalaeontologisches Institutund Museum,ChristianAlbrechts Universitaet, Olshausenstrasse 40-60,

D-2300 Kiel, FederalRepublicof Germany. (ReceivedJuly 15, 1987; revised October 20, 1987;

acceptedOctober22, 1987)