PALEOCEANOGRAPHIC
CURRENTS
PALEOCEANOGRAPHY, VOL. 11, NO. 2, PAGES 147-156,APRIL 1996
Three hundred eighty thousand year long stable isotope
and faunal
records from the Red Sea:
Influence of global sea level change on hydrography
Christoph
Hemleben,
• DieterMeischner,
2 RainerZahn,3 AhuvaAlmogi-Labin,
4
HelmutErlenkeuser,
• andBirgitHiller•
Abstract. Stableisotopeand faunalrecordsfrom the centralRed Seashowhigh-amplitude
oscillations
for thepast380,000years.Positive
• •80anomalies
indicate
periods
of
significantsaltbuildupduringperiodsof loweredsealevel whenwater massexchangewith
the ArabianSeawasreduceddueto a reducedgeometryof the Babel MandebStrait.
Salinitiesas high as 53%o and 55%o are inferredfrom pteropodandbenthieforaminifera
•80, respectively,
forthelastglacialmaximum.
Duringthisperiodallplanktonic
foraminiferavanishedfrom this part of the Red Sea. Environmentalconditionsimproved
rapidlyafter 13 ka as salinitiesdecreased
dueto risingsealevel. The foraminiferalfauna
startedto reappearandwasfully reestablished
between9 ka and 8 ka. Spectralanalysisof
theplanktonic
•80 recorddocuments
highest
variance
in theorbitaleccentricity,
obliquity,
and precessionbands,indicatinga dominantinfluenceof climatically- driven sealevel
changeon environmentalconditionsin the Red Sea. Variancein the precessionbandis
enhance•comparedto the global meanmarine climaterecord (SPECMAP), suggestingan
additionalinfluenceof the Indianmonsoonsystemon Red Seaclimates.
Introduction
Quantitative studiesand isotopic analysesof planktonic
foraminifera and pteropodafrom the Red Sea revealed a
detailedpicture of water massvariation in associationwith
global climatic changesduring the latest Pleistocene(from
isotopestage6 onward) and early Holocene[Herman, 1968;
Chen, 1969; Deuser and Degens, 1969; Deuser et al., 1976;
Risch, 1976; Yusuf, 1978; Reisset al., 1980; Almogi-Labin,
1982;Ivanova, 1985;Almogi-Labinet al., 1986, 1991; Locke
and Thunell, 1988; Thunell et al. 1988]. During the last
glacial maximum (LGM), approximately18,000 years ago,
surface and intermediatewater salinitieswere significantly
higherthantheyare today.Only onepteropodspecies,Creseis
acicula, survived while planktonic foraminifera vanished
almostentirelyfrom the Red Sea, resultingin an Maplanktonic
zoneM [Thunell et al., 1988]. Only during the PleistoceneHolocene transitionperiod did the planktonicfauna slowly
migrate back from the Gulf of Aden into the Red Sea. These
changeshavebeenattributedto the Red Sea, being a marginal
sea, which makes its hydrographyextremely susceptibleto
global climate change. Global sea level variations exert
additional control on Red Sea hydrography in that they
determinethe water massexchangewith the openoceanand
thus residencetime of water massesand salt buildup in the
Red Sea [7hunell et al., 1988; Brydenand Kinder, 1991]
Piston cores up to 22 m long were recovered from the
centralRed SeaandoffshoreSudanduringan expeditionwith
the R/V Meteor in 1987. Here we presentstableisotopeand
faunal records of one of these cores from the central Red Sea
which span380 kyr and significantlyextendthe documented
record of environmentalchangein the Red Sea. The records
show that Red Sea climatesrespondednot only to glacialinterglacialconditionsbut alsoto higher-frequencyvariations
in the orbitalprecessionband. As only fragmentarycontinental recordsare availablefrom the surroundinglandmasses,
the
recordsmay also extendour understanding
of the long-term
climatic history of this region.
•Institute
andMuseumfor GeologyandPaleontology,
Universityof
Ttibingen,Tiibingen,Germany.
2Department
of Sedimentary
Geology,Institutefor Geologyand
Paleontology,Universityof G6ttingen,G6ttingen,Germany.
3GEOMAR,Research
CenterforMarineGeosciences,
Kiel, Germany.
4Geological
Surveyof Israel,Jerusalem.
•Institut
fiir ReineundAngewandte
Kernphysik,
Kiel, Germany.
Material
and Methods
We have generateda planktonicstableisotoperecord and
records of foraminiferal abundanceand faunal variability
along a 21 m long Meteor Core Sta. 174/KL11 from the
centralRedSea(18ø46.3'N, 39ø19.9'E, 825 m water depth),
hereafter referred to as core KL11. For •5•80 measurements
Copyright1996by the AmericanGeophysical
Union.
we haveusedthetropical-subtropical
planktonicforaminiferal
speciesGlobigerinoides
nd,er (forma white), which dominates
the foraminiferalfaunaat the site of core KL11. An age scale
Paper number95PA03838.
0883-8305/96/95 PA-03838510.00
147
148
HEMLEBEN
ET AL.' GLOBAL SEA LEVEL AND RED SEA HYDROGRAPHY
5.5
-2•
[
ca
-1
- ••ßc
n
•
ø
•
0-
•
1.....
7.0•1.
J 909'1
6.06.
.0
...............
I
3.•0
•o
5.2
•
4.2
6.4
••
0
11.0,
10.0
8.3
8.4
6.2
I
2
I
4
I
6
I
8
I
10
depth
•
1
in core
14
16
18
20
I
22
(m)
Figure 1. Pla•to•c (Globi•e•o•e• r•ber, white variety)oxygenisotoperecordof core KL
Numbers•ong •e i•m• r•ord •dica• SPECM• eventsa•er I••
et M. [•984] which were used
for ageconsol(• Table D. At coredepthof 0.9-1.7 m no pl•to•c
forami•fera were availablefor
isotopemeasurements
due to •e "apla•tonJc zone"which was causedby extremelyhigh •inities
during •e last [laci• m•mum •6•
at approximately• 8 ka.
for coreKL11 wasobtainedthroughcorrelationof the isotope
Table 1. Stratigraphic
Fix Pointsfor CoreKL 11
records
withtheglobalmeanSPECMAP5180curve( (Figure
Depth, m
1, Table 1) [lmbrie et al., 1984]. The base of core KL11
reachesSPECMAP interglacialclimatic event 11.2, a shortlived cold spell at approximately375 ka (Figures 1 and 2).
The age model suggeststhat mean sedimentationratesat the
core site are 5.6 cm kyr. The isotopemeasurementswere
carried out at the C 14-Laboratoryat Kiel Universityusinga
FinniganMAT 251 massspectrometer,
which is linked on-line
to a fully automatedcarbonatepreparationdevice. The carbonate samplesare dissolvedin separateglassvials thus minimizingpotentialmemoryeffects.Reproducibilitywas 0.08%0,
and all isotopevaluesare referred to the Peedeebelemnite
(PDB) scale. The planktonicforaminiferalcountswere done
on the size fraction > 150 •m and normalizedto 1 g of
sediment.
Results
Stable Isotopesand Hydrography
Theplanktonic
5•SOprofilealongcoreKL11shows
highamplitudefluctuations
with glacial-interglacialshiftsup to 3.5-
4.0%0(Figure2a).Thecoevalglobalglacial-interglacial
variation
of the world ocean's water masses was 1.1-1.2%0
[Labeyrieet al., 1987; Fairbanks, 1989], and thusplanktonic
•180amplitudes
at coreKL11exceed
themean-ocean
change
bya factorof 3-4.Theglacial-interglacial
•SO amplitudes
are
also about2 to 3 timeslarger(A•SO=3-4%0)thanthose
observed
in planktonic
5•SOrecordsfromthe ArabianSea
(A•SO=1.1-1.7%0) [Niitsumaet al., 1991; Zahn and
Age, ka
Deposition
Rate1em/kyr
SPECMAP
event
0.64
6
2.10
24
1.1
3.23
53
3.90
3.3
3.66
59
7.17
4.0
8.11
3.0
3.88
65
3.67
4.2
4.16
71
4.67
5.0
4.70
80
6.00
5.1
4.85
91
1.36
5.2
5.65
99
10.00
5.3
6.05
109
4.00
5.4
6.63
122
4.46
5.5
6.70
128
1.17
6.0
7.17
135
6.71
6.2
7.61
145
4.40
6.3
8.20
151
9.83
6.4
9.15
174
4.13
6.5
9.64
183
5.44
6.6
9.97
186
11.00
7.0
10.46
194
6.13
7.1
11.00
209
3.60
7.2
11.30
216
4.29
7.3
11.70
228
3.33
7.4
12.63
238
9.30
7.5
12.96
245
4.71
8.0
13.18
249
5.50
8.2
13.60
257
5.25
8.3
14.28
269
5.67
8.4
15.13
287
4.72
8.5
15.67
299
4.50
8.6
16.00
303
8.25
9.0
Pedersen,1991; Sirockoet al., 1993], which may be ascribed
to the concentration
effect(salt buildup) [cf. Morcos, 1970] of
the Red Sea as a marginal sea. In addition, to the glacial-
16.50
310
7.14
9.1
16.80
320
3.00
9.2
interglacialisotopeshiftsthe •sO recorddisplayshigh-
17.48
331
6.18
9.3
amplitudevariationsof more than 1%0 within glacial and
17.93
339
5.62
10.0
interglacialstageswhich exceedthe degreeof variability seen
18.24
341
15.50
10.2
in •sO recordsat open-ocean
sites.For the 'aplanktonic'
19.73
362
7.10
11.0
interval during the LGM no planktonic foraminifera were
20.60
375
6.69
11.2
HEMLEBEN
ET AL.'
GLOBAL
SEA LEVEL
•80
AGE ( 1000 years B.P. )
50 100 150 200 250 300 350 400 •:
',.,' '
' ' ' ;enoe'anX'
' '
o
•[
-2
AND
Z
fø'ø
1.00rn
RED SEA HYDROGRAPHY
149
value of +4.7%0 which was obtained for the LGM
pteropodsampleto the 8•80-scaleof G. ruber (a bulk
sediment
sample
wasmeasured
for reference
andgavea 6•80
valueof +6.3%o).Thus,we arriveat a G. ruber6•0 equivalentof +3.3%o for the LGM. Usingthis value, a
amplitude of 5.5%o is calculatedbetweenthe LGM and the
34
Holocene,whichis the largestamplituderecordedthroughout
10
the entire profile.
In orderto extracta recordof hydrographic
variabilitywe
have computedthe difference between the record of mean-
ocean6•80 (Sw)from Vogelsang
[19901andtheplanktonic
6•80record
fromcoreKL11.Usingbenthic
õ•80records
from
3,3%o
L
W
HIGH
OW
HIGH
LOW
HIGH L
W
SEA LEVEL
the NorwegianSea and assumingthat bottomwater temperature there remained constantthrough glacial-interglacial
times, Vogelsang[1990] composeda recordof mean-ocean
6280(6w)change
forthelast400ka. This5• record(Figure
2a,righthandscale)
documents
global8•O fluctuations
of the
worldocean'swatermasses
whichweredrivenby the growth
anddecayof globalice volumein the courseof glacial-inter-
glacial climaticchange.The mean-ocean
6w valuesvary
between-0.2%o6 •80 and + 1.0%o5 •80 (SMOW,standard
meanoceanwater).Cross-spectral
coherencyis high between
z
0
•
0.05
the mean-ocean
6w recordand planktonic
6•80 from core
KLll, thusallowingdirectcomparison
of bothrecords(Table
2). The recordof planktonic8•80 changein excessof the
mean-ocean
6w changeis shownin Figure2b. This record
showsthatthe offset(A6•80)betweenplanktonic
•80 and
mean-ocean6w was systematically
larger duringglacial
C , i , i i i arii
d
0.025
Lu
0
O -0.025
n' -0.05
humid
0
50
100
150
200
250
300
350
400
Figure2. (a) Planktonic8•80 recordfromcoreKL11 in the
centralRedSea,showingglacial-interglacial
amplitudeswhich
periods,whereasduringfull - interglacialperiodsthe offset
wassimilarto or evenslightlylessthantoday.Superimposed
on the long-term trend are short-livedanomalieswhich cor-
are
up(right-hand
to3times
higher
than
those
ofthe
mean-ocean
index,reflecting
record
scale;from
Vogelsang,
1990]).
The 8••(• relatewith minimain theorbitalprecession
value of + 3.3%o for the aplanktoniczone of the LGM was
obtainedfromplanktonicpteropodsand is plottedout of scale
theinfluence
of precessional
variabilityon low-latitudeclimate
(see below).
(see text for description
of pteropodversusG. ruber•0
The oxygen isotopecompositionof seawater in the Red
Sea changesby 0.29%o for each 1%o changein salinity
oxygen
isotope
stages.
(b)Deviation
(A•O) of theplanktonic [Craig, 1966]. Usingthis slope,we havecomputeda salinity
calibration).Numbersalong the planktonicrecordindicate
•O record
fromthemean-ocean
•wrecord.
Right-hand
scales scalefromthe A6•80 record(Figure2b, right-hand
scale).
give salinityanomaly(AS) relativeto today(modern) (S set
to zero)andabsolutesalinityestimates.Salinitywas estimated
Inferredglacial-interglacial
salinitiesin the Red Sea variedby
up to 10%oduringthe past380,000 years.For the LGM the
wasestimatedby subtracting1.2%ofrom the
represent insolation peaks in precession [Rossfgnol- salinityincrease
of 5.5%obetween
theLGM andtheHolocene
$tHck,1983]. (c) Precessionindex of Berger and Loutre 5180amplitude
to
account
for
global
ice
volume
change. The residue of
[1991].TheA6•O andASanomalies
seenin (Figure2b)are
coherentwith the precessionindex pointingto low-latitude 4.3%o translatesinto a salinity changeof 14.8%o (i.e.,
climatic forcing of the isotope-salinity
anomalies,perhaps 4.3/0.29). Adding 1%osalinityto allow for a globalsalinity
monsoon•variability.The orbitalprecessionrecordhasbeen increase due to the effects of increased ice volume and lower
lagged by 3 kyr to accountfor the phaselag which was
sea level we arrive at an overall salinityincreaseof 15.8%o
determined
by cross-spectral
analysis(seeFigure5 and Table
for theLGM. This numberis reducedby 1.5-2.0%o(salinity)
2).
if we allow for a glacial temperaturedecreaseof 2øC
[CLIMAP Project Members, 1981; Thunell et al., 1988],
availablefor isotopeanalysis.For this intervalwe measured which would be equivalentto a 8•80 increaseof 0.5%o.
a sample containingthree clean aragoniticshellswith no Previousstudiesinferredsimilarsalinityor slighlylower insecondary
aragonite
of theplanktonic
pteropodspeciesCreseis creasesfor the glacial maximum Red Sea [Berggrenand
acicula.Sediment
trapsamples
takenoff Bermudahave shown Boersrna,1969; Deuser and Degens, 1969; Deuser et al.
that the •sO valueof this species
is about0.3%0more 1976;Reisset al., 1980; 7hunellet al., 1988;Almogi-Labin
negativethanthe valueof aragonitewhichis precipitatedin
et al., 1991].Our estimate
is supported
by abundance
changes
isotopicequilibriumwith ambientseawatertemperature
and of planktonicforaminiferal specieswhich are sensitiveto
isotopecomposition
[Jasperand Deuser,1993]. Twenty-six salinitychanges(seebelow).
pairedisotopemeasurements
on Holocenesamples
fromcore
In additionto the planktonic
6•80 we havegenerated
a
usinga ASw:AS
ratioof 0.29 [Craig,19661.Theshadepeaks
KL11 give a meanfi•sOoffsetbetweenC. aciculaand G.
benthic6180recordusinga combined
isotopicrecordof
tuberof-1.4 +_0.12%0 (20). We usethis value to convertthe
epibenthicHanzawaia sp., Cibicidesmabahethiin order to
150
HEMLEBEN
ET AL.'
GLOBAL
SEA LEVEL
AND
RED SEA HYDROGRAPHY
Table 2. Cross-Spectral
CoherenceandPhaseAngles
Coherence
Phase
Coherence
An•le
Mean ocean/•w
a
0.877
-6ø+ 14.8ø
0.749
Phase
Coherence
Phase
An•le
An•le
-15ø+22 ø 0.760
-10ø+22 ø
SPECMAP
0.949
-3 ø+9 ø
0.898
July 65øN
-
-
0.683
-71ø+26 ø 0.903
-3 ø+13 ø
0.966
-50ø+13 ø
July15øN
-
-
-
-50+ 13ø
0.905
-1 ø+9 ø
80% confidenceintervalfor coherencyis 0.439; 95% confidence
intervalis 0.706; band
with is 0.01; time stepis 2 ka; negative
phaseanglesindicatephaselagsof KL11 •J•80
relative to reference record.
aFromVogelsang[1990].
trace the hydrography of the deeper water masses.The
benthie
õ•80recordshows
thesamehigh-amplitude
signalas
the planktonicõ•80 record(Figure3). Even thoughthe
resolution of the benthie record is much lower due to lack of
specimens,one can still observethe close correlation between
õ•sOmaxima.As in theplanktonic
õ180recordthe LGM
valuesare the mostpositivevalues(+4.7%0) observedin the
entirerecord. This suggeststhat the long-termclimate signal
(i.e., the sealevelsignal)was communicatedfrom the surface
to the deeper water masses.Convertingthe benthie isotope
signal into salinity values, our measurementswould correspond to approximately55%0.
The accuracyof our salinityestimatesdependsnot only on
theappliedcorrection
for theG. ruber-C.aciculaõ•0 offset
butalsoonwhetherthemodernõ,•sslopeof 0.29 appliesfor
theextremehydrographicsituationof the LGM and previous
glacials.As Rohling[1994] hypothesizes,
the oxygenisotope
fluxes in the Red Sea may have been different during the
LGM due to evaporationfluxesunderenhancedwind stress.
If true,theassociated
enrichmentin !sOof the surfacewaters
due to evaporationwould have dropped(Rohling suggestsa
dropof some50 %) even if the ratesof evaporationremained
the same as todays. This would imply that our salinity
estimates,
whichwerecalculated
usingthe presentrelationship
between
waterloss
to evaporation
andIsOenrichment,
would
underestimate
theõ•80andsalinityenrichment
duringfull glacial
periods.
Thatis,thesalinityincrease
duringpe,riods
of
maximumglaciationwould have been even larger than the
increasesuggestedaboveof nearly 16%0.
I
,
I
,
I
,
I
,
I
,
I
,
I
I
•o
The direct correlationbetweenmaximuminferred salinity
and full - glacialclimateclearlysuggeststhat sea level was the
drivingforcebehindthe long-term salinitychange[Thunellet
a/., 1988]. As sealevel falls, the geometryof Bab el Mandeb
(thestraitconnecting
the Red Sea and the Gulf of Aden) becomesnarrower,andwater massexchangebetweenthe Red Sea
andthe openoceanbecomeseven more restrictedthan today.
Thus salt builds up in the Red Sea until an equilibriumis
reached between salt import from the open ocean and
freshwaterloss(due to evaporation)and salt exportfrom the
Red Seathroughoutflowto the Gulf of Aden. Figure 4 shows
the correlationbetweensea level and water masssalinity in
the Red Sea. This correlation was calculated using the
numericalmodel of Assafand Hecht [1974], which is similar
to other sea strait models [e.g. Anati, 1980; Bryden and
Stornrnel,1984; Brydenand Kinder, 1991] and predictsRed
Seasalinity as a functionof the sea strait'sgeometry,evaporation, and salt import. For the model's parameterizationwe
useda modernsurfaceareaof the Red Seaof 424 x 103km2,
width and depth of Bab el Mandeb of 37 km and 160 m,
respectively,
anda salinityof 36.6%o of the intowing surface
waters.For the LGM sea level was set to -120 m [Fairbanks,
1989; Brachertand Dullo, 1991], surfacearea to 219 x 103
km2 (i.e. 50 % of the modernsurfacearea),andwidthand
depth of Bab el Mandeb to 12 km and 40 m; water loss to
evaporation
waskeptconstant
(i.e., 6.3 x 10'8m s-I) [Thunell
et al., 1988; Rohling, 1994]. All parameterswere changed
linearily betweentheir LGM and modern valuesas sea level
was raised from glacial maximumlow standto interglacial
high stand.
The commonlypreferreddepthat Bab el Mandeb is 137 m
[Locke and Thunell, 1988; Bethoux, 1988], which would
imply a LGM sill depthof 17 m. Using suchshallowsill
depth, the Assaf and Hecht model is no longer stableand
breaksdown. Thunellet al. [ 1988]useda LGM sill depthof
60 m to be able to run the Assaf and Hecht model. Based on
the sporadicpresenceof planktonicforaminifera and continuous occurrence
'
0
I
50
'
I
100
'
I
150
'
I
200
'
J
'
250
I
300
'
I
350
of benthie
foraminifera
in low numbers
throughoutthe LGM [Risch, 1976; Locke and Thunell, 1988;
400
AGE (1000 ¾oars B.P.)
Figure3. pibenthic
õ!SOrecordfromHanzawaia
sp. and
Yusuf, 1978], we concludethat a moderatewater exchange
throughBab el Mandeb still existedat the LGM and that the
sill musthavebeentectonicallyupliftedsincethen.Despite
these uncertainties, we use the Assaf and Hecht model to
Cibicides mabahethi. The record showsthe same high-
estimatethe effectsof global sea level changeon the Red
amplitudepatternas the planktonicõlsOrecordshownin
Sea's salinity.
Figure2a. The epibenthicfree zone characterizesan interval
during isotopestage3.
60 m during the LGM) a LGM sea level fall of 120 m would
Usingtheaboveparameterization
(includinga sill depthof
HEMLEBEN
Sa
ET AL.' GLOBAL SEA LEVEL AND RED SEA HYDROGRAPHY
I i n ity
40
45
(est.)
periods(Table2). Thishighcoherencyis expectedbecausewe
have usedthe SPECMAP record to calibratethe age model
for core KL11. The high coherenciesand low phaseangles
confirm that the age model of core KL11 reproducesthe
orbital SPECMAP age model to within 0.4-1.3 ka (Table 2).
50
-2O
Thecross-spectrum
showsphasedifferences
between5180
49.9
/
'40
151
and insolation of 71ø_+26 ø and 50 ø_+13 ø in the orbital
S(est)
obliquity and precessionbandswhich are equivalentto time
lagsof 8 kyr and3 kyr (5180lagsinsolation
(Figure5, Table
-6O
2)). UsingJuneinsteadof July 65 øN insolationas the forcing
-80
LGM'
AS = 15.8
S = 53.8
AS
11.9
ß-100
/
e -12o LGM
referencewouldshiftthelag between5180andinsolation
in
the precessionband to 5 kyr. These are establishedlags
between orbitally tunedclimate proxiesand climate forcing
[lmbtieet al., 1984].The planktonic
5180recordfromcore
KL11 thusdirectly reproducesthe globalclimaterecordand
confirms that the record primarly shows global climate
variationwith a strongoverprintof sealevel effectswhich are
documented
in the greatlyenhanced
amplitudes
of the 5180
-140
0
5
10
15
20
signal.
Spectralanalysisof paleoclimaticrecordsfrom the Arabian
Salinity
Sea has shown significantcoherencebetween biological,
chemical,andlithogenictracersand insolationvariancein the
Figure 4. Red Sea salinityas a functionof globalsealevel. precessionband with a phaselag of 122ø correspondingto a
The salinitywasestimated
usingthe empiricalseastraitmodel time lag of 8 kyr [Clemensand Prell, 1990; Clemenset al.,
ofAssafandHecht[1974].Model parameters
are givenin the 1991]. This has been used as evidence for a strong
text.Theglacial-interglacial
5180amplitude
of 5.5%0implies contribution of monsoonalcirculation to the paleoclimatic
a salinity increaseof nearly 16%oat the LGM. This high
signalsin responseto radiativeforcing, cross-equatorial
heat
salinityis not predictedby the waterbalancemodel,pointing
transportand heat releasefrom the Tibetan plateau[Clemens
to climatically - driven salinity changes(evaporationover
precipitation)which add to the sea level effect on the Red et al., 1991].
At presentwe have no independenttracer from core KL11
Sea's salinitychanges.
to test the monsoonal influence on the Red Sea area. Variance
in the precessionband is slightlyenhancedin the planktonic
resultin a salinityincreaseof 12%obringingsalinitiesin the
centralRedSeato 50%o(Figure4). The numericallyderived
salinityincreaseis about4ø6obelow the estimatederived from
5180recordcompared
to theSPECMAPstack(Figure6).
Sincethe KLll 5•80 recordis phase-locked
to globalice
volumeby usingtheSPECMAPagemodel,we cannotusethis
the planktonic5180 enrichment
duringthe LGM. This enhancedvarianceas independentevidencefor monsoonal
discrepancy
may be a resultof our modelparameterization, signals. However, in view of the presence of strong
namely, the assumeddepth of the Bab el Mandeb sill.
monsoonalsignalsin palcoclimaticrecordsfrom the Arabian
However,aswe will showbelow,thepresence
of distinct5180 Sea [e.g., Prell, 1984; Kutzbach and Street-Perrot, 1985;
andfaunalanomalies
throughout
our recordsimpliesthatlocal Prell and van Carnpo,1986; Andersonand Prell, 1992; Prell
climaticinfluences
affectedsalinityin additionto the dominant and Kutzbach,1992] just outsidethe Red Sea, we speculate
thattheenhanced
precessional
5180variancein coreKL11, in
sealevel forcing.
conjunctionwith faunalanomalies(seebelow), is a prelimiSpectral Analysis
nary indicationof palcomonsoon
signalsin the Red Sea.
Faunal Recordsof Paleohydrographyand Paleoclimate
In ordertocompare
statistically
ourplanktonic
5180record
withtherecordof climatechange
we performedcross-spectral The abundance
patternof planktonicforaminiferaalong
analysis
usingthe recordof 65øN July insolation[Bergerand core KLll may be used as an independentindicator of
Loutre, 1991] as climaticforcingreference(Figure5). We palcosalinity,sincemostforaminiferalspecieshavea narrow
have smoothed
the planktonic5180recordwith a Gaussian rangeof tolerancewith respectto salinity.The abundanceof
filter applyinga 6 kyr wide filter window and an effective planktonic
foraminifera
closely
followsthe5•80-climate
signal
time step of 2 kyr. The cross-spectral
procedurefollows in that abundances
decreasefrom full - interglacialto full standard
routinesdescribedby Imbtie et al. [1984]. The cross glacial conditionswhen salinitiesincrease(Figure 7a). The
spectrum shows siginificant coherencebetween the KLll
distribution pattern shows short-term abundance maxima
planktonic
5•80recordand65øNJulyat theorbitalobliquity duringhighsealevel stands,which correlatewith low, close(41 kyr) and precession(23 kyr) periods(Figure6). High- to-modern
salinityvaluesasdeducedfromtheplanktonic5•80
latitudeinsolationvarianceis low in the orbitaleccentricity record,implyingthat rich planktoniccommunitiesdominated
period (100 kyr; (Figure 6, Table 2)), and thusthe crossduringhumidclimateswhen salinitywas low due to increased
spectrumcannotevaluatethe coherenciesin this frequency moisture
transport
by themonsoonal
winds.Duringperiodsof
band.UsingtheSPECMAP5•O stack[ImbrUe
et al., 1984]as enhancedaridity and increasedsalinity the planktonic
a reference
shows
higheohereneies
at all threeprimaryorbital community
wasdecreased
anddominatedby Globigerinoides
152
HEMLEBEN
ET AL.'
GLOBAL
SEA LEVEL
Bandwidth for65 lags
AND
RED SEA HYDROGRAPHY
Coherency cc•fidence
interval
at the5
% level
1.00
+
++
+++'•-++ +
!
++ /
+ +
•+%I
+
+
KLll
6180
+ +++
+ JULY 65•
+
++++-•-
+-•-+
++
+
+ ++++
+
++
+++
.90
.80
Non-zero coherency
.70
•5
level)
.60
.50
pFr•1o
d
.01
100.00
.02
50.00
Bandwidth for65
Phase
.03
33.33
.04
.05
25.00
20.00
Confidence
lags
.06
16.67
interval
.07
14.29
at
.08
12.50
the5%
.09
11.11
.10
10.00
level
124
Angle
93
Precession
31'
o',
.........
:................ .-•.......
. ....
....... ......:.08
.........:.o•
.........'.•o
-31i
100.0025.00
20.D0
I1I'67
14.29
12.50
11.11
10.00
Period
-62!
(aT,kyr)
,C•,
l,i,
,•,,,,t,y
'] Til•
Lag
.................
Obl: AT = 8.04!3.04
kyr
ß
i;;•;
Period
100.00
-
•0 0
33
,,
25.00
20.00
.67
.
Figure 5. Cross-spectral
analysis
of planktonic
5'80 fromcoreKLll and 65øN Julyinsolation
calculated from algorithsof Berger and Loutre [1991]). Both time seriesare coherentat the orbital
obliquityand precession
periods.Phaselagsof 70ø and50ø are observedfor KL11 6'80 whichare
equivalentto timelagsof 8 kyr and3 kyr. Theseare characteristic
lagsbetween6'80 andclimatic
forcing[e.g./mbr/eet al., 1984],thuspointingto high-latitude
climateastheprimarycontrolon 6•80
in core KL11. Positivephaseanglesindicatephaselags.
HEMLEBEN ET AL.' GLOBAL SEA LEVEL AND RED SEA HYDROGRAPHY
153
PREC
% 100 July
insolation,
15øN
Julyinsolation,65øN
95%
%100 SPECMAP,
(•180
90
%100 KL11,
(•180
•
PREC
90
78%
•0
80
,
,
ECC
65%
70
a)
60
6O
•
50
50
>
40
ECC
58%
4O
OBL
OBL
27 %
3O
27 %
PREC
OBL
PREC
16%
14%
6%
ECC
20•
17%
OBL
0.2 %
0.3 %
O"
0•
0
Figure6. Percent
contribution
of variance
in theeccentricity
(ECC),obliquity
(OBL),andprecession
(PREC)periodsto thetotalvarianceof Julyinsolation
at 65ø and 15øN,andof b•80 in theSPECMAP
stack[ImbHeet al., 19841andalongcoreKL11.Variance
in theorbitalprecession
bandis slightly
enhanced
inKL11b•80,pointing
tolow-latitude
climatic
forcing,
possibly
monsoonal
variability.
0
•
'gg 6.•E
._
•-•
'
50
100
150
200
250
300
350
a ' ' ' ' ' '
!:'..,•
,•
4-
i",,",
i l ,.
400
rn
,4
-46 •
-42 •
2-
-38.•
! '1 "1
• 1øø
i -,
54
FFI
502
• 60
o
o
40
..o
20
i - -
0
5'0
1•)0
1•0
2•)0
AGE (1000years
2•;0
",,,,,,.
?,,,:![
42
•
'/ , i-:i 38
•
3•0
3•0
400
B.P.)
Figure 7. (a) Correlation
of planktonic
foraminifera
with salinity.The abundance
distribution
of
planktonic
foraminifera
(solidline)followstheclimaticsignalalongcoreKL11 andcorrelates
withthe
inferredchanges
of paleosalinity
(dotted
line).Theincrease
of salinityduringfull - glacialperiods
resulted
in anabundance
decrease
of planktonic
foraminifera.
(b) Correlation
of Globigerinoides
ruber
withsalinity.G. rubertolerates
maximum
salinities
up to 49%o[Bijrnaet al., 1990].Theabundance
patternof thisspecies(solidline) confirmsour paleosalinity
reconstruction
(dottedline) in that it
disappeared
only duringthe LGM whenour inferredsalinitiesincreasedto 53%oor more.
154
HEMLEBEN ET AL.: GLOBAL SEA LEVEL AND RED SEA HYDROGRAPHY
ruber, which is one of the most flexible species among
planktonicforaminiferaand can toleratesalinitiesbetween22
and49%o [Bijrnaet al., 1990]. The continuouspresenceof G.
ruberthroughoutthe record, exceptduringthe LGM (Figure
7b), supports
our palcosalinity
reconstruction
in that it implies
that salinitiesnever exceeded49%o. It is only during the
LGM, thatG. ruberdisappears
from the centralRed Sea. Our
salinityreconstructions
imply a palcosalinity
above53%ofor
the LGM whichwouldhave been abovethe upper limit of the
salinity range which this speciescan tolerate in laboratory
cultures, thus causingG. ruber to vanish.
By identifying the abundancyof G. ruber, we obtaineda
pattern which may reflect the variationsof the incoming
nutrient rich intermediatewater driven by the SW monsoon
(Morcos, 1970). Today this water massreachesup to 18øN
andaffectsthe distributionpatternof planktonicforaminifera.
During times of weaker or strongermonsoonactivity, this
boundarymay have shiftedfarther north (strongermonsoon)
or south (weaker monsoon).However, this can be verified
only by investigatingcoresin areasnorth and southof our
core location, which will be done in the future.
A general trend toward lower glacial-interglacial
amplitudes is observed in total abundancevariations of
planktonicforaminifera,whichgoesalong with a trend toward
more enhancedchangesin G. ruber, may indicatea general
increasein Red Sea salinitiespotentiallyin conjunctionwith
a tectonicuplift of the sill at Bab el Mandeb during the past
380 kyr.
Between13 and21 ka all planktonicforaminiferavanished
from the Red Sea resultingin a so-calledaplanktoniczone
(Figure 7) [e.g., Locke and Thunell, 1988; Almogi-Labinet
a/., 1991].Duringthisperiod,lithitler aragonitelayers(LAL)
up to 1 m thick [Milliman et al., 1969] were depositedand
seemto be almostbarrenof planktonicforaminifera.It is only
1 cm abovethe LAL, or about180 yearsafter the aplanktonic
zone [cf./llmogi-Labin et al., 1991] that the first planktonic
foraminiferareappearedat site KL11 and another 1-4 cm, or
up to 720 yearsaRer the end of the aplanktoniczone, that the
planktonic foraminiferal community was almost fully reestablished.Apparently, environmentalconditionsimproved
very rapidly after 13 ka due to rising sea level. Thus the
foraminiferal fauna appearedagain, but was presentstill in
low abundances. Between 9 and 8 ka, when climatic
conditionswere at an optimum, the total planktonicand
benthicfaunawas fully reestablished.
During the last deglaciationeastAfrican climateschanged
from arid to fully humidconditionsin responseto an orbitally
- driven increaseof monsoonalrainfall [Kutzbach and Guetter,
1986;Pachurand KrOpelin,1987; COHMAP Members, 1988;
Pachur et al., 1990; Bonnefilleet al., 1990; Gasseet al.,
1990;Lezineand Casanova,1991]. Even thoughthe Red Sea
today receiveslittle or no freshwater from surroundingland
areas,we speculatethatincreasedhumidity may have lowered
evaporation rates over the Red Sea, thereby enhancingthe
effectsof sealevel riseon loweringsalinities.This would have
contributed to the speedat which the faunal communityreestablisheditself during the last deglaciation.Together with
thepresence
of strong
precessional
variancein the/5•80record
from coreKL11, thisleadsus to concludethat duringthe past
380 kyr not only was the hydrography of the Red Sea
controlledby global sea level variation but that variationsin
monsoonal
strengthalsocontributedto the freshwater balance
by altering humidity and evaporation.
Summary
CoreKL11 from the centralRed Sea providesa continuous
sedimentaryrecord for the last 380 kyr showingglacial-
interglacial
5•80amplitudes
thatareupto 3 timeshigherthan
those of the mean-ocean5•80 record. Correctingthe
planktonic5•sO recordfor mean-ocean•S5
O, we derive
salinifiesof up to 53%o or even slightlyhigherfor the LGM.
Thesemeasured
andcalculatedvaluesare approximately
3%o
higher than previouslyassumedvalues. During the glacial
maximum, isotope stage 6 and 10 salinities were also
significantlyincreasedbut lower than the maximumvalues
inferredfor the LGM. The planktonicfaunavaried alongwith
thesalinitychanges
andwasdrasticallyreducedduringglacial
maxima due to the high salinitylevels;duringthe LGM an
aplanktoniczone developedowing to the highestsalinities
inferred for the entire 380 kyr period which apparently
resulted in extremely hostile conditions.Global sea level
variationsare the main factor in controlingthe Red Sea's
salinity. However, sea level cannotexplainthe full rangeof
salinitiychanges
deduced
fromour isotopedata. An additional
climatically - driven componentis neededto add to the sealevel-drivensalinitychanges.This componentis conceivably
linked to variationsin monsoonalstrength.Supportfor this
contentionis provided by enhancedvarianceof planktonic
5•sOin theorbitalprecession
band.
Acknowledgements. We thank the master,crew and scientists
aboardR/V Meteor (cruise5) for their helpandassistance.
We also
thankI. Breitinger,J. Erbaeher,G. Sehmiedl,D. Mfihlen,and R.
Ott,(allat University
of Tfibingen),F. Clasen,B. Gehrke,U. MeliB
(all at Universityof Gtttingen)and H.H. Cordt, I. Klein, and H.
Heeht(allat University
of Kiel)for theirhelpduringdataacquisition
andin preparing
themanuscript
andW. Prell, BrownUniversity,for
using freely distributedsoftware. We are also grateful to D.
Andersonand an anonymous
reviewerfor constructive
criticism
whichhelpedto improvethismanuscript.
This research
wasfunded
bytheDeutscheForsehungsgemeinsehaft
undercontracts
He 697/7
and Me 267/28, and the G.I.F., the German-lsraeliFoundationfor
ScientificResearchand Development.
References
Almogi-Labin,
A., Stratigraphic
andpaleoceanographic
significance
of late Quaternary
pteropods
fromdeep-sea
coresin theGulf of
Aqaba(Elat)andnorthennnost
RedSea,Mar.Micropaleontol.,
7,
53-73,1982.
Almogi-Labin,
A., B.Luz,B.andJ.-C. Duplessy,Quaternary
paleo-oceano•hy, pteropod
preservation
andstable-isotope
recordof
the Red Sea. Palaeogeogr.Palaeoclimatol.
Palaeoecol.,75,
195-211, 1986.
Almogi-Labin, A., Ch. Hemleben, D. Meischner, and H.
Efienkeuser,Palcoenvironmental
eventsduringthe last 13,000
years in the central Red Sea as recordedby Pteropoda:
Paleoceanography,
6, 83-98, 1991.
Anati, D.A., A parameterization
of the geometryof sea straits,
Oceanol. Acta, 3, 395-397, 1980.
Anderson, D.M., and W.L. Prell, The structureof the southwest
monsoon
windsovertheArabianSeaduringthe Late Quaternary:
Observations,simulations,and marine geologicevidence,J.
Geophys.
Res., 97, 481487, 1992.
Assaf,G., andA. Hecht, Seastraits:A dynamicalmodel,Deep Sea
Res., 21,947-958, 1974.
HEMLEBEN ET AL.: GLOBAL SEA LEVEL AND RED SEA HYDROGRAPHY
Berger,A., andM.F. Loutre,Insolation
valuesfor the climateof the
last 10 millionyears,Quat. Sci. Rev., 10, 297-317, 1991.
Berggren,W.A., and A. Boersma,Late Pleistocene
and Holocene
planktonicforaminiferafrom the Red Sea, in Hot Brinesand
Recent Heavy Metal Depositsin the Red Sea, edited by E.T.
Degensand D.A. Ross, pp. 282-298, Springer-Verlag,New
York, 1969.
Bethoux,J.P., RedSeageochemical
budgetsandexchanges
with the
Indian Ocean, Mar. C-hem.,24, 83-92, 1988.
Bijma, J., W.W. Faber Jr., and Ch. Hemleben,Temperatureand
salinity limits for growth and survival of some planktonic
fomminifers
in laboratory
cultures,J. ForaminiferalRes., 20, 95116, 1990.
Bonnet'die,
R., J.C. Roeland,andJ. Guiot, Temperatureand rainfall
estimatesfor the past40,000 yearsin equatorialAfrica, Nature,
346, 347-349, 1990.
~
Bracheft, T.C.,
and W.C.
Dullo, Laminar micrite crusts and
associated
foreslope
processes,
Red Sea,J. Sediment.Petrol., 61,
354-363, 1991.
Bryden,H.L. andT.H. Kinder, Recentprogressin straitdynamics,
U.S. Natl. Rep. lnt. Union Geod. Geophys.1987-1990, Rev.
Geophys.,29, 617-631, 1991.
Bryden, H. L., and H. M. Storereel, Limiting processesthat
determine
basic features of the circulation
in the Mediterranean
Sea, Oceanol. Acta, 7, 289-296, 1984.
Chen, C., Pteropodsin hot brine sedimentsof the Red Sea, in Hot
Brinesand RecentHeavy Metal Depositsin the Red Sea, edited
by E.T. DegensandD.A. Ross,pp. 313-316, Springer-Verlag,
New York, 1969.
Clemens, S.C. and W.L. Prell, Late Pleistocenevariability of
155
Jasper,J.P., and W.G. Deuser,Annualcyclesof massflux and
isotopiccomposition
of pteropodshellssettlinginto the deep
Sargasso
Sea,Deep SeaRes.,40, 653-669, 1993.
Kutzbach,J.E., and P.J. Guetter,The influenceof changingorbital
parameters and surface boundary conditionson elimate
simulations
for thepast18,000years,J. Atmos.Sci., 43, 17261759, 1986.
Kutzbaeh,J.E.,, and F.A. Street-Perrot,Milankoviteh forcing on
fluctuationsin the level of tropicallakesfrom 18 to 0 ka BP,
Nature, 317, 130-134, 1985.
Labeyrie, L. D., J.-C. Duplessy,and P.L. Blanc, Variationsin
modeof formationandtemperatureof oceanicdeepwatersover
thepast125,000years,Nature,327, 477-482, 1987.
Lezine,A.-M., andJ. Casanova,Correlatedoceanicand continental
recordsdemonstrate
pastelimateandhydrologyof NorthAfrica
(0-140 ka), Geology,19, 307-310, 1991.
Locke, S. and R.C. Thunell, Paleoceanographic
recordof the last
glacial/interglacialcycle in the Red Sea and Gulf of Aden.
Palaeogeogr.
, Palaeoclimatol.Palaeoecol.
, 64, 163-187, 1988.
Milliman, J.D., D.A. Ross and T.-L. Ku, Precipitationand
lithifieationof deepseacarbonates
in the Red Sea,J. Sediment.
Petrol., 39, 724-736, 1969.
Moreos,S.A., Physicalandchemicaloceanography
of the Red Sea,
Oceanogr.Mar. Biol. Annu.Rev., 8, 73-202, 1970.
Niitsuma, N., T. Oba, and M. Okada, Oxygenand carbonisotope
stratigraphyat Site 723, Oman Margin, Proc. Ocean Drill.
Program.,Sci. Results,117, 321-341, 1991.
Pachur,H.J., andS. Kr6pelin,Wadi Howar:palcoclimatic
evidence
fromanextinctfiver systemin the southeastern
Sahara,Science,
237, 298-300, 1987.
Arabian Sea summermonsoonwinds and continentalaridity: Paehur, H.J., S.T. Kr/Spelin,P. Hflzmann, M. Gfsehin, and N.
Altmann, Late Quaternaryfluvio-laeeustrineenvironmentsof
Eolian recordsfrom the lithogeniccomponentof deep-sea
westernNubia, Berliner Geowiss.Abh. ,1, 120, 203-260, 1990.
sediments,Paleoceanography,
5, 109-145, 1990.
Clemens,S., W. Prell, D. Murray, G. Shimmield,and G. Weedon, Prell, W.L., Monsoonalelimateof the ArabianSeaduringthe late
Quaternary:a responseto changingsolar radiation,in MilanForcingmechanisms
of the Indian Ocean monsoon,Nature, 353,
kovitchand Climate, editedby A.L. Bergeret al., pp. 349-366,
720-725, 1991.
CLIMAP ProjectMembers,Seasonalreconstruction
of the Earth's
surfaceat the lastglacialmaximum,Geol. Soc.Am., Map Chart
Ser., MC-36, 1981.
COHMAP Members, Climatic changeof the last 18,000 years:
Observations and model simulations,Science, 241, 1043-1052,
1988.
Craig,H., Isotopiccomposition
andoriginof the Red SeaandSalton
Seageothermalbrines.Science,154, 1544-1548, 1966.
Deuser, W.G. and E.T. Degens,O18/O16 and C13/C12 ratiosof
fossilsfromthehot-brinedeepareaof the centralRed Sea, in Hot
Brinesand RecentHeavyMetal Depositsin the Red Sea, edited
by E.T. Degensand D.A. Ross,pp. 336-347, Springer-Verlag,
New York, 1969.
_
Deuser,W.G., E.H. Ross,and L.S. Waterman,Glacialandpluvial
periods:Their relationshiprevealedby Pleistocene
sediments
of
D. Reidel, Norwell, Mass., 1984.
Prell, W.L. andJ.E. Kutzbaeh,Sensitivityof the Indian monsoonto
forcing parametersand implicationsfor its evolution,Nature,
360, 647-652, 1992.
Prell, W.L. and E. van Campo, Coherentresponseof ArabianSea
upwelling and pollen transportto late Quaternarymonsoonal
winds, Nature, 323, 526-528, 1986.
Reiss, Z., B. Luz, A. Almogi-Labin, E. Halicz, A. Winter, M.
Wolf, andD.A. Ross,Late Quaternarypaleoceanography
of the
Gulf of Aqaba(Elat), Red Sea, Quat. Res.,14, 294-308, 1980.
Risch, H., Microbiostratigraphy
of core-sections
of the Red Sea,
Geol. Jahrb. 17, 3-14,1976.
Rohling,
E.J.,Glacial
conditions
in theRedSea,Paleoceanography,
9, 653-660, 1994.
Rossignol-Strick,
M., African monsoons,an immediateclimate
response
to orbitalinsolation,Nature, 304, 46-49, 1983.
the Red Sea and Gulf of Aden, Science, 191, 1168-1170, 1976.
Sirocko,F., M. Sarnthein,H. Erlenkeuser,H. Lange, M. Arnold,
Fairbanks, R.G., A 17,000-yearglacio-eustaticsea level record:
andJ.-C. Duplessy,Century-scale
eventsin monsoonal
climate
Influenceof glacialmeltingrateson the YoungerDryas eventand
overthe past24,000 years,Nature, 364, 322-364, 1993.
deep-oceancirculation,Nature, 342, 637-642, 1989.
Thunell,R.C., S.M. Locke,andD.F. Williams, Glacio-eustaticseaGasse,F., R. T6het, A. Durant, E. Gilbert, and J.-C. Fontes, The
level controlon Red Seasalinity,Nature, 334, 601-604, 1988.
add-humidtransitionin the Saharaandthe Sahelduringthe last
Vogelsang,
E., Pal•.o-Ozeanographie
desEuropiiischen
Nordmeeres
deglaciation,Nature, 346, 141-146, 1990.
an Hand stabiler Kohlenstoff-und Sauerstoffisotope,
Ph.D.
Herman, Y., Evidenceof climatic changesin Red Sea cores, in
thesis,Rep. SFB 313, 23, Kiel Univ. Kiel, Germany,136 pp.,
Means of Correlationsof QuaternarySuccessions,
Proceedings
1990.
Vll Congress
InternationalAssociation
for QuaternaryResearch,
Yusuf, N.,
Mikropal•iontologische und geochemische
vol. 8, editedby R.B. MorrisonandH.E. Wright, pp. 325-348,
Untersuehungenan Borhkernenaus dem Roten Meer, Berl.
Univ. of Utah Press,Salt Lake City, 1968.
Geowiss.Abh. Reihe A, 6, 1-77, 1978.
Imbrie, J., J.D. Hays, D.G. Martinson,A. Mcintyre, A.C. Mix,
Zahn, R. and T. F. Pedersen,Late Pleistoceneevolutionof surface
J.J. Morley, N.G. Pisias,W.L. Prell, and N.J. Shackleton,The
andmid-depth
hydrography
at the OmanMargin:Planktonicand
orbital theoryof Pleistocene
climate:Supportfrom a revised
benthicisotoperecordsat ODP Site 724, Proc. OceanDrill.
chronologyof the marine 61sOrecord, in Milankovitchand
Program,Sci. Results,117, 291-308, 1991.
Climate,editedby A.L. Bergeret al., pp. 269-305,D: Reidel,
Norwell, Mass., 1984.
Ivanova, E.V.,
Late Quaternary biostratigraphy and
palcotemperatures
of the Red Seaand the Gulf of Aden basedon
planktonic
foraminiferaandpteropods.
Mar. Micropaleontol.,9,
335-364, 1985.
A. Almogi-Labin,GeologicalSurveyof Israel, 30 Malkhe YisraelSt.,
95501 Jerusalem,Israel. (e-mail: [email protected])
H. Erlenkeuser,Institutfiir Reineund AngewandteKernphysik,C14-
156
HEMLEBEN ET AL.: GLOBAL SEA LEVEL AND RED SEA HYDROGRAPHY
Labor, University of Kiel, Olshausenstr.40-60, D-24098 Kiel, FR
Germany.(e-mail:[email protected])
C. Hemleben (corresponding
author) and B. Hiller, Institut und
Museum far Geologieund Palfiontologie,University of Tiibingen,
Sigwartstrasse
10, D-72076Tiibingen,FR Germany.(e-mail:Christoph.
[email protected])
D. Meischner,Institutfar Geologieund Pal•iontologie,
Abt. SedimentGeologie, University of G6ttingen, Goldschmidtstrasse
3, D-37077
G6ttingen,FR Germany.(e-mail:[email protected])
R. Zahn, GEOMAR, Research Center for Marine Geosciences,
Wischhofstr.
1-3, D-24148Kiel, FR Germany.(e-mail:RZahn@Geomar.
de)
(Received
May 5, 1995;revisedDecember11, 1995;
accepted
December20, 1995.)