Magnitudes of sea-level lowstands of the past 500,000 years

Abstract

Existing techniques for estimating natural fluctuations of sea level and global ice-volume from the recent geological past exploit fossil coral-reef terraces or oxygen-isotope records from benthic foraminifera. Fossil reefs reveal the magnitude of sea-level peaks (highstands) of the past million years, but fail to produce significant values for minima (lowstands) before the Last Glacial Maximum (LGM) about 20,000 years ago, a time at which sea level was about 120m lower than it is today. The isotope method provides a continuous sea-level record for the past 140,000 years (ref. 5) (calibrated with fossil-reef data), but the realistic uncertainty in the sea-level estimates is around +/-20m. Here we present improved lowstand estimates-extending the record back to 500,000 years before present-using an independent method based on combining evidence of extreme high-salinity conditions in the glacial Red Sea with a simple hydraulic control model of water flow through the Strait of Bab-el-Mandab, which links the Red Sea to the open ocean. We find that the world can glaciate more intensely than during the LGM by up to an additional 20-m lowering of global sea-level. Such a 20-m difference is equivalent to a change in global ice-volume of the order of today's Greenland and West Antarctic ice-sheets.

Full text

Nature © Macmillan Publishers Ltd 1998
8
letters to nature
162 NATURE
|
VOL 394
|
9 JULY 1998
well as through the large-field-gradient positions just above and
below the crown-ether molecules.
We have also made direct measurements of the ionic conductivity
of the crown-ether salt and also of reference materials. We used an
electron-blocking method, in which a compacted disk of the
material under test was sandwiched between Li
+
/polyethylene
oxide disks which were pressed onto Li metal electrodes.
The ionic conductivity of (PEG)-Li(CF
3
SO
3
), where PEG is
polyethylene glycol, was measured to be 1:6 3 10
2 6
Scm
21
at
60 8C, in good agreement with literature values
15
. We measured
values, also at 60 8C, for a range of Ni(dmit)
2
salts: 3 3 10
2 6
Scm
21
for Li
0.6
(15-crown-5-ether) [Ni(dmit)
2
]
2
⋅H
2
O; 2 3 10
2 8
Scm
21
for (C
16
H
33
)
2
(CH
3
)
2
N[Ni(dmit)
2
]; and 3 3 10
2 7
Scm
21
for
(TPP)[Ni(dmit)
2
]
3
, where TPP is tetraphenylphosphonium. Each
measurement was made after the current reached its saturated
value, typically between 30 and 60 minutes after application of
the constant voltage. We have checked that there is no decomposi-
tion (for example, water loss) at 60 8C (by thermal gravimetry), and
consider that the value of ionic conductivity for the crown-ether Li
+
salt indicates clearly that there is significant Li
+
mobility in this
material at this temperature. We have measured significantly lower
values in the (electronically insulating) tetra-alkyl ammonium salt
and the (electronically conducting) tetraphenylphosphonium salt.
We consider that the electron dynamics on the Ni(dmit)
2
chains
are controlled by the dynamics of the Li
+
ions in the crown-ether
columns. Above 250 K the Li
+
ions are mobile, and do not localize
the conduction electrons on the Ni(dmit)
2
stacks, which therefore
show metallic properties. Between 250 and 200 K the Li
+
ions
become immobilized, and provide static coulombic pinning poten-
tials for the conduction electrons on the Ni(dmit)
2
stacks, causing
localization and the magnetic insulator properties evident in Figs 2
and 3. A metal–insulator transition driven by interaction between
ions and electrons in this way is unusual, and the example that we
have illustrated here may provide insights into transport properties
of other mixed ionic/electronic conductors, such as the conjugated
polymer blends recently used by Pei et al.
5
in electroluminescent
diodes.
The electron-transport system with ion channels that we describe
here may allow for the design of electron–cation transport systems
for use in active and selective ion transport. For example, a double
carrier process could be realised in a redox gradient where the
coupled parallel transport of electrons and alkali-metal cations takes
place
16
. M
Received 17 April; accepted 13 May 1998.
1. Je
´
rome, D. The physics of organic superconductors. Science 252, 1509–1515 (1991).
2. Williams, J. M. et al. Organic superconductors—new benchmarks. Science 252, 1501–1508 (1991).
3. Takahashi, T. et al. Solid-state ionics—recent trendsand future expectations. Bull. Electrochem. 11, 1–
33 (1995).
4. Fauteux, D.,Massucco, A.,Mclin, M., Vanburen,M. & Shi, J. Lithium polymerelectrolyte rechargeable
battery. Electrochim. Acta 40, 2185–2190 (1995).
5. Pei, Q., Yu, G., Zhang, C., Yang, Y. & Heeger, A. J. Polymer light-emitting electrochemical cells.Science
269, 1086–1089 (1995).
6. Pregel, M. J., Jullien, L. & Lehn, J. M. Towards artificial ion channels—transport of alkali-metal ions
across liposomal membranes by bouquet molecules. Angew. Chem. Int. Edn Engl. 31, 1637–1640
(1992).
7. Ghadiri, M. R., Granja, J. R. & Buehler, L. K. Artificial transmembrane ion channels from self-
assembling peptide nanotubes. Nature 369, 301–304 (1994).
8. Cassoux, P. et al. Molecular-metals and superconductors derived from metal-complexes of 1,3-
dithiol-2-thione-4,5-dithiolate (dmit). Coord. Chem. Rev. 110, 115–160 (1991).
9. Kim, H., Kobayashi, A., Sasaki, Y., Kato, R. & Kobayashi, H. New radical-anion complex,
[(CH
3
)
4
N][Ni(dmit)
2
]
2
] with metal-semimetal phase-transition. Chem. Lett. 1799–1802 (1987).
10. Guy, D. R. P. & Friend, R. H. Temperaturemeasurement in high pressure cells using a rhodium + 0.5%
iron versus chromel thermocouple pair. J. Phys. E 19, 430–433 (1986).
11. Murata, K. et al. Superconductivity with the onset at 8 K in the organic conductor beta-(BEDT-
TTF)
2
I
3
under pressure. J. Phys. Soc. Jpn 54, 1236–1239 (1985).
12. Estes, W. E., Gavel, D. P., Hatfield, W. E. & Hodgson, D. J. Magnetic and structural characterisation of
dibromo- and dichlorobis(thiazole) copper(II). Inorg. Chem. 17, 1415–1421 (1978).
13. Obertelli, S. D.,Friend, R. H., Talham, D. R., Kurmoo, M. & Day, P. Magnetic susceptibility and EPR of
the Organic conductors a9-(BEDT-TTF)2X, X=AuBr
2
, CuCl
2
and Ag (CN)
2
. J. Phys. Cond. Matter 1,
5671–5680 (1989).
14. Cohen, M. H. & Reif, F. in Solid State Physics (eds Seitz, F. & Turnbull, D.) 321–438 (Academic, New
York, 1957).
15. Robitaille, C. D. & Fauteux, D. Phase-diagrams and conductivity characterization of some PEO-Li
x
electrolytes. J. Electrochem. Soc. 133, 315–325 (1986).
16. Grimaldi, J. J. & Lehn, J. M. Multicarrier transport: coupled transport of electrons and metal cations
mediated by an electron carrier and a selective cation carrier. J. Am. Chem. Soc. 101, 1333–1334
(1979).
17. Burla, M. C. et al. SIR88—a direct-methods program for the automatic solution of crystal-structures.
J. Appl. Crystallogr. 22, 389–393 (1989).
Supplementary information is available on Nature’s World-Wide Web site (http://www.nature.com) or
as paper copy from the London editorial office of Nature.
Acknowledgements. We thank N. Robertson for synthesis of some of the crystals used in this work,
N. Nonose for measurements of the ICP mass spectra, and A. Yap and S. R. Elliott for assistance with the
NMR measurements. This work was supported by Grant-in-aid for Science Research from the Ministry of
Education, Science and Culture, Japan, and the UK Engineering and Physical Sciences Research Council.
One of the authors (A.E.U.) acknowledges support from the British Council.
Correspondence and requests for materials should be addressed to R.H.F. (e-mail: rhf10@cam.ac.uk).
Magnitudes of sea-level
lowstands of the past
500,000 years
E. J. Rohling*, M. Fenton*, F. J. Jorissen
†
, P. Bertrand
†
,
G. Ganssen
‡
& J. P. Caulet§
* Department of Oceanography, Southampton University, Southampton
Oceanography Centre, Southampton SO14 3ZH, UK
†
De
´
partement de Ge
´
ologie et Oce
´
anographie, Universite
´
de Bordeaux I,
CNRS URA 197, Avenue des Faculte
´
s, 33405 Talence Cedex, France
‡
Department of Earth Sciences, Free University Amsterdam, De Boelelaan 1085,
1081 HV Amsterdam, The Netherlands
§ Laboratoire de Ge
´
ologie, National Museum for Natural History,
CNRS URA 723, 43 Rue Buffon, 75005 Paris, France
.........................................................................................................................
Existing techniques for estimating natural fluctuations of sea level
and global ice-volume from the recent geological past exploit
fossil coral-reef terraces or oxygen-isotope records from benthic
foraminifera. Fossil reefs reveal the magnitude of sea-level peaks
(highstands) of the past million years, but fail to produce sig-
nificant values for minima (lowstands) before the Last Glacial
Maximum (LGM) about 20,000 years ago, a time at which sea level
was about 120 m lower than it is today
1–4
. The isotope method
provides a continuous sea-level record for the past 140,000 years
(ref. 5) (calibrated with fossil-reef data
6
), but the realistic uncer-
tainty in the sea-level estimates is around 620 m. Here we present
improved lowstand estimates—extending the record back to
500,000 years before present—using an independent method
based on combining evidence of extreme high-salinity conditions
in the glacial Red Sea with a simple hydraulic control model of
water flow through the Strait of Bab-el-Mandab, which links the
Red Sea to the open ocean. We find that the world can glaciate
more intensely than during the LGM by up to an additional 20-m
lowering of global sea-level. Such a 20-m difference is equivalent
to a change in global ice-volume of the order of today’s Greenland
and West Antarctic ice-sheets.
Our technique relies on evidence of adverse living conditions for
planktonic foraminifera in the glacial Red Sea due to extremely high
salinities, giving rise to so-called ‘aplanktonic’ zones. These contain
only few specimens (mostly Globigerinoides ruber) and are found
not only in our core MD921017 (Fig. 1), but throughout the Red
Sea
7–11
. Planktonic pteropods
10
and benthic foraminiferal faunas
were less affected, although the latter show increased abundances of
high-salinity-resistant miliolid taxa (Fig. 1e; also refs 8, 9). Inorganic
aragonite coatings, cements and concretions indicative of high
salinities are found within the ‘aplanktonic’ zones of stages 2, 6
and 12 (Fig. 1b), in agreement with previous reports
12,13
. Multi-
disciplinary studies of the youngest aplanktonic zone (LGM)
indicate that Red Sea salinities rose to 50 6 2 practical salinity
units (p.s.u.), that is, 10 6 2p:s:u: higher than in the adjacent
ocean
9,10,14
. The highest estimate
11
is 55 p.s.u. Similar faunal reduc-
tions and compositional changes, as well as distinct aragonite

Nature © Macmillan Publishers Ltd 1998
8
letters to nature
NATURE
|
VOL 394
|
9 JULY 1998 163
occurrences, characterize the stage 6 and 12 glacial maxima,
reflecting comparable high-salinity conditions (Fig. 1). Continua-
tion of a more diverse, although severely reduced, fauna through
stages 10 and 8 suggests less harsh conditions, with those during
stage 8 being least restricted.
Arguments for hydraulic control of flow through the Strait of
Bab-el-Mandab (SBM) and conservation of mass and salt
14,15
demonstrate that the glacial freshwater deficit was similar to the
present (2 m yr
−1
), and that the high salinities resulted from
restricted marine exchange through the SBM, which today is only
137 m deep
16
. Continuation of benthic faunas, albeit in reduced
numbers and different compositions, indicates that all glacial sea-
level drops of the past 500 kyr left sufficient communication
between the Red Sea and the open ocean to prevent worse saliniza-
tion and consequent sterilization (Fig. 1c, e). The calculations
14,15
may be rearranged to find the ‘critical’ glacial sill depth (H
crit
) at
which salinities would rise to reconstructed LGM values, and so
trigger an aplanktonic interval. We use the more common range of
estimates for LGM salinities
9,10,14
, with a salinity difference (DS)
across the SBM of 10 6 2 p:s:u:, and in sensitivity tests with
DS ¼ 15 p:s:u: and DS ¼ 10 6 5 p:s:u: we evaluate the potential
effects of the highest-salinity extreme
11
.
A glacial water deficit of 2:0 6 0:5 m yr
2 1
is used, allowing for
uncertainties in sea–air temperature contrasts and freshwater
input of a factor of two, and 62 m s
−1
uncertainty in mean
wind speed variations
14
. As glacial exposure of shelves would
reduce the Red Sea surface area by 50%, volumes of net
glacial evaporation (¼ area 3 deficitÞ amounted to 0:50 6 0:13
times present-day values. Hydraulic control defines Q ¼
W0:375H
crit
ð0:375H
crit
7:4 3 10
2 3
DSÞ
0:5
for the maximum exchange
solution, while Q ¼ ð3:29gQ
p
Þ=DS follows from conservation of
mass and salt
14,15
. Here Q is volume of Red Sea outflow, W is the
glacial width of the shallow passage of the SBM, and g is the ratio of
glacial to present-day net evaporation; subscript p indicates pre-
sent-day value.
Consequently, H
crit
¼ 18 6 5 m, for DS ¼ 10 6 2 p:s:u:, W ¼
11 6 1 km, g ¼ 0:50 6 0:13 and Q
p
¼ 0:32 6 0:03 Sv (1 Sv ¼
10
6
m
3
s
2 1
)
14–17
. Potentially increased mixing of inflow and outflow
for reduced sill depths favours values of H
crit
towards the higher end
of its confidence interval over those towards the lower end. We
combine the present-day sill depth of 137 m with H
crit
to estimate
past sea-level lowstands, but we first evaluate whether—and to what
extent—correction is needed for uplift in the SBM region.
Total planktonic foraminiferal abundances show a distinct
decrease over the interglacial periods of the past 500 kyr in core
MD921017 (Figs 1b, 2a). A similar trend is seen for the past 380 kyr
contained in nearby core KL11
11
. Sedimentation rates throughout
MD921017 are very uniform (see Supplementary Information), so
that the foraminiferal numbers per gram—based on constant
sample thickness—closely approximate true fluxes per unit time.
Figure 1 Results for core MD921017, Red Sea (198 239 240 N, 388 409 840 E, water
depth 570 m). All variables plotted against time in calendar years (kyr
BP)
according to correlation with SPECMAP
22
. a, Oxygen-isotope record (versus
PDB standard) for Globigerinoides ruber in the 250–350 mm size fraction (filled
circles). Numbers refer to SPECMAP events
22
; filled diamonds indicate samples
used for AMS radiocarbon dating (1;666 6 26, 9;681 6 41 and 25;300 6 200 radio-
carbon years
BP, respectively). b, Planktonic foraminifera per gram dry weight (thin
line). Asterisks highlight true aplanktonic zones of oxygen isotope stages 2, 6 and
12 which contain aragonitic coatings, cement and nodules. Heavy line is five-
point moving-average highlighting the general trends. c, Benthic foraminifera per
gram dry weight (thin line), with five-point moving average (heavy line). We note
factor 10 difference between scales in b and c. d, Down-core variation in TOC (%)
(LECO 125-CS elemental analyser). Preparation involved removal of carbonate by
acidification using dilute hydrochloric acid. e, Abundance of miliolids relative to
total benthic foraminiferal fauna. Although of lower resolution than the other
records, the miliolid percentages do show a general increase in abundance
through the record, as well as peaks associated with the aplanktonic zones.
f, Abundance of Globigerinoides ruber relative to total planktonic foraminiferal
fauna. g, Abundance of SPRUDTS group relative to total planktonic foraminifera.
This group represents a cluster of the individually infrequent subtropical species
Globigerinoides sacculifer, Hastigerina pelagica, Globoturborotalia rubescens,
Orbulina universa, Globigerina digitata, Globoturborotalia tenella and
Globigerinella siphonifera
18
. h, Oxygen isotope stages
22
. i, Global sea level
record. The heavy solid line through the past 140 kyr is based on benthic
isotopes
5
and the thin solid line through the past 200 kyr is based on coral reef
terraces and benthic isotopes
4
. Cross-hatched ovals show reported ranges of
interglacial sea-level highstands
2–4,21
. Thick error bars in stages 2, 6, 8,10 and 12
represent ranges of glacial sea-level lowstands according to the model presented
here. The dashed line through stages 8–12 shows schematic sea-level fluctua-
tions sketched through the control points following the main trends in the oxygen-
isotope record.

Nature © Macmillan Publishers Ltd 1998
8
letters to nature
164 NATURE
|
VOL 394
|
9 JULY 1998
Consequently, the decreasing trend in planktonic foraminiferal
numbers reflects: (1) a steady decrease in interglacial productivity
over the past 500 kyr; or (2) a progressive deterioration of living
conditions for foraminifera through increasing isolation of the Red
Sea from the open ocean.
Investigation of other parameters for interglacial periods, to
detect long-term changes unrelated to glacial cycles, provides no
support for the first option. First, the total organic carbon (TOC)
record shows only minor, random fluctuations (Fig. 1d). Second,
there are no long-term trends in the carbon isotope record. Third,
the continuously low-diversity planktonic foraminiferal faunas are
totally dominated by relative fluctuations (within a 100% sum) of
only two taxonomic groups, G. ruber and the SPRUDTS group (see
Fig. 1 legend for details), both of which are typical of warm and
oligotrophic conditions
18
(Figs 1f, g, 2b). Last, there is only a weak,
statistically insignificant, trend in total benthic foraminiferal abun-
dances, whereas a significant decrease would be expected under
conditions of decreasing primary/export productivity (= food
supply) (Figs 1c, 2a).
The second option—increasing Red Sea ‘restriction’— is sup-
ported by a steady increase in the abundances of high-salinity
resistant
8,9
miliolid species in the benthic foraminiferal fauna over
the interglacial periods of the past 500 kyr (Figs 1e, 2c). The
increasing restriction would have occurred within a context of
already limited exchange with the open ocean, as witnessed by:
the lack of successful
19
‘invasions’ into the Red Sea by oceanic species
like Globigerina bulloides, Globorotalia menardii and
Neogloboquadrina dutertrei; the limited and diminishing presence
of Turborotalita quinqueloba (Fig. 2b) (these four species abound
just outside the SBM in core MD921005; 118 359 130 N,
438 319 840 E); and the amplified glacial–interglacial contrasts
throughout the Red Sea oxygen-isotope record relative to those of
the open ocean (Fig. 1a), a typical characteristic of evaporative
marginal basins related to their restriction from the world ocean.
The inferred slow but progressive isolation of the Red Sea, super-
imposed on already restricted communication with the open ocean,
is best explained by slow and fairly continuous uplift of a shallow
sill.
The rate of uplift may be assessed using the present-day sill depth
of 137 m (ref. 16) and previous estimates that sea level stood
around 125 m below present sea level (b.p.s.l.) during stage 6
(135 kyr
BP) (Fig. 1i)
4,5
. Addition of H
crit
gives an estimated stage 6
sill depth of 143 m b.p.s.l., which implies 6 m uplift in 135 kyr
(4.4 cm kyr
−1
). This rate is a factor of 5–10 smaller than the assumed
constant rates of classical reef-terrace-based sea-level studies
1–4
. To
account for uncertainties in the derivation of the rate and/or its
possible temporal nonlinearities, a 50% confidence interval is
added, giving 0:044 6 0:022 m kyr
2 1
. Our model then back-
calculates the stage 6 sea level drop at 137þ ð135 3 0:044Þ 2 H
crit
¼
125 m, with a confidence interval of 66 m based on propagation
20
of
uncertainties in H
crit
(65 m) and uplift rate (60.022 m kyr
−1
).
Validation of the model through determination of the LGM
Figure 2 Trends through interglacial stages. a, Trends in the numbers of
planktonic and benthic foraminifera per gram sediment dry weight, through
interglacial stages 11, 9, 7, 5 and 1. Positions on age axis are simple stage mid-
points. Stage 3 is excluded as this in fact is a warm interstadial within the last
glacial period. Symbols indicate mean values, and error bars an interval of 61
standard error of the mean. Benthic numbers have been multiplied times 10, to
allow plotting on the same scale as planktonic numbers. Equations and r
2
coefficients concern linear fits through the two records. Linear fit for planktonics
shows a significant trend (a ¼ 0:05), for benthics the trend is weak and statisti-
cally insignificant. We emphasise that, as stage 1 is continuing, its values may be
less representative than those of the other (completed) interglacial periods. b,
Mean relative abundances of Globigerinoides ruber and SPRUDTS-group (see
also Fig. 1), and of Globigerinita glutinata and Turborotalita quinqueloba, in
percentages relative to total planktonic foraminiferal fauna, through interglacial
stages 11, 9, 7, 5 and 1. Symbols, error bars and note on stage 3 and 1 values as in
a. c, Trends in high-salinity indicative mioliolid benthic foraminifera—both relative
to total benthic foraminiferal faunta (%), and as absolute abundances (numbers
per gram dry weight)—through interglacial stages 11, 9, 7, 5 and 1. Symbols, error
bars and note on stage 3 and 1 values as in a.
Figure 3 Sensitivity of confidence interval. Changes in the confidence intervals
around the reconstructed mean sea level lowstands for glacial stages, 2, 6, 8, 10
and 12, for glacial Red Sea salinity contrast DS ¼ 10 p:s:u: 6 a variable confidence
interval dDS. Effects are investigated per glacial stage for a range of dDS values
from 0 to 5 p.s.u. Dashed line indicates dDS ¼ 2 as used in this Letter. The sea
level confidence intervals (y axis) essentially change due to modification of the
confidence interval around H
crit
, which by close approximation corresponds to
the line marked ‘‘stage 2’’.

Nature © Macmillan Publishers Ltd 1998
8
letters to nature
NATURE
|
VOL 394
|
9 JULY 1998 165
(20 kyr BP) sea-level lowstand gives 120 6 5 m, in agreement with
fossil reef results
1
and supporting our mean value for H
crit
.
To evaluate the dependence of our lowstand reconstructions on
the main assumption that previous glacial DS values were similar to
LGM values of 10 6 2 p:s:u:, we perform two tests based on the
highest extreme glacial salinity estimate (55 p.s.u.; ref. 11). One
increases DS to 15 p.s.u.; the other keeps DS at 10 p.s.u. but increases
its confidence interval (dDS) from 62 to 65 p.s.u. In the first test,
H
crit
is smaller, and reconstructed sea level drops are greater, by a
maximum of 6 m. In the second test, with dDS increased by a factor
of 2.5, the lowstand confidence intervals remain accurate within a
factor of 2 (Fig. 3). These results justify the use of the SBM uplift rate
with H
crit
to study pre-stage-6 lowstands, noting that our method is
more likely to underestimate rather than overestimate past sea-level
drops, by a few metres.
Accounting for uplift since stage 8 (270 kyr
BP), sill depth was
around 150 m b.p.s.l. Stage 8 does not contain a completely
‘aplanktonic’ interval. Although the total planktonic foraminiferal
numbers are strongly reduced, the main species composition
shows little change that might reflect high-salinity stress (Fig. 1).
We infer that sill depth remained considerably greater than H
crit
.
To allow continuation of all observed planktonic species, Red Sea
salinity should have remained below a maximum of ,45 p.s.u.,
requiring a minimum sill depth of ,30 m (compare ref. 14).
Hence, the maximum conceived stage 8 sea level drop is
150 2 30 ¼ 120 m b.p.s.l.(68 m).
A similar argument to that for stage 8 may be made for stage 10
(340 kyr
BP). However, stage 10 shows a much closer approximation
of a complete ‘aplanktonic’ zone, with disruption of the main
species composition. We infer that sill depth was maintained
between H
crit
(18 m) and 30 m, defining a stage 10 lowstand between
134 and 122 m b.p.s.l. (69 m).
Stage 12 (440 kyr
BP) contains a true ‘aplanktonic’ zone (Fig. 1),
suggesting a sill depth around H
crit
and, consequently, a sea-level
lowstand of 139 m b.p.s.l. (611 m). This mean value implies that
global ice-volume during stage 12 exceeded LGM values by some
15%. This independently derived result validates the only previous
estimate of stage 12 ice-volume, based on benthic oxygen-isotope
records
5
.
Our lowstand values allow assessment of sea-level rises during the
main deglaciations of the past 500 kyr (Fig. 1i), for comparison with
that of 120 m following the LGM
1
. With the maximum stage 5 sea
level ,6 m above the present
2,6
, the stage 6–5 sea-level rise was
around 131 6 6 m. During interglacial stage 7, sea level remained
below the present-day level
4,5
, giving a maximum amplitude for the
stage 8–7 sea-level rise of 120 m, although the actual rise was
probably considerably smaller. The stage 9 highstand reached
0–15 m above the present-day level
4,5
, giving a stage 10–9 sea-level
rise between 122 and 149 m. The largest sea-level rise of the past
500 kyr followed the stage 12 lowstand of 139 6 11 m b:p:s:l: and
culminated in a maximum stage 11 highstand up to 20 m above
present-day sea level
21
.
We conclude that the last glacial–interglacial cycle showed ice-
volume fluctuations that were more than 10% smaller than those
that occurred in three out of four of the immediately preceding
main cycles. The stage 12–11 sea level rise implies that over 30%
greater ice-volume changes were involved in Quaternary glacial–
interglacial cycles than would be expected on the basis of the last
cycle alone.
M
Received 27 October 1997; accepted 22 April 1998.
1. Fairbanks, R. G. A 17,000 year glacio-eustatic sea level record: Influence of glacial melting rates on the
Younger-Dryas event and deep ocean circulation. Nature 342, 637–642 (1989).
2. Radtke, U. & Gru
¨
n, R. Revised reconstruction of middle and late Pleistocene sea-level changes based
on new chronologic and morphologic investigations in Barbados, West Indies. J. Coastal Res. 6, 699–
708 (1990).
3. Pirazzoli, P. A. et al. A one million-year-long sequence of marine terraces on Sumba Island, Indonesia.
Mar. Geol. 109, 221–236 (1993).
4. Bard, E. et al. Pleistocene sea levels and tectonic uplift based on dating of corals from Sumba Island,
Indonesia. Geophys. Res. Lett. 23, 1473–1476 (1996).
5. Shackleton, N. J. Oxygen isotopes, ice volume and sea level. Quat. Sci. Rev. 6, 183–190 (1987).
6. Chappell, J. & Shackleton, N. J. Oxygen isotopes and sea level. Nature 324, 137–140 (1986).
7. Berggren, W. A. & Boersma, A. in Hot Brines and Heavy Metal Deposits (eds Degens, E. T. & Ross, D.
A.) 282–298 (Springer, New York, 1969).
8. Halicz, E. & Reiss, Z. Palaeoecological relations of foraminifera in a desert enclosed sea—The Gulf of
Aqaba. Mar. Ecol. 2, 15–34 (1981).
9. Locke, S. & Thunell, R. C. Palaeoceanographic record of the last glacial-interglacial cycle in the Red Sea
and Gulf of Aden. Palaeogeogr. Palaeoclimatol. Palaeoecol. 64, 163–187 (1987).
10. Almogi-Labin, A., Hemleben, C., Meischner, D. & Erlenkeuser, H. Palaeoenvironmental events during
the last 13,000 years in the central Red Sea as recorded by pteropoda. Paleoceanography 6, 83–98
(1991).
11. Hemleben, C. et al. Three hundred and eighty thousand year long stable isotope and faunal records
from the Red Sea: Influence of global sea level change on hydrography. Paleoceanography 11, 147–156
(1996).
12. Milliman, J. D., Ross, D. A. & Ku, T. L. in Hot Brines and Heavy Metal Deposits (eds Degens, E. T. &
Ross, D. A.) 724–736 (Springer, New York, 1969).
13. Ku, T. L., Thurber, D. L. & Mathieu, G. G. in Hot Brines and Heavy Metal Deposits (eds Degens, E. T. &
Ross, D. A.) 348–359 (Springer, New York, 1969).
14. Rohling, E. J. Glacial conditions in the Red Sea. Paleoceanography 9, 653–660 (1994).
15. Rohling, E. J. & Zachariasse, W. J. RedSea outflow during the last glacial maximum. Quat. Int. 31, 77–
83 (1996).
16. Werner, F. & Lange, K. A bathymetric survey of the sill area between the Red Sea and Gulf of Aden.
Geol. Jahrb. D 13, 125–130 (1975).
17. Siedler, G. in Hot Brines and Heavy Metal Deposits (eds Degens, E. T. & Ross, D. A.) 131–137 (Springer,
New York, 1969).
18. Rohling, E. J., Jorissen, F. J., Vergnaud-Grazzini, C. & Zachariasse, W. J. Northern Levantine and
Adriatic Quaternary planktic foraminifera; Reconstruction of paleoenvironmental gradients. Mar.
Micropaleontol. 21, 191–218 (1993).
19. Ganssen, G. & Kroon, D. Evidence for Red Sea surface circulation from oxygen isotopes of modern
surface waters and planktonic foraminiferal tests. Paleoceanography 6, 73–82 (1991).
20. Squires, G. L. Practical Physics 3rd edn (Cambridge Univ. Press, 1988).
21. Howard, W. R. A warm future in the past. Nature 388, 418–419 (1997).
22. Imbrie, J. et al. in Milankovitch and Climate (eds Berger, A. et al.) 269–305 (Reidel, Hingham, MA,
1984).
Supplementary Information is available on Nature’s World-Wide Web site (http://www.nature.com) or
as paper copy from the London editorial office of Nature.
Acknowledgements. We thank H. Vonhof, M. Dignan and P. Martinez for assistance with stable-isotope
and TOC analyses; J. W. Zachariasse for cooperation within the context of our joint studies of the NW
Indian Ocean; NERC for support to M.F., and the National Museum of Natural History in Paris for
support to E.J.R. during the planning and sampling phase of this work.
Correspondence and requests for materials should be addressed to E.J.R. (e-mail: E.Rohling@soc.soton.
ac.uk).
Megaripple migration in
a natural surf zone
Edith L. Gallagher*, Steve Elgar
†
& Edward B. Thornton*
* Oceanography Department, Naval Postgraduate School, Monterey,
California 93943, USA
†
School of Electrical Engineering and Computer Science, Washington State
University, Pullman, Washington 99164, USA
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Migrating megaripples are bedforms that appear in the surf zone
of sandy coasts
1
. With heights of 0.1–0.5 m and wavelengths of 1–
5 m, they are similar in size and shape to small dunes, large
ripples, or sand waves. Such sedimentary bedforms have been
studied in subaerial
2
, steady-flow
3
and intertidal
4
environments,
as well as in laboratory flume experiments
5
. They affect overlying
currents by introducing hydraulic roughness
4,6
, and may provide a
mechanism for sediment transport
7,8
as well as forming sedimen-
tary structures in preserved facies
9,10
. The formation, orientation
and migration of such bedforms is not understood well
11,12
.
Dunes, for example, can be aligned with their crests perpendicular
to steady unidirectional winds
13
, but in more complex wind fields
their orientation becomes difficult to predict
14–17
. Similarly, it is
not known how sea-floor megaripples become aligned and
migrate in the complex flows of the surf zone. Here we present
observations in the surf zone of a natural beach which indicate
that megaripples do not migrate in the direction of the vector sum
of the currents, but are aligned so that the sediment transport
normal to the bedform crest is maximized
17
. This may need to be
taken into account in modelling morphology change and inter-
preting existing and fossil morphologic patterns.