Content uploaded by Elliot W Jackson
Author content
All content in this area was uploaded by Elliot W Jackson on Dec 15, 2020
Content may be subject to copyright.
electronic structure and tunneling probability of
the different CO molecules. We illustrate this by
calculatingthetunnelingcurrentprobabilityusing
the standard Bardeen approximation [equations
2and3in(23)] and the calculated partial density
of states (DOS) for CO molecules on the cluster
and on the tip (fig. S6) (23). The calculation re-
veals that the CO molecules on the corners indeed
have greater tunneling contributions than the
CO molecules on the edges, qualitatively explain-
ing the experimentally observed contrast of the
STMimageswiththesixbrightspotsplusone
in the center, as shown in Fig. 2, B and D. We
could explain the threefold symmetry of some of
the 19 atom clusters by adding three Cu atoms at
the center of the 19 atoms. These three low-
coordinated Cu atoms, producing the bright
center of the cluster images, can bind three ad-
ditional CO molecules and distort the tilt angles
of the peripheral CO molecules, as shown in Fig.
2, C and E [(details are shown in (23)].
Finally, we investigated the effect of clustering
on surface reactivity for the WGS reaction (i.e.,
CO+H
2
O↔CO
2
+H
2
), which Cu catalyzes. Water
does not adsorb on the Cu(111) surface at room
temperature (Fig. 4B) (32), whereas it dissocia-
tivelyadsorbsonthemoreactiveCu(110)surface
(32).OncethegasphaseCOat1Torrwaspumped
away, the STM images revealed that the Cu
clusters were still present, although atomic res-
olution could not be achieved, likely because
of the absence of CO molecules adsorbed on the
tip in high vacuum (Fig. 3A). In the presence of
2×10
−9
Torr of H
2
O, the cluster-covered sur-
face was very active in dissociating water, as
shown by the increasing oxygen peak in both
the Auger electron spectra (AES) shown in Fig.
3B, and in the XPS spectra shown in Fig. 3C.
The APXPS spectrum indicates that the O peak
is a result of the dissociative adsorption of H
2
O
(Fig. 4A) and that no such peak appears after
experiments at 0.1 Torr of CO because clustering
of the Cu did not occur at lower CO pressures
(Figs. 1B and 3B). A similar effect was also ob-
served during exposure to CO+H
2
Omixtures,as
shown in Fig. 4A. The pristine Cu(111) surface, on
the other hand, not pre-exposed to CO, is inactive
(Fig. 4B).
Our findings open the possibility that other
soft materials (e.g., Ag, Au, and Zn) can similarly
undergo large reconstructions at sufficiently high
pressures of CO (or other molecules). We have also
demonstrated that the inactive (111) face of Cu for
water dissociation, a key step in the water-gas shif t
reaction, becomes highly activated as a result of
the CO-induced clustering. The need for this type
of study to extend our understanding of the working
ofcatalystsunderoperatingconditionsisclear.
REFERENCES AND NOTES
1. G. A. Somorjai, Introduction to Surface Chemistry and Catalysis
(Wiley-VCH, 1999).
2. G. Ertl, Angew. Chem. Int. Ed. 47, 3524–3535 (2008).
3. B. J. McIntyre, M. Salmeron, G. A. Somorjai, Rev. Sci. Instrum.
64, 687–691 (1993).
4. L. Österlund et al., Phys. Rev. Lett. 86, 460–463 (2001).
5. E. Laegsgaard et al., Rev. Sci. Instrum. 72, 3537–3542
(2001).
6. F. Tao, D. Tang, M. Salmeron, G. A. Somorjai, Rev. Sci. Instrum.
79, 084101 (2008).
7. F. Besenbacher, P. Thostrup, M. Salmeron, MRS Bull. 37,
677–681 (2012).
8. C. T. Herbschleb et al., Rev. Sci. Instrum. 85, 083703
(2014).
9. M. Salmeron, R. Schlögl, Surf. Sci. Rep. 63, 169–199
(2008).
10. M. Salmeron, MRS Bull. 38, 650–657 (2013).
11. K. Klier, in Advances in Catalysis, Vol. 31,D.D.Eley,H.Pines,P.B.Weisz,
Eds. (Academic Press, 1982), pp. 243–313.
12. D. S. Newsome, Catal. Rev. 21, 275–318 (1980).
13. J. Szanyi, D. W. Goodman, Catal. Lett. 10, 383–390 (1991).
14. J. Yoshihara, C. T. Campbell, J. Catal. 161, 776–782
(1996).
15. M. Behrens et al., Science 336, 893–897 (2012).
16. G. A. Olah, Angew. Chem. Int. Ed. 52, 104–107 (2013).
17. M. Behrens, Angew. Chem. Int. Ed. 53, 12022–12024 (2014).
18. K. Kambe, Phys. Rev. 99, 419–422 (1955).
19. C. Kittel, Introduction to Solid State Physics, 8th Edition
(Wiley, 2005).
20. S. R. Longwitz et al., J. Phys. Chem. B 108, 14497–1450 2
(2004).
21. D. Tang, K. S. Hwang, M. Salmeron, G. A. Somorjai, J. Phys.
Chem. B 108, 13300–13306 (2004).
22. F. Tao et al., Science 327, 850–853 (2010).
23. See supplementary materials on Science Online
24. B. Eren et al., J. Phys. Chem. C 119, 14669–14674 (2015).
25. J. Lagoute, X. Liu, S. Fölsch, Phys. Rev. Lett. 95,136801
(2005).
26. O. V. Lysenko, V. S. Stepanyuk, W. Hergert, J. Kirschner, Phys.
Rev. Lett. 89, 126102 (2002).
27. M. Mehlhorn, H. Gawronski, K. Morgenstern, Phys. Rev. Lett.
104, 076101 (2010).
28. H. J. Yang, T. Minato, M. Kawai, Y. Kim, J. Phys. Chem. C 117,
16429–16437 (2013).
29. L. Bartels, D. Meyer, K. H. Rieder, Surf. Sci. Lett. 432,
L621–L626 (1999).
30. M. Poensgen, J. F. Wolf, J. Frohn, M. Giesen, H. Ibach, Surf. Sci.
274, 430–440 (1992).
31. M. M. Waldrop, Science 234, 673–674 (1986).
32. S. Yamamoto et al., J. Phys. Chem. C 111, 7848–7850
(2007).
ACKNO WLED GME NTS
This work was supported by the Office of Basic Energy Sciences
(BES), Division of Materials Sciences and Engineering, of the U.S.
Department of Energy (DOE) under contract no. DE-AC02-
05CH11231, through the Chemical and Mechanical Properties of
Surfaces, Interfaces and Nanostructures program (FWP KC3101).
It used resources of the National Energy Research Scientific
Computing Center and the Advanced Light Source, which are
supported by the Office of Science of the U.S. DOE. The
computation used resources from the Oak Ridge Leadership
Computing Facility (OLCF), with time allocated by the Innovative
and Novel Computational Impact on Theory and Experiment
(INCITE) project.
SUPPLEMENTARY MATERIALS
www.sciencemag.org/content/351/6272/475/suppl/DC1
Materials and Methods
Supplementary Text
Figs. S1 to S8
Tables S1 to S3
References (33–41)
16 November 2015; accepted 14 December 2015
10.1126/science.aad8868
OCEANOGRAPHY
Enhanced East Pacific Rise
hydrothermal activity during the last
two glacial terminations
D. C. Lund,
1
*P. D. Asimow,
2
K. A. Farley,
2
T. O. Rooney,
3
E. Seeley,
1
E. W. Jackson,
4
Z. M. Durham
4
Mid-ocean ridge magmatism is driven by seafloor spreading and decompression melting
of the upper mantle. Melt production is apparently modulated by glacial-interglacial
changes in sea level, raising the possibility that magmatic flux acts as a negative feedback
on ice-sheet size. The timing of melt variability is poorly constrained, however, precluding
a clear link between ridge magmatism and Pleistocene climate transitions. Here we present
well-dated sedimentary records from the East Pacific Rise that show evidence of enhanced
hydrothermal activity during the last two glacial terminations.We suggest that glacial maxima
and lowering of sea level caused anomalous melting in the upper mantle and that the
subsequent magmatic anomalies promoted deglaciation through the release of mantle heat
and carbon at mid-ocean ridges.
Sea level–driven pressure variations due to
the growth and decay of ice sheets likely
modulate melt production in the upper
mantle on Milankovitch time scales (1,2).
Model simulations suggest that the mag-
nitude of the resulting signal at mid-ocean ridges
depends on the plate spreading rate, the melt
extraction velocity, and the thermal properties
of the lithosphere (1,3). Because of the slow
rate of melt migration in the upper mantle, the
magmatic signal at ridges probably lags changes
in sea level by thousands of years (1). Surveys of
ridge bathymetry reveal Milankovitch-scale fre-
quencies in abyssal-hill spacing, consistent with
the sea-level hypothesis (3,4). Bathymetry records
are subject to geological damping effects and
478 29 JANUARY 2016 •VOL 351 ISSUE 6272 sciencemag.org SCIENCE
1
Deptartment of Marine Sciences, University of Connecticut,
Groton, CT 06340, USA.
2
Division of Geological and
Planetary Sciences, California Institute of Technology,
Pasadena, CA 91125, USA.
3
Department of Geological
Sciences, Michigan State University, East Lansing, MI 48824,
USA.
4
Department of Earth and Environmental Sciences,
University of Michigan, Ann Arbor, MI 48109, USA.
*Corresponding author. E-mail: david.lund@uconn.edu
RESEARCH |REPORTS
on February 2, 2016Downloaded from on February 2, 2016Downloaded from on February 2, 2016Downloaded from on February 2, 2016Downloaded from on February 2, 2016Downloaded from
substantial age uncertainties, however, and they
therefore require validation with other proxies.
Because hydrothermal activity along ridge sec-
tions is ultimately driven by magmatic heat, sed-
imentary records of hydrothermal output can
be used to assess the sea-level hypothesis and
determine the timing of magmatic anomalies
relative to key Pleistocene climate transitions.
The southern East Pacific Rise (SEPR) has the
fastest spreading rate and the highest magmatic
budget of any ridge in the global mid-ocean ridge
system (5). Due to its elevated magmatism, the
SEPR has over 50 known active vent sites from
5°S to 37°S (6), consistent with the global trend
in plume incidence versus magmatic budget for
ridges spanning a range of spreading rates (5,7).
Intense hydrothermal venting and topographical-
ly steered flow of plumes along the SEPR create
a spatially integrated pattern of metalliferous sed-
iments near the ridge crest (8–10). Compared
with slower ridges, SEPR sediments have anoma-
lously high metal concentrations (8,11), suggesting
that magmatism is the primary factor governing
hydrothermal input to these sedimentary archives
on geologic time scales. Hydrothermal plume par-
ticles are highly enriched in elements that are
derived directly from vents and scavenged from
seawater. Variations in the flux of these elements
to ridge-flank sediments should therefore reflect
long-term changes in hydrothermal activity.
We used a multiproxy geochemical strategy to
reconstruct SEPR hydrothermal activity during
the last glacial cycle. We analyzed a total of sev-
en ridge-crest cores from 6°S and 11°S, whe re
the h alf -spreading rate ave rages 75 mm/year
(Fig. 1). Together with two published records from
near the East Pacific Rise (EPR)–Dietz volcanic-
ridge triple junction at 1°N (12), the locations
span a range of spreading rates, sedimentary en-
vironments, and surface-ocean productivity re-
gimes. To control for spatial heterogeneity in
plume incidence versus magmatic budget (7), the
sampling locations span three separate EPR seg-
ments. At each segment, we analyzed cores from
both sides of the ridge axis to address potential
biases due to horizontal sediment focusing, bio-
turbation, and spatial variability in hydro thermal-
plume direction. Radiocarbon and oxygen isotopic
analyses of planktonic foraminifera provided age
control for each core (13). Major and trace ele-
ment concentrations were determined using x-ray
fluorescence (XRF) and inductively coupled plas-
ma mass spectrometry (ICP-MS). The fluxes of
hydrothermal components were estimated using
both mass accumulation rates and the
3
He nor-
malization method (13). Given that plume parti-
cles primarily consist of Fe oxyhydroxides and
Mn oxides (14), we used sedimentary Fe and Mn
to track hydrothermal inputs. To cross-check the
Fe and Mn results, we also measured arsenic, whi ch
is scavenged from seawater by Fe oxyhydroxides
and varies coherently with Fe in hydrothermal-
plume particles (15)andSEPRsediments(16).
Oxygen stable isotope records from 1°N, 6°S,
and 11°S outline marine isotope stages 1, 2, and 3,
indicating that there has been minimal strati-
graphic disturbance of the cores due to sediment
winnowing or downslope transport (Fig. 2). The
flux of Fe in all nine records peaks between
10 and 20 thousand years before the present
(ky B.P.). Manganese fluxes to EPR sediments
follow a similar pattern, with maximum values
centered at ~15 ky B.P. Arsenic fluxes at 6°S
and 11°S reach a maximum between 10 and 20 ky
B.P., supporting the Fe and Mn results. Offsets
between time series are generally 5 ky or less
(Fig. 2), similar to the age uncertainty associa-
ted with the mass accumulation rate method (13).
Results from the
3
He normalization technique,
which yields fluxes that are insensitive to age-
model uncertainty, show that positive shifts in
metal fluxes at 11°S and 6°S occurred within 2 ky
of one another (fig. S1). Thus, the overall pattern
for the past 50 ky is one of coherent variations
in hydrothermal sedimentation along 1300 km
of the EPR, with maximum metal inputs coincid-
ing with the last deglaciation (Termination I).
Two cores at 11°S span the penultimate de-
glaciation (Termination II), including core Y71-
07-53 on the western flank of the SEPR and core
Y71-07-47 on the eastern flank (Fig. 3). Metal
fluxes are higher in the western-flank core, con-
sistent with the east-west contrast in the shorter
records (Fig. 2) and the spatial pattern in metal
concentrations of late Holocene sediments (8).
In the western-flank core, the flux of each metal
increases markedly at ~140 ky B.P., reaches a
maximum by 130 ky B.P., and then returns to
backgroundlevelsby120kyB.P.(Fig.3).Asim-
ilarpatternoccursintheeastern-flankcore.The
contemporaneous signal at these loca tions indi-
cates that hydrothermal inputs on each side of
the ridge crest varied in phase. The records also
show that the maximum flux of hydrothermal
metals coincided with Termination II, similar to
the pattern for Termination I.
Diagenetic overprinting, horizontal sediment
focusing, and dilution with nonhydrothermal
components can complicate the interpretation
of hydrothermal proxies. Diagenetic remobili-
zation should influence Fe oxyhydroxides and
MnO
2
differently, given the large offset in their
redox potentials (17), yet we observed coherent
down-corevariationsinFe,Mn,andAs,regard-
less of location. Furthermore, down-core Fe/Mn
ratios generally fall within the expected range
for hydrothermal input (8). Anomalously high
SCIENCE sciencemag.org 29 JANUARY 2016 •VOL 351 ISSUE 6272 479
x
+
KLH068
KLH093
68 mm/yr
17 mm/yr
49 mm/yr
79 mm/yr
75 mm/yr
Pacific
Plate
Nazca
Plate
Y71-07-49
Y71-07-51
11.0°S
11.5°S
107°W108°W
6.0°S
6.5°S
Y71-09-115
Y71-09-106
Y71-09-104
Y71-07-53
Y71-07-47
Bathymetry
location
110°W111°W
1°N Sites
6°S Sites
11°S Sites
Fig. 1. Map of sampling sites near the EPR. Core locations include 1°N (cores KLH068 and KLH093)
(12), 6°S (core Y71-09-104 in blue, Y71-09-106 in red, and Y71-09-115 in black), and 11°S (cores Y71-07-47
and Y71-07-49 in magenta and Y71-07-51 and Y71-07-53 in black). Also shown is the location of the
bathymetry record at 17°S (4). Half-spreading rates are shown in white (www.ldeo.columbia.edu/~menke/
plates.html). The half-spreading rate at 1°N (17 mm/year measured at Dietz volcanic ridge) is from (29).
The map was generated using GeoMapApp (www.geomapapp.org).
RESEARCH |REPORTS
Fe/Mn ratios at 1°N are probably due to
suboxic diagenesis and Mn remobilization
(13). At the 1°N locations, near-zero Mn levels
before 20 ky BP are likely driven by MnO
2
reduction and upward migration of dissolved
Mn
2+
. Nevertheless, the overall coherent pat-
tern in Fe records from the high-productivity
equatorial Pacific (1°N) to the northern edge of
the subtropical gyre (11°S) indicates that the
organic carbon flux to the sediments is not a
first-order control on down-core metal varia-
bility. Sediment focusing is an equally unlikely
explanation, given the similar pattern in
multiple cores from a range of sedimentary
environments. Focusing factors estimated using
3
He also show no evidence for anomalous
horizontal sediment transport during Termi-
nation I (fig. S3). Lastly, the
3
He-based metal
fluxes are consistent with the mass accu-
mulation rate results, indicating that carbon-
ate dilution was not a primary driver of the
down-core signal. Taken together, these lines
of evidence indicate that the metal fluxes
primarily reflect the input of hydrothermal-
plume particles to ridge-crest sediments.
The temporal variability in metal fluxes is sim-
ilar to that in seafloor bathymetry at 17°S on
the SEPR, implying that both have a common
driver (Fig. 4). A lowering of sea level due to
ice-sheet expansion would promote decompres-
sion melting in the upper mantle. The resulting
increase in melt delivery to the ridge crest should
result in shoaling of the bathymetry and greater
hydrothermal activity (1,3). Ice-sheet retreat and
rising sea level would have the opposite effect.
Shallower bathymetry on the SEPR generally
corresponds to elevated hydrothermal fluxes,
consistent with the expected pattern (Fig. 4).
The bathymetry record lags the hydrothermal
proxies by ~10 ky, however (fig. S4). The offset is
most likely due to age uncertainty in the bathy-
metrytimeseries,wheretheagemodelisbased
on a half-spreading rate that optimizes the match
between the bathymetry and atmospheric CO
2
records (4). More generally, age constraints for
late Pleistocene oceanic crust are limited to two
control points, an assumed zero age at the ridge
crest and the Brunhes-Matuyama boundary at
780 ky B.P. Even if reliable absolute ages were
available for individual abyssal hills, it is unlikely
that their bathymetry would reflect only the melt
delivery that occurred when that oceanic crust
was at the ridge crest, because of the confound-
ing influences of lower crustal accretion, surface
480 29 JANUARY 2016 •VOL 351 ISSUE 6272 sciencemag.org SCIENCE
0
20
40
60
0
5
10
15
0 10 20 30 40 50
0
0.01
0.02
20
30
40
50
50
100
150
5
10
15
10
20
30
40
0 10 20 30 40 50
0
0.02
0.04
0
0.05
0.1
0.15
0 10 20 30 40 50
0
5
10
15
KLH068
KLH093
δ18O (‰)
Mn flux (µg cm-2 yr-1)
−1.5
−1
−0.5
0
0.5
1
123
Calendar Age (kyr BP) Calendar Age (kyr BP)
Calendar Age (kyr BP)
Y71-09-106
Y71-09-104
Y71-09-115
123
Y71-09-115 * 0.5
Y71-07-49
Y71-07-51
12 3
EPR 6°S EPR 11°S
EPR 1°N
As flux (mg cm-2 yr-1) Fe flux (µg cm-2 yr-1)Mn flux (µg cm-2 yr-1)δ18O (‰)
−2
−1.5
−1
−0.5
0
0.5
−2
−1.5
−1
−0.5
0
0.5
As flux (µg cm-2 yr-1) Fe flux (µg cm-2 yr-1)Mn flux (µg cm-2 yr-1)δ18O (‰)
10
20
30
40
50
Fe flux (µg cm-2 yr-1)
Fig. 2. Metal fluxes for the past 50 ky at three EPR
segments located at 1°N, 6°S, and 11°S. Flux esti-
mates are based on mass accumulation rates and metal
concentrations (13). Data from 1°N (12)areshownin
the first column, including (A) planktonic d
18
O, (B)Fe
flux, and (C) Mn flux. Data from 6°S are shown in the
second column, including (D) planktonic d
18
O, (E)Fe
flux, (F)Mnflux,and(G) As flux. Data from 11°S are
shown in the third column, including (H) planktonic
d
18
O, (I)Feflux,(J)Mnflux,and(K)Asflux.Maximum
fluxes occur during Termination I (gray vertical bar).
Typical errors for each record (crosses in each panel) represent the uncertainty of the flux estimates (vertical line) and age model error (horizontal line).
Calendar-corrected radiocarbon ages are shown as triangles. Approximate time intervals for marine isotope stages 1 to 3 are indicated in the top row of panels.
Arsenic results are not available for cores collected at 1°N because these cores were analyzed using XRF rather than ICP-MS (12).
RESEARCH |REPORTS
lava flows, and vertical and horizontal offsets of
crustal blocks by faulting (3,4,18,19). Although
bathymetric time series are useful for identifying
Milankovitch frequencies, the absolute timing
of events is poorly constrained by these records.
Hydrothermal proxies, on the other hand, can
be accurately dated using radiocarbon and oxy-
gen isotope stratigraphy. As a result, we are able
to infer that intervals of intense hydrothermal
activity on the EPR occurred during the last two
glacial terminations.
The coincidence in timing between hydrother-
mal maxima and glacial terminations implies
that there may be a direct causal relationship
between sea-level rise and hydrothermal activity.
Our understanding of the physical mechanisms
of decompression melting and melt migration
to the ridge axis suggests a more complex rela-
tionship, however. Proxies of magmatic flux should
lag sea-level changes by thousands of years, be-
cause of the slow rate of melt migration from the
magma source region to the ridge axis (1). During
the Last Glacial Maximum, the maximum rate of
sea-level decrease (and hence of pressure release
in the melting regime) occurred between 30 and
25 ky B.P. (20), or 15 ± 5 ky before the inferred
maximum in EPR hydrothermal activity (Fig. 2).
We observed a similar lag between the maximum
rate of sea-level rise at ~15 ky B.P. (20)andthe
late Holocene minimum in metal flux. Assuming
an average melt origin depth of 50 km (21), the
implied melt extraction velocities range from
2.5 to 5 m/year, which is consistent with the rate
of >1 m/year implied by U/Th disequilibrium
in zero-age mid-ocean ridge basalts (22) but
much lower than the estimates of >50 m/year
based on the time lag between deglaciation and
volcanism in Iceland (23).Ourestimateisinde-
pendent of previous methods and provides a
range of constraints for refining models of melt
extraction at fast spreading centers.
Our results support the hypothesis that en-
hanced ridge magmatism, hydrothermal output,
and perhaps mantle CO
2
flux act as a negative
feedback on ice-sheet size (1,4). Although the
modern carbon output from ridges is small (0.02
to 0.2 Pg C/year) (24), the flux probably increased
as a result ofsea-level modulation. Carbon sources
at off-axis locations, backarc basins, and island
arcsmayalsoamplifythemid-oceanridgesignal
(2). The long melt-migration times for carbon-
rich melts may lead to considerable differences
in timing between hydrothermal and carbon-flux
variations, however (25). Another mechanism
whereby magmatic variations may influence
climate is the hydrothermal heat flux itself. En-
hanced geothermal heat flux should warm and
destabilize the deep ocean (26), with excess heat
emerging along isopycnals into the surface South-
ern Ocean (26,27). Temperatures in the deep
SCIENCE sciencemag.org 29 JANUARY 2016 •VOL 351 ISSUE 6272 481
Fig. 3. Planktonic d
18
O
and metal fluxes
spanning Termination
II at 11°S. The time
series are from the
eastern (core Y71-07-47;
magenta) and western
(core Y71-07-53; black)
flanks of the EPR.
(A) Planktonic d
18
O
results superimposed
on a global benthic
d
18
Ostack[LR04(30)]
(gray line), (B)Feflux,
(C)Mnflux,and(D)As
flux. In (A), thin lines
indicate data from dis-
crete samples, and thick
lines indicate time series
smoothed with a three-
point running mean. In
(B) to (D), flux estimates
are based on mass
accumulation rates and
metal concentrations.
Glacial terminations are
indicatedbydashedver-
tical lines. Hydrothermal
metal fluxes peaked dur-
ing Termination II (T2).
Error bars for the west-
ern flank (gray shaded
area) and eastern flank
(thin magenta lines)
reflect the uncertainty of
the flux estimates in (B)
to (D) (13). Metal data
for core Y71-07-53 (16)
were assigned ages
based on the d
18
Ostra-
tigraphy generated for
this work. Arsenic data
are not available for core
Y71-07-47 because it
was analyzed using XRF
rather than ICP-MS.
3
3.5
4
4.5
5
5.5
−1.5
−1
−0.5
0
0.5
40
80
120
160
10
20
30
40
0 50 100 150 200
0
0.1
0.2
0.3
0.4
Calendar Age (kyr BP)
LR04 δ18O Stack (‰)
Planktonic δ18O (‰)
Fe flux (µg cm−2 yr−1)Mn flux (µg cm−2 yr−1)As flux (µg cm−2 yr−1)
Mn flux (µg cm−2 yr−1)Fe flux (µg cm−2 yr−1)
0
10
20
30
40
50
2
4
6
8
10
12
14
0
western
flank
eastern
flank
T2T1
0 50 100 150 200
−2
−1
0
1
2
3
4
Calendar Age (kyr BP)
Normalized Iron
−80
−60
−40
−20
0
20
40
60
0 50 100 150 200
−2
−1
0
1
2
3
4
Calendar Age (kyr BP)
Normalized Manganese
−80
−60
−40
−20
0
20
40
60
0 50 100 150 200
−2
−1
0
1
2
3
4
Calendar A
g
e
(
k
y
r BP
)
Normalized Arsenic
−80
−60
−40
−20
0
20
40
60
17°S Bathymetry Anomaly (m) 17°S Bathymetry Anomaly (m) 17°S Bathymetry Anomaly (m)
Fig. 4. Normalized metal fluxes at 11°S com-
paredwithEPRbathymetry.The hydrothermal
time series are from the eastern (magenta) and
western (black) flanks of the EPR and include (A)
Fe fl ux, ( B)Mnflux,and(C) As flux. We normalized
each record by subtracting the mean and dividing
by the standard deviation of each time series to
facilitate comparison between cores with differ-
ent mean metal concentrations. The results include
both discrete samples (thin lines) and time series
smoothed with a 20-ky-wide Gaussian window
(thick lines) to approximate the resolution of the
bathymetry compilation at 17°S (gray lines) (4).
Fluxes from 0 to 40 ky are based on the results
from Fig. 2; the interval from 40 to 200 ky B.P. is
based on results shown in Fig. 3.
RESEARCH |REPORTS
eastern tropical Pacific and Antarctica peaked
during each of the last two glacial terminations
(28), consistent with the timing of enhanced EPR
hydrothermal activity.
Isolating a mechanistic linkage between ridge
magmatism and glacial terminations will require
a suite of detailed proxy records from multiple
ridges that are sensitive to mantle carbon and
geothermal inputs, as well as modeling studies
of their influence in the ocean interior. The
EPR results establish the timing of hydrothermal
anomalies, an essential prerequisite for deter-
mining whether ridge magmatism can act as a
negative feedback on ice-sheet size. The data
presented here demonstrate that EPR hydro-
thermal output increased after the two largest
glacial maxima of the past 200,000 years, im-
plicating mid-ocean ridge magmatism in glacial
terminations.
REFERENCES AND NOTES
1. D. C. Lund, P. D. Asimow, Geochem. Geophys. Geosyst. 12,
Q12009 (2011).
2. P. Huybers, C. Langmuir, Earth Planet. Sci. Lett. 286, 479–491
(2009).
3. J. W. Crowley, R. F. Katz, P. Huybers, C. H. Langmuir,
S. H. Park, Science 347, 1237–1240 (2015).
4. M. Tolstoy, Geophys. Res. Lett. 42, 1346–1351 (2015).
5. E. T. Baker, G. R. German, in Mid-Ocean Ridges: Hydrothermal
Interactions Between the Lithosphere and Oceans,
C. R. German, J. Lin, L. M. Parson, Eds. (Geophysical
Monograph Series vol. 148, American Geophysical Union,
2004), pp. 245–266.
6. S. E. Beaulieu, E. T. Baker, C. R. German, A. Maffei, Geochem.
Geophys. Geosyst. 14, 4892–4905 (2013).
7. E. T. Baker, Geochem. Geophys. Geosyst. 10, Q06009 (2009).
8. J. Dymond, Geol. Soc. Am. 154, 133–174 (1981).
9. G. B. Shimmield, N. B. Price, Geochim. Cosmochim. Acta 52,
669–677 (1988).
10. K. G. Speer, M. E. Maltrud, A. M. Thurnherr, in Energy and Mass
Transfer in Hydrothermal Systems, P. E. Halbach, V. Tunnicliffe,
J. R. Hein, Eds. (Dahlem University Press, 2003), pp. 287–302.
11. K. Boström, M. N. Peterson, O. Joensuu, D. E. Fisher, J. Geophys.
Res. 74,3261–3270 (1969).
12. M. F rank et al., Paleoceanography 9,559–578
(1994).
13. Material and methods are available as supplemental materials
on Science Online.
14. C. R. German, S. Colley, M. R. Palmer, A. Khripounoff,
G. P. Klinkhammer, Deep Sea Res. Part I Oceanogr. Res. Pap.
49, 1921–1940 (2002).
15. R. R. Cave, C. R. German, J. Thomson, R. W. Nesbitt, Geochim.
Cosmochim. Acta 66, 1905–1923 (2002).
16. T. Schaller, J. Morford, S. R. Emerson, R. A. Feely, Geochim.
Cosmochim. Acta 64, 2243–2254 (2000).
17. S. Emerson, J. I. Hedges, in Treatise on Geochemistry,
K. K. Turekian, H. D. Holland, Eds. (Elsevier, vol. 6, 2004),
pp. 293–319.
18. J. A. Goff, Science 349, 1065 (2015).
19. J. A. Olive et al., Science 350, 310–313 (2015).
20. P. U. Clark et al., Science 325, 710–714 (2009).
21. K. Key, S. Constable, L. Liu, A. Pommier, Nature 495, 499–502
(2013).
22. P. B. Kelemen, G. Hirth, N. Shimizu, M. Spiegelman, H. J. B. Dick,
Philos. Trans. R. Soc. Lond. A 355,283–318 (1997).
23. J. Maclennan, M. Jull, D. McKenzie, L. Slater, K. Gronvold,
Geochem. Geophys. Geosyst. 3,1–25 (2002).
24. P. Cartigny, F. Pineau, C. Aubaud, M. Javoy, Earth Planet.
Sci. Lett. 265, 672–685 (2008).
25. J. M. A. Burley, R. F. Katz, Earth Planet. Sci. Lett. 426, 246–258
(2015).
26. M. Hofmann, M. A. Morales Maqueda, Geophys. Res. Lett. 36,
L03603 (2009).
27. J. Emile-Geay, G. Madec, Ocean Sci. 5, 203–217 (2009).
28. P. Martin, D. Archer, D. W. Lea, Paleoceanography 20, PA2015
(2005).
29. D. K. Smith, H. Schouten, L. Montési, W. Zhu, Earth Planet.
Sci. Lett. 371–372,6–15 (2013).
30. L. E. Lisiecki, M. E. Raymo, Paleoceanography 20, PA1003
(2005).
ACKNOW LEDGM ENTS
We dedicate this paper to J. Dymond, whose 1981 treatise on Nazca
plate sediments made this work possible. We are also indebted
to the Oregon State University Core Repository for carefully
preserving the EPR sediment cores since they were collected in
the early 1970s. We are grateful to L. Wingate at the University of
Michigan and M. Cote at the University of Connecticut for
technical support. This work has benefited from discussions
with J. Granger, P. Vlahos, B. Fitzgerald, and M. Lyle. Data
presented here are available on the National Oceanic and
Atmospheric Administration’s Paleoclimatology Data website
(www.ncdc.noaa.gov/data-access/paleoclimatology-data). Funding
was provided by the University of Michigan and the University
of Connecticut.
SUPPLEMENTARY MATERIALS
www.sciencemag.org/content/351/6272/478/suppl/DC1
Materials and Methods
Supplementary Text
Figs. S1 to S11
Tables S1 to S5
References (31–45)
14 September 2015; accepted 6 January 2016
10.1126/science.aad4296
HISTORY OF SCIENCE
Ancient Babylonian astronomers
calculated Jupiter’s position from the
area under a time-velocity graph
Mathieu Ossendrijver*
Theideaofcomputingabody’s displacement as an area in time-velocity space is usually traced
back to 14th-century Europe. I show that in four ancient Babylonian cuneiform tablets, Jupiter’s
displacement along the ecliptic is computed as the area of a trapezoidal figure obtained by
drawing its daily displacement against time.This interpretation is prompted by a newly
discovered tablet on which the same computation is presented in an equivalent arithmetical
formulation. The tablets date from 350 to 50 BCE. The trapezoid procedures offer the first
evidence for the use of geometrical methods in Babylonian mathematical astronomy, which was
thus far viewed as operating exclusively with arithmetical concepts.
The so-called trapezoid procedures examined
in this paper have long puzzled historians
of Babylonian astronomy. They belong to
the corpus of Babylonian mathematical as-
tronomy, which comprises about 450 tab-
lets from Babylon and Uruk dating between 400
and 50 BCE. Approximately 340 of these tablets
are tables with computed planetary or lunar data
arranged in rows and columns (1). The remaining
110 tablets are procedure texts with computa-
tional instructions (2), mostly aimed at comput-
ing or verifying the tables. In all of these texts the
zodiac, invented in Babylonia near the end of the
fifth century BCE (3), is used as a coordinate sys-
tem for computing celestial positions. The un-
derlying algorithms are structured as branching
chains of arithmetical operations (additions, sub-
tractions, and multiplications) that can be rep-
resented as flow charts (2). Geometrical concepts
are conspicuously absent from these texts, whereas
they are very common in the Babylonian mathe-
matical corpus (4–7). Currently four tablets, most
likely written in Babylon between 350 and 50 BCE,
are known to preserve portions of a trapezoid
procedure (8). Of the four procedures, here labeled
B to E (figs. S1 to S4), one (B) preserves a men-
tion of Jupiter and three (B, C, E) are embedded
in compendia of procedures dealing exclusively
with Jupiter. The previously unpublished text D
probably belongs to a similar compendium for
Jupiter. In spite of these indications of a connec-
tion with Jupiter, their astronomical significance
was previously not acknowledged or understood
(1,2,6).
A recently discovered tablet containing an un-
published procedure text, here labeled text A (Fig. 1),
sheds new light on the trapezoid procedures. Text A
most likely originates from the same period and
location (Babylon) as texts B to E (8). It contains
a nearly complete set of instructions for Jupiter’s
motion along the ecliptic in accordance with the
so-called scheme X.S
1
(2). Before the discovery of
text A, this scheme was too fragmentarily known
for identifying its connection with the trapezoid
procedures. Covering one complete synodic cycle,
scheme X.S
1
begins with Jupiter’s heliacal rising
(first visible rising at dawn), continuing with its
first station (beginning of apparent retrograde
motion), acronychal rising (last visible rising at
dusk), second station (end of retrograde motion),
and heliacal setting (last visible setting at dusk)
(2). Scheme X.S
1
and the four trapezoid procedures
are here shown to contain or imply mathematically
equivalent descriptions of Jupiter’smotionduring
the first 60 days after its first appearance. Whereas
scheme X. S
1
employs a purely arithmetical ter-
minology, the trapezoid procedures operate with
geometrical entities.
482 29 JANUARY 2016 •VOL 351 ISSUE 6272 sciencemag.org SCIENCE
Excellence Cluster TOPOI–Institute of Philosophy, Humboldt
University, Berlin, Germany.
*Corresponding author. E-mail: mathieu.ossendrijver@hu-berlin.de
RESEARCH |REPORTS
DOI: 10.1126/science.aad4296
, 478 (2016);351 Science et al.D. C. Lund
glacial terminations
Enhanced East Pacific Rise hydrothermal activity during the last two
This copy is for your personal, non-commercial use only.
clicking here.colleagues, clients, or customers by , you can order high-quality copies for yourIf you wish to distribute this article to others
here.following the guidelines can be obtained byPermission to republish or repurpose articles or portions of articles
): February 1, 2016 www.sciencemag.org (this information is current as of
The following resources related to this article are available online at
/content/351/6272/478.full.html
version of this article at: including high-resolution figures, can be found in the onlineUpdated information and services,
/content/suppl/2016/01/27/351.6272.478.DC1.html
can be found at: Supporting Online Material
/content/351/6272/478.full.html#ref-list-1
, 6 of which can be accessed free:cites 41 articlesThis article
/cgi/collection/oceans
Oceanography /cgi/collection/geochem_phys
Geochemistry, Geophysics subject collections:This article appears in the following
registered trademark of AAAS. is aScience2016 by the American Association for the Advancement of Science; all rights reserved. The title CopyrightAmerican Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005.
(print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by theScience
on February 2, 2016Downloaded from
