Distinct iron isotopic signatures and supply from marine sediment dissolution

Abstract

Oceanic iron inputs must be traced and quantified to learn how they affect primary productivity and climate. Chemical reduction of iron in continental margin sediments provides a substantial dissolved flux to the oceans, which is isotopically lighter than the crust, and so may be distinguished in seawater from other sources, such as wind-blown dust. However, heavy iron isotopes measured in seawater have recently led to the proposition of another source of dissolved iron from 'non-reductive' dissolution of continental margins. Here we present the first pore water iron isotope data from a passive-tectonic and semi-arid ocean margin (South Africa), which reveals a smaller and isotopically heavier flux of dissolved iron to seawater than active-tectonic and dysoxic continental margins. These data provide in situ evidence of non-reductive iron dissolution from a continental margin, and further show that geological and hydro-climatic factors may affect the amount and isotopic composition of iron entering the ocean.

Full text

ARTICLE
Received 28 Jan 2013 | Accepted 11 Jun 2013 | Published 19 Jul 2013
Distinct iron isotopic signatures and supply
from marine sediment dissolution
William B. Homoky
1
, Seth G. John
2
, Tim M. Conway
2
& Rachel A. Mills
1
Oceanic iron inputs must be traced and quantified to learn how they affect primary pro-
ductivity and climate. Chemical reduction of iron in continental margin sediments provides a
substantial dissolved flux to the oceans, which is isotopically lighter than the crust, and so
may be distinguished in seawater from other sources, such as wind-blown dust. However,
heavy iron isotopes measured in seawater have recently led to the proposition of another
source of dissolved iron from ‘non-reductive’ dissolution of continental margins. Here we
present the first pore water iron isotope data from a passive-tectonic and semi-arid ocean
margin (South Africa), which reveals a smaller and isotopically heavier flux of dissolved iron
to seawater than active-tectonic and dysoxic continental margins. These data provide in situ
evidence of non-reductive iron dissolution from a continental margin, and further show that
geological and hydro-climatic factors may affect the amount and isotopic composition of iron
entering the ocean.
DOI: 10.1038/ncomms3143
OPEN
1
Ocean and Earth Science, University of Southampton, National Oceanography Centre, European Way, Southampton SO14 3ZH, UK.
2
Department of Earth
and Ocean Sciences, University of South Carolina, Columbia, South Carolina 29208, USA. Correspondence and requests for materials should be addressed to
W.B.H. (email: wbh@noc.soton.ac.uk).
NATURE COMMUNICATIONS | 4:2143 | DOI: 10.1038/ncomms3143 | www.nature.com/naturecommunications 1
& 2013 Macmillan Publishers Limited. All rights reserved.

I
ron (Fe) inputs to the surface ocean may stimulate photo-
synthesis and have an impact on the uptake of carbon dioxide
in the ocean on glacial to inter-glacial timescales of climate
change
1
. Global ocean reservoir-flux models
2
indicate that 90% of
Fe used by marine phytoplankton in the present day surface
ocean is supplied from the deep water below, but the sources of
dissolved Fe to this deep water are still poorly constrained.
Therefore, quantifying and tracking iron supplied to the ocean
will provide key information to resolve climate models and
sensitivity to the Fe cycle
3,4
.
Measurable differences in the isotopic composition of Fe
between various sources to the ocean have prompted widespread
interest in seawater Fe isotope determintions
5–7
, which can
potentially be used to track Fe inputs and assess the relative
importance of different sources of dissolved Fe to the oceanic
reservoir. Microbial sediment respiration supports a major flux of
dissolved and isotopically light Fe to the global ocean
8–10
,by
catalysing the reductive dissolution (RD) of Fe oxyhydroxide
minerals during organic matter decomposition
11
. Reduction of Fe
oxyhydroxide enriches soluble Fe(II)
(aq)
in sediment pore water,
which diffuses into bottom water when the oxygenated layer of
surface sediment is adequately shallow
9,12
, most notably from
oxygen-deficient continental margins
8–10
. Benthic fluxes of Fe are
mixed in bottom waters and can be transported to open ocean
and surface waters
13,14
, where Fe may control the efficacy of the
biological carbon pump
15,16
.
Dissolved Fe(II)
(aq)
produced by RD initially has d
56
Fe values
0.5–2.0% lighter than the original substrates
17
, and at isotopic
equilibrium, experiments show d
56
Fe(II)
(aq)
is  1.05 to  3.99%
relative to the common reactive Fe oxides haematite
17
, goethite
18
and ferrihydrite
17,19,20
. Similar light d
56
Fe values (  1.82 to
 3.45%) have been observed in both the pore waters
21–23
and
overlying seawater
9,24
of river-dominated and dysoxic margins,
and light Fe isotopic compositions are recorded in ocean
basin sediments coeval with past episodes of ocean oxygen
deficiency, consistent with seawater transport of light Fe from
ferruginous shelf sediments to ocean basins
25
. Thus, benthic
fluxes of isotopically light Fe appear to be distinguishable from
other sources of Fe to the ocean, such as atmospheric dust
dissolution (d
56
Fe ¼þ0.13
±
0.18%)
26
and river discharge
(d
56
Fe ¼þ0.14
±
0.28%)
27
.
Paradoxically, however, equatorial Pacific seawater originating
from the continental margin of New Guinea contains elevated Fe
concentrations with heavy Fe isotopic compositions (d
56
Fe ¼
þ 0.37
±
0.15%)
28
. These and other seawater isotope
measurements have led to the proposition of an additional
‘non-reductive dissolution’ (NRD) mechanism for Fe
28,29
, albeit
with existing Fe isotope evidence from continental margin
sediments indicating otherwise
9,24
. These findings coincide with
a growing need to evaluate the geographical variability of benthic
Fe fluxes to effectively model carbon cycling in the ocean
3,4
,
where models presently rely on global extrapolations from
potentially unrepresentative regions.
Here we characterise the pore water isotopic composition and
corresponding flux of dissolved Fe from the Cape margin, South
Africa—a semi-arid passive margin derived from deeply weath-
ered saprolite soils and surrounded by oxygenated South Atlantic
seawater. These sites are distinct from most previous sites of
benthic Fe flux investigation, which have focused on active
margins next to areas of rapid uplift with oxygen-deficient shelf
waters (Fig. 1). This study reveals that the amount of dissolved Fe
released from the Cape margin is less than predicted by benthic
Fe flux relationships
8
widely used to model ocean Fe–CO
2
interaction
3,4
. We report solid-phase compositional data that
suggests that the small pore water Fe flux reflects geological and
hydro-climatic influences on reactive Fe substrate delivery to the
shelf. Isotopically heavy Fe present in ‘oxidizing’ pore waters of
the Cape margin—a zone previously beyond analytical
resolution—provides in situ evidence for the role of ‘NRD’ of
Fe proposed by Radic et al.
28
These discoveries have implications
for past and present oceanic Fe cycles and the parameterization of
ocean biogeochemical models.
Results
Geological setting. Cape margin sediments originate from
Palaeozoic sedimentary sequences of the Cape Supergroup
(quartz arenites, subgreywackes, shales and sandstones) that were
formed by weathering of ancient Kalahari, Patagonia and Rı
´
ode
la Plata Cratons and deposited along rifted margins of Southern
Gondowanda B480–300 Ma (ref. 30). The modern Berg and
Olifants Rivers drain the Cape Supergroup and outflow 100–
200 km north of our study area, slowly delivering mature
siliciclastic material to the Cape shelf (o3 cm kyr
 1
) (ref. 31),
where shelf deposits have been further re-worked by Neogene
sea-level fluctuations
32
. Consequently, winnowed glauconitic
sands and quartzose muddy sands occupy the outer shelf and
upper-shelf slope
31
, from where we collected surface sediments
from three sites at 733, 1,182 and 2,602 m water depth (Fig. 1).
South Atlantic water masses intersecting the sample sites are
replete with oxygen, with a small oxygen minimum
(174 mmol l
 1
) associated with Upper Circumpolar Deep Water
intersecting the 1,182 m study site.
Cape margin sediment composition. The presence of quartz,
plagioclase and K-feldspars, calcite and clay fractions of illite and
glauconite were confirmed by XRD and scanning electron
microscopy (SEM) analyses. Iron was often associated with Ti
and O as illmenite minerals, whereas Fe-S phases were not
identified by XRD or energy-dispersive X-ray spectroscopy (EDS)
point analysis. The Cape margin sediments contained sub-
stantially fewer reducible Fe oxide minerals (Fe
HR
) than have
been observed in river-dominated margin sediments from around
the world (Fig. 2; Table 1). Furthermore, total Fe in Cape margin
sediments largely reflects the average crustal abundance in con-
trast to areas influenced by river discharge or dysoxic ocean
waters that are typically enriched above crustal Fe concentrations
and therefore have a more replete Fe reservoir for dissolution
processes. Thus, Cape margin sediments provide a previously
missing case study to link earth surface processes with the nature
of seawater Fe supply from the continents.
Cape margin pore water composition. The upper-shelf slope of
the Cape margin (733, 1,182 and 2,602 m water depth) contains
sufficient organic carbon (means 2.2, 1.7 and 1.4 wt%; 0–20 cm)
to consume dissolved O
2
within ju st a fe w millimetres of the
sediment–water interface (Fig. 3; Table 1). Subsequent down-
core chemical zonations of dissolved NO
3

,MnandFeinthe
pore water are consistent with control by organic matter
decomposition ( Fig. 3). Sub-surface dissolved Fe maxima
(between 3 and 6 mmol l
 1
) are clearly identified in what have
been termed ‘ferruginous’ zones bet ween relatively ‘oxidiz ing’
(O
2
and NO
3

in the pore water) and ‘sulphidic’ (Fe depletion
consistent with SO
4

reduction
33
) zon es. However, dissolved
Fe values remain 10–100 times lower than pore waters from
California
8,9,12, 21, 23
,Oregon
9,12,21
and Peru
10
margins, where
previous benthic Fe flux investigations have been fo cused.
The speci ation of dissolved Fe in the sampled pore water
is not kn own, b ut t he near-surface abundance of electron
acceptors m eans dissolved Fe in t he oxidizing z one may be
present as either Fe(II)
(aq)
colloidal oxyhy droxides
34
or organic
complexes.
ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms3143
2 NATURE COMMUNICATIONS | 4:2143 | DOI: 10.1038/ncomms3143 | www.nature.com/naturecommunications
& 2013 Macmillan Publishers Limited. All rights reserved.

Cape margin iron isotope signatures. We observe a wide range
in the Fe isotopic composition of Cape margin pore waters
(d
56
Fe ¼3.09 to þ 1.22% relative to IRMM-14), but with
distinct and reproducible down-core behaviour between
sulphidic, ferruginous and oxidizing zones in the sediments
(Fig. 3; Table 1). We find d
56
Fe of the bulk solid phase to be
þ 0.08
±
0.2% (n ¼ 11), equal to the average weathering product
of the continental crust ( þ 0.09
±
0.7%) described by previous
South Africa
Section
This study
1
2
3
18°E 20°E16°E 22°E14°E
a
b
32°S
34°S
36°S
38°S
3
2
1
0
125 100 75 50
Distance from shore (km)
Cape margin
Cape margin
733 m
1,182 m
2,602 m
Water depth (km)
Figure 1 | Location of benthic Fe flux determinations from ocean margins. The inset map (a) and corresponding cross-section (b) show the location of
Cape margin sample sites (733 m, 1,182 m and 2,602 m) and their corresponding data markers (diamond, square and circle) used in accompanying figures.
Models of ocean productivity and C-export
3,4
assume that bio-essential Fe supplied from ocean margins corresponds to previous flux constraints from
dysoxic borderland basins of Southern California (1)
8,9,44
, and supported by studies of seasonally dysoxic river-fed margins of California and Oregon (2)
9,12
and the Peruvian margin oxygen-minimum zone (3)
10
. The Cape margin sites reveal a pronounced variability to the amount and isotopic composition of Fe
that may be supplied from ocean margins between regions, which is not yet well accounted for by global ocean biogeochemical models.
San
Pedro
Basin
60
Peru margin
59
Santa
Monica
Basin
60
Average global river margins
47
Average upper cont. crust
59
20
10
0
Depth (cm)
1.51.00.50.0
0.90.80.70.60.50.40.3
Cape margin, South Africa (this study): 733 m 1,182 m 2,602 m
Fe
T
/AlFe
HR
(wt %)
Figure 2 | Fe abundance in Cape margin sediments. The amount of highly reactive Fe (Fe
HR
) and the proportion of total Fe to Al is shown for the
Cape margin—a semi-arid passive-tectonic margin—compared with global averages and sites previously used to characterize the benthic Fe flux to the
oceans. Grey bars represent the two times the s.d. of mean values from literature; coloured bars represent the range of abundances determined across
corresponding sediment depths. Fe
HR
is liberated by a Na-dithionite extraction
55
and used here to show that the reducible Fe oxide reservoir in Cape margin
sediments is depleted relative to the global average of river-dominated margins. Upper-shelf slope sites also contain crustal Fe/Al ratios, indicating little
enrichment of authigenic Fe compared with sites used for previous benthic Fe flux determinations.
NATURE COMMUNICATIONS | DOI: 10.1038/ncomms3143 ARTICLE
NATURE COMMUNICATIONS | 4:2143 | DOI: 10.1038/ncomms3143 | www.nature.com/naturecommunications 3
& 2013 Macmillan Publishers Limited. All rights reserved.

workers
36
. Relatively heavy ‘sulphidic’ zone pore water Fe
isotopes at depth in these sediments are consistent with control
by low-solubility sulphide minerals observed elsewhere
21
, and
supported by experimental studies of pyrite formation, in which
pore water Fe(II)
(aq)
may reach þ 1.3% when not in equilibrium
with the solid phase
37
. We observe the lightest Fe isotopic
compositions (  0.34 to  3.09%) in the ferruginous zone at
each study site, which is characteristic of microbially catalysed RD
seen in marine sediments elsewhere
9,21,23
and supported by
incubation experiements
17,20
. However, the isotopic composition
of Fe in the most oxidizing layer of surface sediments shows a
systematic and previously unidentified transition towards heavier,
near-crustal values at the sediment–water interface.
The pore water Fe isotopic concentration variations that we
observe could not have been generated by oxidation or sorption
of the reduced Fe pool, and nor could they be oxidation-related
sampling artifacts. Precipitation of Fe(II)
(aq)
by oxidation and/or
sorption has a kinetic isotope effect that lowers the residual
d
56
Fe(II)
(aq)
(ref. 17). We hypothesize that the near-surface
transition to heavier Fe isotopic compositions reflects mixing
with a heavier end-member Fe input, which is only discernable
because of the very low abundance of dissolved Fe supplied by
Table 1 | Summary of results from Cape margin pore water and sediment samples.
Pore water Sediment
Latitude Longitude Water
depth
Pore water
depth
O
2
*
NO
3

þ NO
2

Mn Fe d
56
Fe Sediment
depth
d
56
Fe Fe
HR
Total Fe Total Al Organic C
Deg. S Deg. E m cm l mol l
 1
l mol l
 1
l mol l
 1
l mol l
 1
% s.e. cm % s.e. wt% wt% wt% wt%
34°20’ 17°37’ 733 olw 201 nd 0.09 0.09 0.36 0.19 0–1 0.08 0.01 0.27 1.39 3.24 1.25
34°20’ 17°37’ 733 0.5 60.3 21.6 0.07 0.10  1.08 0.11 1–2 0.12 0.01 0.07 1.39 3.37 1.31
34°20’ 17°37’ 733 1.5 o0.3 10.3 0.50 1.33  1.91 0.02 2–3 nd nd 0.12 nd nd 1.47
34°20’ 17°37’ 733 2.5 o0.3 1.3 0.72 2.20  3.09 0.02 3–4 nd nd 0.13 nd nd 1.37
34°20’ 17°37’ 733 3.5 o0.3 0.4 nd nd nd nd 4–5 nd nd 0.12 nd nd 1.52
34°20’ 17°37’ 733 4.5 o0.3 0.7 0.49 2.57  1.31 0.02 6–8 nd nd 0.11 1.42 3.35 1.38
34°20’ 17°37’ 733 5.5 o0.3 1.0 0.41 5.19  1.10 0.02 8–10 0.12 0.02 0.15 1.41 3.48 1.36
34°20’ 17°37’ 733 6.5 o0.3 1.2 0.35 1.64  1.01 0.01 12–14 0.16 0.01 0.09 1.28 3.28 1.00
34°20’ 17°37’ 733 7.5 o0.3 0.9 0.33 2.61  0.6 0.02 16–18 nd nd 0.08 nd nd 0.91
34°20’ 17°37’ 733 8.5 o0.3 0.9 nd nd nd nd 20–22 0.16 0.02 0.14 1.35 3.42 1.10
34°20’ 17°37’ 733 9.5 o0.3 nd 0.29 0.87  0.34 0.02 24–26 nd nd nd nd nd 1.17
34°20’ 17°37’ 733 11.5 o0.3 0.8 0.29 0.52 
0.48 0.02
34°20’ 17°37’ 733 13.5 o0.3 nd 0.24 0.22  0.17 0.02
34°20’ 17°37’ 733 15.5 o0.3 0.7 nd nd nd nd
34°20’ 17°37’ 733 19.5 o0.3 nd 0.16 0.16 0.7 0.02
34°20’ 17°37’ 733 23.5 o0.3 0.0 nd nd nd nd
34°20’ 17°37’ 733 27.5 o0.3 0.0 nd nd nd nd
34°20’ 17°37’ 733 28.5 o0.3 nd 0.14 0.19 1.22 0.03
34°22’ 17°33’ 1,182 olw 174 nd 0.06 0.17  0.29 0.05 0–1 nd nd 0.85 1.73 3.34 2.29
34°22’ 17°33’ 1,182 0 .5 9.2 24.9 0.38 0.24  0.07 0.04 1–2 nd nd 0.33 1.81 3.41 2.18
34°22’ 17°33’ 1,182 1.5 o0.3 12.2 0.06 0.21  0.48 0.09 2–3 nd nd 0.30 1.75 3.44 2.21
34°22’ 17°33’ 1,182 2.5 o0.3 1.7 0.28 0.58  1.66 0.02 3–4 nd nd 0.36 nd nd 2.38
34°22’ 17°33’ 1,182 3.5 o0.3 0.4 0.67 2.22  1.61 0.01 4–5 nd nd 0.25 nd nd 2.10
34°22’ 17°33’ 1,182 4.5 o0.3 0.3 0.44 3.48  1.34 0.01 6–8 nd nd 0.19 1.68 3.38 2.20
34°22’ 17°33’ 1,182 5.5 o
0.3 0.4 0.63 3.60  1.14 0.02 8–10 nd nd 0.21 1.55 3.18 2.13
34°22’ 17°33’ 1,182 7.5 o0.3 0.4 0.47 5.33  0.81 0.02 12–14 nd nd nd 1.74 3.36 2.38
34°22’ 17°33’ 1,182 9.5 o0.3 nd 0.44 2.09  1.41 0.02 16–18 0.00 0.02 0.18 1.74 3.36 2.27
34°22’ 17°33’ 1,182 11.5 o0.3 0.2 0.33 2.78  0.84 0.03 20–22 nd nd 0.19 1.78 3.35 2.04
34°22’ 17°33’ 1,182 13.5 o0.3 0.1 0.33 1.60  1.23 0.02 24–26 nd nd nd nd nd 2.11
34°22’ 17°33’ 1,182 16.5 o0.3 0.0 0.37 1.05  0.73 0.02
34°22’ 17°33’ 1,182 19.5 o0.3 0.0 0.35 0.70  0.03 0.02
34°22’ 17°33’ 1,182 22.5 o0.3 0.2 0.35 1.08 0.71 0.04
34°22’ 17°33’ 1,182 25.5 o0.3 0.1 0.33 0.51 0.51 0.02
34°37’ 17°03’ 2,602 olw 225 nd nd nd nd nd 0–1 nd nd 0.05 1.88 2.98 1.07
34°37’ 17°03’ 2,602 0.5 160 32.7 0.05 0.65 0.04 0.09 1–2  0.03 0.01 0.09 1.83 2.89 1.05
34°37’ 17°03’ 2,602 1.5 89.1 33.5 0.02 0.16 0.8 0.29 2–3 nd nd 0.10 1.92 3.06 0.97
34°37’ 17°03’ 2,602 2.5 20.6 24.6 0.04 0.15 nd nd 3–4 nd nd 0.08 1.93 2.96 1.05
34°37’ 17°03’ 2,602 3.5 o0.3 12.1 0.03 0.05 nd nd 4–6 nd nd 0.08 1.83 2.87 0.93
34
°37’ 17°03’ 2,602 5 o0.3 3.3 0.04 0.27  0.12 0.04 6–8 0.04 0.02 0.06 1.72 2.81 1.08
34°37’ 17°03’ 2,602 7 o0.3 1.1 1.34 0.35  1.11 0.03 10–12 0.02 0.02 0.05 1.57 2.59 0.97
34°37’ 17°03’ 2,602 9 o0.3 0.6 1.90 0.92  2.04 0.03 14–16 0.15 0.02 0.07 1.57 2.74 0.99
34°37’ 17°03’ 2,602 11 o0.3 4.2 2.23 2.36  1.72 0.02 18–20 nd nd 0.11 1.56 2.69 nd
34°37’ 17°03’ 2,602 13 o0.3 0.7 2.15 2.33  2.06 0.02 22–24 0.08 0.01 0.08 1.48 2.57 0.98
34°37’ 17°03’ 2,602 15 o0.3 1.2 2.29 2.36  2.14 0.02
34°37’ 17°03’ 2,602 17 o0.3 0.6 2.32 2.95  2.20 0.02
34°37’ 17°03’ 2,602 19 o0.3 nd 2.36 1.85  2.69 0.02
34°37’ 17°03’ 2,602 21 o0.3 0.9 2.38 2.36  2.46 0.03
34°37’ 17°03’ 2,602 23 o0.3 nd 2.42 1.56  2.77 0.02
Deg. E, degree east; Deg. S, degree south; Fe
HR
, highly reactive Fe oxides operationally defined as Fe leachable by Na dithionate (see text for method); nd, not determined; olw, overlying water.
Olw includes seawater collected within 30 m of the seafloor by Niskins for measuring [O
2
], and seawater trapped by Mega Corer within 0.4 m of the seafloor for metal determinations.
*[O
2
] in overlying water was measured by Winkler titration. Each pore water value reported is the mean of 20 microsensor measurements across
±
0.1 cm; o0.3 mM is below detection limit.
ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms3143
4 NATURE COMMUNICATIONS | 4:2143 | DOI: 10.1038/ncomms3143 | www.nature.com/naturecommunications
& 2013 Macmillan Publishers Limited. All rights reserved.

RD at these sites compared with previous continental margin
studies (Fig. 4).
Potential sources of heavy Fe isotopes include dissolution of
sulphide, oxide and silicate minerals. The equilibrium isotopic
fractionation between Fe(II)
(aq)
and FeS
(s)
has been experimen-
tally determined (  0.32
±
0.29%)
37
, but previous field
observations of Fe(II)
(aq)
in the presence of FeS
(s)
have been
restricted below the ferruginous zone
21
, where pore waters do not
reflect equilibrium conditions
37
. In surface sediments, FeS
(s)
is
commonly unstable and readily forms Fe oxyhydroxide
minerals
38
. Multiple solid-phase spectroscopic analyses did
not identify any FeS or FeS
2
minerals in the oxidizing
surface layer of the Cape margin study sites. Therefore, any Fe-
S minerals physically entrained in the surface layers have
probably already contributed to the authigenic pool of reactive
Fe oxide minerals.
Isotopically heavy Fe in pore water Fe could be attributed to an
equilibrium isotopic effect during NRD of oxide and/or silicate
weathering products on the Cape margin; for example, the
isotopic composition of Fe in oxidizing pore waters is also heavy
( þ 0.16
±
0.05%)
23
in deep-sea volcanogenic turbidites, where
dissolved Fe is dominated by colloids formed by inorganic
dissolution of volcanic minerals
34
. Other disparate lines of
evidence exist for a common equilibrium isotopic effect; first,
dissolved Fe released to seawater from atmospheric dust is heavy
(d
56
Fe of þ 0.13
±
0.18%)
26
, and confirmed beneath Saharan
dust plumes in the North Atlantic where elevated surface ocean
Fe concentrations have a d
56
Fe of þ 0.33
±
0.05% (ref. 39);
second, dissolved Fe in river water is heavy (d
56
Fe of
þ 0.14
±
0.28%)
27
and persevered through the estuarine mixing
zone despite intense removal during flocculation, indicative of
isotopic equilibration with suspended solids; and finally, dissolved
Fe in New Guinea coastal waters influenced by sediment re-
suspension has a d
56
Fe of 0.37
±
0.15%, 0.2% heavier than the
suspended particles
28
. The remarkable consistency in dissolved Fe
isotopic compositions across a diverse set of oxygenated
sediment–seawater interactions is used here to predict the mean
isotopic fingerprint of dissolved Fe released by NRD at our study
sites (mean d
56
Fe ¼þ0.22
±
0.18%), and we find this consistent
with the predictions of Radic et al.
28
We consider the amount of Fe in surface sediment pore waters
that may originate from NRD on the Cape margin using the
estimated end-member isotopic composition of dissolved Fe
supplied by reductive and NRD processes. We derive isotope-
mixing lines in Fig. 4 with a standard two-component mixing
calculation, where the slope (a) and intercept (b) describe the
relationship between d
56
Fe and [Fe] in equation (1). Equations
(2) and (3) define terms a and b, where the two end-member
isotopic compositions and their respective concentrations in the
pore water are set by reasoned constraints from the literature and
values befitting to the Cape margin data set;
d
56
Fe ¼ a= Fe½b ð1Þ
a ¼ Fe
RD
½Fe
NRD
½ðd
56
Fe
NRD
 d
56
Fe
RD
Þ= Fe
RD
½Fe
NRD
½ðÞ
ð2Þ
b ¼ Fe
RD
½d
56
Fe
RD
 Fe
NRD
½d
56
Fe
NRD
ð3Þ
Fe supplied by RD (Fe
RD
) is defined as having d
56
Fe
RD
¼
 3.0% (refs 17,20), and a hypothetical concentration of
1.75 mmol l
 1
, whereas d
56
Fe
NRD
¼ 0.22
±
0.18%. When
[Fe
NRD
] is set between 0.07 and 0.6 mmol l
 1
in the surface
sediment (a majority of the observed pore water [Fe]), these
parameters provide two mixing lines (Fig. 4), which approximate
the observed relationship between d
56
Fe and [Fe] between
30
20
10
0
Depth (cm)
2001000
a
b
c
403020100
543210
30
20
10
0
Depth (cm)
403020100
2001000
543210
30
20
10
0
Depth (cm)
543210
403020100
2001000
30
20
10
0
–3 –2 –1 0 1
30
20
10
0
–3 –2 –1 0 1
30
20
10
0
–3 –2 –1 0 1
Average solid
Average solid
O
2
NO
3
–
Mn
Fe
Sulfidic
Oxidizing
Bottom water
Ferruginous
Fe and Mn (μmol l
–1
)
δ
56
Fe (‰)
O
2
(μmol l
–1
)
NO
3
–
(μmol l
–1
)
δ
56
Fe (‰)
Fe and Mn (μmol l
–1
)
O
2
(μmol l
–1
)
NO
3
–
(μmol l
–1
)
δ
56
Fe (‰)
Fe and Mn (μmol l
–1
)
O
2
(μmol l
–1
)
NO
3
–
(μmol l
–1
)
Average solid
Figure 3 | Pore water geochemical profiles for Cape margin. Down-core
distribution of dissolved O
2
,NO
3

, Mn and Fe in Cape margin sediment
pore water from (a) 733 m, (b) 1,182 m and (c) 2,602 m water depths.
Chemical zones are described with respect to dissolved Fe as ‘oxidizing’
(O
2
and/or NO
3

in the pore water), ‘ferruginous’ or ‘sulphidic’ (Fe
depletion consistent with SO
4

reduction). Corresponding pore water iron
isotopic compositions express distinct behaviour across chemical zones,
which are broadly reproduced at each site (see text for further details).
NATURE COMMUNICATIONS | DOI: 10.1038/ncomms3143 ARTICLE
NATURE COMMUNICATIONS | 4:2143 | DOI: 10.1038/ncomms3143 | www.nature.com/naturecommunications 5
& 2013 Macmillan Publishers Limited. All rights reserved.

ferruginous and oxidizing zones on the Cape margin. Pore water
Fe isotopic data and the implied mechanism of pore water Fe
supply have implications for the isotopic composition and flux of
Fe from the Cape margin, as well as the geographical distribution
and isotopic diversity of Fe supply to the global ocean.
Iron supply from the Cape margin. A one-dimensional
diffusion-oxidation calculation using pore water Fe and O
2
con-
centration data
40
has previously been used to estimate the benthic
flux of Fe from sites on the California and Oregon shelves, where
it was validated by direct field comparison with in situ and ex situ
incubation techniques
12
. We use the same approach to estimate
the benthic flux of Fe to bottom waters from the Cape margin.
The flux (mol cm
 2
s
 1
)ofFe
2 þ
(J) is described in equation (4),
where j is the sediment porosity, L is the thickness (cm) of the
oxygenated layer of surface sediment and C
p
is the concentration
(g cm
 3
)ofFe
2 þ
in the pore water beneath L .
J ¼ jðD
s
k
1
Þ
0:5
C
p

= sinh ðk
1
=D
s
Þ
0:5
L

ð4Þ
The rate constant for Fe
2 þ
oxidation is described by k
1
(s
 1
)
in equation (5), as a function of bottom water O
2
concentration
(mol kg
 1
), pH and the value of k (3.54  10
 1
kg mol
 1
s
 1
)
derived from the relationships between temperature (6 °C) and
salinity (34) with Fe
2 þ
oxidation in seawater
41
.
k
1
¼ k O
2
½OH

½ ð5Þ
The diffusion coefficient
42
of Fe
2 þ
in muddy shelf sediment
pore waters (cm
2
s
 1
) is described in equation (6), as a function
of j, corrected for tortuosity, and temperature T (°C).
D
s
¼ j
1:7
3:31 þ 0:15TðÞ10
 6
ð6Þ
We use a j value of 0.77 determined by Multi-Scan Core
Logging of archived sediment cores at the National Oceano-
graphy Centre (NOC), and in the absence of direct determina-
tion, pore water pH is assumed to be 7.5
±
0.1 (ref. 12). Bottom
water [O
2
] is derived from shipboard Winkler titrations of near-
bottom water samples of 201 and 174 mmol l
 1
at 733 and
1,182 m, respectively. Values of L are derived from O
2
microsensor determinations from sites 733 and 1,182 m of 1.05
and 0.62 cm, respectively, and we use a depth-corrected pore
water [Fe] beneath L to derive corresponding C
p
values of 0.84
and 0.24 mmol l
 1
. The calculated fluxes of Fe to bottom water
from the 733 and 1,182 m Cape margin sites are 0.11
±
0.13 and
0.23
±
0.17 mmol m
 2
d
 1
, respectively.
We find values of benthic Fe flux to be 2–4 orders of
magnitude smaller than widely cited and most recent constraints
from active-tectonic and dysoxic continental margins
8–10,43
.
Furthermore, the flux of Fe is substantially lower than predicted
by the application of the Elrod et al.
8
relationship to organic
C oxidation rate; Elrod et al.
8
show that benthic Fe fluxes
correlate with rates of organic C oxidation for numerous sites
along the California margin, and the relationship is used to
extrapolate benthic fluxes of Fe from organic C fluxes for ocean
biogeochemical models
3,4
. Using a one-dimensional steady-state
O
2
diffusion-consumption model, we approximate the rate of
organic C oxidation by fitting calculated outputs to the observed
down-core profiles of O
2
concentration determined by
microsensors (Fig. 5). The approach follows Berner
44
, in which
a single pool of reactive organic C is assumed to be the only
mechanism of O
2
utilization (for example Papadimitriou et al.
45
),
and the influence of biophysical mixing, seasonal accumulation
rates and porosity structure are ignored. The rates of organic C
oxidation for sites 733 and 1,182 m (3.5 and 5.1 mmol m
 2
d
 1
)
are derived from the modelled flux of O
2
and the stoichiometry of
organic matter remineralization, and are equivalent to rates of
organic matter decomposition from previous sites of benthic Fe
flux determination
8
. Following Elrod et al.
8
, we predict a benthic
Fe flux of 2.4–3.5 mmol m
 2
d
 1
from the Cape margin (Fig. 5),
and find this an order of magnitude more than we calculate from
pore water Fe data. Thus, Cape margin sediments appear to deviate
from the relationship between sediment respiration and benthic Fe
supply rates observed on North American river-dominated and
dysoxic margins of the Pacific Ocean.
Discussion
Cape margin sediments indicate that the supply of Fe to the
southeast Atlantic Ocean is smaller and isotopically heavier than
current models of Fe cycling would suggest
8,9,24
(Fig. 6). Both of
these findings have widespread implications for the marine Fe
cycle. It appears that the relationship proposed by Elrod et al.
8
,
which for a decade has provided the most widely used constraint
on the benthic flux of Fe to the global oceans, may overestimate
benthic Fe flux from margin sediments with lower reactive Fe
inventories or less-effective mechanisms for Fe enrichment.
Continental margin sediments are a reservoir for reactive iron
mineral substrates, which supply dissolved Fe to the oceans and
thereby support ocean life. Globally, rivers provide three quarters
–4
–3
–2
–1
0
1
0.1 1 10 100
M
ixing
RD: –3.0 ‰
IRMM-14
Oxidizing
bottom water
Oxidizing
pore water
Ferruginous
pore water
NRD: 0.2 ‰
Dissolved [Fe] (μmol l
–1
)
δ
56
Fe (‰)
Riverine/dysoxic
margins
Crozet Is. 3,400 m
23
Crozet Is. 4,200 m
23
Semi-arid/oxic
(cape margin)
733 m
1,182 m
2,602 m
Eel river
9,23
Umpqua river
9
Santa monica basin
9
Volcanic/oxic
deep sea
Figure 4 | Benthic d
56
Fe signatures of marine sediment dissolution.
The log-normal scatter plot compares the concentration and isotopic
composition of dissolved Fe observed between the oxidizing and
ferruginous zones of pore waters from the semi-arid Cape margin (this
study), river-dominated sites of California and Oregon margins, and
volcanogenic deep-sea sediments of the Southern Ocean. The relatively low
abundance of dissolved Fe on the Cape margin is attributed to the scarcity
of reducible Fe oxides in these sediments. The isotopic composition of
dissolved Fe in the ‘oxidizing’ pore waters of the Cape margin are best
described by mixing between RD and NRD end-member Fe sources. The
high abundance of Fe with NRD isotopic compositions in the volcanic and
oxygenated deep sea is consistent with more rapid alteration of the primary
volcanic minerals derived from the Crozet Island basalts
34
.
ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms3143
6 NATURE COMMUNICATIONS | 4:2143 | DOI: 10.1038/ncomms3143 | www.nature.com/naturecommunications
& 2013 Macmillan Publishers Limited. All rights reserved.

of the particulate Fe content of continental margins
46
, where on
average sediments comprise 1 wt% highly reactive Fe oxides
(Fe
HR
)
47
. Tectonic uplift enhances sediment transport to the
ocean from active orogenic belts due to many factors (fractured
and brecciated rocks, over-steepened slopes and seismic and
volcanic activity) in addition to elevation/relief
48
. Hydro-climatic
conditions provide a second-order influence on sediment
transport to the oceans
48,49
, where continental run-off intensity
increases the Fe
HR
enrichment of sediment carried by rivers
47
.
For example, Indian margin sediments record monsoon intensity
with enrichments of Fe
HR
and total Fe (Fe
T
) to Al ratios nearly
double the continental average on decadal timescales
50,51
.
Therefore, tectonics and climate are effective ways to mediate
the supply and enrichment of Fe
HR
at ocean margins.
The Cape margin is depleted in Fe
HR
(0.17 wt%, 0–20 cm)
compared with the global average of continental margins
47
(Fig. 2). Prolonged tectonic stability in this region and a semi-
arid climate have probably contributed to the relatively slow rate
of sediment accumulation
31
—perhaps allowing the rate of Fe
HR
reduction and dissolution to meet or exceed the rate of Fe
HR
supply and produce the low Fe
HR
content observed. We consider
the limited abundance of reactive Fe oxide minerals as the most
likely means of restricting the flux of dissolved Fe to seawater in
this region. Thus, shifting patterns in tectonic and hydro-climatic
conditions might have influenced the Fe inventory of margin
sediments in the past, with unknown impact on regional and
global inputs of dissolved Fe to seawater.
The limited amount of RD on the Cape margin reveals the co-
existence of a NRD process, releasing Fe in the oxidizing layer of
surface sediments. The discovery is consistent with predictions
based on seawater isotopic compositions of Fe and Nd measured
elsewehere
28,52
. Dissolution rates may be slower than microbially
catalysed RD, but the process could be widespread. Non-reductive
Fe dissolution is likely to reflect physical and compositional
variations in sediments influenced by geological provenance and
weathering rates, as indicated by the high abundance of Fe in
oxidizing volcanic sediments around the Crozet Islands (Fig. 4).
Using a sediment respiration parameter to estimate Fe supply
from marine sediments is unlikely to account for the distribution
and magnitude of Fe released by NRD of marine sediment.
Readily weathered volcanic and oxygenated sediments are
prevalent across ocean basins
34
, and dilute sediment
suspensions can be transported hundreds of kilometres offshore
where they influence primary production
13,53
, so multiple
mechanisms for Fe dissolution may have a far-reaching
influence on seawater isotopic compositions and productivity.
Cape margin Fe isotopic data remain consistent with inter-
pretations of isotopically light Fe in modern marine and ancient
sediment records
25,54
in which anoxic seawater is suggested to
shuttle isotopically light Fe from ferruginous margin sediments to
ocean basins. However, the measured or inferred absence of light
Fe isotopic compositions in seawater would be a poor assessment
of the oceanic Fe inventory and its impact on primary
productivity, given that we now need to consider the variables
of non-reductive Fe dissolution.
Cape margin sediments shed light on Fe exchange between a
semi-arid tectonically passive continental margin and oxygenated
ocean. The sediments host distinct mechanisms of Fe dissolution,
resulting in a smaller, isotopically heavier input to seawater than
predicted. Semi-arid passive margin environments are common,
and their distribution varies over time, thus requiring appraisal
when reconstructing past ocean conditions. In addition to the
distribution of ocean anoxia, we predict that the proportion of
dissolved Fe supplied to the ocean by reductive and non-reductive
sediment dissolution will reflect patterns in continental weath-
ering and transport to ocean margins. Slow rates of reactive Fe
1.50
1.25
1.00
0.75
0.50
0.25
0.00
Depth (cm)
200150100500
Site 1,182 m
Site 733 m
Measured data
Model fits
Pore water [O
2
] μmol l
–1
Modelled O
2
flux = 4.5–6.1 mmol m
–2
d
–1
C oxidation rate = 3.5–5.1 mmol m
–2
d
–1
Fe flux predicted = 2.4–3.5 μmol m
–2
d
–1
Figure 5 | Benthic Fe flux prediction from C oxidation rates. Comparison
of pore water O
2
concentration versus depth profiles with model fits to data
from steady-state pore water O
2
consumption by organic C oxidation
44
.
The corresponding rates of organic C oxidation are used to predict benthic
fluxes of Fe from the Cape margin based on correlated C oxidation and Fe
flux observations from the California margin
8
. However, despite comparable
rates of organic C oxidation between these regions, this approach predicts
benthic Fe fluxes higher than calculated from pore water Fe data in this
study.
Santa monica basin
8
Eel/Ump. river margins
12
–5
–4
–3
–2
–1
0
1
0.1 1 10 100 1,000
IRMM-14
Cape margin (semi-arid/oxic)
733 m
1,182 m
Previous studies (riverine/dysoxic)
S. Oregon (umpqua river) margin*
N. California (eel river) margin*
Benthic flux dissolved Fe (μmol m
–2
d
–1
)
δ
56
Fe (‰)
Peru anoxic margin
10
San pedro basin
8
Figure 6 | Benthic flux and isotopic composition of dissolved Fe. The
benthic flux of dissolved Fe and its corresponding isotopic composition is
shown for the semi-arid and passive-tectonic Cape margin of South Africa
(this study) compared with previous studies of river-dominated, dysoxic
and active-tectonic margin settings. Coloured bars correspond to the range
of benthic Fe flux determinations where Fe isotopic composition was not
also determined.
NATURE COMMUNICATIONS | DOI: 10.1038/ncomms3143 ARTICLE
NATURE COMMUNICATIONS | 4:2143 | DOI: 10.1038/ncomms3143 | www.nature.com/naturecommunications 7
& 2013 Macmillan Publishers Limited. All rights reserved.

substrate delivery to ocean margins may limit the benthic flux of
Fe by RD relative to organic C oxidation rates. In addition, young
volcanic terrains that are easily weathered are likely to release
greater amounts of Fe by NRD compared with the mature
sediment lithologies of the Cape margin. Regional constraints on
oceanic Fe supply are missing links for modelling the coupled
ocean-atmosphere carbon cycle
3,4
. Therefore, evaluating Fe
supply from additional and diverse ocean boundaries remains
an essential goal for modelling climate.
Methods
Sediment and pore water sampling . A Bowers and Connelly Megacore and Box
core sampled shallow (o0.4 m) sediment and pore water from three sites on the
Cape margin (733, 1,182 and 2,602 m; Fig.1, Table 1) during the UK GEOTRACES
A10 expedition from the RRS Discovery (D357) in October–November 2010. Sub-
samples for oxygen profiling and pore water extraction were collected by poly-
carbonate push core, and shipboard processing was performed at ambient bottom
water temperatures.
Rhizon samplers collected dissolved pore water and overlying seawater
constituents for elemental analysis (filtration cut-off of 0.15 mm). Rhizon samplers
(50  2.5 mm) were pre-soaked in 18.2 MO de-ionized water and inserted through
pre-drilled holes in core tubes at 1–3 cm intervals down-core. A ‘BD Discardit’
20 ml syringe (pre-cleaned: 72 h 10% Decon; 72 h 6 M HCl; 72 h 6 M HNO
3
; rinsed
by 18.2 MO de-ionized water) and secured to each Rhizon by Luer-lock connection.
A brace inserted between each syringe housing and plunger applied suction to
Rhizons. The first 0.5 ml of sample was discarded. Rhizons drew B6 ml of pore
water (10–20 min) and were divided for macronutrient and metal analysis.
The aliquot for metals was acidified (pHo2) directly through the syringe tip
(6 ml of 6 M quartz-distilled (Q-)HCl per ml of pore water) and transferred to
low-density polyethylene (LDPE) pots. Residual sediments were divided for
elemental analyses by extruding and slicing cores with a Teflon sheet at 1–2 cm
depth intervals. Sediments were later freeze-dried and homogenized by agate
pestle and mortar.
Sediment digestion and leaching procedures
. Heated Aqua Regia and combined
HF-HClO
4
acid dissolved sediment samples following an established protocol at
the NOC in Southampton
34
. Digested sediment residues were re-dissolved in
6 M HNO
3
in preparation for analysis by Inductively Coupled Plasma-Mass
Spectrometry (ICP-MS). Digestion of the certified standards SCO-1, SGR-1
(United States Geological Survey) and BCSS-1 (National Research Council Canada)
allowed for the recovery of Fe, Mn and Al within consensus values (Table 2).
Highly reactive Fe (Fe
HR
) is operationally defined as Fe liberated by dissolution
during a 2-h reaction with Na dithionite designed to target reducible Fe oxide
phases (for example, ferrihydrite, lepidocrosite, goethite and haematite)
55
.A10ml
aliquot of 50 g l
 1
Na-dithionite solution buffered to pH 4.8 with 0.35 M acetic
acid and 0.2 M Na citrate was reacted at room temperature with B100 mg of dry
sediment sample. The resultant solution was spun for 8 min by centrifuge at
9,000 g. Supernatant containing dissolved Fe was decanted into LDPE before
analysis by ICP-MS.
Element abundance by ICP-MS
. A Perkin Elmer Element X2 ICP-MS measured
the concentration of Fe and Mn in pore water, and Fe and Al in sediment solutions
at the NOC. Sediment digests were diluted 1,000 times, and pore water samples 100
times, with a 0.48 M Q-HNO
3
solution containing 2 ng g
 1
Re, Rh and Sc and
5ngg
 1
Be as internal standards. External calibration standards were prepared
from certified stock solutions. For dilute pore water analyses, standards were
matrix matched with 1% seawater (NASS-5; National Research Council
Canada), effectively adding a small known amount of Fe (35 pmol l
 1
) and Mn
(160 pmol l
 1
) to the standards that is corrected for during external calibration.
Samples were introduced by an Elemental Scientific SC-4 DX Autosampler and
PC
3
Peltier cooled inlet system with integrated cyclonic spray chamber at
100 mlmin
 1
. Masses
56
Fe,
55
Mn and
27
Al were measured in medium resolution
mode. Internal standards were monitored throughout and used to correct for the
reduction in signal intensity over time. The accuracy of the method was verified by
the intermittent analysis of blank-bracketed SLRS-5 within certified values
(Table 2), with a relative s.d. o0.5%. The detection limits (3 s.d. of analytical
blanks, n ¼ 11) for Fe, Mn and Al were 24, 4 and 14 nmol l
 1
, respectively. Mean
procedural blanks (n ¼ 4) for pore water sampling and analyses were
85
±
3.5 nmol l
 1
for Fe and below detection for Mn.
Fe isotope determinations by Multi-Collector ICP-MS
. The isotopic composi-
tion of Fe in pore waters and sediments was assessed using a modification of John
and Adkins
7
. Sample solutions containing 17–115 ng of Fe were quantitatively
spiked with a
57
Fe–
58
Fe double spike
56
using a spike:sample ratio of 2:1. Spike-
sample mixtures were dried in Savillex PFA Teflon vials and re-dissolved with 5 M
Q-HCl þ 0.001% v/v Fisher Scientific Optima. A 135 ml aliquot of acid-cleaned
AG-MP1 anion exchange resin was used in LDPE columns (pre-cleaned: 72 h 10%
v/v decon, 1 week 6 M HCl) for the separation of Fe from sample matrices.
Resin-filled columns were rinsed with 2M Q-HNO
3
and conditioned with 5 M
Q-HCl þ 0.001% v/v H
2
O
2
before loading spiked samples in 100 ml aliquots.
Loaded columns were rinsed by 12  100 ml aliquots of 5 M Q-HCl þ 0.001% v/v
H
2
O
2
. Fe was eluted by 800 m l of 1 M Q-HCl into Savillex PFA Teflon vials, dried
and re-dissolved in 2 ml of 0.1 M Q-HNO
3
before analysis by Multi-Collector (MC)
ICP-MS. Column calibrations assessed procedural blanks (2.7
±
0.6 ng of Fe, n ¼ 2)
and re covery of Fe (495%). Calibrations also confirmed the effective separa tion of
Fe from major salts (Ca) and interferences (
58
Ni and
54
Cr).
A Thermo Scientific (Neptune) MC ICP-MS-measured Fe isotope ratios at the
University of South Carolina. Samples and standards were introduced by a Teflon
PFA nebulizer and an (ESI) Apex-Q desolvating system at 150 mlmin
 1
, with an
Al ‘X’ skimmer cone. High-resolution mode resolved Fe from polyatomic
interferences (ArN
þ
,ArO
þ
and ArOH
þ
). Signal intensity was measured for
atomic masses 53, 54, 56, 57, 58, 60 and 61, with
53
Cr and
60
Ni used to correct for
isobaric interferences on
54
Fe and
58
Fe, respectively. Signal intensity was measured
over 50 cycles of 4.2 s. The first 12 cycles were discarded due to uptake and
stabilization time. Any cycles with ratios more than 3 s.d. of the remaining 38
cycles were also discarded. Memory effects were minimized by a 3-min rinse
(0.32 M Q-HNO
3
) between analyses. All sample intensities were blank-corrected
with the mean of 38 cycles from periodic 0.1 M HNO
3
analyses. Fe isotope ratios
were calculated using a double-spike data-reduction scheme based on the iterative
approach of Siebert et al.
57
, and are expressed relative to IRRM-14 using standard
delta notation (d
56
Fe):
d
56
Fe %ðÞ¼
56
Fe=
54
Fe

sample
=
56
Fe=
54
Fe

IRMM-014
 1

1;000
Sample ratios are expressed relative to the average of IRMM-14 standards
mixed with the
57
Fe–
58
Fe double spike in equivalent proportions and concentra-
tions as samples. Standard-spike ratios and concentrations were assessed for
deviation in IRRM-14 determination but none was found. Each sample was
analysed twice, and the average is shown (Table 1). Uncertainty for d
56
Fe is
expressed as the mean s.e. of the isotope ratio over each 160 s analysis, based on
previous demonstration that uncertainty of Fe isotopic measurement of a natural
sample by double-spike MC ICP-MS is dominated by internal error
56
.
Pore water O
2
profiling. Unisense equipment was used to for shipboard O
2
determination in surface sediments as previou sly described
34
. Linear calibration s
were performed between aerated seawater and anoxic (N
2
saturated) seawater
before each use, with a detection limit o0.3 mmol l
 1
. A micromanipulator and
SensorTracePro software controlled down-core profiling at 100 mm depth intervals.
Data were converted to dissolved O
2
concentration using empirical constraints for
O
2
saturation in seawater. Pore water profiles are surface-normalized to bottom
water O
2
determinations from shipboard Winkler titrations.
Table 2 | Measured elemental abundance of certified reference materials.
CRM Measured (
±
1 s.d.) Certified (
±
1 s.d.)
Fe Mn Al Fe Mn Al
SLRS-5* (n ¼ 5, p.p.b.) 93.4
±
1.6 4.53
±
0.10 51.8
±
1.9 91.2
±
5.8 4.33
±
0.18 49.5
±
5.0
SCO-1
w
(n ¼ 1, wt%) 3.76 nd 7.15 3.59
±
0.13 nd 7.23
±
0.22
SGR-1
w
(n ¼ 1, wt.%) 2.04 nd 3.17 2.12
±
0.10 nd 3.34
±
0.11
BCSS-1* (n ¼ 1, wt%) 3.20 nd 6.07 3.29
±
0.10 nd 6.26
±
0.22
nd, not determined; p.p.b., parts per billion.
*National Research Council Canada (SLRS-5, river water; BCSS-1, marine sediment).
wUnited States Geological Survey (SCO-1, Cody Shale; SGR-1, Green River Shale).
ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms3143
8 NATURE COMMUNICATIONS | 4:2143 | DOI: 10.1038/ncomms3143 | www.nature.com/naturecommunications
& 2013 Macmillan Publishers Limited. All rights reserved.

Pore water NO
3

and NO
2

determination. Pore water samples (2 ml) were
diluted with 18.2 MO de-ionized water to a volume of 30 ml. A 5-channel Bran and
Luebbe AAIII segmented flow colorimetric autoanalyser was used to determine
NO
3

and NO
2

concentration using standard analytical techniques at sea
58
. Data
presented are combined NO
3

þ NO
2

, with an analytical uncertainty of
±
0.2 mmol l
 1
. Accuracy was verified by determination of nutrient standards
(Ocean Scientific International) within 5% of certified values.
Sediment mineralogical description
. Polarized light microscopy, XRD and SEM
with EDS assessed bulk sample mineralogy. A Philips X’Pert pro instrument with
Cu-Ka radiation performed XRD. Elemental composition of targeted mineral
grains was assessed from EDS generated from a Princeton Gamma Tech (IMIX-
PTS) X-ray beam with a 2–3 mm diameter connected to an LEO 1450VP SEM
operated at 15 keV.
Sediment organic C determination
. Total organic carbon concentrations were
calculated from the difference between coulometric determination (UIC 5012
Coulometer) of total carbon (TC) and total inorganic carbon (TIC) content of dry
homogenized sediments. TC was calculated from CO
2
released during sample
combustion, and TIC was calculated from CO
2
released during heated sample
reaction with 1.5 M H
3
PO
4
. Accuracy of TC and TIC determinations was assessed
with anhydrous CaCO
3
powder, with a mean recovery of 100.4
±
0.8% (1 s.d.,
n ¼ 15). The limit of detection (3 s.d. of blanks) was o10 mg C, equivalent to o0.03
wt% total organic carbon.
References
1. Boyd, P. W. & Ellwood, M. J. The biogeochemical cycle of iron in the ocean.
Nat. Geosci. 3, 675–682 (2010).
2. Lefe
`
vre, N. & Watson, A. J. Modeling the geochemical cycle of iron in the
oceans and its impact on atmospheric CO2 concentrations. Global Biogeochem.
Cycles 13, 727–736 (1999).
3. Moore, J. K. & Braucher, O. Sedimentary and mineral dust sources of dissolved
iron to the world ocean. Biogeosciences 5, 631–656 (2008).
4. Tagliabue, A., Bopp, L. & Aumont, O. Evaluating the importance of
atmospheric and sedimentary iron sources to Southern Ocean biogeochemistry.
Geophys. Res. Lett. 36, L13601 (2009).
5. Anderson, R. F. et al. GEOTRACES—an international study of the global
marine biogeochemical cycles of trace elements and their isotopes. Chemie der
Erde Geochem. 67, 85–131 (2007).
6. Lacan, F. et al. Measurement of the isotopic composition of dissolved iron in
the open ocean. Geophys. Res. Lett. 35, L24610 (2008).
7. John, S. G. & Adkins, J. F. Analysis of dissolved iron isotopes in seawater. Mar.
Chem. 119, 65–76 (2010).
8. Elrod, V. A., Berelson, W. M., Coale, K. H. & Johnson, K. S. The flux of iron
from continental shelf sediments: a missing source for global budgets. Geophys.
Res. Lett. 31, L12307 (2004).
9. Severmann, S., McManus, J., Berelson, W. M. & Hammond, D. E. The
continental shelf benthic iron flux and its isotope composition. Geochim.
Cosmochim. Acta 74, 3984–4004 (2010).
10. Noffke, A. et al. Benthic iron and phosphorous fluxes across the Peruvian
oxygen minimum zone. Limnol. Oceanogr. 57, 851–867 (2012).
11. Froelich, P. N. et al. Early oxidation of organic matter in pelagic sediments of
the eastern equatorial Atlantic: suboxic diagenesis. Geochim. Cosmochim. Acta
43, 1075–1090 (1979).
12. Homoky, W. B. et al. Dissolved oxygen and suspended particles regulate the
benthic flux of iron from continental margins. Mar. Chem. 134-135, 59–70
(2012).
13. Lam, P. J. & Bishop, J. K. B. The continental margin is a key source of iron to
the HNLC North Pacific Ocean. Geophys. Res. Lett. 35, L07608 (2008).
14. Siedlecki, S. A., Mahadevan, A. & Archer, D. E. Mechanism for export of
sediment-derived iron in an up-welling regime. Geophys. Res. Lett. 39, L03601
(2012).
15. Murray, R. W., Leinen, M. & Knowlton, C. W. Links between iron input and
opal deposition in Pleistocene equatorial Pacific Ocean. Nat. Geosci. 5, 270–274
(2012).
16. Pollard, T. R. et al. Southern Ocean deep-water carbon export enhanced by
natural iron fertilization. Nature 457, 577–580 (2009).
17. Crosby, H. A., Roden, E. E., Johnson, C. M. & Beard, B. L. The mechanisms of
iron isotope fractionation produced during dissimi latory Fe(III) reduction by
Shewanella putrefaciens and Geobacter sulfurreducens. Geobiology 5, 169–189
(2007).
18. Beard, B. L. et al. Iron isotope fractionation between aqueous ferrous iron and
goethite. Earth Planet. Sci. Lett. 295, 241–250 (2010).
19. Wu, L. et al. Stable iron isotope fractionation between aqueous Fe(II) and model
Archean ocean Fe-Si coprecipitates and implications for iron isotope variations in
the ancient rock record. Geochim. Cosmochim. Acta 84, 14–28 (2012).
20. Wu, L., Beard, B. L., Roden, E. E. & Johnson, C. M. Stable Iron Isotope
Fractionation Between Aqueous Fe(II) and Hydrous Ferric Oxide. Environ. Sci.
Technol. 45, 1847–1852 (2013).
21. Severmann, S., Johnson, C. M., Beard, B. L. & McManus, J. The effect of early
diagenesis on the Fe isotope compositions of porewaters and authigenic
minerals in continental margin sediments. Geochim. Cosmochim. Acta 70,
2006–2022 (2006).
22. Bergquist, B. A. & Boyle, E. A. Iron isotopes in the Amazon River system:
weathering and transport signatures. Earth Planet. Sci. Lett. 248, 54–68 (2006).
23. Homoky, W. B. et al. Pore-fluid Fe isotopes reflect the extent of benthic Fe
redox recycling: evidence from continental shelf and deep-sea sediments.
Geology 37, 751–754 (2009).
24. John, S. G., Mendez, J., Moffett, J. & Adkins, J. The flux of iron and iron
isotopes from San Pedro Basin sediments. Geochim. Cosmochim. Acta 93,
14–29 (2012).
25. Owens, J. D. et al. Iron isotope and trace metal records of iron cycling in the
proto-North Atlantic during the Cenomanian-Turonian oceanic anoxic event
(OAE-2). Paleoceanography 27, PA3223 (2012).
26. Waeles, M., Baker, A. R., Jickells, T. & Hoogewerff, J. Global dust
teleconnections: aerosol iron solubility and stable isotope composition. Environ.
Chem. 4, 233–237 (2007).
27. Escoube, R., Rouxel, O. J., Sholkovitz, E. & Donard, O. F. X. Iron isotope
systematics in estuaries: the case of north river, Massachusetts (USA). Geochim.
Cosmochim. Acta 73, 4045–4059 (2009).
28. Radic, A., Lacan, F. & Murry, J. W. Iron isotopes in the seawater of the
equatorial Pacific Ocean: new constraints for the oceanic iron cycle. Earth
Planet. Sci. Lett. 306, 1–10 (2011).
29. Jeandel, C. et al. Ocean margins: the missing term in oceanic element budgets?
Eos 92, 217–224 (2011).
30. Fourie, P. H. et al. Provenance and reconnaisance study of detrital zircons of
the Palaeozoic Cape Supergroup in South Africa: revealing the interaction of
the Kalahari and Rı
´
o de la Plata cratons. Int. J. Earth. Sci. 100, 527–541 (2011).
31. Compton, J. S. & Wiltshire, J. G. Terrigenous sediment export from the western
margin of South Africa on glacial to interglacial cycles. Mar. Geol. 266, 212–222
(2009).
32. Wigley, R. & Compton, J. S. Oligocene to Holocene glauconite-phosphorite
grains from the Head of he Cape Canyon on the western margin of South
Africa. Deep-Sea Res. II 54,
1375–1395 (2007).
33. Canfield, D. E. & Thamdrup, B. Towards a consistent classification scheme for
geochemical environments, or, why we wish the term ‘suboxic’ would go away.
Geobiology 7, 385–392 (2009).
34. Homoky, W. B. et al. Iron and Manganese diagenesis in deep sea volcanogenic
sediments and the origins of pore water colloids. Geochim. Cosmochim. Acta
75, 5032–5048 (2011).
35. Jones, M. E., Beckler, J. S. & Taillefert, M. The flux of soluble organic-iron(III)
complexes from sediments represents a source of stable iron(III) to estuarine
waters and to the continental shelf. Limnol. Oceanogr. 56, 1811–1823 (2011).
36. Beard, B. L. et al. Application of Fe isotopes to tracing the geochemical and
biological cycling of Fe. Chem. Geol. 195, 87–117 (2003).
37. Wu, L. et al. Experimental determination of iron isotope fractionations among
‘Fe2 þ (aq)-FeS(aq)-Mackinawite at low temperatures: Implications for the rock
record. Geochim. Cosmochim. Acta 89, 46–61 (2012).
38. Luther, I. I. I., George, W., Giblin, A., Howarth, R. W. & Ryans, R. A. Pyrite and
oxidized iron mineral phases formed from pyrite oxidation in salt marsh and
estuarine sediments. Geochim. Cosmochim. Acta 46, 2665–2669 (1982).
39. John, S. G. & Adkins, J. The verticle distribution of iron stable isotopes in the
North Atlantic near Bermuda. Global Biogeochem. Cycles 26, GB2034 (2012).
40. Raiswell, R. & Anderson, T. F. Reactive iron enrichment in sediments deposited
beneath euxinic bottom waters: constraints on supply by shelf recycling. Geol.
Soc. 248, 179–194 (2005).
41. Millero, F. J., Sotolongo, S. & Izaguirre, M. Oxidation kinetics of Fe(II) in sea
water. Geochim. Cosmochim. Acta 51, 793–801 (1987).
42. Grant, W. D. & Madsen, O. S. The continental-shelf bottom boundary layer.
Ann. Rev. Fluid Mech. 18, 265–305 (1986).
43. Berelson, W. M. et al. A time series of benthic flux measurements from
Monterey Bay, CA. Continental Shelf Res. 23, 457–481 (2003).
44. Berner, R. A. Early Diagenesis: A Theoretical Approach (Princeton University
Press, Princeton, NJ, 1980).
45. Papadimitriou, S., Kennedy, H. & Thomas, D. N. Rates of organic carbon oxidation
in deep sea sediments in the eastern North Atlantic from pore water profiles of O2
and the d13C of dissolved inorganic carbon. Mar. Geol. 212, 97–111 (2004).
46. Raiswell, R. et al. Contributions from glacially derived sediment to the global
iron (oxyhydr)oxide cycle: implications for iron delivery to the oceans.
Geochim. Cosmochim. Acta 70, 2765–2780 (2006).
47. Poulton, S. W. & Raiswell, R. The low-temperature geochemical cycle of iron:
from continental fluxes to marine sediment deposition. Am. J. Sci. 302, 774–805
(2002).
NATURE COMMUNICATIONS | DOI: 10.1038/ncomms3143 ARTICLE
NATURE COMMUNICATIONS | 4:2143 | DOI: 10.1038/ncomms3143 | www.nature.com/naturecommunications 9
& 2013 Macmillan Publishers Limited. All rights reserved.

48. Milliman, J. D. & Syvitski, J. P. M. Geomorphic/tectonic control of sediment
discharge to the ocean: the importance of small mountainous rivers. J. Geol.
100, 525–544 (1992).
49. Gislason, S. R. et al. Direct evidence of the feedback between climate and
weathering. Earth. Planet. Sci. Lett. 277, 213–222 (2009).
50. Pattan, J. N., Parthiban, G., Gupta, S. M. & Mir, I. A. Fe speciation and Fe/Al
ratio in the sediments of southeastern Arabian Sea as an indicator of climate
change. Quaternary Int. 250, 19–26 (2012).
51. Mazumdar, A. et al. Sulfidization in a shallow coastal depositional setting:
Diagenetic and paleoc limatic implications. Chem. Geol. 322-323, 68–78 (2012).
52. Lacan, F. & Jeandel, C. Neodymium isotopes as a new tool for quantifying
exchange fluxes at the continent-ocean interface. Earth Planet. Sci. Lett. 232,
245–257 (2005).
53. Nishioka, J. et al. Iron supply to the western subarctic Pacific: importance of
iron export from the Sea of Okhotsk. J. Geophys. Res. 112, C10012 (2007).
54. Severmann, S. et al. Modern iron isotope perspective on the benthic iron
shuttle and the redox evolution of ancient oceans. Geology 36, 487–490 (2008).
55. Poulton, S. W. & Canfield, D. E. Development of a sequential extraction
procedure for iron: implications for iron partitioning in continentally-derived
particulates. Chem. Geol. 214, 209–221 (2005).
56. John, S. G. Optimizing sample and spike concentrations for isotopic
analysis by double-spike ICPMS. J. Analyt. Atomic Spectrom. 27, 2123–2131
(2012).
57. Siebert, C., Na
¨
gler, T. F. & Kramers, J. D. Determination of molybdenum
isotope fractionation by double-spike multicollector inductively coupled plasma
mass spectrometry. Geochem. Geophys. Geosyst. 2, 1032 (2001).
58. Armstrong, F. A. J., Stearns, C. R. & Strictland, J. D. H. The measurement of
upwelling and subsequent biological processes by means of the Technicon
AutoAnalyzer and associated equipment. Deep Sea Res. 14, 381–389 (1967).
59. Rudnick, R. L. & Gao, S. in The Crust Vol. 3 (ed. Rudnick, R. L.) 1–64 (Elsevier-
Pergamon, 2003).
60. Siebert, C. et al. Molybdenum isotope signatures in continental margin marine
sediments. Earth Planet. Sci. Lett. 241, 723–7 33 (2006).
Acknowledgements
We thank the captain and crew of the RRS Discovery and UK GEOTRACES expedition
D357. We are grateful to Y.-T. Hsieh and G.M. Henderson (University of Oxford) for
supporting the collection of samples, and E.M.S. Woodward (Plymouth Marine
Laboratory) and S. Reynolds (CSIRO Marine and Atmospheric Research) for shipboard
analysis of nitrate and bottom water oxygen. J.E. Thompson (University of South-
ampton) provided valuable assistance leaching and digesting sediment samples.
A.J. Milton (University of Southampton) and A.D. Rosenberg (University of South
Carolina) provided essential technical support of analyses by ICP-MS and
MC ICP-MS, respectively. This work was supported by the UK Natural Environment
Research Council (NE/F017197/1 and NE/H004394/1) and the US National Science
Foundation (OCE-1131387).
Author contributions
W.B.H. and R.A.M. jointly conceived this study. W.B.H. designed and conducted the
approach to sampling and performed all analyses, with the exception of the method for
Fe isotope purification and analysis designed and assisted by S.G.J. and T.M.C. The
manuscript was written and edited by W.B.H., with intellectual contributions throughout
from R.A.M., S.G.J. and T.M.C.
Additional information
Competing financial interests: The authors declare no competing financial interests.
Reprints and permission information is available online at http://npg.nature.com/
reprintsandpermissions/
How to cite this article: Homoky, W. B. et al. Distinct iron isotopic signatures and
supply from marine sediment dissolution. Nat. Commun. 4:2143 doi: 10.1038/
ncomms3143 (2013).
This article is licensed under a Creative Commons Attribution 3.0
Unported Licence. To view a copy of this licence visit http://
creativecommons.org/licenses/by/3.0/.
ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms3143
10 NATURE COMMUNICATIONS | 4:2143 | DOI: 10.1038/ncomms3143 | www.nature.com/naturecommunications
& 2013 Macmillan Publishers Limited. All rights reserved.