news & views

can arise in part from the interpretation coast of Venezuela, where circulation with deposition under an anoxic, sulphate-
of different geochemical proxies used to the open ocean is restricted by physical poor system11, but we know they were
reconstruct redox conditions, from genuine barriers, resulting in localized reducing deposited in well-oxygenated, sulphate-rich
physiographic separations between different conditions at depth. Spatial redox variability marine waters. In this instance, the proxies
ocean basins around the globe, or as the can also arise in the open ocean in the form instead reflect the continued reworking of
result of secondary alteration of the original of oxygen minimum zones10. There, bursts these sediments as they migrate along the
geochemical signals8. of biological productivity in the upper ocean coast of South America. As this example
One way to wring order from this chaos lead to localized depletion of oxygen at demonstrates, a detailed understanding
is to use several adjacent locations to intermediate depths, as the sinking biomass of the depositional context is essential
reconstruct a three-dimensional framework is eventually degraded and respired. Oxygen for building robust interpretations of
of redox variability over time. Recently, this minimum zones often occur where nutrient- geochemical data.
approach was applied to the Ediacaran- rich waters from the deep ocean upwell to Poulton and colleagues4 have provided
aged Nanhua Basin in China, which led the surface, such as off the western coast intriguing evidence for early spatial
to suggestions that Ediacaran oceans may of Chile. variability in redox structure, which will
have been doubly redox-stratified, with Thus there are modern physical undoubtedly spur renewed efforts to model
a wedge of sulphidic waters separating mechanisms that could plausibly be invoked early ocean chemistry. However, much work
deep ferruginous waters from oxic to generate spatial gradients in redox integrating geochemical and stratigraphic
surface waters7. conditions in the ancient ocean, in addition data sets remains ahead of us before we can
Poulton and colleagues4 have used to any conditions perhaps unique to such have a true sense of the three-dimensional
a similar approach to investigate redox ancient biogeochemical cycling 4,8. However, redox structure in the oceans, and how this
conditions more than one billion years the expected duration of these oceanic has varied through time. ❐
earlier, examining several stratigraphic redox gradients remains an open question.
sections in the 1.8 billion-year-old Animikie Presumably the conditions would have David Fike is in the Department of Earth and
basin of North America to look for spatial needed to persist for millions of years to be Planetary Sciences, Washington University, St.
variation associated with the transition preserved in the Animikie sediments, but Louis, Missouri 63130, USA.
to sulphidic conditions. The distribution transient conditions lasting for short periods e-mail: dfike@levee.wustl.edu
of redox-sensitive phases of iron clearly cannot necessarily be ruled out.
indicates a stratified redox structure, with The apparent spatial variability found in references
a near-shore wedge of sulphidic waters redox signals at such critical biogeochemical 1. Canfield, D. E. Nature 396, 450–453 (1998).
intruding between overlying oxic surface junctures in Earth history may put many 2. Poulton, S. W., Fralick, P. W. & Canfield, D. E. Nature
waters and deeper ferruginous waters. These ill at ease. It is difficult to tell whether these 431, 173–177 (2004).
3. Brocks, J. J. et al. Nature 437, 866–870 (2005).
results imply the coexistence of ferruginous reports reflect spatially varying conditions
4. Poulton, S. W., Fralick, P. W. & Canfield, D. E. Nature Geosci.
and sulphidic conditions in spatially in the ancient ocean itself or from some 3, 486–490 (2010).
distinct zones, similar to those found in the other means, such as chemical alteration 5. Canfield, D. E. & Teske, A. Nature 382, 127–132 (1996).
Ediacaran oceans (Fig. 1). of the signal during or after deposition7. 6. Canfield, D. E. et al. Science 321, 949–952 (2008).
In both cases, it remains unclear exactly A note of caution comes from the analysis 7. Li, C. et al. Science 328, 80–83 (2010).
what physical (and biological) conditions are of many of our most trusted redox proxies 8. Kaufman, A. J., Hayes, J. M., Knoll, A. H. & Germs, G. J. B.
Precambrian Res. 49, 301–327 (1991).
necessary to support these redox gradients (iron speciation, sulphur isotopes and
9. Lyons, T. W., Werne, J. P., Hollander, D. J. & Murray, R. W.
in the water column, and how long such carbon/sulphur ratios) in samples from Chem. Geol. 195, 131–157 (2003).
stratified conditions are likely to remain the mobile mud belts of the Amazon River 10. Paulmier, A. & Ruiz-Pino, D. Prog. Oceanogr. 80, 113 (2009).
stable. For this we can look to modern ocean delta11. Taken at face value, the geochemical 11. Aller, R. C., Madrid, V., Chistoserdov, A., Aller, J. Y. & Heilbrun, C.
settings such as the Cariaco Basin9 off the redox indicators of these sediments suggest Geochim. Cosmochim. Acta doi:10.1016/j.gca.2010.05.008 (2010).

Structural geology

to fault or not to fault

Most of the oceanic crust has a simple layered structure. The discovery that slow-spreading ridges exhibit a
comparatively complex crustal structure and some of the largest extensional faults on Earth is leading to the
recognition of a new mode of seafloor spreading.

Michael cheadle and craig grimes

M
id-ocean ridges are the locations However, in 1983, large faults that expose in May 2010, geoscientists and biologists
where the Earth’s crust is renewed. rocks from deep below the mid-ocean ridge concluded these oceanic fault systems
Here, the tectonic plates that form were discovered at the Mid-Atlantic Ridge2. should be recognized as a fundamentally
the sea floor spread apart and new oceanic These structures have since been identified distinct mode of seafloor spreading.
crust is generated from upwelling magma. in most ocean basins. At the American Conventionally, oceanic crust is thought
Studies of ancient oceanic crust have led Geophysical Union Chapman Conference3 to have a relatively constant thickness of
to the idea that it forms in simple layers1. on ‘Detachments in oceanic lithosphere’ 6–7 km and a simple layered structure. The

454 nature geoscience | VOL 3 | JULY 2010 | www.nature.com/naturegeoscience

© 2010 Macmillan Publishers Limited. All rights reserved

 news & views

basal layer of the oceanic crust forms as Spre

adin
minerals crystallize within a deep magma g dir
ectio
n
chamber. Above this layer, melt is injected Detachment
vertically into the overlying crust, supplying fault surface
pillow lavas that are erupted onto the sea
floor, forming a third layer. As the tectonic N
plates move apart, new crust is added in
the same configuration, creating a laterally
continuous, layered structure. This model Corrugations
km
explains much of what we see in the Pacific 45 is
e ax
Ocean where new crust is formed at a dg
Ri
mid-ocean ridge, spreading at fast rates of
30 km
more than 7 cm yr−1. However, it has long
been recognized4,5 that slower-spreading
ridges, such as the Mid-Atlantic and the
Southwest Indian ridges, do not entirely fit 5,000 m 4,000 m 3,000 m 2,000 m 1,000 m
this simple model. In places, the crust does
not consist of layers of igneous material, as
Figure 1 | A three-dimensional view of the Kane segment of the Mid-Atlantic Ridge at 23° N. At the
expected. Instead, heterogeneous mixtures
Chapman conference3, detachment faults were defined as low-angle extensional faults that generate large
of mantle peridotite and crystallized
displacements as they slip. The faults uplift blocks of heterogeneous crust that is composed of mantle
magma are found on the sea floor, exposed
peridotite and crystallized magma. The dome-shaped uplifted blocks form oceanic core complexes. Here,
by movement on large detachment faults
the Kane oceanic core complex shows topography with a vertical extent exceeding 3 km. The exposed
with significant displacement. These faults
detachment-fault surface extends between the black dashed lines.
run parallel to the mid-ocean ridge and
open up in a direction perpendicular to the
ridge. Typically, the detachment faults can
accommodate 30–40 km of plate extension. 10–15% of plate separation. Detachment de Physique du Globe de Paris;
However, in exceptional cases they can faults represent a mechanism of mid-ocean G. Fruh-Green, ETH Zurich). Because of
accommodate extension on the scale of ridge extension that is distinct from the their weakness, these altered rocks may
up to 125 km, making them some of the classic mode of seafloor spreading whereby localize deformation and facilitate gradual
largest known extensional faults on Earth. extension is accommodated by the injection slip at low angles as the faults roll over
The displacement on these faults can also of magma. (A. Morris, University of Plymouth; B. John,
generate extremely high topography that The lithosphere at mid-ocean ridges University of Wyoming). Thus, water plays a
forms up to 6 km of mountainous relief. is hot and weak. As a detachment fault key role in allowing detachment faults in the
Much of the variability in the structure slips, it lifts up blocks of lithosphere, and brittle crust to slip at low angles. At depth,
of the oceanic crust, ranging from a simple mantle peridotite and crystallized magma slip is probably accommodated by ductile
layered crust to faulted and heterogeneous are accreted at the base of the blocks. The flow and indeed rocks that exhibit ductile
crust, can be explained by variations in faults are not strong enough to maintain a deformation are found on the surfaces
the rates of plate spreading and magma high angle of slip. They flatten over time, of these faults (M. Cheadle, University
supply. At fast-spreading ridges, the rate dipping at low, almost horizontal angles at of Wyoming).
of magma injection balances the rate the sea floor. The lithospheric blocks uplifted Importantly, intense hydrothermal
at which the plates separate and new by the detachment faults form relatively activity seems to be strongly correlated with
ocean crust is formed by shallow magma smooth, sometimes dome-shaped sea floor, detachment faulting. Some of the largest
injection. In contrast, at slow-spreading with small ridges or corrugations that run known hydrothermal systems, such as the
ridges, the lithosphere has a more variable parallel to the direction of plate spreading. Trans-Atlantic Geotraverse hydrothermal
thickness. Where the lithosphere is thick, At slow-spreading mid-ocean ridges, these field, are associated with oceanic core
the magma is prevented from reaching structures may form up to 60% of the sea complexes8 (A. McCaig, University of Leeds;
the Earth’s surface and the magma supply floor 7. The underlying mantle peridotite R. Sohn, Woods Hole Oceanographic
is insufficient to accommodate the full and crystallized magma are uplifted and Institute). Fluids circulating along these
amount of seafloor spreading. The classic exposed by the detachment faults, in fault systems extract and mobilize metals,
rift valley found at slow-spreading ridges topographic features that are called oceanic such as copper, lead and zinc, from the
has faults that can only account for 10–15% core complexes. surrounding rock and concentrate them
of plate separation6, the remaining 85–90% Traditional rock mechanics suggests that as ore minerals near the sea floor. These
is accommodated by magma injection. It faults in the brittle crust should form and hydrothermal systems are also uniquely
now seems that in areas of reduced magma slip at high angles, yet detachment faults are rich in hydrogen and methane, suggesting
supply, these extensional faults continue observed to slip at unusually low angles. This that these systems could host abundant and
to slip, forming faults with large offset implies that the fault is weak. Detachment unique deep-sea life. Yet the diversity of the
that can accommodate up to 100% of the faults probably originate as high-angle faults, deep sea and subsurface biosphere here has
plate separation. extending to depths of 5–8 km in the brittle only just begun to be recognized (D. Kelley,
At the Chapman conference3, oceanic crust, but are later weakened. The ingress University of Washington; A-L. Reysenbach,
detachment faults were formally defined of sea water into the detachment faults to Portland State University; M. Perner,
as low-angle extensional faults with large depths of several kilometres can lead to University of Hamburg).
displacements (Fig. 1). These faults account the formation of weak, highly altered and We do not know why a detachment fault
for significantly more than the classic metamorphosed rocks (M. Cannat, Institut might continue to slip for up to four million

nature geoscience | VOL 3 | JULY 2010 | www.nature.com/naturegeoscience 455

© 2010 Macmillan Publishers Limited. All rights reserved

news & views

years without breaking up9. Plate separation The Chapman conference3 raised as many references
could more easily be accommodated by questions as answers. Whatever the details 1. Penrose Field Conference Participants Geotimes 17, 24–25 (1972).

several smaller, successive faults, whose surrounding the origin and development 2. Karson, J. A. & Dick, H. J. B. Mar. Geophys. Res. 6, 51–98 (1983).
3. Chapman Conf. Detachments in Oceanic Lithosphere (AGU,
combined displacement equals that of the of detachment faults, they do not fit our Cyprus, 2010); http://www.ipgp.fr/chapman2010
slip on one larger fault. It seems that for textbook model of crustal formation at 4. Cannat, M. J. Geophys. Res. 98, 4163–4172 (1993).
large detachment faults to slip, a particular mid-ocean ridges. A consensus emerged 5. Dick, H. J. B., Lin, J. & Schouten, H. Nature 426, 405–412 (2003).
set of thermal conditions and volume of at the meeting that oceanic detachment 6. Escartin, J. et al. J. Geophys. Res. 104, 10421–10437 (1999).
magma input may be required10. Those faults and oceanic core complexes should be 7. Smith, D. K., Cann, J. R. & Escartin, J. Nature

areas with very low magma injection rates considered a distinct and fundamental mode 442, 440–443 (2006).
8. Escartin, J. et al. Nature 455, 790–794 (2008).
form smaller detachment faults, similar of seafloor spreading. ❐ 9. Ohara, Y., Yoshida, T., Kato, Y. & Kasuga, S. Mar. Geophys. Res.
to those observed at the slow-spreading 22, 47–61 (2001).
Gakkel Ridge in the northern Atlantic and Michael Cheadle is in the Department of Geology 10. Tucholke, B. E., Behn, M. D., Buck, W. R. & Lin, J. Geology
the far eastern Southwest Indian Ridge. and Geophysics, University of Wyoming, Laramie, 36, 455–458 (2008).
In contrast, large detachment faults are Wyoming 82071-3006, USA. Craig Grimes is in
formed at spreading ridges with moderate the Department of Geosciences, Mississippi State acknowledgements
Citations relating to the Chapman conference indicate
magma supply, such as the structures seen University, Mississippi State, Mississippi those speakers who were invited to give overview talks.
at the Mid-Atlantic Ridge (M. Cannat, 39762-5448, USA. The ideas discussed here therefore represent a summary of
Institut de Physique du Globe de Paris). e-mail: Cheadle@uwyo.edu ideas from various workers.

econoMic geology

Volatile destruction
Direct evidence for the role of volatiles in magmatic ore formation has been elusive. Magma degassing at Merapi
volcano in Indonesia is found to be directly linked to the selective leaching of metals from sulphide melts that
ultimately form ore deposits.

Bruno Scaillet

S
ulphur carried by magmas plays with an immiscible sulphide melt that the density difference and accelerating
a pivotal role in many geological hosts numerous economically interesting the settling out of sulphides. Yet, hydrous
processes. A great chemical affinity elements, such as copper, nickel, gold and magmas are not observed to produce layers
exists between sulphur and metals that is the platinum-group elements5. According to of accumulated sulphides. Instead, field
important for ore formation. Thus many ore the classic mechanism of magmatic sulphide evidence shows that sulphur-bearing rocks
deposits of economic interest are intimately ore formation6, sulphide melts readily settle related to arc magmas accumulate above
associated with magmatic activity. The out of the main molten body of magma the main silicate–magma intrusion, and not
sulphur and metals merge together at the because they are much denser than the in it, forming the so-called hydrothermal
magmatic stage to yield immiscible sulphide surrounding silicate melt. They accumulate ore deposits. One of the reasons for this
melts. At some stage during magma at the base of the magma chamber or observation lies in the degassing processes
evolution, the ore-forming elements are intrusion (Fig. 1a). This mechanism works at work in such settings, in particular
removed or concentrated, but the reasons well for dry iron- or magnesium-rich (mafic) the loss of water by the magma. During
for this, particularly in subduction-zone magma, which is typical for magmatic their ascent to the surface, magmas lose
settings, remain obscure. The volatile activity at intraplate or extensional rift- their volatiles continuously because water
content of the magma has long been related settings. These tectonic settings are solubility in silicate melts decreases with
thought to play an important role in the associated with many metallic deposits of decreasing pressure1. The escaping gas rises
separation process1, but direct evidence has huge economic interest, such as at Norilsk towards the upper reaches of the magma
been lacking. Writing in Nature Geoscience, in Siberia7. There, some 250 million years plumbing system. As the volatiles percolate
Nadeau et al.2 report observations from ago profuse intraplate volcanism created the through the magma they scavenge elements
Merapi volcano, Indonesia, that indicate largest nickle–copper–palladium deposits in and essentially destroy the sulphide melt
that present-day arc magmas, produced the world. (Fig. 1b). The volatiles thus provide a
during the subduction of water-rich The abundance of water in arc magmas, mechanism for the removal and transport
oceanic crust, owe their ore potential to the however, adds an intriguing complexity of metal elements, depositing them either
abundance of volatiles associated with this to this scenario. As far as we know, water above the main magma intrusion or
tectonic setting. does not significantly affect the solubility of carrying them into the atmosphere as a
After water and carbon dioxide, sulphur sulphur in silicate melts8,9, at low pressures gaseous emission.
is the third most abundant dissolved volatile at least. Therefore, hydrous magmas, such as Nadeau et al.2 present detailed
in magmas. Concentrations of sulphur range those formed in arc settings, should follow petrological analyses of rocks ejected during
from a few hundred to a few thousand parts the same pattern of sulphide removal as that an explosive event at Merapi volcano in
per million3,4. When placed in the upper of dry magmas. In fact, water should lower 2006. They provide a comparison between
crust, the majority of magmas soon saturate the density of silicate liquids, amplifying some preserved sulphide melt, the main

456 nature geoscience | VOL 3 | JULY 2010 | www.nature.com/naturegeoscience

© 2010 Macmillan Publishers Limited. All rights reserved