CRUSTAL ACCRETION
Benoît Ildefonse
Geosciences Montpellier, CNRS and University of
Montpellier, Montpellier, France
Synonyms
Formation of ocean crust
Definition
Crustal Accretion. Formation of ocean crust at mid-ocean
ridges, by a combination of magmatic, tectonic, and
hydrothermal processes.
Introduction
Since the early 1970s, the classical model of a uniformly
layered ocean crust, with from top to bottom basaltic lava
flows, basaltic sheeted dikes, and gabbros, known as the
“Penrose” model (Anonymous, 1972), has considerably
evolved, with complementary contributions from
ophiolite studies, seafloor geology, marine geophysics,
and scientific ocean drilling. At the global scale, there is
substantial variability of the mid-ocean ridge morphology
and hence of crustal architecture. This results from various
modes of accretion that are controlled by magma supply to
the ridge, which itself primarily depends on the spreading
rate. This article briefly describes the current end-member
models for crustal accretion at fast-spreading
(>~80 mm/year) and slow-spreading (<~40 mm/year)
ridges.
Fast-spreading ridges: a nearly continuous
magma chamber
At fast-spreading (and part of intermediate-spreading)
ridges, the ocean crust is assumed to be relatively simple
and continuously layered, in accordance with the “Penrose” model, as a result from the presence of nearly continuous and steady-state axial magma chamber, along the
axis. Most of the first models for mid-ocean ridge magma
chambers, in the 1970s and early 1980s, proposed rather
large magma bodies, ~10–20 km large across the ridge
axis and ~4 km thick at the axis, lying below the sheeted
dike complex (e.g., Cann, 1974; Pallister and Hopson,
1981). Using thermal constraints, Sleep (1975) proposed
that most of the chamber was filled with magmatic mush
and that melt-rich magma would be limited to the top,
below the sheeted dikes. This prediction was confirmed
about 10 years later, with the first seismic reflection
images of an axial magma chamber beneath the Valu Fa
Ridge in the Lau Basin (Morton and Sleep, 1985) and
beneath the East Pacific Rise (e.g., Hale et al., 1982;
Detrick et al., 1987; Kent et al., 1994). This shallow and
thin magma lens then became a key element of all models
for accretion at fast-spreading ridges. Details of the relationships between the morphologic segmentation of the
ridge, the along-axis variability (depth and melt content)
of the magma lens, volcanic activity, lava chemistry, and
hydrothermal cooling in the upper crust are discussed in
many studies published over the last three decades (e.g.,
Macdonald et al., 1988; Hooft et al., 1997; Singh et al,
1998; Goss et al., 2010; Carbotte et al., 2013; Xu et al.,
2014). In intact fast-spread crust, the deepest stratigraphic
level reached by drilling is the transition zone between the
sheeted dike complex and the gabbros of the lower crust,
in the ODP (Ocean Drilling Program) Hole 1256D on
the Cocos Plate (Wilson et al., 2006; Teagle et al., 2012;
Ildefonse et al., 2014). This zone, also described in
ophiolites (e.g., Gillis, 2008; France et al., 2009), marks
the thermal boundary layer between the magmatic and
hydrothermal systems at the ridge axis. It displays a complex petrological record of the interplay between temporally and spatially intercalated magmatic, hydrothermal,
partially melting, and metamorphic processes.
Below the axial magma lens is a seismically attenuated
domain, which is believed to correspond to a melt-poor
magmatic mush (e.g., Sinton and Detrick, 1992; Dunn
et al., 2000). Two competing models are proposed for
building the lower, gabbroic crust in this domain. The
“gabbro glacier” models (e.g., Henstock et al., 1993;
Phipps Morgan and Chen, 1993; Figure 1a) postulate that
the entire lower crust is formed by the flow of mushy
material downward and outward from the single, shallow
axial magma lens. In contrast, “sheeted sill models,”
largely based on observations made in the Oman ophiolite
(e.g., Kelemen et al., 1997; Figure 1b), involve in situ formation of the lower crust by sill intrusions. Boudier
et al. (1996) proposed a combination of these two
end-members. Deep, lower crustal seismic reflectors,
interpreted as magma sills accreting the lower igneous
crust, have now been imaged below the axial magma lens
at the Juan de Fuca Ridge (Canales et al., 2009) and at the
East Pacific Rise (Marjanović et al., 2014). Accretion
models can be used for predicting contrasted vertical
trends of chemical composition, deformation, and alteration in the gabbroic crust, which still remain to be fully
tested in ophiolites and by deep drilling in intact fastspread crust (e.g., Teagle et al., 2012; Ildefonse et al.,
2014). While the gabbro glacier model does not require
deep hydrothermal circulation close to the axis, the latter
is needed to sustain a sheeted sill model (Figure 1b; e.g.,
Maclennan et al., 2005). The IODP (Integrated Ocean
Drilling Program) Expedition 345 recovered, for the first
time, significant sections of layered gabbros from
disrupted lower crust exposed at Hess Deep in the equatorial Eastern Pacific (Gillis et al., 2014), where young
(~1 Ma) lower fast-spread crust is tectonically exposed
at the tip of the Cocos-Nazca rift.
Slow-spreading ridges: a variable, composite crust
In contrast to fast-spread crust, ocean crust created at
slow- and ultraslow-spreading ridges is spatially heterogeneous, both along and across the ridge axis. Along parts of
slow-spreading ridges (e.g., the centers of ridge segments
CRUSTAL ACCRETION
133
Benoît Ildefonse
Geosciences Montpellier, CNRS and University of
Montpellier, Montpellier, France
Synonyms
Formation of ocean crust
Definition
Crustal Accretion. Formation of ocean crust at mid-ocean
ridges, by a combination of magmatic, tectonic, and
hydrothermal processes.
Introduction
Since the early 1970s, the classical model of a uniformly
layered ocean crust, with from top to bottom basaltic lava
flows, basaltic sheeted dikes, and gabbros, known as the
“Penrose” model (Anonymous, 1972), has considerably
evolved, with complementary contributions from
ophiolite studies, seafloor geology, marine geophysics,
and scientific ocean drilling. At the global scale, there is
substantial variability of the mid-ocean ridge morphology
and hence of crustal architecture. This results from various
modes of accretion that are controlled by magma supply to
the ridge, which itself primarily depends on the spreading
rate. This article briefly describes the current end-member
models for crustal accretion at fast-spreading
(>~80 mm/year) and slow-spreading (<~40 mm/year)
ridges.
Fast-spreading ridges: a nearly continuous
magma chamber
At fast-spreading (and part of intermediate-spreading)
ridges, the ocean crust is assumed to be relatively simple
and continuously layered, in accordance with the “Penrose” model, as a result from the presence of nearly continuous and steady-state axial magma chamber, along the
axis. Most of the first models for mid-ocean ridge magma
chambers, in the 1970s and early 1980s, proposed rather
large magma bodies, ~10–20 km large across the ridge
axis and ~4 km thick at the axis, lying below the sheeted
dike complex (e.g., Cann, 1974; Pallister and Hopson,
1981). Using thermal constraints, Sleep (1975) proposed
that most of the chamber was filled with magmatic mush
and that melt-rich magma would be limited to the top,
below the sheeted dikes. This prediction was confirmed
about 10 years later, with the first seismic reflection
images of an axial magma chamber beneath the Valu Fa
Ridge in the Lau Basin (Morton and Sleep, 1985) and
beneath the East Pacific Rise (e.g., Hale et al., 1982;
Detrick et al., 1987; Kent et al., 1994). This shallow and
thin magma lens then became a key element of all models
for accretion at fast-spreading ridges. Details of the relationships between the morphologic segmentation of the
ridge, the along-axis variability (depth and melt content)
of the magma lens, volcanic activity, lava chemistry, and
hydrothermal cooling in the upper crust are discussed in
many studies published over the last three decades (e.g.,
Macdonald et al., 1988; Hooft et al., 1997; Singh et al,
1998; Goss et al., 2010; Carbotte et al., 2013; Xu et al.,
2014). In intact fast-spread crust, the deepest stratigraphic
level reached by drilling is the transition zone between the
sheeted dike complex and the gabbros of the lower crust,
in the ODP (Ocean Drilling Program) Hole 1256D on
the Cocos Plate (Wilson et al., 2006; Teagle et al., 2012;
Ildefonse et al., 2014). This zone, also described in
ophiolites (e.g., Gillis, 2008; France et al., 2009), marks
the thermal boundary layer between the magmatic and
hydrothermal systems at the ridge axis. It displays a complex petrological record of the interplay between temporally and spatially intercalated magmatic, hydrothermal,
partially melting, and metamorphic processes.
Below the axial magma lens is a seismically attenuated
domain, which is believed to correspond to a melt-poor
magmatic mush (e.g., Sinton and Detrick, 1992; Dunn
et al., 2000). Two competing models are proposed for
building the lower, gabbroic crust in this domain. The
“gabbro glacier” models (e.g., Henstock et al., 1993;
Phipps Morgan and Chen, 1993; Figure 1a) postulate that
the entire lower crust is formed by the flow of mushy
material downward and outward from the single, shallow
axial magma lens. In contrast, “sheeted sill models,”
largely based on observations made in the Oman ophiolite
(e.g., Kelemen et al., 1997; Figure 1b), involve in situ formation of the lower crust by sill intrusions. Boudier
et al. (1996) proposed a combination of these two
end-members. Deep, lower crustal seismic reflectors,
interpreted as magma sills accreting the lower igneous
crust, have now been imaged below the axial magma lens
at the Juan de Fuca Ridge (Canales et al., 2009) and at the
East Pacific Rise (Marjanović et al., 2014). Accretion
models can be used for predicting contrasted vertical
trends of chemical composition, deformation, and alteration in the gabbroic crust, which still remain to be fully
tested in ophiolites and by deep drilling in intact fastspread crust (e.g., Teagle et al., 2012; Ildefonse et al.,
2014). While the gabbro glacier model does not require
deep hydrothermal circulation close to the axis, the latter
is needed to sustain a sheeted sill model (Figure 1b; e.g.,
Maclennan et al., 2005). The IODP (Integrated Ocean
Drilling Program) Expedition 345 recovered, for the first
time, significant sections of layered gabbros from
disrupted lower crust exposed at Hess Deep in the equatorial Eastern Pacific (Gillis et al., 2014), where young
(~1 Ma) lower fast-spread crust is tectonically exposed
at the tip of the Cocos-Nazca rift.
Slow-spreading ridges: a variable, composite crust
In contrast to fast-spread crust, ocean crust created at
slow- and ultraslow-spreading ridges is spatially heterogeneous, both along and across the ridge axis. Along parts of
slow-spreading ridges (e.g., the centers of ridge segments
CRUSTAL ACCRETION
133