Magma generation at mid-ocean ridges
Overview
Separation of tectonic plates causes compositionally heterogeneous mantle beneath mid-ocean ridges (e.g.,
Allègre et al., 1984) to rise and decompress, partially melting the upper mantle (e.g., Klein and Langmuir, 1987;
Kelemen et al., 1997) and producing basaltic parent melts
(e.g., Elthon, 1979). Magmatic compositions are modified
by multiple processes during ascent and storage prior to
eruption (e.g., Klein and Langmuir, 1987; Sinton and
Detrick, 1992; Rubin and Sinton, 2007). Mid-ocean ridge
magma compositions therefore reflect a combination of
magmatic processes related to (a) the composition and
temperature of mantle beneath the ridge, (b) the volcanic
plumbing system within ridge volcanoes, and (c) cooling
from the infiltration of overlying seawater along faults
and fissures in the upper crust leading to magma differentiation (e.g., Macdonald, 1982; Phipps Morgan and Chen,
1994). Variations in the rate and style of magma production, volcanism, and faulting result from upper mantle
temperature and composition variations (e.g., Dalton
et al., 2014) and a change in seafloor spreading rate from
<1 cm/year to 20 cm/year through the ridge system, causing differences in igneous ocean crustal thickness (e.g.,
Bown and White, 1994), structure and morphology (e.g.,
Small, 1998), and composition (e.g., Rubin and Sinton,
2007).
Away from hotspots, magma supply to ridges from
mantle melting covaries positively with spreading rate,
so that high spreading rate generally equates to high
magma supply. Superimposed on this plate-spreadingdriven magma supply, differences in mantle composition,
melting, and melt transport further modulate melt supply
to the ridge axis along its length (e.g., Langmuir et al.,
1986; Sinton et al., 1991), in turn producing physical
and chemical segmentation of the mid-ocean ridge at
10–100 km length scales. Upward migration of MORB
melts through the mantle, followed by accumulation in
crustal magma bodies called magma chambers, results in
magma differentiation and crystallization (e.g., Klein and
Langmuir, 1987; Grove et al., 1992; Herzberg, 2004)
and perhaps to melt-rock reactions (e.g., Lissenberg and
Dick, 2008; Lissenberg et al., 2013) prior to either eruption or emplacement as a pluton within the oceanic crust.
Structural and thermal conditions within the mantle and
crust produce magma supply and magmatic differentiation
variations at even smaller spatial scales (i.e., 1–10 km
along the ridge axis), producing spatial patterns in erupted
magma compositions akin to those found within and
between neighboring terrestrial volcanoes (e.g., Reynolds
et al., 1992; Perfit et al., 1994; Perfit and Chadwick,
1998). Studies of individual eruption deposits (e.g., Rubin
et al., 2001; Bergmanis et al., 2007; Goss et al., 2010; Colman et al., 2012) and dikes that feed eruptions (e.g., Pollock et al., 2009) reveal spatial geochemical patterns that
record variations in the delivery of individual magma
batches to, and magma mixing and differentiation
conditions within, subaxial magma reservoirs from which
they erupt.
Magma and mantle source compositions
Like all mafic magma generation on Earth, the major element composition of MORB magmas is largely set by the
composition of the parent rock in the mantle source, the
melting depth, and the degree to which the source rock is
partially melted. MORB major element compositions vary
less than basalts erupted on land, in part because the aforementioned melting parameters and source compositions
are comparatively homogeneous beneath ridges and in
part because the thinner and compositionally more uniformly oceanic crust imparts significantly less chemical
variation to MORB magmas during magma accumulation
and differentiation before eruption. Still, MORB display
geographical variations in major element, trace element,
and radiogenic isotopic compositions that reflect regional
differences in mantle composition and melting conditions
around the globe (e.g., Dupre and Allègre, 1980; Hamelin
and Allègre, 1985; Klein and Langmuir, 1987; Mahoney
et al., 1989; Asimow et al., 2004; Hanan et al., 2004;
Niu and O’Hara, 2008), as well as the influence of nearridge hotspots (e.g., Schilling, 1991; Cushman et al.,
2004).
Globally, the trace constituents in MORB (trace elements and radiogenic isotopes) vary more significantly
than the major elements. Petrologists use these constituents to understand the mineralogy, chemical composition,
and history of a magma’s mantle source rock, as well as to
help track preeruptive magma differentiation and magmahost rock reactions in the crust. MORB have been widely
used to identify variations in upper mantle composition
although such mantle signatures in MORB still need to
be distinguished from variations imparted to the magmas
by melting, magma transport, melt-rock reactions, mixing,
and magma differentiation (e.g., see review by Rubin
et al., 2009). Of particular importance are the degree to
which different types of mantle rocks (i.e., lithologies),
also known as “mantle heterogeneities,” melt at different
rates and proportions (e.g., Ito and Mahoney, 2005;
Stracke et al., 2006; Russo et al, 2009) and the extent of
subsequent magma mixing and homogenization (Rubin
and Sinton, 2007). The latter works to partially or fully
obscure inherent mantle source rock variability in aggregated melts of the mantle, becoming more pronounced
as magma supply increases.
Magma chamber processes
Nearly all MORB erupt bearing the chemical and mineralogical signatures of magma differentiation (e.g., olivine
and plagioclase crystallization) and partial loss of initial
melt magmatic gas loads, essentially requiring that these
magmas accumulated and partially crystallized in magma
bodies at one or more locations in the crust before eruption. There is considerable debate as to the nature of these
magma bodies, where they reside, and how they impart
MID-OCEAN RIDGE MAGMATISM AND VOLCANISM
503
Overview
Separation of tectonic plates causes compositionally heterogeneous mantle beneath mid-ocean ridges (e.g.,
Allègre et al., 1984) to rise and decompress, partially melting the upper mantle (e.g., Klein and Langmuir, 1987;
Kelemen et al., 1997) and producing basaltic parent melts
(e.g., Elthon, 1979). Magmatic compositions are modified
by multiple processes during ascent and storage prior to
eruption (e.g., Klein and Langmuir, 1987; Sinton and
Detrick, 1992; Rubin and Sinton, 2007). Mid-ocean ridge
magma compositions therefore reflect a combination of
magmatic processes related to (a) the composition and
temperature of mantle beneath the ridge, (b) the volcanic
plumbing system within ridge volcanoes, and (c) cooling
from the infiltration of overlying seawater along faults
and fissures in the upper crust leading to magma differentiation (e.g., Macdonald, 1982; Phipps Morgan and Chen,
1994). Variations in the rate and style of magma production, volcanism, and faulting result from upper mantle
temperature and composition variations (e.g., Dalton
et al., 2014) and a change in seafloor spreading rate from
<1 cm/year to 20 cm/year through the ridge system, causing differences in igneous ocean crustal thickness (e.g.,
Bown and White, 1994), structure and morphology (e.g.,
Small, 1998), and composition (e.g., Rubin and Sinton,
2007).
Away from hotspots, magma supply to ridges from
mantle melting covaries positively with spreading rate,
so that high spreading rate generally equates to high
magma supply. Superimposed on this plate-spreadingdriven magma supply, differences in mantle composition,
melting, and melt transport further modulate melt supply
to the ridge axis along its length (e.g., Langmuir et al.,
1986; Sinton et al., 1991), in turn producing physical
and chemical segmentation of the mid-ocean ridge at
10–100 km length scales. Upward migration of MORB
melts through the mantle, followed by accumulation in
crustal magma bodies called magma chambers, results in
magma differentiation and crystallization (e.g., Klein and
Langmuir, 1987; Grove et al., 1992; Herzberg, 2004)
and perhaps to melt-rock reactions (e.g., Lissenberg and
Dick, 2008; Lissenberg et al., 2013) prior to either eruption or emplacement as a pluton within the oceanic crust.
Structural and thermal conditions within the mantle and
crust produce magma supply and magmatic differentiation
variations at even smaller spatial scales (i.e., 1–10 km
along the ridge axis), producing spatial patterns in erupted
magma compositions akin to those found within and
between neighboring terrestrial volcanoes (e.g., Reynolds
et al., 1992; Perfit et al., 1994; Perfit and Chadwick,
1998). Studies of individual eruption deposits (e.g., Rubin
et al., 2001; Bergmanis et al., 2007; Goss et al., 2010; Colman et al., 2012) and dikes that feed eruptions (e.g., Pollock et al., 2009) reveal spatial geochemical patterns that
record variations in the delivery of individual magma
batches to, and magma mixing and differentiation
conditions within, subaxial magma reservoirs from which
they erupt.
Magma and mantle source compositions
Like all mafic magma generation on Earth, the major element composition of MORB magmas is largely set by the
composition of the parent rock in the mantle source, the
melting depth, and the degree to which the source rock is
partially melted. MORB major element compositions vary
less than basalts erupted on land, in part because the aforementioned melting parameters and source compositions
are comparatively homogeneous beneath ridges and in
part because the thinner and compositionally more uniformly oceanic crust imparts significantly less chemical
variation to MORB magmas during magma accumulation
and differentiation before eruption. Still, MORB display
geographical variations in major element, trace element,
and radiogenic isotopic compositions that reflect regional
differences in mantle composition and melting conditions
around the globe (e.g., Dupre and Allègre, 1980; Hamelin
and Allègre, 1985; Klein and Langmuir, 1987; Mahoney
et al., 1989; Asimow et al., 2004; Hanan et al., 2004;
Niu and O’Hara, 2008), as well as the influence of nearridge hotspots (e.g., Schilling, 1991; Cushman et al.,
2004).
Globally, the trace constituents in MORB (trace elements and radiogenic isotopes) vary more significantly
than the major elements. Petrologists use these constituents to understand the mineralogy, chemical composition,
and history of a magma’s mantle source rock, as well as to
help track preeruptive magma differentiation and magmahost rock reactions in the crust. MORB have been widely
used to identify variations in upper mantle composition
although such mantle signatures in MORB still need to
be distinguished from variations imparted to the magmas
by melting, magma transport, melt-rock reactions, mixing,
and magma differentiation (e.g., see review by Rubin
et al., 2009). Of particular importance are the degree to
which different types of mantle rocks (i.e., lithologies),
also known as “mantle heterogeneities,” melt at different
rates and proportions (e.g., Ito and Mahoney, 2005;
Stracke et al., 2006; Russo et al, 2009) and the extent of
subsequent magma mixing and homogenization (Rubin
and Sinton, 2007). The latter works to partially or fully
obscure inherent mantle source rock variability in aggregated melts of the mantle, becoming more pronounced
as magma supply increases.
Magma chamber processes
Nearly all MORB erupt bearing the chemical and mineralogical signatures of magma differentiation (e.g., olivine
and plagioclase crystallization) and partial loss of initial
melt magmatic gas loads, essentially requiring that these
magmas accumulated and partially crystallized in magma
bodies at one or more locations in the crust before eruption. There is considerable debate as to the nature of these
magma bodies, where they reside, and how they impart
MID-OCEAN RIDGE MAGMATISM AND VOLCANISM
503