best-documented examples are Hawaii and the Galapagos.
Tatsumoto (1978) and Stille et al. (1986) showed that the
Hawaiian hot spot consists of two volcanic chains: Loa
and Kea, whose Pb isotopic compositions are distinct
and have remained so for about 5 Ma. Subsequent studies
(e.g., Abouchami et al., 2005; Bryce et al., 2005; Weis
et al., 2011) fully support that early work and show the difference extends to other isotope ratios as well. Parallel
chains having distinct radiogenic isotopic characteristics
have also been identified in the Marquesas, the Society
Islands, Samoa, and the Tristan da Cunha–Gough hot spot
(Desonie et al., 1993; Huang et al., 2011; Chauvel et al.,
2012; Payne et al., 2013; Rohde et al., 2013). In the
Galapagos, four distinct plume components, northern,
central, southern, and a depleted eastern one, can be distinguished in the mixing trends (White et al., 1993;
Harpp and White, 2001). Hoernle et al. (2000) showed
that these same domains occur in the same relative positions in a geochemical profile across the Cocos Ridge
(the Galapagos hot spot track on the Cocos Plate) indicating that the spatial zonation of the Galapagos hot spot has
continued for at least 14 Ma. As they point out, this
implies “that plume material can ascend from the lower
mantle, possibly from the core-mantle boundary, with little stirring occurring during ascent, and that zonation in
hotspot lavas may in some cases reflect spatial heterogeneity within the lower mantle source.” These zonations
may in turn reflect chemical zonation and structure in the
lowermost mantle (Weis et al., 2011; Farnetani et al.,
2012; Harpp et al., 2014).
In summary, plumes carry a variety of material from the
deep mantle to the surface, including anciently subducted
oceanic and continental crust and very primitive noble gas
signal preserved from the earliest days of Earth’s history.
Lithological composition of plumes
The lithology of plumes, which is their mineralogical
makeup, is more difficult to assess because it cannot be
directly observed, only inferred. Without question, the
mantle consists predominately of peridotite. However,
populations of mantle xenoliths always contained a small
but significant fraction of rocks, such as eclogite and
pyroxenite, similar in composition to basalt. Allègre and
Turcotte (1986) pointed out that continuous production
and subduction of oceanic crust over billions of years
was likely to produce a “marble-cake” mantle with ubiquitous stringers of mafic material embedded in an ultramafic
matrix. Indeed, recent numerical simulations suggest 97 %
of the mantle has been processed through the oceanic crust
cycle (Huang and Davies, 2007). If, as isotopic and trace
element studies suggest, plumes are particularly rich in
recycled crustal materials, one might expect them to be
lithologically distinct from asthenospheric mantle and particularly rich in mafic materials such as eclogite and
pyroxenite. Since this material melts at lower temperature
than peridotite, might the mantle plume melts be derived
predominantly from pods and stringers of mafic material
rather than the more refractory peridotite in which they
are embedded?
This question remains unresolved and debated. Stracke
et al. (1999) argued against the existence of garnetpyroxenite or eclogite in the Hawaiian source on the basis
of combined Hf-Nd-Th isotope and trace element systematics. On the other hand, Reiners (2002) argued that
“temporal-compositional trends in the form of decreasing
incompatible elements and MgO, with increasing SiO 2 , as
the eruption proceed”. . . “reflect sequential eruption of
melt from increasing depths in mantle melting regions that
are compositionally zoned because of different solidi and
melt productivities of distinct peridotite and pyroxenite
lithologic domains.”
A particularly interesting, provocative, and consequently controversial idea was proposed by Sobolev
et al. (2005), who argue on the basis of SiO 2 , Ni, and
Mn concentrations that part of the Hawaiian source is
olivine-free. They suggested that eclogite, which originates as recycled oceanic crust, composes up to 30 % of
the Hawaiian plume. As a consequence of its much lower
solidus, the eclogite melts at greater depth than peridotite.
This melt then reacts with surrounding peridotite to form
pyroxenite, which subsequently melts at shallower depth.
Sobolev et al. (2007) argue that recycled oceanic crust in
the form of eclogite stringers, and consequently this
two-stage melting process, is pervasive in the mantle;
according to them even MORB are composed of
10–30 % pyroxenite-derived melts. A number of other
studies have also concluded that pyroxenites play an
important role in the genesis of Hawaiian magmas. Putirka
et al. (2011), however, argue that high Ni contents and
high FeO/MnO ratios in Hawaiian basalts compared to
MORB reflect higher temperatures (by 300
) and pressures (4–5 GPa vs. 1–3 GPa) of melting under Hawaii
than beneath mid-ocean ridges and that both Hawaiian
basalts and MORB are generated by melting of peridotite
rather than pyroxenite or eclogite. They suggest that
subducted material is thoroughly mixed back into the peridotite matrix and does not survive “high temperatures and
potentially long residence times” in the deep mantle as a
distinct lithological component.
Evidence that plumes are hot and buoyant
If plumes are hotter than the depleted upper mantle, this
should produce petrologic and chemical distinctions. To
begin with, one would expect them to melt more. Because
melting in both plumes and at mid-ocean ridges results
from adiabatic decompression of rising mantle, the extent
of melting depends both on temperature, which determines the depth at which melting begins, and lithospheric
thickness, which determines the height to which mantle
can rise, and consequently the total amount of melting
for mantle of a given temperature and composition. At
mid-ocean ridges, the lithosphere is thin to nonexistent
and melting continues to shallow depth (<30 km). Older
lithosphere, for example, beneath Hawaii, may approach
322
HOT SPOTS AND MANTLE PLUMES
Tatsumoto (1978) and Stille et al. (1986) showed that the
Hawaiian hot spot consists of two volcanic chains: Loa
and Kea, whose Pb isotopic compositions are distinct
and have remained so for about 5 Ma. Subsequent studies
(e.g., Abouchami et al., 2005; Bryce et al., 2005; Weis
et al., 2011) fully support that early work and show the difference extends to other isotope ratios as well. Parallel
chains having distinct radiogenic isotopic characteristics
have also been identified in the Marquesas, the Society
Islands, Samoa, and the Tristan da Cunha–Gough hot spot
(Desonie et al., 1993; Huang et al., 2011; Chauvel et al.,
2012; Payne et al., 2013; Rohde et al., 2013). In the
Galapagos, four distinct plume components, northern,
central, southern, and a depleted eastern one, can be distinguished in the mixing trends (White et al., 1993;
Harpp and White, 2001). Hoernle et al. (2000) showed
that these same domains occur in the same relative positions in a geochemical profile across the Cocos Ridge
(the Galapagos hot spot track on the Cocos Plate) indicating that the spatial zonation of the Galapagos hot spot has
continued for at least 14 Ma. As they point out, this
implies “that plume material can ascend from the lower
mantle, possibly from the core-mantle boundary, with little stirring occurring during ascent, and that zonation in
hotspot lavas may in some cases reflect spatial heterogeneity within the lower mantle source.” These zonations
may in turn reflect chemical zonation and structure in the
lowermost mantle (Weis et al., 2011; Farnetani et al.,
2012; Harpp et al., 2014).
In summary, plumes carry a variety of material from the
deep mantle to the surface, including anciently subducted
oceanic and continental crust and very primitive noble gas
signal preserved from the earliest days of Earth’s history.
Lithological composition of plumes
The lithology of plumes, which is their mineralogical
makeup, is more difficult to assess because it cannot be
directly observed, only inferred. Without question, the
mantle consists predominately of peridotite. However,
populations of mantle xenoliths always contained a small
but significant fraction of rocks, such as eclogite and
pyroxenite, similar in composition to basalt. Allègre and
Turcotte (1986) pointed out that continuous production
and subduction of oceanic crust over billions of years
was likely to produce a “marble-cake” mantle with ubiquitous stringers of mafic material embedded in an ultramafic
matrix. Indeed, recent numerical simulations suggest 97 %
of the mantle has been processed through the oceanic crust
cycle (Huang and Davies, 2007). If, as isotopic and trace
element studies suggest, plumes are particularly rich in
recycled crustal materials, one might expect them to be
lithologically distinct from asthenospheric mantle and particularly rich in mafic materials such as eclogite and
pyroxenite. Since this material melts at lower temperature
than peridotite, might the mantle plume melts be derived
predominantly from pods and stringers of mafic material
rather than the more refractory peridotite in which they
are embedded?
This question remains unresolved and debated. Stracke
et al. (1999) argued against the existence of garnetpyroxenite or eclogite in the Hawaiian source on the basis
of combined Hf-Nd-Th isotope and trace element systematics. On the other hand, Reiners (2002) argued that
“temporal-compositional trends in the form of decreasing
incompatible elements and MgO, with increasing SiO 2 , as
the eruption proceed”. . . “reflect sequential eruption of
melt from increasing depths in mantle melting regions that
are compositionally zoned because of different solidi and
melt productivities of distinct peridotite and pyroxenite
lithologic domains.”
A particularly interesting, provocative, and consequently controversial idea was proposed by Sobolev
et al. (2005), who argue on the basis of SiO 2 , Ni, and
Mn concentrations that part of the Hawaiian source is
olivine-free. They suggested that eclogite, which originates as recycled oceanic crust, composes up to 30 % of
the Hawaiian plume. As a consequence of its much lower
solidus, the eclogite melts at greater depth than peridotite.
This melt then reacts with surrounding peridotite to form
pyroxenite, which subsequently melts at shallower depth.
Sobolev et al. (2007) argue that recycled oceanic crust in
the form of eclogite stringers, and consequently this
two-stage melting process, is pervasive in the mantle;
according to them even MORB are composed of
10–30 % pyroxenite-derived melts. A number of other
studies have also concluded that pyroxenites play an
important role in the genesis of Hawaiian magmas. Putirka
et al. (2011), however, argue that high Ni contents and
high FeO/MnO ratios in Hawaiian basalts compared to
MORB reflect higher temperatures (by 300
) and pressures (4–5 GPa vs. 1–3 GPa) of melting under Hawaii
than beneath mid-ocean ridges and that both Hawaiian
basalts and MORB are generated by melting of peridotite
rather than pyroxenite or eclogite. They suggest that
subducted material is thoroughly mixed back into the peridotite matrix and does not survive “high temperatures and
potentially long residence times” in the deep mantle as a
distinct lithological component.
Evidence that plumes are hot and buoyant
If plumes are hotter than the depleted upper mantle, this
should produce petrologic and chemical distinctions. To
begin with, one would expect them to melt more. Because
melting in both plumes and at mid-ocean ridges results
from adiabatic decompression of rising mantle, the extent
of melting depends both on temperature, which determines the depth at which melting begins, and lithospheric
thickness, which determines the height to which mantle
can rise, and consequently the total amount of melting
for mantle of a given temperature and composition. At
mid-ocean ridges, the lithosphere is thin to nonexistent
and melting continues to shallow depth (<30 km). Older
lithosphere, for example, beneath Hawaii, may approach
322
HOT SPOTS AND MANTLE PLUMES