Geoscience Reference
In-Depth Information
FeO, would decrease the olivine content of the
mantle.
Earth models based on cosmic abundances,
with Mg/Si approximately 1 (molar), give rela-
tively low total olivine contents. Since the upper
mantle appears to be olivine-rich, this results
in an even more olivine-poor transition region
and lower mantle. It is usually assumed, how-
ever, that the basaltic fraction of the Earth is still
mostly dispersed throughout the mantle. This is
the assumption behind the pyrolite model and
most geochemical models of the mantle. Basalts
were probably liberated during accretion of the
Earth and concentrated in the upper mantle.
Efficient remixing or rehomogenization can be
ruled out.
Samples of the Earth's upper mantle are dis-
tinct in composition from that of any kind of
extant meteorite. A hypothetical so-called prim-
itive upper mantle (PUM) -- a mixture of conti-
nental crust and ultramafic upper mantle rocks --
also differs from meteorites in significant ways.
Pyrolite -- a hypothetical rock constructed from
basalts and peridotites from the upper mantle --
also does not have the same ratios of major ele-
ments as do meteorites. Several suggestions have
been made to explain the elevated Mg/Si and
Al/Si ratios in the Earth's shallowest mantle rela-
tive to undifferentiated meteorites. These include
sequestering Si in the core of the Earth, raising
the Mg/Si and Al/Si ratios in the silicate mantle,
or appealing to the possibility that the mantle
is chemically stratified with the lower mantle --
which is the bulk of the Earth -- having a dif-
ferent composition from the upper mantle. The
latter is the most plausible suggestion.
There is no particular reason why the Earth
should have accreted olivine in preference to
pyroxene. There is no compelling evidence that
Si is or can be extracted into the core and the
idea is also inconsistent with the upper man-
tle abundances of V, Cr and Mn; they should be
depleted far below observed abundances if con-
ditions were sufficiently reducing to allow Si to
dissolve in molten iron.
Estimates of the Earth's upper mantle com-
position are distinct from any known type of
meteorite, but there are numerous petrological
reasons why the Earth's mantle should not have
Table 3.8
Mineralogy of Earth's mantle
Whole--Mantle Models
Upper
Mantle
Species
(1)
(2)
(3)
(4)
Olivine
47.2
36.5
37.8
51.4
Orthopyroxene
28.3
33.7
33.2
25.6
Clinopyroxene
12.7
14.6
11.8
11.0
Jadeite
9.8
2.2
1.8
0.65
Ilmenite
0.2
0.5
0.24
0.57
Garnet
1.53
11.6
14.2
9.6
Chromite
0.0
1.6
0.94
0.44
(1) Equilibrium condensation (BVP, 1980).
(2) Cosmochemical model (Ganapathy and
Anders 1974).
(3)
Cosmochemical
model
(Morgan
and
Anders, 1980).
(4) Pyrolite (Ringwood, 1977).
Table 3.9
Simple Earth model based on cos-
mic abundances
Molecular
Weight
Oxides Molecules
Weight
Grams Fraction
MgO
1.06
40
42.4
0.250
SiO
2
1.00
60
60.0
0.354
A1
2
O
3
0.0425
102
4.35
0.026
CaO
0.0625
56
3.5
0.021
Na
2
O
0.03
62
1.84
0.011
Fe
2
O
0.45
128
57.6
0.339
Total
169.7
1.001
are composed of nickel--iron and olivine. Olivine
is a major constituent in all the chondrites except
the enstatite chondrites and some of the car-
bonaceous chondrites, and it is present in some
achondrites. For these reasons -- and ignoring the
enstatite chondrites -- olivine is usually consid-
ered to be the major constituent of the mantle.
The olivines in meteorites, however, are generally
much richer in FeO than mantle olivines. The
olivine compositions in chondrites generally lie
in the range 19 to 24 mole% fayalite. Removal
of iron from meteoritic olivine, either as Fe or
