Geoscience Reference
In-Depth Information
largely different iron content and hence differ-
ent electrical conductivity. However, the volume
fraction of eclogite is small in most regions (
1300 1200 1100 1000
900
800
700 (K)
amphibolite
5%
for MORB source regions), so eclogite has little
effect on the bulk conductivity, and the variation
in electrical conductivity in the mantle reflects
the variation in water content more directly.
However, in some source regions of ocean is-
land basalts, the eclogite might occupy
∼
1
Hydrous
10
-2
HS
+
Dry
HS
−
10
-4
HS
+
10-20%
(Sobolev
et al
., 2007) and in these regions the
high conductivity of eclogite will have some ef-
fects on the bulk conductivity if eclogite bodies
are interconnected.
Figure 5.15 shows models of electrical conduc-
tivity in the Earth's upper mantle and the transi-
tion zone assuming the pyrolite composition. The
model is not constructed for the lower mantle
because there is no experimental data to evaluate
the influence of hydrogen on electrical conduc-
tivity in lower mantle minerals. In this model, we
assumed a typical adiabatic geotherm (plus small
changes due to the latent heat release effect). We
assume the pyrolite composition (homogeneous
composition). Even though the bulk chemistry is
assumed to be constant, oxygen fugacity changes
with depth due to self-buffering effect. Oxygen
fugacity in Earth's upper mantle is controlled
by a reaction among skiagite (
Fe
2
3
Fe
3
2
Si
3
O
12
garnet), olivine and orthopyroxene (in the garnet
peridotite) and decreases with depth relative to
the commonly used buffering reactions such as
QFM (Frost &McCammon, 2008). Consequently,
the oxygen fugacity in the upper mantle decreases
significantly with depth (relative to the QFM
buffer). This has an important effect on electrical
conductivity: it decreases iron-related conduction
but it increases hydrogen-related conduction.
Such effects are included in Figure 5.15 (such
effects were not considered by Fullea
et al
., 2011;
Yoshino, 2010; Baba
et al
., 2010; Khan and Shank-
land, 2012, which resulted in the systematic error
of conductivity of
∼
Yang
et al.
2011
10
-6
HS
−
10
-8
8
10
12
14
16
10
4
/ T(K)
Fig. 5.14
Electrical conductivity in the lower crust: a
comparison of laboratory data with geophysical
inference (from Wang
et al
., 2012a). Hatched regions
correspond to a range of conductivity of the
continental lower crust inferred from geophysical
studies. The lines with ''Hydrous'' corresponds to
minerals containing
0.04 wt % water, ''Dry''
correspond to water-free samples.
HS
+
,
−
correspond to
the Hashin-Shtrikman upper and lower bounds
respectively. Thick lines are for amphibolites. Above
∼
∼
800K, conductivity of amphibolite increases strongly
with temperature due to oxidation of iron caused by
dehydration. Reproduced with permission of Springer.
amphibole enhances electrical conductivity at
high temperature (T
>
800 K) in amphibolites.
This enhancement is not due to the production
of aqueous fluids but due to the change in the
oxidation states of iron. This leads to highly
temperature sensitive conductivity, and the high
conductivity (10
−
2
to 10
−
1
S/m) can be explained
by the modest
800-900 K)
without invoking partial melting or the role of
aqueous fluids (Figure 5.14).
temperatures
(
∼
(b) The upper mantle and the transition zone
In
contrast to the lower crust, the distribution of
temperature and major element chemistry in the
mantle is rather uniform except for the litho-
sphere. However, there is some evidence that the
mantle is a mixture of peridotite and eclogite (e.g.,
Sobolev
et al
., 2007). These two components have
one order of magnitude). The
depth variation in water content in the conti-
nental lithosphere inferred by Fullea
et al
. (2011)
could be due to the influence of depth variation
of oxygen fugacity in otherwise dry continental
lithosphere. Similarly, the partitioning of key
∼
