Geoscience Reference
In-Depth Information
of incompatible (or fusible) elements with respect to compatible (or refractory) elements
against an ideal average mantle model (Bulk Silicate Earth or BSE). The opposite of
“depleted” is of course “enriched.” Depletion can be determined for the time immedi-
ately before melting using the analysis of trace elements: for example, less-than-chondritic
La/Sm ratios signal a refractory source depleted in incompatible elements. It may also
be determined for long periods that precede melting by analyzing isotopic compositions of
elements containing a radiogenic isotope (
87
Sr,
143
Nd,
176
Hf,
206
Pb,
207
Pb, etc.) In MORB,
Nd more radiogenic (
143
Nd
144
Nd
11) than in BSE (
143
Nd
144
Nd
/
=
0.5132 or
ε
Nd
=+
/
=
0.51265 or
0) indicates that the asthenospheric mantle was depleted with respect to
a chondritic mantle at least 1-2 billion years ago.
The MORB source is therefore depleted in the more incompatible elements both in the
indistinguishable from the asthenosphere, and so must be most of the upper mantle. Since
granites, a major constituent of continental crust, are enriched in both the long and short
terms, a core concept has emerged according to which depletion of the upper mantle reflects
the extraction of continental crust throughout geological time. In contrast, the OIB source
is generally depleted in the long term (as indicated by radiogenic isotopes) but enriched in
the short term (as shown by their incompatible element distributions). A popular idea is that
the OIB source has been enriched before melting by fluid and magma percolation reacting
with the rock matrix: this process is known as metasomatism. In most cases, however,
modern models of melting make this ad hoc interpretation rather unnecessary.
The orogenic magmas arising from the subduction zones, such as andesites, have a more
complex history. If they are produced at a continental margin, such as the Andes, they
clearly exhibit interaction between mafic magmas from the mantle and anatectic melts
from the continental crust. If they are produced in oceanic convergence zones, the result
is more surprising: the isotopic composition of their Nd is very similar to that of the Nd
of descending lithospheric plates. By contrast, their
87
Sr/
86
Sr ratio is particularly radio-
genic, suggesting that marine Sr is present in the mantle source of these magmas. This is
often taken as evidence that dehydration of sediments and altered basalts of the downgoing
plate produces fluids which trigger melting in the overlying mantle wedge. Andesites are
often seen as the resulting melts. This interpretation is confirmed by higher
ε
Nd
=
18
O values
δ
(
), therefore requiring the pres-
ence in their source material of rock components that have been subjected to weathering
or alteration at low temperature. Andesites are richer in Si than basalts, which is consistent
with the petrological concept stated earlier that water increases the Si content of melts.
A number of trace elements exhibit abundance levels typical of such an environment: in
particular the extreme depletion in Nb and Ta, which are thought to be locked in highly
refractory minerals such as titanium oxide, in contrast to many incompatible elements (Rb,
K, Ba, Sr, Pb, La, Th, etc.), which are returned toward the surface by dehydration of the
plates. Petrology indicates that the origin of andesites cannot be ascribed to simple melt-
ing of the subducted crust as this process would produce lava far too rich in silica. This
liquid is a special kind of granitic melt known as dacite. It can therefore be imagined that
these liquids released by melting of the hydrated ocean crust react with the mantle located
above the plate. Reaction between the mantle and the felsic liquids generated by hydrated
fusion of the plate modifies the mantle overlying subduction zones: andesites are thought to
>
7
) than for the average of the ordinary mantle (
≈
5.2
