Geology Reference
In-Depth Information
(a)
(b)
SAMPLE:
SAMPLE:
Composition
Composition
SOURCE:
SOURCE:
mixing
Composition
Composition
Figure 10.15. Contrasting conceptions of mantle sources. (a) Generation of a sam-
ple by mixing material from two reservoirs before eruption. (b) Direct sampling
of a heterogeneity of intermediate composition.
a primitive lower mantle [107]. The concept is illustrated in Figure 10.15(a). In
contrast, if the mantle is conceived of as being heterogeneous, with a range of
compositions already existing in the mantle, then the Nd-Sr array can be viewed
as directly sampling that range of compositions, as in Figure 10.15(b).
Some of the reasons for expecting the mantle to be heterogeneous have been
covered in Section 10.4. In addition, we can note that Figure 10.2(a) would require
at least three source types to span the two-dimensional array of data. Indeed,
geochemists soon defined four or more source types required to span the data in
higher dimensions [177, 178, 212]. Thus two-reservoir mixing is not plausible.
10.6.2 No primitive mantle
The identification of the less depleted source as primitive was always in conflict
with the lead isotope data (Figure 10.3(a)), which implies that this reservoir would
have
206
Pb/
204
Pb of 22 or more, well away from the primitive 'geochron' locus,
whereas the putatively depleted source plots around the geochron. This is the reverse
of the relationship that would be expected, as I noted in 1984 [115]. There is no
significant evidence for a primitive reservoir [60], and later trace element ratios
such as Nb/U (Figure 10.7) reinforce the point made from the lead isotopes. Nor
do we expect a significant amount of primitive mantle, according to the discussion
of Section 10.4.3.
The later proposal for a deeper reservoir that is enriched but not necessarily
primitive [108] avoids some of these problems. Indeed, its origin was ill-defined
enough that it could be a receptacle for virtually any source type required to span the
data. However, this layer would be required to contain half of the Earth's uranium
and would be expected to generate a large plume flow, as discussed in Section 8.3
(see Figures 8.6 and 8.7), and the Earth's topography is inconsistent with this.
