Geology Reference
In-Depth Information
the geophysical requirement that the mantle is heated more within than from below
(Chapters 7 and 8). The significantly lower concentrations obtained in previous
estimates for the 'depleted MORB mantle' are plausibly explained by the failure
to consider the full implications of mantle heterogeneity and by a continuing
tendency to think in terms of mixing from end-member reservoirs or components,
as discussed in Sections 10.4 to 10.6. The remaining significant discrepancy is
between the heating required by conventional thermal evolution models (Sections
9.1 and 9.2) and the global abundances of heat source elements inferred from
cosmochemistry, as discussed in Section 9.3.
Dynamical models of the mantle, of the kind presented in Chapter 9, have by
now reproduced the four main features of the refractory incompatible elements.
We can therefore consider that the physical and chemical models of the mantle
have been reconciled regarding the important features of these elements, except
for the level of radioactive heating required for evolving models. The basis for this
conclusion will now be outlined.
10.7.1 Heterogeneity and the MORB-OIB difference
Christensen and Hofmann [122] showed the way in 1994 by modelling subducted
oceanic crust and depleted mantle in a convection model using tracers. They showed
that some crustal material accumulated at the bottom, thus supporting Hofmann
and White's hypothesis [74] that plumes incorporate additional oceanic crust from
the D
layer. They showed that residence times would yield apparent ages similar
to the lead observations. They also obtained a range of isotopic heterogeneity com-
parable to the observations. Their model was limited in significant respects, both
by the computer power available at the time and by some of their assumptions, but
they established an approach that has proven very fruitful. Subsequent modelling
extended this approach to the full mantle Rayleigh number and full age of the Earth
[203], to three dimensions [215-217] and to fully evolutionary models [156, 218]
like those of Figures 9.5 and 9.10.
Davies [203] discussed the approach and added several considerations. The
Christensen and Hofmann [122] (C&H) models were run for only 3.6 Gyr at
steady conditions, but about 18 Gyr are required at steady conditions to achieve
the required number of transits (Section 10.4.3), or evolving models with decay-
ing heat sources are needed, as later became feasible. Most of the C&H models
were mainly bottom-heated, rather than mainly internally heated, and the resultant
strongly upwelling sheets may have affected residence times. C&H sampled the
chemical tracers throughout the model at the end of the run, whereas geochemists
sample only the oceanic crust, formed at the top, and OIBs, drawn from the bottom,
so the models should be sampled accordingly.
