Geoscience Reference
In-Depth Information
convection, and the mechanisms of heat removal;
the present styles of mantle convection and plate
tectonics were unlikely to have been operating.
of the Earth. The continental crust contributes
5.8--8.0 TW to the total energy budget of the
Earth.
The estimates of radioactivity in the conti-
nental crust are incompatible with geochemi-
cal models that attribute its origin to extrac-
tion from a chondritic or primitive upper mantle,
leaving behind an undifferentiated lower mantle.
There is not enough material in the upper man-
tle portion of undifferentiated mantle to provide
the concentrations inferred for the more incom-
patible elements. One-hundred-percent efficient
extraction, and absence of recycling replenish-
ment, is unlikely. It appears that the portion of
the mantle from which the crust was extracted
was already enriched in the LIL, most likely as
a result of a radial zone-refining process concur-
rent with accretion. A corollary is that crustal
extraction did not have to approach 100% effi-
ciency, and that there is still U, Th and K in
the upper mantle. The deep mantle may have
little heat-production capability; it may be bar-
ren or sterile. If indeed most of the radioactive
heating in the mantle is shallow then there will
be a smaller lag between heat production and
heat flow and the bulk of the slab cooling effect
will be in the region of main heat production.
Some of the arguments against layered convec-
tion assume that the lower mantle is U-rich and
will therefore overheat. The core would also lose
heat less efficiently in this kind of layered model,
which may pose problems for the growth of the
inner core and maintenance of the dynamo.
Distribution of radioactive elements
There is a strong decrease in the concentrations
of the radioactive elements as one goes from
the upper crust to the lower crust to the upper
mantle. Accretion and early differentiation prob-
ably swept most of the radioactivities toward the
surface. Present recycling involves dehydration
at depths of the order of 200--300 km, which
removes incompatible elements from the slab
and places them in the mantle wedge above the
slab. The residual slab carries LIL-poor material to
greater depths. A vigorously stirred mantle would
tend to be homogenous. In a layered mantle with
dehydration, fluid migration and partial melting
we expect the processes of fractionation, gravi-
tational separation and differentiation to dom-
inate over processes of convective homogeniza-
tion, except where melts are co-mingled prior to
eruption. Plates, continents, downward increase
of viscosity, phase changes, non-Newtonian rhe-
ology and chemical stratification all serve to
decrease the vigor of convection and to prevent
efficient homogenization. The net effect could be
a rapid, or exponential, decrease of K, U and Th
content with depth, rather than the completely
homogeneized and depleted upper mantle and a
U-rich lower mantle, as in current geochemical
models.
Continental lithosphere
Continental crust
Based on xenolith studies, the continental litho-
sphere (CL) may have radioactivities as high as
10% of the level of average crustal rocks. The
CL may contribute half or more of the heat
flow in older or low-heatflow continental ter-
rains. Peridotitic xenoliths represent only part of
the subcrustal mantle, possibly the most depleted
part.
Large parts of the Archean cratonic
lithosphere appear to be enriched in
UandTh
. About 15 mW/m
2
of heat is gener-
ated in the continental keel (Vitorello and Pol-
lack, 1980). This requires radioactivities about
Estimates of the U and Th concentrations of
continental crust range from 0.9--1.3 ppm and
3.5--9.0 ppm, respectively (Rudnick and Fountain,
1995). Estimates of the K
2
O content show a simi-
lar range, 1.1--2.4 wt.%. The inferred heat produc-
tions and heatflows range from 0.58--1.31 mW/m
3
and 23--52 mW/m
2
. The large range reflects, in
part, uncertainty in the composition of the lower
crust. More than 30% of the most incompatible
elements in BSE are in the continental crust.
This is a measure of the differentiation efficiency
