Geoscience Reference
In-Depth Information
subduction history. But this mantle will also
correlate with hotspots, even if only 'normal'
mantle is present. Some hotspots correlate
with some tomographic maps at some depths
[
scoring hotspots
].
trating into the lower mantle. Thermal coupling
could be an alternative explanation of the tomo-
graphic results.
Correlations with heat flow
and geoid
Chemical stratification
The possible chemical stratification of mantle
convection is an issue of current interest, which
I will keep coming back to. It is hard to prove or
disprove simply by looking at tomographic cross-
sections. The endothermic nature of the phase
change at 650 km depth may cause slab flatten-
ing and delay slab penetration but it cannot pre-
vent large masses of cold material -- of the same
composition as the surrounding mantle -- from
episodically cascading across the boundary. The
combination of a negative Clapyron slope at 650
km, a chemical barrier near 1000 km, and an
increase of viscosity with depth may serve to con-
fine the plate tectonic cycle and recycling to the
outer 1000 km of the mantle.
Below some -- not all -- subduction zones,some --
not all -- seismic tomographic cross-sections con-
tain bands of fast anomalies that extend into
the mantle below 650 km. These visual anoma-
lies are generally interpreted as slabs penetrat-
ing through the 650 km seismic discontinuity,
and as evidence in support of 'whole-mantle con-
vection flow' (Grand
et al.
, 1997; Bijwaard
et al.
,
1998). However, thermal coupling between two
flow systems separated by an impermeable inter-
face provides an alternative explanation of the
tomographic results.
The dynamical interpretations of the geoid
showthatthemodelwithanimpermeableinter-
face at 650 km or deeper can satisfy the data
equally well as the whole-mantle model (Wen
and Anderson, 1997) and some resistance to up-
and downgoing flows can reduce the amplitudes
of surface dynamic topography (Thoraval
et al.
,
1995). If there is an impermeable interface at 650
km, cold subducted lithosphere that has accumu-
lated above the boundary at 660 km can cool
the underlying material and initiate a down-
welling instability into the lower mantle. In such
a case, it would be difficult on the basis of seismic
tomography alone to discriminate between 'ther-
mal slabs' (with no material exchange between
the upper and lower mantles) and slabs pene-
Love-wave velocities are well correlated with
heat flow and, therefore, with surface tectonics.
Shields are areas of low heat flow and exhibit
high Love-wave velocities, in spite of the thick
low-velocity crust. From the Love-wave data one
would predict that southeast Asia and the Afar
region are characterized by high heat flow.
The geoid has a weak correlation with surface-
wave velocities, consistent with a deep origin for
the causative mass anomalies. The correlation
between surface-wave velocities and the geoid is
weak.
Regions of generally faster than average
velocity occur in the western Pacific, the western
part of the African plate, Australia--southern
Indian Ocean, part of the south Atlantic, north-
eastern North America--western North Atlantic
and northern Europe. These are all geoid lows.
Dense regions of the mantle that are in isostatic
equilibrium generate geoid lows. High density
and high velocity are both consistent with cold
mantle. The above regions may be underlain
by cold subducted material. Geoid highs occur
near Tonga-Fiji, the Andes, Borneo, the Red Sea,
Alaska, the northern Atlantic and the southern
Indian Ocean. These are generally slow regions
of the mantle and are therefore presumably
hot. The upward deformation of boundaries
counteracts the low density associated with the
buoyant material, and for uniform viscosity
the net result is a geoid high (Hager, 1983). An
isostatically compensated column of low-density
material also generates a geoid high because of
the elevation of the surface.
Features of the geoid having wavelengths of
about 4000 to 10 000 km are generated in the
upper mantle. Geoid anomalies of this wave-
length generally have an amplitude of about
20 to 30 m. An isostatically compensated den-
sity anomaly of 0.5% spread over the upper
mantle would give geoid anomalies of this size. It
