Geoscience Reference
In-Depth Information
including partial melting, and dislocation relax-
ation. Small-scale heterogeneity such as slabs and
dikes scatter seismic energy and this mimics
intrinsic anelasticity. Allowance for anelastic dis-
persion increases the inferred high-frequency
velocities in the low-velocity zone determined
by free-oscillation and surface-wave techniques,
but partial melting is still required to explain
the regions of very low velocity. Allowance for
anisotropy results in a further upward revision
for the velocities in this region, as discussed
below. This plus the recognition that subsolidus
effects, such as dislocation relaxation, can cause
a substantial decrease in velocity has complicated
the interpretation of seismic velocities in the
shallow mantle because one wants to compare
results with laboratory data. Velocities in tectonic
regions and under some oceanic regions, how-
ever, are so low that partial melting is implied.
In most other regions a subsolidus mantle com-
posed of oriented olivine-rich aggregates can
explain the velocities and anisotropies to depths
of about 200 km. Global tomographic inversions
involve a laterally heterogenous velocity and
anisotropy structure to depths as great as 400 km.
Mantle tomography averages the seismic
velocity of the mantle over very long wavelengths
and travel distances. Sampling theory tells us
that extremes of velocity are averaged out in this
procedure. As paths get shorter and shorter the
variance goes up and sections of the path become
much slower or much faster than inferred from
global tomography. The minimum shear veloc-
ity found along ridges and backarc basins -- and
probably elsewhere -- in high-resolution studies,
is smaller than inferred from tomography. If
global shear velocities approach the minimum
shear velocities in solid rocks, then a reasonable
variance of 5% placed on top of this probably
means that the low-velocity regions require some
melting.
The rapid increase in velocity below 220 km
may be due to chemical or compositional changes
(e.g. loss of water or CO
2
) or to transition from
relaxed to unrelaxed moduli. The latter expla-
nation will involve an increase in
Q
, and some
Q
models exhibit this characteristic. However, the
resolving power for
Q
is low, and most of the
seismic
Q
data can be satisfied with a constant-
Q
upper mantle, at least down to 400 km.
Tomographic results show that the lateral
variations of velocity in the upper mantle are
as pronounced, and abrupt, as the velocity varia-
tions that occur with depth. Thus, it is mislead-
ing to think of the mantle as a simple layered
or 1D system. Lateral changes are, of course,
expected in a convecting mantle because of vari-
ations in temperature and anisotropy due to crys-
tal orientation. They are also expected from the
operation of plate tectonics and crustal processes.
In particular, lithologic heterogeneity is intro-
duced into the mantle by subduction and lower
crustal delamination. But phase changes and par-
tial melting are more important than tempera-
ture and this is why the major lateral changes
are above about 400 km depth.
The geophysical data (seismic velocities, atten-
uation, heat flow) are consistent with par-
tial melting in the low-velocity regions in the
shallow mantle. This explanation, in turn, sug-
gests the presence of volatiles in order to depress
the solidus of mantle materials, or a higher-
temperature mantle than is usually assumed. The
top of the low-velocity zone may mark the cross-
ing of the geotherm with the wet solidus of peri-
dotite, or the solidus of a peridotite--CO
2
mix.
Its termination would be due to (1) a crossing
in the opposite sense of the geotherm and the
solidus, (2) the absence of water, CO
2
or other
volatiles or (3) the removal of water into high-
pressure hydrous phases or escape of CO
2
.In
all of these cases the boundaries of the low-
velocity zone would be expected to be sharp.
Small amounts of melt or fluid (about 1%) can
explain the velocity reduction if the melt occurs
as thin grain-boundary films. Considering the
wavelength of seismic waves, magma-filled dikes
and sills, rather than intergranular melt films,
would also serve to decrease the seismic veloc-
ity by the appropriate amount. Slabs that are
thin and hot at the time of subduction may
stack up in the shallow mantle. The basaltic parts
will melt since the solidus of basalt is lower
than the ambient temperature of the mantle.
At seismic wavelengths this will have the same
effect as intergranular melt films. Anelasticity
and anisotropy of the upper mantle may also be
due to these mega-scale effects rather than due
to crystal physics, dislocations, oriented crystals
and so on.
