Earth as a Planet: Surface and Interior 207
6.2 Upper Mantle: Lithosphere and Asthenosphere
(25–400 km Depth)
The boundary between the crust and the upper mantle was
discovered in 1909 by a Yugoslavian geophysicist Andreiji
Mohorovicic. It represents a 30% increase in seismic ve-
locities and some 15% increase in density. It is a chemical
boundary with the mantle material primarily composed of
minerals olivine and pyroxene, being much richer in heavier
elements, such as magnesium and iron.
The terms lithosphere and asthenosphere refer to the
rheological properties of the material. The lithosphere,
strong and brittle, is characterized by very high viscosity.
It is often modeled as an elastic layer. It includes the crust
and some 30 to 100 km of the upper mantle. The astheno-
sphere is hotter (>1573 K by convention), its viscosity much
lower, and in modeling is represented by yielding. Under
loads, such as glacial caps, the lithosphere bends elasti-
cally, whereas the asthenosphere flows. The difference of
rheological properties is explained by differences in tem-
perature: the viscosity is an exponential function of tem-
perature. The lithosphere is relatively cool; the transport
of heat is mostly through conduction. The asthenosphere
is hotter, and the convective processes are believed to be-
come important. Low viscosity of the asthenosphere is used
to explain the mechanical decoupling between the plates (in
the plate tectonic theory) and the underlying mantle. The
depth of this decoupling varies with position: it is shallow
near midocean ridges and increases as the plate cools with
time and its lithosphere grows in thickness.
The continents, with its very old and cold shield regions,
may be significantly different. If the hypothesis of the “tec-
tosphere” is correct, it may have roots that are 400 km deep
and move as coherent units over long periods of the Earth’s
history. The depth of roots is still subject to a debate (most
recent results would indicate their depth extent as 200–250
km). As the seismic velocities decrease with increasing tem-
perature, the vertical gradient of seismic velocities in the
transition between the lithosphere and the asthenosphere
may become negative. This is called the “low velocity zone”;
its presence creates a shadow zone in seismic wave propaga-
tion, making interpretation of data complex and nonunique.
Measurements of attenuation of seismic waves led to the
determination of models of Q (quality factor) for the shear
and compressional energy. Anelastic dissipation of shear
energy, due to grain boundary friction, is most important.
Attenuation in the range of depths corresponding to the
low velocity zone is several times stronger than in the litho-
sphere.
6.3 Transition Zone (400–660 km Depth)
Knowledge of the composition of the transition zone is es-
sential to the understanding of the composition, evolution,
and dynamics of the Earth. In seismic models, this depth
range has been known for a long time to have a strong veloc-
ity gradient; much too steep for an increase under pressure
of the elastic moduli and density of a homogeneous ma-
terial. It was first postulated in the 1930s that this steep
gradient may be due to phase transformations: changes in
the crystal lattice that for a given material take place at
certain temperatures and pressures.
In the 1960s, when major improvement in seismic instru-
mentation took place, two discontinuities were discovered:
one at 400 km and the other at 670 km (the current best
estimate of the global average of their depth is 410 and
660 km, respectively). Their existence has been well doc-
umented by nearly routine observations of reflected and
converted waves. There is still some uncertainty of how
abrupt the velocity changes are: the 410-km discontinuity
is believed to be spread over some 5–10 km, whereas the
660-km discontinuity appears to be abrupt. The estimates
of the velocity and density contrasts are still being studied
by measuring the amplitudes of the reflected and converted
waves; the values of these contrasts are important for un-
derstanding the mineralogical composition of the transition
zone.
In general terms, the seismological models are consistent
with the hypothesis that olivine is the main (up to 60% ) con-
stituent of the upper mantle. Laboratory experiments under
pressures corresponding to depths up to 750 km show that
olivine undergoes phase transformations to denser phases
with higher seismic wave speeds. At pressures roughly cor-
responding to 400 km depth, theα-olivine transforms into
β-spinel. The latter will transform toγ-spinel at about
500 km depth, with only a minor change in seismic ve-
locities. Indeed, a seismic discontinuity at 520 km has been
reported, although some studies indicate that in some parts
of the world it may not be substantial enough to be de-
tected. At 660 kmγ-spinel transforms into perovskite and
magnesio-w ̈ustite.
Although olivine may be the dominant constituent, it is
not the only one. The presence of other minerals compli-
cates the issue. Also, there are other hypotheses of the bulk
composition of the upper mantle: “piclogyte model,” for
example.6.4 Lower Mantle (660–2890 km)
The uncertainties in the mineralogy of the upper mantle and
the bulk composition of the Earth have created one of the
most stubborn controversies in the Earth sciences: are the
upper mantle and lower mantle chemically distinct? A “yes”
answer means that there has not been an effective mixing
between these two regions throughout the Earth’s history,
implying that the convection in the Earth is layered. The
abrupt cessation of seismic activity at about 660 km depth,
coinciding with the phase transformation described earlier,
and geochemical arguments—mostly with respect to differ-
ences in isotopic composition of the midocean ridge basalts