210 Encyclopedia of the Solar System
FIGURE 22 Maps of lateral variations of S velocities at four
depths in a shear velocity model of Ekstr ̈om and Dziewonski.
The yellow/red colors indicate slower than average velocities and
blue – the faster. The range of variations is about 7% at 70 km
and 3% at near the core–mantle boundary.
seismographic stations is now a real, even though expen-
sive, possibility.
Figure 22 is a collection of maps of the shear velocity
anomalies from a recent model of the mantle by Ekstr ̈om
and Dziewonski published in 1998, built using a wide range
of types of data (travel times, surface wave dispersion and
waveforms). The nominal resolution of this model is about
100 km in depth and 1500 km horizontally near the surface.
At 70 km depth, the model agrees with the predictions
of the plate tectonics and the thermal history of the conti-
nents. The stable continental areas (old and cool) are very
fast (up to+7% ), whereas material under the midocean
ridges is much slower than normal (up to−7% ). This
negative anomaly decreases with the increasing age of the
oceanic plate to become faster than average for ages greater
than 100 Ma. The depth to which the anomalies associated
with the midocean ridges persists in the tomographic maps
(>200 km) puzzles geodynamicists who think that mido-
cean ridges are passive features and that conditions below
about 100 km depth are isothermal.
Figure 23 shows a cross-section through the upper
mantle of the Pacific from the model of Ekstr ̈om and
Dziewonski; the direction of the cross-section follows the
direction of motion of the Pacific plate. Going from East
to West, we see higher seismic velocities associated with
subduction under South Africa; very slow lithosphere at
the East Pacific rise; increase in velocities with the distance
from the ridge, and subduction under the Mariana trench;
the red dots are earthquakes. It is clear that the veloci-
ties change with age to depths below 200 km. The map in
Figure 22 at 200 km depth shows diminished variability of
velocities under ocean but still very strong anomalies under
the continents; the old cratons, in particular.
The map in Figure 22 at 500 km depth represents aver-
age shear velocity anomalies in the transition zone. The most
characteristic features are the fast anomalies in the west-
ern Pacific and Eastern Asia, under South America, the
Atlantic, reaching to western Africa. In the western Pacific
they can be associated with subduction zones, although they
are much wider than an anomaly associated with a 100-km-
thick slab. Studies of the topography of the 660-km dis-
continuity show that the areas of high seismic velocity are
correlated with a depressed boundary, yielding credence to
an interpretation that these anomalies are indicative of an
accumulation (temporary, perhaps; see earlier discussion
on the models of flow in the mantle) of the subducted ma-
terial in the transition zone. Figure 24 shows comparison
of lateral variation in velocities obtained in a model named
S362D1. The two maps one just above and the other just
below the 660-km discontinuity are very different; the map
representing the transition zone shows features similar to
that at 500-km map in Figure 22; the lower mantle map
is quite different and has distinctly different spectral con-
tent: it is dominated by relatively short wavelength features.
This result, and similar results obtained by other modeling
groups, supports the concept of a separation–perhaps not
absolute–between the upper mantle and lower mantle.
In the middle mantle the anomalies are not well orga-
nized. This observation contrasts with the results of “high-
resolution” tomography which in this depth range shows
two narrow high velocity features: one stretching from the
Hudson Bay to Bolivia and the other from Indonesia to theFIGURE 23 Cross-section through the upper
mantle of model of Ekstr ̈om and Dziewonski. Note
that velocities change as a function of distance from
the East Pacific Rise (proportional to age of the
plate) to depths greater than 200 km Red dots
indicate earthquakes. There is vertical exaggeration
by a factor of about 20.