206 Encyclopedia of the Solar System
FIGURE 20 The preliminary Earth model
(PREM) of Dziewonski and Anderson (1981)
describing the compressional velocity(νs), shear
velocity (νS), and density (ρ). (a) Model for the
entire Earth and (b) an expansion of the
uppermost 1000 km. From Moho to 220 km depth
the model is characterized by transverse
anisotropy, in which the waves propagating in the
vertical (solid line) and horizontal (dashed lines)
planes have different velocities. Parameterη,
characterizing the propagation of P waves at
intermediate angles, is unity in an isotropic
medium and is about 0.95, just under the Moho.
Below 220 km depth the model is isotropic.of equations governing the wave propagation in liquid and
in solid are different. The commonly adopted solution is to
introduce a layer of water whose thickness is such that the
total volume of water in all the oceans and that calculated for
the SSEM are equal. It is a reasonable decision, but it will
be necessary to introduce corrective measures even when
constructing the model, as practically all seismographs that
record ground motion are located on land.
This chapter uses the preliminary reference Earth model
(PREM) published in 1981 by Dziewonski and Anderson
as an example. It has been derived using a large assem-
bly of body-wave travel time data, surface wave dispersion
and periods of free oscillations, collected through the end
of 1970s. An effort to revise it is now under way: a large
body of very accurate data has been assembled in the nearly
20 years since the publication of PREM. However, with the
exception of the upper mantle, no substantial differences
are expected. A reference model designed to fit the travel
times of body waves (ak135) has been developed by Kennett
and Engdahl in 1995.
Figure 20a shows the density, compressional velocity,
and shear velocity in the model PREM. To illustrate the
complexities in the uppermost 800 km of the model, its
expansion is shown in Figure 20b. In what follows, we shall
give a brief summary of our knowledge and significance of
the individual shells in the Earth’s structure.
6.1 Crust
This is the most variable part of the Earth’s structure, both
in terms of its physical properties as well as history. Large
areas of the Earth’s surface are covered by soils, water, and
the sediments. These provide support for life and economic
activity. However, the vast proportion of what is called “the
crust of the Earth” consists of crystalline rocks, mostly of
igneous origin.
The primary division is between the continental and the
oceanic crust. The former can be very old, with a signifi-
cant fraction being older than 1.5 Ga. It is light, with an
abundance of calcium, potassium, sodium, and aluminum.
Its average thickness is 40 km, but varies substantially, from
about 25 km in the areas of continental thinning due to
extension (the Basin and Range province in the Western
United States, for example) to 70 km under Tibet, in the
area of continent—continent collision.
The oceanic crust is thin (7 km, on average, covered
by some 4.5 km of the ocean), young (from 0 to 200 Ma),
and somewhat more dense, with a greater abundance of
elements such as magnesium and iron. It is created at the
midocean ridges and is consumed in subduction zones, with
trenches being their surficial manifestation. The difference
between oceanic and continental crusts is called by some
the most important fact in Earth sciences, as it is related
intimately to plate tectonics. The thinner, denser oceanic
crust provides conditions more favorable for initiation of
the subduction process.
Overall, crustal thickness follows the Airy’s hypothesis
of isostasy closely: thick roots under mountains and a thin
crust under “depressed” areas—oceans. The seismic veloc-
ities in the crust increase with depth. It is a subject of a
debate whether this increase is gradual or the crust is lay-
ered; recently, the latter view has begun to prevail.