Theory of the Earth
Don L. Anderson
Chapter 3. The Crust and Upper Mantle
Boston: Blackwell Scientific Publications, c1989
Copyright transferred to the author September 2, 1998.
You are granted permission for indi
vidual, educational, research and
noncommercial reproduction, distribution, display and performance of this work
in any format.
Recommended citation:
Anderson, Don L. Theory of the Earth.
Boston: Blackwell Scientific Publications,
1989.
http://resolver.calt
ech.edu/CaltechBOOK:1989.001
A scanned image of the entire book may
be found at the following persistent
URL:
http://resolver.caltech.edu/CaltechBook:1989.001
Abstract:
T he structure of the Earth's interior is
fairly well known from seismology, and
knowledge of the fine structure is impr
oving continuously. Seismology not only
provides the structure, it also provid
es information about the composition,
crystal structure or mineralogy and physical state. In subsequent chapters I will
discuss how to combine seismic with othe
r kinds of data to constrain these
properties. A recent seismological model
of the Earth is shown in Figure 3-1.
Earth is conventionally divided into crust,
mantle and core, but each of these has
subdivisions that are almost as fundamen
tal (Table 3-1). The lower mantle is the
largest subdivision, and therefore it dominates any attempt to perform major-
element mass balance calculations. The crust is the smallest solid subdivision,
but it has an importance far in excess of it
s relative size because we live on it and
extract our resources from it, and, as we sh
all see, it contains a large fraction of
the terrestrial inventory of many elements. In this and the next chapter I discuss
each of the major subdivisions, starting wi
th the crust and ending with the inner
core.
The
Crust and
Upper Mantle
Z
O
E
:
Come
and
I'll
peel off.
B
L
O
O
M
:
(feeling
his
occiput dubiously with
the
unparalleled embar
-
rassment
of
a harassed pedlar
gauging
the
symmetry
of
her
peeled
pears)
Somebody
would
be
dreadfully jealous
if she
knew.
-
JAMES
JOYCE,
ULYSSES
T
he structure
of
the
Earth's
interior is fairly
well known
from seismology,
and
knowledge
of
the
fine
structure
is improving continuously. Seismology
not
only
provides
the structure,
it also
provides
information
about
the com
-
position, crystal structure or mineralogy
and
physical state.
In
subsequent chapters I will discuss
how
to
combine seis
-
mic with
other
kinds
of
data
to
constrain these properties.
A
recent seismological
model
of
the Earth is
shown
in
Fig
-
ure 3
-
1.
Earth is conventionally divided into crust,
mantle
and
core,
but
each
of
these
has
subdivisions that are
almost
as fundamental
(Table 3
-
1). The
lower
mantle is the largest
subdivision,
and
therefore
it dominates
any
attempt to per
-
form major
-
element
mass
balance calculations. The crust is
the smallest solid subdivision,
but
it has
an
importance far
in excess
of
its relative size because
we
live
on it
and
extract
our resources from
it,
and,
as
we
shall see,
it contains a
large fraction
of
the terrestrial inventory
of
many
elements.
In
this
and
the
next
chapter I discuss each
of
the major sub
-
divisions, starting
with
the crust
and
ending
with
the inner
core.
THE
CRUST
The
major divisions
of
the
Earth's
interior
-
crust,
mantle
and
core
-
have
been
known
from seismology
for
about 70
years. These are
based
on the
reflection and
refraction
of
P
-
and S
-
waves. The boundary
between the
crust
and mantle
is called the
MohoroviCiC
discontinuity (M
-
discontinuity or
Moho for short) after the Croatian seismologist
who discov
-
ered
it in
1909.
It separates rocks
having
P
-
wave velocities
of
6
-
7
km/s
from
those having
velocities
of
about
8
km/s.
The term
"
crust
"
has been used
in
several ways.
It initially
referred
to the brittle
outer shell of
the
Earth
that
extended
down
to the asthenosphere
(
"
weak layer
"
);
this is
now
called the lithosphere
(
"
rocky
layer
"
). Later it
was used
to
refer to the rocks occurring
at or near
the surface
and
acquired a petrological connotation. Crustal rocks
have
distinctive physical properties that
allow
the crust to
be
mapped by
a variety
of
geophysical techniques.
The term
"
crust
"
is now
used
to refer
to that
region
of
the Earth
above the
Moho.
It represents
0.4
percent
of
the
Earth's
mass.
In
a strict sense, knowledge
of
the existence
of
the crust is
based
solely
on
seismological data. The
Moho
is
a sharp seismological
boundary
and in
some
re
-
gions
appears to
be
laminated. There are three
major
crustal
types
-
continental,
transitional
and
oceanic. Oceanic crust
generally ranges from 5 to 15
km
in
thickness
and
com
-
prises 60 percent
of
the total
crust
by
area
and
more than 20
percent
by
volume.
In some
areas,
most notably
near oce
-
anic fracture zones,
the
oceanic crust
is
as
thin
as
3
km.
Oceanic plateaus
and
aseismic ridges
may
have
crustal
thicknesses greater
than 30
km.
Some
of
these appear to
represent large
volumes
of
material generated at oceanic
spreading centers or hotspots,
and
a few
seem
to
be
conti
-
nental
fragments. Although these
anomalously
thick
crust
regions constitute only about
10
percent
of
the area
of the
oceans,
they
may
represent
up
to
50
percent
of
the
total
volume
of
the oceanic crust. Islands,
island arcs
and
conti
-
nental
margins are collectively referred to as transitional
crust
and
generally range from
15
to
30
km
in thickness.
Continental crust generally ranges from
30
to
50
krn
thick,
Depth
(km:)
FIGURE
3
-
1
The
Preliminary
Reference Earth
Model
(PREM).
The
model is
anisotropic
in the
upper
220
km,
as shown in
Figure
3
-
3. Dashed
lines are
the
horizontal
components
of
the seismic
velocity
(after
Dziewonski
and
Anderson,
198
1).
but
thicknesses up
to
80
km
are
reported
in
some conver
-
gence regions.
Based
on
geological and
seismic
data,
the
main
rock type in the
upper
continental crust is granodiorite
or
tonalite
in
composition.
The
lower crust
is probably
dio-
rite,
garnet granulite
and
amphibolite. The average com
-
position
of
the continental
crust is
thought
to
be
similar
to
andesite
or
diorite. The upper part
of
the continental crust
is enriched
in
such
"
incompatible
"
elements
as
potassium,
rubidium, barium,
uranium and
thorium
and has a marked
negative europium
anomaly
relative
to
the mantle.
Niobium
TABLE
3
-
1
Summary
of
Earth Structure
Region
Depth
Fraction
Fraction
of
Total
of
Mantle
(km)
Earth
Mass
and
Crust
Continental crust
0
-
50
0.00374
0.00554
Oceanic
crust
0
-
10
0.00099
0.00147
Upper
mantle
10
-
400
0.103
0.153
Transition
region
400
-
650
0.075
0.111
Lower
mantle
650
-
2890
0.492
0.729
Outer
core
2890
-
5150
0.308
-
Inner core
5150
-
6370
0.017
-
and
tantalum are
apparently depleted
relative
to
other
nor
-
mally
incompatible trace elements.
It has recently been rec
-
ognized
that
the terrestrial
crust
is unusually thin compared
to
the
Moon
and
Mars
and compared
to
the
amount
of
po
-
tential
crust
in the mantle.
This is related
to the
fact that
crustal
material converts to
dense
garnet
-
rich assemblages
at relatively shallow
depth. The
maximum theoretical thick
-
ness
of
material
with crust
-
like physical
properties is
about
50
-
60
km,
although the crust may
temporarily
achieve
somewhat
greater
thickness because
of
the
sluggishness
of
phase
changes
at
low
temperature.
Clomposition
M[ineralogically,
feldspar
(K
-
feldspar, plagioclase)
is the
most abundant
mineral in
the
crust,
followed
by quartz
and
hydrous minerals (such as
the
micas and amphiboles) (Table
3
-
2).
The minerals
of
the
crust and
some
of
their
physical
properties are
given in Table
3
-
3.
A
crust
composed
of
these
minerals will have an average density
of
about
2.7
g/cm
3
.
There is enough
difference in the velocities and
VJV,
ratios
of
the more
abundant minerals
that
seismic velocities pro
-
vide a
good
mineralogical discriminant.
One
uncertainty
is
the
amount
of
serpentinized
ultramafic
rocks in the lower
crust since serpentinization
decreases
the
velocity
of
olivine
THE
CRUST
47
TABLE
3
-
2
Crustal Minerals
-
--
Mineral
Composition
Range
of
Crustal
Abundances
(vol.
pct.)
Plagioclase
31
-
41
Anorthite
Ca(Al,Si,)O,
Albite
Na(AI,Si,)O,
Orthoclase
K
-
feldspar
K(AI,Si,)O,
Pyroxene
Hypersthene
(Mg,Fe2+)
Si0
,
Augite
Ca(Mg,Fe2+)(Si0,),
Olivine
(Mg,Fe2+),Si0,
Oxides
Sphene
CaTiSiO,
Allanite
(Ce,Ca,Y)(Al,Fe),(siO,),(OH)
Apatite
Ca,(PO,,CO,),(F,OH,Cl)
Magnetite
FeFe
,04
Ilmenite
FeTiO
,
to
crustal values. In some regions the seismic
Moho
may
not
be
at
the
base
of
the basaltic section but
at
the base
of
the serpentinized
zone
in
the mantle.
Estimates
of
the composition
of
the oceanic
and
con
-
tinental crust are given
in Table
3
-
4;
another that covers the
trace
elements
is given
in Table
3
-
5.
Note
that the continen
-
tal crust is richer
in
SiO,,
TiO,,
Al,O,,
Na,O
and
K,O
than
the
oceanic crust. This means that the continental crust
is
richer
in
quartz
and
feldspar
and is
therefore
intrinsicalIy
less
dense
than the oceanic crust.
The
mantle under stable
continental
-
shield crust has seismic properties that suggest
that
it
is intrinsically less dense than mantle elsewhere.
The
elevation
of
continents is controlled primarily
by
the
density
and
thickness
of
the crust
and
the intrinsic density
and temperature
of
the underlying mantle. It
is
commonly
assumed that
the
seismic
Moho
is also
the
petrological
Moho,
the boundary between sialic
or
mafic
crustal rocks
and
ultramafic mantle rocks.
However,
partial melting, high
pore
pressure
and
serpentinization can reduce the
velocity
of
mantle
rocks,
and
increased abundances
of
olivine
and
pyroxene
can
increase the
velocity of
crustal rocks. High
pressure also increases the
velocity of mafic
rocks,
by
the
gabbro
-
eclogite phase change,
to
mantle
-
like values.
The
increase
in velocity
from
"
crustal
"
to
"
mantle
"
values
in
regions
of
thick continental crust
may
be
due, at
least
in
part,
to the appearance
of
garnet
as
a stable phase.
The
situation
is
complicated further
by
kinetic considerations.
Garnet
is
a
common metastable
phase
in
near
-
surface
jntru-
sions
such
as
pegmatites and metamorphic teranes. On the
other
hand,
feldspar
-
rich rocks
may
exist
at
depths greater
than
the
gabbro
-
eclogite equilibrium boundary
if tempera
-
tures
are
so
low
that the reaction
is sluggish.
The
common assumption that
the
Moho is
a chemical
boundary is
in
contrast to
the
position
commonly
taken
with
regard
to
other
mantle discontinuities. It is almost
univer
-
TABLE
3
-
3
Average
Crustal Abundance,
Density
and
Seismic
Velocities
of
Maior
Crustal Minerals
Mineral
Volume
p
VP
vs
percent (g/cm
3
)
(km/s)
(km/s)
Quartz
12
2.65
6.05
4.09
K
-
feldspar
12
2.57
5.88
3.05
Plagioclase
39
2.64
6.30
3.44
Micas
5
2.8
5.6
2.9
Amphiboles
5
3.2
7.0
3.8
Pyroxene
11
3.3
7.8
4.6
Olivine
3
3.3
8.4
4.9
TABLE
3
-
4
Estimates
of
the Chemical Composition
of
the
Crust
(Weight
Percent)
Oxide
Oceanic
Continental Crust
Crust
(1)
(2)
(3)
SiO,
47.8
63.3
58.0
Ti02
0.59
0.6
0.8
A1z03
12.1
16.0
18.0
-
Fez03
-
1.5
FeO
9.0
3.5
7.5
MgO
17.8
2.2
3.5
CaO
11.2
4.1
7.5
Na,O
1.31
3.7
3.5
K
20
0.03
2.9
1.5
H
20
1
.O
0.9
-
(1)
Elthon
(1979).
(2)
Condie
(1982).
(3)
Tayor and
McLennan
(1985).
sally assumed that
the
major mantle
discontinuities
repre
-
sent equilibrium solid
-
solid phase changes
in
a homoge
-
neous
material. It
should
be
kept
in
mind
that chemical
changes
may
also occur
in
the
mantle. It
is hard
to
imagine
how
the Earth
could
have
gone
through
a high
-
temperature
TABLE
3
-
5
Composition
of
the
Bulk
Continental
Crust,
by
Weight
accretion
and
differentiation process
and
maintained
a ho
-
mogeneous
composition throughout. It
is probably
not
a co
-
incidence that the maximum
crustal
thicknesses are
close
to
the
basalt
-
eclogite boundary. Eclogite is
denser
than
peri-
dotite, at least
in the shallow
mantle,
and
will tend
to
fall
into
normal
mantle,
thereby
turning a phase
boundary
(ba
-
salt
-
eclogite)
into
a chemical boundary
(basalt
-
peridotite).
Seismic
Velocities
in
the
Crust and
Upper
Mantle
Seismic
velocities
in
the
crust
and
upper mantle are typi
-
cally determined
by
measuring the
transit time
between
an
earthquake
or
explosion
and
an
array
of
seismometers.
Crustal compressional
wave
velocities in
continents,
be
-
neath the sedimentary
layers,
vary
from about
5
kmls
at
shallow
depth
to
about
7
kmls
at a
depth
of
30
to
50
km.
The
lower velocities reflect
the
presence
of
pores
and
cracks
more than
the
intrinsic
velocities
of
the
rocks.
At
greater
depths
the pressure
closes cracks
and
the remaining pores
are fluid
-
saturated. These effects cause a considerable
in
-
crease
in velocity.
A
typical crustal velocity range
at depths
greater
than
1
km
is
6
-
7
kmls.
The corresponding
range
in
shear
velocity is about
3.5
to
4.0
kmls.
Shear
velocities can
be
determined from both body
waves and
the
dispersion
of
short
-
period surface
waves.
The top
of
the
mantle
under
SiO,
TiO,
'4120,
FeO
MgO
CaO
Na20
K,O
Li
Be
B
Na
Mg
A1
Si
K
Ca
Sc
Ti
v
Cr
Mn
Fe
57.3
pct.
0.9
pct.
15.9
pct.
9.1
pct.
5.3
pct.
7.4
pct.
3.1
pct.
1.1
pct.
13
PPm
1.5
pprn
10
PPm
2.3
pct.
3.2
pct.
8.41
pct.
26.77
pct.
0.91
pct.
5.29
pct.
30
PPm
5400
pprn
230
pprn
185
pprn
1400
pprn
7.07
pct.
29
PPm
105
ppm
75
PPm
80
PPm
18
PP~
1.6
pprn
1.0
pprn
0.05
pprn
32
PPm
260
pprn
20
PPm
100
pprn
11
PP
1
PP
1
PP~
80
PP~
98
PP~
50
PP~
2.5
pprn
0.2
pprn
1
PPm
250
pprn
16
PPm
33
PPm
3.9
pprn
16
PPm
3.5
pprn
1.1
pprn
3.3
pprn
0.6
pprn
3.7
pprn
0.78
pprn
2.2
pprn
0.32
pprn
2.2
pprn
0.30
pprn
3.0
ppm
1
PPm
1
PPm
0.5
ppb
0.1
ppb
3
PP~
360
ppb
8
PP"
60
PP~
3.5
pprn
U
0.91
ppm
Taylor
and
McLennan
(1985).
THE
SEISMIC
LITHOSPHERE
OR
LID
49
continents usually
has
velocities
in
the range
8.0
to
8.2
kmls
for
compressional
waves
and
4.3
and
4.7
kmls
for
shear waves.
The compressional
velocity near
the base
of
the oce
-
anic crust usually falls in the range
6.5
-
6.9 kmls.
In some
areas a
thin
layer at the
base
of
the crust
with
velocities as
high
as
7.5
kmls
has been
identified. The oceanic upper
mantle
has
a P
-
velocity
(P,)
that
varies from about
7.9
to
8.6
kmls.
The
velocity
increases with oceanic age, because
of
cooling,
and varies with
azimuth, presumably due
to
crystal orientation. The fast direction is generally close to
the inferred spreading direction. The average velocity is
close to
8.2
kmls,
but young
ocean
has
velocities as
low as
7.6
his.
Tectonic regions also
have low
velocities.
Since water does
not
transmit shear
waves and
since
most
velocity measurements use explosive sources, it
is dif
-
ficult
to
measure the shear velocity in the oceanic crust
and
upper mantle. There are therefore fewer measurements
of
shear
velocity, and
these
have
higher uncertainty
and
lower
resolution
than
those for P
-
waves. The shear
velocity in
-
creases from about
3.6
-
3.9
to
4.4
-
4.7
km/s
from the base
of
the crust to the top
of
the mantle.
Ophiolite sections found at some continental margins
are thought
to
represent upthrust or
obducted
slices
of
the
oceanic crust
and
upper mantle. These sections grade down
-
ward
from
pillow
lavas to
sheeted
dike swarms, intrusives,
pyroxene
and
olivine gabbro,
layered
gabbro
and
peridotite
and,
finally,
harzburgite
and
dunite. Laboratory velocities
in
these rocks are
given
in Table
3
-
6. There
is good
agree
-
ment
between
these velocities
and
those actually observed
in the oceanic crust
and
upper mantle. The sequence
of
ex
-
trusive~,
intrusives
and
cumulates is consistent
with what
is
expected
at a midocean
-
ridge
magma
chamber. Many
ophi-
olites apparently represent oceanic crust formed near island
arcs
in
marginal basins. They
might
not
be
typical
of
crustal
sections formed at mature midoceanic spreading centers.
Marginal basin basalts,
however,
are
very
similar to
mid-
ocean
-
ridge basalts, at least
in
major
-
element chemistry,
and
the bathymetry, heat
flow
and
seismic crustal structure
TABLE
3
-
6
Density, Compressional
Velocity
and
Shear
Velocity
in Rock
Types
Found
in Ophiolite Sections
Rock
Type
P
v~
V,
Poisson's
(g/cm
3
)
(km/s)
(kmls)
Ratio
Metabasalt
2.87
6.20
3.28
0.31
Metadolerite
2.93
6.73
3.78
0.27
Metagabbro
2.95
6.56
3.64
0.28
Gabbro
2.86
6.94
3.69
0.30
Pyroxenite
3.23
7.64
4.43
0.25
Olivine gabbro
3.30
7.30
3.85
0.32
Harzburgite
3.30
8.40
4.90
0.24
Dunite
3.30
8.45
4.90
0.25
Salisbury
and
Christensen
(1978),
Christensen
and
Smewing
(1981).
in marginal basins are similar to
values
for the major
ocean
basins.
The velocity contrast
between
the
lower
crust
and
up
-
per mantle is commonly smaller beneath
young
orogenic
areas
(0.5
to
1.5
kmls)
than
beneath cratons
and
shields
(1
to 2
kmls).
Continental rift systems
have thin
crust (less
than 30 km)
and
low
P,
velocities (less
than
7.8
kmls).
Thinning
of
the crust
in
these regions appears to take place
by
thinning
of
the
lower
crust.
In
island arcs the crustal
thickness ranges from about 5
km
to 35 km.
In areas
of
very
thick crust
such
as in
the Andes
(70 km)
and
the
Himalayas
(80 km), the thickening occurs primarily
in the
lower
crustal
layers. Paleozoic orogenic areas
have about
the same range
of
crustal thicknesses
and
velocities as platform areas.
THE SEISMIC
LITHOSPHERE
OR
LID
Uppermost mantle velocities are typically
8.0
to
8.2 kmls,
and
the spread is about
7.9
-
8.6 kmls.
Some
long
refraction
profiles
give evidence for a deeper
layer
in
the lithosphere
having a velocity
of
8.6
km/s.
The seismic lithosphere, or
LID,
appears to contain at least
two
layers.
Long
refraction
profiles
on continents
have
been
interpreted
in
terms
of
a
laminated
model
of
the upper 100
km with
high
-
velocity
layers,
8.6
-
8.7
km/s
or
higher, embedded
in
"
normal
"
material (Fuchs, 1977). Corrected
to
normal conditions
these velocities
would be about
8.9
-
9.0
kmls.
The P
-
wave
gradients are often
much
steeper
than
can
be
explained
by
self
-
compression. These
high
velocities require oriented ol
-
ivine or large amounts
of
garnet. The detection
of
7
-
8 per
-
cent azimuthal anisotropy for
both
continents
and oceans
suggests that the shallow mantle at least contains
oriented
olivine. Substantial anisotropy
is
inferred
to
depths
of
at
least
50
km
depth
in
Germany (Bamford, 1977).
The average
values
of
V,
and
V,
at 40 km
when
cor
-
rected
to
standard conditions are
8.72
kmls
and
4.99
km/s,
respectively. The corrections for temperature
amount to
0.5
and
0.3
kmls,
respectively, for
V,
and
V,.
The pressure cor
-
rections are
much
smaller. Short
-
period surface
wave
data
have
better resolving
power
for
V,
in
the
LID.
Applying
the
same corrections to surface
wave
data (Morris
and
others,
1969),
we
obtain
4.48
-
4.55
kmis
and
4.51
-
4.64
kmls
for
5
-
Ma
-
old
and
25
-
Ma
-
old oceanic lithosphere.
Presumably,
velocities can
be
expected to increase further
for
older re
-
gions. A
value
for
V,
of
8.6
kmls
is commonly
observed
near
40
km depth
in
the
oceans. This corresponds
to about
8.87
kmls
at standard
conditions. These
values can be
com
-
pared
with
8.48
and
4.93
kmls
for olivine
-
rich aggregates.
Eclogites
have
V,
and
V,
as
high
as
8.8
and
4.9
kmls
in
certain directions
and
as
high
as
8.61
and
4.86
kmls
as
average values.
All
of
the
above
suggests that corrected velocities
of
at
least
8.6
and
4.8
km/s,
for
V,
and
V,,
respectively, occur
in
the lower lithosphere,
and
this requires substantial amounts
of
garnet at relatively
shallow
depths.
At
least
26
percent
garnet is required
to
satisfy the compressional velocities.
The
density
of
such
an
assemblage
is
about
3.4
g/cm
3
. If
one is to
honor
the higher seismic velocities,
even
greater
proportions
of
garnet are required. The lower lithosphere
may
therefore
be
gravitationally unstable
with
respect to the
underlying mantle, particularly
in
oceanic regions. The
up
-
per mantle under shield regions is consistent with a very
olivine
-
rich peridotite
which
is buoyant and therefore stable
relative to
"
normal
"
mantle.
Most refraction profiles, particularly at
sea,
sample
only the uppermost lithosphere.
P,
velocities
of
8 .O
-
8.2
kmls
are consistent
with
peridotite or harzburgite, thought
by
some to
be
the refractory residual after basalt removal.
Anisotropies are also appropriate for olivine
-
pyroxene as
-
semblages. The sequence
of
layers, at least in oceanic re
-
gions, seems to
be
basalt, peridotite, eclogite.
Anisotropy
of
the
upper
mantle is a potentially useful
petrological constraint, although it can also
be
caused
by
organized heterogeneity, such
as
laminations or parallel
dikes
and
sills
or
aligned partial melt zones,
and
stress
fields.
Under
oceans the uppermost mantle, the
P,
region,
exhibits
an
anisotropy
of
7 percent (Morris
and
others,
1969). The fast direction
is
in the direction
of
spreading,
and the magnitude
of
the anisotropy
and
the high velocities
of
P,
arrivals suggest
that
oriented olivine crystals control
the elastic properties. Pyroxene exhibits a similar
anisot-
ropy, whereas
garnet
is
more isotropic. The preferred ori
-
entation is presumably due to the emplacement or freezing
mechanism, the temperature gradient or
to
nonhydrostatic
stresses. A peridotite layer at the top
of
the oceanic mantle
is consistent
with
the observations.
The average anisotropy
of
the upper mantle is
much
less
than
the
values
given above. Forsyth
(1975) studied the
dispersion
of
surface
waves
and
found shear
-
wave
anisotro-
pies, averaged
over
the upper mantle,
of
2 perent. Shear
velocities in the LID vary from
4.26
to 4.46
kmls,
increas
-
ing with age; the higher
values
correspond to a lithosphere
10
-
50
Ma
old.
This can
be
compared
with
shear
-
wave
velocities
of
4.30
-
4.86
kmls
and
anisotropies
of
1
-
4.7
percent found
in
relatively unaltered eclogites (Manghnani,
et
al,
1974). The compressional velocity range
in
the same
samples is
7.90
-
8.61
kmls,
reflecting
the large amounts
of
garnet. Surface
waves
exhibit
both
azimuthal
and
polariza
-
tion anisotropy for at least the upper 200 km
of
the mantle.
On
the basis
of
data available
at
the time,
Ringwood
suggested
in
1975 that the
Vp/Vs
ratio
of
the lithosphere
was
smaller
than measured
on
eclogites.
He
therefore ruled out
eclogite as
an
important constituent
of
the lithosphere. The
eclogite
-
peridotite controversy regarding the composition
of
the suboceanic mantle is
long
standing and still
unre
-
solved.
Newer
and
more complete data
on
the
Vp/Vs
ratio
in
eclogites
show
that the
high-V,/V,
eclogites are generally
of
low
density
and
contain plagioclase or olivine. The
higher-
density eclogites are consistent
with
the properties
of
the
lower lithosphere. Garnet
and
clinopyroxene
may
be
impor
-
tant components
of
the lithosphere.
A
lithosphere composed
primarily
of
olivine
and
(Mg,Fe)SiO,,
that is, pyrolite
or
lherzolite, does
not
satisfy the seismic data for the
bulk
of
the lithosphere. The lithosphere, therefore,
is not just
cold
mantle, or a thermal boundary layer alone.
Oceanic crustal
basalts
repiesent
only part
of
the
ba
-
saltic fraction
of
the upper mantle. The peridotite
layer rep
-
resents depleted mantle or refractory cumulates but
may
be
of
any
age.
Basaltic material
may
also
be
intruded
at depth.
It is likely that the upper mantle is layered,
with
the
vola-
tiles
and
melt products concentrated
toward its
top.
As
the
lithosphere cools,
this
basaltic material
is incorporated onto
the base
of
the plate,
and
as the plate thickens
it eventually
transforms to
eclogite,
yielding
high velocities
and
increas
-
ing
the thickness
and
mean
density
of
the oceanic plate.
Eventually the plate becomes denser
than
the underlying
asthenosphere, and conditions become appropriate for sub
-
duction
or
delamination.
O'Hara
(1968)
argued
that erupted
lavas
are not the
original liquids
produced
by
partial
melting
of
the
upper
mantle,
but are residual liquids from processes that
have
left
behind
complementary eclogite accumulates in the upper
mantle. Such a model is consistent
with
the seismic
obser-
Previous estimates
of
seismic
Age
of
oceanic
lithosphere
(Ma)
FIGURE
3-2
Tlne
thickness
of the
lithosphere
as
determined
from
flexural
loading
studies
and
surface waves.
The
upper edges
of the open
boxes
gives
the
thickness
of
the
seismic
LID
(high
-
velocity
layer,
or seismic
lithosphere).
The lower
edge gives
the
thickness
of the
mantle
LID
plus
the oceanic crust
(Regan
and
Anderson,
1984).
Tlne
triangle
is
a
refraction
measurement
of oceanic seismic
litho-
splhere
thickness
(Shimamura
and
others, 1977).
The
LID
under
continental shields
is about
150
km
thick
(see
Figure
3
-
4).
thick
(see
Figure
3
-
4).
vations
of
high velocities for
P,
at
midlithospheric depths
and with
the
propensity
of
oceanic lithosphere
to
plunge
into the asthenosphere. The latter observation suggests that
the
average density
of
the
lithosphere
is greater than
that
of
the asthenosphere.
The thickness
of
the seismic lithosphere, or
high-
velocity
LID,
is
about
150
km
under continental shields.
Some surface
-
wave results give a
much
greater thickness.
A
thin
low
-
velocity
zone
(LVZ)
at depth,
as
found from
body
-
wave
studies,
however,
cannot
be
well
resolved
with
long
-
period surface
waves.
The
velocity reversal between
about 150 and
200
km
in shield areas
is
about the depth
inferred for kimberlite genesis,
and
the
two
phenomena
may
be
related.
There
is very
little information about the deep oceanic
lithosphere
from
body
-
wave
data. Surface
waves
have
been
used
to infer a thickening
with
age
of
the oceanic litho
-
sphere to depths greater
than
100
km
(Figure 3
-
2).
How
-
ever,
when
anisotropy is
taken
into
account,
the
thickness
may
be
only about
50
krn
for
old
oceanic lithosphere
(Regan
and
Anderson, 1984). This is
about
the thickness inferred
for
the
"
elastic
"
lithosphere from
flexural bending
studies
around oceanic islands
and
at trenches.
The seismic velocities
of
some upper
-
mantle minerals
and.rocks
are given in
Tables
3
-
7 and
3
-
8,
respectively.
Garnet
and
jadeite
have
the highest velocities,
clinopyrox-
ene
and
orthopyroxene the lowest. Mixtures
of
olivine
and
orthopyroxene (the peridotite assemblage) can
have
veloci
-
ties similar to mixtures
of
garnet-diopside-jadeite
(the
eclo-
gite assemblage). Garnet
-
rich assemblages,
however, have
velocities higher
than
orthopyroxene
-
rich assemblages. The
TABLE
3
-
7
Densities and Elastic
-
wave Velocities
in
Upper
-
mantle Minerals
Mineral
P
v~
vs
vp
/vs
(g/cm3)
(kmls)
-
Olivine
Fo
3.214
8.57
5.02
1
.?I
Fo93
3.311
8.42
4.89
1.72
Fa
4.393
6.64
3.49
1.90
Pyroxene
En
3.21
8.08
4.87
1.66
En
530
3.354
7.80
4.73
1.65
Fs
3.99
6.90
3.72
1.85
Di
3.29
7.84
4.51
1.74
Jd
3.32
8.76
5.03
1.74
Garnet
PY
3.559
8.96
5.05
1.77
A1
4.32
8.42
4.68
1.80
Gr
3.595
9.31
5.43
1.71
Kn
3.85
8.50
4.79
1.77
An
3.836
8.51
4.85
1.75
Uv
3.85
8.60
4.89
1.76
Sumino
and
Anderson
(1984).
TABLE
3
-
8
Densities and Elastic
-
wave
Velocities
of
Upper
-
mantle Rocks
Rock
P
v~
vs
vplvs
Garnet
3.53
8.29
4.83
1.72
lherzolite
3.47
8.19
4.72
1.74
'
3.46
8.34
4.81
1.73
3.31
8.30
4.87
1
.70
Dunite
3.26
8.00
4.54
1.76
3.31
8.38
4.84
1.73
Bronzitite
3.29
7.89
4.59
1.72
3.29
7.83
4.66
1.68
Eclogi
te
3.46
8.61
4.77
1.81
3.61
8.43
4.69
1.80
3.60
8.42
4.86
1.73
3.55
8.22
4.75
1.73
3.52
8.29
4.49
1.85
3.47
8.22
4.63
1.78
Jadeite
3.20
8.28
4.82
1.72
Clark
(1966),
Babuska
(1972),
Manghnani
and
others
(1974),
Jordan
(1979).
v@,
ratio
is greater
for
the
eclogite minerals
than
for
the peridotite minerals. This ratio
plus
the anisotropy
are
useful diagnostics
of
mantle
mineralogy. High
velocities
alone
do
not necessarily discriminate
between
garnet
-
rich
and olivine
-
rich assemblages. Olivine
is
very
anisotropic,
having
compressional velocities
of
9.89,
8.43
and
7.72
km/
s
along
the principal crystallographic axes. Orthopyroxene,
likewise,
has
velocities ranging from
6.92
to
8.25
kmls,
depending
on
direction.
In
natural olivine
-
rich aggregates
(Table
3
-
9), the
maximum
velocities
are about
8.7
and
5.0
krnls
for P
-
waves
and
S
-
waves,
respectively.
With
50
per
-
cent orthopyroxene the velocities are reduced
to
8.2
and
4.85
km/s,
and
the composite
is nearly
isotropic. Eclogites
are also
nearly
isotropic.
The
"
standard model
"
for
the
oceanic lithosphere
as
-
sumes 24 km
of
depleted peridotite, complementary
to
and
contemporaneous
with
the basaltic crust,
between
the crust
and the
presumed fertile peridotite upper mantle. There
is
no direct evidence for
this
hypothetical model. The
lower
oceanic lithosphere
may
be much
more basaltic or eclogitic
than
in
this
simple model.
THE
LOW
-
VELOCITY
ZONE
OR
LVZ
A
region
of
diminished
velocity
or negative
velocity
gradi
-
ent in the upper
mantle was
proposed
by
Beno
Gutenberg
in
1959.
Earlier, just after
isostasy had been
established, it
had
been concluded that a
weak
region
underlay the rela
-
tively strong lithosphere. This
has been
called the astheno
-
sphere. The discovery
of
a low
-
velocity zone strengthened
TABLE
3
-
9
Anisotropy
of
Upper
-
mantle Rocks
Mineralogy
Direction
Vp
vs,
vs2
vplvs
Peridotites
100
pct.
01
I
8.7
5.0
4.85
1.74
1.79
2
8.4
4.95
4.70
1.70
1.79
3
8.2
4.95
4.72
1.66
1.74
70
pct.
01,
1
8.4
4.9
4.77
1.71
1.76
30
pct.
opx
2
8.2
4.9
4.70
1.67
1.74
3
8.1
4.9
4.72
1.65
1.72
100
pct.
opx
1
7.8
4.75
4.65
1.64
1.68
2
7.75
4.75
4.65
1.63
1.67
3
7.78
4.75
4.65
1.67
1.67
EcUogites
51
pct.
ga,
1
8.476
4.70
1.80
23
pct.
cpx,
2
8.429
4.65
1.81
24
pct.
opx
3
8.375
4.71
1.78
47
pct.
ga,
1
8.582
4.91
1.75
45
pct.
cpx
2
8.379
4.87
1.72
3
8.30
4.79
1.73
46
pct.
ga,
1
8.31
4.77
1.74
37
pct.
cpx
2
8.27
4.77
1.73
3
8.11
4.72
1.72
Manghnani and others
(1974),
Christensen and Lundquist
(1982).
the concept
of
an
asthenosphere,
even though
a weak layer
is not
necessarily a
low
-
velocity
layer.
Gutenberg
based
his conclusions primarily on ampli
-
tudes
and
apparent
velocities
of
waves
from earthquakes
in
the
vicinity
of
the low
-
velocity
zone.
He
found that at dis
-
tances from
about
lo
to
15"
the
amplitudes
of
longitudinal
waves
decrease
about
exponentially
with
distance.
At
15
"
they
increase
suddenly
by
a factor
of
more than
10
and then
decrease at greater distances. These
results
can
be
explained
in
terms
of
a low
-
velocity region,
which
defocuses seismic
energy,
underlain
by
a higher gradient
that
serves to focus
the
rays.
Most
recent
models
of
the
velocity
distribution in the
upper
mantle
include a
region
of
high
gradient
between
250
and
350
km
depth. Lehmann (1961) interpreted her results
for
several
regions in terms
of
a discontinuity at 220
km
(sometimes
called
the Lehmann discontinuity),
and many
subsequent studies give high
-
velocity gradients near this
depth.
It is difficult to study
details
of
the
velocity
distribution
in and just
below
a low
-
velocity zone,
and
it is still not clear
if the base
of'the
low
-
velocity
zone
is
gradual
or
abrupt.
Reflections have
been
reported from depths between 190
and
250
km
by
a number
of
authors (Anderson, 1979). This
situation is further complicated
by
the extreme lateral
het
-
erogeneity
of
the
upper 200
km
of
the mantle. This
region
is also
low
Q
(high attenuation)
and
anisotropic. Some re
-
cent results are
shown
in
Figures
3
-
3
and
3
-
4.
Various
interpretations
have
been
offered
for the
low-
velocity
zone. This is
undoubtedly
a region
of
high thermal
gradient, the boundary
layer between
the near surface where
heat
is
transported
by
conduction
and
the deep interior
where heat is transported
by
convection.
If the temperature
gradient is
high
enough, the
effects
of
pressure can
be
over
-
come
and velocity
can decrease
with
depth. It can
be
shown,
however,
that a
high
temperature gradient alone
is
not
an
adequate explanation. Partial melting
and
dislocation
Velocity
(km/s)
FIGURE
3
-
3
Velocity
-
depth profiles
for
the average
Earth, as
determined from
surface waves
(Regan
and
Anderson,
1984).
From
left to
right,
the
graphs
show
P
-
wave
velocities (vertical and horizontal),
S
-
wave velocities (vertical and horizontal),
and
the anisotropy pa
-
rameter
7
(see Chapter
15),
where
l
represents isotropy.