Chapter
4
The
outer
shells
of
Earth
When
I
use
a
word,
it
means
just
what
I
choose
it
to
mean
-
neither
more
nor
less.
The
question
is,
whether
you
CAN
make
words
mean
so
many
different
things.
The
question
is,
which
is
to
be
master-
that's
a l
l.
Humpty
Dumpty
and
Alice
Plate
tectonics
involves
the
concepts
of
plates,
lithospheres,
cratonic
keels
and
thermal
bound-
ary
layers;
these
are
not
equiva
l
ent
concepts.
Lithosphere
means
rocky
shell
or
strong
layer
.
It
can
support
sign
i
ficant
geologic
loads,
such
as
moun-
tains,
and
can
bend,
as
at
trenches,
without
sig-
nificant
time-dependent
deformation;
it
behaves
elastically.
Plates
are
not
necessarily
strong
or
elastic.
Thermal
boundary
layers
conduct
heat
out
of
the
underlying
regions;
they
are
defined
by
a
particular
thermal
gradient
.
P
la
tes
A
plat
e
is
a
region
of
the
Earth's
surface
that
translates
coherently.
The
word
plat
e
as
in
plate
t ectonics
implies
strength,
brittleness
and
perma-
nence
and
often
the
adjectives
rig
id
and
elas-
tic
are
appended
to
it.
But
plates
are
collages
,
held
together
by
stress
and
adjacent
portions
rather
than
by
intrinsic
stren
g
th
.
Plates
break
at
suture
zones
(former
plate
boundaries),
frac-
ture
zones
and
subplate
boundaries,
often
gener-
ating
volcanic
chains
in
the
process.
A
plate
can
be
rigid
in
the
sense
that
relative
plate
motions
can
be
described
by
rotations
about
Euler
poles
on
a
sphere
but
can
still
have
meter-wide
cracks
,
which
is
all
that
is
needed
to
create
volcanic
chains
from
the
hot
underlying
mantle
and
its
low-viscosity
magmas
.
What
is
meant
by
rigid-
ity
is
relative
coherence
in
motions,
not
absolute
strength.
Because
rocks
are
weak
under
tension,
the
conditions
for
the
existence
of
a
plate
proba-
bly
involve
the
existence
of
lateral
compressive
forces.
Plates
have
been
described
as
rigid
but
this
implies
long-term
and
long-range
strength.
They
are
better
described
as
coherent
entities
organized
by
stress
fields
and
rheology.
The
corol-
lary
is
that
volcanic
chains
and
plate
boundaries
are
regions
of
extension.
Plates
possibly
organize
themselves
so
as
to
minimize
dissipation
.
The
term
plate
itself
has
no
agreed-upon
for-
mal
definition.
If
plate
is
defined
operationally
as
that
part
of
the
outer
shell
that
moves
coherently
then
several
interpretations
are
possible.
(1)
Plates
are
strong
and
rigid
(the
conventional
interpretation).
(2)
Plates
are
those
regions
defined
by
lat-
eral
compression
since
plate
boundaries
are
formed
by
lateral
extension.
(3)
Plates
move
coherently
because
the
parts
experience
similar
forces
or
constraints.
With
the
first
definition,
if
there
is
to
be
spread-
ing
and
volcanism,
local
heating
or
stretching
36
T
':"lE
OUTER
SHELLS
OF
EARTH
must
overcome
the
local
strength;
this
reason-
ing
spawned
the
plume
hypothesis.
With
the
second
definition
the
global
stress
field,
dictated
by
plate
boundary
and
subplate
conditions
,
and
cooling
plates,
controls
the
locations
of
stress
conditions
appropriate
for
the
formation
of
dikes
and
vol-
canic
chains,
and
incipient
plate
boundaries
;
the
underlying
mantle
is
already
at
or
near
the
melt-
ing
point.
This
is
the
plate
hypothesis
.
Plates
are
not
permanent;
they
are
tempo-
rary
alliances
of
subplates
. Global
plate
reorga-
nization
processes
episodically
change
the
orien-
tations
of
spreading
centers,
the
directions
and
speeds
of
plates,
and
redefine
the
plates.
Plates
annex
and
lose
territory
to
adjacent
plates
and
they
break
up
or
coalesce
. New
plate
bound-
aries
do
not
form
all
at
once
but
evolve
as
age-
progressive
chains
of
volcanoes.
Volcanic
chains
can
also
be
extinguished
if
lateral
compression
takes
over
from
local
extension.
Volcanism
can
be
turned
on
and
off
by
changing
stress
but
it
is
not
so
easy
to
turn
off
plume
volcanism,
or
to
suddenly
reduce
the
temperature
of
the
mantle.
I
mportant
aspects
of
plate
tectonics
are
the
necessity
for
ridges
and
trenches
to
migrate,
for
triple
junctions
and
boundary
conditions
to
evolve
,
and
for
plates
to
interact
and
to
reconfig-
ure
when
boundary
conditions
change
.
Second
order
features
of
plates
and
plate
boundaries
(e
.g.
fracture
zones,
a
ccre
t e d t
e
rr
a n
es
,
trans-
form
faults,
broad
diffuse
zones,
swells,
sutures,
lithospheric
architecture
and
microplates)
and
boundary
reorganizations
are
actually
intrinsic
and
provide
the
key
for
a
more
general
view
of
plate
tectonics
than
contained
in
the
rig id
plate-
fixed
hot
s
pot
scheme
.
T he ma
ny
li
thosp
he
res
The
lithosphere
is
that
part
of
the
cold
outer
shell
of
the
Earth
that
can
support
stresses
elas-
tically.
The
lithosphere
is
defined
by
its
rheolog-
ical
behavior.
There
are
other
elements
of
the
outer
shell
that
involve
lateral
motions,
buoy-
ancy,
chemistry,
mineralogy
or
conductivity
and
these
may
or
may
not
be
part
of
the
lithosphere
.
Lith
osph
ere
is
not
the
same
as
thermal
boundary
la
yer
or
plat
e.
Since
mantle
silicates
flow
readily
at
high
temperatures
and
flow
more
rapidly
at
high
stress,
the
lithosphere
appears
to
be
thicker
at
low
stress
levels
and
short
times
than
it
does
for
high
stress
levels
and
long
times.
Thus,
the
elastic
lithosphere
is
thick
when
measured
by
seismic
or
postglacial-rebound
techniques
. At
longer
times
the
lower
part
of
the
instantaneous
elastic
litho-
sphere
relaxes
and
the
eff
ec
ti
ve
el
ast
i c
thi
c
kn
ess
decreases
.
Thus,
the
elastic
litho-
sphere
is
relatively
thin
for
long-lived
loads
such
as
seamounts
and
topography.
Estimates
of
the
flexural
thickness
of
the
lithosphere
range
from
10-35
km
for
loads
having
durations
of
millions
of
years.
A
more
complete
definition
of
the
litho-
sphere
is
that
part
of
the
crust
and
upper
mantl
e
that
deforms
elastically
for
the
load
and
time
scal
e in
qu
estion
.
The
viscosity
and
strength
of
the
man-
tle
depend
on
composition
-
including
water
content
-
mineralogy
and
crystal
orientation
as
well
as
on
temperature
and
stress.
If
the
upper
mantle
is
compositionally
layered
,
then
the
lithosphere-asthenosphere
boundary
may
be
controlled
by
factors
other
than
tempera-
ture.
For
example
,
if
the
subcrustal
l
ayer
is
dry
olivine-rich
harzburgite
,
it
may
be
stronger
at
a
given
temperature
than
a
damp
peridotite
,
or
a
clinopyroxene-garnet-rich
layer
.
If
the
latter
is
weak
enough,
the
lithosphere-asthenosphere
boundary
may
represent
a
chemical
boundary
rather
than
an
isotherm.
Likewise,
a
change
in
the
preferred
orientation
of
the
dominant
crystalline
species
may
also
markedly
affect
the
creep
resistance.
The
boundary
may
represent
a
dehydration
boundary
-wet
minerals
are
weak.
The
effective
elastic
thickness
of
the
lithosphere
depends
on
many
parameters
but
these
do
not
necessarily
include
the
parameters
that
define
plates,
thermal
boundary
layers
and
cratonic
keels
.
The
layer
that
translates
coherently,
the
plate
of
plate
tectonic
s,
is
often
taken
to
be
identical
with
the
elastic
lithosph
e
re.
This
is
probably
a
valid
approximation
if
the
stresses
and
time
scales
of
the
experiment
that
is
used
to
define
the
flexu-
ral
thickness
are
similar
to
the
stresses
and
time
scales
of
plate
tectonics
.
It
must
be
kept
in
mind
,
however,
that
mantle
silicates
are
anisotropic
in
their
flow
and
thermal
characteristics,
and
that
the
stresses
involved
in
plate
tectonics
may
have
different
orientations
and
magnitudes
than
the
stresses
involved
in
surface
loading
experiments.
We
do
not
know
the
thickness
of
the
plate
or
how
well
it
is
coupled
to
the
underlying
mantle.
We
do
not
even
know
the
sign
of
the
basal
drag
force.
In
a
convecting
or
cooling
mantle
there
is
a
surface
thermal
boundary
layer
(TBL)
through
which
heat
must
pass
by
conduction.
The
thickness
of
the
thermal
boundary
layer
is
controlled
by
such
parameters
as
conductivity
and
heat
flow
and
is
not
related
in
a
simple
way
to
the
thickness
of
the
elastic
layer
or
the
plate.
Since
temperature
increases
rapidly
with
depth
in
the
conduction
layer,
and
viscosity
decreases
rapidly
with
temperature,
the
lower
part
of
the
boundary
layer
probably
lies
below
the
elastic
lithosphere;
that
is,
only
the
upper
part
of
the
thermal
boundary
layer
can
support
large
and
long-lived
elastic
stresses.
Unfortunately,
the
con-
duction
layer
too
is
often
referred
to
as
the
litho-
sphere.
In
a
chemically
layered
Earth
there
can
be
TBLs
between
internal
layers.
These
TBLs
act
as
thermal
bottle-necks
and
slow
down
the
cooling
of
an
otherwise
convective
mantle.
Continental
cratons
have
high
seismic
velocity
roots,
or
keels,
extending
to
200
-
300
km
depth.
These
are
referred
to
as
archons.
They
persist
because
of
low
density
and
high
viscosity,
and
because
they
are
protected
from
high
stress.
They
are
more
than
just
part
of
the
strong
outer
shell.
These
keels
can
last
for
billions
of
years.
Most
models
of
the
Earth's
mantle
have
an
upper-mantle
low-velocity
zone,
LVZ,
overlain
by
a
layer
of
higher
velocities,
referred
to
as
the
LID.
The
LID
is
also
often
referred
to
as
the
litho-
sphere.
Seismic
stresses
and
periods
are
much
smaller
than
stresses
and
periods
of
geological
interest.
If
seismic
waves
measure
the
relaxed
modulus
in
the
LVZ
and
the
high-frequency
or
unrelaxed
modulus
in
the
LID,
then,
in
a
chem-
ically
homogenous
mantle,
the
LID
should
be
much
thicker
than
the
elastic
lithosphere.
If
the
LID
is
chemically
distinct
from
the
LVZ,
then
one
might
also
expect
a
change
in
the
long-term
rheological
behavior
at
the
interface
.
If
the
LID
and
the
elastic
lithosphere
turn
out
to
have
the
same
thickness,
then
this
would
be
an
argument
THE
MANY
LITHOS
P
HERE$
37
for
chemical,
water
or
crystallographic
control.
rather
than
thermal
control,
of
the
mechanical
properties
of
the
upper
mantle.
In
su
mmary,
the
following
'lithospheres'
appear
in
the
geodynamic
literature
('When
I make
a word
do
a lot
of
work
like
that,'
said
Humpty
Dumpty,
'I always
pay
it
extra
.'
).
(1)
The
elastic,
flexural
or
rheological
lithosphere.
This
is
the
closest
to
the
classical
definition
of
a
rocky,
or
strong,
outer
shell.
It
can
be
defined
as
that
part
of
the
crust
and
upper
mantle
that
supports
elastic
stresses
of
a
given
size
for
a
given
period
of
time.
The
thickness
of
this
lithosphere
depends
on
stress
and
load
duration.
(2)
The
plate.
This
is
that
part
of
the
crust
and
upper
mantle
that
translates
coherently
in
the
course
of
plate
tectonics
.
The
thickness
of
the
plate
may
be
controlled
by
chemi-
cal
or
buoyancy
considerations
or
by
stress,
as
well
as
by
temperature,
but
there
is
no
known
way
to
measure
its
thickness.
Plates
are
ephemeral.
(3)
The
chemical
or
compositional
lithosphere.
The
density
and
mechanical
properties
of
the
lithosphere
are
controlled
by
chemical
com-
position
and
crystal
structure
as
well
as
tem-
perature.
If
chemistry
and
mineralogy
dom-
inate
,
then
the
elastic
lithosphere
and
LID
may
be
identical.
If
the
lithosphere,
below
the
crust,
is
mainly
depleted
peridotite
or
harzburgite,
it
may
be
buoyant
relative
to
the
underlying
mantle
. A
cratonic
root,
or
archon
,
is
often
called
the
continental
lithosphere
or
sub-
continental
lithospheric
mantle
(SCLM)
and
has
been
proposed
as
a
geochemical
reservoir
.
(4)
The
thermal
boundary
layer
or
conduction
layer
should
not
be
referred
to
as
the
litho-
sphere,
which
is
a
mechanical
concept,
but
if
the
lithospheric
thickness
is
thermally
controlled,
the
thickness
of
the
lithosphere
should
be
proportional
to
the
thickness
of
the
thermal
boundary
layer
.
If
TBLs
get
too
thick,
they
can
sink,
or
delaminate.
(5)
The
seismic
LID
is
a
region
of
high
seismic
velocity
that
overlies
the
low-velocity
zone.
At
high
temperatures
the
seismic
moduli
measured
by
seismic
waves
may
be
relaxed,