Temperature, Heat, and Energy of the Earth

1
Heat in the Earth

Generation and transfer of heat in the Earth control various physical and
chemical processes, including volcanoes, magmatic intrusions, earthquakes,
mountain building, metamorphism, and plate tectonics.

The activity of the lithosphere and asthenosphere

The development of the basic structure of the Earth
The differentiation of the Earth

Earth’s thermal budget

2
Energy Source
1. Heat arrives at the surface of the Earth from the Sun
• Arriving at the Earth is 2x1017 W
• Averaged over the surface this is 4x102 W/m2
• Yet most of the heat from the Sun is radiated back into space. But it
drives the surface water cycle, rainfall, erosion, and keeps the surface
temperature

2. Heat from the Earth’s Interior

• 4x1012 TW, 8x10-2 W/m2
• Governs the geological evolution of the Earth, controlling plate
tectonics, igneous activity, metamorphism, the evolution of the core,
and hence the Earth’s magnetic field

3
Heat Transfer Mechanisms
• Conduction
Transfer of heat through a material by atomic or molecular interaction within
the material

• Convection
Transfer of heat by the movement of the molecules themselves

• Radiation
Direct transfer of heat as electromagnetic radiation

• Advection
A special form of convection, corresponding to the horizontal/lateral heat
transfer --- wind
In the deep Earth, when a hot region is uplifted by tectonic events or by erosion
and isostatic rebound, heat (called advected heat) is physically lifted up with the
rocks.
4
Heat Transfer Mechanisms in the Earth
Within the Earth, heat moves predominantly by conduction through the
lithosphere (both oceanic and continental) and the solid inner core.

Although convection cannot take place in rigid solids, over geological

times the Earth’s mantle appears to behave as very-high-viscosity liquid,
which means that slow convection is possible in the mantle.

Actually, heat is generally thought to be transferred by convection

through most of the mantle as well as through the liquid outer core.

In addition, hot lava radiates heat, as do crystals at deep, hot levels in the
mantle, radiation is a minor factor in the transfer of heat within the Earth.

5
Thermal Conduction
Heat flows from hot things to cold things.

Heat flow is proportional to the temperature gradient in a material.

(T + T ) − T
Temperature gradient:
z
T
Heat flow: Q( z ) = −k
z

Q: W/m2
k: thermal conductivity, W/(m·oC)

Large temperature gradient – higher heat flow

Small temperature gradient – lower heat flow
6
Thermal Conduction
T
Q( z ) = −k
z
k: thermal conductivity of solids vary widely
Silver: 418 W/(m·oC) Magnesium: 159 W/(m·oC) glass: 1.2 W/(m·oC)
Rock: 1.7-3.3 W/(m·oC) wood: 0.1 W/(m·oC)

The minus sign arises because the temperature is increasing in the positive z
direction, while heat flows from a hot region to a cold region. It flows in the
negative z direction.
In the Earth, z denotes depth beneath the surface.
Since z increases downwards, a positive
temperature gradient means that there is a net
flow of heat upwards out of the Earth.
Measurements of temperature gradients and
thermal conductivity in near-surface boreholes
and mines can provide estimates of the rate of
loss of heat from the Earth.
7
Consider a small volume of height δz and cross-sectional area a.
Any change in temperature δT of this small volume in time δt depends on:
1. The flow of heat across the volume’s surface (net
flow is in or out)
2. The heat generated in the volume
3. The thermal capacity (specific heat) of the
material
The heat per unit time entering the volume across its
face at z is aQ(z), whereas the heat per unit time
leaving the element across its face at z + δz is aQ(z +
δz)

Expanding Q(z + δz) in a Taylor series gives:

Q (z ) 2  2Q
Q( z + z ) = Q( z ) + z + +
z 2 z 2

Ignore the higher order terms 8

In this case, the net gain of heat per unit time is:
Q
aQ( z ) − aQ( z + z ) = −az
z

Suppose that heat is generated in this volume element at a rate A per unit
volume per unit time. The total amount of heat generated per unit time is:

Aaz

*Remember, radioactive heat is the main internal heat source for the Earth as
a whole.

In this case, the total heat gains in per unit time to the first order is:
Q
Aaz − az
z

9
Specific heat cp
The amount of heat necessary to raise the temperature of 1 kg of the material by
1oC with unit J/(kg·oC).

Assuming the material has density ρ and specific heat cp, and undergoes a
temperature increase δT in time δt, the rate at which heat is gained is:

T
cP az
t
In this case, we have:
T Q
cP az = Aaz − az
t z
T Q
cP  = A−
t z

10
Combining the following two equations together:
T Q T
cP  = A− and Q( z ) = −k
t z z
We have:
T  2T
cP  = A+ k 2
t z

or T A k  2T
= +
t cP cP z 2

In three dimension:
T A k  2T  2T  2T
= + ( 2 + 2 + 2)
t cP cP x y z

T A k
or = +  2T
t cP cP
11
 T A k
= +  2T
t cP cP

The temperature in a column of rocks is controlled by several parameters.

Internal Parameters External Parameters

Conductivity heat flow into the column
Specific heat surface temperature
Density erosion or deposition
Radioactive heat generation

12
Now consider how the temperature of the small volume changes with time if it is
in relative motion through a region where the temperature varies with depth.

Assume that the volume element is moving with velocity uz in the z direction. In
this case, the temperature change in time t can be described as:

T z T T
= = uz
t t z z

The variation of temperature with time can be finally written as:

T A k  2T T
= + − u
t cP cP z 2 z
z

or
T A k
= +  2T − u  T
t cP cP

13
Two cases
1. For a steady-state situation, when there is no change in temperature with time.
A k
0= +  2T
cP cP

or A
 T =−
2

2. When there is no heat generation, we have:

T k
=  2T
t cP

This is called the diffusion equation.

14
Equilibrium Geotherms

Temperature-depth profiles within the Earth is called geotherms.

In a one-dimensional column with no erosion or deposition and a constant

heat flow, the column may eventually reach a state of thermal equilibrium in
which the temperature at any points is steady
-----The temperature-depth profile is called an equilibrium geotherm.

 2T A
= −
z 2 k
With known thermal boundary conditions, this second-order differential
equation can be solved.

15
Case one
 2T A
i. Temperature T = 0 at z = 0 =−
z 2
k
ii. Surface heat flow Q = -kǝT/ǝz = -Q0 at z = 0

*Surface heat flow is negative because heat is assumed to be flowing upwards

out of the medium which is in the negative z direction.
T Az Q0
=− +
z k k
Doing the integration again:
A 2 Q0
T =− z + z+c
2k k

Based on the boundary conditions, c = 0. We thus have:

A 2 Q0
T =− z + z
2k k
16
Case Two
i. Temperature T = 0 at z = 0
ii. Heat flow Q = -Qd at z = d

*This could be used to estimate equilibrium crustal geotherms if d was the

depth of the crust/mantle boundary and Qd was the mantle heat flow in
to the base of the crust.
T Az T Az Qd + Ad
=− + c1 =− +
z k z k k

Doing the integration again:

A 2 Qd + Ad
T =− z + z + c2
2k k
Based on the boundary conditions, c2 = 0. We thus have (0 ≤ z ≤ d):
A 2 Qd + Ad
T =− z + z
2k k 17
One-Layer Model
A 2 Qd + Ad
T =− z + z
2k k
Assuming (curve a)
50-km thick conductivity: k = 2.5 W/(m·oC) radioactive generation: A = 1.25 μW/m3
Heat flow into the base of the column: 21 × 10-3 W/m2
Shallow level: 30oC/km deep level: 15oC/km

Curve b: k = 1.7 W/(m·oC)

Geotherm: 45oC/km

Curve c: heat generation = 2.5 μW/m3

Geotherm: 50oC/km

Curve d: basal heat flow = 42 × 10-3 W/m2

Geotherm: 40oC/km

Curve e: basal heat flow = 10.5 × 10-3 W/m2

Geotherm: 27oC/km
18
Two-Layer Model
For the Earth, more realistic models have a layered crust with the heat generation
concentrated towards the top.

In the first layer, 0 ≤ z ≤ z1

T=0
 2T A1
=−
z 2
k
In the second layer, z1 ≤ z ≤ z2
 2T A2
= −
z 2 k

A1 2 Q2 A2 A1 z1
T =− z + ( + ( z 2 − z1 ) + )z 0 ≤ z ≤ z1
2k k k k
A2 2 Q2 A2 z 2 A1 − A2 2
T =− z +( + )z + z z1 ≤ z ≤ z2
2k k k 2k
19
Timescale of conductive heat flow
Geological structures such as young mountain belts are not usually in thermal
equilibrium because the thermal conductivity of rock is so low that it takes many
millions of years to attain equilibrium.

The equilibrium time can be calculated through the diffusion equation assuming
no internal heat generation

20
Timescale of conductive heat flow

T k k
=  2T =
t cP c P
Thermal diffusivity

Assuming suddenly increasing the

basal heat flow from 21 to 42 × 10-3
W/m2 (from case a to d)

Initial temperature is 567oC

It takes 20 Ma to reach 580oC

It takes 100 Ma to reach 700oC, close

to the equilibrium temperature
734oC.
21
*Still 50-km thick
Heat budget

22
Heat flux at the core-mantle boundary (CMB)
• Secular cooling
• Latent heat from inner-core crystallization
• Compositional energy due to chemical separation of the inner core
• Potential radiogenic heat generation

Growth rate of the inner core: 300 m/Gy

23
 Global Terrestrial Heat Loss

Pollack et al. (1993) 24

 Heat budget

Gravitational Energy Internal Heat Production

Kinetic Energy per Unit Volume
Internal Energy Surface Heat Flux

Energy Transfer to or
from External System
Energy From
Contraction

25
 Heat budget

Gravitational Energy
Kinetic Energy
Internal Energy

dominant

26
Oceanic Heat Flow

Oceanic crust is formed by intrusion of basaltic magma from below

Two mechanisms for heat flowing through the permeable rock and
sediments
• Conduction through the solids
• Water flow through pores and fracture because the fresh basalt is very
permeable and the heat drives water convection

27
Methods for measuring the oceanic heat flow
• Using a rigid rod carrying a thermistor chain which is shoved into sediments
—— strongly depends on the measurement environment and requires thick
sedimentary cover
• Measure temperature in deep-sea drillholes
—— very time consuming
—— The thermal environment will be disturbed during drilling operations

28
Heat flow is higher over young oceanic crust, although it is much
scattered than older oceanic basins.

The decrease of heat flow with age is to be expected if we consider hot

volcanic material rising along the axis of the mid-ocean ridges and plates
cooling as they move away from the spreading centers.

Observed heat flows (dots) for three oceans vs modeling (lines)

29
The very scattered heat-flow values are a consequence of the
hydrothermal circulation of sea water through the crust.

Contact with sea water causes rapid cooling of the new crust, and many
cracks form in the lava flows and dykes.

With increasing age, deep-sea sediments have a low permeability and

thus yield reliable estimates of the actual heat flow.

Observed heat flows (dots) for three oceans vs modeling (lines)

30
For ages less than 20 Ma, there is a simple relation between bathymetric
depth d (km) and lithosphere age t (Ma):

d = 2.6 + 0.365t 1/ 2
For ages between 20 Ma and 55 Ma, there is:

d = 5.65 − 2.47e −t / 36
For ages greater than 55 Ma:
d = 48 + 96e − t / 36

Observed heat flows (dots) for three oceans vs modeling (lines) 31

Models of plate formation and cooling
A 2D thermal conduction equation can be used to model the creation of
the lithospheric plate at the axis of a mid-ocean ridge.

The plate is cooled as it moves away from the ridge axis

T A k
= +  2T − u  T
t cP cP

32
A Simple Model
Assumption
1. The lithosphere is the cooled asthenospheric material
----- no temperature variation with time
2. The ridge is at a constant temperature Ta.
3. No heat generation
Advection only occurs
0 along the horizontal axis

0 T A k
= +  2T − u  T
t cP cP

k  2T  2T T
( 2 + 2 )=u
cP x z x

33
A Simple Model
Further assume that the horizontal heat conduction is insignificant
compared with the horizontal advection and vertical conduction of heat

k  2T  2T T
( 2 + 2 )=u
cP x z x

k  2T T
= u
cP z 2 x

34
Re-introduce the term of time by using t = x/u

k  2T T T T
=u =u =
cP z 2
x (ut ) t

or k  2T T
=
cP z 2
t
With the boundary condition: T = Ta at x = 0, and T = 0 at z = 0
It has the solution:
z 2
T ( z, t ) =T aerf ( ) Error function: erf ( x) =
2 t x
 e − y2
dy
0

35
The lithospheric thickness L can be estimated by using:
1. The temperature of the asthenosphere at the ridge axis to be 1300oC
2. The temperature at the base of the lithosphere is 1100oC.
L
1100 = 1300 erf ( )
2 t

Using  = 10-6 m2/s, there is: L = 11 t

At 10 Ma, the thickness L is 35 km; at 80 Ma, the thickness L is 98 km.

36
Two types of boundary models
Boundary condition
T = Ta at x = 0, and T = 0 at z = 0
Half-space cooling model
• The rigid lithosphere is assumed to be cooled
asthenosphere
• The base of the lithosphere is defined by an isotherm
• Fits the observed depths to about 60-70 Ma
• Lower predicted heat flow value at greater age

37
Two types of boundary models
Boundary condition
T = Ta at x = 0, and T = 0 at z = 0
Plate model
• Oceanic lithosphere has a constant thickness of L
• Base of the lithosphere is at the same constant
temperature Ta as the vertical ridge axis

38
Thermal parameters for oceanic-lithospheric models

70-80 Ma

39
GDH1
(Global depth and heat)
Considers the dependence of ocean
depth on age

40
Thermal structure of the oceanic lithosphere
Two parts of the lithosphere
• An upper rigid layer
• A lower viscous thermal boundary layer
• Become unstable about 60 Ma and develop small scale convection

41
Laboratory experiments with a moving rigid upper boundary
Flow in the upper mantle mantle could take this type of form

42
Continental Heat Flow
Harder to understand than oceanic heat flow
Harder to fit into a general theory of thermal evolution of the continents
or of the Earth.

It is affected by
Erosion Deposition Glaciation Any tectonic events
• Local concentrations of heat-generation elements in the crust
• The presence or absence of aquifers
• The drilling of holes in which the measurements were made

43
Continental Heat Flow

Over 10,000 heat flux measurements

over the continents and their
margins contribute to determine the
continental heat flow.

Average heat flux is 80 mW/m2

44
Continental Heat Flow

A strong bias
• most high heat flux values from
geothermal areas, such as
western US.
• No data coverage in large areas,
including Antarctica, Greenland,
parts of the shields in Brazil and
Africa

Average heat flux is 65 mW/m2

Total energy loss of 14 TW

45
Continental heat flow = heat produced by the radioactive decay
+ heat from the mantle

Canadian Shield

46
Using the samples from the exposed cross-section produced by tectonic
processes, together with the seismic data, the total heat production of the
continental crust can be determined.
Based on these results, the average mantle heat flux ranges from 11 to 18
mW/m2.
47
 Younger continental regions have a higher heat production

Archean: 太古宙，3.8-2.5 By
Proterozoic: 元古宙，2500 My to 542 My
Phanerozoic: 显生宙，after Proteorzoic

48
In some specific areas known as heat-flow provinces, there is a linear relationship
between surface heat flow and surface radioactive heat geneation.

They can be expressed by a linear function:

Q0 = Qr + A0 D
where Qr and D are two constants

Averaged heat flux versus average surface heat production for major geological
provinces of North America 49
Continental Heat Source

Uranium
Refractory lithophile elements
Can be directly measured
Thorium
Refractory lithophile elements

Potassium
Has to be estimated from
Moderately volatile elements
the K/U ratio

50
51
Geotherms in the Continental Crust
The transfer and generation of heat are also affected by metamorphism.

Understanding of the thermal constraints on metamorphism is important

in attempts to deduce past tectonic and thermal settings from the
metamorphic evidence available to geologists today.

Three 2D models:
• Burial metamorphism
• Intrusion metamorphism
• Overthrusting metamorphism

52
Burial metamorphism
Typical burial terrain
Sediment deposits within the
片麻岩
granitic country rock

Initial state
equilibrium temperature gradient

Sediments
• 100oC throughout the whole
body
• 0.84 μW/m-3 radioactive heat
generation
--- similar to sedimentary through
formed on a continent above a
subduction zone

53
Burial metamorphism
Typical burial terrain

片麻岩

• The sediment rapidly

equilibrates towards the
temperature of the
surrounding country rock

20 Ma • Strongly influenced by the heat

later production in the surrounding
granite

54
Intrusion metamorphism
Large igneous rock is intruded into the country rock
At the beginning, hydrothermal convection cells occur around the hot body
because of a much higher temperature, especially if the intrusion is in a
relatively wet country rock.
Basic Intrusion
Intrusion temperature: 1000oC radioactive heat generation: 0.42 μW/m-3
The country rock is heated The basalt body has solidified
and cooled

1 Ma 20 Ma

55
Granite Intrusion
Intrusion temperature: 700oC radioactive heat generation: 4.2 μW/m-3

Less contact metamorphism

than the basic intrusion

Beneath and around the

intrusion at depth, some
1 Ma 5 Ma local partial melting may
occur in the country rock

2 Ma 20 Ma

56
Overthrust metamorphism
Large overthrust slice of granite gneiss material emplaced over mafic rock
--- Suduction zone or eastern Alps
Thrusting is assumed to be instantaneous
Immediately after thrusting, the hot base of the overthrust block re-
equilibrates with the cool underthrust rocks beneath.

1 Ma
57
Overthrust metamorphism
If the cool lower slab is rich in volatiles, rapid retrograde metamorphism
takes place in the upper rock
Meanwhile, equally rapid prograde high-pressure metamorphism occurs in
the lower slab
The produced partial melt will cause re-distribution of radioactive elements.

1 Ma
58
Erosion and Deposition
Both erosion and deposition can change a geotherm rapidly.

Erode the rock at a rate of 1 km/Ma for 25 Ma

It gradually relaxed towards the new A sedimentation rate of 0.5 km/Ma

equilibrium
After sedimentation ceases, the
temperature gradually relaxed
towards the new equilibrium

59
The adiabat and melting in the mantle
For the mantle and outer core, conduction is not the primary methods of heat transfer.

Convection is believed to occur in the mantle and the outer core.

In the mantle and outer core, the material is under adiabatic expansion, in which
entropy is constant for the system.
----- The system can be imagined to be in a sealed and perfectly insulating rubber bag.

Consider a situation
• Rock at depth z and temperature T is suddenly raised up to depth z’
• When it reaches position z’, is hotter than the surrounding rocks
• But because it was previously at a higher pressure, it expands and in so doing, cools.
----- If the temperature to which it cools as a result of this expansion is the temperature
of the surrounding rocks, then the temperature gradient in the rock pile is adiabatic.
----- An adiabatic gradient is essentially the temperature analogue of the self-
compression density model.
----- Temperature gradients in a convecting system are close to adiabatic.
60
Let’s work together to describe the adiabatic temperature gradient.

Considering the rate of change of temperature with pressure

T T S
( ) S = −( ) P ( )T
P S P
With the Maxwell’s thermodynamic relations:
S V
( )T = −( ) P
P T
We have
T T V
( ) S = −( ) P ( ) P
P S T

We also have:
1 V
= ( )P
V T
and S
mass mcP = T ( )P
T

61
 T TV
( )S =
P mcP
With the density of the rock, we have:
T T
( )S =
P cP
For the Earth, we have:
dP
= − g
dr
The change in temperature with radius r is therefore given by:

T T dP T g
( )S = ( )S =−
r P dr cP
For the uppermost mantle, the adiabatic temperature gradient is 0.4oC/km.

It reduces to ~0.3oC/km at greater depths because of a reduced thermal expansion

coefficient of rocks with depth.

62
Example of the temperature
profile of the mantle and core

63
Temperature constraints

64
 Sample assemblage

The olivine to wadselyite phase transition has a large temperature

dependence of 4 MPa/K

65
Katsura et al. 2004
For 410-km discontinuity
Depth between 411 and 418 km gives a temperature of 1710 and 1780 K

Average depth of the discontinuity of 414 km corresponds to a temperature

of 1740 K.

The uncertainty of these temperature estimations is 35 K.

66
Katsura et al. 2004
Ringwoodite to perovskite and ferropericlase phase transition

67
Katsura et al. 2003
Perovskite to Post-perovskite phase transition

68
 Phase diagram of Fe

69
Anzellini et al. 2013