Chapter 1
Introduction
Convection is the main form of heat transfer in fluid in addition to conduction and radiation.
Convection is the term used to describe the slow, creeping movement of solid silicate rocks in
geological time within the Earth's mantle, driven by buoyancy forces as a result of temperature
variations in the mantle. This motion drives fundamental geodynamic forces such as
continental drifting, plate tectonics, volcanism, mountain building, and earthquakes. The
knowledge of the origin of convection and its stability is significant for the interpretation of
thermal and mechanical behavior of the Earth's interior. In particular, the study of the highly
viscous mantle convection allows studying the transport of heat from the deep interior to the
surface and its impact on the long-term evolution of the Earth. In this dissertation, we focus on
the analysis of the initiation and the stability of mantle convection in constant viscosity and
variable viscosity fluids in domains of different aspect ratio. The information on critical
Rayleigh numbers, flow patterns, and the effect of viscosity stratification on convection
patterns from this type of analysis is important for the understanding of plate generation, mantle
dynamics and lithospheric deformation. [1]
1.1 Structure of the Earth’s interior
The Earth has a dynamic and changing nature and it is supported by its complex internal
structure, through which a wide range of geological and geophysical processes occur. Such an
internal architecture is essential to geodynamics, seismology, mineral physics, and planetary
science. The planet has a stratified structure with three main layers that include the crust, mantle
and the core characterized by distinct physical properties, chemical composition and
rheological behavior. These layers interact on geological time scales, and provide processes
that alter the surface of the Earth and affect long-term thermal and mechanical evolution. [2]
1.1.1 Crust
The outermost solid layer on the earth is the crust. It is also relatively thin in comparison with
deeper layers and can be divided into two categories namely continental crust, crust of the
ocean. [3]
1|Page
Continental Crust: It varies between 30 km and 70 km. It consists primarily of silica-aluminum-
rich granitic rocks (called sial). It has a density of approximately 2.7 𝑔 /𝑐𝑚3. It is older and
more diverse than oceanic crust, and it is more than 4 billion years old in parts of the cratons.
[3]
Crust of the Ocean: It is usually 5 km to 10 km thick. It consists mainly of magnesium and
iron rich basaltic rocks (sima). It has density of approximately 3.0𝑔/𝑐𝑚3 and is therefore
denser than continental crust. The crust of the oceans is geologically recent, forming to a
maximum age of 200 million years and older oceanic crust is cycled back into the mantle at
subduction zones. [4]
The boundary between the crust and the mantle is referred to as the Mohorovicic discontinuity
(Moho) which was discovered by Andrija Mohorovicic in 1909 during the study of seismic
waves. It is associated with a significant acceleration of seismic velocity as the crustal rocks
are replaced by peridotite which is a form of mantle rock. [5]
1.1.2 Mantle
Mantle is under the crust and it is found up to a depth of about 2,900 km. It constitutes
approximately 84 percent of the bulk of the earth and approximately 68 percent of the mass.
Silicate minerals, especially olivine, pyroxenes, garnet and their high-pressure polymorphs,
make up the bulk of the mantle. There are two types of it. [6]
Upper Mantle (including Lithosphere and Asthenosphere)
The upper mantle is the region of the mantle that lies between the crust base and the depth of
about 660 km that includes the lithospheric mantle and the underlying asthenosphere. It
consists mainly of the ultramafic rocks which include peridotite with higher concentration of
olivine and pyroxene and traces of garnet and spinel depending on the pressure-temperature
regimes. The upper, which is combined with the crust, makes the lithosphere, which is a rigid
layer and is involved in the tectonic movement of plates. Below this is the asthenosphere, which
is mechanically weaker, and partly molten, and which allows lithospheric movement to occur
by ductile flow. There are great temperature and pressure gradients in the upper mantle, with
temperatures varying between (500 − 900)°𝐶 along the lithosphere-asthenosphere boundary
and 1600°𝐶 along the transition zone. Seismic investigations indicate that there is a distinct
zone of low-velocity (LVZ) in the asthenosphere, which demonstrates partial melting or greater
ductility. The upper mantle convection is a major contributor to plate motions, volcanism, and
2|Page
mantle plume activity and, therefore, is a key layer in the study of geodynamic processes.
Increasing pressure causes a strong mineralogical change in Transition Zone, especially the
conversion of olivine to wadsleyite at a depth of about 410 km and ringwoodite to bridgmanite
and ferropericlase at a depth of about 660 km. These transitions influence mantle convection
dynamics, slab stagnation, and seismic velocity discontinuities. [6] [7] [8]
Lower Mantle (𝟔𝟔𝟎 km − 𝟐, 𝟖𝟗𝟎 )
This area is mainly composed of bridgmanite (𝑀𝑔𝑆𝑖𝑂3 ), perovskite, ferropericlase(𝑀𝑔, 𝐹𝑒)𝑂,
which is stable at high pressure. It is stronger than the upper mantle in a mechanical sense, and
the flow is sluggish and slow because of the higher viscosity. Recent research also implies that
iron-spin transitions in minerals influence their seismic behaviour and density. [3]
1.1.3 Core
The core extends from a depth of approximately 2,890 km to the Earth’s center at 6,371 km
and is composed predominantly of iron, with lesser amounts of nickel and light elements such
as sulfur, silicon, and oxygen. It is subdivided into two distinct regions: the outer core and the
inner core, each with unique physical states and dynamic roles in Earth’s interior processes. [9]
Outer Core (𝟐, 𝟖𝟗𝟎 km− 𝟓, 𝟏𝟓𝟎 km): The outer core is liquid known by the fact that there
was no S-wave propagation and attenuation of P-waves was large. Its dynamo generates the
geomagnetic field of Earth by convection. Outer core temperatures vary between 4,000℃ and
6,000℃ and pressures lie between 135 GPa and 330 GPa. [10]
Inner Core (𝟓, 𝟏𝟓𝟎 km − 𝟔, 𝟑𝟕𝟏 km): This solid inner core is hot because of the extreme
pressures (more than 330 Gigapasigas). It is mainly made of crystalline iron-nickel alloy,
perhaps in a hexagonal close-packed (hcp) structure. The anisotropy in the inner core,
indicating that the crystal is preferred, is found through seismic techniques, and increases
gradually when the earth cools down to the extent that the outer core convects, releasing latent
heat. [11]
1.1.4 Seismic Evidence and Discontinuities
Seismic wave propagation has helped most in the study of the internal structure of the Earth.
The primary waves (P-waves) are compressional and can travel through solids, liquids, and
gas, and the Secondary waves (S-waves) are shear waves and can only occur in solids and this
is the essential evidence of the existence of liquid layers in the Earth. There are a number of
essential discontinuities that are observed seismically. At an average depth of about 35 km, the
3|Page
boundary between the crust and mantle is represented by the Mohorovic discontinuity (Moho).
In the mantle, the phase changes take place at depths of about 410 km and 660 km i.e.
mineralogical transitions of olivine and related phases. The interface between the solid silicate
mantle and the liquid iron-rich outer core is the Core-Mantle Boundary (CMB), at a depth of
about 2,890 km. The features of this boundary are complex, such as ultra low velocity zones
(ULVZs), which are defined as places of partial melt or great chemical heterogeneity. Last but
not least, the Inner Core Boundary (ICB) at the depth of approximately 5,150 km, signifies the
phase and composition change between the liquid outer core and the solid inner core. [12] [13]
1.1.5 Geothermal Gradient and Heat Transfer
The geothermal gradient is 25 − 30 ℃/𝑘𝑚 in the crust but gets lower in the mantle as a result
of the convective heat transfer, and thus the temperature profile is approximately adiabatic. The
sources of heat on the Earth are Primordial heat of accretion and differentiation, Radiogenic
decay of isotopes (U, Th, K) and Latent heat release on solidification of the inner core.
Convection is the most important mode of heat transfer in the mantle and conduction and
convection are important in the core with the outer core convection producing the geomagnetic
field. [14]
Figure 1.1: Structure of Earth’s interior (Source: [Link]/figure/The-internal-structure-of-the-Earth-Geosphere-is-layered-in-
spherical-shells_fig1_334132194).
4|Page
1.2 The Earth’s Mantle and its Properties
The mantle is a large, almost solid mass that stretches downward through the base of the crust
(the Mohorovicic discontinuity) to the core-mantle boundary at a depth of about 2,890 km. It
is approximately 84 percent of the total volume, and about 68 percent of the total mass of the
earth, so it plays a leading role in the regulation of planetary heat transfer, chemical
differentiation, and dynamic evolutionary processes. [15]
1.2.1 Composition and Mineralogy
Silicate minerals that contain magnesium and iron constitute the major part of the mantle. At
deeper levels, on its upper portion, it is primarily composed of olivine [(𝑀𝑔, 𝐹𝑒)2 𝑆𝑖𝑂4 ],
orthopyroxene [(𝑀𝑔, 𝐹𝑒)𝑆𝑖𝑂3 ], clinopyroxene [(𝐶𝑎, 𝑀𝑔, 𝐹𝑒)𝑆𝑖2 𝑂6 ], and garnet. As depth and
pressure increase, these minerals become high-pressure polymorphs, Olivine turns into
wadsleyite, which is more dense and higher seismic velocity, then Wadsleyite turns into
ringwoodite, then ferropericlase (𝑀𝑔, 𝐹𝑒)𝑂, the lower mantle begins. Bridgmanite is the
primary mineral (~80%), followed by ferropericlase and small amounts of calcium silicate
perovskite, in the lower mantle. Seismic discontinuities and convection patterns are largely
affected by the mineralogical transitions. [16]
1.2.2 Mechanical Behavior and Rheology
The mantle is mainly solid, but it presents a viscoelastic behavior on geological time scales, so
that it can deform and flow due to creep processes under permanent stress. The two main
processes of deformation are diffusion creep wherein, the strain is accommodated by atomic
diffusion, and dislocation creep wherein, the dislocation moves between crystal lattices under
differential stress, and are the most common in fine-grained and coarse-grained peridotite
lithologies, respectively. The rate of deformation and the large-scale convective flow are
controlled by the effective viscosity of the mantle, which therefore plays a fundamental role in
controlling the lithospheric plate motion, mantle plume ascent, slab subduction, and the global
thermal change of the planet. [17]
1.2.3 Mantle Viscosity: Dependence and Variability
The viscosity of mantle is not constant but it varies significantly with temperature, pressure,
grain size, water content, and mineralogical composition. According to the models of
geophysical constraints based on post-glacial rebound studies, geoid modeling, and mantle
flow modeling, the upper mantle generally has viscosities of the order of~10²¹ Pa·s, with
5|Page
higher viscosities of ~10²² Pa s as one approaches the transition zone. The lower mantle has
viscosities estimated to be in the range of 1022-1024 Pas, with various models indicating a
gradual scale change by factor of 10-100 across the 660 km seismic discontinuity. [18]
1.2.4 Arrhenius-type Viscosity Dependence
The temperature- and pressure-dependent viscosity of the mantle can be modeled as an
Arrhenius-type relationship, developed through experimental research in mineral physics and
rheology, that takes the form of the equation-
𝐸 ∗ +𝑝𝑉 ∗
𝜂 = 𝜂0 exp ( ) (1.1)
𝑅𝑇
where 𝜂 denotes the effective viscosity (𝑃𝑎 · 𝑠), η0 is the reference viscosity at a specified
temperature and pressure, 𝐸 ∗ is the activation energy (𝐽 · 𝑚𝑜𝑙⁻¹), 𝑝 is the pressure (𝑃𝑎), 𝑉 ∗ is
the activation volume (m³·mol⁻¹), RRR is the universal gas constant (8.314 J·mol⁻¹·K⁻¹), and
𝑇 is the absolute temperature (𝐾). The activation energy 𝐸 ∗ quantifies the temperature
sensitivity of creep, with values for mantle minerals typically ranging from 300 to 600 𝑘𝐽 ·
𝑚𝑜𝑙⁻¹. The activation volume 𝑉 ∗ reflects the sensitivity to pressure, influencing the increase
in viscosity with depth as the free volume available for atomic movement decreases under
compression. Theoretically, as pressure rises up and down, the effect is partly negated by the
corresponding rise in temperature, and this results in complex depth-dependent viscosity
profiles in the mantle. [19]
1.2.5 Factors Affecting Mantle Viscosity
a. Temperature
The viscosity in the mantle shows an exponential rate of decrease as temperature increases,
producing regions of thermal anomalies like mantle plumes that are much more deformable
than colder and seismically fast regions linked with subducting slabs. Spatial variations in
temperature thus have a primary control on the strength of mantle convection by changing local
Rayleigh numbers and hence the strength of convective heat transport. [20]
b. Pressure
In general, increasing pressure tends to enhance mantle viscosity through the compression of
crystal lattices, a relationship captured by the positive activation volume term in the Arrhenius
formulation. However, this trend is disturbed by the effects of mineral phase changes,
6|Page
electronic spin change in iron bearing phases and the existing thermal gradient and any of these
factors can change the overall pressure viscosity relationship in the mantle. [21]
c. Composition and Phase
The compositional heterogeneity of the mantle, such as differences in iron content and basaltic
enrichment of the recycled oceanic crust, has a profound effect on the rheological behavior of
mantle minerals. Additionally, the mineral phase changes, including olivine, wadsleyite and
ringwoodite, cause significant viscosity changes, which are caused by the restructuring and
increment in density. [22]
d. Water Content
Water plays a major role in lowering the viscosity of mantle minerals by increasing movement
of defects in the crystal structure. Experimental studies show that water in the form of even
trace levels, tens to hundreds of parts per million, can reduce the viscosity of olivine by about
one order of magnitude, thus leading to the ductile creep of the asthenosphere and plate tectonic
processes. [23]
e. Grain Size
The grain size critically controls the deformation mechanisms of the mantle with smaller grains
performing diffusion creep and hence reducing viscosity, and larger grains performing
dislocation creep. Moreover, processes like recrystallization and grain growth during
deformation are dynamic, and consequently result in the changing sizes of grains, which, in
turn, dynamically change the effective viscosity of the mantle. [24]
1.2.6 Viscosity Structure and Convection
The nature of mantle convection is subject to a fundamental control by the viscosity structure
of the mantle. High viscosity stratification, which is the large difference in viscosity between
upper and lower mantle, is likely to favor layered convection thus limiting material exchange
across the 660 km seismic discontinuity. On the contrary, low viscosity stratification promotes
whole-mantle convection that supports the transport of masses and heat on large masses of the
mantle. In geodynamic models, either non-Newtonian rheology (where stress-dependent
viscosity, as found in dislocation creep, is taken into account) or Newtonian rheology (where
stress-independent viscosity, as found in diffusion creep-dominated regimes, is used) is
common.
7|Page
−𝐸 ∗ + 𝑝𝑉 ∗
𝜀̇ = 𝐴𝜎 𝑛 𝑒𝑥𝑝 ( ) (1.2)
𝑅𝑇
The strain rate, denoted as 𝜀̇, is measured in (𝑠 −1 ). The parameter 𝐴 represents the pre-
exponential factor, which is a material constant. The differential stress is symbolized by 𝜎 is
measured in pascals (𝑃𝑎). The stress exponent, 𝑛, is a dimensionless quantity that characterizes
the nonlinearity of the creep behavior. The activation energy, 𝐸 ∗ , is expressed in (𝐽/𝑚𝑜𝑙),
while the pressure 𝑝 is given in pascals (𝑃𝑎). The activation volume, 𝑉 ∗ has units of cubic
meters per mole (𝑚³/𝑚𝑜𝑙). The universal gas constant, 𝑅, has a value of 8.314 (𝐽/𝑚𝑜𝑙 · 𝐾),
and the absolute temperature, 𝑇, is measured in kelvin (𝐾). [17] [25]
1.2.8 Implications for Geodynamics
The viscosity of the mantle is very important to regulate various basic geodynamic processes.
It controls the effectiveness of the convective heat transfer and as a result, it affects the cooling
rate of the earth as well as the surface heat flux. Plate tectonics and lithospheric mobility are
also influenced by the viscosity structure; a weak asthenosphere promotes the mechanical
decoupling of tectonic plates. Also, low-viscosity areas allow volcanism through the mantle
plumes that cause hotspots. In subduction regions, descending slabs are so viscous that they
can slip into the lower mantle, causing deep mantle convection. Lastly, viscosity in the region
of the core-mantle boundary affects the thermal conductivity between the core and mantle,
which is critical in maintaining long-term processes of geodynamic and magnetic field
generation. [26]
Figure 1.2: Variation of mantle velocity and density with depth (Source:
[Link]/articles/35087500 ).
8|Page
1.3 Convection and Rayleigh-Bénard Convection
The concept of convection is a major mechanism of heat transfer in fluids due to buoyancy
forces that develop as a result of density differences caused by temperature variations. When
fluid is heated below and cooled above, the density contrasts formed cause the warmer, less
dense fluid to rise and the colder or denser fluid to fall. An example of such a process is
Rayleigh-Benard convection, in which a layer of fluid that is heated at the bottom develops
highly structured convective cells as soon as the temperature gradient rises above a critical
value, with the dimensionless Rayleigh number. This convective behaviour is fundamental to
many natural and man-made systems, including atmospheric circulation, oceanic currents, and
the geodynamic convection that takes place in the mantle of the Earth.[27]
1.3.1 Thermal Convection
Buoyancy forces have to overcome the stabilizing effect of viscosity (which prevents motion
of fluids) and thermal diffusion (which tends to eliminate temperature gradients) in thermal
convection. The ratio between these forces is measured by the Rayleigh number (Ra) which is
a non-dimensional number that determines the development and strength of convection. Ra
Rayleigh number is defined as.-
𝜌𝑔𝛼𝛥𝑇𝑑3
𝑅𝑎 = (1.3)
𝜅𝜂
Where; 𝜌 is Fluid density (kg/m³), 𝑔 is the gravitational acceleration (m/s²), α is known as the
thermal expansion coefficient (𝐾⁻¹),𝛥𝑇 is the temperature difference across the layer (𝐾), 𝑑
is the depth or thickness of the fluid layer (m), 𝜅 is the thermal diffusivity (m²/s), 𝜂 is defined
as dynamic viscosity (𝑃𝑎 · 𝑠). [28]
1.3.2 Rayleigh-Bénard Convection
The most basic form of thermal convection is called Rayleigh-Benard convection that takes
place in a horizontal layer of fluid between two infinite parallel plates with the warmer plate at
the bottom and the colder plate at the top. It is an ideal representation of convectional patterns
and heat transfer of fluids and planetary mantles. At a critical value of the Rayleigh number
(𝑅𝑎𝑐 ), the statical conductive state loses its stability, and convection establishes itself. Critical
Rayleigh number is related to the boundary conditions: in rigid-rigid boundaries, (𝑅𝑎𝑐 ) is about
1708; in free-free boundaries, it is about 657; and in rigid-free boundaries, it is about 1100.
When the Rayleigh numbers is lower than the critical value, the conduction heat transfer occurs
9|Page
without any fluid motion. But as the Rayleigh number just exceeds the critical value,
convection starts.[29]
1.3.3 Convection Pattern Evolution
As the Rayleigh number increases beyond the critical threshold, the nature of convection
undergoes distinct changes. At the critical number of Rayleigh, the convection is steady and
laminar with symmetric patterns. Beyond moderate Rayleigh numbers, the flow is no longer
steady but oscillatory. Convection at high Rayleigh numbers also changes to a turbulent regime
of plumelike structures which are complex and chaotic. This turbulent convection is observed
in the mantle of the earth, the Rayleigh number usually falls between 107 and 106 . [30]
Physical Interpretation
The Rayleigh number quantifies the ratio of destabilizing buoyancy forces to stabilizing
viscous and diffusive forces within a fluid. The numerator, expressed as 𝜌𝑔𝛼𝛥𝑇𝑑 3 , represents
the buoyancy driving force, where fluid density, gravitational acceleration, thermal expansion
coefficient, temperature difference, and fluid layer depth collectively contribute to the upward
force. The denominator, consisting of the product of thermal diffusivity (𝜅) and dynamic
viscosity (𝜂), characterizes the stabilizing effects of thermal diffusion and viscous resistance.
Increasing either the temperature difference 𝛥𝑇 or the depth 𝑑 raises the Rayleigh number by
improving the buoyancy forces, or the volume scale of the system, the latter being cubically
dependent. On the other hand, when dynamic viscosity is low, viscous damping is small the
Rayleigh number is large. Conversely, higher thermal diffusivity will allow more conduction
of heat in the fluid layer stabilizing it and lowering the Rayleigh number. [6] [15]
Importance in Geodynamics and Engineering
Rayleigh-Benard convection is a significant model in geophysical fluid dynamics and offers
fundamental understanding of the processes in the mantle and core. This model gives the
explanation of the convective process of heat transfer in the interior of the earth that leads to
major phenomena like plate tectonics and formation of the mantle plumes in the mantle.
Likewise, at the outer core of the earth, the dynamic process that leads to the formation of the
geomagnetic field is based on convection. [31]
10 | P a g e
Industrial applications: The industrial uses of Rayleigh-Benard convection are very diverse
that involves numerous processes in which buoyancy-based heat transfer is an essential factor.
These are design and optimization of heat exchangers, where effective thermal management is
required to achieve energy conservation and system performance. Cooling systems used in
electronics and power plants are based on controlled convective flows to ensure that operational
temperatures are maintained and overheating is avoided. Moreover, the concepts of Rayleigh-
Benard convection are used in crystal development technologies and the temperature
homogeneity and fluid flow are essential in the development of high-quality crystals with fewer
defects. Knowledge and management of convection patterns in this environment enhance
efficiency of process, quality of products and use of energies. [32]
Figure 1.3: Schematic of Rayleigh-Bénard convection (Source
[Link]/figure/A-sketch-of-Rayleigh-Benard-convection-in-a-
cylindrical-cell-with-unit-aspect-ratio-The_fig2_264536303).
1.4 Onset of Convection and Stability Analysis
The beginning of the convection marks the basic change in the process of pure conduction,
when heat is carried only by molecular diffusion to convection, fluid movement under the
effects of buoyancy promotes the efficiency of heat transfer. The critical Rayleigh number
controls this transition and above this number the static conductive equilibrium is no longer
stable, and convective motions appear. [15]
11 | P a g e
Stability analysis: It gives a mathematical context in which the critical conditions of how an
infinitesimal perturbation increases temperature and velocity fields instead of decays. Stability
analysis is also essential in mantle geodynamics to determine when and how convective
overturn starts, which in turn has a direct impact on the cooling rates of the planet, lithosphere,
and plume formation. [6]
Highly viscous mantle materials: In this kind of materials the critical Rayleigh number is
greatly influenced by the viscosity structure. Constant and variable viscosity profiles change
the resistance to flow which also change the onset conditions. In addition, boundary conditions,
e.g. no-slip or free-slip surfaces, are of great importance in determining stability limits and
convection cell patterns.. [15]
Aspect ratio (width to height ratio): The flow patterns and the critical Rayleigh numbers are
affected by the aspect ratio of the domain or convection cell. A low aspect ratio contributes to
the stability of large-scale rolls whereas a high aspect ratio contributes to multiple narrower
convection cells. The aspect ratio effects are fundamental for the modelling of realistic mantle
convection geometries that helps to understand the convective features in the interior of the
earth. [33]
1.5 Variable Viscosity Effects
In real models of mantle convection, viscosity is not uniform and greatly depends on
temperature and depth. The temperature dependence of viscosity is specifically important
because mantle rocks show an exponential temperature dependence of viscosity, consistent
with Arrhenius-type relations. This variance has important consequences on convective
dynamics. The asymmetrical flow configurations are created by the presence of high viscosity
between cold and inflexible lithospheric slabs and the hot low-viscosity mantle plumes. Cold
layers that are thick and stagnant do not deform due to the viscous nature, whereas the hot,
low-viscosity areas that are thin and active allow the horizontal thermal plumes to develop
within the mantle. The depth dependence of viscosity due to pressure, changes not only the
mineral phase but also changes the convection style by changing the resistance to flow at the
bottom relative to the top of the mantle. However, these natural viscosity variations are
important to include in a geodynamic solution for the natural observation of mantle circulation,
heat history, and surface tectonic expressions. [25] [34]
12 | P a g e
1.6 Boundary Conditions: No-slip and Free-slip
In geophysical models and fluid dynamics, boundary conditions are the central reason for onset,
flow structure and stability of convection.
No-slip boundary conditions: Here, we consider the velocity of the fluid to be zero at the
boundary, which is to say that the fluid is stuck to the boundary surface without any tangentially
or normal motion. This special condition has physical implications of realistic situations where
the mantle is exposed to a rigid lithosphere or core-mantle boundary causing increased shear
stress and increased viscous dissipation at the boundaries. As a result, the critical Rayleigh
number of convection onset is larger under no-slip than under free-slip, meaning it has more
resistance to initiating the flow.
Free-slip boundary conditions: In contrast to no-slip boundary conditions in this case, we
consider the fluid to slide tangentially along the boundary, but with zero normal velocity. This
causes a reduction in the shear stress at the boundary, which leads to easier flow start and
decreases the critical Rayleigh number of the start of convection.
The knowledge of the impacts of these boundary conditions is crucial in stability analysis and
numerical simulations in order to provide a realistic model of mantle convection behavior for
different geological conditions and aspect ratios.[35]
1.7 Mathematical Modelling and Computational Methods for Mantle Convection
This section describes the numerical techniques and the mathematical formulation used to
analyze the onset of convection and stability in highly viscous mantle convection, focusing on
different aspect ratio cells with both constant and variable viscosity.
Numerical Techniques
The incompressible Stokes equations of flow and the thermal energy equation of mantle
convection are nonlinear coupled equations, and it is challenging to find an analytical solution
to such equations. The complex interactions, especially viscosity variability and realistic
boundary conditions, thus require numerical techniques that can solve the complex interactions.
The most common approaches to spatial discretization are the Finite element and Finite volume
methods because their adaptability allows smooth integration to complex geometries and
variable material properties. These techniques divide the computational field into smaller
volumes or elements. The thermal energy equation is often time integrated using implicit or
13 | P a g e
semi-implicit schemes to ensure that the numerical system is stable, particularly due to the
large time scale difference between slow mantle flow and thermal diffusion. Iterative solvers
are commonly used to solve the velocity and pressure fields simultaneously to solve the Stokes
equations. Nonlinear viscosity models adds further computing complexity and requires strong
convergence criteria to guarantee accuracy of solutions. Alternatively, adaptive mesh
refinement methods can be used to refine resolution in the areas with sharp gradients, e.g. a
thermal boundary layer or a plume channel, to optimize the use of computational resources
without compromising the reliability of the solution. In general, the integration of highly
discretized algorithms, iterative solvers, the inclusion of nonlinear rheology, and adaptive
meshing create a strong suite of computational tools that allow the accurate prediction of the
processes of mantle convection. [36] [37]
1.8 Importance of Studying Onset and Stability in Mantle Convection
The mechanism and stability of mantle convection is one of the main factor for understanding
the internal processes of the Earth and its long-term thermal history. Mantle convection
controls the process of heat transfer between the deep interior and the surface, and affects plate
tectonics, volcanic eruptions, and seismicity. The change in the conductive heat transfer to the
convective motion is a vital shift in the redistribution of energy in the Earth. The analysis of
the conditions under which convection initiation takes place helps us to understand the limits
within which mantle substance starts to deform and flow. This is especially necessary when
dealing with highly viscous mantle materials, where it is difficult to initiate motion due to a
tremendous resistance. Stability analysis also helps us to determine how initial convective
motions will sustain, amplify or dissipate, aiding predictions of flow patterns and the possibility
of mantle plumes or slab sinking. The study of viscosity variations and different aspect ratio
geometries help us to understand the real phenomena of Earth’s mantle because mantle
viscosity varies greatly with temperature and pressure, and Earth’s interior is in complex
shapes. These studies help to increase the predictability of geodynamic models, which leads to
improved our knowledge on thermal and mechanical behavior of the earth. This understanding
finally helps us to explain the geological phenomena. [15] [34]
14 | P a g e
References
[1] G. F. Davies, Dynamic Earth: Plates, Plumes and Mantle Convection., Cambridge University
Press., 1999.
[2] F. D. D. P. M. Stacey, Physics of the Earth, Cambridge University Press, 2008.
[3] C. M. R. Fowler, An Introduction to Global Geophysics, Cambridge University Press, 1990.
[4] R. J. T. Eldridge M. Moores, Tectonics, W. H. Freeman, 1995.
[5] H. K. G. ,. A. C. ,. E. R. E. ,. R. K. ,. A. M. ,. S. R. ,. K. S. ,. S. U. Kusumita Arora,
Encyclopedia of Solid Earth Geophysics, Springer, 2021.
[6] D. T. Gerald Schubert, Geodynamics, Cambridge University Press, 2014.
[7] K. A. K. ,. F. J. V. Philip Kearey, Global Tectonics, Wiley-Blackwell, 2009.
[8] K. C. Condie, Plate Tectonics and Crustal Evolution, Butterworth-Heinemann, 1997.
[9] D. L. A. Adam M. Dziewonski, "Physics of the Earth and Planetary Interiors," Preliminary
reference Earth model, vol. 25, no. 4, pp. Pages 297-356, 1981.
[10] B. A. Buffett, "Earth's Core and the Geodynamo," Science, vol. 288, no. 5473, 2000.
[11] S. P. J. P. &. L. M. J. L. Labrosse, "Earth and Planetary Science Letters:The age of the inner
core," sciencedirect, vol. 190, no. 3-4, pp. 111-123, 2001.
[12] T. C. W. Thorne Lay, Modern Global Seismology, Academic Press, 1995.
[13] P. M. Shearer, Introduction to Seismology, Cambridge University Press., 2009.
[14] C. L. S. &. M. J. C. Jaupart, "Temperatures, Heat, and Energy in the Mantle of the Earth,"
Research Gate, vol. 7, pp. 223-270, 2007.
[15] D. L. Turcotte, "Mantle Convection in the Earth and Planets," Research Gate , 2009.
[16] A. E. Ringwood, Composition and petrology of the Earth's mantle, McGraw-Hill Inc.,US, 1975.
[17] S.-i. Karato, Deformation of Earth Materials:An Introduction to the Rheology of Solid Earth,
Cambridge University Press, 2008.
[18] "A new inference of mantle viscosity based upon joint inversion of convection and glacial
isostatic adjustment data," vol. 225, no. 1–2, pp. 177-189.
[19] G. Ranalli, Rheology of the Earth, Chapman & Hall, 1995.
15 | P a g e
[20] D. K. Greg Hirth, "Rheology of the Upper Mantle and the Mantle Wedge: A View from the
[Link] the Subduction Factory," vol. 138, pp. 83-108.
[21] S.-I. &. J. H. Karato, "Effects of pressure on high-temperature dislocation creep in olivine,"
ResearchGate, vol. 83, pp. 404-414, 2003.
[22] "Mantle Convection and Plate Tectonics: Toward an Integrated Physical and Chemical Theory,"
vol. 288, no. 5473, pp. 2002-2007, 2000.
[23] G. &. K. D. L. Hirth, "Water in the oceanic upper mantle: Implications for rheology, melt
extraction, and the evolution of the lithosphere. Earth and Planetary Science Letters," Science
Direct, vol. 144, no. 1-2, pp. 93-108.
[24] W. M. B. G. H. L. T. P. A. Speciale, "Rates of Olivine Grain Growth During Dynamic
Recrystallization and Postdeformation Annealing," Advancing Earth And Space Sciences, vol.
125, no. 11, 2011.
[25] U. R. Christensen, "Convection with pressure and temperature-related non-Newtonian
rheology," Geophysical Journal International, vol. 77, no. 2, pp. 343-384, 2007.
[26] "Plate tectonics, damage and inheritance," Nature, vol. 508, no. 7497, pp. 513-516, 2014.
[27] S. CHANDRASEKHAR., Hydrodynamic and Hydromagnetic Stability., Oxford University
Press, 1961.
[28] B. G. R. L. M. Y. Jaluria, Buoyancy-Induced Flows and Transport, Springer, 1988.
[29] E. L. Koschmieder, Bénard Cells and Taylor Vortices, Cambridge University Press, 1993.
[30] E. D. Siggia, "High Rayleigh Number Convection," Annual Review of Fluid Mechanics, vol. 26,
no. 1, pp. 137-168.
[31] P. H. &. K. E. M. Roberts, "On the genesis of the Earth’s magnetism.," Reports on Progress in
Physics., vol. 76, no. 9, 2013.
[32] A. Bejan, Convection Heat Transfer, Wiley, 2013.
[33] "On the effect of continents on mantle convection," JOURNAL OF GEOPHYSICAL
RESEARCH, vol. 100, no. B12, pp. 24217-24238, 1995.
[34] P. J. Tackley, "Mantle Convection and Plate Tectonics: Toward an Integrated Physical and
Chemical Theory," Science, vol. 288, no. 5473, pp. 2002-2007, 2000.
[35] A. C. Fowler, "Toward a Description of Convection with Temperature-and-Pressure-Dependent
Viscosity," Studies in Applied Mathematics journal, vol. 88, no. 2, pp. 113-139, 1993.
[36] L.-N. &. S. V. S. Moresi, "Numerical investigation of 2D convection with extremely large
viscosity variations.," Physics of Fluids, vol. 7, no. 9, pp. 2154-2162, 1995.
16 | P a g e
[37] T. S. Khaleque, A. C. Fowler and P. D. Howell, "Numerical studies of thermal convection with
temperature- and pressure-dependent viscosity at extreme viscosity contrasts," Physics of
Fluids, vol. 27, no. 7, 2015.
17 | P a g e