Climate Modelling

Isaac Tchuwa, PhD

Email: [email protected]
Tel: +265881370317
 Factors influencing climate
 Incident solar radiation - variation with latitude

 Closeness to large water bodies - distribution of land & water

 Mountain barriers

 Altitude

 Ocean temperature and currents

 Land cover

 Atmospheric composition
 Interactions

The non-linear interaction among the components leads to

climate variability at a range of spatial and temporal scales
Pic . NOAA
 Review of Basics:
Climate System

Added
warming by
human
intervention

The non-linear interaction among the components leads to

climate variability at a range of spatial and temporal scales
 How do we quantify the response
of the climate?

• The response of the climate system to this

forcing agents is complicated by:
 feedbacks
 the non-linearity of many processes
 different response times of the different components to a
given perturbation

• The only means available to calculate the

response is by using numerical models of the
climate system.
 Climate Model

• A climate model is a computer simulation that uses mathematical

equations to represent the processes that drive Earth's climate.
• It helps to understand the past, present, and future climate by
predicting responses to changes in factors like greenhouse gases.
Purpose:
• Understanding Climate Processes: Climate models help to analyze how
energy flows through the atmosphere, ocean, and other parts of the
climate system.
• Predicting Future Climate: They are used to project future climate
scenarios based on different assumptions about emissions and other
factors.
• Risk Assessment: Models assess the potential impact of climate
change on natural and human systems.
 Observations

Theory
Model
s

McGuffie,
Warner K. and Henderson-Sellers,
(2011) Numerical WeatherA.and
(2005) A Climate
Climate Modelling
Prediction. Primer.
Cambridge 3rd ed., Press.
University Wiley.
Climate model - an attempt to simulate many
processes that produce climate

The simulation is accomplished by describing the

climate system of basic physical laws.

Model is comprised of series of equations

expressing these laws.

Climate models can be slow and costly to use,

even on the faster computer, and the results can
only be approximations.

The objective is to understand the processes

and to predict the effects of changes and
interactions.
The processes of climate system interact with
each other, producing feedbacks, which in turn
involves great deal of computation to simulate.

The solutions start from some “initialized” state and

investigate the effects of changes in different
components of climate system.

Boundary conditions – solar radiation – set from

obs. data, but since data itself aren‟t that
complete, hence inherent uncertainty exists.

2 sets of simplifications
–Involving process
–Involving resolution of model in time and
space
Process simplification –

• Treating some processes in detail and

approximating others due to their
Inadequate understanding or lack of
computer resources.

• E.g.- treating radiation process in detail,

but approximating the horizontal energy
flows associated with regional – scale
winds.

• These approximations may be approached

either by using available obs. data, some
empirical approaches, or through
simplifications of physical laws.
Time and space simplifications –

• Resolution of a model should be used

appropriately.

• If process involved is larger than model resolution,

finer resolutions for that model may be avoided.

• Temporal resolution or “timestep‟ approach may

have constraints imposed by data, computational
ability and model design

• Process allowed to simulate for a certain time 

new conditions calculated  process repeated with
new values  continues till conditions at the
required time have been established.
 Components of Climate models
• Radiation – input and absorption of solar radiation and
emission of infrared radiation handled.
• Dynamics – horizontal movements of energy around the
globe (low to high lat.) and vertical movements
(convection etc.)

• Surface processes – inclusion of land/ocean/ice and the

resultant change in albedo, emissivity and surface-
atmosphere energy interactions.

• Resolution in both time and space – the time step of the

model and the horizontal and vertical scales resolved.
 Climate Model
Climate models divide the Earth into a grid and simulate various climate
processes.
They include components for:
Atmosphere: Governs the movement of air, the distribution of
temperature and moisture, and the interaction with solar radiation.
Ocean: Models ocean currents, temperature, salinity, and their role in
transporting heat and carbon.
Land Surface: Represents the exchange of energy, moisture, and carbon
between the land and the atmosphere, including processes like
evapotranspiration and vegetation dynamics.
Cryosphere: Simulates the interaction of ice sheets, glaciers, and sea ice
with the climate system.
Biosphere: Includes vegetation and ecosystems, simulating carbon
exchange, photosynthesis, and other biological processes.
Framework for a Model

Source: MPI, Germany

Components of a weather/climate model

McGuffie, K. and Henderson-Sellers, A. (2005) A Climate Modelling Primer. 3rd ed., Wiley.
 Modelling Climate

Vertical exchange between layers

of momentum, heat and moisture

15° W
Horizontal exchange
60° N
3.75° between columns
of momentum,
heat and moisture
2.5°

Vertical exchange
between layers
of momentum, 11.25° E
heat and salts Vertical exchange between layers 47.5° N
by diffusion, by diffusion and advection
convection
and upwelling Orography, vegetation and surface characteristics
included at surface on each grid box
 Types of Climate Models
Climate models differ in their complexity and the number of processes they include. The main
types are:
Energy Balance Models (EBMs)
Simple Models that calculate the balance between incoming solar radiation and outgoing
terrestrial radiation.
Used to provide a broad understanding of climate sensitivity and equilibrium temperature
changes.
Zero-Dimensional and One-Dimensional Models
Zero-Dimensional EBMs treat the Earth as a single point and determine an average global
temperature.
One-Dimensional Models account for temperature changes with altitude (vertical) or across
latitudes, providing a more detailed but simplified representation.
General Circulation Models (GCMs)
Also known as Global Climate Models, these are the most sophisticated models, providing a
three-dimensional representation of the atmosphere and oceans.
Atmosphere-Ocean General Circulation Models (AOGCMs) are often used to simulate climate
over long time periods and include interactions between the ocean and atmosphere.
 Types of Climate Models

Earth System Models (ESMs)

Extended GCMs that incorporate additional processes like the
carbon cycle, vegetation, and biogeochemical interactions.
Used to assess long-term feedbacks, such as how increasing CO₂
levels might affect oceanic uptake of carbon.
Regional Climate Models (RCMs)
Focus on specific regions, offering higher-resolution simulations
to analyze local climate impacts and processes.
Typically "nested" within a global model to provide detailed
information for smaller areas.
 Basic Equations in General Circulation Models
• Climate models are built on fundamental equations that describe the physical and thermodynamic
processes occurring in the Earth's climate system.
• These equations form the foundation of how atmospheric, oceanic, and land-surface interactions
are simulated, allowing us to understand and predict climate behavior.
• Here, we explore the key equations that govern climate models, including their purpose,
applications, and underlying principles.
• The equations governing climate models primarily consist of conservation laws—the conservation
of momentum, mass, and energy—and describe the movement of fluids (air and water),
thermodynamic changes, and radiative transfers.
• The major types of equations are:
• Navier-Stokes Equations (for momentum conservation)
• Continuity Equation (for mass conservation)
• Thermodynamic Energy Equation (for energy conservation)
• Equation of State (relating pressure, temperature, and density)
• Radiative Transfer Equations (for radiation processes)
• Moisture Conservation (for humidity and phase changes)
• Tracer Transport Equations (for chemical components like CO₂)
• These equations are coupled and numerically solved to simulate the atmosphere, ocean, land, and
ice dynamics.
 Basic Equations in General Circulation Models
Navier-Stokes equations, which are the fundamental equations for fluid flow, including
atmospheric motion. In a simplified form, these equations can be written for the three-
dimensional flow of air (in Cartesian coordinates):
 Basic Equations in General Circulation Models
The Navier-Stocks equations describe atmospheric circulation (e.g., wind patterns) and
ocean currents, which are critical for understanding climate variability and transporting
heat globally.

Continuity Equation (Mass Conservation): the continuity equation ensures that mass
is conserved within the climate system, meaning that air and water cannot be created
or destroyed but can move and change density.
 Basic Equations in General Circulation Models
• The continuity equation ensures that any movement of air or water results in changes
in density or distribution, which is essential for simulating pressure systems, ocean
currents, and atmospheric phenomena.
• Thermodynamic Energy Equation (Energy Conservation): The thermodynamic energy
equation accounts for the changes in temperature and energy within the climate
system.
• It includes terms for heat capacity, advection, latent heat, radiative processes, and
phase changes.
 Basic Equations in General Circulation Models
• The thermodynamic equation is used to model temperature changes and energy
exchange processes, including radiative heating, latent heat from phase changes
(evaporation, condensation), and conduction.
• Equation of State (Ideal Gas Law): the equation of state relates the pressure,
temperature, and density of air, acting as a closure relation for the above equations.
• The equation for the perfect gas law (also called the ideal gas law) is:
PV=nRTPV
Where:
• P= Pressure of the gas (in pascals, Pa)
• V= Volume of the gas (in cubic meters, m³)
• n = Number of moles of the gas (in moles)
• R = Universal gas constant, approximately 8.314 J/mol\k
• T = Temperature of the gas (in Kelvin, K)
• This equation is used to determine the density or pressure of air based on other
variables, providing crucial information for the dynamics of the atmosphere.
 Basic Equations in General Circulation Models
Radiative Transfer Equations
Purpose: The radiative transfer equations govern the absorption, emission, and
scattering of radiation as it passes through the atmosphere.
• They determine the energy balance at the Earth's surface and within the atmosphere.
Equation (simplified form):

These equations are essential for modelling incoming solar radiation, outgoing longwave
radiation, greenhouse gas effects, and cloud interactions, which influence the Earth's energy
budget.
 Basic Equations in General Circulation Models
Moisture Conservation Equation: the moisture conservation equation accounts for the
distribution and phase changes of water vapor in the atmosphere. It describes
evaporation, condensation, and transport processes.

This equation tracks the moisture content in the atmosphere, affecting cloud formation, precipitation, and
latent heat release, which are crucial for simulating weather and climate.
 Basic Equations in General Circulation Models
Tracer Transport Equations: Tracer transport equations describe the movement and
distribution of chemical components (e.g., CO₂, aerosols) in the atmosphere and ocean.

This equation is used to simulate the transport of gases and particles,

which influences radiative forcing, cloud properties, and air quality.
 Basic Equations in General Circulation Models
Parameterization of Subgrid Processes: many important climate processes occur at
scales smaller than the grid size of a climate model and must be parameterized.

Examples include:
Cloud Processes: Cloud formation, condensation, and precipitation are
parameterized using empirical relationships that relate temperature, humidity, and
other factors to cloud cover and rainfall.
Turbulence and Convection: Atmospheric convection and turbulence are
parameterized to represent the mixing of heat, moisture, and momentum on small
scales.
Surface Fluxes: Heat and moisture exchanges between the surface (ocean/land)
and the atmosphere are parameterized to account for processes like
evapotranspiration and sensible heat flux.
 Basic Equations in General Circulation Models
Coupling Between Equations
In climate models, these equations are coupled together to represent interactions
between different components of the climate system:
Atmosphere-Ocean Coupling: the exchange of heat, momentum, and gases
between the ocean and the atmosphere is simulated to represent processes like
ocean heat uptake and carbon cycling.
Atmosphere-Land Coupling: Vegetation, soil moisture, and land use changes affect
surface fluxes of energy and water, impacting climate patterns.
Radiative Feedbacks: Changes in temperature, cloud cover, and greenhouse gases
feed back on radiation, affecting the energy budget.
Numerical Solution of Climate Model Equations
Discretization: The equations are discretized on a three-dimensional grid,
with the atmosphere and ocean divided into cells in the horizontal and
vertical dimensions.
 Basic Equations in General Circulation Models

Finite Difference Methods: Numerical methods, such as the finite difference method, are
used to solve the partial differential equations at each grid cell.
Time Integration: The equations are solved using time-stepping methods, which advance
the solution forward in time from an initial state.
Computational Aspects and Challenges
Resolution: The grid resolution determines how well small-scale processes are captured.
Finer resolution improves accuracy but increases computational cost.
Boundary Conditions: Models require appropriate boundary conditions (e.g., solar
radiation, land surface properties) to accurately simulate the climate system.
Stability and Accuracy: Numerical methods must be stable and accurate. Techniques like
adaptive time-stepping and implicit schemes are used to ensure that simulations are
both efficient and realistic.
 Practical Exercise in Climate Modelling: Investigating
Temperature Trends Using a Simple Energy Balance
Model (EBM)
Objective:
To explore how changing radiative forcing, such as increased
greenhouse gases, affects the global mean temperature. This
exercise introduces the concept of energy balance and how
simplified climate models can provide insight into
temperature trends.
Scenario:
You will use a zero-dimensional Energy Balance Model (EBM)
to estimate changes in global mean temperature in response
to increased radiative forcing, such as from CO₂. The model
balances incoming solar radiation with outgoing infrared
radiation to maintain energy equilibrium.
Model Description
The zero-dimensional EBM assumes that the Earth can be
modeled as a single point, and the temperature is uniform
over the entire surface. The equation for the energy balance
model is:
Basic Equations in General Circulation Models
Basic Equations in General Circulation Models
Basic Equations in General Circulation Models
Basic Equations in General Circulation Models
Basic Equations in General Circulation Models

Discussion
•Impact of Forcing: The exercise demonstrates how an increase in radiative forcing,
such as from increased greenhouse gases, leads to a rise in global temperature.
•Sensitivity: The climate sensitivity can be estimated from this exercise. For a forcing of
4 W/m2, the temperature increase was 1.3 K, suggesting a sensitivity of approximately
0.325 K per W/m².
•Simplifications: This model is highly simplified and does not account for feedback
mechanisms (e.g., water vapor, cloud cover, ice-albedo feedback) or spatial variability.
More complex models (General Circulation Models) are used in actual climate research
to capture these complexities.
Practical Exercise: Applying the Navier-Stokes Equation in Climate
Modelling
Objective:
To understand how the Navier-Stokes equations can be applied to model fluid dynamics within the
atmosphere, specifically looking at wind flow and the impact of temperature gradients on pressure and
velocity.
Background
The Navier-Stokes equations describe the motion of fluid substances and are fundamental to the
understanding of weather patterns, ocean currents, and atmospheric dynamics. In the context of climate
modeling, these equations help simulate how air moves, which is critical for predicting weather and climate
systems.
Navier-Stokes Equation
In three dimensions, the Navier-Stokes equation can be expressed as:

Where:
u: Velocity vector of the fluid (m/s).
t: Time (s).
•ρ: Density of the fluid (kg/m³).
•p: Pressure (Pa).
•ν: Kinematic viscosity of the fluid (m²/s).
•f: External forces acting on the fluid (e.g., gravity).
Practical Exercise: Applying the Navier-Stokes Equation in Climate
Modelling
Practical Exercise: Applying the Navier-Stokes Equation in Climate
Modelling
Practical Exercise: Applying the Navier-Stokes Equation in Climate
Modelling
Practical Exercise: Applying the Navier-Stokes Equation in Climate
Modelling
 Basic Equations in General Circulation Models

Discussion
Stability Considerations: The results highlight that naively applying the Navier-
Stokes equations without considering boundary conditions or initial disturbances
can lead to non-physical results.
Complexity: Real atmospheric models consider more complex interactions, such as
turbulence and stratification, which can stabilize the equations.
Applications: This exercise illustrates the fundamentals of applying fluid dynamics
to climate modeling, focusing on the impacts of pressure gradients and wind flow
in the atmosphere.
 Grid Boxes and Time Steps in Climate Models

• Climate models work by dividing the Earth's atmosphere, oceans, and land
into a 3-dimensional grid.
• These grid cells, often referred to as grid boxes, are the fundamental
building blocks of climate models.
Characteristics of Grid Boxes:
• Spatial Resolution: Each grid box represents a region of the Earth’s surface
with specific horizontal and vertical dimensions.
• The spatial resolution is defined by the size of these grid boxes.
• For instance, a model might have a resolution of 100 km x 100 km in the
horizontal direction, meaning that each grid box represents a 100 km by
100 km area.
• Horizontal Resolution: This refers to the dimensions of the grid box across
the Earth’s surface.
 Grid Boxes and Time Steps in Climate Models
• Vertical Resolution: Models also have a vertical component, dividing the
atmosphere into different layers (e.g., 30 or 50 layers from the surface to the
stratosphere).
• State Variables: Each grid box contains values for different variables that
describe the state of the atmosphere or ocean, such as temperature, humidity,
wind velocity, and more.
• These values are averaged over the entire grid box.
Types of Grids:
• Rectangular Grids: The simplest and most common grid type, used for global
and regional models.
• Curvilinear Grids: These grids may follow geographic features, such as
coastlines.
• Adaptive Grids: In some models, the grid boxes can change size or shape
dynamically to improve resolution in areas of interest (e.g., regions with strong
gradients, like fronts).
Numerical Solution: Time steps and Grid boxes

All the physical processes occurring in the climate system are resolved at
individual grid and the coupling occurs at these grids. Source: NASA
 Grid Boxes and Time Steps in Climate Models
Parameterization
• It is a crucial concept in climate modeling, allowing us to represent physical processes
that cannot be resolved directly by the model due to computational constraints or
scale limitations.
• This is particularly important for processes occurring at scales smaller than the model’s
grid resolution, such as cloud formation, turbulence, and convection.
• Parameterization is the process of representing the effects of sub-grid scale processes
by using simplified relationships or empirical formulas that depend on the resolved-
scale variables.
• Climate models divide the Earth into grid boxes that have limited spatial resolution
(e.g., 100 km x 100 km).
• Many physical processes, such as cloud microphysics, small-scale turbulence, and
radiation interactions, occur at scales much smaller than the size of a grid box and
cannot be directly resolved.
• Resolving these processes explicitly would require extremely high resolution and thus
be computationally expensive and impractical.
 Grid Boxes and Time Steps in Climate Models
Examples of Parameterized Processes
• Convection: Convection involves vertical transport of heat and moisture,
typically occurring in the atmosphere over small areas. Models do not have
enough resolution to explicitly resolve convection in individual clouds, so
convective parameterizations are used to represent their impact on the larger-
scale atmosphere.
• Cloud Formation: Cloud microphysics and cloud formation involve processes
like condensation, evaporation, and precipitation, which occur on very small
spatial scales. Cloud parameterizations approximate the impact of clouds on
atmospheric radiation, precipitation, and energy balance.
• Turbulence and Boundary Layer Processes: The planetary boundary layer (PBL)
is where friction and turbulent mixing are most active. PBL parameterizations
represent the effects of small-scale turbulence on momentum, heat, and
moisture transport between the surface and the atmosphere.
 Grid Boxes and Time Steps in Climate Models

• Radiation: The transfer of shortwave (solar) and longwave (terrestrial)

radiation through the atmosphere is influenced by gases, clouds, and
aerosols.
• Due to the complexity of radiative transfer calculations, radiation
schemes are used to parameterize these effects in an efficient way.
• Surface Processes: Surface exchanges of moisture, heat, and momentum
between the land/ocean surface and atmosphere involve processes like
evaporation, transpiration, and roughness effects.
• These are parameterized to account for the interactions between the
surface and atmospheric boundary layer.
 Grid Boxes and Time Steps in Climate Models
Types of Parameterization
• Empirical Parameterizations: These use empirical relationships derived from
observational data.
• They are often simple to implement but may be less generalizable to different
conditions.
• Example: Using an empirical relationship between surface temperature and
evaporation rate to estimate latent heat flux.
• Diagnostic Parameterizations: These determine the parameterized quantity based on
the current state of resolved variables in the model without considering the evolution
over time.
• Example: Cloud cover calculated as a function of relative humidity in a grid box.
• Prognostic Parameterizations: These involve equations for the evolution of the
parameterized variable over time, adding a dynamic aspect to the simulation.
• Example: Predicting cloud water content as a function of temperature, humidity, and
vertical velocity.
 Grid Boxes and Time Steps in Climate Models
Parameterization Challenges
• Scale Mismatch: The grid scale versus process scale mismatch creates challenges in accurately
representing sub-grid processes.
• The simplifications may lead to biases or inaccuracies.
• Example: Convective clouds are highly localized, but their parameterization assumes an effect
over an entire grid box, which might miss the localized dynamics.
• Complex Interactions: Different parameterizations interact with each other, which can lead to
unexpected behaviors.
• For example, the interaction between cloud and radiation parameterizations may affect
surface temperatures.
• Tuning and Calibration: Parameterizations involve empirical constants or coefficients that
need to be tuned to match observed data.
• The tuning process can be subjective, and a set of parameters that works well in one region or
condition may not work well in another.
• Computational Trade-offs: The accuracy and complexity of parameterizations must be
balanced with computational efficiency.
• More sophisticated schemes are often computationally expensive.
 Grid Boxes and Time Steps in Climate Models
Time Steps in Climate Modeling
• Time steps represent the increments in time at which the model equations are solved.
• The entire evolution of climate variables is simulated by updating the state of each grid
box at each time step.
Characteristics of Time Steps:
• Temporal Resolution: The time step represents how often calculations are made for
each variable.
• A model might use a time step of minutes to hours, depending on the type and scale of
the model.
• Short-term Weather Forecast Models: These models often use smaller time steps
(e.g., seconds to minutes) to capture rapidly changing atmospheric conditions
accurately.
• Climate Models: These models often use longer time steps (e.g., 15 minutes to an
hour) because they are interested in long-term trends rather than immediate
atmospheric changes.
 Grid Boxes and Time Steps in Climate Models
Considerations for Time Steps
• Stability of Numerical Integration: The choice of time step is crucial for the numerical
stability of the model.
• If the time step is too large, the model might become unstable and produce inaccurate
results.
• For stability, models often adhere to the Courant-Friedrichs-Lewy (CFL) condition,
which constrains the time step based on the speed of atmospheric processes and the
size of the grid box.
• Physical Processes: Different processes in the climate system occur at different
timescales.
• For instance:
• Fast Processes: Processes like cloud formation, convection, or atmospheric turbulence
need small time steps to be accurately represented.
• Slow Processes: Processes like ocean circulation, soil moisture changes, or carbon
cycling can be represented with longer time steps.
 Grid Boxes and Time Steps in Climate Models
• The CFL condition is a stability criterion for numerical solutions of partial differential
equations, especially in fluid dynamics. It essentially ensures that the numerical time
step size (Δt) and the spatial step size (Δx) are chosen so that the information does not
move further than one grid cell during a single time step.
• The CFL condition can be written as:

If this condition is not satisfied, numerical solutions may become unstable, leading to incorrect or divergent
results.
 Grid Boxes and Time Steps in Climate Models
Combining Grid Boxes and Time Steps in Climate Models
• In climate models, the dynamical core of the model uses grid boxes and time steps to solve the
fundamental equations of fluid motion (e.g., the Navier-Stokes equations) and energy balance.
Step-by-Step Simulation:
• Initialization: The model is initialized with observed data for all grid boxes, representing the
initial state of the atmosphere, oceans, and land.
• Time Integration: The model equations are then numerically integrated over time, using the
prescribed time steps.
• At each time step: the model calculates fluxes of heat, momentum, and moisture between
neighboring grid boxes.
• Physical processes, like radiation, convection, and condensation, are calculated for each grid
box.
• Update State Variables: After each time step, the state variables (e.g., temperature, pressure,
wind speed) are updated in each grid box to reflect the new state of the atmosphere or ocean.
• Output: The model writes output data at specified intervals (e.g., every 6 hours or every day) to
GRIB files for analysis and visualization.