Climate system

Global climate models

Climate models are mathematical representations of the Earth that help us predict changes in the climate.

They take into account conditions affecting the main processes in our climate system; these effects are based on observations and on established physical laws such as the conservation of mass, energy and momentum.

Climate models provide simplified versions of the real world that can be used to test our understanding of how the climate system will respond to changes in conditions. They allow us to estimate the response, for example, to more volcanic eruptions, less sunlight and more greenhouse gas emissions.

Schematic of Global Climate Model
Climate models are systems of differential equations based on the basic laws of physics, fluid motion, and chemistry. To “run” a model, scientists divide the planet into a 3-dimensional grid, apply the basic equations, and evaluate the results. Atmospheric models calculate winds, heat transfer, radiation, relative humidity, and surface hydrology within each grid and evaluate interactions with neighbouring points. (NOAA)

Global climate models (GCMs) work by dividing the global climate system into three-dimensional boxes, or grid cells, of different sizes. This division occurs both horizontally (like the pages of a street directory) and vertically (from the land surface up into the atmosphere and down to the bottom of the ocean). Horizontal grid cells are typically 100 to 250 km per side, whereas vertical grid cells correspond to known layering of the air or ocean.

GCMs produce outputs for each grid cell, such as temperature, precipitation, pressure, humidity and wind speed. The climate can be modelled as it evolves over periods of time, which is why GCMs are sometimes known as ‘dynamical’ models.

Limitations of global climate models for regional planning

GCMs have some important limitations in providing projections suitable for regional planning.

Simulating the climate across the globe requires an enormous volume of data to be gathered from grid cells. To manage this task, the number of cells needs to be limited, so that each covers hundreds of square kilometres.

Such a large-scale matrix makes GCMs unsuitable for projecting climate conditions at a regional scale. They are not able to represent accurately the details of the geographical features that influence climate, such as coastlines and mountain ranges.

There are a range of techniques available for overcoming the limitations of GCMs so that they can be used to generate regional-scale climate projections.

First, climate projections can be based on simulations from a range of different GCMs. Although this means there is no single answer to the question of what the future climate of a region may be like, it better reflects the range of possible future climates that arise because of climate modelling uncertainty. This is particularly important for users wanting to use the projections to manage risk.

Second, the output of a GCM can be translated into finer-scale data for a particular region of the globe through a process called ‘downscaling’, of which there are three main types:

Scaling observations is a very simple method that uses GCM output to generate future climate conditions at regional scales:  future climate changes from a GCM are used to adjust current observations from a region. This technique can be easily applied to many GCM simulations without the need to use powerful computers. However, this simple technique may not capture the interactions of important regional climate processes that can affect regional climates.

Statistical downscaling takes as its foundation data on the existing climate, rather than the physical principles of climate drivers. The existing climate might be represented by a series of temperature and rainfall measurements in a particular location over time: a ‘time series’. Statistical techniques are used to apply changes to this time series on the basis of future climates from GCM outputs. A drawback of this method is that it assumes that the relationship between weather variables and the main climate drivers will remain the same in the future, and this may not be the case.

Dynamical downscaling uses regional climate models (RCMs), which are built by using the same principles as GCMs but contain additional information about a range of factors, such as topography, that are important in determining regional climate. Because they are built on physical principles, dynamical RCMs allow for changes in the existing relationship between weather variables and climate drivers. Dynamical models create a new time series rather than adjusting an existing one.

For more information, see the Intergovernmental Panel on Climate Change’s FAQ section, question 9.1: ‘Are Climate Models Getting Better, and How Would We Know?’