Climate models are the essential tools that allow scientists to understand past climatic shifts, attribute current warming to specific causes, and project future changes under different emission scenarios. These models are built on fundamental physical principles, including the conservation of energy, mass, and momentum, expressed through mathematical equations that describe the behaviour of the atmosphere, oceans, land surface, and ice. Running on powerful supercomputers, climate models divide the Earth into a three-dimensional grid of cells, sometimes with vertical layers that extend from the deep ocean to the upper stratosphere, and solve these equations forward in time. The resolution of these grids has improved dramatically over recent decades, moving from hundreds of kilometres to around twenty-five kilometres for global models and even finer for regional downscaling, allowing for more realistic representations of weather systems, topography, and coastlines.
Advertisement
The physical processes represented in climate models range from the large-scale circulation of the atmosphere and ocean currents to the complex interactions of clouds, aerosols, and radiation. Some of these processes, such as the general flow of westerly winds in mid-latitudes, are well understood and can be expressed with high confidence. Others, including the formation and dissipation of clouds, occur at scales far smaller than the model grid cells and must be parameterised—represented by simplified, empirically derived formulas. This parameterisation introduces uncertainty, as small changes in cloud cover can have a significant amplifying or dampening effect on warming. Aerosol particles from volcanic eruptions, industrial emissions, and dust also interact with radiation and serve as cloud condensation nuclei, adding further layers of complexity. The Intergovernmental Panel on Climate Change (IPCC) regularly assesses the performance of dozens of models from independent research centres worldwide, comparing their outputs against historical observations to gauge reliability.
One of the most powerful demonstrations of model reliability comes from their ability to simulate the observed warming of the past century and a half only when both natural and human-caused forcings are included. Models driven solely by volcanic and solar variability fail to reproduce the sustained temperature increase seen since the mid-twentieth century. When greenhouse gas emissions from fossil fuel combustion, land use change, and industrial processes are added, the simulated temperature curves match the instrumental record closely. This attribution fingerprint extends beyond surface temperatures to include the cooling of the stratosphere, the warming of the ocean’s subsurface layers, and the reduction of Arctic sea ice, each of which bears the distinct signature of an enhanced greenhouse effect. Such multi-variate agreement between models and observations provides a robust basis for the conclusion that human activity is the dominant driver of recent climate change.
