Mathematical modelling serves as an indispensable tool in climate science, enabling researchers to simulate and understand the Earth's complex climate system. Models are built upon fundamental physical laws, including thermodynamics, fluid dynamics, and radiative transfer, which are expressed as mathematical equations. These equations describe interactions between the atmosphere, oceans, land surfaces, and ice sheets. By solving these equations numerically using powerful computers, scientists can create virtual representations of the climate system. These models allow for experiments that are impossible in the real world, such as examining the effects of doubling atmospheric carbon dioxide or altering solar radiation.
Without such models, it would be exceedingly difficult to predict future climate conditions or attribute observed changes to specific causes. They provide a framework for integrating diverse observations and testing hypotheses, forming the bedrock of modern climate research and informing international policy decisions. Climate models vary in complexity, ranging from simple energy balance models to sophisticated general circulation models and Earth system models. Energy balance models represent the Earth's energy budget with a few equations, useful for conceptual understanding. General circulation models (GCMs) divide the globe into a three-dimensional grid and simulate atmospheric and oceanic circulation patterns using fluid dynamics equations.
These models capture large-scale features like jet streams and ocean currents. Earth system models (ESMs) extend GCMs by incorporating biogeochemical cycles, such as the carbon cycle and nitrogen cycle, and interactions with vegetation and ice sheets. Each type of model serves a specific purpose: GCMs are used for weather prediction and short-term climate projections, while ESMs are essential for long-term studies of climate change and feedbacks. The choice of model depends on the research question and computational resources available. Data from diverse sources—satellites, weather stations, ocean buoys, and ice cores—are used to initialize and constrain climate models.
They provide a framework for integrating diverse observations and testing hypotheses, forming the bedrock of modern climate research and informing international policy decisions.
For instance, sea surface temperatures, atmospheric pressure fields, and greenhouse gas concentrations are input to simulate the current state. Models then project future scenarios based on greenhouse gas emission pathways, such as those from the Shared Socioeconomic Pathways (SSPs). These simulations generate outputs like temperature changes, precipitation patterns, and sea level rise. By comparing model runs with different assumptions, scientists can assess the range of possible outcomes. A critical component is the representation of processes that occur at scales too small to be resolved directly, such as cloud formation and turbulence.
These processes are 'parameterised' using empirical relationships, introducing some uncertainty. Nevertheless, models have successfully reproduced past climate changes, building confidence in their projections. Despite their power, climate models have inherent uncertainties. One major source arises from the parameterisation of sub-grid-scale processes, as mentioned. For example, cloud formation remains a challenge because it involves complex microphysics that cannot be explicitly simulated in coarse grid cells. Different models may parameterise clouds differently, leading to variations in climate sensitivity—the equilibrium temperature response to doubled CO2. Another uncertainty stems from natural variability, such as volcanic eruptions and solar cycles, which can temporarily offset or amplify human-induced changes.
Additionally, models rely on future emissions scenarios, which are not deterministic but depend on human choices. To address these uncertainties, scientists use ensemble simulations, running a model multiple times with slightly different initial conditions or parameter values. The spread in outcomes provides a probabilistic forecast. Despite limitations, models consistently project warming and increased frequency of extreme events, lending robustness to their conclusions. Key findings from climate models have profoundly shaped our understanding of global warming. Models consistently show that the planet has warmed by approximately 1. 1°C since pre-industrial times, with most warming attributable to human activities, particularly greenhouse gas emissions.
They project that under high-emission scenarios, global temperatures could rise by 3–5°C by 2100, leading to severe impacts such as more intense heatwaves, droughts, and flooding. Models also indicate that Arctic sea ice is declining rapidly, and that sea levels will continue to rise due to thermal expansion and melting ice sheets. Furthermore, they predict shifts in precipitation patterns, making wet regions wetter and dry regions drier. These projections have been validated by observed trends, such as the increase in record-breaking temperatures and the loss of glacier mass.
Thus, models serve as early warning systems, highlighting the urgency of reducing emissions. Validation of climate models involves comparing their outputs with historical observations, a process known as hindcasting. For instance, models are run for the 20th century using known forcings (e. g. , solar variability, volcanic eruptions, and anthropogenic emissions) and their results are checked against temperature records, rainfall data, and ice extent. A model that accurately reproduces past climate—including the cooling after major volcanic eruptions like Mount Pinatubo in 1991—is more trusted for future projections. To further improve reliability, scientists use multi-model ensembles, averaging outputs from many independent models to reduce individual biases.
The Coupled Model Intercomparison Project (CMIP) coordinates this effort, with dozens of modelling centres worldwide contributing. Through these rigorous validation exercises, confidence in model projections has increased, though refinements continue. Uncertainties are quantified and communicated transparently, allowing policymakers to make informed decisions under risk. Ultimately, mathematical models are essential for informing climate policy and adaptation strategies. Governments and international bodies, such as the Intergovernmental Panel on Climate Change (IPCC), rely on model projections to assess risks and set emission reduction targets. For example, the Paris Agreement's goal to limit warming to well below 2°C is underpinned by model calculations showing the carbon budget—the total allowable CO2 emissions to stay within that threshold.
Models also guide adaptation planning, such as designing coastal defences against projected sea level rise, modifying agricultural practices for changing rainfall patterns, and preparing health systems for heat-related stresses. Ongoing research aims to reduce uncertainties, improve spatial resolution, and incorporate more processes (e. g. , ice sheet dynamics). As computational power grows, higher-resolution models will provide even more detailed regional projections. In summary, mathematical modelling is not just a technical exercise; it is a cornerstone of our collective response to climate change.