The climate of the future: what scenarios do scientists already know?
· 8 min read
How can scientists estimate what the climate might look like decades—or even centuries—from now? They use mathematical models of the Earth system to explore different possible futures, from scenarios in which greenhouse gas emissions fall rapidly to those in which they remain very high.
Climate models combine fundamental laws of physics with enormous volumes of observational data. Increasingly, artificial intelligence is also being used to make these models faster and more detailed. But how reliable are climate projections, what do they tell us about Europe, and could humans ever deliberately influence the climate?
The planet has been warming since the late 19th century, primarily because human activities have increased the concentration of greenhouse gases in the atmosphere. Carbon dioxide, methane and other greenhouse gases absorb part of the infrared radiation emitted by the Earth's surface, reducing the amount of energy that escapes into space.
As a result, additional energy accumulates within the climate system. Most of this excess heat is absorbed by the oceans, while the atmosphere, land and cryosphere also warm.
One consequence is rising sea levels, driven by both the thermal expansion of seawater and the melting of glaciers and ice sheets. Climate change is also altering the frequency and intensity of many extreme events, particularly heatwaves, droughts, heavy rainfall and flooding. Cold extremes still occur, but they are generally becoming less frequent and less severe as the global climate warms.
Europe deserves particular attention. According to the 2025 European State of the Climate report from the Copernicus Climate Change Service and the World Meteorological Organization, Europe has been warming roughly twice as fast as the global average since the 1980s. Over the 30 years from 1996 to 2025, European temperatures increased at around 0.56°C per decade, making Europe the fastest-warming continent.
The effects are already visible from the Mediterranean to the Arctic. In 2025, at least 95% of Europe experienced above-average annual temperatures, while the continent also faced major heatwaves, drought, glacier loss and record wildfire activity in some regions.
What scientists do not expect is a scenario similar to the disaster film The Day After Tomorrow, in which the global climate changes almost beyond recognition within days or weeks. Climate systems can contain tipping points and abrupt changes, but a planet-wide shift of the kind portrayed in the film is inconsistent with the physics and timescales of the processes involved.
For people, shorter-term weather variability can nevertheless feel increasingly disruptive. Large-scale atmospheric circulation can transport unusually warm or cold air across Europe, producing substantial temperature swings over relatively short periods.
It is important, however, to distinguish weather from climate. A sudden cold spell, heatwave or temperature swing cannot automatically be attributed to climate change. Scientists use statistical attribution studies to determine whether global warming has changed the probability or intensity of a particular type of event.
There is no single prediction for the climate of 2100.
Future warming will depend on factors such as population, economic activity, energy demand, technological development, land use and, above all, how quickly humanity reduces greenhouse gas emissions.
Scientists therefore use scenarios rather than a single forecast.
In the IPCC's very low-emissions scenario, SSP1-1.9, global carbon dioxide emissions fall rapidly and reach net zero around the middle of the century. The best estimate of global warming for 2081–2100 relative to 1850–1900 is around 1.4°C.
Under the very high-emissions SSP5-8.5 scenario, the best estimate rises to approximately 4.4°C, with a very likely range of 3.3°C to 5.7°C. An intermediate scenario, SSP2-4.5, produces a best estimate of about 2.7°C.
These numbers may appear small compared with the temperature changes people experience between seasons or even between day and night. For the global climate system, however, a change of several degrees in the planet-wide annual average represents an enormous redistribution of energy.
The difference between a world that stabilises close to 1.5–2°C of warming and one approaching 4°C or more would be profound for sea levels, ecosystems, agriculture, water availability, human health and the frequency of extreme heat.
Higher warming also increases the risk of crossing climate tipping points—thresholds beyond which parts of the Earth system may undergo large, self-sustaining or extremely difficult-to-reverse changes.
Climate models therefore do not tell us that one particular future will happen. Instead, they show how the climate system is likely to respond under different choices and emissions pathways.
At the heart of a climate model are mathematical equations describing the physical behaviour of the atmosphere, oceans, land and ice.
They include equations governing the movement of air and water, conservation of mass and energy, temperature, pressure and moisture.
One of the underlying principles is familiar from school physics: Newton's second law, which describes the relationship between force, mass and acceleration. But modern Earth-system models go far beyond a small set of basic equations.
They also represent processes such as:
To run such a model, scientists divide the atmosphere and oceans into a three-dimensional computational grid.
For every grid cell, powerful computers repeatedly calculate how temperature, pressure, humidity, wind, ocean currents and many other variables evolve over time.
The result is not simply a statistical extrapolation of past temperatures. The model simulates the behaviour of the climate system according to physical laws.
Scientists can then change the model's external conditions—for example, atmospheric CO₂ concentrations—and examine how the simulated Earth responds.
Global climate models have to simulate the entire planet for decades or centuries. Even on some of the world's most powerful supercomputers, there is therefore a limit to how detailed they can be.
A typical global model may have atmospheric grid cells tens of kilometres across, while some lower-resolution models use cells around 100 km wide. A city such as Berlin, Paris or Madrid may therefore occupy only a small part of a single model cell.
This is a problem when researchers need information for urban planning, flood protection, agriculture or infrastructure.
Artificial intelligence can help bridge the gap.
Machine-learning models can learn relationships between large-scale atmospheric conditions and local observations, allowing scientists to downscale global climate simulations into much higher-resolution regional information.
For example, a global model might calculate conditions for a 100 × 100 km area. AI-assisted downscaling can help estimate how temperature, precipitation or wind may differ between individual valleys, cities or coastal areas within that grid cell.
AI is increasingly being explored for several other climate applications as well, including faster weather forecasting, representation of processes such as clouds, identification of extreme events and the creation of computationally inexpensive approximations—or emulators—of traditional physical models.
This does not mean replacing physics with AI. In many of the most promising applications, machine learning complements physical models and observational data rather than operating independently.
Humanity already influences the climate—unintentionally—through greenhouse gas emissions and changes in land use. Deliberately controlling the entire climate system would be a very different challenge.
Directly manipulating a system as large as the Earth's atmosphere and oceans would require interventions on an enormous scale.
Nevertheless, researchers are studying technologies collectively known as solar radiation modification, which would attempt to reduce the amount of solar energy absorbed by the planet.
One frequently discussed concept is stratospheric aerosol injection. Reflective particles would be introduced into the upper atmosphere to scatter a small proportion of incoming sunlight back into space.
Researchers have modelled different possible materials, ranging from sulphate aerosols to highly reflective particles. Some theoretical studies have even examined materials such as diamond dust because of their optical and chemical properties.
Other speculative proposals involve brightening marine clouds or deploying enormous reflective or light-blocking structures in space.
None of these concepts currently offers a proven method for safely controlling the global climate. Their potential side effects, costs, governance requirements and geopolitical implications remain major unresolved questions.
Most importantly, solar radiation modification would not remove CO₂ from the atmosphere. It would therefore not directly solve problems such as ocean acidification and would require continued intervention to maintain its cooling effect.
Changing weather over a limited area is a somewhat different question.
Scientists can use atmospheric models to investigate whether modifying local conditions could theoretically affect cloud formation or rainfall. Experiments have examined how changes in vertical air movement, aerosols or cloud microphysics might influence precipitation in particular regions.
There are also real-world weather-modification techniques such as cloud seeding, although their effects are far more limited and uncertain than the idea of "controlling the weather" might suggest.
Regional experiments can nevertheless be valuable. They allow researchers to test how sensitive rainfall and cloud formation are to specific changes without implying that entire climate systems could be easily engineered.
Climate modelling is ultimately less about predicting exactly what the weather will be on a particular day in 2080 than about understanding risks.
A model can show how Europe's probability of extreme heat changes in a warmer world, where drought risks may increase, how heavy rainfall could intensify, how glaciers may retreat or how sea levels might evolve along European coastlines.
The further into the future scientists look, the more the outcome depends on decisions that have not yet been made.
That uncertainty is not a weakness of climate science. It reflects the fact that part of the future climate is still determined by human choices.
The physics establishes the range of possible outcomes. Emissions, technology and policy will help determine which part of that range the world eventually experiences.
illuminem Voices is a democratic space presenting the opinions of leading Sustainability Thought Leaders, their views do not necessarily represent those of illuminem.
The world needs sustainability knowledge. At illuminem, no interest group or shareholder can influence our work. Thank you for supporting our mission to make high-quality and independent sustainability information free for all. Every contribution helps. Thank you for donating today.
Yury Erofeev
Climate Change · AI
Robert Muggah
Public Governance · Corporate Governance
illuminem briefings
Ethical Governance · Climate Change
The Conversation
Effects · Climate Change
Climate Home News
Effects · Climate Change
Wired
Effects · Climate Change