How supercomputers turn current weather observations into tomorrow's forecast using complex mathematical models.

Nature · Published 20 October 2026 Swipe more like this →

In short: Modern weather forecasting relies on numerical weather prediction, a process that feeds real-time atmospheric data into powerful supercomputers. By solving complex equations, these systems simulate the future state of our planet's atmosphere. While these models are incredibly sophisticated, they remain limited by the chaotic nature of the weather itself.

Numerical weather prediction (NWP) transforms current observations—gathered from satellites, radiosondes, and other systems—into future forecasts. By assimilating this data into a mathematical framework, computers predict hundreds of meteorological elements, from surface temperatures to precipitation. The Met Office, for instance, utilizes the Unified Model across various configurations to produce both global medium-range forecasts and high-resolution regional predictions. These regional models are particularly adept at representing intricate surface features like coastlines and orography, which are essential for localized accuracy.

The process involves complex physical simulations, but it is not without constraints. Because the atmosphere is governed by chaotic partial differential equations, exact solutions are impossible to calculate. Small errors in initial conditions inevitably grow over time, which is why ensemble forecasting has become a vital tool. By running multiple versions of a model with slightly perturbed initial conditions or physical parameters, meteorologists can quantify uncertainty and gauge the confidence level of a specific forecast. This approach helps extend the useful range of predictions beyond what a single deterministic model could achieve.

Operational systems, such as those implemented in the Met Office's Operational Suite cycle 47, distinguish between weather prediction and climate modeling. While climate models might incorporate long-term factors like atmospheric chemistry or the carbon cycle, NWP focuses on shorter timescales where these processes have less immediate impact. Instead, NWP prioritizes high-resolution data assimilation to capture dynamic shifts. Despite these advancements, the fundamental limit of predictability remains a challenge, as the chaotic nature of the atmosphere dictates that even with perfect models and data, accurate forecasting is likely capped at approximately 14 days.

General information, not professional advice. Spotted an error? Flag the card or see corrections and complaints.

Intellect Plus is a visual educational discovery platform for exploring fascinating facts and ideas. Brain Food, Not Brain Rot. What is Intellect Plus?

Related

• The massive supercomputers predicting the atmosphere's next 16 days
• Why even the most advanced supercomputers cannot perfectly predict tomorrow's weather
• Without parallel computing, tomorrow's weather would be predicted 1,000 years later
• How do scientists see a century into the future?
• Can a machine learning model outperform massive, expensive weather forecasting supercomputers?
• Oktibbeha County, Mississippi, derives its name from a Choctaw phrase meaning icy creek.
• Decommissioning a nuclear plant is a multi-decade, billion-dollar challenge with no easy exit.
• From Cape Town mutual insurer to digital banking: the evolution of Old Mutual
• A Mississippi county named for a Choctaw word meaning to stink faces complex history.
• The high-altitude air war where physics and improvisation met
• Earth's highest clouds shine at night because they are made of frozen space dust
• Earth stays cool by shedding invisible heat into the vastness of space

More in Air · More in Nature · Latest · A–Z · All topics