Course 3, lesson 27 of 100, Ages 7+
Forecasting the weather
How AI helps predict rain and storms
Like I’m 5
Weather forecasters study clouds, wind and temperatures to guess tomorrow's weather. AI helps them spot patterns in mountains of weather data.
The big idea
Weather satellites, balloons and ground stations measure temperature, wind, pressure and humidity all over the world, all day. That's far too much data for people to read by hand.
Traditional forecasts use physics equations run on supercomputers. Newer AI models learned from decades of past weather and can produce forecasts in minutes instead of hours. Forecasters combine both to warn people early about storms.
Examples
- Rain alerts: 'Rain starting in 10 minutes' comes from patterns in radar data.
- Cyclone warnings: Earlier, better forecasts help people move to safety.
- Farmers: Knowing when rain is coming helps them plan sowing and harvesting.
How it works
- Sensors measure the weather everywhere.
- Models compare today with patterns from past weather.
- Forecasters check the predictions and share warnings.
Check your understanding
- Why does weather forecasting need computers?
- Options: There's far too much data for people to read by hand; Clouds are made of computers; Forecasts are random.
Answer: There's far too much data for people to read by hand. Computers and AI process millions of measurements quickly. - What does '70% chance of rain' mean?
- Options: Rain is likely, but not certain; It will rain for 70 minutes; 70% of the sky is blue.
Answer: Rain is likely, but not certain. Forecasts give chances, because the weather is never perfectly predictable.
Remember
Forecasts combine huge amounts of weather data with physics and AI to predict what's likely.
Talk about it
Look at today's forecast. How sure is it, and was it right?
Go deeper
Machine-learning weather models such as GraphCast and similar systems are trained on decades of reanalysis data. They can match or beat traditional numerical models on many measures while running far faster.