Step one: measure the whole sky
Every forecast starts with a snapshot of the atmosphere as it is right now. That snapshot comes from tens of thousands of surface weather stations, weather balloons launched twice a day from hundreds of sites, airliners reporting conditions along their routes, ocean buoys, radar, and satellites watching clouds, moisture and temperature from orbit. No single country could pay for all of it, so the data gets shared worldwide through the World Meteorological Organization. It's one of the quieter success stories of countries working together.
Step two: run the physics forward
The atmosphere obeys physical laws, mainly fluid dynamics and thermodynamics. A weather model splits the sky into a three-dimensional grid of millions of cells and applies those laws over and over to push the whole system forward a few minutes at a time, out to hours and then days. Doing that fast enough to be useful takes some of the biggest supercomputers on Earth. The best-known global models are Europe's ECMWF and America's GFS. Regional models add detail for the next day or two. Forecast sites and apps, including the weather page here, which uses the open Open-Meteo service, show a blend of those model outputs for your exact spot.
Why forecasts go fuzzy after a few days
In the 1960s a meteorologist named Edward Lorenz found that tiny differences in the starting snapshot, smaller than any instrument can measure, grow into completely different weather within about two weeks. That's the butterfly effect, and it puts a hard ceiling on forecasting. No amount of computing power will ever give you a reliable day-by-day forecast a month out. Modern forecast centers deal with this by running ensembles: dozens of forecasts from slightly different starting points. When the members agree, confidence is high. When they scatter, an honest forecast says so.
What "70% chance of rain" means
It does not mean it will rain for 70% of the day, or over 70% of the area. It means that in setups like this one, measurable rain reaches your spot about 7 times in 10. In practice: under 20%, leave the umbrella. 30 to 60%, have a plan B. Over 70%, assume rain and be pleased if it doesn't.
How far ahead to trust it
- Days 1 to 3: temperature and rain forecasts are solid. Plan on them.
- Days 4 to 7: the broad pattern, warm or cold, settled or stormy, is usually right. The exact timing and amounts wobble. Check again as the day gets closer.
- Past day 7: treat it as a trend, not a promise. Even the 7-day view on the forecast page is best read as the shape of the week.
It's worth remembering how far this has come. Today's 5-day forecast is about as accurate as a 1-day forecast was in 1980. The forecast on your phone is one of the most sophisticated things modern science produces, refreshed several times a day, for free.