It’s partially done by looking at historical norms for the region (e.g. you know it’s more likely to rain when the wind is from a certain direction etc), and partially done by using big computers to analyse weather patterns and predict how they’ll change. Accuracy for short-term forecasts isn’t really guaranteed after three or four days, though, and that’s because the weather is a chaotic system–that doesn’t mean it’s entirely impossible to predict, but that the slightest error in your starting conditions will rapidly diverge from the actual weather.
To give an example: let’s say you had a magic device that could measure the wind speed, temperature and air pressure at every point one metre apart in the atmosphere. You feed those values into a perfect supercomputer that can exactly simulate every interaction there is and let it run. The resulting prediction will *still* be totally wrong after a month, because you didn’t measure the bits in between those 1m points and have no idea what was happening there!
Latest Answers