There are now three basic methods of predicting weather.
The first, synoptic forecasting, combines information from weather stations within a region and extrapolates, or takes from that information, likely weather patterns.
The data is studied by meteorologists who use computers to generate short-term predictions of wind, temperature, air pressure, and humidity changes, and possible storm fronts.
Statistical forecasting is based on historical knowledge of weather and its patterns in a given area. Meteorologists apply mathematical equations created from studying past weather patterns to foretell likely changes in the next 5 to 90 days.
Numerical forecasting stems from the knowledge of physics and the atmosphere. Mathematical equations are drawn from the physical laws of nature, such as Newton’s laws of motion and thermodynamics (the conservation of energy), applied to weather conditions.
Computers work out the mathematics—the equations were impossible to solve in a timely fashion before computers existed—and the information is primarily used for 5-day forecasts.