Weather and climate are most often predicted using numerical models.
Numerical models start from a snapshot of the current atmospheric conditions in the area of interest, from the surface to the upper atmosphere, at points on a three-dimensional grid.
A set of atmospheric variables, such as wind speed, temperature, pressure and humidity in each grid box are stored and a set of equations are solved for each grid box to predict the values at that point a short time later.
A high-resolution rainfall forecast
Image above: An example of a high-resolution rainfall forecast from the Met Office UK variable resolution (UKV) model. The Met Office runs this model with a 1.5 km grid length over the UK in order to produce improved forecasts of small-scale weather such as thunderstorms and winter snow. Read more about the UK variable resolution (UKV) forecast model in this Research News article from 2012.
Forecast lead time
The process of generating the forecast is repeated many times; each time the forecast stepping a few minutes further into the future to produce either a weather forecast for the next few days or a climate prediction for the coming 100 years.
The set of equations which are solved when creating a forecast fall broadly into two categories.
The dynamical core solves the equations of motion for a fluid, on a rotating sphere, to calculate the evolution of the atmospheric flow.
Alongside the dynamical core, a large number of other atmospheric processes and parametrizations operate to:
- warm/cool or moisten/dry the atmosphere;
- form clouds and precipitation;
- represent both the weather which we experience and the effect of that weather on the evolution of the atmospheric flow.