But during the First World War a group of Norwegian meteorologists realised that the way to forecast the weather was to work out the movement of the atmosphere as a set of mathematical equations.

The idea was to create a mathematical model of the atmosphere - put in the weather observations, work out the equations and hey presto! A forecast by numbers - or numerical weather prediction (NWP) to give it its correct title.

In theory this is easy to do. The laws which govern the way the air and water vapour in the atmosphere behave are fairly straightforward.

The first person to try it was Lewis Fry Richardson back in 1922. He got a team of mathematicians together to work out the sums. But their first attempts were spectacularly unsuccessful.

Richardson and his team set themselves the goal of predicting the weather for a particular day several days ahead. Unfortunately, they were still working out their calculations long after the day in question.

Even more disappointingly their forecast bore little resemblance to the day for which they were trying to forecast. On the face of it Richardson’s attempt had been a total flop!

Yet, 80 years later a wing of the Met Office headquarters at Bracknell was being named after Richardson. This is because, many years after, it was recognised that (NWP) was indeed the way forward. Richardson and his team had failed, not because they had used the wrong equations but because they, like the rest of us, made simple arithmetical errors.

Their forecast had taken so long to produce because what they really needed was a team of thousands of mathematicians to work out the calculations - or a supercomputer.

The first commercial computers could manage only about 1000 operations per second. Nowadays they can do several trillion and the Met Office has some of the largest computers in the world. But despite their increasing sophistication, they still use the same principles that Richardson applied all those years ago.

In fact when Richardson’s work was re-examined many years later and his calculations were computed correctly, the forecast was found to be correct. The accuracy of weather forecasts increases year on year.

Whilst improved satellite, radar and observational data play their part in improving the forecasts, the continuing improvements in computers and the way we model the weather using those same basic equations are the key factors.