From Babylon to Google: a history of weather forecasting

Recent scientific breakthroughs allowing forecasters to better predict the weather are just the latest in a long line of meteorology developments.

Google’s artificial intelligence (AI) arm DeepMind has developed a system allowing forecasters to predict the chance of rain within the next couple of hours with much higher precision.

But people have been attempting to work out whether it is going to chuck it down or not for thousands of years. As far back as 650BC, the Babylonians, in modern-day Iraq and Syria, tried to divine the weather based on cloud patterns and astrology.

By 350BC the Greek philosopher Aristotle was describing weather patterns in texts, while even Jesus Christ himself had a crack at forecasting in the New Testament.

However, it was in the 19th century that the science of weather forecasting truly began, with the invention of the electric telegraph in 1835. Soon after, the Royal Navy officer Francis Beaufort developed the wind force scale. It would later become known as the Beaufort scale.

The sinking of the Royal Charter ship in a storm off the north coast of Anglesey in October 1859 inspired Beaufort’s Royal Navy protege Robert FitzRoy to develop weather charts which he described as “forecasts”, the first known usage of the term.

He went on to help set up 15 land stations, which used a telegraph to transmit daily weather reports and led to the creation of a gale warning service. In 1861, the first daily weather forecasts were published in the Times.

In the 20th century, advances in understanding atmospheric physics led towards modern-day numerical weather prediction.

Since then, forecasting techniques have included analysing data relating to pressure, air speed, precipitation and temperature. These are collected from around the world and fed into supercomputers for analysis.

Hannah Fry, a professor at University College London’s Centre for Advanced Spatial Analysis, explained that 21st-century weather forecasting services all interpret the data differently, leading to variable forecasts.

Writing in the New Yorker magazine in 2019, she said: “They don’t yield sensible answers like ‘it will rain on Wednesday in Soho’ or ‘hailstorms will start at 2pm’.”

She added: “The apps differ wildly in how they handle all that uncertainty, and some will be much more pessimistic than others.”

As technology continues to evolve, more observations will be captured, allowing more complex equations to provide forecasts for smaller areas at a greater speed.