'Great Storm could be repeated'

In the early hours of October 16, 1987, winds peaked at more than 120mph, killing 18 people, damaging buildings and felling 15 million trees in the south-east of England. At the time, people believed the storm - the worst since 1703 - was a one-in-200-year event, but just over two years later the country was hit by another vicious storm on Burns Day, 1990.

And with climate change making it more likely we will see such events more often, those at the frontline have used the 1987 storm to learn lessons on how to cope with violent weather.

The Met Office believes forecasters would be unlikely to fail to predict the Great Storm and warn people about the dangers if it happened now - in the way they did in 1987. But they have warned errors were still possible in predicting the weather and that more powerful data and computer systems are needed to pinpoint weather on a local level - for example to predict the rainfall which caused the Boscastle flooding in 2004.

Technological improvements - including computers which are several hundred times more powerful than those used in 1987 - would make it "highly likely" the storm would be predicted today, chief meteorologist Ewen McCallum said.

Scientists have also since learned that the phenomenon which caused the devastating winds was a "sting jet" of high momentum, high energy air which acts like a scorpion's tail lashing through the area of depression.

The sting jet was only 30 miles wide - the size of just one grid point on the models forecasters used at the time - but created the winds which hit a top speed of 122mph on the night of October 15/16.

"I'd like to think, should a great storm or major depression occur again, the risk of it happening would be picked up and communicated to the public rather than the very deterministic 'yes' or 'no' we had in 1987," he said.

The Met Office has re-run simulations of the storm with current forecast models around 50 times, and the "vast majority" of the runs predicted the Great Storm. But Mr McCallum warned: "With the super-computers we've got it's highly likely we'd predict the storm but there still could be errors."

Just a small change in the data can generate a very different response, so meteorologists run a series of models to try and predict the risk of an event happening. The biggest changes in weather forecasting in the past 20 years include satellite data, better radar networks for ground level weather information and more powerful computers.