By Mark Kinver
Science and nature reporter, BBC News
The Met Office has said advances in technology and forecasting mean its failure to detect the Great Storm is unlikely to be repeated.
The agency was strongly criticised for failing to warn the nation about the arrival of the hurricane-force winds in the early hours of 16 October 1987.
By dawn, 18 people had been killed and 15 million trees uprooted.
But better satellite data, computer models and communications have improved things immeasurably, forecasters say.
Forecasts earlier in the week preceding the storm had predicted severe weather. However, it was expected to track its way along the English Channel and miss mainland Britain.
But in the early hours of 16 October, the weather system veered north-eastwards, taking it across the south-eastern corner of England.
The region was rocked by 50mph (80km/h; 43 knots) mean wind speeds and gusts between 90 and 122mph (144 and 196km/h; 78 and 106 knots).
In approximately four hours, the worst storm since 1703 had left a trail of devastation estimated to cost more than £1bn.
The Met Office's chief meteorologist, Ewen McCallum, admitted that they had been "caught out".
"It is still not clear as to exactly why it was a one-in-300-year event, as opposed to a one-in-50-year event," Mr McCallum explained.
"In 1987, there was obviously a very strong temperature gradient."
These gradients are caused by the difference in temperature in equatorial regions, where the Earth absorbs most of the solar energy that warms the planet, and those at the poles.
Energy from the Sun warms air over the equatorial regions, which is then transferred to cooler polar regions.
GREAT STORM TIMELINE
Five days before
Forecasters predict severe weather for Thursday or FridayFew days before
Models suggest severe weather will only hit Channel and coast15 October (afternoon)
Winds very light over UK; gale warning issued for Channel15 October (evening)
Late TV forecast focuses on heavy rain over UK mainland16 October (early hours)
Storm turns inland; emergency services are alerted
Larger differences in temperature between the two areas result in a steeper temperature gradient, leading to stronger winds.
"In this particular case, there were some remnants of high momentum air at high levels from Hurricane Floyd," he recalled.
"Some of the energy (from Floyd) at higher levels was maintained, and that basically helped increase the thermal contrast in the atmosphere.
"But why it was unprecedented is very difficult to say, to be honest. It was a fairly typical type of storm that we get in this country several times a year, but it was just more severe."
One contributing factor, he suggested, could have been a "sting jet", which is air that descends from the stratosphere; the downward momentum of the air feeds more energy into a weather system.
Later studies of satellite images of the storm showed evidence of a sting jet, but the phenomenon's role in strengthening mid-latitudinal storms was unknown in 1987.
21st Century forecasting
"Twenty years ago, scientists would have said that the main source of information would have been from [weather balloons]. Now, they'd tell you it is from satellites," Mr McCallum said.
He added that the amount of satellite data now available to forecasters in the Met Office's operations room made a huge difference.
"I am not talking about pictures, but data about the characteristics of the atmosphere.
"To forecast the weather, you need to know what the atmosphere is doing right now.
"From that, you then try to predict what it is going to do in the future. But to do that, you have to have a very accurate representation of what is happening now."
This includes precise measurements of temperature, humidity and wind speeds at a wide range of heights within the atmosphere.
But he said that it was not only access to more detailed data that had improved as a result of the fallout from the 1987 failure.
"It is also the way we assimilate the data into our models, which have got better as well. We can extract the signal from the data much better than we used to do 20 years ago."
The models, which contain the latest scientific knowledge of how the atmosphere works and behaves, are run on the Met Office's supercomputer at its HQ in Exeter, Devon.
An independent inquiry, conducted after the Great Storm, urged that forecasters had access to the latest machines.
"We would stress the importance of ensuring that the Met Office always has at its disposal the most powerful computer possible," It concluded. "Under-computing would be a false economy."
"The other aspect that has been improved is the communication of risk," Mr McCallum revealed.
"One of the big lessons from 1987 was how to get your message across."
People are now able to access information on the weather 24 hours a day. The BBC Weather Centre provides updated bulletins on the web, interactive TV and mobile phones, as well as on TV and radio.
The Great Storm led to the government funding the establishment of a National Severe Weather Warning Service.
"For the people who matter, such as the police and other emergency services, planners etc, we have got a system to make sure that they get updated weather warnings quickly."
When it was first introduced, the service offered regional coverage. Today, because of the technological and forecasting improvements, it operates at county level.
Are all of these developments going to be put to the test more regularly in the future?
Climate models predict Britain experiencing extreme weather events more often in the coming decades as the planet warms.
"It's possible, but you can never give a definite answer to that sort of question," Mr McCallum replied.
"It is like if we have a big storm tomorrow, you'll get the same answer out of places like the Met Office, where they will say that it certainly fits with the [future] scenarios, but to blame any one event on climate change is facile.
"It is only when you look back over time and you look at global trends, can you make comments like that."