We are often described as a nation preoccupied by weather, and we've certainly had plenty to talk about over the last few weeks. The wind and rain continue their relentless assault, and the headlines focus, quite rightly, on the plight of those worst affected - what should be done to help? Meanwhile flood waters continue to rise and this is not unexpected. The weather forecasters have done their job well: no Michael Fish moments to distract our attention from what's important. Indeed, the last few winters have been marked by extremes - from snowbound Gatwick Airport to the St Jude Day Storm - and the weathermen have stayed ahead of the game. So is Mother Nature's number finally up when it comes to blowing us away with a storm from out of the blue?
The science and the technology that underpin weather forecasting have advanced in leaps and bounds since that fateful event in October 1987. In the mid-1980's the performance of the Met Office supercomputer was about one hundred thousand times slower than the machine in use today, and the amount of information available to forecasters from satellites and weather radar has increased dramatically. The deluge of weather data is like water coming from a firefighter's hose - but you can't drink from a fire hose.
One of the keys to successful forecasting lies not in number-crunching prowess, but in our ability to identify the large-scale processes affecting next week's weather. This video published by EUMETSAT (the European Meteorological Satellite) shows the entire year of weather across our planet for 2013. It makes for fascinating viewing, and one of the striking features is the near-repetitiveness of the whirling patterns that are responsible for much of the weather we experience in Britain.
No two swirling weather systems are ever quite the same, but many of them have broad features in common. Meteorologists quantify what orchestrates these recurring patterns, and weather forecasters use this information to make sure their next forecast gets the important details right.
The storms that have battered our shores and drenched towns and villages this winter are a phenomenon we understand fairly well. The computer models do a good job at predicting the track of depressions and fronts, and this is affirmed by the accuracy of the forecasts we've all benefitted from in recent weeks.
Way back in 1972 the American meteorologist Edward Lorenz epitomized the Achilles' heel of forecasting when he said that the flap of a butterfly's wing over Brazil could trigger a tornado in Texas: in other words, unless we have perfect knowledge of the state of the atmosphere at one moment in time, prediction of future weather will always be subject to the curse of chaos. This allegorical view led to what became a prevalent anti-theory of weather prediction, and the Great Storm of October 1987 was seen as a typical "forecast bust".
So have the weathermen vanquished chaos? Despite recent successes, the answer is no. We cannot guarantee that Mother Nature will not outwit us in the future, because chaos is a consequence of the laws of physics on which weather forecasting is based. But, as the EUMETSAT video shows, there are well-defined patterns and computers simulate the mechanisms that create such patterns with increasing reliability.
One of the next big challenges the forecasters face is to improve predictions of severe thunderstorms. These storms are very different from the low pressure systems that dominate our weather for much of the year, and the computer models have some way to go before reliable forecasts of thunderstorms become the norm.
So in the meantime we need to hang on to our umbrellas and, more importantly, learn how to use timely information about flood risk to alleviate the misery that so many have experienced this winter.Suggest a correction