Weather Forecasters: What a shower

The Met Offfice's new £27.5m computer is supposed to reduce forecasting errors. So why did it get last weekend's weather so wrong? Jonathan Brown reports on a history of inaccuracies

Being a weather forecaster has never been easy, as Robert FitzRoy, the commander of HMS Beagle, discovered when he was given the task of heading Britain's first weather service back in 1854. As imperial trade approached its zenith, the mandarins at the Board of Trade were becoming increasingly disgruntled about the losses of ships and their cargo. FitzRoy invented a cheap, serviceable barometer, to which he gave his name, and set up a weather bureau with characteristic Victorian zeal.

Being a weather forecaster has never been easy, as Robert FitzRoy, the commander of HMS Beagle, discovered when he was given the task of heading Britain's first weather service back in 1854. As imperial trade approached its zenith, the mandarins at the Board of Trade were becoming increasingly disgruntled about the losses of ships and their cargo. FitzRoy invented a cheap, serviceable barometer, to which he gave his name, and set up a weather bureau with characteristic Victorian zeal.

Four years later, catastrophe struck when the passenger vessel Royal Charter was wrecked off Anglesey. It was returning from Australia with almost 500 passengers on board, most of them miners carrying the spoils of their labours. Foundering in heavy seas just a few hundred yards from shore, 459 people lost their lives.

FitzRoy, with a staff of only three, needed to act fast. Within two years he had established 15 coastal stations, which provided visual gale warnings for ships at sea. The masters of the stations would erect canvas- covered wooden frames in the shape of cones or drums to indicate impending storms or changing wind patterns. At night, lanterns were hung to show the outline of the frames. Sailors could now check the weather without coming ashore. Thus the first weather forecast was born. But mistakes led to criticism from no less a place than the august salons of the Royal Society and for FitzRoy, the strain of being Britain's first weather forecaster was too much. Physically exhausted and severely depressed after a lifetime of public service, he committed suicide.

Today weather forecasters have a dazzling array of technology at their disposal, but one thing remains the same - the public's unstinting ruthlessness in its demands for accuracy. Typical was last Sunday, when 10m people living in London had been told to brace themselves for a rainy day. Instead they woke to clear blue skies and temperatures nudging 17C (70F), pleasant for the time of year.

For the Meteorological Office, the successor to FitzRoy's small and put-upon department, the result was newspaper headlines denouncing another "weather gaffe". "We have to hold up our hands up and say that we got it wrong," a spokesman said.

Sunday's rain dutifully arrived to douse commuters returning to work on Monday. It had been held at bay around Gatwick for 24 hours longer than predicted.

Memories are long. People still cling to the recollection of the devastating storms of 1987, when 16 people died and 15 million trees were felled. On its eve, Michael Fish remarked in his television forecast that "a lady has rung in to ask if there is going to be a hurricane tonight ... there is not."

Fish was correct. He was referring to a tropical cyclone in the West Atlantic that eventually drifted out to sea harmlessly. The storm itself - it never reached true hurricane proportions - developed too rapidly for forecasters to predict.

That was 20 years ago. But mistakes still happen, said Andrew Bond of Metcheck.com, one of the new breed of private forecasting companies that has developed in recent years to meet the growing demands of business and the public to know what is in store. He said: "Sometimes we get it wrong because weather is not an exact science. We get it right much more often, but people only seem to remember the times when we fail."

The Met Office's failure in getting the weather right on Sunday resulted in nothing more catastrophic than a pleasant surprise, but unfortunately for the Met, it chose this week to trumpet the success of its new £27m supercomputer NEC SX-8.

The accuracy of forecasts, it said, had increased by 11 per cent since its introduction, compared to an improvement of only three per cent across modelling centres in the world's leading forecasting nations. Earlier this month the Met Office announced that it had doubled the capacity of the computer.

Ewen McCallum, the Met's chief meteorologist, said, however, the devil remains in the detail. "Even though we have made enormous advances in terms of prediction, we still have a problem. What people want to know is 'will it rain in my back garden in London or will it be in Cambridge?' We still have a little more work to do to answer that."

Refining weather prediction techniques has been the work of millennia. Once the job was the preserve of holy men, who interpreted the weather as a sign of the community's standing in relationship to its gods, based on little more than superstition or folklore. More recently, such knowledge has traditionally been the province of sailors or farmers. Weather historian Brian Taylor said : "It is sometimes accurate, but more often than not wrong."

It is also specific to location. The old adage of "red sky at night, shepherd's delight, red sky in the morning, shepherd's warning", has bearing only in the Northern Hemisphere. A red sunset showed a mass of dry air moving towards you. A red dawn meant it had passed, Mr Taylor said.

The Chinese, the Egyptians and then the Greeks devised calendars central to understanding the weather patterns. Aristotle is responsible for the word meteorology, taken from his collected writings on the natural world, Meteorologica. In the fourth century Hippocrates, the father of modern medicine, declared: "Whoever wishes to pursue the science of medicine must first investigate the seasons of the year and what occurs in them."

But it was not until the mid-19th century that weather forecasting began with any accuracy. It was made possible by a series of technological advances. Rain gauges had been around for 2,000 years, but other instruments were revolutionary. By 1593 Galileo had bequeathed the world the thermometer, then nearly 50 years later his student Torricelli invented the barometer. Pascal meanwhile had made the connection between the weather and air pressure. The first hygrometer to measure humidity was been built by Ferdinand the Second of Tuscany.

By the time FitzRoy was putting together his forecast, both Daniel Farenheit and Anders Celsius had devised their temperature scales and weather balloons were already in use. There was also to be seen a growing mood of international co-operation, particularly at the Brussels Maritime Conference in 1853. The International Meteorological Organisation was born in 1873. With the advent of the telegraph and the first synoptic charts, drawn up for the North Atlantic in 1882, the modern weather forecast was becoming a reality.

Mr McCallum said that the next great advances came with the evolution of frontal theory among the "Norwegian school" of forecasters. "During World War One they realised that the weather was a battle between warm air and cold air, like two opposite armies," he said. The confrontation mirrored the fronts on the killing fields of Flanders.

The next big advances were again due to war, this time in the air. As aircraft took an increasingly central role in the Second World War, meteorologists were asked to explain why so many Spitfires were running out of fuel and crashing into the sea on their return to Britain. This led to the realisation that the forces of the atmosphere, particularly the jet stream, were at play. These, researchers surmised, must play a central part in the world's weather patterns.

Radar followed, but perhaps the weatherman's finest hour came in 1944. After a day's stomach-churning delay, Group Captain James Stagg, a Met Office employee, briefed General Eisenhower of the impending weather window to allow the D-Day landings to go ahead. The rest is history.

The post-war years saw the forecast develop around the twin poles of computers and satellites. As the US began to lead the world in both technologies, in Britain the weather permeated the popular culture irrevocably in 1954 when George Cowling presented the first television forecast.

Today the debate in the meteorological circles revolves around who should benefit from the government-funded global data collected, the private companies who sell their services to commercial enterprises or the public ones.

Global warming and climate change have also created a new interest in the weather, although as Mr Bond points out this remains largely the preserve of climatologists rather than the forecasters. "Britain is the worst place to monitor global warming simply because our weather is so changeable. For that you need a steady climate like the Sahara desert of Antarctica."

So with changes afoot in the world's weather system, predicting Britain's weather could be about to get even harder, which forecasts yet more stormy times ahead for the poor meteorologists.

Test the Met Office this weekend

The Met Office issued this long-range forecast yesterday.

The Bank Holiday weekend should see the warmest day of the year so far, as mainly dry weather and spells of sunshine look set to bring a brief taste of summer to much of the UK.

Met Office forecasters are predicting that the unsettled weather affecting many areas this week will gradually move north by the weekend, allowing higher temperatures to spread across most of the country. Although there will be scattered showers for some areas throughout the weekend, temperatures of 25C are likely across south-east England.

While the temperature inland may well encourage a spate of barbecues and picnics, the sea temperature around our shores is still pretty low and a few coastal areas will be quite chilly and misty at times.