Ever since Vice-Admiral Robert FitzRoy set up the world’s first national meteorological office in England in 1854 forecasters have been blamed for getting the weather wrong. In fact, FitzRoy was so widely ridiculed whenever he made a wrong forecast that he committed suicide.

So you have to feel sorry for Michael Fish. Every anniversary of the Great Storm of October 1987, out trot the repeats of his legendary forecast about “don’t worry, there is no hurricane on the way”, just hours before the big storm struck. Actually he meant a hurricane in Florida, and added it would be windy in England that night. But everyone went to bed thinking there was nothing to worry about. So there was national horror when 100mph winds felled 15 million trees, power and transport links collapsed and 19 people died, the most devastating storm in England for more than 250 years.

The Met Office was roasted. It had forecast the storm, but thought it would largely miss England. Crucial weather observations were missing after weather ships had been withdrawn as part of financial cutbacks. Though its big weather forecasting computer was not up to tracking highly explosive storms, the human forecasters were too enthralled by it to question its results.

Things changed afterwards. Satellites beamed down more information, a bigger computer was bought and run on better models, and forecasters dissected the 1987 storm to learn every lesson. Only three years later the Met Office correctly forecast the huge Burns Day storm that battered the country with even greater devastation.

Forecasts are now better, and we are told another 1987-type fiasco is highly unlikely. But blunders still happen. In January 2003 the Midlands was brought to its knees when the Met Office predicted snow and councils gritted their roads.

Then unexpected rains washed away the grit and a cold snap froze the roads into sheets of ice. In 2004 seaside resorts were left seething on May Bank Holiday when heavy rain was predicted, only to see glorious sunshine but no visitors.

The list can go on and, of course, the forecasts that go wrong are more memorable than the majority that are right. But the Met Office is now entering some murky waters with its seasonal forecasts made months in advance. Last winter was billed as milder than average but colder than in recent years, which led some media to conclude that we were heading for an Arctic freeze with icebergs floating down the North Sea. And what happened to predictions back in May that this summer’s temperatures would be above average, with “70 per cent certainty”? As it turned out, temperatures were just below average, but that was the least of our worries when the rains crashed down.

The problem is that forecasters are not speaking the same language as the rest of us. The Met Office’s seasonal forecasts are still experimental projects, as newfangled as daily forecasts were a century ago. They should be interpreted carefully rather than taken at face value.

Even the daily television weather forecast has become an information battleground, as weather, places, times and dates come thick and fast in a blizzard of data. A breathless forecaster gabbles in a frenzy, trying to keep up with graphics that whiz around Britain like some sort of computer game, one moment in Plymouth, next flying over Glasgow and with a stomach-churning U-turn race down to London at furious speed. By the end of it the average viewer is probably feeling queasy and utterly bamboozled. What we need is less gee-whiz graphics, fewer words and clearer language.

Does any of this really matter, though? For a daily forecast, not knowing to take an umbrella out before it rains is a nuisance, but not forecasting gale-force winds could be life-threatening. Seasonal forecasts are probably too vague to make much difference to most people, but they do highlight a much more serious problem about understanding the climate, with considerable consequences.

Climate scientists are growing increasingly frustrated that their message about man-made global warming is still not getting through to everyone. Look, they say, the world is growing warmer at breakneck speed, and the only explanation is greenhouse gas pollution, a concept recognised for 180 years. Forget about the swings and roundabouts of the past climate, set off by volcanic eruptions, ocean currents, the waxing and waning of the Sun or wobbles in the Earth’s axis and orbit.

None of those natural fluctuations can explain what is happening now, but there is no black-and-white proof, in the same way that lightning can be proved to be electrical or the Earth is round.

Arguments about man-made climate change rely more on probabilities, rather like trying to judge someone innocent or guilty based on a range of probabilities from the evidence presented in court. There is no 100 per cent evidence that the greenhouse effect has taken grip of the world, but the probabilities of guilt are growing stronger each year, until eventually the sceptics will have dwindled into obsolescence, although by then it will probably be too late to do anything about it.

The same uncertainties are true for predicting the impacts of our future climate. When scientists talk of Britain growing warmer that doesn’t mean we will never see snow or even a full-blown Siberian freeze again, it just means the chances are fewer. Summers could still be wet, as this summer showed, but the odds are they will grow drier and droughts worse.

There is no reason why these uncertainties and probabilities cannot be better explained, given that gambling odds at a horserace are widely appreciated. The Met Office needs to choose the right language to talk to the public, otherwise it leads to confusion and misunderstanding.

After all, look what happened to Michael Fish.

Times Recommends

• We must lead the world to stability
• The recession's a blessing in disguise
• The guilty men who sent mum to Iceland