These are worrying times for the arts and humanities. There seems to be a belief in government and in much of business that knowledge can be cut into discrete blocks and that the ones that matter most are those of science and technology, engineering and maths.

In tough economic times, this belief that they are the subjects that will create the graduates and research the country needs has deepened. We believe that they are vital to the economy. But we believe other research areas are just as important. Even in narrow economic terms it must be wrong to neglect the importance of the creative economy and the importance of a rich and vibrant museums, galleries and cultural sector for tourism.

The challenges facing the country and the world cannot be addressed without the arts and humanities. People's complexity comes from their language, identities, histories, faiths and cultures. Without understanding that complexity we cannot address these challenges. Subjects such as literature, philosophy and history teach students to look at the world from a different perspective, to challenge ideas and to communicate effectively, to bring the flexibility and imagination that employers need and welcome.

We fear for the future of the arts and humanities. Our cultural capital is in danger of decline if the government fails to protect them.

Prof Geoffrey Crossick, warden of Goldsmiths, University of London; Nicholas Penny, director of the National Gallery; Prof Rick Trainor, principal, King's College London; Sir Nicholas Kenyon, managing director, the Barbican; Prof Nigel Carrington, rector of University of the Arts, London; Sandy Nairne, director of the National Portrait Gallery; Prof Edward Acton, vice-chancellor, University of East Anglia; Andrew Burnett, deputy director, British Museum; Prof Malcolm Gillies, vice-chancellor, London Metropolitan University; Prof Colin Riordan, vice-chancellor, University of Essex; Prof Paul Curran, vice-chancellor, Bournemouth University; Colin Jones, president, the Royal Historical Society; Richard Mantle, general director, Opera North; Clare Matterson, Director of Medical Humanities , Wellcome Trust

Climate sceptics need context

I am pleased to accept Dr Peiser's apology for his use of a false quotation ("unless we announce disaster, no one will listen") that bolstered his accusation that both I and the IPCC deliberately exaggerated the evidence for human induced climate change and its likely consequences.

The new quote Dr Peiser has found is from an interview in 1995: "If we want a good environmental policy in the future we'll have to have a disaster. It's like safety on public transport. The only way humans will act is if there's been an accident." The first sentence requires the second two sentences to provide the context for the whole quotation. It is wrong to describe the false quotation as derivative from or supported by the quotation from 1995. Their contexts are very different as is what they say. The 1995 quotation describes how attitudes might change in response to disasters after they have actually occurred. It cannot be used to prove that I am alarmist or that I promote exaggeration .

Sir John Houghton

Meteorological Office, Exeter

Don't defend Koestler's deeds

In his review "The double life of Arthur Koestler, intellectual and sexual adventurer",Robert McCrum wrote: "Twelve years ago, rival biographer David Cesarani's account of the 'rape' of Jill Craigie (Mrs Michael Foot) caused headlines. Now, it is just a contentious part of Koestler's opinion that, as he wrote to his second wife, 'without an element of initial rape there is no delight'." A couple of points. 1. What does this sentence actually mean? 2. My great-aunt was raped, violently, in her own home, by Koestler. She was not "raped", and Koestler was not a "sexual adventurer", but a rapist.

John Foot

London WC1

It's an ill wind…

Touring New Zealand last year we discovered that the TV weather forecasts were incredibly inaccurate. We even joked that by contrast the British Met office weather forecasts seemed quite good. I am therefore amazed to read in "Rain or Shine" that the BBC is considering using the New Zealand-based service to provide the UK forecasts. If they can't even predict their own weather, what chance have they got of predicting ours?

Kay Bagon

Radlett, Herts

Beckett was no villain to Lucia

In her article '"Alone in despair: the tragic end of James Joyce's only daughter", on the troubled life of James and Nora Joyce's daughter, Lucia, Vanessa Thorpe quotes Helen MacTaggart, Lucia's hospital visitor from 1977 to 1982, as follows: "She was still in love with him. And I do think, as many do, that Beckett had used her." This accusation must be treated with extreme caution. Lucia Joyce's years of illness, an illness which would have occurred with or without Beckett's involvement with the Joyces, were 50 years of loss and extremely sad to behold, but it's simplistic to cast the young Samuel Beckett as the villain of the piece.

Bruce Ross-Smith

Oxford

They are among the most respected, scientific and accurate forecasters in the world. Yet to the British public they are a joke. Tim Adams visits the Met Office's HQ in Exeter to meet the people for whom the outlook is always gloomy

The walls of the hi-tech head-quarters of the Met Office in Exeter are decorated with wisdom about the weather. The words tend to act as a comic counterpoint to the work that goes on in the building. The meteorologists who wander the glassy corridors with one eye on the ever-changing Devon skies outside will tell you that the job of forecasting is becoming incrementally more exact with every new satellite and software update, but the walls invariably tell you something different. In the lobby, a quotation from Hansard of 1854 recalls parliament's reaction to the apparently wild suggestion of Vice-Admiral Robert FitzRoy that, with the appliance of scientific study, "we might know in this metropolis the condition of the weather 24 hours beforehand": uproarious laughter. As I am being led through some of the rigorous logic of the current mapping systems by the heirs to FitzRoy (the manic-depressive commander of Darwin's Beagle), I can't help glancing over their shoulders to a quotation from Paul Cézanne: "We live in a rainbow of chaos."

Perhaps because it remains a part of the Ministry of Defence (albeit a pretty much self-sufficient part), the Met Office borrows some of the language of the military ("warm fronts" and "cold fronts" were coined during the First World War, when such waves of offence were much on meteorologists' minds). This mindset is useful because, from the outside at least, the Met Office can seem almost permanently embattled. When meteorologists are not engaged with the elements, they are traditionally in conflict with the great British press (or, on occasion, with each other). One of the few things that any British weather forecaster can predict with absolute certainty is the round of headlines that will accompany any forecast that goes awry.

This winter they have seen something of a perfect storm of such coverage. Following on from its "odds on" suggestion that 2009 would be a barbecue summer, the Met Office suggested back in October that we were likely to be due for an unusually mild winter. January's record-breaking Big Freeze, which, as it actually unfolded, the Met Office got pretty much spot on, nevertheless became another opportunity to blame the forecasters for the vagaries of the weather. The Sun, a title that loves a tale of meteorological meltdown, offered its readers a weather dart board in the belief it might do a better job than the Met Office's new supercomputer. Elsewhere the mild-winter forecast and the Great British Blizzard were tenuously linked by sceptics with the failures of the Copenhagen summit and the leaked emails from the Climate Research Unit at East Anglia (a Met Office partner) to suggest the Met Office was so preoccupied with potential warming in 50 years' time that it had lost sight of next week's wind chill.

This cycle of stories gathered pace and deepened when it was revealed that the BBC, another institution experiencing some ongoing turbulence, had made the decision to put the contract for its weather broadcasting, which the Met Office has held for 90 years, out to tender, with whispers that the New Zealand-based service Metra is being considered to replace the current provider. All in all, it has been the most unsettled couple of months in the institution's history since Michael Fish's hurricane in 1987, and the Bill Giles bullying scandal which followed close behind (after allegations by weather forecasters of a climate of fear that existed beyond the Velcro clouds – allegations that were overturned on appeal).

If Rob Varley, the recently appointed director of forecasting, feels himself to be in the eye of this storm, then he doesn't show it. Varley was born to this job. His father was a forecaster with the Met Office for 34 years, and he has already put in nearly three decades himself. He has overall responsibility for getting the weather right for a range of clients that includes half the world's commercial aviation and the military in Afghanistan, as well as commuters on the M6 and schoolboys hoping their match won't be called off, and he is impressively sanguine about the task. His ultimate responsibility, he suggests, is winning the daily struggle to make perfect sense of a vast global atmosphere of swirling fluid. "If you went up to Dartmoor," he says, "and you dropped a small stick in the River Dart at Dartmeed and you drove to Newbridge, 10 miles down the road, and waited for the stick to arrive, and then had a go at predicting precisely where along the bridge the stick would pass through, bearing in mind all the factors and currents and obstacles that might affect it on the way, that is about what we are trying to do every hour…"

Meteorologists, as every weatherman's favourite joke goes, do it with models. To aid in the understanding of this wantonly fluid system – Edward Lorenz proved chaos theory using infinitesimal modifications to weather systems – the Met Office is now equipped not only with 1,800 staff but a £33m IBM computer that can make 1,000,000,000,000,000 calculations a second, through code that enables it to project weather systems forward in time an hour, a day, a week, a month, a year and a century – and all places in between.

At any one point – though the central problem with the weather is that it does not have a beginning or an end, it is all middle – several dozen potential weather scenarios will be running through this endlessly refined programme, each one based on the incoming minute-by-minute data from thousands of weather stations across the globe, supplemented with observations from aircraft and radar and weather balloons and anchored vessels in the oceans, all set against a framework of patterns from orbiting satellites in the heavens. And even then, as Varley concedes, no one alive can tell you absolutely for certain whether you'll need an umbrella if you go out at lunchtime.

One result of this is that criticism, like that which has come the Met Office's way recently, goes with the territory. "When you get a foot of snow in one place overnight there will inevitably be disruption. That's just life. But of course it's frustrating and we want to blame someone for it. We are an easy target." The long-range forecasts, of barbecue summers and mild winters, were unfortunate, but Varley stresses, "for most people, we try to make it clear there is no usefulness whatsoever in those forecasts. If I'm booking a holiday, I would love to know whether June will be warmer than August, but the seasonal forecast will not tell you."

So why do them?

"If you look over a number of years, roughly two times out of three the thing we say most likely turns out to be right," he says, "so it is better than chance – and in some industries that margin is crucial." He would much rather the forecasts never made it into the press, however, "but the information is out there, so what can you do?"

Varley doesn't believe, though, that negative publicity has contributed to the BBC's decision to consider their relationship. "The BBC reviews its contracts from time to time, and rightly so," he says. "The point is, we don't want to be delivering services out of some obligation, but because people know we are the best at it in the world."

What really annoys Varley is the suggestion that because of the seasonal forecasts there were "some aspersions about the quality of our climate science" – the "If they can't tell what it's going to be like this winter, how do they know what it will be like in 2050?" argument, which ignores the fact that long-term temperature trends and knowledge about the effects of CO₂ in the atmosphere are much more reliable than multifaceted local weather systems. The Hadley Centre, the Climate Change wing of the Met Office which was established by Margaret Thatcher 20 years ago, was recently deemed in an independent study to be the most reliable of 43,500 geoscience institutes around the world. "You don't get to that position by being imprecise," Varley says, precisely. "The fact is, the things we are trying to do are very difficult. Our reputation around the world is second to none, and yet in the UK we are all too often a target for criticism. That is so painful to us."

It is perhaps inevitable that we should require the bringers of our weather news to suffer for their science: we need them to share our pain. There is a good deal of evidence to suggest that all of our behaviour is intimately linked to small shifts in the weather. Recent research suggests the stock market is far more bullish on cold days than on warm ones (a fact that holds true from Taiwan to Sweden). Our moods come and go like scattered showers. I was moving house from one end of the M4 to the other in the week of January's heavy snow, watching the hour-by-hour forecast unfold with the zeal of an entrail-gazing druid, now clicking on snowflake icons, now desperately shovelling grit on to the road as a hopeless offering, now stopping to thrill at the view.

The BBC understands this love-hate obsession like nothing else. It persists in calming us to sleep on our damp island with the Forties Cromarty Rockall lullaby of the shipping forecast; offering the comfort of far-off storm warnings. On the night of the winter's heaviest snow, the corporation's reporters were predictably sent en masse to survey the nation's drifts in their North Face jackets and their Berghaus hats in order to say nothing at all – just so we could see them being snowed on for once in the land of dull drizzle and patchy cloud. The ritual – like the belittling of forecasters – is part of the nation's idea of itself.

The uncertainty of our weather engenders both a need to know what comes next and a deep-seated anxiety that we will never know, not exactly; it is in this gap that the Met Office lives. In recent years our obsessive weather watching has taken on a medieval, apocalyptic tone as each extreme-weather event seems a portent, and all of them at least partly our fault. The weather has become personal. As Martin Amis pointed out in an interview earlier this month: "Don't you sense the incredible potential for violence in the weather, already — the storms, the snow? You can see the nature of what the future will be, and it's all terror and boredom all over again. You will be massively inconvenienced and appalled by the power of the weather…"

Well, maybe, you might say. The institution of the Met Office is a strangely seductive idea precisely because it is a physical dramatisation both of our knowledge of the unpredictability of the future and our stubborn efforts to make it known. It attempts, in a somewhat heroic way (from its earliest incarnation as a lifesaver for seamen), to bureaucratise chaos, to systematise doubt, and it sometimes comes close to doing so. It not only shows us on a weekly basis the limits of science, but also the human need for the ingenuity to overcome these limits.

It is not an exact discipline but it is certainly an exacting one. Anyone who believes that dis-crepancies in climate models are down to some arcane global conspiracy should, for example, meet Stuart Goldstraw, observations manager at the Met Office, and have him explain with diehard enthusiasm the extraordinary lengths he explores in order to eliminate error from global weather statistics.

Goldstraw outlines at length the detail of this endeavour to me, standing on a small hillside behind the looming Met Office building, among measuring instruments old and new. On the one hand, massive dishes stand alert for satellite information from the upper atmosphere; on the other are thermometers and rain gauges in louvred boxes. Collecting weather data, which for more than a century was a resolutely human enterprise, has lately become almost wholly mechanised, and the shift is his biggest headache.

How do you create machines that can exactly replicate the human eye? Which can see fog in a valley or judge visibility? You can't – but the important thing is to know exactly where the limitations lie and to weight the information accordingly. Goldstraw points me toward a "present weather centre" which sends an infra-red beam out between two arms and examines what interrupts that beam. The machine is confident when it comes to snow or drizzle or rain, but sleet is beyond it. "In that case it will say 'precipitation of an unknown type'," he says, "and that's when the forecasters have to guess the uncertain boundaries between snow and rain and the no man's land of sleet."

Temperature, which can be influenced by all sorts of local factors – from urban creep to flaking paint on instrument boxes – is often even harder to ascertain. "Climate with a capital C is our big challenge," Goldstraw concedes. For comparative purposes the climatologist is looking for accuracy to at least 0.1 of a degree, and any error in the way data is collected totally undermines that. "It's only the very best," Goldstraw suggests, somewhat despairingly, "that can meet that challenge over a long period."

That means out of 12,000 global sites producing weather data they "can generally trust only 500, and the subset of those we absolutely rely on is much reduced again, and sites drop out even of that group; it is a continual, daily battle, and then there are gaps – particularly in Africa, where we don't have much reliable data at all".

If Stuart Goldstraw is the pragmatist of this process, then the blue-sky thinker is Brian Golding, head of research, who takes all that data and runs it through the model that describes chaos. Golding takes me down to the basement level of the Met Office building, where the supercomputer, in its grey rows of tombstone cabinets, hums to itself. Golding has been developing this model for 37 years. He is a meteorologist who can do maths, rather than vice versa, he says, so it is a practical model, the only integrated one of its kind, by which he means it can be used to study not only what will happen in half an hour, but also, by asking it very different questions, what will happen in 50 years' time.

Golding is a spirited, focused man, and his work has focused the understanding of what occurs in the swirls of air approaching our island. When he started out, the reference points on the UK's mapping grid were spaced at 100km; in the 1980s Golding led the team to a global first, a grid length of 15km; he has just implemented the third generation of this, which is a 1.5km model for the whole of the country.

The effect of this, he suggests, should eventually be a fineness and accuracy of information about what will almost certainly be the increasing wetness of Britain: "If you look at how we did in the Cumbrian floods [in 2009], the forecasts were brilliant, but they were still at the scale of Cumbria. With the 1.5km scale model we could come down to individual river catchments, which means you could accurately predict the height a river will rise to – which might have meant the bridge the poor policeman [PC Bill Barker] was standing on was closed before it collapsed."

The weather forecast might soon, he says, be able to move from the level of depressions and fronts to very local squalls and thunderstorms, to say whether it is going to rain over this hill or that one, but that is a whole new strata of unpredictability. "Mostly," Golding says, brightly, explaining his life's work, "you have a fairly predictable uncertainty up to some point, and then you get to an unpredictable event, and there is a dividing of possibility and then there are two groups of solutions and then one or both of those will divide into more. And typically, chaos happens with these multiple bifurcations." Generally this first splitting of possibility happens after two or three days. And after that, it will always be trickier to know.

It is somehow gratifying to know that despite all this computer power, the forecasts we hear still come down to a bloke sitting in front of a map and wondering whether on this occasion the model's billions of calculations have really got it exactly right. All of the data collection and exponential number crunching eventually ends up in front of the chief forecaster, who, on this day, is a man called Frank Saunders, who is staring at a screen in the middle of the operations room – from which half the world's weather is forecast, from which aircraft are grounded and gritters dispatched – and looking vaguely alarmed.

When I visit the Met Office in the first week of February, there is some snow around in the north of England, but the temperature is hovering around freezing point, so trying to work out where it is falling is an informed guessing game. Saunders is trying to work out exactly how Manchester is going to receive the "unknown precipitation event" while keeping an eye on storm surges around the coast and the changing information about overnight ice. "My intuition," he says, "suggests it is being slightly overdone in the observations and it will fall as rain." Get that intuition wrong, though, and what falls as a centimetre of rain will work out to 10cm of snow, the difference between drizzle and gridlock.

It is Saunders's job to convey the nuance of this message to all the nation's broadcasters in an hourly conference-call briefing. For all the science, if he gets the emphasis wrong, or a word out of place, sunshine can quickly become showers. The chief forecaster is not for nothing known as God. As Saunders prepares for his briefing he accidentally knocks a cup of water over his desk, and I wonder for a moment how the subliminal stress of that particular precipitation event will play out in the nation's skies.

Einstein once observed: "Before I die, I hope someone will clarify quantum physics for me. After I die, I hope God will explain turbulence to me." An explanation for turbulence is what we have come to expect, though, and we have little patience when our prophets fail us. In an effort to deflect criticism when what falls from the air does not match what we have been told to prepare for, the Met Office is planning to introduce a more "probabilistic" element to forecasting, offering you percentage chances rather than a definitive symbol of cloud or rain. One of the elements of their ongoing contract discussions with the BBC is how this can best be done.

Almost uniformly the people I speak to at the Met Office believe the probabilistic method will finally allow them to convey the subtleties of the systems they are describing, the probabilistic will allow them to be right nearly all the time, the probabilistic will prevent the forecast coming back to bite them. I hesitate to point out that in America, where this style has long been employed, it was recently shown that only 50% of people understood what a 30% chance of rain actually meant and were therefore more likely to take their umbrella on the off-chance than ever before. As the man said, there is never any such thing as bad weather, just the wrong clothes.

When they got it wrong

August 2004 The Cornish village of Boscastle was deluged with more than a month's rain in seven hours. The medium-term forecast had been a modest 30-50mm of rainfall, but 133mm fell between 11am and 6pm. As villagers paddled down the streets, the Met Office declared: "No one could have forecast 133mm in seven hours. This was a freak event."

May 2009 Bournemouth tourism officials were furious about a bank holiday forecast that predicted pouring rain and thunderstorms for the southeast. Town traders claimed they lost 25,000 would-be visitors – who would have basked in 22C sunshine.

June 2001 In May the Met Office was spreading joy – it was going to be a long hot summer. Three months later Scotland was in the grip of horrendous electrical storms: 9,000 homes without electricity, hailstones the size of golf balls, transport chaos and, horror of horrors, 500,000 viewers left without TV for more than three hours when lightning struck a major television transmitter.

When they got it right

December 2004 A bad winter for the bookies – but a good winter for the Met Office. Scotland enjoyed its first official white Christmas in three years and bookies cried into their porridge. The Met Office had predicted that "much of Scotland, Northern Ireland, northwest England and Wales are likely to see a white Christmas".

June 1997 As Pete Sampras limbered up for his fourth Wimbledon title, Britain was on course for the wettest June in six years. While fans dreaded the inevitable Cliff Richard (right) sing-alongs, the Met Office had a spring in its step: "We forecast showers for 30 days, and we were right for 22 of them."

January 1990 Forty-seven people died in the Burns Day Storm, when 104mph winds uprooted 3m trees, buildings collapsed, and the damage cost insurers £3.37bn. Unlike the infamous hurricane of '87, however, the Met Office had given plenty of warning – four days before the storm hit, severe gale warnings were issued on TV.