Michael McCarthy, The Independent 19 Jan 10;
Why are we asking this now?
Rather than renewing its current weather forecasting contract with the Met Office automatically, when it expires in April, the BBC is putting it out to tender – for the first time since 1922, when national broadcaster and national forecaster first became partners. No one on either side says how much the contract is worth.
Why would the BBC drop a national institution?
Allegedly because it is looking for a cheaper alternative from among the many independent weather forecasting companies which have sprung up in the last 20 years. But the Beeb's move also coincides with a period in which the Met Office's forecasting accuracy has come under unprecedented fire. There is speculation that the two may be linked.
Have recent forecasts been inaccurate?
Some of them have. Most recently was the snow in London and the south-east last Wednesday morning, which was heavier than had been predicted, and caused widespread disruption. But beyond that, the Met Office failed to predict this year's Big Freeze as a whole.
The winter seasonal forecast for 2009-10, issued on September 29 last year, said that "winter temperatures are likely to be near or above average over much of Europe including the UK. Winter 2009/10 is likely to be milder than last year for the UK, but there is still a one in seven chance of a cold winter". As it turned out, we are in the middle of the coldest winter for 30 years. And then there was the famous case of the "barbecue summer".
Can we recap on that?
Last April, the Met Office issued its seasonal forecast for summer 2009, and said it was "odds-on for a barbecue summer", in a tremendously resonant phrase which made big headlines everywhere, not least because it was such a terrific piece of good news after the washout summers of 2007 and 2008. Chief forecaster, Ewen McCallum, said at the press conference: "We do not see the London bus syndrome of three wet summers coming in a row. The likelihood of that happening is extremely small." That was a hostage to fortune if ever there was one: July turned out to be one of the wettest summer months on record.
By the end of it, the resentment from a public whose hopes had been firmly raised for hot dry evenings on the patio was so intense, that the Met Office felt obliged to issue a public apology.
Why were they wrong?
The funny thing is, on one level, they didn't; the key phrase was "odds-on", and the odds Mr McCallum was talking about were precise: they were 65:35. That meant that the Met Office supercomputer had run 50 different simulations of the weather over the coming summer, in what is known as an "ensemble" of forecasts, and 65 per cent of these had indicated it would be warmer and drier than average, while 35 per cent had indicated the opposite.
So the Met Office did say there was a 35 per cent chance or rain, which is how it turned out – but that was entirely lost on the public in the forceful catchiness of the "barbecue summer" phrase, which, of course, was chosen to make headlines. The Met Office got its headlines, but it paid a very high price, in image terms, for getting them wrong: barbecue summer will take a lot of forgetting.
What does the Met Office say in its defence?
That dealing with a chaotic system such as the earth's atmosphere means one can never make forecasts with complete accuracy, especially a season in advance. Although it is accepted that "barbecue summer" was a big blunder, the Met men assert that most of their forecasts are right most of the time, and although last Wednesday's snow might not have been fully predicted, most of the episodes of the big freeze have been accurately called.
Who might take over as BBC forecaster of choice?
The Beeb is said to be talking to Metra, the commercial arm of New Zealand's state-owned national weather forecaster. Weather Commerce, Metra's UK subsidiary, already supplies forecasts to Tesco, Sainsbury's, Marks & Spencer and Waitrose, which help with sales predictions and weather-related distribution issues. Metra is clearly panting to win the contract (it was obvious from news reports at the weekend that Metra itself was the source of the story). It could do so.
In the past, only national weather forecasting services, such as those of the UK, the US or Japan, had the resources to perform full-scale weather prediction, which needs immensely-complex mathematical models of the global atmosphere and supercomputers costing billions of pounds on which to run them (and which have to be replaced with newer and even faster models every couple of years).
But the advent of the internet has meant that these forecasts can be available and downloaded, sometimes free, sometimes paid for, by many smaller firms, who can then tailor individual forecasts to specific clients – such as north sea oil companies. One potential advantage such smaller companies may have is that they take into account a wider range of forecasts than a single national weather service relying principally on its own weather model. But you have to be fairly big to fill the BBC bill: only companies with a turnover of more than £10m have been invited to apply.
How do you win the contract?
You have to provide the BBC with meteorological data for five years, at a competitive price, obviously. But there's something more tricky, too. You have to provide them with weathermen (or weatherwomen, or weatherpersons) – a cohort of 20 meteorologically-trained TV presenters, who will become household names and household faces. You have to provide the next Bill Giles (he of the tough reputation). The next John Kettley (he of his own pop song). The next Michael Fish (he of his own hurricane).
Is that so difficult?
It's something that the Met Office, with its big staff, can do easily, but it may be a much tougher call for a small company. Live broadcasting is a challenge. So the Met Office is still in with a good chance of recapturing the BBC contract; don't be surprised if it carries on. It said yesterday: "We consider we are in the best position to provide the BBC with accurate and detailed weather forecasts for the UK, and we hope this successful relationship continues."
Wind of change
1854 Met Office founded to provide information on the weather and marine currents to the marine community by Robert Fitzroy, captain of Darwin’s ‘HMS Beagle’ and later Governor of New Zealand
1909 Transatlantic shipping starts to use wireless telegraphy to transmit weathermessages ashore
1914-1918 Military personnel become dependent on Met Office forecasts for war planning
1939 Second World War sees introduction of radio sondes – ‘a collection of balloon-borne sensors transmitting data on pressure, temperature and humidity to receiving sites on land’
1940 The Met Office moves from London to wartime accommodation at Dunstable
1953 Major floods in south-east England, caused by storms in the North Sea, lead to the construction of the Thames Barrier
1954 The first live BBC Television forecast, lasting five minutes, was made by Met Office forecaster George Cowling
1959 London Weather Centre opens in Bracknell
1964 The first operational satellite images become available
1972 An IBM 360/195 computer is installed in the Richardson Wing of the Bracknell headquarters
1974 The Met Office takes part in the first global observation experiment
2003 One of the world's fastest supercomputers – the NEC SX-6 – isinstalled at the Met Office
2004 The Met Office's new headquarters in Exeter is fully operational
Would the BBC be justified in abandoning the Met Office?
Yes...
*If it could secure a substantial saving for the same quality of service
*Some recent Met Office forecasts have been very wide of the mark, such as last year's barbecue summer and the recent big freeze
*The internet means a national weather service is no longer needed for serious forecasting
No...
*The Met Office's general forecasting record is nearly 90 per cent accurate. Could others do better?
*It has more experience than any other weather service in the world (it dates back to 1854)
*It is a respected national institution and the appropriate weather source for the national broadcaster
