Noon on Thursday June 9 in the Met Office operations room in Exeter. From desk to ceiling, brightly coloured computer screens show past weather and future predictions. Showers speckle a rainfall radar map of the British Isles. A temperature chart shows tongues of warm orange air sticking into a pool of cool blue over the North Atlantic.
In the middle sits Martin Young, the chief forecaster, facing a dilemma about the weather three days ahead. He has known since the beginning of the week that an Atlantic depression was likely to reach Britain at the weekend, but now, with millions of people making plans, he must issue more specific guidance about when and where the rain is going to fall on Sunday.
"We have techniques for merging models on screen," he says. "I've blended the early part of one model with the later part of another one. But at the moment we're leaning more towards the Met Office model." That would mean rain reaching London at about 2pm on Sunday and northern England by nightfall – we'll see later what actually happened.
. . .
For all the prodigious processing power and observational resources of the worldwide meteorological system, forecasting decisions still come down to human skill and experience. The Met Office would never rely solely on a computer-generated forecast, says Young, who has worked there for 32 years – moving from the old headquarters in Bracknell, Berkshire, to the new Exeter centre.
For Young's boss, chief meteorologist Ewen McCallum, today's uncertainty about what will happen in three days' time illustrates the improvement in forecasting over the past generation. When he joined the Met Office 37 years ago, forecasters frequently faced similar or worse uncertainty about what would happen the next day.
"A four-day forecast today is about as accurate as a one-day forecast was when I started," says McCallum, in an accent as Scottish as his name. "Then, we had no operational access to weather satellites, no radar and very slow computers."
Public service Changing face of the Met Office The Meteorological Office was set up in 1854 as a small department within the Board of Trade by Robert FitzRoy, who had commanded HMS Beagle during Charles Darwin's famous voyage round the world in the 1830s (pictured above). Although its main role was to provide a service to mariners, the new Met Office soon started serving the wider public as well. The first newspaper forecast appeared in The Times in 1860, followed by the first weather map in 1875. The focus switched from oceans to air with the coming of aviation early in the 20th century. In 1920, the Met Office became part of the Air Ministry and began to set up observation and forecasting posts at RAF airfields, where many of them remain today. BBC radio included weather forecasts from the start of public service broadcasting in 1922. When the Air Ministry was folded into the Ministry of Defence in 1964, the Met Office followed it. It remains part of the MoD, though operating semi-independently – becoming an executive agency in 1990 and a government trading fund in 1996. In 2003, the Met Office moved from a series of 1960s office blocks in Bracknell, Berks, to new purpose-built headquarters on the edge of Exeter, designed by international architects Broadway Malyon. The relocation, at a total cost of £105m, was carried out on time and to budget. The MoD is still the biggest single funder of the Met Office, receiving in return a comprehensive range of support for the armed forces, including Mobile Met Units that provide critical weather advice wherever military action is taking place. The Met Office financial results for 2010-11, released this week, show income up by 2 per cent to £195.6m. Of this, £32.2m came from commercial business (up 10 per cent) and the remaining £163.4m from a range of government departments.
>
"I have been at the Met Office for 28 years; my father joined in 1950 and was a forecaster for 34 years," says Rob Varley, operations director. "So the Met Office has been an intimate part of my life since before I was born."
Brian Golding, head of forecasting research and development, will soon celebrate 40 years at the Met Office. The raw materials in a successful forecast have improved beyond recognition over that period, he says, and will continue to do so in the future.
The first key ingredient is the fundamental physics of the atmosphere and how it interacts with oceans and land masses to produce weather. This is encapsulated in increasingly sophisticated models, as computing power grows. The £33m Met Office supercomputer – a twinned IBM Power 6 machine installed in 2009 and about to be upgraded – can carry out trillions of calculations a second. It sits in two huge halls, shrouded by what look like plastic shower curtains. These are intended not to preserve the modesty of the energy-guzzling machine but to reduce the need to cool in the immediate vicinity.
Forecasts become more accurate as the model's spatial resolution increases. The latest global model cuts the atmosphere into 70 horizontal levels, with a spacing of 25km between points at each level, though the resolution of this grid can be improved to as little as 1.5km for short-term forecasting over the British Isles.
The other ingredient in forecasting is having accurate data to feed into the computer models. "The World Meteorological Organisation facilitates a remarkable global system for sharing data," Golding says. "Nowadays satellites provide the bedrock of our global observations, though radiosondes [weather balloons] are still used extensively and direct data from aircraft is becoming more important."
For UK forecasting, observations from the network of local weather stations remain essential, while 18 weather radars across the British Isles give forecasters a picture of where rain or snow is falling. One of the most important developments in recent years is ensemble forecasting. Because the observations are far from being a perfect description of the atmosphere – and because the atmosphere is in scientific terms a "chaotic" system in which very small errors in the initial state can lead to big errors in its development a few days ahead – the ensemble technique runs the computer model many times with slightly different starting conditions.
Ensemble forecasting gives meteorologists a much better idea of the likelihood that a particular event will occur than a single computer run. "At the moment we can run the forecast 24 times for our ensemble," says Golding. If the ensemble runs vary a lot, then the forecaster knows there is a lot of uncertainty about what will happen. If they cluster together, he or she will have more confidence in predicting a particular event.
One occasion on which this technique gave the Met Office enough confidence to issue a local severe weather warning three days ahead was the 2009 Cumbrian flooding. All the ensemble runs showed a stationary weather front over Cumbria embedded in a very moist south-westerly airstream – a recipe for prolonged downpours over the hills. In the event, Seathwaite recorded 31.6cm of rain on November 19, an all-time UK record.
The Met Office currently offers the public forecasts up to a month ahead, couched in general terms, because day-by-day predictions become unreliable more than a week or so ahead. They are produced by the ensemble method, running the model many times and averaging out the results.
Here is a typical example: "The generally unsettled conditions look to continue as June comes to a close, with further showers or rain likely to affect many parts of the country, but especially Scotland and Northern Ireland. However, there should also be some drier, brighter and slightly warmer periods, these most likely in the south of the UK, but particularly across south-eastern parts of England…"
. . .
The trouble was that the nuanced predictions of the seasonal forecasts were spun first by Met Office staff and then again by the media. For instance, the forecast at the end of April 2009 said there was a 50 per cent chance that temperatures from June to August would be above average, a 30 per cent chance that they would be around average and a 20 per cent chance that they would be below average.
Ewen McCallum, talking to journalists when the forecast was issued, could not resist commenting: "The chances of getting the barbecue out are much higher than last year." And that quickly translated into "barbecue summer" headlines, although McCallum emphasised at the time that the seasonal prediction was based on probabilities and there was a substantial risk of its being wrong.
However, the Met Office still sells seasonal forecasts to large commercial customers such as energy companies, which believe they can make worthwhile business decisions on the basis of probabilistic forecasts. Small businesses and individuals such as farmers and gardeners, who might also be sophisticated enough to make planting decisions based on the forecasts, lose out.
Behind the scenes the Met Office has an intensive programme to make its seasonal and other long-range forecasts accurate enough for presentation to the public again in the future. Adrian Scaife, who is in charge of the research, points to several improvements in the pipeline.
A second improvement comes from working out the consequences for the northern hemisphere of the famous oscillation in the tropical Pacific Ocean between warm El Niño and cold La Niña. Although the impact in the southern hemisphere is dramatic, as Australia's La Niña floods showed earlier this year, the indirect effects on Europe are only now emerging. The latest research shows that El Niño tends to produce high pressure over the polar regions of the north Atlantic and low pressure further south – leading to cold winters in northern Europe.
Third, scientists at the Met Office and elsewhere are beginning to understand the effect of the 11-year solar cycle on climate. When sunspots and other solar activity are at a minimum, the effect is similar to that of El Niño: more easterly winds and cold winter weather for Britain.
"We now believe that [the solar cycle] accounts for 50 per cent of the variability from year to year," says Scaife. With solar physicists predicting a long-term reduction in the intensity of the solar cycle – and possibly its complete disappearance for a few decades, as happened during the so-called Maunder Minimum from 1645 to 1715 – this could be an ominous signal for icy winters ahead, despite global warming.
Moving beyond the next season, Met Office scientists are also working on annual and, particularly, decadal forecasts – the territory between weather and climate prediction. Business and government planners are used to working on timescales 10 years ahead, but horizons of 50 or more years, often discussed by climate change researchers, may be too long term.
To communicate effectively with the outside world – the public, commercial customers and government departments – good internal communications are essential. Besides the usual computer-based methods, a new means of personal communication has emerged since the move to Exeter: the Thursday "Street brief". After lunch, Met Office staff emerge from their open-plan offices to congregate on the internal balconies of the building's elongated atrium, an area known as the Street. At 1.45pm, forecaster Tom Morgan embarks on a detailed disquisition about previous weather and the Met Office's success in predicting it – which has been good over the past week.
Then, projecting charts onto the white wall across the Street, he moves to the future, expressing doubts not just about Sunday's weather but about the very next day. "By midday tomorrow there's a trough over southern Britain but we have a large degree of uncertainty, considering the short range," Morgan says. He predicts heavy rain for Exeter early on Friday afternoon – correctly as it turns out.
The rest of the briefing includes notable weather events around the world, such as the first hurricane of the season in the east Pacific, and a review of Britain's remarkable spring weather: the driest season ever recorded in parts of south-east England and East Anglia, and the wettest May on record in the Outer Hebrides.
. . .
Though the Met Office is an agency of the Ministry of Defence, the atmosphere in its headquarters seems open and relaxed – with little evidence that the visitor is in a defence establishment. Privatisation, an option raised briefly last year, is off the agenda, at least for the time being, and the Met Office's arm's length "trading fund" status seems to be shielding it from the worst of the cuts threatening the MoD and other government departments.
Nor do the staff seem too concerned about the pasting the Met Office has received recently in parts of the media. "Bashing the weatherman is part of the British psyche," says McCallum.
However, some of the recent antagonism is linked to the Met Office's deserved reputation as a champion of research into climate change – and its scientists' unrepentant calls for urgent action against man-made global warming. Some climate-change sceptics seem to attack Met Office weather forecasts as a way of undermining its climate predictions. Still senior staff insist that, judging from their personal experiences of talking to non-meteorologists and from its surveys of public trust, the reputation of the Met Office is higher than ever. "Our trust scores are about 82 per cent, which is phenomenal for any organisation," says Varley. "I find it heart-warming that, when it comes to the crunch, people trust the Met Office."
So what happened on Sunday June 12, three days after the FT's visit to Exeter? Anyone who was attempting outdoor activities in England and Wales will remember that the rain arrived sooner and moved north more quickly than forecast. In London it started to pour before breakfast, preventing any play in the men's tennis final at Queen's Club, which had been brought forward by two and a half hours in a fruitless effort to beat the weather.
Clearly it was not the Met Office's best moment of 2011, but whether you regard it as a noteworthy forecasting failure depends on your perspective. "We picked up the general signal that it was likely to be a miserable wet day five or six days previously – and that was right," McCallum tells me afterwards. While such a forecast would have been a triumph a generation ago, people expect much more today.
Whether more accurate forecasting will outpace expectations in the years ahead remains to be seen. The outlook is bright, but no one in the Met Office expects to win universal love and approval.
Clive Cookson is the FT's science editor. To comment on this article please e-mail magazineletters@ft.com
No hay comentarios:
Publicar un comentario