A relationship that lasted 93 years has hit the rocks. The UK’s Met Office is not on the shortlist for renewal of the contract it has held since 1922 to provide weather forecasts to the BBC. This is unlikely to affect the visible face of the weather forecast dramatically, but is the underlying change in the data provider a good idea?
The Met Office has its origins in a service for mariners, set up in 1854 under the leadership of Robert FitzRoy, who had previously captained HMS Beagle on Darwin’s famous voyage. The organisation developed as part of the Ministry of Defence, with an important role to play in both world wars. Weather prediction helped the 1944 D-Day landings succeed, but equally armies have often waited for bad weather to attack, such as during the Ardennes counter-offensive later that year.
For most of the 20th century, the Met Office provided a public service in supplying weather forecasts to the BBC and conducting atmospheric science research. It has acquired an obligation to generate commercial revenue, however, and in 2011 it became part of the Department for Business, Innovation and Skills.
The Met gets its income through providing a range of specialised forecasts, for example on behalf of the aviation and insurance industries, but the BBC contract remains a key part of its work. At a time when government is reducing spending, it is natural that the Met Office should seek to maximise income and the BBC should seek to cut costs, leading to the current schism.
Should anyone care?
The clichéd conversational opener about the state of the weather is invariably followed by the comment that it wasn’t predicted correctly. Yet anecdotal evidence is misleading. The “skill” of a weather forecast is determined by measuring the correlation between a forecast weather map and what actually happened. This can be judged against the skill of a forecast based on statistical summaries of past records, known as climatology.
Since the 1980s, computer-based forecasting has steadily improved at a rate of about one forecast day per decade. This means that a five-day forecast made now is about as skillful as a two-day forecast was in 1985. The World Meteorological Organisation compares all national services and the UK’s forecast quality ranks among the very best, despite the notorious unpredictability of the country’s weather.
This notable improvement is a result of two related factors. First, the data from a vast network of surface and satellite observations is shared and distributed to national services. This is vital: knowing what the whole atmosphere is doing right now is the first step to predicting how it might change. Second, the weather models that project that information forward in time have been improved. This requires supercomputers, which are becoming increasingly accessible. In principle, any meteorological service that runs a global model of the atmosphere could offer a forecast for any region of the world.
Global models are needed to forecast more than a few days into the future, but models that focus on a smaller area are also required to account for local factors, such as hills and valleys. For UK forecasts, a large-scale model of the North Atlantic and a still smaller-scale model of the UK and Ireland alone can be embedded into the global picture, each working down to finer scales. Other meteorological services can run similar local-area models focused on the UK, but it is not a trivial task and requires computing resources roughly equivalent to those for the global model.
There is also a human aspect. Computer results are rarely presented raw, but are first interpreted by a human forecaster with local knowledge and experience. Different audiences also understand weather information communicated in different ways and we’re still not sure if forecasts are best presented as a single, most-likely case or as a range of probabilities. It is a brave move to change format, and impossible to please everyone.
The BBC has an obligation to find the best value. But value should not be confused with price, and cheaper bidders must provide a forecast of comparable quality. Viewers may be reassured by retaining the same presenters and graphical presentation; subtle changes in forecast quality and communication will be more difficult to assess.
Then there are strategic issues for the country. The Met Office will inevitably lose income, a net loss to the British economy, as well as public recognition. Since the 19th century, the UK has built world-leading atmospheric science expertise, of vital importance in applications ranging from defence to pollution and climate monitoring. Any erosion of capability in these areas may later be a cause for regret.