Toy Story: Mystic Met needs swanky new kit, swoon MPs

Media to blame for believing us reporting our forecasts

Comment The Mystic Met has bewitched MPs who have recommended the forecasters are given the swanky new supercomputer they want.

Despite the meteorologists getting 10 of their last 11 winter forecasts wrong, MPs say the Met needs to improve its communication and PR, rather than its scientific methodology. Amazingly, MPs even blame the media for being "irresponsible" with the tablets of wisdom they receive from Exeter.

The enquiry is significant for criticisms it chose not to address. The report contains no recommendation either to privatise the agency - a move with cross-party support that was actually suggested by the Treasury under the previous Labour administration in 2009 – or how to make it more commercially efficient.

The Met answers to the Business Department these days, and it's officially a 'Trading Fund' which is required to operate on a commercial basis. But even the Office's local authority – Devon Council – chose not to renew its contract with the Met, in preference of a private commercial operator.

The issue of executive compensation – the head receives a £200,000 salary (including bonuses) – was not examined either.

Questions of institutional bias - the Met has forecast 10 of the past 11 winters to be warmer than they turned out - and conflicts of interest were likewise ignored.

The Great British Weather Forecast

The Commons Select Committee set up an enquiry into the Meteorological Office, a publicly funded agency that answers to the Minister of Defence, last July, in response to criticism after the harsh winter of 2010-2011 and the "barbecue summer" of 2009, which turned out to be one of the wettest on record.

MPs promised the enquiry would examine how well it was performing its public service duty and look into the robustness of computer models it uses. Overall, the committee took an uncritical approach, leading to some unintentional humour:

"We were pleased to find that science is very much at the heart of the services provided by the Met Office," they report.

MPs discovered that Met Office executives were reluctant to jump off the gravy train, which may not be the most surprising discovery a committee has ever made:

"We heard that privatisation would be almost impossible due to the Met Office’s reliance on international partners to provide it with data, which in some cases would not be allowed if it were a commercial organisation," MPs report. They were also told that “the Met Office only owns 4 per cent of its data”. During the course of the committee's enquiry, the Met found a powerful ally in LibDem MP Edward Davey, now the Energy and Climate Change and self-styled "eco-minister". Davey told the committee that the agency was already “pretty efficient” and warned of terrible consequences if it was made more commercially accountable:

"[Davey] was, however, clear that the services that the Met Office provided to the public sector were 'absolutely critical' and that while it was possible to expand the Met Office’s commercial activities, that should not put services for the public sector at risk," we learn. Those risks are not elucidated.

The Met Office employs 1,700 staff and costs were £186.7m in 2010/11 and revenues £196.1m. However, much of the "revenue" is non-commercial, and five-sixths of the income comes from the taxpayer: £90m for weather forecasts and £17m for climate-related science. Only £32.2m comes from commercial operations.

Toy Story

The Met's long-running campaign to invest in computer resources was sympathetically received. According to the Met's chief scientist, “the science is ready and waiting”, like Sleeping Beauty awaiting to be awoken - or ravished. Playing the part of the handsome prince in this scenario is a £50m supercomputer.

It may seem like only yesterday that the Met took delivery of a new supercomputer, a £33m monster from IBM. But in fact, that was in 2008. The £50m upgrade is needed.

MPs heard how the Met had “slipped down the league table in terms of its computing resource” and that it would be “impossible to deliver world class weather and climate science without access to adequate computing capacity”. Met boss John Hirst claimed that £0.5bn of economic benefits would result from £50m of spending in the next five years on supercomputing resources.

Yet MPs also heard how the Met Office had let the UK's observational network fall into disrepair: there are only around 3,000 stations reporting, half of the number reporting in 1975.

MPs then recommended the Office needs its new supercomputer and told it to go back and buff up its business case.

"We note that the climate model did not accurately predict the extent of the flattening of the temperature curve during the last 10 years," MPs noted, but the Met Office replied that: "what the early models predicted has largely come to pass", and MPs agreed.

If it wasn't for these pesky journalists…

One intriguing passage is the subject of the Met Office's seasonal forecasts. MPs say the Met should express the uncertainty of seasonal forecasts more often. Without explaining the probability behind a forecast, the press may jump to the wrong conclusions.

The problem with this analysis is that the Met does garnish predictions with probabilities, and these are faithfully reported. "After two disappointingly wet summers the signs are much more promising this year," Ewan McCallum, the Met's chief press officer said in 2009. "This year we can expect times when temperatures will be above 30 degrees Celsius, something we hardly saw at all last year. The chances of getting the barbecue out are much higher than last year," said in April 2009. McCallum added that "the temperature differential between southern Europe and the UK was likely to close. All this was accompanied by a specific probability reference. The Telegraph clearly reported that:

"… he warned that it was only a forecast and that while there was a two in three chance it would be right, therefore, there was a one in three chance it would be wrong."

MPs did recommend that the Met Office make seasonal forecasts available to the public – something it stopped doing after criticism.

Speaking of probabilities, the odds of a forecaster getting its prediction wrong 10 times out of 11 – predicting warmer winters than materialised – are left as an exercise for the reader. Seasonal forecasts now contain much hedging.

Something or other may happen: The Met's advice to government

Serving the FCO

Bureaucrats are likely to stick together, and MPs heard the most effusive praise for the Met from the government itself.

"The Met Office provides support on accessing, understanding and interpreting the latest climate science which allows the FCO to talk with an authoritative voice on climate change from a sound scientific evidence base," said the government in written evidence.

"The Met Office has also been able to provide unbiased scientific answers where there is disagreement on the basic issues of the science which have proved sticking points in discussion," we learn [our emphasis].

And there's the opportunity for budding 'communicators' to accumulate Air Miles:

"Met Office climate scientists have accompanied FCO representatives on country visits, speaking at events organised by the FCO to raise awareness and engagement on climate change issues," according to written evidence.

Singled out for praise by the government is the Met's "4 Degree Map". The tool designed to show "How the world could look under high-end climate change" is lauded as "an extremely useful influencing tool worldwide."

At issue is whether policy is founded upon evidence and rational arguments, or whether pre-determined policy requires the manufacture of sympathetic "evidence" - as we've seen in health and social policy areas under successive governments. The committee would do well to allay concerns that the Met has such political utility, it's beyond serious criticism. ®

Related link

The committee's report

Sponsored: Minds Mastering Machines - Call for papers now open

Biting the hand that feeds IT © 1998–2018