Ill Met By Moonlight

FORECASTERS AIM FOR CONSISTENCY. BUT THE MET OFFICE’S CONSISTENCY IS IN GETTING ITS PREDICTIONS WRONG

Forecasting is an inexact science, no matter what the field. It’s said, for instance, that economists have successfully predicted 14 of the last five recessions. Yet Britain’s Met Office has achieved an extraordinary consistency in its long-range forecasts – by getting them wrong every time.

It has been wrong in its predictions for the past three summers and the past three winters. You may recall that in April 2009, the Met Office forecast a “barbecue summer” for Britain. If there was a summer day suitable for unearthing the barbie, I must have missed it. Tthen, in October, they prophesied a mild winter. Perhaps nobody should have been too surprised that we were soon plunged into the coldest winter for 20 years.

This is despite the Met Office costing the public purse getting on for £200m a year, installing a new £30m supercomputer the size of two football pitches in 2009 and giving its directors pay rises of a third, with the Chief Executive now getting more than the Prime Minister. Its record has probably cost British industry and tourism millions and it makes one wonder how accurate are its predications on climate change, a field in which it is incredibly influential.

Despite the Met Office’s supercomputer, I’m more inclined in future to pay attention to Harry Kershaw. Harry, an amateur weather forecaster from Sale, predicted in late December that we were heading for one of the harshest winters in years. Maybe it was simply a lucky guess. Lucky or not, Harry’s is still a better record than the bloated, massively-funded, climate change-obsessed Met Office. How appropriate that the speed of the Met Office’s new computer, one of the 20 most powerful in the world, is “one petaflop”.