Let’s save it, rain or shine

Does anyone want to buy a weather forecaster? With Liam Fox overseeing the Ministry of Defence and promising to make cuts "ruthlessly and without sentiment", it is becoming ever harder to believe our national forecasting service will survive much beyond this glorious - dare I say it, barbecue - summer.

The trouble is, the Met Office is a soft target. We are more sceptical about scientists' ability to predict the weather than we are about an octopus's ability to predict the outcome of a football match. This is largely to do with our own fear of complexity.

Few of us get enough information to judge the quality of the forecast. As I write, one forecast says the overview for the day is "a good scattering of showers mixed in with brighter weather for many of us". Snow or hail would be a shock; beyond that, the words are fairly meaningless.

But, in fact, we don't want our forecasters to be more specific. The most scientifically accurate statements that a forecaster can make involve probabilities, but probabilities leave us floundering. A study in the United States, for example, showed that most people thought "a 50 per cent chance of rain" meant that the forecasters hadn't a clue whether it would rain or not.

What it really means is that, in a given set of conditions, it rains half of the time. But who has time to think through the implications of that when Newsnight is about to start? It's far easier just to let something concrete settle in our minds and, when the next day rolls around and it doesn't happen, complain that the forecast was wrong.

The meteorologists' own science of "forecast verification" is no paragon of detached objectivity. Across the world, various weather agencies measure their success in different ways. The UK Met Office, for example, takes its computer program's forecast and compares the numbers it produces against actual measurements from the weather stations - rainfall, pressure, humidity, and so on. These are combined to give what is called the "numerical weather prediction" index.

Many meteorologists turn their noses up at this, muttering darkly about simplistic approaches to measuring forecast accuracy. But the World Meteorological Organisation thinks we have something worth holding on to: it consistently rates the Met Office as one of the world's top two (Japan is also blessed with accurate forecasters).

Perhaps that accolade alone should make us think twice about selling off the Met Office. To me, however, there is an even more compelling reason.
For most of us, the weather doesn't matter much - generally, we do what we do, come rain or shine. Accurately forecasting and monitoring climate change, on the other hand - a less obvious part of the Met Office's remit - matters to everybody. The idea of making that function a slave to market forces sends a cold front down my back.