Source: BBC Radio 4: Today Programme
URL: N/A
Date: 28/03/2011
Event: Roger Harrabin discusses the Weather Test project
People:
John Humphrys: Who'd be a weather forecaster, eh? Get it right, and nobody notices, get it wrong and, well... Do I need to remind you about the "barbecue summer", which we spent huddling under umbrellas, watching the barbecues fill with rain. Anyway, that's all going to change... maybe. You may not have heard of it but the BBC Weather Test is about to begin, and the full weight of this programme is behind it! It's tasked, no less, than to assess the reliability of our forecasts. Our Environment - [cough] 'scuse me - Analyst Roger Harrabin has been working on it for the past year now, a whole year, Roger...
Roger Harrabin: Yes, er, pretty pathetic, John, when I first mentioned this idea I thought we'd sort it out in a few weeks. But it is [sic] been a long and detailed collaboration with the Royal Meteorological Society, Royal Statistical Society, Royal Astronomical Society and various other experts. And what we're trying to do is to come up with a protocol so we can measure different weather forecasters against each other, and find out how accurate they are, so...
John Humphrys: Can't we just look out of the window?
Roger Harrabin: Well you could look out of the window but it wouldn't be very helpful for tomorrow or the summer, and that's what we'd be looking at, so if, for instance, say the weather girl says okay, take a brolly today, how much should you trust that forecast - 100%, 80%, 70%? We all have an instinct, but we want the statistics. Or then, you know, if somebody forecasts a "barbecue summer", how much should you trust that - 10%, 5%, 15%? Well, the truth is we don't know, and what we've done now is devised a protocol, which we'll be launching this evening at the Royal Institution in Central London. It's an open meeting, people are welcome to come, and we hope this will be agreed as an accurate protocol by which we can measure the different forecasters.
John Humphrys: The trouble is, even the different forecasts themselves are - even when you look at a specific forecast, like, for instance, we've had one this morning, haven't we, from the front page of the Telegraph which...
Roger Harrabin: "Brolly and sunblock", Positive Weather Solutions...
John Humphrys: Ah, exactly, and it depends which version you read - in the Mail, they say it's going to be shocking and dreadful and we're all doing to drown, um, in our beds. And in the Telegraph it's a "barbecue August". I mean, you know... What do you make of it?
Roger Harrabin: I think that's partly a function of news reporting and us being selective, but what we're really trying to do here, and it has been difficult, is to work up a formula by which we can accurately test people. So this "brolly and sunblock", for instance, you know, if you have a phrase like that - it's a phrase, how can you test and measure a phrase against another pos - another weather forecaster's forecast? And hopefully that is what we'll be able to do over several years, but we're not there yet. We still have to, after agreeing the protocol, we still have to get the different participants to take part. The Met Office are a bit nervous, because they see their reputation, potentially, on the line. So there is a long way to go yet, John, and I fear that one year may just be the start of it.
John Humphrys: Do you know what, there are many of us hoping you'll fail, you take all the fun out of it, that's the point, it's much more fun this way. Anyway - it's on the website, give us the website.
Roger Harrabin: Well, it's on the Today Programme website and you can click through there to the Weather Test - I'm sorry, to the Weather Club website, where you'll be able to find all the details.