I was at the dentist yesterday, leafing through a copy of Private Eye in the waiting room. I think my scepticism (of weather forecasters) and derision (of loud and inaccurate weather forecasters) – especially amateur ones, or those who predict Armageddon every time they give a forecast – is well documented.
But this article caught my eye (Monbiot’s blog is their source):
Back in January Guardian columnist George Monbiot took issue with the use of “extreme” weather forecasts by the Mail and Express to generate eye-catching headlines. Spurning the Met Office, both titles relied on these often erroneous predictions on Positive Weather Services (PWS), run by one Jonathan Powell.
Monbiot discovered that the people who Powell claimed were helping put together the “service” did not, in fact, exist. Having been rumbled, Powell explained that “as I can’t find anything to back anyone up then quite frankly PWS is now more trouble than it’s worth and in debt. Therefore, I have taken the decision after six years to close the business forthwith.”
But you can’t keep a good weatherman down. Powell popped up in the Express a few months later, this time under the guise of Vantage Weather Services.
“NEW COLD SNAP TO LAST FOR A MONTH,” howled the paper’s front page on 15 May. Hack Nathan Rao quoted Powell’s warning to readers: “It looks like this is going to continue well into June. Summer is really on hold, with no sign of it in the near future.”
Whoops! Four days later Rao was back on the front page, but this time to report: “GLORIOUS SUMMER IS ON WAY.” He even turned to Powell for his verdict: “We will get a glimpse of summer next week with temperatures reaching 73F in the South [in fact they were up to 80]. It is going to turn cool again after that, but this extremely cold May is a precursor to some extremely hot spells during the summer.”
Phew, what a scorcher!
In actual fact, although May began quite coolly, as most people will recall it ended with very high temperatures indeed. Overall, May was just “average” in terms of temperature, rainfall, and so on. Any extremes were localised. But that didn’t stop the Mail predicting that May would be the “coldest for 100 years” back in April (using PWS as its “expert source”). You can see an example of another PWS forecast here. History obviously reveals how humongously wrong it really was.
It’s worth reading the full Monbiot blog post. Private Eye’s article doesn’t do it justice, and the obvious deceptions taking place are quite astounding. Monbiot asks if the demise of PWS and the deceptions by its founder will stop the Mail and Express using Powell for weather forecasts.
Well, no. Yesterday, the Express – once again using Powell and Vantage Weather Services (which appears to have closed down as of November 2014) as its weather guru(s) – predicted “downpours” in July and August. Much of their evidence seems to be based what the weather is doing at the moment – and Powell’s previous howlers listed by Private Eye and Monbiot followed an identical pattern.
In that same Express article, another group, Netweather, neatly side-stepped the issue of dodgy forecasting and said:
You’d be forgiven for thinking we’d skipped summer entirely and arrived in the middle of the autumn storm period.
So please take care if out and about, and watch for some treacherous driving conditions and possible floods.
It looks like this lot are diversifying into traffic advice instead of the weather. It’s also worth pointing out that Michael Fish is their senior forecaster. Remember 1987, anyone?
But the likes of AccuWeather – the agency which provides forecasting on Android devices – is no better. The information it provides is only accurate for stuff that happened 10 minutes ago – so it’s more of a “postcast”, really. Anything ahead of time is nonsense. Only yesterday (12 June) it showed this on its 4-day forecast:
Already, one day later, it is forecasting this – the first icon in the image below is the same day as the second icon in the one above:
This is minor compared to what it can show sometimes, but the forecast for today was totally wrong only yesterday, and AccuWeather has updated that “forecast” so it was right! The icon showing today’s weather is just a statement of what actually happened.
EDIT: On the morning of 14 June, this is what it says about today (Thursday) – that first icon (they represent “today”, “tonight”, and “tomorrow”) appears primarily because that’s what it’s like outside at the moment… sunny:
Of course, it didn’t say that yesterday or the day before, so its forecast was completely wrong.
Google says this at the moment (and I can’t really comment on its accuracy because I haven’t tried it before, but it’s wrong about today):
Incidentally, with today being 13 June, I predict that the weather will remain like it is now until around 17 June, then it will pick up a little (it won’t be wall-to-wall sunshine, but it’ll at least be a bit sunny) and will stay that way for at least another 4 days after that.
Let’s see if my prediction comes true. If it doesn’t, I don’t care – because I’m not a weather forecaster. But it will be fun to see what happens. If I get it right I’ll see if I can do it again using my information sources. Or I might get bored.
Edit 17/6/2012: Hey, so far so good. Just adding to that prediction, I reckon the way the weather is today is how it is going to remain for at least the next 7 days. Not wall-to-wall sun, but less wet than of late (showers), and with higher temperatures. This mainly applies to the central part of the UK – Scotland might still be shitty (it has a head start anyway) and so might the extreme south.
Edit 18/6/2012: An update to the above. I now reckon that we might have some unsettled weather again around 22nd June, which will last a couple of days, then it will go back to like it is now. I’ve had to revise my previous forecast because the data I am using have also changed.
Edit 6/7/2012: OK. I got bored. But I was correct with all my forecasts unless the source data changed. And when the source data DOES change, it means making a totally different forecast. Any weather forecaster has the same problem, so they can never be totally accurate. Or – as is the case with the “long range” forecasters – they just take a guess.