Weather forecasts

[Kathyan]

Why do they give a five day forecast only for it to change every day? I looked at the five day forecast for our area yesterday and it was cloudy/rain on Monday, fine on Tuesday and cloudy/rain on Wednesday. I've looked at it again today and it still says cloudy/rain for today but the next two days are fine! What's the point?

[jake-the-peg]

Sometimes the forecasts are more reliable than others because of the weather systems involved.

I believe the Met Office re-run predictions with slightly different values to what is currently being seen, a degree out here, a few millibars there wind a little stronger.

If the model predicts much the same weather as the proper one they know the forecast's pretty reliable but if it's way out then they know there's a lot of luck involved.

Unfortunately they don't seem to think that the general public is able to cope with being told that the forecast is say 60% reliable or 80% reliable - maybe they think we'd think they were not up to the job or something.

Personally I tend to think we could well do with knowing how much we can rely on the forecast.

[Quizmonster]

A professor of statistics claimed a few years ago that if you say...

"Tomorrow's weather will be pretty much like today's"

...you'll be right about two thirds of the time! If true, that certainly can't be any worse than the forecasts offered by the media.
Try ignoring all the forecasts, Kathy, for a week on the above basis and see how it goes. Seems about right to me.

[Clanad]

One needs to understand that all weather forecasters, at least here in the U.S. (and they are brothers and sisters under the skin worldwide) have two jobs... the second being writers for newspaper horoscopes.

In aviation weather, at least, meterologists own up to the fact that the forecast for about a 24 hour period is fairly accurate, but degrades significantly beyond that point. The general trend can be forecast, but specifics are very changeable, especially with the recognition and study of what they therm micro-climates; areas of weather phenomena that are very local in nature...

Enjoying the summer, Q?... any boulangerie la Francaise recently?

[OBonio]

My dad was a meteorologist in the RAF for over 40 years. Quite a good one at that. Went to college with Michael Fish to boot. He freely admits anything further than 5 day forecasts are pretty pointless. The models used by the computers tend to predict very stupid things any further in the future. Hence there's still a large degree of human calibration of the computer's results. Sometimes my dad even ignored the computer's ideas completely and went with his own experience. Sometimes for better, sometimes for worse. All in all, as Clanad mentioned, the 24 hour forecasts are going to be more accurate as they have more data to base predictions on.

And my dad has never written for any newspaper horoscopes.

[Quizmonster]

Kathyan, my apologies for hijacking your thread to reply to C.
Summer, C? Clearly - though you obviously know lots about aviation's weather-patterns - you are not au fait with Britain's recently...we've scarcely seen the sun since April!
I was actually due to hit la belle France tomorrow with a friend. However, his mother has been taken to hospital so that trip is off...next one due on 7th August. We'll make up for it then by eating and drinking twice as much! Cheers

[kwicky]

Have you heard of chaos theory? Thats where a butterfly flapping its wings in a distant country causes a tornado in yet another country some days hence..
The weather works on a similar principle. A small drop of pressure joined with infinite number of variables produces unexpected results which accumulate. Massive cray computers at the weather centre try to extrapolate all the known variables mathematically to produce simultaneous equations hoping to provide an answer. On a similar theme there is a computer language to feed in these variables and is called S.L.A.N.G. and which is used by missile designers.

[THECORBYLOON]

Different forecasters use different fir cones and seaweed so their forecasts are bound to differ.

[heathfield]

One of the problems for forecasters is estimating the rate at which weather systems will move. I remember a particular week last year when, for our area, for a full seven days, all the 24hr forecasts were wrong - simply because changeable weather came through much faster or slower than forecast. They got the weather right, but got the timing very wrong.

In the UK, even a three-day forecast is unlikely to be correct in the end.

[Azalian]

The rule of thumb for forecasting is follows:-

Up to six hours out; accuracy = 95%
twelve hours= 90%
24 hours = 75%

[smssms]

Surely the short answer to the original question is:

They keep updating their prediction as new information comes in. It is a ROLLING 5-day forecast....

[Quizmonster]

H, I have to say that - when I read your comment above about forecasters who "got the weather right but the timing very wrong" - I was immediately reminded of Eric Morecambe's wonderful comment at the piano to the effect that he was "playing all the right notes but not necessarily in the right order."
The forecasts concerned were, therefore, about as effective as his playing...ie utterly useless!