Long range forecast

[david_e]

What are the places to find out the likely weather patterns for the period from 17th October for 10 days. With half term approaching and Hurricane season kicking off am trying to do some forward planning.

 

[rhinorhino]

More than five days ahead is really too far ahead to get any accurate forecasts but you will find all the data you need neatly gathered togeather on www.greatweather.co.uk

 

[david_e]

Thanks, interesting site, have put it in my favourites. Lets hope they get it wrong as it looks pretty horrible for half term!

 

[Guest]

Regardless of soothsayers, witch doctors, government long range forecasts, donkey's tails, St Swithun's Day and all the rest, no forecast longer than about five days is worth the paper it is written on.

 

[qsiv]

... and I think you'll find that the best tehy'll claim for 120hr (5 days) f'cast is about 20%

Personally I find the US Navy (FNMOC/NOGAPS) surface wind and wave to be good, and reasonably accurate unless the picture is very complex. The USAF 'pretty' pictures give good indications of fronts and cloud, and then back to FNMOC and NOGAPS for the relative humidity for fog and gentle precip. For charts, bracknell still gets my vote - and they update the surface analysis frequently.

All in all we're spoilt for choice with data from the web (dont forget GRIB data too although you need the software to interpret). Some GRIB sources go further forward than any published charts, wich can be useful, provided you add sufficient salt.....

 

[david_e]

You could say the same down to a few hours I suppose. My thinking was along the lines that with all the technology and knowledge available these days I would have thought that predictions on pressure movements and cycles would be fairly good and be able to give a reasonable long range synopsis. It might be the case that they have the info or maybe they don't. If they do I just want to know about it!
That said, the site mentioned above makes the long range prediction which is helpful.

 

[ronniewood]

Try www.weatheraction.co.uk. They do a 14 day forecast for a fiver over the net . Have used it a couple of times and had better than 75% accuracy.

 

[rhinorhino]

Just out of interest what do you rate as accurate?
Even the shipping forecast isn't 75% accurate

 

[johndf]

Try www.wetterzentrale.de which gives a 9 day pressure chart forecast. Not quite as far ahead as you want to go, but forecasting beyond ten days ahead is not reliable.

Depending on the weather situation, five days ahead may be quite reliable, or alternatively totally unreliable. Some forecasts add an indication of the likely reliability of a given forecast. This is done by running the computer model several times,with slightly different inputs, and lookng at whether the resulting forecasts show much variation. If not then the forecast is likely to be fairly reliable. If there is a lot of variation then things are in an unstable situation (from the forecasting viewpoint), and so forecast will be less reliable.

For the week ahead, the BBC Countryfile forecast (Sunday between 1130 and 1200 - usually around 1150 on BBC1) gives a good picture of the likely weather for the next 5 to 7 days and the expected reliability of the forecast.

 

[kgi]

The trouble with weather forecasting is that they use computer models, well computers operate on logic, the weather operates on chaos theory.!!!!!!!! keith

 

[johndf]

...which is why they run the model with a variety of inputs to see if the current situation is stable or chaotic.

Five day forecasts have improved enormously in the last 20 years thanks to the increase in computing power allowing finer meshes to be used and multiple simulations to be run.

 

[kynaston]

Are you sailing in UK or Carribean waters? I was in Virgin Islands recently and found www.stormcarib.com to be a good site though there are others. They have reports from individuals on each island.

If you are in UK waters our weather is not really stable enough to forecast beyond 72 hours in my opinion. Others have explained that here. Good luck with wherever you are though.

 

[oldharry]

Forget it! Any forecast that far ahead will be purely hypothetical. Feed the same data into the same software 3 times over and you will get 3 totally different predictions.

And yes Wetterzentrales ' 6 day forecasts are amongst the best going at present - and even they get it wrong a lot of the time! All the other Long range forecasts I have seen have been so far off the mark this year that I wonder why they bother...