I know living in an upland area we have variable localised weather, but yesterday was supposed to be dry. We had rain by lunch, drizzle all afternoon, evening & night, and drizzle this am. No sign of rain on the BBC forecast..current weather showing dry with sunny spells. No its not. And early morning frost yesterday, although the day dawned bright & sunny it only lasted 5 hours.
Friend of mine sold his place in France earlier this 'summer'. After 4 weeks of lousy weather he flies out to Ibiza today for a month to search for a place to live. He's had enough, but will keep a base here!
|
I'm a dedicated follower of fash weather since the days when I worked outside for 14 years.
I noticed a BIG change when we moved from sowf lunden down to Hastings in 87 and wondered what the blimmin hell I'd done!
The Canaries really do have the ideal climate - better than mainland Spain / Greece / Cyprus / Florida etc. etc.
Coming back from Tenerife in 96 and settling in Cornwall, well, rain, rain, and more forking rain :)
The Met office are a bunch blimmin muppets if you ask me - a total waste of money IMO.
I check out the wev daily in Cyprus / Tenerife / Javia/Xabia / Malaga / Brisbane (I told y'all I was dedicated!)
This is my favourite weather site:
www.meteoxabia.com/datos_combinados.htm
Followed by: www.malagaweather.com/
And: cyprus-storms.webs.com/
|
|
My favourite weather site is the Met Office. It gives the forecast for every three hours. www.metoffice.gov.uk/weather/
|
My fave is the Norwegian site:
www.yr.no/
Fill in your place name and away you go. Seems to be more accurate than our met office.
|
They do a "Live" weather widget that you can download free to your desktop (or website)!
I`ve got it, its more accurate than T.V News/Weather! - (for our region anyway!), and you can programme in the nearest weather station to you!
yowindow.com/download.php
|
>> My favourite weather site is the Met Office. It gives the forecast for every three
>> hours.
Recently they started giving one hour forecasts for today and tomorrow (UK).
|
>> >> My favourite weather site is the Met Office. It gives the forecast for every
>> three
>> >> hours.
>>
>> Recently they started giving one hour forecasts for today and tomorrow (UK).
Sorry - I'm thinking of the BBC weather, although AFAIK they get their info from the met office so presumably the latter could give hourly forecasts if they wanted to. Perhaps the BBC has to pay extra for it.
|
>> hours. www.metoffice.gov.uk/weather/
>>
The new version www.metoffice.gov.uk/public/weather/forecast/ is even better, and it allows you to zoom in and select your map display from the options below
Weather
Temperature
Temp. map
Feels like temp.
Rainfall
Surface pressure
Cloud and rain
Wind
Wind gust
UV
Cloud cover
Pollen count
The rainfall radar history gives you to make a good forecast of your own as to where the rain is heading for in the next few hours.
|
Weather forecasting in France seems even more haphazard than the UK.
Recently, I started using the specialised three-hourly forecast for farmers but it doesn't seem any more accurate than the rest and always seems to be forecasting fine weather.
When I mentioned this to a French neighbour he said 'well, farmers work hard so they deserve a good forecast'.
As I have said before, France is a parallel universe.
|
I find the hour-by-hour forecast on here to be pretty accurate up to a day in advance:
uk.weather.com/weather/today-Coalville-LE67
Combined with this: www.xcweather.co.uk/
And this: www.raintoday.co.uk/
I can usually get a reasonable picture of what the day holds.
|
|
Actually I used to find the relatively imprecise Ceefax forecast pretty good - it just showed you a lo res picture of the UK split into 3 or 4 weather regions with day and night forecasts. KISS :)
|
|
Rain from the West, spreading to all most parts by dawn, they probably just use the same chart over and over again!
|
They got the weather forecast wrong here as well, its the same everywhere.
Its the middle month of winter and was supposed to be 24 degrees, in fact it turned out to be 27!!
Unforgivable.
|
>> Actually I used to find the relatively imprecise Ceefax forecast pretty good
The BBC and ITV use the Met Office (part of the MoD?)... how it's reported/displayed has nothing to do with the Met Office source data.
Recently they computers said storms over London and the forecasters said it was wrong and broadcast wrong info.
The current Met Office computers are very powerful and if the model still being used is accurate the forecasts should be too..... but it assume the model is still right.
|
>> The current Met Office computers are very powerful and if the model still being used
>> is accurate the forecasts should be too..... but it assume the model is still right.
Unless the weather is in a 'chaotic' state, when no matter how accurate the data, the prediction can still be way off. But I don't think that applies to forecasts for the next day or 2.
|
>>the Met Office (part of the MoD?)...
Run as a P&L within Department for Business Innovation and Skills
|
|
Did I remember incorrectly that they were even once paid for by the MoD?
|
No, it was part of the MOD, in the late 80s I think, and only for about 10 years, then some government shuffle or other decided it should be less military and more commercial. Essentially it wasn't broken so it obviously needed to be fixed.
Originally, it was part of the Air Ministry for a looong time.
Last edited by: No FM2R on Sat 1 Sep 12 at 23:44
|
>No, it was part of the MOD, in the late 80s I think, and only for about 10 years,
Up until last year it was still part of the MoD but operated on a commercial basis.
|
SPOILER - Boring explanation of weather forecasting follows but I'm going to do it anyway.
>how it's reported/displayed has nothing to do with the Met Office source data.
rtj is correct.
The output from an ensemble forecast is a range of possible outcomes with the distribution representing the uncertainty in the forecast. If atmospheric conditions are chaotic then the range of outcomes and therefore the uncertainty can be quite wide. Most of the outcomes though will tend to group around a particular value which is statistically the most probable outcome.
Forecasts for public consumption invariably present the most probable outcome. They rarely show you the whole range of possible outcomes.
For example, the BBC forecast will often say something like "overcast with occasional showers", they do not say "cloud cover between 6 and 8 okta in legacyland, precip is most likely to be around 3mm/6hr but there's a 25% chance it could be nothing or 6mm, and a 10% chance of 9mm".
Would it be better if Joe Public was given the whole range? Do they even care beyond "Is it going to rain this weekend?" I don't know.
>but it assume the model is still right.
The Met Office, like most forecast centers, run multiple models and generally two (or more) versions of each model. One version is the official "production" model, the other versions are run in parallel and contain slight tweaks to see if accuracy is improved. Only after a long test period will a tweaked version showing an improvement become the production model and the cycle start again.
Last edited by: Kevin on Sun 2 Sep 12 at 21:07
|
We live on a fairly small island, stuck out on a limb in the middle of one of the worlds most complicated series of currents, winds and temperatures.
We aint got a hope in hell of getting anything like an advance accurate weather forecast no matter what "models" you use, due to too many variables. Add to that of course climate change which throws your models a complete curve ball.
Ask them why this year has been the wettest ever, and they can tell you the what, the mechanism, but not the why.
|
|
You have mail Z. Not for public consumption.
Last edited by: Kevin on Sun 2 Sep 12 at 23:42
|
Got it
For gods sake man, how many times do I have to tell you, I do not want to see you in your school girl uniform.
|
|
You said you liked navy blue knickers and knee-length white socks. :-(
|
>>Z. Not for public consumption.
He certainly isn't.
|
|
Get the mods to send me an email address and your location and I'll extend the experiment to South America.
|
Kevin,
My profile is visible, but for ease my e-mail addess is no_fm2r@outlook.com
|
Mark,
What is your nearest village/town/city/point-of-geographical-interest?
|
>> Mark,
>>
>> What is your nearest village/town/city/point-of-geographical-interest?
Its a lump of goat poo with an interesting array of flies on it.
|
The flies aren't very interesting.
Last edited by: No FM2R on Mon 3 Sep 12 at 12:25
|
>Santiago, Chile.
Ahh. That works better.
I entered "Goat Poo" as a location at least half a dozen times and each time it took me to some place in Surrey with a Mitsubishi parked outside.
|
Easy mistake Kevin. it was almost certainly defaulting to "old goat"...
:-)
|
Kevin,
Today, temperature 4 degrees higher than the maximum indicated
Cloud cover wrong, even tho they covered their bums at midday, by spreading the forecast from 2 octa to 8 octal.
|
The other week the forecasters ignored what the computer said (thunder storms)... for once it seemed the couldn't blame the computer.
I know the forecasts are based on solving massive amounts of differential equations for the data they collect and should be fairly accurate. But at times the model (based on the diff eq) seems wrong... so the climate itself seems to be in a right state.
Last edited by: VxFan on Tue 4 Sep 12 at 21:16
|
Kevin;
Keeping a rough and ready diary. Send it to you in 10 days.
|
>Keeping a rough and ready diary. Send it to you in 10 days.
Thanks Mark.
If it does go wildly inaccurate I'd forget keeping a record after that (other than for your own interest). Once it goes wrong it rarely recovers.
I'll take another look tomorrow and see if anything has changed.
|
Its consistently a bit warmer, somewhat cloudier, and less windier.
You better hope it rains on Thursday or I'm afraid its not going very well.
|
>Its consistently a bit warmer, somewhat cloudier, and less windier.
That's good. I expected the temperature bias.
>You better hope it rains on Thursday or I'm afraid its not going very well.
Tomorrow will be the real indicator though.
|
Kevin, Today 13:00
Temp 3c above the forecast, Cloud cover (again they manage to forecast a spread from 2 octa to 8 octa) 0 octa, wind - too variable, here anyway, to gauge. Precipitation correct (0)
Last edited by: Zero on Wed 5 Sep 12 at 13:34
|
>Temp 3c above the forecast,
You sure?
www.met.reading.ac.uk/Data/fieldsite/cgi-bin/graph_page.cgi?RL=1&
>Cloud cover (again they manage to forecast a spread from 2 octa to 8 octa) 0 octa
Measure against the blue line. That is the expected real value. For today that would be 0 okta up until about midday increasing to about 6 okta at 6pm tonight then dropping below 4 around midnight.
Will you give Pat the bad news for next week?
|
Why does the blue line not follow the % median? And I am not in Reading, its 20c here, and still 0 Okta.
next week, its a lot isnt it. I am not telling her, she will go off on one.
|
Well, you must be wrong!
XC Weather, Accuweather and the met office say it's going to be lovely.
BBQ Saturday night with the blokes cooking and me and daughter in law directing operations with a glass in hand!
After that we eat out so any recommendations for a good restaurant around Bourton on the Water?
Pat
|
>Why does the blue line not follow the % median?
Apparently the most likely outcome at any particular point is not always the median value. I don't know why - I'll have to ask someone in research.
>its 20c here, and still 0 Okta.
Probably all the hot air from Westminster, PM's questions today.
Light cloud forming here in Reading from the West. Don't think it will reach Sunday's prediction of 6 okta though and it's a few hours late. Clouds are like women you see, always late and forever getting changed.
|
>You better hope it rains on Thursday or I'm afraid its not going very well.
It didn't rain did it? Or not very much.
Pretty cloudy though.
|
Actually it did rain, and a fair amount - but it was early - 5am ish for proper rain and drizzele until about 9am. Then damp for the rest of the morning.
But you're not doing bad.
|
|
Did you get my email, Kevin?
|
|
It was so good he's looking forward to number 16.
|
>Did you get my email, Kevin?
Just got it. Thanks for the feedback.
Three questions for you and Zero though:
1) Overall, how accurate do think it was?
2) How easy was it to understand the format?
3) Would Joe Public understand/appreciate the extra information?
|
>> >Did you get my email, Kevin?
>>
>> Just got it. Thanks for the feedback.
>>
>> Three questions for you and Zero though:
>>
>> 1) Overall, how accurate do think it was?
On the whole, not bad. But I suspect its been a freak week this week, very settled. Its been under on the temp scale all week and the cloud cover forecast varies from one end of the scale to the other during the day.
>>
>> 2) How easy was it to understand the format?
Not very to be honest. Those % probability bars will be martian to most folks. Its a good way of providing a 10 day forecast in a single snapshot, but it tries to cover its bases too much
>>
>> 3) Would Joe Public understand/appreciate the extra information?
Nah - too technical for them. As above.
>>
|
>On the whole, not bad. But I suspect its been a freak week this week, very settled.
Predicting a freak week after what we've had so far is bad?
Will you be more convinced if the rain moves in late Monday and Tuesday as predicted by the forecast I gave you last weekend?
|
>> Will you be more convinced if the rain moves in late Monday and Tuesday as
>> predicted by the forecast I gave you last weekend?
you want me to congratulate you for forecasting rain? go whistle.
|
1) Overall I was impressed. To be honest I started sceptical and so didn't pay it too much heed. By the time we got to the end of Thursday I was getting quite impressed.
2) awful. I found it difficult to build a picture in my head without studying it, there was no chance of a quick glance working. Its ok for sailers and the like, but the general public would not be prepared to invest sufficiently.
3) No. Firstly my perception would be that the public do not want more detail, they want more perceived accuracy.
Secondly there is the issue of perception versus reality. People take away from any forecast what they want. So, sunny with a chance of rain = sunny. Thier feelings about the weather forecast are emphasised by their feelins about the weaher, which in turn is driven by its relevance.
So, tell them it'll be sunny on a Tuesday afternoon when they're in the office, then the rain will barely register. Tell them its going to be sun and showers on a sunday and then greet them with an hour of rain and the barbecue will increase their displeasure with the weather, this will flow into the forecast, and the frustraiton will hit the forecaster on the basis that without him there truly is no other scapegoat.
Its not like they can move their barbecue to a working Monday because you've said the weather will be ok. Its Sunday afternoon or not at all.
So, for people who need to weather; sailors, event organisers and the like, then more accuracy will be welcome.
For the general public you'd be better off just making it bland and entertaining.
And its not just the British, the Brasilians have a joke that the name for the first hot sunny day after two consecutive rainy days is "Monday".
So you're knackered, basically.
Last edited by: Webmaster on Wed 19 Sep 12 at 01:33
|
>Today, temperature 4 degrees higher than the maximum indicated
>Cloud cover wrong, even tho they covered their bums at midday, by spreading the forecast from 2 octa to 8 octal.
Wrong enough for me to shame myself and beg for forgiveness in a public forum?
Cloud cover was pretty much spot on in Reading (the location I used). Even the threat of slight precip was correct.
I expected the temp to be slightly out, I've seen a fairly constant 1deg under-bias over the last few months. Reading is also usually a degree or two lower than inside the M25.
Another thing I expect is a slight time shift.
|
Can you make me a good one for the Cotswolds next week please Kevin?
Pat
|
>> We live on a fairly small island, stuck out on a limb in the middle
>> of one of the worlds most complicated series of currents, winds and temperatures.
And that's before you factor in effects of local topography. In the Hebrides the weather can differ between Tarbert and Obbe, never mind between the Butt of Lewis and Barra Head.
You can see Sandown from Shanklin but they quite often report differences in sunshine or rainfall.
|
We live just beyond the southern end of the Cheshire Gap.
en.wikipedia.org/wiki/Cheshire_Plain
Makes weather forecasting very difficult.
There is , however , a local very well known forecasting method as we look out to see a local hill: Mow Cop.
en.wikipedia.org/wiki/Mow_Cop
When you can see Mow Cop, it's not raining.
If Mow Cop is visible but covered in cloud, it's going to rain.
If Mow Cop is not visible, it's raining.
:-)
|
I take all weather forecast with a pinch of salt very difficult to preduct in the UK.
The prevailing winds here South West not to bad on the Eastcoast when storm arrives here.The West Coast takes most of the bad weather.This area is in a dip protected a little by the Humber.I look at the Barometer when pressure drops.The weather girls are nice looking do.>:)
|
|
i use the METARS for east mids aiport , thay are usually spot on ...planes need the info
|
|
Its not a forecast tho is it, its a current weather report.
|
|
works for me zeddo...i just take a look at the clouds from my window and do a simple relative bearing.. job done
|
|
but you dont know what your window will be telling you in 8 hours.
|
i update on the move , the metars update every hour...i just make sure im zulu
|
>> Its not a forecast tho is it, its a current weather report.
True, but most Metar reports also include a TAF
|
Let's not forget that the met office produce forecasts that Kevin tells us are pretty much 100% accurate at 48 hours out, and 95% accurate at 5 days.
www.car4play.com/forum/post/index.htm?v=e&t=11180&m=248202
>>Forecasts provided by the Met Office and ECMWF currently have a 3 Monthly mean accuracy
>>of around 90% at 5 days - the most accurate available anywhere. Accuracy at 2 days is
>>pretty much spot on.
Suddenly in this thread,
>>"I expected the temp to be slightly out, I've seen a fairly constant 1deg under-bias over the
>>last few months. Reading is also usually a degree or two lower than inside the M25."
This even though, Zero says "They covered their bums at midday, by spreading the forecast from 2 octa to 8 octal."
That's not a forecast, that's nearly a dead cert - and they even managed to get that wrong, if it was zero.
|
I thought I'd posted a reply to this last Wednesday but obviously not.
>Let's not forget that the met office produce forecasts that Kevin tells us are pretty much
>100% accurate at 48 hours out, and 95% accurate at 5 days.
>This even though, Zero says "They covered their bums at midday, by spreading the forecast
>from 2 octa to 8 octal."
Half an hour before you made this post I explained that the expected cloud cover was 0 okta up until around midday. Which it was. It was my fault for not making it clear how to read the forecast not the forecast that was wrong.
>That's not a forecast, that's nearly a dead cert - and they even managed to get that wrong,
>if it was zero.
As I said in the thread you provided a link to, I think you need do a bit of research into how forecast accuracy is measured.
You're quite content to carp from the sidelines but unwilling to make the effort to understand the subject. Why is that?
|
Wrong weather forecast again - Dog
