I normally use http://www.metoffice.gov.uk/.
It only goes up to five days though.
There’s mention of 30 day forecasts, but I can’t find them. Am I missing them somewhere ?

Is there a long range forecast online anywhere ?

They often struggle to get it right the next day so 30 days and you’re really pushing it.

A 30 day forecast, regardless of what website you use, is going to be pretty useless.

If I’m planning a ride i use metoffice, metcheck and accuweather, and just decide the best forecast is whats going to happen. I have about a 50% success rate.

As others have said, long range forecasts are not accurate, and beyond 3 days is pretty much guess work.

Weather underground, and yr.no both give good long range forecasts (in the 5-7 day region) which are less inaccurate than some.

Accuweather seems reasonably accurate but
Gavin Partridge does some really nice analyses on youtube, explaining what some of the forecasting models are doing.
http://www.gavsweathervids.com/

Go here:

http://www.theweatheroutlook.com/twoother/latest.aspx

This is the weather forecaster guy discussing the models for up to a week or so ahead. Much more information than a simple picture of a sun or some raindrops, and he talks about the possibilities that are ahead and their likelihoods. Very interesting, and a good insight into forecasting too.

If you want an accurate 14 day forecast then you’re out of luck, they haven’t been invented yet.

As others have said, long range forecasts are not accurate, and beyond 3 days is pretty much guess work

Nah, five days is pretty good these days. But you have to understand what you’re looking at. If it says ‘rain on Friday’ and it doesn’t, then you might think it’s wrong, but what the forecasters are really saying is ‘a breakdown in the weather is on its way, which might hit just before the weekend’. So it may not rain on Friday but there’s a pretty good chance it’d be on Thursday night or Saturday.

The above link covers all of that stuff.

If you look at this chart

http://www.netweather.tv/index.cgi?action=ensviewer;sess=

You can see how much the forecast models split after 3 days, then see how much after 5. That kind of data spread is useless to make an accurate forecast with.

If it says ‘rain on Friday’ and it doesn’t, then you might think it’s wrong, but what the forecasters are really saying is ‘a breakdown in the weather is on its way, which might hit just before the weekend’. So it may not rain on Friday but there’s a pretty good chance it’d be on Thursday night or Saturday.

What a load of guff. If it says rain on Friday and it doesn’t then it’s wrong, however you dress it up. If you’re making decisions based on the forecast then a day or so either way is generally not good enough. Currently 3 days is about as good as you can get for a fairly accurate forecast, especially for somewhere like the UK which is a pretty complex place to predict the weather for due to all the variables.

What a load of guff. If it says rain on Friday and it doesn’t then it’s wrong, however you dress it up.

My point is that it shouldn’t say ‘rain on friday’ it should say ‘a rain event approaching, arriving at your house at 1200gmt friday +/- 6 hours’. But people can’t handle that, they want a little picture of sun or clouds.

What I’m saying is, it’s useful to learn what’s going on rather than relying on the little pictures.

You can see how much the forecast models split after 3 days, then see how much after 5. That kind of data spread is useless to make an accurate forecast with.

They don’t always split, sometimes there is more consistency than others. That’s why sometimes they tell you there’s snow for next week (say) and sometimes they only talk about tomorrow.

Look at all of them and then pick the one that most closely matches the weather you most want to have?

I use the pressure/cloud charts from this portal

Charts

if you re-visit you can see how they evolve.

I’ve found them very accurate for the general picture however 50 miles out 7 days ahead can make a big difference to your local experience.

I realise accuracy will decline with timescale.
Even short term, MetOffice for my area is currently showing a picture of a cloud with no raindrop or snowflake for 1500 today.
For those who want a simple yes/no forecast, that means it won’t rain and they can feel a satisfied indignation when it does.
I prefer to look at the next column where it shows <20% chance of precipitation, which I take to mean it probably won’t rain, but don’t be surprised if it does.

Putting Worcestershire and Precipitation or Average Temperature in to http://www.netweather.tv/index.cgi?action=ensviewer;sess= , even without fully understanding all the terminology, shows it probably won’t rain much and probably won’t rise much above freezing over the next 14 days.
I’d rather take a probablity from someone who seems to know what they are doing than simply dismiss it as “useless”.

In true stw style, not exactly what you asked for but this is my favourite weather forecast site.
I like the style, and somehow they’re usually more accurate than more local based forecasters.
Although “long term” to them is 10 days, not the 30 you asked for, they do seem to know what they’re doing.

http://www.yr.no/place/United_Kingdom/England/Worcester/long.html

My point is that it shouldn’t say ‘rain on friday’ it should say ‘a rain event approaching, arriving at your house at 1200gmt friday +/- 6 hours’. But people can’t handle that, they want a little picture of sun or clouds.

Agreed. Timing and location are not always perfect.
The best thing to look at for short term forecasts is the rainfall radar.

it should say ‘a rain event approaching, arriving at your house at 1200gmt friday +/- 6 hours’.

So that means rain between 0600 and 1800 on Friday, or put another way “rain on friday”. And if it doesn’t rain on Friday it’s wrong.

I heard on the radio this morning that because this year is starting like it did in 1976 we’ll have the same heat wave summer as 1976.

I won’t hold my breath on this

ECMWF is useful but ‘only’ goes out 10 days. Gives some clue as to what systems are around and where they might be, but my rule of thumb for anything beyond short term is to check a few different sites and only give them any credibility when they start to converge on a similar prediction.

As above http://www.theweatheroutlook.com/twoother/latest.aspx is where it’s at

The data and charts and different results from the computer models shows just how difficult it is to be specific about our weather but the Easter weekend is looking apocalyptic at the moment

So that means rain between 0600 and 1800 on Friday, or put another way “rain on friday”. And if it doesn’t rain on Friday it’s wrong.

Alright then, 0000 GMT +/- 6 hours. It was just an example.

My point is that the weather forecast can do a great job of predicting what’s happening, but a small variation in the prediction can make ‘wrong’ technically.

So it pays to learn about what’s happening.

In true stw style, not exactly what you asked for but this is my favourite weather forecast site.
I like the style, and somehow they’re usually more accurate than more local based forecasters.
Although “long term” to them is 10 days, not the 30 you asked for, they do seem to know what they’re doing.

http://www.yr.no/place/United_Kingdom/England/Worcester/long.html

+1