UKC

Looking at the forecast for tomorrow, so would expect relative accuracy and consistency, but the Met Office gives a temp of 13c versus 18c on the BBC and weather.com.

Frequently see such differences in the short term forecasts. Who to trust? 

Treat the variance in models as an indicator of the uncertainty of the forecast.

We know there are lots of downpours and thunderstorms at the moment, but exactly when and where they hit is very unpredictable.

So you could have a sunny afternoon at 18C, and then a storm comes in (or doesnt ) and it drops to 13C

Another answer is to use the Met Office forecast rather than some random American website that happens to have a good name. Anyway, it'll be 20C tomorrow if you are in most of England. 

One issue that I have noticed on the met office app is that " current location" may not be at the altitude you are actually are situated.

On the app, Torver, cumbria is 12 degrees ATM, Torver, coniston is 20 degrees. Must be up the hill somewhere.

Lapse rate is 6C/1000m. Its not (just) elevation.

Maybe a weather balloon then at 1100m.

Perhaps but the BBC? 

All forecasts are for sunny clear skies. I wonder if the Met Office accounts for the altitude of the location (Amurlee in this case). 

I see similar differences in wind speed forecast.

At the moment, none of the forecasts seem very reliable with regard to wind and rain. We've been expecting heavy showers at various intervals in the last few days, only about half of which have actually arrived. 

If you type Amulree in the "Find a forecast" box at https://www.metoffice.gov.uk/ you get a forecast for Ben Chonzie. 

Which is the correct answer.

glenfinnan is another one to watch out for

Climbed lower grade stuff on Bowden Doors today. Forecast indicated we would be rained off around 13:00 with increasing chances of showers thereafter. A big rain cloud missed us by a few km (aside from a few spots) and we climbed rain free. Lovely conditions when we left. I'm not complaining forecasters sometimes get things wrong.

On a slightly related note, has anyone else noticed that the Nant Peris temperature forecast on metoffice is consistently 4 or 5 degrees lower than similar, nearby areas. I can factor that in but not sure if the rest of the forecast is off as well. 

Same problem - you are getting the "nearest forecast" to Nant Peris which happens to be for the top of Y Garn. The website tells you this, but the phone app doesn't. 

I agree this sounds like a misleading anomaly. 

However, anyone wanting mountain area forecasts, including in the valleys, should look specially for them. 

MWIS

Met office mountain weather forecast.

Metcheck hobbies - climbing

All three will give you a good idea about how conditions change with height.

As Dan said, the variance between forecasts is an indicator of low confidence / unstable conditions.

Also as mentioned the spot forecasts may not all be for locations at the same altitude.

There are several widely used forecast models, they vary in their resolution (how large an area is covered in one sample) and timescale (both how frequently they are updated and the period covered by each prediction).  On top of this some models include localised effects caused by terrain eg sea breezes, katabatic winds, inversions, others don't.  Some interesting gen here: https://windy.app/blog/what-are-weather-models.html and here https://windy.app/support/windy-app-weather-forecast-models.html

I like to use the feature on windy.com that allows you to compare forecasts from different models side-by-side, eg: https://www.windy.com/multimodel/53.156/-4.03?53.109,-3.878,11,i:pressure

If you regularly compare the different forecasts for the same area you'll develop a feel for which tend to be most accurate for a given location.

Garve too, which gives you a forecast for the summit of Ben Wyvis; and Struy which gives the forecast for whichever Strathfarrar Munro is closest. This is obvious on the desktop version of the website but on the app there is no indication that the forecast is being shown for a different location to the one you entered.

Another useful way of understanding confidence in a forecast are the ensembles.  

A meteorologist can explain better than me I think... These run the forecast model from many different starting points, each randomly perturbed in relation to the uncertainty on the inputs, to give an indication as to how sensitive the forecast is to the current inputs.

https://www.netweather.tv/charts-and-data/ecmwf-ensembles

Every day is a school day!  That's very cool and interesting, thanks!

Ha haa, not so long ago I would have been surprised by that comment, maybe not even have understood it. But if that confuses anyone BBC weather is a redistribution of meteogroup weather data, which is part of dtn.com (Mailing Address: 11400 Rupp Drive, Burnsville MN 55337).

Met office is probably as good as you will get considering the nature of predicting stuff. Followed second by my late grandad's banjo barometer weather station (not sure whatever happened to it), followed in third place by "BBC" weather.

Thank you, this is useful to know. I wonder why they chose there and how many other summits have their own forecast. 

Would be great to see a map of met office forecast locations, does anyone know of one? 

The forecasts are based on volumetric grids, and you can (if you have access) interrogate them at the nearest point to any given latitude, longitude and altitude (or pressure).  The same for other forecast sources like the US GFS and ECMWF.  

Intermediaries like the BBC get to choose how they sample that grid (nearest point? Interpolation?) and how they present that information.  Providers don’t tend to make their method clear.  

All very fraught in mountain areas which is where the human expertise in interpreting the models comes in to its own with things like MWIS.

I think all of the Munros plus a few other key summits in areas where there aren’t any Munros have their own forecast. Not sure how many English and Welsh summits do - presumably quite a few. The Met Office website does have a map of all forecast locations (hills and towns/villages) somewhere but I can’t remember how to find it offhand - must check next time I’m on my laptop. The app doesn’t - in fact it doesn’t even show you the location of each forecast (a bit iffy when there are multiple places with the same name).