why are weather forecasts different

According to the BBC it’s supposed to be raining all day but hyperlocal weather shows sunny with some clouds.
Don’t they get all their info from the met office?

No point watching weather forecasts at the moment. Its just rain rain rain, build a boat to get down the road, more rain rain rain, can’t build a boat, all the woods warped due to the rain. Rain rain rain. Storms coming. Everywheres flooded.

Don’t they get all their info from the met office?

Nope. There are multiple weather forecasting models. Ventusky is a good way of comparing some of them. The BBC get their forecasts from Meteogroup, not the Met office.

Different models from different supercomputers. Local weather can be different to national forecast or regional forecast. Many a day it has been wet in Rammy when the forecast said dry. Over the near hill will probably be dry on those days. ☔

BBC declined to renew their contract with the Met Office a few years ago and went with Meteogroup as a cheaper option.

BBC always er towards doom and gloom scenario IME.

I use x3 apps – if two agree that’s what it’s going to be! 🙂

And why does the BBC now require sign-in for their weather app. It’s just a bleedin’ weather app, I’m not watching anything!

I usually find the Met Office to be the most accurate.  BBC weather has been less reliable since they swapped to MeteoGroup.

I use x3 apps – if two agree that’s what it’s going to be! 🙂

That is a good idea

So far BBC has been wrong for today and hyperlocal correct

Can the supercomputers model the very high ocean temperatures?

I’m guessing they use years of data and now that the oceans are so warm, its throwing everything out of whack.

BBC has been rubbish for Hampshire forecasts since Meteo Group replaced Met Office ~5+ years ago.

I use Met Office site/app to look days ahead and Weather Underground site/app for realtime.

for some sense of long range predictions I use WX–

https://www.wxcharts.com/?panel=default&model=gfs,gfs,gfs,gfs&region=europe&chart=overview,850temp,wind10mkph,snowdepth&run=12&step=003&plottype=10&lat=51.500&lon=-0.250&skewtstep=0

for the next five days forecast, I use Met Office https://www.metoffice.gov.uk/  and Met Éireann https://www.met.ie/ they rarely ‘agree’

for a forecast for the next five hours I use Met Office rainfall radar maps, which is usually accurate enough:

https://www.metoffice.gov.uk/weather/maps-and-charts/rainfall-radar-forecast-map#?bbox=%5B%5B54.1173828217967,-7.091674804687501%5D,%5B55.035892710948424,-3.5760498046875004%5D%5D&model=ukmo-ukv&layer=rainfall-rate&timestep=1666191600000

all good fun 🙂

No point watching weather forecasts at the moment. Its just rain rain rain, build a boat to get down the road, more rain rain rain, can’t build a boat, all the woods warped due to the rain. Rain rain rain. Storms coming. Everywheres flooded.

Alright Noah, no need to panic

Physics controls the climate system but how you compute that physics, in different models, and with different input data means you get differing outputs. Some interesting comparisons here: https://meteologix.com/uk/model-charts

Physics controls the climate system but how you compute that physics, in different models, and with different input data means you get differing outputs. Some interesting comparisons here: https://meteologix.com/uk/model-charts

I like to watch the ‘deep dive’ forecasts by the Met Office, the forecasters go into the different models/scenarios, which obviously differ especially with the longer-range forecasts:

Can the supercomputers model the very high ocean temperatures?

I’m guessing they use years of data and now that the oceans are so warm, its throwing everything out of whack.

The sea temperature measurements are pulled in to operational weather forecasting models on an hourly basis – – indeed on a time scale of minutes for some data sources (a process called assimilation).  forecasting and observing sea temperatures are important for many weather users – not least of which are military fast jet pilots who need to know sea temperatures before they decide whether or not they fly in uncomfortable full immersion suits.

Observations of sea temperatures in real time come from a range of different sources including satellites, merchant and military ships,  floating weather buoys in the middle of nowhere in the oceans – and in some cases seals with temperature sensors glued to their heads.

The fact that oceans are currently warmer than they were historically makes no difference to operational weather forecasting as the physics is very well understood and consistent.
Weather forecasting models don’t work by looking at last x years of weather records for a location and extrapolating forward as to what tomorrows weather will be like as a result – it’s done by measuring the weather we have now across the world now (*) and then discretely calculating how the currently measured and distributed energy and moisture quantities measured will change according to laws of physics and fluid dynamics.

* ok so there’s a 4 dimensional assimilation process in some cases that blends last few hours worth of model runs with real weather observations to help fill time and space gaps.

So – no the oceans warming aren’t throwing weather forecasting models out of whack, although the models are tending to more regularly predict things that weren’t seen a few years/decades ago – but as we continue to warm the planet that is to be expected – just makes it harder for humans who issue forecasts based of weather model data as model results at the edge of or beyond the experience envelope occur

Different models, different methods, different data. There are large similarities across the different forecast centres in all of these, but they aren’t identical. ECMWF has traditionally been the best but it’s a while since I had much real interest in this stuff (and if you’re looking at one specific location, it’s possible that someone else is best there).

Then you get the Corbyns who just make stuff up like a horoscope.

If you look at three different forecasts and take the average, you’ll do better in the long run than if you just take any one of the individual forecasts.

If you actually need to see a forecast, look at a time-lapse map based one.

Hour by hour postcode forecasts are rubbish.  It might say 50% chance of rain at 1pm, but gives you zero context whether that is 50/50 it might be early or late by an hour but there’s a 1hour shower at some point between 12 and 2.  Or it might be a localized shower at 1pm, with a 50/50 chance of it landing on you.

So you’re sat at your window watching the weather, it’s clear at 12:59, but unless you’ve actually looked at a map you have no idea if that means you should be leaving to go for a ride or not.

There’s an illusion of accuracy because you can put in a very local postcode and you except a very local forecast to be accurate.  But weather doesn’t work like that.

Thankyou tractionman that looks very interesting.

Met Office forecast here in the Surrey Hills is sunny with clouds and that is exactly spot on.

I don’t touch BBC weather and focus on the Met Office maps: rainfall radar timelapse for what’s coming our way and can us soon,  and the other tab for a display of rain for next day or so: then you can get a feel for how localised it’s likely to be.

Fun fact: most modern forecasts from individual centres are combinations of numerous forecasts with slightly different start conditions (perturbations) which are called ensembles.

No one model is correct everywhere at every lead time. ECMWF is very strong 5-10 days, MO is better 3-5. But it depends on how you measure it and where in the world you measure it.

Talking about ocean temperatures a lot of the global models are now coupled with corresponding ocean models so you get a more complete modelling of the physics impacting both (somewhat artificially separated) systems.

The documentation here is a pretty good started talking about how forecasts are made and explaining some of the variability.

BBC has been a waste of time for a while. Met office is best with accuweather for rain over the next few hours (but even tracking the predicted rainfall radar is often wrong) oh and AccuWeather doesn’t seem to count drizzle as rain, which in Scotland is a major issue.

http://www.windy.com allow you to compare 5 different models for any location.

I have found the met office forecasts have been pretty accurate recently, although I don’t bother looking more than 48hrs ahead.