The Met Office

[spikey]

I always suspected that the Met Office exists in a parallel universe, but now I know for sure. According to a thing on the BBC News site, those rainfall probabilities don't mean what most folk think they mean. Apparently, "40% rain at 1200hrs" actually means that at 1300 hrs, there will be a 40% chance of rain having occurred since 1200hrs ...

 

http://www.bbc.co.uk/news/uk-44238426

Edited May 28, 2018 by spikey

[peter220950]

I thought the Met Office no longer supplied the bbc with weather information?

 

https://blog.metoffice.gov.uk/2018/02/06/bbc-and-met-office-an-enduring-partnership/

 

I think the wonderful new style hourly information was introduced shortly after they lost the contract, it doesn't impress, I suspect the new 'cheapest is best' outfit are running the simulations on an Amstrad 1512.

 

Peter

[spikey]

I thought the Met Office no longer supplied the bbc with weather information?

Erm ... I'm on about the forecasts on the Met Office website.

[Quarryscapes]

Erm ... I'm on about the forecasts on the Met Office website.

 

OK, but you posted an article that was specifically about BBC weather, not the Met Office. 

 

I thought the Met Office no longer supplied the bbc with weather information?

 

https://blog.metoffice.gov.uk/2018/02/06/bbc-and-met-office-an-enduring-partnership/

 

I think the wonderful new style hourly information was introduced shortly after they lost the contract, it doesn't impress, I suspect the new 'cheapest is best' outfit are running the simulations on an Amstrad 1512.

 

Peter

 

Yep you'd be right - it says in the link that the BBC get their data from MeteoGroup. 

[The Stationmaster]

I always suspected that the Met Office exists in a parallel universe, but now I know for sure. According to a thing on the BBC News site, those rainfall probabilities don't mean what most folk think they mean. Apparently, "40% rain at 1200hrs" actually means that at 1300 hrs, there will be a 40% chance of rain having occurred since 1200hrs ...

 

http://www.bbc.co.uk/news/uk-44238426

 

The Met Office website has always been very clear what the percentage numbers mean by saying 'chance of precipitation'(in the hour block the forecast shows. If the BBC has been getting it wrong that's down to the BBC.

[spikey]

Apologies for the confusion. So when the Met Office website says 40% chance of precipitation at noon, am I correct in assuming that means there is a 40% chance it will be raining at noon?

[Dunsignalling]

I always suspected that the Met Office exists in a parallel universe, but now I know for sure. According to a thing on the BBC News site, those rainfall probabilities don't mean what most folk think they mean. Apparently, "40% rain at 1200hrs" actually means that at 1300 hrs, there will be a 40% chance of rain having occurred since 1200hrs ...

 

http://www.bbc.co.uk/news/uk-44238426

Which means exactly the same as "At 1200, there will be a 40% chance of rain falling in the next hour"; which is what I always thought it meant.

 

At 1300, there isn't a "probability" of rain falling between 1200 and 1300, because it will have already happened, or not. Hindsight makes it either 100% or zero.

 

The clue is in the word "forecast"

 

John  

Edited May 28, 2018 by Dunsignalling

[spikey]

I blame the heat. The lady wife says it's my age and the fact that I still haven't found what's causing the short in my sidings. She says I'll be better tomorrow, so there's hope ...

[LBRJ]

I do not know if it just my perception but "on the whole" weather forecasts for the short term seem to have got a whole lot more accurate over the last two/three years.

[Co-tr-Paul]

The forecast here in Cornwall is so localised that the forecast for the next hour here means looking upwards......!

[Mike Storey]

If it helps any, the forecasts here, over the past 6 months, have been about as accurate as a stopped clock, no matter which channel or website you consult - they will all differ. People swear by them - I know I do.

 

On Saturday afternoon, we had FR2 and The Weather Channel, and the more local radio forecasts swearing blind that it is slightly cloudy with a bit of sun, right now, right in the middle of a hailstorm and lightning bolts, which took out the electrickery around here for between 2 to 4 hours, depending on which village you were in. An hour later, I had a text from EdF warning me to unplug everything as a storm was approaching. Nice.

 

To be fair, the weather has been so weird and unpredictable lately, as well as highly localised, that I would suspect their computerised prediction models are having nervous breakdowns. I am supposed to be laying concrete for the next section of my garden layout - fat chance at the moment.

[peter220950]

I had to smile this morning, I have the BBC weather app on my tablet, and have a couple of locations listed where we visit as well as for home.

 

For some reason I have managed to get both Droitwich and Droitwich Spa listed, though it's the same place, one merely an abreviation of the other, however it had different forecasts at the same time today, so they can't even manage to agree on that.

 

 

 

A minute ago it had different temperatures for days next week, I don't know whether to stay here or move to here to get better weather.

 

Peter

[Norm81]

I think they've got less accurate, on more than one occasion I've checked the forecast for a location which has said no more than 10% rain for most of the day. Only to find it completely chucking it down and for over an hour. I appreciate it's a forecast and they get it wrong, but when it says 10% and the reality is 100% that's getting it very wrong.

[34theletterbetweenB&D]

 The Met Office performance as presented by the Beeb's forecast service was proving pretty robust. The BBC's current meteogroup based forecast is a step back to ten years ago in performance, no more helpful than looking at the sky, movement of the barometer, and making your own mind up. Yet more evidence that the BBC has lost the plot.

Apologies for the confusion. So when the Met Office website says 40% chance of precipitation at noon, am I correct in assuming that means there is a 40% chance it will be raining at noon?

 No. From noon to 13.00 the forecast is a 40% chance that some rain will fall during that hour. That's all that forecast says, taken in isolation.

[Mike Storey]

I don't think anyone is finding it easy to accurately forecast with the weather patterns that have emerged this year. Unfair to point the finger just at the Beeb.

 

I follow three different forecasters here now, FR2, FranceMeteo and The Weather Channel. They rarely agree with each other, and none of them have been much use in telling me whether I can safely do some concreting or put out the washing or whatever, even on the same day, for months. The most exasperating thing is that rain has been forecast (including thunderstorms) here almost continuously for weeks and weeks (and is still forecast for the week or so ahead). But every other day, there is no or little rain. So another day wasted because I thought I could not get on with something. And then I risk it the next day, only for the heavens to open.

 

The weather is being more than a little strange right now.

[jonny777]

As an ex-Met Office employee of over 35 years I read this thread with some amusement. 

 

Don't mistake this for arrogance, but there always has been ignorance of meteorological sciences by the wider public, and this thread appears to prove (to me anyway) that the knowledge is not improving. 

 

As can be seen from this thread; everyone is an expert and they never remember when the forecast is correct, but never forget when the forecast is wrong.

[Mike Storey]

As an ex-Met Office employee of over 35 years I read this thread with some amusement. 

 

Don't mistake this for arrogance, but there always has been ignorance of meteorological sciences by the wider public, and this thread appears to prove (to me anyway) that the knowledge is not improving. 

 

As can be seen from this thread; everyone is an expert and they never remember when the forecast is correct, but never forget when the forecast is wrong.

 

You are missing the point. The issue is that forecasting seems to have become much more unreliable this year, when I think most would agree it had become much more accurate over the previous decade or so. It is easier to remember the few days it was right, than on how many days it has been wrong, lately.

 

Your comment about that, given your experience and expertise, would be rather more welcome.

[jonny777]

I can't comment on the reliability of the forecasts this year, or for the last few years. I don't watch or listen to them. 

 

I look at the synoptic charts online and make my own mind up. 

 

However, I have found this site to be useful as a general guide to temperature and rainfall risk - http://www.xcweather.co.uk/forecast/charente_maritime

 

(I took the liberty of entering your location into the forecast search box).

 

It is based on the GFS (Global Forecast System) output and not the Met Office, and it has to be remembered that the hourly data shown is raw computer output. Therefore the weather variables will be simply as calculated for that particular position at that particular time and on that particular computer run. The program is run at 6 hourly intervals. 

 

If one run has a 70% risk of 3.5mm of rain in a shower at 1200 tomorrow, and the next has a 0% risk for the same time, it does not necessarily mean the shower risk has vanished; just that the computer did not develop a shower in that precise area at that time. There may have been a heavy one developed 5 miles away, but as with showers/thunderstorms - one end of a town can have a deluge and the other end remain virtually dry. 

 

I would not base as much confidence on the forecast for 6 days ahead, as the one for tomorrow; although they are presented in the same tabulated way. You will learn to recognise when the general continuity of forecast conditions remains constant from run to run. 

[Mike Storey]

Thanks for that. I have narrowed down to my village and compared what the XC site says (although it does not seem to give rain probability forecasts, just mm expected) to weather.com, francemeteo and the BBC (this cannot be localised so much). For the next two hours here, they all say something very different. For the next twelve hours they more or less concur, except that two predict thunderstorms and two just light rain by the early hours.

 

I am therefore, none the wiser.

 

But thanks anyway.

[EddieB]

Yes, of course we all have some degree of bias based on our own experiences, but there does seem to be a consensus building that forecasting reliability has fallen recently (based on BBC weather app in my case).

 

If needed, I could make a lot of sense out of synoptic charts, to the point that I could give a fairly reasonable forecast for at least a couple of days hence. Most of that comes from a grounding in my first year at secondary school (when one of my classmates was selected in a TV audience to ask a question of a BBC weatherman - Jack Scott, I think - about showing such charts on TV for longer). Obviously a professional meteorologist should do far better than me.

 

However, I would expect computers to far exceed both our human capabilities, given the increased accumulation of historic weather data, and their ability to perform the myriad calculations needed to model the whole range of scenarios that might develop. That was where I thought we had come, until quite recently, so much data to draw from that short-range forecasting accuracy was asymptotic to perfection.

 

But that was before last week. I checked the next day forecast before booking rather expensive tickets for a family day out. Light cloud it said. The next morning dawned wet, the forecast had changed completely and we all got thoroughly soaked (financially and meteorologically).

 

Perhaps next time I’ll just listen to Marty Feldmann...

[4630]

 

Perhaps next time I’ll just listen to Marty Feldmann...

 

 

Or Jeremy Paxman;

 

https://www.youtube.com/watch?v=oQx3OSRj23Y

 

[Mike Storey]

Well, its gone 10pm here in downtown Charente Maritime, and every single weather station, at c. 4pm,  forecast cats and dogs by now, or several hours ago according to some. It is still as dry as an aborigine's armpit...... I could have concreted to my heart's content. This is holding up construction of the Belle End railway, much to the mirth of the unbelievers in these here parts.

 

Even my bl88ding seaweed has curled up in disgust. And as for my witch's nose....

[Edwin_m]

Is it just me or do the BBC now forecast about 95% of days to be "Breezy", with wind speed ranging from about 5min to about 30mph? 

[jonny777]

I do wonder if many people are expecting too much of the weather forecast, both broadcast and online. 

 

Presumably most people are interested in if it is going to rain or not during daylight hours; but they expect it to be precise for their position on the globe, something that even the most powerful computer is not able to do. The last few weeks have been a case in point. In a situation where thunderstorms are expected to develop, the forecast may say there is a 40% chance of getting one at any point, but the rest of the day will be warm and sunny. However, that does not mean that you could not have a deluge when you do get one and your garden be flooded. 

 

Thunderstorms are thunderstorms and always have been. They can be split into two general groups - isolated and trough. The development of the first (isolated, which by the way does not mean there will not be a lot of them, but simply means they will develop independently) is almost impossible to predict precisely, even in the morning of the day they develop. A forecast can give general areas where they are likely (over high ground, over urban area heat islands, on the forward edge of sea breezes etc.,) but others can develop simply from the outflow of winds from a storm which is already under way.

 

And, contrary to many long held beliefs, thunderstorms due not necessarily develop when the sun is highest in the sky that day. Sunday was a good example.

 

Most people living well inland had a decent day last Sunday. It was hot, dry and sunny; but the winds from the east and south coasts came together over W London at about 7pm and a large thunderstorm developed from nothing in about 45 minutes (not unusual for thunderstorms), and then moved slowly south and southwest towards Reading and Farnborough areas. If the forecast said there was only a 15% chance of a thundery shower, for which area was that incorrect?

 

For those who find it difficult to believe a tiny cloud can develop into a flood-worthy thunderstorm in less than an hour here are some archive rainfall radar images from April 3rd this year - 

 

This was the position at 1430 in the afternoon, a number of showers around but nothing too heavy apart from a couple of small ones over the South Wales coast. 

 

 

Would anyone dare to forecast from that image how much rain people living in Clevedon are going to get in the next 20 minutes?

 

 

 

 

 

This is the position just 5 minutes later at 1435, the South Wales ones are looking more threatening and have expanded, and some slightly heavier rain appears in the Weston Super Mare area. The red radar echoes are rainfall rates of between 10 and 20mm per hour. The browns are between 20 and 50mm per hour, a serious downpour for anyone). 

 

 

 

 

Another 5 minutes pass, and this next radar picture is at 1440. The pale brown/grey echoes are rates of 50-100mm an hour which is monsoon-like for this country. Serious flooding can occur if this continues for more than a few minutes. 

 

 

 

 

Finally the 1445 image (and remember that this is only a quarter of an hour after the first one I posted) - 

 

 

 

And yet, the entire area between Wells, Shepton Mallet and Bristol has remained dry. How does anyone forecast a percentage rainfall risk for Bristol and Somerset at 8am that day, and expect the residents to say "well done, that was 100% correct" by 3pm.  And hindsight is a wonderful addition to perceived forecasting ability.

 

 

The second version of thundery weather is ones that form on a trough, the most notable of which tend to move north from France late in the evening and overnight. These are less difficult to predict for the southern half of the UK because there tends to be vast amounts of lightning, the darkness of night makes the lightning visible up to between 80 and 100 miles away (yes, really), the quietness means thunder can be heard up to 20 miles away (yes, really), and most people are in bed and not caught out in the sudden downpours. 

 

I cannot comment on the accuracy (or lack of) forecasts recently, because as I said earlier in the thread I take no notice of them. 

 

However, on the point of if the weather has been rather abnormal this year, my answer would be 'yes' and the reason for that would be the Sudden Stratospheric Warming event which took place in early February and gave rise to "the beast from the east" as the MSM like to portray it; but that is a far more complex scientific issue and something we can do nothing about. 

 

 

 

Edited, to add that I should have credited Netweather.tv for the radar images which are at about 0.5km resolution and updated every 5 minutes; although I do have to subscribe on a yearly basis to get this level of detail. 

Edited June 6, 2018 by jonny777

[Mike Storey]

Thank you Jonny.

 

I see why you ignore the weather forecasts now. If the weather is indeed abnormal this year, then weather forecasting models will have trouble predicting accurately, because the interactions and probabilities their models have stored from previous patterns over many years, will mean little.

 

It is entirely reasonable to say that very localised variations are extremely difficult to predict, but what is most noticeable this year so far, is that entire areas are being poorly forecast.

 

If we are expecting "too much" of weather forecasts (and we probably are at the moment), then what are they for?