Uk Met Office - Worse Or Better Than 10 Years Ago?

[Terminal Moraine]

It's easy to say the Met' Office forecasts are worse now than 10 years ago but without having kept detailed records of the accuracy of the forecasts it's little more than a stab in the dark.

I can remember occasions from 10, 20, 30 or 40 years ago when 'interesting' events were forecast only to be nothing out of the ordinary when the time came or to miss us altogether . The difference between then and now is, particularly for the inhabitants of forums like this, that we all have huge amounts of information at our fingertips ( as much as the Met' Office had 30 years ago, if not more ) and we are able to analyse and dissect to our hearts content exactly where every storm, snowfall or shower is going and what it's doing, so often we're aware of a forecast going wrong before it's actually happened. By the time it's become a non-event and the warnings have been removed we've all had several hours of wallowing in disappointment and the recriminations have begun in earnest.

We need to remember that the Met' Office are forecasting for the general public, not for us alone, and the majority of the public aren't hanging on every nuance of a forecast, they're only interested in a general overview of what may or may not happen.

If the Met' Office decide there's a 50% chance of severe storms, or whatever, affecting a certain area they issue an advisory and the forecasters begin to mention the possibility of 'weather to watch out for' in the forecasts. The general public see/ hear the forecast and either take no notice of it at all or bear in mind it might happen. By this time we've already been aware of the possibility of severe weather occurring for a couple of days and the fact the Met' Office issues an advisory makes it, in the eyes of many, a matter of certainty that it will occur, it's just a matter of when and how much.

The Met' Office will maintain the advisory until such time as all threat has passed whereas we've been aware that the threat has passed some consderable time before and are already annoyed that the advisories are still in place. The general public meanwhile go abut their lives as normal, probably pleased that whatever it was ' missed us' but without a deep seated sense of injustice that the forecast was wrong.

The time for major criticism of the Met' Office from the public comes when something severe happens that wasn't forecast and it's this driver which leads the Met' to err on the side of caution and issue advisories well before the percentage chance of any particular event occurring has reached anything more than a possibility.

In summary then I don't think the Met' Office forecasts are worse than they were 10 years ago, there will always be forecasts which go wrong, sometimes badly wrong. The major difference now compared to 10 years or so ago is that we have access to a huge amount of information and we're able to examine each forecast in minute detail as it unfolds and criticise it accordingly.

Edited August 23, 2011 by Terminal Moraine

[Thundery wintry showers]

My opinion has, for a while, been that most of the BBC national forecasts have gone downhill in recent years, though in the meantime the regional North East ones have improved. I find the biggest issues to be time allocation and trying to reassure the public when the next "good" weather will happen (both of which started to creep in at the BBC around 2002, but were rife in ITV forecasts well before that) and the "touring of the country at 8am" (which was introduced in May 2005 along with the current graphics package). However, this is down to the BBC rather than the Met Office- indeed there was an article from 2005 where the MetO specifically condemned the use of value judgements on the weather.

I haven't noticed any recent trends in the Met Office forecasts but I find John Holmes's insights interesting and consistent with personal experience. Perhaps the closing of the regional weather centres may also have had an impact as it potentially reduces the amount of input from humans with a strong understanding of a region's local climate, as I do find their short-range forecasts for south-west England to be more accurate than those for the north-east, for example (while the regional discrepancy vanishes once we get to the 3-7 day range). Most of it, though, is probably much as Terminal Moraine said above.

[The watcher]

Why pick on the met office, half the forecasters on here also got it wrong did they not?

I look at it as a close but no cigar situation. The storms were there, they weren't far away, but they just didn't make it far enough North.

No forecaster will ever get 100%...such is weather, and that's what makes it interesting.

Forecasting is not as simple as most of you think, it's not a 1 man job. Human input, models, etc are pieced together, minor forecasters glaze the specifics and then one man takes light of these to create what he thinks is as accurate as possible a prediction. When push comes to shove, I'm positive that most forecasts for certain areas can sometimes be 50-100 miles out within 12 hours around 50% of the time. If we are so concerned about accuracy, try making your own forecasts 12 hours out with the wonderful information we have at hand here at Netweather and beyond.

I believe the problem in recent forecasts has been too much effort put into detail. It would have sufficed to say to the public or indeed us that "there is a risk some heavy thundery rain currently in NW France may impede Southern areas of the UK overnight".

[Geordiesnow]

A Major FAIL, no words or excuses for technology back in the day or even now would explain why the forecast was so wrong. My Rabbit could have forecast what was coming, trouble is I don't have a rabbit!!!

To be fair, yesterday's prediction was a very bold one, its not like the forecasters to use the term "thunderstorms" so often but in that situation, the potential was there for thunderstorms to be embedded with the rain band. It looks like the shallow low that was bringing that particular weather was further East than what the models predicted 24-48 hours ago but thats the nature of shallow lows.

Too me its not like the main theme of the forecasts are wrong, e.g last night, a band of rain was coming up from the continent, that did occur but the thunderstorms didnt it would appear but if you are telling a normal person who lives down the South was the forecast right for this morning, they would probably say yes because "its raining".

In some situations, forecasting in 48 hours in advance can be tricky as this situation with the shallow low has prooven. The low to our West is more simplier in that in general the forecast for the next few days will be one that it will be unsettled for some with rain or showers at times, pinpointing where the PPN will be is more difficult but I tend to find they are more or less accurate.

[cyclonic happiness]

Most every rain event they've forcast over the last 6 months has fizzled out before it hits our region, even 6 hours before the event?

It does seem like forecasts 10-20 years ago, were better and more accurate than today.

No doubt it was because we had an excellent met head in the shape of John Holmes to help keep their botties in line

The modern graphic are ok, but the precipitation is very strangely represented in my opinion, rather than moveing smoothly across the map, it slowly oozes from one place only to reappear slightly further on in a different area of the map.

This makes it look as though it has jumped an area, and I still cannot really understand why it does that?

Like some folk say on here, it's probably 'rose tinted' specs, but I find I have to constantly check forums and wesites to find out any correct forecasts and they are ALWAYS a damn site better than the stuff the beeb churns out.

And they really should do something with their severe weather warnings, last year we had a really big (for the region) snowfall, which the beeb didn't even register until the snow had been falling for an hour here.

The forums on the other hand, had been following the evnt for several hours and had given the readers here a good few hours notice of the severe conditions.

Now one has to ask one's self, why a big organisation like the met office in this day and age, cannot organise a change to their pre-recorded or pre scripted weather forecasts, to accommodate such an obvious event when the likes of us can?

[Weather-history]

I know its daft but why dont they go back to the 70s and 80s of forcasting.. they were dead on with nearly every weather event that happened back then (except the storm of 87)

That with all due respect is selective memory. They were caught out back then as well especially when it came to snowfalls and thaws. They were caught out in December 1981 at least once with unexpected snowfalls and failed thaws. I can recall a failed thaw back in early January 1986. I can recall snow being predicted and it never happened.

[Radiating Dendrite]

Most every rain event they've forcast over the last 6 months has fizzled out before it hits our region, even 6 hours before the event?

It does seem like forecasts 10-20 years ago, were better and more accurate than today.

No doubt it was because we had an excellent met head in the shape of John Holmes to help keep their botties in line

The modern graphic are ok, but the precipitation is very strangely represented in my opinion, rather than moveing smoothly across the map, it slowly oozes from one place only to reappear slightly further on in a different area of the map.

This makes it look as though it has jumped an area, and I still cannot really understand why it does that?

Like some folk say on here, it's probably 'rose tinted' specs, but I find I have to constantly check forums and wesites to find out any correct forecasts and they are ALWAYS a damn site better than the stuff the beeb churns out.

And they really should do something with their severe weather warnings, last year we had a really big (for the region) snowfall, which the beeb didn't even register until the snow had been falling for an hour here.

The forums on the other hand, had been following the evnt for several hours and had given the readers here a good few hours notice of the severe conditions.

Now one has to ask one's self, why a big organisation like the met office in this day and age, cannot organise a change to their pre-recorded or pre scripted weather forecasts, to accommodate such an obvious event when the likes of us can?

I think the graphics are rubbish! The precipitation graphic is useless as you never really know when the rain will start or end.

The original radar graphics were much better and the symbols!

It was supposed to be dry and 23c here in London today. It is currently cool and raining, thanks Met!

[Methuselah]

Have to say I agree with John Holmes about this. Anyone who thinks the forecasts 30 or 40 years ago were better than the current ones is definitely wearing rose tinted glasses.

Perhaps before having a go at the Met Office about its forecast for the last 24 hours or so you should click on the Alert bar on the Netweather home page and read what it says.

Absolutely agree, PeteG; today's forecasts are far more accurate than those of yesteryear. The trouble, I think, lies more with the issuing of warnings than with the actual forecasts themselves. Once a warning goes out, us lot and the media conspire to ramp things a tad out of proportion. And, let's face it, in the aftermath of a non-event, shooting the messenger is much more comfortable than an egg-splattered face??

Mea culpa!

A Major FAIL, no words or excuses for technology back in the day or even now would explain why the forecast was so wrong. My Rabbit could have forecast what was coming, trouble is I don't have a rabbit!!!

I'm looking forward to seeing your rabbit's-eye view of the coming winter...Bugs Bunny Predicts?

[Thundery wintry showers]

Re. a couple of recent posts, the graphics change was down to the BBC rather than the Met Office. They chose to adopt Metra's graphics package back in May 2005 (Metra being a weather company based in New Zealand), but appear to use only a small subset of its capabilities.

http://www.metraweather.com/

Regarding the original question, I'm pretty sure that the Met Office's forecasts at 4+ days have improved over the last 10 years, as medium-range forecast technology has advanced considerably in that time and the Met Office's own model, after previously being quite dodgy at T+144, has moved into a close second place behind the ECMWF at both T+120 and T+144 over the last few years. For example, I recall numerous examples of BBC forecasts in the 1990s that suggested that the UKMO model had overdone northerlies at as little as 3 days' notice (e.g. I remember at least two from March 1997 alone)- these days the "downgrades" usually start at 5 days' range at minimum.

I haven't monitored their 1-3 day forecasts closely over the last 10 years, but the "erring on the side of caution" issue re. severe weather warnings has existed ever since I started observing weather forecasts back in the early 1990s. My suspicion is that some areas will have improved (due to improved model outputs) and others got worse (due to reduced human input). The FAX charts retain human input, so for those with a good understanding of synoptic charts, it is always worth supplementing the Met Office's warnings with interpretation of their FAX charts which give a strong indication of why the warnings are being issued.

Our media seems to be becoming more and more critical of weather forecasts' accuracy as time goes on (except when they are sensationalist and make good headlines and aren't made by the MetO- I'm looking at you, PWS).

[Methuselah]

And the sad irony in all that, Ian, is that most (if not all) of these alternative forecasts are next to useless...

[Radiating Dendrite]

What are peoples views on their nowcasting?

I think this can be a struggle for them. I remember last winter when they were forecasting snow for the NE (late Dec) and yet it was raining. I just wish they would be a bit quicker in changing the forecast when it has/is going wrong.

Also, an apology when they do screw up would be nice. A great example would be the rain from last week, they just covered their a*** by saying that it "was too localised"......rubbish. If it is too localised, why bother having models with a resolution of 5km (as JH mentioned previously I believe).

Edited August 24, 2011 by Radiating Dendrite

[chionomaniac]

I suspect that the accuracy of the Met office forecasts has improved over the last decade, but so has the technology and availability of computer generated forecasts to the masses, that allows a scrutiny that previously wasn't available prior to the internet age. To compare to football or cricket - do we think that referees or umpires are making more mistakes than years ago or is it just that technology has advanced so much that we are able to pinpoint any error from every angle available. Of course that is the case!

Having said that, the forecast plume that didn't materialise was one of those events that is notoriously difficult to forecast and there was considerable model divergence. What to me was of more concern was the failure to put out adequate warnings for the flooding in Bournemouth last week when there was far greater model consensus. This was compounded by not reacting to the situation as it evolved, and replying with tinpot excuses about the flooding not being widespread enough, whilst we were watching images of police cars afloat! The reasons for not reacting need to be addressed from the top of the organisation as it was obvious to any novice looking at the radar that a lot of rain was falling and there was a lot more to come and flooding would likely be a real probability.

[SteveB]

I find myself watching the forecasts & thinking it's not going to happen, especially when they go out to three days.

Take for instance a forecast you see on Monday evening which shows the weather for Wednesday. I have noticed if there is rain in the forecast for Wednesday, it will almost certainly be different come Wednesday, whether its lighter or heavier or none at all it always seem wrong.

I can't remember if the forecast were better 10yrs ago, but I am sure they couldn't have been worse.

[johnholmes]

the latest outputs from GFS and NAE (MetO Fine Mesh model) shows differences.

GFS 06z had 5mm north of me and 13mm south of me, too course to get the exact spot location.

NAE 12z just out, finer mesh with location practically over me has 33mm, 15 east, 26 north and south with 20mm west.

Be interesting to see what 12z GFS shows and also Net Wx Extra and NMM. Even more interesting to see what actual rainfall I get by 18z tomorrow. (all totals are to 18z quoted above)

updating on the 12z Extra - it shows a total of 6mm for here but 21mm on the spot location east of me.

NMM at 06z gave 15mm for here, 6mm west, 12mm north, 43mm east and 27mm south.

In unstable type bands of rain getting the total correct for a spot location will always be prone to error.

I'll put in the NMM value when it updates with 12z data.

here is the NMM chart for 12z giving the total out to 18z Thursday.

my spot location has 23mm but over 30 east and south of here.

Edited August 24, 2011 by johnholmes

[Anti Boring Weather]

I agree that the met office and bbc get forecasts wrong at short notice, many times for my area , but the models GFS, UKMO and ECM are just as bad so we can't soley blame the met office all the time.

[The PIT]

I think they have improved in some areas but others not. I feel the general forecast has improved.

Weather warnings tend to be over the top and are very slow to be adjusted. The reply I got from the met office for the storm that never was was that the person who updates the warnings had gone home for the night!!! However that didn't explain why they were putting warnings up when clearly nothing was happening anyway.

They're still very poor in predicting rainfall amounts which of course has the knock on effect for snow depths. New tech may help but NMM is totally useless and it's been wrong every time I used it.

They are pretty good in predicting wind speeds probably the strong area.

Storms are pretty tricky to predict so it's going to be a few years before that improves.

One problem they seem to suffer from is when they make a forecast and it isn't happening they still seem to stick to the forecast just delaying matters slightly before finally giving up about 12 hours later.

They've now got a large amount of amateurs on wow and although some of the data will be rubbish it should give a general indication how the forecast is going and how fast things are developing or not depending on the case. I wonder they intend to use such data in the future.

[SteveB]

I agree that the met office and bbc get forecasts wrong at short notice, many times for my area , but the models GFS, UKMO and ECM are just as bad so we can't soley blame the met office all the time.

I agree that the met office and bbc get forecasts wrong at short notice, many times for my area , but the models GFS, UKMO and ECM are just as bad so we can't soley blame the met office all the time.

But I think thats one of the problems, the meto etc rely too much on the models & maybe lack that human touch the forecasters of old had & needed.

Edited August 24, 2011 by SteveB

[Cheese Rice]

the latest outputs from GFS and NAE (MetO Fine Mesh model) shows differences.

GFS 06z had 5mm north of me and 13mm south of me, too course to get the exact spot location.

NAE 12z just out, finer mesh with location practically over me has 33mm, 15 east, 26 north and south with 20mm west.

Be interesting to see what 12z GFS shows and also Net Wx Extra and NMM. Even more interesting to see what actual rainfall I get by 18z tomorrow. (all totals are to 18z quoted above)

updating on the 12z Extra - it shows a total of 6mm for here but 21mm on the spot location east of me.

NMM at 06z gave 15mm for here, 6mm west, 12mm north, 43mm east and 27mm south.

In unstable type bands of rain getting the total correct for a spot location will always be prone to error.

I'll put in the NMM value when it updates with 12z data.

here is the NMM chart for 12z giving the total out to 18z Thursday.

my spot location has 23mm but over 30 east and south of here.

Its no wonder they have problems when the NAE and NMM both place todays PPN in the same position yet it is further north and west...

Thus making the Meto warning invalid as it doesn't cover the appropriate areas.

Now while I can't say I'm fussed that the warnings haven't been updated, I can guarantee if it was snow there would be an onslaught of peeved of members who missed out on the snow. Bewildered that the warnings haven't been updated and the BBC graphics haven't changed.

This is my issue with the Met. It appears no-one is in the office from 4pm onwards as there have been too many occasions to count when the weather warnings don't correspond with the actual weather.

An over reliance on computers when looking out the window would suffice.

Edited August 25, 2011 by Cheese Rice

[johnholmes]

an update re my post No 40

rainfall so far is 10.4mm, to that add 20-25% as the rain guage under reads consistently by that amount when checked with 2 other sites close by.

[matty8]

Maybe if they broaden the warning area's out stating ' Within the highlighted area, there is a risk of snow/rain/storms on Tuesday', much like the US do with their warnings, and then once a warning has been issued continue to modify it as situations change. Prime example would be today, London and East Anglia saw torrential rain this morning - did they put a warning out? NO! Yet they put a warning out for heavy rain that never came on Tuesday and to a certain extent lied to the media afterwards stating they were never wrong and London did get rain (about 3mm in truth).

Overall I don't sympathise as their forcasters are on mega wages and should have understood this by now. The statement in this thread about short term forecast quality not being good enough is spot on, there needs to be more brain power not computer power behind these <48hour forecasts. We deserve better!

[johnholmes]

http://www.metoffice.gov.uk/about-us/jobs/current-vacancies/002165

http://www.prospects...gist_salary.htm

is the advert the union for Met forecasters where they give a factual idea rather than the fanciful comment in the post above, the first link is the Met O advert curently running for a forecast position

Edited August 25, 2011 by johnholmes

[Thundery wintry showers]

Good links John, it's fair to say that meteorology is not a profession that people enter in order to get "mega wages".

Going by experiences of doing forecasting of my own here on N-W, it often strikes me that there is so much model output-related information out there that it is easy to get overwhelmed, particularly when different models show different trends and a forecaster has to determine the most likely outcome from the various options. I imagine that this will be even more true at the Met Office HQ where they have access to many additional outputs that we don't have.

I think there are, indeed, significant downsides to the "slow-updating" alerts system (I recall one re. a gale on 11 November last year for instance) but there is also the problem that if warnings are updated more frequently, no doubt the media and public alike will come out with things like "see, you didn't get it right until it was already happening and by the time it did, it was too late as I'd already planned my trip up into the affected area!". As long as forecast models continue to throw up issues (such as the NMM/NAE getting the positioning of rain wrong at 12 hours out) forecasters will always, to a certain extent, be damned if they do and damned if they don't.

In particular I remember, regarding the infamous Norfolk "dry corridor" on 30 November-2 December 2010, that the NMM and NAE and GFS were consistently forecasting that snow showers would move down into the Norwich area in 12-18 hours' time, and they never did. The general patterns are usually accurately modelled but the specific locations of severe weather, unfortunately, often aren't.

[J10]

Regular posters on here will know that I have defended the BBC with their new graphics and forecasts. I also think that forecasts have become more accurate over the past 10 years. However I have become slightly disillusioned with the forecasts this summer. Many times at 24 hes notice they have got frontal positions wrong, which has meant it staying dry and bright here when cloud and rain has edged across England. Conversely there was a period when it was always forecast to be sunny in 2 days time, also for it to change the following day. However I don't think the weather type this year has been that easy to get spot on. Perhaps though the point above about the graphics indicating a level of accuracy above that of the models is a valid issue

One bug bear I d have though is the SE of England bias, either in terms of future days, and even reporting of that days weather. Many times this summer, there has been a west east split and the weather here has been considerably different to that further east. So turning on the 6 0 clock news, i have often seen will the weather get better tomorrow, when we have had a nice day, and we would be happy for the next day to stay the same, or them commenting on a heatwave in SE parts when it has been dull and wet here. Maybe I have become more sensitive to it, but this does appear to be getter worse.

I would also have more respect for forecasters if they admit they (or even the models) got it wrong, but this seldom happens, and there is a lack of now casting, when it is clear that the forecast is going wrong, e.g. by looking at the radar.

[jamesthemonkeh]

My perception is that the accuracy of their forecasts went downhill when they moved from Bracknell to Exeter. Which I am sure was done for political reasons, thanks to New Labour. It surely makes sense to have the national meteorological office near where most of the population is?

Most of my friends rely on me for weather forecasts as they do not trust the Met Office or the BBC, and are not inclined to spend the effort researching the weather like I am, for it is a passion of mine, and for them it is just part of their life that they have to deal with.

I don't have a degree in meteorology, nor can I afford access to the best models, but I do what I can from what is freely available and blog about it, for my local area. I do get a lot of positive feedback from friends who appreciate it, and most of them do say they would rather rely on me than the Met Office/BBC.

I personally think that the future is kind of like David Cameron's Big Society idea, that individuals will give weather forecasts for the area that they are in, and use the media available to them, such as blogs to do so, as opposed to a central agency.

I only do my forecasts for fun, I like the challenge. I don't always get it right and I do admit when I have been wrong.

[ZONE 51]

Some interesting views on here.

If one model laid down the output to be the exact weather to happen by the hour its stating then we would never get it wrong. the point is the models take us for a ride and at times a bumpy ride, if 3 different models place precipitation in 3 zones 50 to 100miles apart(example)then why is this? is it why we have different models? they all take the data from the atmosphere but place a different outcome. so some setups get harder to predict. so i then understand why things go wrong.