Model Output Discussion 00z 19/07/13

[phil nw.]

The last T72hrs fax from this morning shows the difficulty with the period from mid-week

 

That next low approaching from the SW with it.s attendant trailing front separating the hot continental air from the cooler Atlantic incursion.

The probability that the front will buckle and another shortwave will run up it towards us complicating a clean evolution for another 2-3 days.

These shallow features are a devil to forecast when the upper air(500hPa)pattern is so slack with little momentum.

I think eventually the highest uppers over the near continent will be progressively modified and the flow will go more from the west or south west.

[shedhead]

No beating round the bush, ECM is pretty poor compared to recent output from all models. Much more progressive with the Atlantic push

 

Very ordinary brief spell of warmth in the south (high twenties) and likely a breakdown of patchy rain.

Suppose with 16C close by you might squeeze 30C if it's sunny

Warmth all but gone by midday Friday.

Indeed CS, but I guess it looks even worse due to the fact it's predecessor was so extreme.  Let's get the overnight runs in before writing this off completely, but it is looking increasingly likely that only a few parts of the UK will see a brief hot spell next week, with the next 7-10 days appearing decidedly average overall.... at least from a temperature perspective.

Edited July 27, 2013 by shedhead

[Cloud 10]

It's what happens when you take charts at 168 hrs out as gospel.

 

No one taking 168 hrs charts as gospel CC,that only applies in winter

when a -15 screaming easterly is "nailed on". 

[Uncle_Barty]

It's what happens when you take charts at 168 hrs out as gospel.

 

No-one I saw was taking the runs as gospel - we all know they were up to 168h away and therefore subject to big change. The point that someone was making is that the solution had been showing for 3-4 runs before the change on the latest one, so I'm really not sure what point you are trying to make here

 

Interestingly, the GEM raises heights over Scandi which may open up the possibility of warmer air being advected off the Continent in time and serve to keep lows that bit further west. Dont dismiss that solution lightly, GEM did quite well with today, I think.

Edited July 27, 2013 by Uncle_Barty

[Cheshire Freeze]

No-one I saw was taking the runs as gospel - we all know they were up to 168h away and therefore subject to big change. The point that someone was making is that the solution had been showing for 3-4 runs before the change on the latest one, so I'm really not sure what point you are trying to make here

 

Interestingly, the GEM raises heights over Scandi which may open up the possibility of warmer air being advected off the Continent in time and serve to keep lows that bit further west. Dont dismiss that solution lightly, GEM did quite well with today, I think.

 

3 or 4 runs? It first showed on the ECM 12z last night and was largely gone by the 0z. Yes, the posts I read last night were ramping away about possible 40C heat and how it was likely to happen as it was gathering support. Sometimes it's not what is said in an obvious way but the subtleties within the post which suggest this to be the case.

 

People were being called trolls for less a couple of days back! I've not really got involved with it all, but the double standards of some people shocks me at times.

 

I suppose it all comes down to preference of weather types!

[Captain Shortwave]

3 or 4 runs? It first showed on the ECM 12z last night and was largely gone by the 0z. Yes, the posts I read last night were ramping away about possible 40C heat and how it was likely to happen as it was gathering support. Sometimes it's not what is said in an obvious way but the subtleties within the post which suggest this to be the case.

 

People were being called trolls for less a couple of days back! I've not really got involved with it all, but the double standards of some people shocks me at times.

 

I suppose it all comes down to preference of weather types!

Actually the ECM started showing it two days back, I commented on it at the time saying it was probably a rogue run. 

I think you are being a bit harsh here considering.

ECM, GFS, UKMO, GEM, JMA, NAVGEM all backed this evolution and seemed to agree for 2 consecutive runs. Developing consistency and cross model agreement.

That's really not hopecasting to be honest.

Edited July 27, 2013 by Captain shortwave

[Man With Beard]

GFS and ECM T144 ensemble means are very similar and summarise most output pretty well - suggesting rather more confidence in a hot end to the week with temps reaching 30C, maybe 32C if surface conditions are favourable.

[damianslaw]

Its a slack pressure set up for the foreseeable future but broadly southwesterly/westerly in origin, meaning staying relatively warm but rather unsettled with showery downpours interspersed with lengthy sunny dry periods. The SE closest to the heat over the near continent looks like staying preety warm over the coming days with a chance of proper heat spilling into the SE by next weekend - but only the SE,no return to widespread heatwave conditions.

 

I've been quiet in the forum these past few weeks, enjoying the summer weather conditions.

 

Will August be as good as July? - probability wise unlikely.

[Polar Maritime]

Its a slack pressure set up for the foreseeable future but broadly southwesterly/westerly in origin, meaning staying relatively warm but rather unsettled with showery downpours interspersed with lengthy sunny dry periods. The SE closest to the heat over the near continent looks like staying preety warm over the coming days with a chance of proper heat spilling into the SE by next weekend - but only the SE,no return to widespread heatwave conditions.

 

I've been quiet in the forum these past few weeks, enjoying the summer weather conditions.

 

Will August be as good as July? - probability wise unlikely.

 

Agree, and one of months along with September that are the hardest to forecast.

[forecaster]

The problem is that model run to run consistency does not correlate with accuracy. 

 

 

As everyone else has mentioned, the EC det has changed. Now there is only 4C difference between the ensemble mean and the deterministic, compared with 8C-10C two runs ago.

 

To clarify what I said earlier, there is no scientific correlation between the run-to-run consistency of a deterministic model and its accuracy. Jumpiness in models is a necessity for them. But human forecasters do not have to behave like models and jump all over the place. The great value of ensembles is that they can pick trends and allow human forecasters to steer a steady course without jumping all over the place and degrading the end users confidence in the forecast. 

 

[Weather Boy]

As everyone else has mentioned, the EC det has changed. Now there is only 4C difference between the ensemble mean and the deterministic, compared with 8C-10C two runs ago.

 

To clarify what I said earlier, there is no scientific correlation between the run-to-run consistency of a deterministic model and its accuracy. Jumpiness in models is a necessity for them. But human forecasters do not have to behave like models and jump all over the place. The great value of ensembles is that they can pick trends and allow human forecasters to steer a steady course without jumping all over the place and degrading the end users confidence in the forecast. 

 

I've always been told that consistency doesn't equal accuracy but never really understood it as I shall explain. The GFS runs four times per day and the other models les frequently. Imagine a scenario where the GFS jumps from a to b to c and then back to a again in one day. Ok that doesn't mean it's wrong but what is the run you are supposed to be judging? If a model is consistent at least you can tell its right or wrong, but if it says a different thing each run, how can you still have the same regard for it when you don't know what it's trying to say? Can you clarify the point please as it's a thing that really troubles me when model watching.

Secondly and arguably off topic, can you, as a professional, clarify why the METO were so badly wrong today? We had 29C as against a forecast 22-24C. Yet the models were showing a hot plume for the weekend for ages, leaving me at a loss as to why such low temperatures were forecast. Very unusually it was 1-0 to me over the METO this weekend....but do you know why such an obviously (in my eyes) wrong forecast was produced?

Grateful for your input in helping me learn thank you.

[Swave Snow]

I've always been told that consistency doesn't equal accuracy but never really understood it as I shall explain. The GFS runs four times per day and the other models les frequently. Imagine a scenario where the GFS jumps from a to b to c and then back to a again in one day. Ok that doesn't mean it's wrong but what is the run you are supposed to be judging? If a model is consistent at least you can tell its right or wrong, but if it says a different thing each run, how can you still have the same regard for it when you don't know what it's trying to say? Can you clarify the point please as it's a thing that really troubles me when model watching.Secondly and arguably off topic, can you, as a professional, clarify why the METO were so badly wrong today? We had 29C as against a forecast 22-24C. Yet the models were showing a hot plume for the weekend for ages, leaving me at a loss as to why such low temperatures were forecast. Very unusually it was 1-0 to me over the METO this weekend....but do you know why such an obviously (in my eyes) wrong forecast was produced?Grateful for your input in helping me learn thank you.

 

 

On the temperature front i would say that they expected the storm much earlier originally and as was the case the temps (for example in Hendon went from 28c to 23c in half an hour) went much lower, but after the heat of the day had built, if the storms had occurred as originally forecast then the temp i think would have been nearer the forecasted max.  As it was associated with a warm plume I guess we actually got the heat associated with that warm air mass.  

[Cecil Hogmanay Melchett]

^{UKMO again looks pretty poor for summer weather fans.GFS looks slightly better but the optimism of a few days ago from my perspective is slowly ebbing away.}

^{IObviously August is still 4 days away but one is starting to get the feeling August will be very different to July.}

[Gordon Webb]

^{UKMO again looks pretty poor for summer weather fans.GFS looks slightly better but the optimism of a few days ago from my perspective is slowly ebbing away.}

^{IObviously August is still 4 days away but one is starting to get the feeling August will be very different to July.}

 

could do with increasing the size of your writing , I could hardly read it

[shedhead]

^{UKMO again looks pretty poor for summer weather fans.GFS looks slightly better but the optimism of a few days ago from my perspective is slowly ebbing away.}

^{IObviously August is still 4 days away but one is starting to get the feeling August will be very different to July.}

Indeed HD...the reality of the overall situation looking forward is now starting to dawn I think.

 

UKMO looks a little disappointing regarding midweek plume potential, at least away from the E.

 

 

 

GFS offers the prospect of somewhat wider warmth and localised heat for England and Wales on Thu/Fri, but unfortunately much of Scotland and N Ireland largely miss out.

 

 

 

 

Looking farther ahead there is little if any sign of widespread cool, wet conditions, but a NW/SE split does look set to develop, with most places seeing pretty bog standard Aug weather....which to be fair should still be very pleasant and useable for much of the time.

 

The longer range 500mb anom charts certainly want to rev things up a bit in the N Atlantic....

 

Edited July 28, 2013 by shedhead

[Frosty.]

Hotting up again on the Gfs 00z later in the week and becoming very humid with sunny periods and scattered thunderstorms, even by next weekend when the hotter uppers have cleared east, the surface conditions will continue on the warm side with temperatures into the mid 70's F but for a few days later in the week we are looking at high 20's to low 30's celsius once again, however, the far west and north of the uk are probably going to have pulses of heavy and possibly thundery outbreaks of rain for most of the week ahead but with some sunnier gaps and becoming warm and humid. FI  shows warm anticyclonic conditions predominating across the southern half of the uk but rather more unsettled further north at times.

Edited July 28, 2013 by Frosty039

[Cecil Hogmanay Melchett]

Indeed HD...the reality of the overall situation looking forward is now starting to dawn I think.

 

UKMO looks a little disappointing regarding midweek plume potential, at least away from the E.

 

 

 

GFS offers the prospect of somewhat wider warmth and localised heat for England and Wales on Thu/Fri, but unfortunately much of Scotland and N Ireland largely miss out.

 

 

 

 

Looking farther ahead there is little if any sign of widespread cool, wet conditions, but a NW/SE split does look set to develop, with most places seeing pretty bog standard Aug weather....which to be fair should still be very pleasant and useable for much of the time.

 

The longer range 500mb anom charts certainly want to rev things up a bit in the N Atlantic....

 

ecm is ever so slightly better than ukmo more akin to gfs but IMo the ecm has been a bit eratic recently so not sure what to make of it,esp the 144h chart which again looks odd.

ECM does bring in some very warm uppers to the SE corner late this coming week but the main plume is heading for the Benelux countries.

[Captain Shortwave]

Don't quite know what to think of this, it's better than its previous output but completely different to all the other models

The ecm if verified would bring a decent hot spell to the south east in particular with Thursday-Sunday pushing 30C

@Happydays - It can show a direct plume hit for us at 6 hours away and it would still end up in the Benelux countries 

Edited July 28, 2013 by Captain shortwave

[mushymanrob]

3 or 4 runs? It first showed on the ECM 12z last night and was largely gone by the 0z. Yes, the posts I read last night were ramping away about possible 40C heat and how it was likely to happen as it was gathering support. Sometimes it's not what is said in an obvious way but the subtleties within the post which suggest this to be the case.

 

People were being called trolls for less a couple of days back! I've not really got involved with it all, but the double standards of some people shocks me at times.

 

I suppose it all comes down to preference of weather types!

 

ive seen / read nothing that wasnt model based, even if the run was an outlier if something is displayed on a run its elegible for discussion on here... true there was some ramping going on, over enthusiastic predictions/statements, but thats no different to what goes on in winter when something snowy is a possibility...

 

slight hints this morning that after next weeks possibly thundery few hours (2 days max) that there might be some sort of pressure rise. ok, the 00z gfs doesnt show it, but neither does it show deep lows over us, there is scope imho for pressure rise. the ecm's version is abit odd, but it does show pressure rise this morning.

 

the mjo is all over the place atm, the anomaly charts appear to be moving away from deep troughing close to us, citing it further north, this might indicate a future movement towards higher pressure esp to our south... more runs needed obviously and i might be wrong in my interpretation!

[Man With Beard]

I've always been told that consistency doesn't equal accuracy but never really understood it as I shall explain. The GFS runs four times per day and the other models les frequently. Imagine a scenario where the GFS jumps from a to b to c and then back to a again in one day. Ok that doesn't mean it's wrong but what is the run you are supposed to be judging? If a model is consistent at least you can tell its right or wrong, but if it says a different thing each run, how can you still have the same regard for it when you don't know what it's trying to say? Can you clarify the point please as it's a thing that really troubles me when model watching.Secondly and arguably off topic, can you, as a professional, clarify why the METO were so badly wrong today? We had 29C as against a forecast 22-24C. Yet the models were showing a hot plume for the weekend for ages, leaving me at a loss as to why such low temperatures were forecast. Very unusually it was 1-0 to me over the METO this weekend....but do you know why such an obviously (in my eyes) wrong forecast was produced?Grateful for your input in helping me learn thank you.

I don't claim to be a forecaster in the league of some on here, but I think the bbc/met made a basic mistake yesterday which was trying to make too precise a forecast based on a thundery incursion. It was clear that not getting the volatile situation absolutely spot on could lead to temps in high 20s again. Models this morning - all seems set in stone up to Thursday, Mon/Tues cooler changeable, drying and warming up Wed/Thurs. Inconsistencies continue thereafter, ECM has another go at a tropical storm like chart at T168. GFS ensemble mean chart consistent with yesterdays, showing hot weather on Friday begin to be pushed away on Saturday

[johnholmes]

The convection thead last evening was rather like the snow thread in winter, full of hype, a lot of nonsenical posts and a lot of misunderstanding of meteorology. The replacement of so many toys that came out of prams will be quite costly for some! Reading the actual weather reports on the BBC News site shows how many places had torrential rain, the electrical side was always of secondary importance in spite of the posts by a fair number of folk on Net Wx. Viewing the lightning strike accumulation also shows how far out some comments were in reality.

 

The 500mb anomaly charts continue to show differences so on a longer term, beyond the short term 24-72 hour forecasts, the models will continue to switch from one version to another. This will continue until the position of the main trough is decided, west of or almost over the UK. Less widespread heat for sure but don't be surprised by an occasional hot day. Also the odds are on it becoming less settled with eventually the Atlantic having more say than for some time. However posts writing off August are similar to those writing the summer off in June, having no scientific backing whatever.

 

from BBC News

A number of flood warnings and alerts are in place. There has been flooding in the High Street in Market Harborough, Leicestershire.

The average rainfall for the whole of July is 78mm (3in) but on Saturday evening, there was more than 60mm of rain in parts of the county.

There are also yellow warnings of heavy rain - meaning be aware - for the East of England, London and the South East, Strathclyde, south-west Scotland and Lothian and Borders.

The Environment Agency has three flood warnings, meaning flooding is expected, and more than 35 flood alerts in place in England, meaning flooding is possible.

The warnings relate to the River Spodden, at Whitworth, Lancashire; and Blackburn Brook and Charlton Brook, in South Yorkshire, and Hilton Brook at Sutton on the Hill and Uttoxeter Road, Hilton, Derbyshire.

There are 23 flood alerts in north-east England, eight in the Midlands and one in the Anglian region.

Edited July 28, 2013 by johnholmes

[shedhead]

Don't quite know what to think of this, it's better than its previous output but completely different to all the other models

The ecm if verified would bring a decent hot spell to the south east in particular with Thursday-Sunday pushing 30C

@Happydays - It can show a direct plume hit for us at 6 hours away and it would still end up in the Benelux countries 

Ditto CS, that is a weird looking evolution thrown up by the ECM to say the least. Think we need to view the postage stamps later, but if it gets decent support I might start getting some wood together for my ark...

Edited July 28, 2013 by shedhead

[shedhead]

The convection thead last evening was rather like the snow thread in winter, full of hype, a lot of nonsenical posts and a lot of misunderstanding of meteorology. The replacement of so many toys that came out of prams will be quite costly for some! Reading the actual weather reports on the BBC News site shows how many places had torrential rain, the electrical side was always of secondary importance in spite of the posts by a fair number of folk on Net Wx. Viewing the lightning strike accumulation also shows how far out some comments were in reality.

 

The 500mb anomaly charts continue to show differences so on a longer term, beyond the short term 24-72 hour forecasts, the models will continue to switch from one version to another. This will continue until the position of the main trough is decided, west of or almost over the UK. Less widespread heat for sure but don't be surprised by an occasional hot day. Also the odds are on it becoming less settled with eventually the Atlantic having more say than for some time. However posts writing off August are similar to those writing the summer off in June, having no scientific backing whatever.

 

Who has written off Aug JH?  I haven't seen one post in here to that effect, but have you seen it said elsewhere?

 

Edited July 28, 2013 by shedhead

[forecaster]

I've always been told that consistency doesn't equal accuracy but never really understood it as I shall explain. The GFS runs four times per day and the other models les frequently. Imagine a scenario where the GFS jumps from a to b to c and then back to a again in one day. Ok that doesn't mean it's wrong but what is the run you are supposed to be judging? If a model is consistent at least you can tell its right or wrong, but if it says a different thing each run, how can you still have the same regard for it when you don't know what it's trying to say? Can you clarify the point please as it's a thing that really troubles me when model watching.

Secondly and arguably off topic, can you, as a professional, clarify why the METO were so badly wrong today? We had 29C as against a forecast 22-24C. Yet the models were showing a hot plume for the weekend for ages, leaving me at a loss as to why such low temperatures were forecast. Very unusually it was 1-0 to me over the METO this weekend....but do you know why such an obviously (in my eyes) wrong forecast was produced?

Grateful for your input in helping me learn thank you.

 

 

Your question is a very good one but I don't consider myself an expert in this. There's a Swedish meteorologist called Anders Persson who also worked at ECMWF for a while. I will try to find his ecmwf guide, which I think it available to everyone, and in it he basically runs through some situations "this is agreeing with this, but this is jumpy, what do you do?". He freely admits that there is one or two situations where it's so rare that it's difficult to know how to proceed. 

 

"The GFS runs four times per day and the other models les frequently. Imagine a scenario where the GFS jumps from a to b to c and then back to a again in one day. Ok that doesn't mean it's wrong but what is the run you are supposed to be judging? If a model is consistent at least you can tell its right or wrong, but if it says a different thing each run, how can you still have the same regard for it when you don't know what it's trying to say? Can you clarify the point please as it's a thing that really troubles me when model watching."

 

I've never been involved in model verification so I don't know how they treat it, but I definitely see your point. I think the trick is that if the det. is jumping all over the place, you should probably not be latching onto any and every change and treating them as reliable. This is why, if the model jumps, the ensemble should be really helpful because it usually won't jump. Now the ensemble is not going to turn out to be 100% accurate, but it does provide you with the general flow. It also filters out features that are not predictable (which you will see appear and disappear and move around all over the place on the deterministic runs).

An example might be the current situation. The EC backed off on the heat for Friday, maybe in later runs it puts it back in, it's possible. But the ensemble kind of consistently gave a solid above average signal, without getting too carried away. You could write a decent long range forecast mentioning above average temperatures, then add the detail as it becomes reliably predictable.

The lack of a heatwave in the ensemble generally suggests that the heatwave was not predictable. This however relies upon the axiom that the ensemble is correctly capturing the uncertainty inherent in the base atmosphere. We really hope it does, but sometimes it doesn't (especially at short ranges). 

The ensemble mean gives you the general flow, the probabilities in an ensemble show you what was filtered out in the averaging. 

 

On a Sunday morning, I'm afraid this is too rambling, sorry. 

 

As for yesterday's forecast, I don't know enough about the situation to explain why it was wrong. I was out and about and also disappointed by the forecast. I mean, 29C in Heathrow. That was not expected. I copped it from the family. I told them I don't work for the MetO but that never works. Forecasters are all the same to them.

 

 

Page 8 in here: http://www.ecmwf.int/publications/newsletters/pdf/129.pdf

 

Well worth a read. Hopefully more illuminating than the garbled things I've written above. I shouldn't go on the internet before breakfast.

Edited July 28, 2013 by forecaster

[Captain Shortwave]

The convection thead last evening was rather like the snow thread in winter, full of hype, a lot of nonsenical posts and a lot of misunderstanding of meteorology. The replacement of so many toys that came out of prams will be quite costly for some! Reading the actual weather reports on the BBC News site shows how many places had torrential rain, the electrical side was always of secondary importance in spite of the posts by a fair number of folk on Net Wx. Viewing the lightning strike accumulation also shows how far out some comments were in reality.

 

The 500mb anomaly charts continue to show differences so on a longer term, beyond the short term 24-72 hour forecasts, the models will continue to switch from one version to another. This will continue until the position of the main trough is decided, west of or almost over the UK. Less widespread heat for sure but don't be surprised by an occasional hot day. Also the odds are on it becoming less settled with eventually the Atlantic having more say than for some time. However posts writing off August are similar to those writing the summer off in June, having no scientific backing whatever.

 

from BBC News

A number of flood warnings and alerts are in place. There has been flooding in the High Street in Market Harborough, Leicestershire.

The average rainfall for the whole of July is 78mm (3in) but on Saturday evening, there was more than 60mm of rain in parts of the county.

There are also yellow warnings of heavy rain - meaning be aware - for the East of England, London and the South East, Strathclyde, south-west Scotland and Lothian and Borders.

The Environment Agency has three flood warnings, meaning flooding is expected, and more than 35 flood alerts in place in England, meaning flooding is possible.

The warnings relate to the River Spodden, at Whitworth, Lancashire; and Blackburn Brook and Charlton Brook, in South Yorkshire, and Hilton Brook at Sutton on the Hill and Uttoxeter Road, Hilton, Derbyshire.

There are 23 flood alerts in north-east England, eight in the Midlands and one in the Anglian region.

Please note that the Anglia flood alert is not in any way related to last nights rain (tidal surge)

To be fair yes the forecast for Northern England was correct, but there are many people in Central/Southeastern areas, especially business owners for some of the main tourist attractions who must be spitting blood. There is a huge difference between a cloudy day with outbreaks of rain with temperatures around average, and glorious sunshine until the evening and temperatures nearly hitting 30C. Given the times we are in, that's a pretty massive error.

Edit - From the convective thread it turns out most of the forecast yesterday was based on the NAE high res model which seemed to go against pretty much most output, including other hi-resolution models which actually correctly predicted certain areas to remain mostly dry. It kind of goes to show when you ended doing tv forecast which are only based on one set of data (bbc use the NAE), if it is wildly wrong then that forecast is next to useless. Given  the constant underestimation of temperatures countrywide this week. I will only say that the NAE has had a complete shocker of a week.

Anyway rant over, lets move on and put a lid of stormgate round 60,078,903,340

Edited July 28, 2013 by Captain shortwave