model discussion Model output discussion 11th July onwards

[Paul]

47 minutes ago, mushymanrob said:

NOAA anomaly chart accuracy..

Ok, ive gone back over the last 3 months of my own postings to discover how these charts have performed.

Ive used them 42 times, with the 8-14 day chart being used twice as many times as the 6-10.

They have been accurate or very accurate 70% of the time.
They have been inaccurate 30% of the time with totally wrong 7 out of 12 inaccurate occassions.

Very accurate = 22/42
accurate = 8/42
not very accurate = 5/42
wrong = 7/42

Most of the inaccurate readings came in late June/early July when for some reason there was a cluster of bad predictions.

They are very good at picking up the signals for a high pressure dominated pattern, and these signals are picked up at the 8-11 day range as a rule.

These findings support the findings @johnholmes did when he first researched these charts.

Its true that using the operational runs by themselves anything past around t144 becomes uncertain, but what these charts do do is extend that, so we can see the general pattern with some certainty, most of the time, out to t240, ten days. This is especially true in the 8-10 day timeframe when the two charts share those dates, and the 8-14 day chart follows on from the 6-10 day chart.

No model suite is perfect, these arent, but are, id suggest, the most accurate for the timeframe they cover and the stats suggest this to be true.

 

13 minutes ago, MATTWOLVES said:

Like I've said before Mushy,I don't dispute your independent research into those charts...my point as always been all data beyond that 10 day window becomes very uncertain,and there will always be that element of bias towards climatology the further ahead we go. And this as been the answer I have received from other forecasters in the profession! The noaa charts are know doubt a very useful tool,but certainly not the be all and end all of compiling a forecast.

As much as it's been great fun following this back and forth over the last few weeks (months, years?). Maybe it's time to do as Queen Elsa says and let it go?

 

[MATTWOLVES 3]

1 hour ago, Ed Stone said:

I don't really know what to make of this morning's GFS 06Z, Matt; inasmuch as it shows a slight -ive anomaly over Scandi and a weak +ive anomaly over Greenland, it seems to agree with the Anomalies? But, it also shows weakish ridging from the Azores, across southern areas, at times? Though, of course, each run can flip and flop, without necessarily wandering too far outwith the anomalies' envelope?

    

    

    

There are good reasons why Exeter make little effort at predicting day-to-day 'deviations' past ten days?

Yes Pete that repeated ridging gains more traction towards the South as we move into the 2nd week of August! I'm always skeptical beyond 10 days,but the signal is there for perhaps an improving picture at this stage,with less robust Heights over Greenland and lower Heights less evident across the Scandy location.. What are the chances of a re run of the last week popping up again towards mid August? Perhaps not as Hot,but summery potentials surely! 

Edited July 25, 2021 by MATTWOLVES

[mushymanrob]

58 minutes ago, MATTWOLVES said:

Like I've said before Mushy,I don't dispute your independent research into those charts...my point as always been all data beyond that 10 day window becomes very uncertain,and there will always be that element of bias towards climatology the further ahead we go. And this as been the answer I have received from other forecasters in the profession! The noaa charts are know doubt a very useful tool,but certainly not the be all and end all of compiling a forecast.

Id remind you that john was "in the profession", lol.. however, im not arguing about a timeframe beyond ten days as such, days 12-14 on the anomaly charts are always going to be ropey. Its the assertion that you cant look past t144 that i certainly dispute, i do, with success.

Ill ignore Pauls record of choice, its terrible, lol, and quote ABBA instead... "move on".

[johnholmes]

1 hour ago, mushymanrob said:

NOAA anomaly chart accuracy..

Ok, ive gone back over the last 3 months of my own postings to discover how these charts have performed.

Ive used them 42 times, with the 8-14 day chart being used twice as many times as the 6-10.

They have been accurate or very accurate 70% of the time.
They have been inaccurate 30% of the time with totally wrong 7 out of 12 inaccurate occassions.

Very accurate = 22/42
accurate = 8/42
not very accurate = 5/42
wrong = 7/42

Most of the inaccurate readings came in late June/early July when for some reason there was a cluster of bad predictions.

They are very good at picking up the signals for a high pressure dominated pattern, and these signals are picked up at the 8-11 day range as a rule.

These findings support the findings @johnholmes did when he first researched these charts.

Its true that using the operational runs by themselves anything past around t144 becomes uncertain, but what these charts do do is extend that, so we can see the general pattern with some certainty, most of the time, out to t240, ten days. This is especially true in the 8-10 day timeframe when the two charts share those dates, and the 8-14 day chart follows on from the 6-10 day chart.

No model suite is perfect, these arent, but are, id suggest, the most accurate for the timeframe they cover and the stats suggest this to be true.

Yes, the early summer mushy is quite often, along with early/mid autumn, their worst period for accuracy. Something to do with the hemisphere's change over between the winter and summer. Also early autumn is the height of the hurricane season and no model deals very well with them even at short range in some cases.  I've not found my data other than one sheet, about 70-75% accuracy for 500 mb charts at 6-10 and 60-65% for 10-14, so your recent checks are pretty similar. I just hope folk will try to use them as they are meant to be used, rather than the sniping that occurs at times. It is after all only the weather and it makes monkeys out of all of us at times.

[johnholmes]

1 minute ago, johnholmes said:

Yes, the early summer mushy is quite often, along with early/mid autumn, their worst period for accuracy. Something to do with the hemisphere's change over between the winter and summer. Also early autumn is the height of the hurricane season and no model deals very well with them even at short range in some cases.  I've not found my data other than one sheet, about 70-75% accuracy for 500 mb charts at 6-10 and 60-65% for 10-14, so your recent checks are pretty similar. I just hope folk will try to use them as they are meant to be used, rather than the sniping that occurs at times. It is after all only the weather and it makes monkeys out of all of us at times.

Sorry Paul I had not seen your post otherwise I would not have posted, delete it if you feel it should go. I agree with your comment.

[Frosty.]

There continues to be tentative signals from the GEFS 6z for an improving weather pattern from towards mid August, especially across southern parts of the u k with increasing Azores high / ridge influence, as I say, it’s tentative because there are still a number of unsettled ensemble members persisting in that extended period...I will be happier when the models reach mid august which will give a better idea of the direction of travel for the last few weeks of this meteorological summer and I’m hoping we see a spell similar to the past week which would at least mean there’s a great end to a predominantly decent summer! ⛈ ? ? 

 

Edited July 25, 2021 by Jon Snow

[Mike Poole]

Heading back in the right direction at T168 on the ICON, I feel now might be a good time for the models to start smelling the coffee !

I’m still seeing a week of unsettled weather, not too wet on most days excepting the middle of next week, and then a gradual improvement.  

[mb018538]

20 minutes ago, Jon Snow said:

There continues to be tentative signals from the GEFS 6z for an improving weather pattern from towards mid August, especially across southern parts of the u k with increasing Azores high / ridge influence, as I say, it’s tentative because there are still a number of unsettled ensemble members persisting in that extended period...I will be happier when the models reach mid august which will give a better idea of the direction of travel for the last few weeks of this meteorological summer and I’m hoping we see a spell similar to the past week which would at least mean there’s a great end to a predominantly decent summer! ⛈ ? ? 

 

I don't want to have a pop or anything - but I pointed out on Thursday that just cherry picking out a load of GFS ensemble members at T+384 and saying it looks like a pattern change doesn't really mean anything. Well - it's now Sunday (3 days later), and these charts you've posted are all at T+384 again, which kind of proves my point 
 

The morning ECM clusters aren't overly promising in the 10-15 day range, pretty much all low pressure dominated. 

 

The latest 500mb anoms (from Thursday) have the trough stuck here or very close by out to mid August:



Pattern changes can sneak up unannounced...but I'm not really expecting much joy for us until mid August at least. 

Edited July 25, 2021 by mb018538

[MATTWOLVES 3]

42 minutes ago, Mike Poole said:

Heading back in the right direction at T168 on the ICON, I feel now might be a good time for the models to start smelling the coffee !

I’m still seeing a week of unsettled weather, not too wet on most days excepting the middle of next week, and then a gradual improvement.  

I'm with you on this Mike my boy,what do you think of UKMO at day 6! Trough being pegged back to the NE or still being a pain in the butt..I'll post the day 7 as soon as its available..

[Frosty.]

48 minutes ago, mb018538 said:

I don't want to have a pop or anything - but I pointed out on Thursday that just cherry picking out a load of GFS ensemble members at T+384 and saying it looks like a pattern change

I used the word tentative, twice and tentative means without confidence / hesitant / uncertain etc etc etc... if you had read my post properly you would have noticed that minor detail.  

Edited July 25, 2021 by Jon Snow

[Mike Poole]

1 minute ago, MATTWOLVES said:

I'm with you on this Mike my boy,what do you think of UKMO at day 6! Trough being pegged back to the NE or still being a pain in the butt..I'll post the day 7 as soon as its available..

What I do know, is that it is a damn sight better than the GFS:
 

Yes, again the T168 on the UKMO will be interesting.  I’m still not seeing unsettled lasting that long into August as some are.  But that’s maybe partially location - if I’m honest, a push from the Azores would see the south favoured first.  

[Methuselah]

GFS 12Z operational run is still mostly within the anomalies' window (I think!) . . . It doesn't look too inclement either, to me:

    

It wouldn't take much though, would it!

 

[MATTWOLVES 3]

Day 7 UKMO hints perhaps...not to bad tbh.

[Sunny76]

3 hours ago, MATTWOLVES said:

Yes Pete that repeated ridging gains more traction towards the South as we move into the 2nd week of August! I'm always skeptical beyond 10 days,but the signal is there for perhaps an improving picture at this stage,with less robust Heights over Greenland and lower Heights less evident across the Scandy location.. What are the chances of a re run of the last week popping up again towards mid August? Perhaps not as Hot,but summery potentials surely! 

I think the second half of august will provide us with some very warm or hot weather for a change, with some more thunderstorms, but more widespread. 

[mb018538]

19 minutes ago, MATTWOLVES said:

Day 7 UKMO hints perhaps...not to bad tbh.

Be a bit of a shock to the folks in Scandi after the summer they’ve had! Back to reality!

[NApplewhite]

1 hour ago, Mike Poole said:

Heading back in the right direction at T168 on the ICON, I feel now might be a good time for the models to start smelling the coffee !

I’m still seeing a week of unsettled weather, not too wet on most days excepting the middle of next week, and then a gradual improvement.  

2 weeks of unsettled weather ahead from Monday onwards. Here is the chart for 8th August from GFS. The trough hangs around for 16 days. 

Edited July 25, 2021 by NApplewhite

[NApplewhite]

34 minutes ago, Sunny76 said:

I think the second half of august will provide us with some very warm or hot weather for a change, with some more thunderstorms, but more widespread. 

Some hints of slighly above average temps in late August 

Edited July 25, 2021 by NApplewhite
Add chart

[damianslaw]

1 hour ago, Ed Stone said:

GFS 12Z operational run is still mostly within the anomalies' window (I think!) . . . It doesn't look too inclement either, to me:

    

It wouldn't take much though, would it!

 

Mmm these charts don't suggest a quick return to settled. Azores high positioned and angled poorly, a westerly/north westerly flow most likely with trough forced down on the UK.

All eyes on the coming days how the jet behaves and whether it retracts further north in time to allow azores high ridging and a more typical westerly pattern, which at this time of year usually brings drier warmer conditions to the south at least.

[mb018538]

30 minutes ago, damianslaw said:

Mmm these charts don't suggest a quick return to settled. Azores high positioned and angled poorly, a westerly/north westerly flow most likely with trough forced down on the UK.

All eyes on the coming days how the jet behaves and whether it retracts further north in time to allow azores high ridging and a more typical westerly pattern, which at this time of year usually brings drier warmer conditions to the south at least.

Yeah - there will be some quieter interludes as the ukmo run shows tonight, but at the moment the void left behind will be eventually filled by another low in due course rather than high pressure. GEM sort of shows this, but in keeping with the slow theme, it takes a number of days for the next low to arrive, leaving a sunshine and showers regime to dominate i between. Not the worst.

[Singularity]

Just spotted that the CFSv2 AAM / GWO forecasts are updating again.

It's making a decent effort of showing the MJO Pacific transit scenario; AAM is driven back upward for a time (A), which is where the notion of the Azores High returning to our shores comes from. Sadly (for warm dry weather seekers), this is looking a pretty slow process, so we could be waiting until over 10 days from now for this to happen.

I'm not going to rule out a more vigorous MJO driving the next AAM orbit a bit quicker, though. I think that would result in the Azores High continuing its gradual eastward tendency that begins next weekend, rather than retreating back west as the 12z GFS goes with.

Longer-term, CFS shows gradually lowering AAM cycles, which ties in well with a gradually developing La Nina base state. Not sure just how quick or slow the descent will be - the slower it is, the more chance we have of further fine spells in the 2nd half of August. 2016 showed us what's possible if AAM is in the neutral range, despite a negative long-term trend.

 

[Mike Poole]

ECM T240 coming in to my way of thinking. Very happy with this chart at T240:

No big heights over Greenland, pattern moving for the ridge to edge closer to the UK.  One run, we need to see this backed up on the 0z runs…we will see…

[mb018538]

Crazy differences between the 00z and 12z at day 10 as you can see. 00z lots of northern blocking still around, 12z has it all gone with low pressure back over the pole. Much better prospects for the UK of the 12z is correct.

[Mike Poole]

12 minutes ago, mb018538 said:

Crazy differences between the 00z and 12z at day 10 as you can see. 00z lots of northern blocking still around, 12z has it all gone with low pressure back over the pole. Much better prospects for the UK of the 12z is correct.

Was what I was saying earlier re smelling the coffee, AAM still generally high and rebounding according to the only completely unreliable site I have access to from the CFS, but it is telling nonetheless.  

The Azores ridge should push on now, actually, I think faster than todays 12z runs.  And even if it is sluggish, the runs today show the trough lifting out quite quickly to our NE, so even if clear skies take a little longer to arrive, the washout scenario looks for the bin.

Edited July 25, 2021 by Mike Poole

[ANYWEATHER]

Looking at the long term trend  slim picking for Summer lovers  ,in fact if your camping out in northern UK especially Scotland Jack Frost could may make an early appearance as early as this coming weekend, although generally unsettled it looks as though most of the rainfall in the next ten will be showers rather than frontal rain...

[MATTWOLVES 3]

35 minutes ago, Mike Poole said:

Was what I was saying earlier re smelling the coffee, AAM still generally high and rebounding according to the only completely unreliable site I have access to from the CFS, but it is telling nonetheless.  

The Azores ridge should push on now, actually, I think faster than todays 12z runs.  And even if it is sluggish, the runs today show the trough lifting out quite quickly to our NE, so even if clear skies take a little longer to arrive, the washout scenario looks for the bin.

Still hints of an improvement for me,but perhaps more towards the 2nd week... The op was a little out of kilter with those pressure rises.. So at this stage more runs needed...where have I heard that one before