swilliam
-
Posts
498 -
Joined
-
Last visited
Content Type
Forums
Blogs
Gallery
Events
Learn About Weather and Meteorology
Community guides
Posts posted by swilliam
-
-
30 minutes ago, nick sussex said:
Of course the ensembles can be helpful when they're not plucking a solution that looks totally against the background signals and theres a decent clustering around an outcome but still they're prone to some wild swings. Personally I'd shred all GFS operational runs past T240hrs and use them as cat litter! With 4 runs a day it gets lauded sometimes for picking up on the right solution when in fact all its done is throw 4 different solutions out and the odds eventually favour that it will be right eventually!
Anyway this counts now as two GFS moans from me so I'd better move on!lol
I agree with Mucka on this one. I do not really pay much attention to the Op (count it as one of the ensembles). But looking at the 10-15 day ensembles it is often apparent that there are some tendencies appearing. I am not a fan of the mean at this range as this just normally shows the climatological average - i.e zonal westerlies. However when it shows something different to this then it is useful. In fact this is the case just now - below is the mean at +384 showing a pronounced ridge in the vicinity of the BI. What this means to me is that there is likelihood of significant height rises in our area at this range. Now as this an average the exact position is completely undefined. Looking through the members there are UK highs, Scandi ridges and height rises to the NW. There are are of course zonal options also toning down this signal. The most common options over the last few runs (of those showing height rises) are the UK high and the Scandi ridges rather than the Greenland high though. But any of these are possible. The main point is if these options were not being shown then we would know that nothing is likely in the next 15 days . So this is the main point in my view. If no meridional solutions are being shown then more zonal is highly likely. If we have some more blocked options then this is much more possible and we can start looking how this may develop and where it may or may not set up.
It seems to me if we do not use these then all we have is the Meto 30 dayer and some info from IF about the other models (Glosea EC32 etc). So I more I am more than happy to look at these as it is the best we have (not the OP in isolation though - comments on this are pointless).
-
3
-
-
29 minutes ago, Catacol said:
Hehe - GP vs Glosea5. A great way to start the New Year. Here's hoping human interpretation outplays the computer says no....
HNY one and all.
Yes - very interesting - numerical modeling and picking the average or most common solution or using teleconnections - I must say I am pretty skeptical about both at this range. If either predict something other than zonal and mild (as GP has done ) and they turn out to be mostly correct I will be impressed.
-
Still plenty of cold options on EC ensemble with the OP near the top
-
2
-
-
39 minutes ago, THE CHOSEN ONE said:
Certainly still an option & well within the realms of possibility -
The slider low at 144 would ideally need more trough disruption & to seperate into 2 systems moving apart - this always allows the CAA over the top an easier pathway west...
is it out there yet?
look & PTB 6 & 10!!!!!!
look at PTB 10 @138 then UKMO 144
s
Yes it depends what folks want. The Det gave a snow fest for many but never brought the real cold near us. The 2 perturbations S mentions get the cold pool in but do not deliver much snow (based on the current run - I know the oft touted saying here get the cold in the snow will come - but this is not always true - e.g. Feb 86). In either event though it looks like the disrupted nature of the NH profile will allow opportunities further down the line.
Concerning means and other LRF models (beyond D10) I think generally they are not worth the paper they are written on in terms of the weather wee will experience in our little locale. This especially true when the set up is volatile as now. People hanging on to the Meto 30 day outlook should really need to relax. This is only going to represent the largest cluster(s) and as the sum of these is probably less than 50% so is more likely to be wrong as right. If it says there is no sign or low probability of a prolonged cold spell then that is exactly the case - based on what is before them. However this does not mean a cold spell will not happen - just that it is not shown in the clusters.
Finally someone asked why was the forecast model output was not the same as the analogue composite shown for the forecast MJO state for previous occasions. If LRF was that simple we would not all be on this forum!
-
3
-
-
5 minutes ago, Hotspur61 said:
Is the 18z a trend away from a much colder set up??. only time will tell but I don't think we will know till Weds at the earliest as the models get a grip on how strong the High is to the NE and how strong the Atlantic train is. I find it fascinating times ahead watching how the models handle things and it makes a pleasant change from the monotonous weather the UK has endured( particulary the poor souls in the NW) over the last month.
Unless it goes pear shaped then I do not think this will be resolved until about T96 which by my reckoning will be about Friday
-
4 minutes ago, summerplume18 said:
When do the ECM ensembles come out?
Short ensembles out - not much support for the OP again
-
2
-
-
For me whether we get anything interesting ( from a cold perspective) depends on how far west the block is (i.e. out over and West of Norway). This has two benefits; it stops the cold pool lifting out northwards (round the end of the block) before it gets to us and it means the small lows to the south will head SE and not NE or N. A couple of examples where the block has spread or extended
You can see from the ens spread this is where the uncertainty is
-
1
-
-
1 minute ago, Allseasons-si said:
Re the ensembles
just like to ask because when viewing them they always seem to scatter after days 5-6(uncertainty) when a cold spell is shown on the models,when i looked at them a week ago they was showing a straight line through them with no cold outliers(consistency)
Mainly because when they are showing mild mush we do not scrutinise them - we just look and say no cold. When cold is showing they are examined to death and large spread is apparent. When it mild we do not care much if is slightly mild, quite mild or very mild.
-
2
-
-
10 minutes ago, Aaki Khan said:
Wasnt It The GEM Which Picked Out This Trend Few Days Back?
Well it was shown by several GEFS members for several runs before that in F1 (adn derided on here by some). Then the ECM showed it once and then went away from it, then the GEM showed it for a couple of runs (and it was derided for that as being cannon fodder). Now the ECM is showing it but not really in the ensemble and the UKMO is showing the correct set up for it. Meanwhile the GFS is showing the same general pattern but not the cold outcome for us. Currently GEFS are all showing roughly the same pattern but with a wide range of outcomes most of which do not lead to cold here. (some do though). So no model was right or wrong. This is due to the volatility of the this kind of set up and small changes have large effects. What it shows is that no runs should be ignored or binned as we were told to do so several days ago. Each run should be given the same credibility and assessed in terms of the overall developments. As I said yesterday following means for this kind of development is pretty useless until it is almost certain.
Just to add re the post from Jason M the most likely outcome is that we will miss out but what the current outputs show us is that it is possible. Likely we will not know whether we hit the jackpot until about +96 Hrs but even then it can go wrong as we have seen previously. This is because a lot of (small) things have to fall in place. The most likely outcome I would say is that we do not get the trough disruption and trigger low which would lead to the cold lifting out into Scandinavia (as if they need it). This will not be known until much nearer the time.
-
3
-
-
15 minutes ago, Jason M said:
By day 10 most GEFS are 'near misses'
Still, if you had offered me the charts as they stand tonight a couple of days okay i'd have banked them every time.
2 and 17 appear to be the best but all are interesting if not delivering cold for us.
-
Just now, bluearmy said:
This run seems to have stalled. Not sure where it goes by day 10. The renewed jet could head ne or re enforce an effective undercut.
Jet profile looks good
-
1
-
-
7 minutes ago, Mark N said:
Fascinating, good to see mr murr back on here! ECM a peach.
ECM ensembles soon?
Actually still not that many supporting the Op - at least the control does
-
1
-
-
33 minutes ago, forecaster said:
That's called "confirmation bias". Ensembles aren't perfect and they are underdispersive. That's what you need 100-200 members for. They are NOT "as prone to flipping as the op". Science backs this up.
Ensembles and their means are averages so as you say are not prone to swings. They are extremely useful if the weather turns out to be near average as of course it often does. However when the weather deviates far from average then ensembles are well pretty average at predicting this. This has been highlighted IMO by this month which has been exceptional by any standards. However I do not recall any extended products predicting this level of extreme temperature excursion. Extended range forecasting based on ensembles is good at saying it will be a bit colder/warmer wetter/drier etc but in terms of predicting significant deviations from this are currently fairly poor. So if you are looking for any kind of significant deviation from the norm the ensembles IMO are not that helpful - only in identifying trends. I tend to use them as many do on here as to what the possibilities are going forward are rather than what they are showing as a mean as that is nearly always not far from average (or the default zonal set up). The current situation is a good example of this. There is clearly potential; for a significantly colder/drier spell but based on ensembles you would not state this publicly in forecast (if that is what you do professionally). Hence the current forecast from Meto but to most people what is stated means there will not be any significantly colder weather in the next month. So if it transpires that there is then this will be a busted forecast as far as most people are concerned.
On a separate note as a physical scientist in another field it seems strange to me that modelling is done by feeding in what is effectively adjusted (or duff) data and seeing what the result would be (i.e. the ensembles) and then take an average as prediction.
-
2
-
-
Whilst the 18Z has been fun to look at I think that it is all it is likely to be. Looking at the ensembles whilst there is quite a bit of support for the overall pattern there is is none for the extreme evolution of the OP so it likely to be an outlier. Most of them lift the cold pool out before reaching the UK. Not to say it won't happen like that though.
-
1
-
-
Interesting control run on the ECMWF control although not much support it has to be said.
-
1
-
-
ECMWF forecast with a model based climatology updated and goes into 7 in F1 albeit at a rather low amplitude.
-
1
-
-
5 minutes ago, IDO said:
Hi
The 06z OP is a shocking outlier amongst the ensembles and I would just discard it:
Looking at the GEFS ensembles, the Control supports the OP re HLB but only one other member really going with that setup and the mean keeps any pressure build as more of the same at D16:
Though as we would expect, temperatures will be moderated, as I dont expect the next seven days +7c over average could continue! A 48 hour cooler period just before the big day but the uppers on the rise again from Christmas Day:
GFS clearly not handling the likely more meridional flow after D11 and more time needed before it gets its act together!!
Not really shocking in terms of the overall pattern though - in fact quite close to the mean
-
2
-
-
20 minutes ago, MPG said:
FI and extremely short lived on the GFS.
Of course it is FI he is discussing matching the MO with the Meteo 30 day outlook
-
1
-
-
Has there ever been a mean like this at just t144 - what is happening in the Atlantic - nothing?
Looks like the control and mean were the coldest at this time frame and did not phase the two lows - then became the mildest
-
1
-
-
One thing is for sure - we are going to see some HP according to the GEFS - Op off the scale and mean at 1030
with some matching cold temps
-
4
-
-
Woke up to 3 rumbles - looked out the window and saw these beauts being lit from below by the sun (looking NE)
Out the front middle level cells demonstrated nicely how you do not need surface heating for this stuff.
Nice line of storms moving NE across biscay
http://de.sat24.com/de/fr/infraPolair
Just had a loud rumble and have some warm rain just now (5:40)
-
2
-
-
Nice easterly in the S in the JMA
-
2
-
-
Yes the variable ensembles reflected in the means at 192
Although temps look ok till then
-
1
-
-
Looking at the Jetstream profiles - forecast is for a split jet, this is crucial as it will reinforce the euro trough and prevent heights from sinking too far south, a very likely scenario is further energy from the NW sinking SE forcing the azores high out to the west whilst any strong energy off the eastern USA seaboard is deflected well to the north west on the northern arm of what will be a much weaker jet overall - as I said good chance of a rinse and repeat scenario - BBC certainly suggesting this, with atlantic attacks coming unstuck to the west and forced SE into central Europe/Italy.
Yes - nice split flow at the end of the high res run with more energy going into the southern arm
Would expect to see more disruption from that.
Model output discussions pm 31/12/15
in Forecast Model Discussion
Posted
Ye I was going to comment this in that this seems be a major change overnight. The big low in the Atlantic was due to move slowly over us bringing lots of rain and wind. It now seems to to want to stay out in the Atlantic and is if anything moving West rather than east. We do still get LP coming over us but more in the form of trough disruption with and an elongate finger of the LP. I would say this is goodnews for those that have been flooded as whilst not being dry by any means it should be drier.
This can be seen on alm models
UKMO
ECM
A couple of days the models were showing the Atlantic low right over us at that timescale.
Looking at the NH profile it seems the main reason for this is that the direction of lifting out of the low heights from Eastern Europe is more Northerly and not S Norway. Hence there is no link up with the Atlantic trough and this is not then pulled East