Trom

Yes that's what I meant. Sorry if my post confused. What I meant was you can't use the ensembles to add support to the GFS main run at the moment. I think it a third quarter update on the ensembles.

True that the GFS has overcooked the westerly progression in the past but remember this is the new model. It remains to be seen if it shares this bias with its predecessor. I keep making this mistake myself especially when looking at the ensembles - then remembering that essentially they are now from a different model.

Don't forget though that this is a new model - so whether it has the same tendencies remains to be seen

Definitely worse places to be diverted too!

Hat's off to that - that makes it easier to see the top down effect than anything I've seen.

Yes I agree this is certainly not a blocked pattern but it's trend has taken the colder route at each key junction so far which makes it interesting. It clearly to demonstrates that for a lot of the country wintry output is possible without a SSW event or a HLB. What I puzzle over is the impact of the minor warmings we've seen in the strat and how they relate to the trop in any predictable way. They've certainly impacted the PV even if it has been quick to reform. The other key input is the Azores high and it's westward retreat with some amplitude north. It's a significant change as we move through January. Of course we have the low pressure front tracking through Scotland on Thursday and and bringing very strong winds to areas further south. All in all given the variables in play I've been surprised how little variation there has been between control and ensembles on the GFS and between GFS and other models.

Very low dispersion up to the 19th - so the GFS seems confident in it's own output. A bit of cross model agreement will certainly increase the excitement levels for lots.

Actually I think their shouldn't be an in depth thread as it just seems stagnant. I think the mods need to shift more posts. To many posts that don't even mention the models. I think we should be encouraging posters to demonstrate some evidence for their posts. I still want more links to the learner section, This area should be more dynamic and newer users should be made more aware of it. There are some great resources there. In particular the understanding ensembles thread is a great tutorial thread.

It's interesting to note that the charts showed dispersion from 11th of January onward for about a week. This gave the opportunity for cold outcomes. It's nice to that as we've arrived at this date that opportunity seems to have verified. I'm certainly not claiming clairvoyance as all I saw was the volatility. Not deep cold but that was never on the cards, but for cold we got the best outcome on offer. Now we are here at that range on the GEFS ensembles the trend seems to flatten with less dispersion going forward. That sees us getting cold to the 15th and then flattening. ECM better if you want a continued downward trend.

I don't think you can remove emotion from forums and secondly it would be a lot duller place if you did. Part of the skill set you develop on here is the ability to spot biased and non biased posters. It doesn't take too long to develop this skill and also your ability to read the models improves as you do so. If you can't spot this then I'm not sure you will be getting much out of looking at the charts anyway. The only issue I have on these boards is finding where the relevant threads actually are. For example the Strat thread not being in model output and discussion seems strange. There is such a lot of content on these boards that being aware of what is available can be an issue. So my main issue is navigation. There are some great leaning threads discussing the models and a pinned link in the MOD forum would be great in my opinion.

Well I'm off to Wisconsin next week and it was -16 today! Ouch - but you can go ice fishing on the Mississippi in LaCrosse which is where I will be. Turns out that that's not too much fun as you can't really venture out to enjoy it for any period. I remember a ciggie freezing to my lip in Minneapolis in the days before I quit that habbit. A long way from cold verifying in the far South of the country but I'm sure lots are happy that at least the charts show it as a possibility. The GFS ensembles start to show significant dispersion from the 19th but reasonably consistently downward track before this date. As somebody who flies a lot for work it only takes a small event to grind the South to a halt as I can attest to on multiple occasions! Makes sense really as you can see from this year so far the north of the country can get wintry synoptics from the North West but the high pressure to the South prevents that in anything but extreme situations, so the case for spending millions on specialist equipment is that much lower. In the short run some cracking low pressure events to hit the country if you like your weather on the extremes. I will put my hands up here and admit to loving the extremes of our weather in all its guises.

Slightly ironic if it's the azores high migrating north that gives us the blocking after how troublesome it's been for cold synoptics this winter thus far.

Well interesting that that both GFS and ECM show volatility in the ensembles post 11th or 12th of Jan. For the GFS this has been a present for quite some time. Interesting that the ECM shows dispersion at the same range. Lets see what happens win we get there.

Noticeable that both runs have a tight spread of ensembles until the 11th and then quite a dramatic dispersion.

I'm all ready struggling with how the strat affects the troposphere in terms of predictability. Can you explain how the MJO has this effect on the strat and it's follow on potential repercussions at a lower level? I'm just trying to get a handle on this.

Thanks for that - What data collection period do they use for the averages (I assume they are moving averages) and are they available anywhere? Thanks

Yes last winters awesome low pressure features that seemed to batter us from December to Feb were really quite something. Not cold but certainly extreme. I'm a volatility lover - but easy to say when you don't live on a flood plane. Models not showing much in terms of extremes at the moment. Even the westerly procession for the next weeks looks to be fairly benign. That stubborn azores high really needs to go to give us some excitement regarding potential.

Trouble is a trend without a theory is just data mining potentially. Still has not stopped the masses in the financial worlds believe in chartism.

Yes it remains a problem of highs to the SW if you are looking for cold for the whole of the UK. Far North to be hit with transient colder periods. Not much to bottle up the atlantic. The SSW effects seem to be, based on my prior observations, to be like smacking the white in to a pack of reds in snooker - a big variable but how it shakes out in the troposphere seems complicated. I'll leave that to those with more knowledge on the subject than me.

This is exactly how I started - using Gibby's summaries and then going to the charts to see if I could see how he'd reached his conclusions.

Sorry I meant to imply that, what was it from the initial variables caused the switch and how quickly in the model the flip occurred. So I want to see the next control and ensembles from the ECM to see how confident it is relative short run, as we have a run of 1 for GFS (even if it is in the direction most posters want). Best mate is from Seattle (so knows our weather), has lived in Texas and now lives with the polar opposite in WI. He's in his 12th year and hates it - but they're paying him enough to live with it!

Yes it has been a bias that we've seen before - it's also been correct before. My point was really it's hard to have confidence in a model that shows that amount of change between two runs at such a short timescale (relatively speaking). What would be interesting is to know exactly what variable caused such a sea change. It's not as if it was a gradual adjustment - more of a wham. Lets see where it settles now it's picked up the ECM/UKMO trends. Personally I have more confidence in the models that have evolved to this solution. I'm not sure how GFS got to its output in the last run. Which variable changed and at what timescale? Anyway all good fun and very addictive watching those models. I'm lucky as my work splits 50:50 UK and Wisconsin so I'm guaranteed all sorts. Over in WI they are always hoping for mild! It just shuts off outside as a place you want to be for any period of time - compensated by more reliably hot summers. If you are obsessed by red meat, shooting things, pick up trucks but like a beer I'd recommend it. Last year was quite exceptional too with the mighty low pressure events battering us time and time again. I think Imight just be a volatility lover.

It's interesting to note that the GFS was being rubbished for being an outlier early today - which in fact looks to have turned out to be correct. But now it's showing the coldest outcomes there's little criticism. I sense there's a fair amount of cognitive dissonance on here tonight but understandable in the excitement. Will be interesting to see the dispersion on the ensembles as we move forward. Just yesterday we had huge divergence outside the 72-86 hr time frame - it will be interesting to see if this is now significantly lower.

Entropy is a measure of informational content within a data set. So in the context of models it is typically a used as a measure of uncertainty rather than certainty. Shannon is just one particular measurement of entropy. A high level of Shannon entropy means low levels of predictability or put in the context of the models, that they have low levels of predictive informational content. Very certain outcomes still have entropy values (the more certain the closer to zero). I only mention it as it seems to be a buzz word on here.

The spread post 27th is dramatic suggesting the models are struggling. Picking a model or run or ensemble member is too tough to call. The volatility run to run and between models is extreme right now with more cold trends appearing but far from being convincing on anything..