Trom

I'd be surprised if we get a SSW by the end of Jan/early Feb but I'm not sure we need it given then the forcing from the trop. All we need is is the continued trop to strat influence and a bit of fingers crossed if you want cold. The signs are in place regarding the MJO and the constriction of the PV shown over current runs. What happens next, if at at all, is like a ball hitting the reds from a snooker break i.e. no guarantee that they end up where you want them. So a bit of short term pain with the constriction and then who knows? Certainly the NWP models have been guessing.

Thats exactly at the core of the issues the models face. As we know the numerical models essentially apply algorithms based on fluid and thermal dynamics to initial variables to attempt to predict a future state. We also know that the robustness of the algorithms is tested by using historic data (analogues) to predict the past (a process known as backtesting or in the weather world as hind casting). So in other words can the model predict the past - this is essentially the process used to build and refine the model. Given in this current period we are dealing with an El nino event with warm waters in both indian and pacific oceans which is a very rare event (I think I read another poster saying the last occurrence was 1963) and a January hurricane in the North Atlantic (last experienced in 1938), it gives you some insight regarding why all the models are struggling and throwing out large changes run to run. Essentially the algorithms are not really calibrated to cope.

The basic concept is the models are deterministic i.e. they follow a set of algorithms based on past conditions. So there is a fair bit of back fitting with the models. Given this they will take the current base conditions and project it forward according to the models parameters. The ensembles are what we see when we tweak those starting conditions. What the models lack is common sense. The skill that at present can't be replaced by pure equation at the current point in time. Experience and the luxury to be able to focus on a specific area with greater available info than we see. What we are also seeing is a hugel El Nino coupled with warm water in the pacific and indian oceans. This is not a normal situation and as such not one the models equations know the bounds of the reliable future even a few days in.

Yes but the issue is how well are the models actually coping with a situation that they've not encountered before. We've not had El nino conditions with such warm waters in both indian and pacific oceans in decades. Presumably the deterministic models lack the experience (for want of a better word) of modelling this situation. Also there's a fair chunk of winter left and if Tamara and Stuart are correct then a rapid movement through early MJO phases could have us looking at wintery synoptics improving the later we go.

Tamara - what did you make of that strange loop the GFS threw up in phase 8 the other day if you don't mind me asking? It certainly seems to explain why the GFS has been at odds with other models. It's interesting to see that you think that speed of the MJO is underplayed in the early phases. If correct it means Feb could be very interesting.

One way of looking for certainty or at least high probability is to look at how tightly the ensemble members are clustered and at what point they start to diverge significantly. So look at for example the GFS ensemble charts 850 Temps. If you do that right now you can observe they are tightly clustered to the 13th of Jan and then start to diverge. This gives you an idea that the ensembles strongly support the control up to this date. This therefore tells you that the GFS is fairly confident in the pattern out until the 13th of Jan. Also you want to look at clustering i.e. what proportion of the ensemble members are tracking down and which appear to be outliers. Comparing the individual ensemble members to the mean is useful for this. The next thing to look for is how similar the various models i.e. GFS, ECM etc appear to be. When the models diverge and ensembles diverge you will hear people talk on entropy which just means a lot of volatility in the output so hence more uncertainty. How far into the future you have to go until you see significant divergence varies depending on how the various models and ensembles are handling the current set up. More consistency basically means more certainty.

Good post on an area that I've been exploring myself. Regarding mountain torques I had a good exchange with Tamara on this subject - it's certainly a interesting topic. The model thread often gets focused on the stratospheres impact on the troposphere particularly via warnings but of course it works two ways and the Rossby waves deflecting upwards are an important influence and driver of the warmings. There's actually a nice wiki entry on the subject http://en.wikipedia.org/wiki/Rossby_wave I still have plenty to learn on this subject but thought your post was a very tidy and clear summary.

I was in Amsterdam the last couple of days and it was bitter. Interesting ECM and positives from the GFS 18Z - albeit they take very different routes. Seems to be indications of potential variability from about six days out. Until then some dry days so get outside!

Trust me there is such a thing as too cold. Just came back from LaCrosse Wisconsin and it was hitting -24 at night and not above -6 in the day. I notice it's a bit warmer in the mid west right now. Interesting to see the ECM evolution tonight but without any support it's hard to to believe it. GFS also showing interest but with a very different evolution. Leaves us in the usual area of "more runs needed" but some volatility is better than none if your looking to shift that high.

They must test the algorithms and adapt them based on historical data to prove the model one would assume. If the model can't predict whats happened in the past then it would be no use in predicting the future. So I would imagine some back testing using historic data must be carried out as a proof of the algorithms. Edit: sorry mods I see this is correct from other posters - please feel free to delete.

Looks like the best Southerly location for snow is Berkshire/Wiltshire. It's that damned M4 - normally it's the dew points but this time the direction. Some how that road is cursed for snow lovers in the South.

Wouldn't worry to much - looking at the radar it's only just reaching St. Albans nowish

This is going to be a near run thing for Streatham. Pretty much South of central London but with a SW post code. Needs that little curve. Looks good for Berks and Surrey.

The point is he's given a possible scenario and backed it with charts. If people can't suggest evolutions of runs then this thread would be pointless. I don't see any claims of likelihood but just possible evolutions of the current synoptics. There have been a few posts recently that have stressed the importance of this site as a platform for learning which I whole heartedly agree with. I have no issue with someone pointing to a possible direction we could take down the line, especially when the post is in the realms of FI anyway. We all look to the next period of volatility for a take and the ensembles suggest that's next weekend. I've seen a lot of indicators that the MJO may be begin from this point forward (5th Feb). I'd be interested in posters views of the significance of this for the Atlantic going forward.

Yes it does seem to have been a pattern over the last few days. Let's see if it repeats.

Well I don't work for the papers but I'm guessing extremes of anything from celebrities to weather helps shift more papers. "Azores ridging leads to some slush events in the south" doesn't have quite such a pull to their potential readership. If you are getting your facts about the world from a tabloid then I'm not to sure you are concerned with fact and bias free reporting. Why people expect the weather reports would differ in quality from any of their other reporting of global events of any type mystifies me. Their sole remit is to sell copy and stay marginally the right side of ending up in court (a line which is often crossed).

No unfortunately for the far South it looks slightly the wrong side of marginal but certainly a never say never scenario. I'd say sledging with the kids would be an unexpected bonus for London. The rest of the country looks set for a week of cold. The ensembles for London are showing remarkable consistency for temps out to the 4th/5th of Feb - 9-10 days. With some notable posters missing today there's been a slight lack of analysis of the drivers behind the models. Tamara's post was a good read earlier https://forum.netweather.tv/topic/82366-model-output-discussion-23rd-jan-12z-onwards/page-46

Maybe the charts should carry the same disclaimer that financial services have, "past performance is no guarantee of future performance"

Surely there would be a lag in activity and the impact on our planet. Presumably what ever mechanisms are in play don't travel the galaxy at the speed of light? Edit: A google search reveals a CME (coronal mass ejection) event takes on average 98 hours and a solar flare between 24 and 48 hours.

Yes that seems to be a critical time. Only 8 ensembles kept us with Westerlies at that date. The rest showing NW to NE. So it looks from this that if we are to see another cold spell it's going to be northerly rather than easterly based.

Whilst it's late and so I'm unlikely to derail the thread too much with an off topic post, could I ask why the strat thread is not in the models area of the site? It seems to me that it would be better pinned next to the MOD thread to make it easier to flick between developments in the strat and trop. It just seems that it's odd that the strat thread lurks in "Weather Discussion and Chat" rather than "Forecasting Model Discussion". A very minor gripe - I know. As a relative newcomer to this site I managed to miss the strat thread as I would have never thought of looking for it in its current location. Currently I'm still developing my understanding of both strat and trop modelling but in order to understand some of the posts on both threads you need to be able to cross reference. For example Tamara's mountain torques have a direct upwards effect on the strat. Sorry Tamara if I keep mentioning your torques but your explanation of Rossby wave breaking was really helpful in understanding the interaction from trop to strat and suddenly a lot more of your posts make sense to me. We genuinely have some amazing posters on this site. I'd like to give a little kudos to Gibby, Holmes and Tamara on this thread and Recretos on the strat. All of you have advanced my understanding more than you know. Also Ian F has good posts (albeit it's easier to play poker when you know the other player cards but they don't know yours!). Sorry Ian not a dig at all but really just wish we had the detail of the GFS on the other models. I guess pay up or shut up.

Due to the western tracking caused by the earths rotation we will always have a tendency to westerly driven weather and as such it's easier to predict. Getting the conditions to prevent the westerly Atlantic dominating and therefore cold is a much harder task and prone to much greater uncertainty. Cold requires a number of elements to fall into place simultaneously and as such is a harder beast to predict and more prone to breaking down as variables change.

On the strat's implication on the troposphere I found this model hugely illuminating in terms of the direction of cause and effect. Source: Recretos

Thanks for that. So we're looking at that impacting on the charts most likely to be at the end of the range but something to be aware of in terms of cold in Feb. particularly interesting if we can get our current cold snap to extend which we've seen steps towards. Really exciting model watching with short term fluctuations making big differences and interesting macro indicators moving. I leave Tamara and her torques to explain that. I've been working on understanding her posts and it's really interesting and I'm certainly getting a better understanding. It's took a lot of reading and google searches mind!

Isn't it a little early for the warming to have an effect on the trop? I was reading an academic paper the other day that these take from 2 weeks to 2 months.