Recretos

Add climate change to that as well. This is an example of FIM analysis sounding over the north pole (90N,0E).

Just a side note to the person that made the "mouse-over popups" for MJO, GFS, etc... MJO is Madden–Julian oscillation, not Madden–Julien oscillation. You wont see much more ens diversity past day 10, especially not with 51 members. The spread increases and the mean gets flattened really bad. Unless the member consensus would be strong. That is why GEFS sometimes has stronger anomalies at day 16 with only 21 members.

The MJO seems to be heading into the phase 7 soon, as mentioned before, so I made the phase 7 composite for January, ENSO filtered and amplitude above 1. I didnt apply any lag, because the dates are mixed, some are in the middle of the phase, some at the end,... So in average a default lag time is already integrated. There is some similarity between the MJO composite and the EPS forecast. You could say there is an MJO signal "present" in the forecasts. As for anomalies, it also doesn't differ that much from the composite. Tho we must realize that there are other factors at work even in the MJO composite, since that is not the only thing that affected the pattern in those days that are in the composite. Nonetheless, the similarities are there which is an indicator of the MJO signal being present. Regards

Well, +5 or -1 could also be a difference of 1500m geopotential height in the core, or vortex core being 2000-3000km more to the south, which means the "anti-vortex" (the high ) being more over the pole, etc,... It might not sound like 6 m/s is a big deal, but it can be, in the big picture. Lets just keep to the 60N 10mb zonal wind reversal as a requirement. Every m/s less helps, once under 10m/s zonal mean. While I am at zonal means, I think that it is almost my duty to make a reality check. The trend seems fairly firm by now. The warming will not keep increasing at 384h. It gets closer and then diminishes. So there is nothing wrong with the models by making the 384h forecasts "colder" than in the last few days. That is just simply the evolution of the current trend. And basically what that means, is that we would need another warming trend. Tho there is still plenty of time for that to develop. GEFS for example, which has had a ridiculously good consistency, kinda sums it up, but of course on a relatively stronger-than-actual level, since it is an ensemble mean. We will have this first warming, dealing a good blow, but the vortex recovers, just to be hit by the second stronger warming, which also doesn't seem to do the trick. Control looks a bit better, but it still isn't enough. Even GFS is not that convincing. The cross-section is not really encouraging either. And the ECMWF is now starting to grab onto the vortex recovery phase I mentioned above. I don't want to be seen as a pessimist, but I am trying to put a bit more emphasis on the zonal wind aspect, since it is just as important as the warming. And the zonal wind forecasts are not really close to a technical SSW, so I wont even start speculating about possible effect of this warming on the troposphere. Lets hope the runs eventually improve on this aspect aswell. The displacement on itself can be helpful (so to speak), but I will wait for the "warming+after warming" period to get within the ECMWF domain, so we can see some model-to-model consensus. Best regards.

First I have to apologise for not plotting the MJO composites, because I had some other priorities. To answer blizzard: I get data from the NCEP FTP servers. http://www.ftp.ncep.noaa.gov/data/nccf/com/ Regards.

Come on now guys, comments/discussions like that (about Santa and ponies) do not help this thread in any way. More like the opposite of help. Anyhow, The post: http://forum.netweather.tv/topic/78161-stratosphere-temperature-watch-20132014/?p=2881764 That EAMT event was forecasted to be more baroclinic biased in nature. And MJO was/is still sleeping. But nonetheless, that very EAMT event that I was mentioning then, did verify by ECMWF and GFS. Of course it is not as strong as that notorious 06z run had it (not to mention the differences in the strat. model output since then). And to date, that 6z run was still the best run this entire season. By 180h, the ECMWF completely neutralizes it, while GFS holds barely on to it, but eventually it will go back into negative mode. 6Z GFS also tries to recover the NAMT in about a week time, which might be good to note (or not), at least when looking at the latest 6z strat output (speaking of which, the 6z might be a loner, or a start of a new trend, which should be more obvious by the 18z run). And at the same time it holds the EAMT still in place, but not for that long. I added the zonal flux of gravity wave stress, with which I am going to experiment in the coming days in regards to MT events. As for the MJO, it is/was sleeping for quite a while now, but it looks like it might come back to life in the 6. or more like 7. phase. At least GEFS seems to be allover the phase 6. I will try to do some MJO phase 6/7 composites (in regards to surface patterns for MT events) when I get back from work around 22 UTC. Best regards. Edit: I know this post seems to fit more in the model thread, but I am linking this to the stratosphere output (or at least I am trying to), so it fits better in here.

Well, GEFS does play a role of an "underdog" in this game, with its top at 2mb only. And the NCEP models in general seem to be faster in zonality than ECM (in the stratosphere). So the question is, if the ECM is too slow or are GFS/GEFS too fast. You could argue the GEFS by being low topped (in a certain way, but GEM ensembles are even lower topped at 10mb), but GFS is much higher and also fast. So you can mainly blame it simply on the physics. Because changes in strength/positioning of the pattern is what affects the outcome. And the "pattern" is what the zonal mean zonal flow is calculated from (or actually its a U-component of the wind field. ).And the comparison of the ECM and GFS 18z and GEFS 18z. Note that GEFS only goes up to 10mb. It is obvious that ECM is much more aggressive. And the difference in resolution is quite obvious. Well... And the GEFS output (I am testing some new colour combinations): With the GFS going over 0°C quite constantly now, I dont really expect the GEFS to follow, since a lot of the game is played at 1mb and above. 2mb is not too bad for a model top, but it can make a difference in strong top down events. Tho the bias corrected version does reach max. -11, which I think is the highest on GEFS so far. Looks pretty much like the Yin/Yang from last season temperature charts. And to finish with more colours (), the time-series of zonal wind and temperature from GEFS: Suma summarum: The trend is most definitively good, but as it can be nicely seen from the GFS zonal mean cross-sections above, nothing really "warming related" is getting below 10mb. So don't hold your breath to see warming induced effects in the trop. model output just yet. Just a friendly reminder. And compared to ECM at mid range, I can only imagine how aggressive the ECM would be if it would go out till 384h. A bit more colorful post, but its much better then text only. Cheers. Edit; Added the GFS 18z 10mb anomalies. Based on NCEP-R2 Climatology.

GEFS 06z continues the trend, now reaching max. -14 deg. with ensemble mean at 384h. And the control ramps it up to +11. Note that this is the bias corrected control run. I couldn't plot the bias-corrected ensemble mean above, because it is not available yet and I am late for work. Regards.

As I suspected. Thank you. Anyhow, GEFS plays along with GFS (or vice-versa) and slightly increases the warming. Important notice: GEFS does not postpone the warming like GFS sometimes does by always keeping it in the 384h period. It still has the same amount of warming that was there 4 days ago in FI at the same date, and just evolves it further. If you would see the animation, it is slowly increasing the warming and moving it more to the pole. So GEFS has almost a surreal consistency. Not to mention that this is an ensemble mean, and specific perturbations go up to +5 max. temp at 10mb. I also made verification by comparing old GEFS runs (ensemble means) to the analysis for current time, and they were mainly colder than the reality, of course due to the fact that it is an ensemble mean and a lower resolution, so it can also be less intense than operational or control runs. Speaking of control runs, they weren't bad actually for the FI when I did the comparison of current state and the FI forecasts 2 weeks ago. But the control seemed to have a really too cold vortex core. Of course I have to add that this was verification on a short scale and does not necessarily reflect the long term statistics. Tho if we look at the verification for the last month, GEFS mean does indeed have a negative temperature and height bias, basically meaning that it has a slightly too strong vortex in the FI. I do believe that this is the ordinary GEFS version being verified and not the bias-corrected version. Ok, now for the actual GEFS charts. The control run does seem to have big personality issues with the vortex. Speaking of personality issues and biases, the GEFS BC (Bias Corrected) run seems to mitigate just that what I was noticing in the first part of the post (about the GEFS bias), as one would expect of course, since it is a bias correction.

Since correlations seem to be a point of interest at the present time, I also decide to have a go at them. Without further ado, i present some basic, but interesting QBO correlations plots for January. I decided to plot for January, because that is mainly where most of the intense SSW related stuff occurs (so to speak ) and because we are in January at the moment. I can always plot correlations for any month if the need be. I used the NCEP R2 dataset. I decided to use the 80-13 period, because of the greater reliability of stratospheric data in the satellite era. I filtered the negative and positive phase from the NOAA QBO dataset, to get correlations specific for one phase or the other. First I present the correlations for the east or negative QBO. 10mb temperature correlation: An important thing to consider here, is that the colour table is reversed. That is because I am correlating a negative value. So a positive correlation of negative only values, means that the other one is also negative. So the warm (cold) colours in the legend, tell you where it was warmer (colder) in the east QBO at 10mb in January. And the values tell you how strong is the correlation for the "warmer or colder" area. If I put it in layman's terms. Specifically this correlation tells me that there is a very decent correlation between January SSW events and the east QBO phase, which has been established quite some time ago. 10mb Geopotential height. The same as above regarding the legend colours and the interpretation. And again the same, only this time for 30mb. Now for the west (positive) QBO phase. I will add December geopotential height, since we were in a west phase this season. As you will see better for January, west QBO seems quite wave2 "supportive". January. I wont add 30mb, since there isn't really such a big difference. ________________________________________________________________________________________________________________________________________________ I know that the concept of correlations in meteorology might not be so simple to understand for just anyone, so let me give you a quick crash-course in correlations. When you compare two values (an index and a parameter in meteorology), you can determine what is their relationship, if any. A positive correlation means, that as one value goes up, so does the other. And a negative correlations means that as one value goes up, the other one goes down. Max. correlation is 1, which you could interpret as 100%. The best example to demonstrate, is the correlation between NAO and 500mb height. We know that the NAO index is calculated from the pressure differences between Iceland and Azores. Now when we look at the correlation, we see strong correlations (~0.9 out of max 1). This is a correlation of NAO vs 500mb height, so looking at the Azores, we see a strong positive correlation. That means that when pressure increases in the Azores, the NAO index will also go up, while at the same time the opposite is true for the Iceland area (hence the negative correlation, because as pressure here drops, the NAO goes up). Now we are comparing the index to a variable that is pretty much the core of the index itself, because the index is calculated from this pressure difference, so the index follows the pressure difference (that is why we cant say that NAO determines the pattern in N. Atlantic, since it is exactly the opposite ), and so we get an expected high correlation. Its not 1, because there are more variations for the NAO phasing than just this classical one, tho it is by far the most common. The important thing to note, is the fact that even tho these correlations tells us a possible relationship between different parameters, they dont tell us what exactly is the relationship, or more importantly, who is the dominant one. For example, if I correlate the QBO and 30mb zonal wind, I know that the wind is the dominant over the index, since the index is calculated from that exact parameter. Or for example NAO index and surface temperature, where NAO index is "dominant" since it is calculated from the pressure differences or the "pattern" which determines the surface temperature. The correlation with the sfc. temperature is weaker (tho still very high), since the surface temperature is much more sensitive to the pattern differences, while the area that "calculates" NAO has much more tolerance to these differences. This is the same way I made correlations last year, when I was searching for SSW precursor patterns simply by "cross-linking" different indexes/parameters. So all in all, these correlations are a great tool, but some basic knowledge in meteorology is required, so you can correctly interpret the results. I read something about the linkage between the 30mb vortex and the tropospheric vortex positioning. Now for fun, I made the "cross-linkage", by correlating the NAO (which has a specific tropospheric pattern) to the 30mb geopotential height. The correlation is ~+/-0.6, which is way too much to be ignored and indicating from this perspective a long term correlation between the positioning of the 30mb vortex and the tropospheric vortex. The 30mb pattern even looks very "NAO-ish", and spatially lagged. So I would not go so far to say that there is no (or no obvious) linkage between the trop/strat vortex positioning. Since I already spent 2 hours on this, I might as well spent 15 more minutes, and add the latest ensemble guidance. Cheers.

Sorry I was at work (the night shift) so I couldn't post the 18z mean. But now I just got home, and I simply have to look at all the latest runs before I take a shower and eat my breakfast (or make that dinner ). So. The 18z GEFS was basically the same as 12z. And 00z is pretty much similar. i wont include the bias-corrected version this time, because they are practically without a difference in the 00z run. Regards. Edit: I also made an animation of 10mb temperature from GEFS.

Uhm... Funny enough, if I keep the same legend scales as on the previous graphics, the bias-corrected version of GEFS looks even better, with a weaker vortex and a bit stronger wave, as one would expect, and the warming only a degree weaker. We might actually be onto something here, since the GFS and GEFS seem to actually agree this time. And for those of you that like dealing with anomalies (personally I don't like dealing with anomalies at all, when it comes to the stratosphere): Regards.

GEFS sure doesn't joke this time around.

I don't know for you guys, but personally I am more interested in the actual wave 1 disposition or the "embedment", than the potential warming.

Well, there isn't much for me to post about at the present time. By the way, I am "he", not "it". Nevertheless, here is some model output, which pretty much tells a similar story. NAEFS+CMC GEFS Suite NOAA Reforecats v2 Other models (with 48h delay): Best regards

What thickness exactly are we talking about here? What levels?

I somehow missed your post yesterday. I already answered that question in the topic from last season. But to give you a summary: The GFS 300-384 mid-upper stratosphere forecasts, are generally as "reliable" as 500mb forecasts in the 6-8 day range. But since the stratosphere has its features on a much larger scale compared to the troposphere, the skill scores can automatically be a bit higher due to the verification process.

Not to take anything away from the FU-Berlin archive, but I think you have very comparable data and analyses in the freely accessible NCEP R2 reanalysis. Or perhaps even better with my favourite, the ECMWF ERA reanalysis project. The tricky part about the ERA data, is that you also have to register and the data is in gridded format only (at least as far as I know), so you cant make online plots like on the PSD site for the NCEP dataset. You have to plot the ECMWF reanalysis data manually. As for the 1991 winter: I was again beaten to it, this time by s4lancia and Interitus, but nonetheless I will present some random stuff from ERA-Interim, since my post was already almost complete at the time they posted. Temperature at 10mb peaked on 10.1. 00z, at -1°C. It is worth noting how the wave 1 was of Atlantic origin, instead of the more common Pacific origin. The mean zonal wind at 10mb was never reversed at 60N. There was a reverse in February at higher latitudes. The mesosphere was also "boiling" from all the likely wave breaking and energy released. There are indications of a possible EAMT in mid-late December 1990. But exact calculations would have to be made to be totally sure. Now finding some tropospheric responses is another thing. As you can see, there were strong height rises in the begging of February, with a quasi-wave 2 burst, briefly splitting the vortex. So looking at the timeseries, the early January warming was not really the main direct driver for the February winter, as far as the stratosphere role is considered, because the vortex pretty much reformed after that and there were only brief waves until the late January wave burst. The 500mb Azore ridge went all the was up to Scandinavia, so I plotted the timeseries for the 15E-30E region. It began in the troposphere, went up into the stratosphere, and came "back down" on steroids and at the same time preparing the "terrain" for the February UK winter mayhem (as for its part in it at least). I will wrap this up since the amount of graphics is getting out of hand, and this topic is called "2013/2014 stratosphere watch" and not "1991 case study". And besides, colleagues before me have already pointed out some of the other important stuff about the 1991 winter. As I said already, if you need any specific day-to-day or month-to-month comparisons from 1000mb to 1mb (and some to 0.1mb), just say so. I can plot it and I'm sure it would be interesting for others too. Best regards.

Well, I am also interested in the cause of the run-to-run variations. Now, at 10mb, there are differences, but less defined than in the upper strat, especially at 1mb. That of course indicates the differences in wave braking calculations. Which downstream relates to MT events. So I think that this could be a place to start. Now of course that means we have to go into the lower troposphere, where GFS skill scores drastically decrease beyond 192h ( and the reduced resolution decreases it even faster), where the main MT event(s) is(are) occurring that is relevant to the FI strat forecast. Looking at today's 6z run and yesterdays 6z, there is quite a difference in the EAMT department. Around 192-240 there seems to be a MT in the works, more defined on the 6z yesterday. At this time, the upper strat situation seem relatively similar, with no major anomalies. Now, the 6z from today doesnt really intensify the MT event from 192-300, and kinda backs off slightly. Yesterdays 6z on the other hand, really kicked it into overdrive from 200-324. And looking at the 384h 1mb temperature, there is a certain linkage between yesterdays 6z EAMT and its FI upper strat output, with the main action cenetered just downstream of East Asia, since waves do not "fly" straight up vertically. So the next thing that comes to mind is of course ECMWF. Currently there is a negative EAMT in progress. But buy day 10, the situation is reversed with a possible quasi-baroclinical +ve EAMT onsetting (at least in the model). And this actually looks quite different than the latest 6z GFS op, and a bit more similar to the yesterdays 6z for the same period. And the EPS generally supports it. Please forgive me if I am not allowed to post these aspects of forecasting here, but since I am "coupling" this directly to the stratosphere model output I do believe it fits here more than in the model thread (which I must admit I don't really read , so I don't know if someone has pointed this out or not already). And I have to add that the EAMT is by far not the only thing that affects the strat model output, but it definitely plays a big role. And in the latest outputs there is definitely a EAMT related signal. Best regards

I am pretty sure that he can do plots for any layer that is available after post-processing, which is 10-20-30-50-70-100 for GEFS stratosphere. Maybe I don't understand your sentence, but anyway, the so-called GFS ensembles and GEFS are basically the same thing. So the limitations that me and bluearmy were discussing earlier, apply to these plots. And even tho these products are really welcome because they provide an additional perspective, I am very dubious about the general usefulness of GEFS in stronger top-down events. But as I said many times now, the real test is yet to come. This is my first season that I am actually making custom plots, so I have no operational observations of GEFS stratosphere dynamics. And I am also monitoring CMC ensembles and the combination of both (NAEFS), and none of them delivers anything worth mentioning. But (there is always a but), I totally forgot about the navy ensembles, which have 42 vertical levels (tho the same as GEFS), but the model top is at 0.04hpa. And there is a real difference between these ensembles and GEFS. I am still trying to plot the ensemble mean again, but I am having some difficulties with the OPeNDAP data protocol. Lets just say that mostly all ensembles are warm pretty much over the pole. Tho the geopotential heights do not really seem supportive for such a scenario, its at least seeing some dynamics. I was also doing some ECMWF-NAVGEM comparisons recently. And NAVGEM is actually quite similar as the ECM. Temperature, height and U-wind wise. Aside of these general parameters, I was mainly interested in the EP-Flux, which I obviously cant produce for NAVGEM. But I can plot vertical velocity all the way up to the mid stratosphere. On the flux chart, we can see the strength peak on 1st, gradual weakening to around 7th, and then very slightly picking up around 9/10. Now for a comparison with the 30mb omega (the 60N lat. line is on the south tip of Greenland. So I generally look close to this line). Worth pointing out is the fact, that just like the EP-Flux, so does the omega intensity generally reduce with height (logically, since the pressure decreases with height). So that's why we look at those times when both the omega and especially EP-Flux exceed the expected norm. If I plot the omega vector magnitude, you can also get the idea, tho this is obviously not really the same as EP-Flux (and it cant be), which is a much more complicated derived quantity. And some short term lower mesosphere forecasts. It fits perfectly with the ECM' wave forecast (as one would expect I guess ). Wave 2 is not really that present in the lower mesosphere, while wave 1 looks quite decent there, as also obvious from the quasi-textbook displacement pattern. Wave 2 reaches max. peak between 3-5mb in this period. Now this isn't really of any big direct importance to the dynamics in the 300-384 period for example. But nevertheless I took the opportunity to try and present even more practical "explanations/examples" of the ECM wave charts that might be confusing for some. Tho it is actually a derived calculated parameter, you can still easily find clues and "correlations" on the basic parameter maps, from which these special products really originate from. @Aginob: Most (if not all) of the sites that provide non-standard ECMWF data, are pay-per-use, because the ECMWF data is not for free (except the WMO essentials), like for instance the NCEP' data. On a side note: Whose idea was to change Joe's surname into laminate floori? Ok, enough of my spamming for a while. Took me two hours for a post that you can read in two minutes (graphics want their time). Cheers.

Yes, this is mainly the difference in the QBO between these two years. Nonetheless, looking at the 20.12-30.12 periods, the 2013 was still a bit stronger than 2008. This is the difference between 2013 and 2008 for the 20.12-30.12 period. It shows where the 2013 period was stronger/weaker when compared to the 2008 period. The equatorial zonal flow basically corresponds to the QBO phasing, and the zonal flow at higher latitudes corresponds to the state of the polar vortex. So I am not presenting anything new actually. If you would like to see 2013 compared to any specific year, just say so. Regards. p.s.: I was looking at the JMA weekly 10mb temperature change, so I decided to make one too. I also added one for geopotential height. Basically wave 1+2 on the increase in this period. Tho wave 1 intensified a bit more in this period, it naturally also affects wave 2 calculations. Temperature waves following accordingly in retrospect. p.s.2: GEFS 12z mean still with no real dynamics, tho I have actually found 2 perturbations that are worth mentioning. But the waves look a bit off (compared to higher topped models), which I do believe is because of the low model top.

It depends on how the stratospheric picture looked like in GEFS last year, and in what lead time did it start showing the response. And lets not forget that the SSW last year was different than what this one might be (if at all), with the later split going very fast throughout a very deep layer and with much more direct interaction with the troposphere (hence the lesser sensitivity to the model top), so the ensembles had a slightly easier job than this year (as you mentioned the top-down or bottom-up). This year GEFS obviously has no dynamics whatsoever in the later stages of the current runs, so if it wont grab onto something in the coming week and will continue with this p.v. stability, than it will be interesting to see how it will evolve the tropospheric patterns compared to the ECM EPS. But of course considering if there will actually be such wave activity as it is being suggested by the mean of GFS op runs for the FI. As I said, this will be a very good test for GEFS, especially now when I/we can actually see how it looks like in the stratosphere. I hear ya on the ECM vs GFS op U comparison.

I was thinking the same thing. Compared to the general trend of the past runs, 00z run was quite suspicious with its strong negative deviation, almost like a run error. On the other hand, we now have the 6z which is one bomb of a run. But since it has a strong positive deviation from the "mean trend" I am of course taking it with huge reserves for now too, tho I would not call it a run error by any means. Nevertheless, here are again some 3D products. 2 animations, that repeat 2 times. 6Z GFS, geopotential height topography. 3 levels, 10mb, 30mb and 100mb. By the end of the run, we can see the obvious wave domination at 10mb, with height rises and the vortex getting the split features. But (there's always a but), it doesnt not yet reflect that much below 20mb. Of course there are some features at 30mb, like the weakening vortex core, and the Atlantic wave coming into picture in the later stages of the run. This is of course because of the downward propagation that I will present later. On the 100mb level however, the vortex remains stable with no disruptions to the mean circulation, except the existing ones. But that is of course because the propagation has not reached this low. In the 6Z, the split features only starts slowly appearing on the 10mb "plate" in the later stages of the run. The perspective is the standard "USA in front". gfs6.mov And the side view. The perspective is from Europe. So if you look at the animation above, this one is like looking from the right side in. Most notable is the substantial weakening of the vortex at 10mb. gfs6_1.mov Now, as for the upper strat, it took me 1 hour, but I managed to get a basic animation together. Geopotential height topography from 300-384. The levels from top down are again like the last time: 1-2-5-10-30. Wave activity is really substantial in the upper strat. The upper vortex gets quickly dominated. There is also an obvious and fast downward propagation, that slows down in the mid strat 10-30, due to the normal reduced wave activity amplitude at lower altitude. The perspective is again the classical "USA in front". Upper_strat.mov And images of the 384h timeframe, showing the same layers as above in the animation. The perspectives are "USA in front" ("From Europe") in the first (second) image. I dont think I have to repeat the wave amplitude/height "relationship" again. The zonal wind is reversed in the upper strat, but it actually slightly increases with time at the end of the run in the mid strat @60N. But if the run would go further out, it would eventually fall back down. The geopotential height anomaly at 10mb is quite substantial. And on the temperature mean it can be seen how the temperature wave propagated down, but has hit a "wall" in the mid/lower strat. Well, I am monitoring the GEFS output the whole time, and lets just say that I am mainly ignoring it. The model has the top at 2mb. That is too low to grab onto the strong wave breaking in the later stages, of which about 50% is in the lower mesosphere. Here is how the GEFS mean looks like: As I have said before, the only ensemble output that I have any "faith" in, is the ECMWF EPS. It is the only ensemble output that meets the required "pre-processing" conditions to be a viable option. Unless GEFS is right and nothing will happen. Personally I have tested GEFS all this time, and it doesnt seem appropriate as the wave amplitude increases. So my bet would be on the general trend of the operational GFS runs. You can now trust either my opinion which I backed up, or the general "weight" that people put on ensemble forecasts. You have to realize that we are dealing with a different "level" of numerical forecasting here, which unlike its tropospheric counterpart, has much more sensitivity to the model depth. But of course the main GEFS test will come in the next 2 weeks. Best regards.

Funny that you ask, because I was just trying to present the latest ECM32 (or mainly the control run) in this direction. Now this model has 91 vertical levels with the top at 0.01hpa, so it has the basic conditions required I guess. We are going into deep FI so, keep that in mind at all times. Up to around the 17th, there is no major change in my view. At least not anything that would be related to the recent GFS scenarios. Funny that you mention this, because in about week time, there are new dynamics induced into the tropospheric mean circulation, which might have something to do with the activity above. I personally believe that without some external dynamics from the stratosphere, the tropospheric pattern could not really shift this hard in one week time simply on its own, with no changes to the pre-existing stratosphere conditions. And a couple days later, further (and stronger) disruptions of the mean circulation. The AO index of this control run probably tanks at -5 with ease. I am certain that the control run has major stuff going on up in the stratosphere, similar to the CFSv2. But of course the ensemble mean doesnt want to hear anything about new tropospheric pattern dynamics. But what else do you expect from an average of 50 ensembles at 768h time. Regards. Post scriptum: Thanks to MPG for providing a good dialogue with the questions.

Thanks Lorenzo. Trust me, I would be in a really bad mood if those telecom people wouldn't show up ASAP. By the way, I am also working on potential vorticity animations like those, but I am having a slight problem with my RAM. It wants to burn out. It will take further optimization to reduce the load on the computer, so I can make some 3D animations. Lets have some more fun (I say "more" because the 18z kinda began the real fun and games). If a technical SSW is what you are after, than look no further than the CFSv2. The most optimistic scenario so far. Take a note that it has a technical SSW around 22/01 (pretty much on 21/01 I think), with an awesome residual vortex displacement right over Europe. This is actually the first time I have noticed the CFSv2 with a technical SSW. It is interesting how quickly it cools down the top strat. afterwards. It is also worth mentioning that it completely reforms the upper strat. vortex within a week. Regards.