Recretos

Hi Bish. It is a demonstration of power. An epic battle for supremacy, between the polar vortex (fighting for survival) and the "polar High" (on the major offence). So to speak.

The soundings at 192h, 12Z run. Through the 10mb warm core: Through the 2mb warm core. I swear I can see sparks flying in there! Best regards.

The latest runs are of course still swinging a bit here and there, but an improving trend and progress are defiantly there! Interesting times ahead I could say. Happy holidays everyone.

No problem Chionomaniac. I was playing around with the latest GFS, 12z, converting it into the zonal mean graph @10mb. Looking at Meteociel charts, we have a slight temperature change till 216h in the 65N-90N area. I drew an approximate line for the 65N. I would say the mean at 12h to be around -75, and around -70/-65 at 216h. From 216h to 372h, there is a constant warm offensive. The power of the warm core is reducing along the way, but the cold core is also retreating, so the trend would be positive to up until 360h, maybe 372h. I drew my view of the 12z idea. Of course the time span is much longer than it should be, but it is hard to draw in a such tight space. So basically ignore the time scale. Blue line is up to around 200h. The starting warming trend would go up fast, but then slightly reducing the pace from 348 to 384. Basically not a bad warming on the 12Z, and I think my graph is a bit "pessimistic". The calculated means could be up to around -25 at max. Just another run, I know, but I decided to put it into this graph, for a comparison, to see just how good/bad it really is. At least in my view of course. Best regards.

I know. But I was comparing the exact runs and the same forecast time (240h).

When waiting for troposphere response, I think this graph might prove useful. Its the blocking strength/location forecast from CFSv2. Since CFSv2 is also a troposphere/stratosphere coupled model with downward wave coupling integrated, it should get the idea of a possible SSW induced blocking. Given the verification, it should start picking it up around 6-8 days in advance. CFSv2 has some problems with simulating downward wave propagation from the strat, but on a shorter range when thing are about to "blow", I think (I hope) it should see something. Some easy to understand info in here: http://www.cpc.ncep.noaa.gov/products/ctb/meetings/2012/CFSv2/Perlwitz.pdf I find it really interesting how ECMWF just doesn't want to buy this top strat warming. My money goes to ECM, due to its reputation, but my hopes are on GFS regarding this matter. But I think they will both correct themselves, ECM firing up a bit, and GFS maybe watering it down a notch or two.

I did a correlation between the stratosphere/polar vortex dynamics and North Hemisphere snow cover a while ago, how snow cover affects the stratosphere. But since the main question at the moment is how the stratosphere is going to affect us, I decided to do some reverse correlating. Usually we take a parameter like geopotential height for example, and we correlate it with certain indices, like ocean temperatures, wind oscillations, pressure oscillations, etc... and we apply a certain time lag to the parameter, to see how a certain index affects the parameter over time, to see how dominant it is, if at all. Main example being SST Indices like ENSO, and how and where they affect pressures rising or falling. But this time I decided to do reversed correlating, by correlating the index to a parameter. And since we are working around with the stratosphere, I applied a 1 month lag to the index, so the stratosphere activities happened up to 1 month prior to the calculation of the specific index. That should be a good time setup for SSW induced effects. Now there are 2 main problems. - I cannot choose specific years, meaning that I have to apply a certain time span in which the parameter and the index are correlated. The problem being, that SSW and non-SSW years get mixed up, and we get a very noisy correlation. I am sure that if I could apply specific years (at least 7 years with SSW events, because correlations of specific years or less than 3 years have no meaning), I could get a much more clear correlation and I could also make separate correlations for split and displacement events. - I can only apply monthly lag. I cant apply like a 2 week lag. 1 month is a long period in meteorology, so I could overshoot or undershoot a correlation maximum. With all that taken into consideration, I must say that in general, these correlations are nothing special and don't present any new ideas, besides from what has already been mentioned in here, and also by Chionomaniac a post or two ago. So despite the fact we can all read the papers (still trying to find time to do that, because I think I will soon have enough basis covered to dig into it), I sometimes like to go a step further, to "look behind the curtain". Down to the charts. I only found 2 indices that seem to have a better response to major strat. activity, more specifically polar vortex displacements. The TNA (Tropical North Atlantic) and AMM (Atlantic Meridional Mode). I think I could find more, but I would have to apply specific years. First up is TNA. You can see how noisy the correlation is, but there is a still defined signal for a P.V. displacement. So i could say that this basically represents a positive TNA mode in general. Now if we look how wintertime positive TNA correlates with 500mb geopotential heights. Next up is AMM (Atlantic meridional mode) correlation. Very noisy correlation again, but the signal is there. And looking at the right image, with only Jan. presented, the displacement signal is a bit more obvious. And how wintertime AMM correlates with 500mb heights. Now I must point out that both TNA and AMM are indices based on surface parameters like pressure, wind or SST. They are kinda similar in a way. But the point I am trying to make here is, that I was correlating stratosphere with indices that are semi-dominated by the 500mb pattern. So basically I used these indices merely as "agents" to correlate stratosphere and 500mb pattern. I could simply correlate the stratosphere directly to a geopotential height index like NAO, but the correlation is just too noisy and overgeneralised. So I basically "filtered" and bypassed that by correlating to specific indices that respond to patterns and then re-correlated the indices and 500mb height, to see the corresponding pattern for the specific phase of the index that correlated with the displaced vortex. I hope I haven't over complicated the explanation. So basically the real picture here is, that stratosphere dynamics, which in this case translates to an SSW, affect the pattern below. And that pattern semi-affects different indices. So by looking how certain indices and patterns changed after SSW events in the satellite era, you can go back before satellite era to reconstruct patterns using this key combinations. So now that you know all this (some already know all that ), I can show you the overall pattern resulting in respected AMM mode. Above is the positive mode and negative below. So to have a positive mode, you see what kind of a "global setup" is required, wanted or dominant for the specific mode. To filter it for you: the above correlation for displacement and positive AMM and TNA is basically a correlation between generally displaced polar vortex (probable SSW displacement) and the global setup you see in the upper part of this graphic. So I only used AMM and TNA to filter through all the general data noise. Furthermore, we can see negative AAM (Atmospheric angular momentum) above, and positive below. How about splits,I hear you ask? Well, it is hard extract certain SSW mode connected indices because the correlations are made over a certain time span (1980-2012 in my case), but there is one I can show you, so you can at least see how the split signal looks like. In this case, it looks quasi-connected to TNH (Tropical/Northern Hemisphere). The respected TNH phase translates into the 500mb pattern on the left. On the right, I added a composite of split SSW Jan/Feb. Of course there are differences between specific years, but just as the correlations are a multi year extraction, the same has to be applied to the corresponding season pattern composite. And besides that, we have here only one of many indices from which we could extract 500mb anomalies to make for more specific connections. Problems can emerge when it comes to practicality, because it is hard to generalise specific SSW modes response based on some correlations. Every event is specific in its own way, having slightly different dynamics, different global setups that lead to different tropospheric patterns prior to SSWs, and so on.... And as I have pointed out the two problems in the start, I cant fine tune the signal for either displacement or split. A much easier way is of course to just look at past events and use reanalysis to see how it resulted. But as Chionomaniac has pointed out, Greenland seem like an "average" target, regardless if we are talking about a SSW split or displacement. And of course I was only looking for some connections between SSW and 500mb patterns, and not really how exactly the SSW makes the pattern change. I think there have been many scientific papers linked already that explain that matter, and I must admit that I cant wait to go through all of em. I tried to read some, but I thought that it would be better to get some "first hand" basics first, so it will be easier for me to assimilate the info in the papers. We will see how this years events will turn out. The fact is that in this "age of technology", we have lots of good tools to analyse and predict these dynamics, so I think we can all learn a lot from these events, regardless if we are newcomers or if you are pros. Cheers.

WoW! Thanks everyone for a positive feedback on my GFS skill post. I really appreciate it. Its nice to see your work being accepted in the community. To throw a comment or two on the latest runs. We all see how things are shaping up, so I am just going to point out something interesting in the last runs, more evident in the last 12z. Something wants to sneak in through back door, to put it that way. Basically I am talking about the actual downwelling of warmth from the top, maybe less, maybe more expected. 288h-312h, blowtorching at the top. At 288h, we can see the main "warm core" at 3mb and some warmth around, but still no secondary warm core just yet. In the meantime at 5mb, the area of focus is chillin for now. Going further on, a secondary warm core starting to appear on 3mb. 7mb still chillin in that area at the time. At 360h, finally reaches 5mb. At 372h appearing on 7mb. And at 384h, some clues on 10mb, but this is too much of a detail to be focusing on at this time (384h). And the animation. Now I don't want go give the appearance of being too hasty. It looks good and all, but I don't want to give it too much credibility for now, at least not until it becomes a consistency. It is getting a bit more defined in the last few runs, with the 12z being the most "aggressive" on this downwelling so far. Another thing worth considering is the ECMWF. It clearly showed that its not buying the super warming from GFS in the uppermost stratosphere, at least not for now. And I think it is still too early to start speculating about possible effects on the troposphere if everything would come into fruition. Lets first get a clear picture what exactly is going to come into fruition. There is certain variance from event to event, but for a general idea, you can look at this composite from Reichler&Baldwin. And about the MJO, its in a weaker phase 3 for now, but the models are basically taking it all over the place. Tho there is not really that much similarity between the phase 3 and the current pattern and the pattern forecasted for the near future. Maybe in some points, but the phase 3 composite is basically just a composite, so we should look at it in the same spirit we are looking at ensemble means. In the meantime, 12z GEFS wants to bring AO above the "surface". If it really comes above in this time frame, I really hope the stratosphere slams it back down, in the whack'a mole style. (a bit of wishcasting there ) And just on a side note, it is interesting to see there was some upwelling recently in the ENSO 3.4 area. Best regards. Edit: GP beat me to the downwelling part.

I wrote a post or my opinion about GFS stratosphere accuracy a few days ago, but I will further address that matter. I will add skill scores and observed vs. forecast, for the PVO, or the Polar Vortex Oscillation. So what's this all about. This is a part of abstract from Ren & Ming: "The EOF and regression analyses indicate that the PVO can be described by a pair of poleward and downward propagating modes. These two modes together account for about 82% variance of the daily potential vorticity anomalies over the entire Northern Hemisphere. The power spectral analysis reveals a dominant time scale of about 107 days in the time series of these two modes, representing a complete PVO cycle accompanied with poleward propagating heating anomalies of both positive and negative signs from the equator to the pole. The strong polar vortex corresponds to the arrival of cold anomalies over the polar circle and vice versa. Accompanied with the poleward propagation is a simultaneous downward propagation." Basically we are going to look at GFS 16-day skill for PVO in two winters, 06/07 and 05/06 (which featured a split-type SSW). PVO is kinda like a polar vortex strength index, with positive values featuring stronger and organised vortex, and negative values indicate a weak vortex or an SSW event if the values are more negative (look at it kinda like a stratospheric AO). So PVO is derived from the polar vortex and is following its dynamics, so we cant really use the PVO index as a forecast index, because it only reflects polar vortex dynamics. But it is possible to try and forecast the PVO index. Sounds complicated? "It describes a polar vortex oscillation (PVO) between a strong polar vortex (a positive phase event of the annular mode) and a weak polar vortex (a stratosphere warming event ). In spite of using daily data, the time series of PVO (the PVO index) has a remarkable long time scale and exhibits only 1-2 PVO events within a winter season. The interannual variability of the PVO index is associated with both the duration variations of PVO events and their intensity variations." Ok. Down to the charts. So basically since PVO index is derived from the state of the polar vortex, a model needs to forecast the stratosphere polar vortex activities correctly if it means to accurately forecast the PVO. Logical, isn't it? First is the comparison between the 06/07 analysed PVO and 16-day forecasts (The 2006 looks like 2008, due to lower graphical resolution). Each red spot represents the forecast of the PVO for the specific date, issued 16 days (384h) prior to its analyse. And the average skill score for the 0-16 day period. Basically a 16-day (384h) skill of 0.8 is the same accuracy as the GFS and ECMWF have for the 500mb height in the 7 day range. The overall GFS skill for the troposphere in 06/07 Dec/Jan, was not really that impressive. Next up is the 05/06 winter, with an SSW in mid Jan-early Feb. The GFS 16-day accuracy of the PVO was quite phenomenal, to say the least, with very good prediction of the oncoming SSW. Notice the fast decline, corresponding to the P.V. disruption and an SSW. And the skill score. A score of 0.97 is really high, especially when considering the time span. When compared to 07, GFS had a better 5-day skill of the 500mb heights in the NH (20N-80N) in the period prior to the SSW (Nov-Jan). This is important, because it had a slightly better idea of the dynamics and consequently a better idea of the coupled trop-strat dynamics. This year, the model skill is fairly good, with ECMWF having a very decent average day-5 score. You can notice a skill drop in late November, around the time of the great Blocking/Highs moving around the pole and into the pole, giving models a bit of a hard time, figuring out their exact position. Keep in mind that this is a day-5 skill, meaning that the weak skills in the late Nov. belong to the forecasts issued/run in the mid November. Now I will try to write a summary based on all above. I am first going to add an abstract from Shin&Cai: "The NCEP operational GFS still has a remarkable prediction skill of (stratospheric) polar vortex oscillation (PVO) index at the lead time of day 16, despite a very poor skill in forecasting tropospheric circulation anomalies beyond day 7. The largest gain of prediction skill w.r.t. the persistence is over the stratospheric polar region. The remarkable skill comes from the signal of systematic poleward propagation of thermal anomalies from the equator to the pole in the stratosphere associated with the global mass circulation variability (intensity/time scale)" So to try and simplify all above: GFS stratosphere gph,temp. and wind forecasts in the 336h-384h period, are in average as "accurate" as the 500mb forecasts on day 5. I first wrote 6-7 days, but after doing some of this "research" I noticed it is actually better. The PVO is/was forecasted from more layers of the stratosphere, so basically it means that the forecasts for a single layer should be just as good. But as we have seen in just two cases, the skill or accuracy has its variance from year to year. It depends how well the model picks up the tropospheric dynamics and processes that affects the stratospheric polar vortex. So the GFS FI for mid and upper stratosphere would be,... well... lets say from 336h on. Now I am not saying that the strat. forecasts beyond 240 are still very accurate and reliable, but I am saying they are decent, because they usually pick up trends fairly good, and show highly likely scenarios. There is always variance from run to run in such a time extent, but looking in general, GFS has a good idea of the overall dynamics in the 12-16 day range. Of course I say overall, because details like warm core or cold core max/min temperatures are something to be more accurately shown under 192h range. And even tho the runs are jumping from split to displacement and vice versa in the 13-16 range, the overall idea is constantly the same, and it has little "delaying", meaning when the model shows a certain scenario in the works, but then it keeps delaying it, like it usually happens with snow scenarios down here in the troposphere. All non quoted text above and especially the summary, is just based on my observations and a little research. You can easily do some verification yourself. Just go a few pages back, like around 50 for example, and search for GFS charts for the 384h, and compare them with latest runs. If I can do one myself. Besides the slightly less warm Canada, the overall idea is very similar. And this is basically quite amazing, when considering the left graphic is a 384 forecast, and the right one only 72h. 13 days have passed, and a resolution increase, but the original idea was very much in place. The warming at this period seems to be even slightly more intense than the original forecast and the Canadian High a bit more defined. But given the fact it was a day-16 forecast on a low resolution, I think it did a great job, with very little delay, not moving the scenario away into the future. That's about it. I thought I was going to write this in like 30 minutes, but it actually took me 4 hours and 5 cans of beer. I guess that's the downside of being a perfectionist. But I must say that I really enjoyed every minute of it. I really like writing posts like this, because it also helps me to gain some more knowledge along the way. And I must add that I only have around 6 months in total of "stratosphere expertise" gathering, in contrast to some of the pros in here who are following stratospheric activities for a few years now. http://forum.nwstatic.co.uk//public/style_emoticons/default/smile.png As I said in a post a few days ago, this time last year I didn't even knew what and SSW is. I only had the idea that there is some sort of a cyclone high up in the stratosphere, but no clue just how important it is. So basically I am really trying to learn things fast to keep up with the pace and to try and compensate for my stratos. experience deficit. But a big part of my strat. education process is basically this thread itself. There are some really knowledgeable people in here that I learned from (and to which I would like to say Thank you), besides other sources. I guess my data assimilation system is much better than that one of the GFS. http://forum.nwstatic.co.uk//public/style_emoticons/default/biggrin.png I also have around 4 years of "general meteorology" experience, which helps me to understand things a bit faster and to put them into the big picture. Of course all those 4 years are sadly or luckily self learned too. I never had the will to go to "meteo school". The "learn through experience" motto worked fine for me I guess. I am going waay off topic here. Cheers. http://forum.nwstatic.co.uk//public/style_emoticons/default/smile.png

I almost feel sorry for the guy. (the vortex that is)

192h sounding on the exact location where EC had its Tmax at 10mb.

"The epic battle began. Powers collided, with no end in sight. Which one shall prevail? Which one shall dominate? Which one shall overcome the other and rule over the north?" I just had to add a little Recretos' touch of epicness.

Well, my opinion is, that given all the model output so far, and the reanalysis done, correlations, and overall patterns setting up, a displacement type SSW (@10mb) is more likely than a split type event. And in your chart, the heat influx would more than likely split the vortex. One of the best or "textbook" split SSWs recently, was the 2009 event. I used ERA, because it basically looks a lot better than the standard NCEP reanalysis.

Introducing, stratosphere forecasted soundings. FIM model, 192h. Location: In close proximity of the warm core at 192h. Nothing special of course, but looks interesting. Will be even more interesting, once more decent warming gets under 192h timeframe.

The way I understand this, is that they took this(example) --> , and put it in a 3D mode, basically giving it spatial presentation, to see not just how strong the EP Flux is, but where exactly is it occurring, prior to the respected SSW mode (split / displacement). How couldn't I see that before? I guess actually reading the paper might help me. So I guess that the pattern chats and the WAF charts from Cohen, are basically in the same "family" so to speak. His EP flux charts should indeed correspond with the pattern constructs prior to SSW modes. Imho of course.

@Lorenzo: Well, by looking at it and the units (m2s-2), I don't think there is a pre-defined option for this in the reanalysis tools. To do an exact same plot, you would have to download datasets and basically manually "write" the algorithm for this parameter and plot it. I am not really sure how to get close to this. Maybe with some combination of vertical pressure gradients, Omega and/or 100mb eddy heights. But I think that is not really close to it. I understand it as a vertical wave propagation, but given the derived unit m2/s-2 this also has to do something with kinetic energy. Or maybe my physics are a bit rusty. I only had physics for 2 years in elementary school, so... yea... I am lacking basic theory in the area most important (or connected to) in my main interest, which is meteorology. Perhaps 250mb streamfunction can be helpful. Maybe GP as a pro, can shed some light on this.

Judging by your description and my experience, it was 100% a positive CG lightning discharge. A quite common type of lightning for cold part of the year, or final stages of thunderstorms. And about the damage, I don't think one bang will do you much harm. I've experienced many close bangs, some made me deaf for couple of second, yet my hearing is still great, or so I was told by the doctor 6 months ago, when I had to take a full health check for my new job.

This chart probably doesn't belong in here, but just to give some support to Mr. Chionomaniac. My opinion: Having a negative AO going into the timeframe of major stratospheric vortex disruptions, is nothing but positive (no pun intended) in my view. EDIT: I was playing around with the reanalysis tools and I made an analog of January 500mb anomaly with conditions: December 30mb CW (Canadian Warming) in an east QBO and a Polar vortex disposition in December. Also added is the SLP anomaly.

Its not any of my business, but shouldn't you wait for December to end, before bashing analog forecasts? (I'm speaking in general and not to a specific person) GP's analog forecast seems legit to me. My analog is basically quite similar (posted in the seasonal forecast thread). There are some differences regarding the blocking, which is fairly weaker, mainly because it is not present throughout the whole month. And you should look at these analog forecasts the same way you look at ensemble means. There is a respected deviation. I don't remember which years I've put into the analog picture, and I have edited the originals (Stupid me, I know). But that doesn't really matter. Basically, I am in "GP's camp" with my analogs. Best regards.

About that: CFS is propagating the Novaya Zemlya blocking west and into Scandinavia, weakening it along the way and forming a new one just W/SW of Greenland. Basically quite the idea that ECM is also promoting. About 1968/69. This was the December pattern. This is December so far, and the ECM ensemble forecast. And the strat+AO comparison between 68/69 and this year. Best regards

@Sundog: Well, actually they are. The troposphere is much more complex. You have terrain interaction, not to mention the water, then a very complex Rossby wave global system, different forcing from the tropics, different factors on the meso-gamma and meso-delta scale that affect the overall picture, etc,... Basically a ton of variables that affect the weather and which must be taken into the calculations of a model. While the stratosphere is in a certain way much less complex, basically on a "flat surface" (so to speak) with less intense activity, and the features being on an overall large scale (synoptic-global), while troposphere has features all the way down to the storm scale. Of course I simply have to mention the troposphere-stratosphere interaction. Basically the activity in the troposphere, reflects in the stratosphere in a certain way or through succession. (it can also be the other way around, which is basically the primary reason why this thread is up and running ). So looking from the model perspective, it is easier to forecast stratosphere activity, than the troposphere activity. This is the "skill" (Anomaly Correlation) of the global models on day 6, for the last month, with ECMWF in the lead (as always). For the mid and upper stratosphere, I would say the day 10 skill score would be around 0.94 and around 0.88 from day 11 to 13 in average, and around 0.82 from day 13-15. Just my assumption. Nothing official. This is a summary from a report from Shin and Cai. http://www.esrl.noaa...ic_modeling.pdf "The NCEP operational GFS still has a remarkable prediction skill of (stratospheric) polar vortex oscillation (PVO) index at the lead time of day 16, despite a very poor skill in forecasting tropospheric circulation anomalies beyond day 7. The largest gain of prediction skill w.r.t. the persistence is over the stratospheric polar region. The remarkable skill comes from the signal of systematic poleward propagation of thermal anomalies from the equator to the pole in the stratosphere associated with the global mass circulation variability (intensity/time scale)" So to simplify: GFS stratosphere gph and temp. forecasts in the 300h-384h period, are as "reliable" as the 500mb forecasts on day 6-7. Maybe some of the pros in here can add a thing or two on this subject Best regards.

In high resolution. Presented in color are sea level pressure anomalies, and mean values with lines. a,b,c - prior to vortex displacement, d,e,f - prior to vortex split 500mb pattern and MSLP forecast, for comparison. Regardless of the latest GFS runs slightly reducing the warming intensity, I still think that the runs are great, because of the continuing trend of the weakening cold core.

http://www.meteociel.fr/modeles/gfse_cartes.php?ech=6&code=code&mode=10&mode3h=&runpara=0&carte=1 http://www.instantweathermaps.com/GFS-php/showmap-strat.php?run=2012121512&var=TMP&lev=10mb&hour=384

Well I wasn't really referring to it as a bad thing, but just keeping it in the picture, so people don't expect the forecasted warming at upper and mid strat to have an immediate effect on the tropospheric patterns.

The 1400m wave 1 would certainly surpass the last one. But even if everything goes as planned and in the best way possible, we would still have to deal with the lag. Like Chionomaniac has pointed out yesterday, and as it is shown in this graphic from T. Reichler.