I have been taking a look at what has happened to the CET in the period after significant past SSW events to see if there is any trend worthy of note. Of course it is important to point out up-front that there are many factors that influence temperature and so any consideration of just one or two variables cannot in and of itself suggest that those variables may be the key influence of temperature differences.
In order to identify significant SSW events I made use of the following article from 2007 by Charlton and Polvani. This has a very handy table of SSW events on page 453 covering the period 1958 to 2002.
Charlton and Polvani article
I then turned to other online sources to identify SSW events since 2002. In particular this site by Patrick Martineau was very helpful. I have only used events that, on his index, are -2.9 NAM or less.
http://p-martineau.com/ssw-animations/
This gives me the following table of SSW events since 1958. The first column shows the central date of the SSW (or in the case of events since 2002 the date of the weakest Polar vortex state). The next column indicates whether the SSW caused the Polar vortex to be merely displaced or actually caused a split in the vortex. The third column shows whether the SSW event was a relatively weak one or a relatively strong one.
At this point it is worth referring back to the Charlton and Polvani paper. If you go to page 463 and refer to diagrams a) and b) these show the 1000hPa geopotential height anomalies for the 60 days from the central date of the SSW event. Chart a) is for vortex displacement events and b) for split vortex events. What we see is a clear positive height anomaly over the Pole and a negative anomaly over the Azores suggesting an anomalous easterly flow for the UK. The anomalies are higher for split vortex events. The data seems very clear on this.
So if we then go back to my table I was interested to see what effect this anomalous easterly flow had on the CET. So I compared the actual CET for each of the 60 days following the central date of each SSW event to the 1971-2000 mean temperature for each date. This gave me an average CET anomaly for the 60 days which is presented in column 4 of the table. Column 5 then gives the number of days within the 60 day period where the CET was more than 2C below the 1971-2000 mean.
The results show that the CET mean is, on average, about one-third of a degree below the 1971-2000 mean in the 60 days following an SSW event. This is not a very large anomaly but is still interesting. There seems to be no difference in the temperature anomaly between vortex displacement and vortex splitting events.
There is, however, a larger temperature anomaly with strong SSW events compared to weaker events. For these purposes I define a strong event as one where the temperature anomaly in the final column of the table on page 453 of the Charlton and Polvani paper is +8 Kelvin or greater. For later events I take SSW events shown in the Martineau paper with a minimum NAM figure of more than -4.
I have also analysed the temperature impact by the month in which the SSW occurs. What we see here is that the earlier in the winter the SSW occurs the higher the temperature anomaly on average. For those SSWs that occurred in March (of which there are not many) the temperature anomaly was in fact on average positive. The vast majority of SSWs occur in January and February. In recent times almost all have occurred in January.
I then got to thinking about what other factors might significantly influence the temperature changes that might be triggered by a SSW. An obvious factor is the QBO. As noted above, a SSW typically results in an anomalous easterly flow over the UK. If this occurs during a westerly (positive) QBO phase then the easterly flow that the SSW attempts to trigger may well be offset partly or entirely by the westerly QBO effect. Conversely during periods where the QBO is easterly (negative) this is likely to be enhanced by the anomalous flow resulting from a SSW.
When we look at the temperature anomalies split by QBO phase we see a very significant result, albeit this is not really surprising. The final column of my table gives the QBO figure for the month in which the central date of the SSW occurs.
When a SSW occurs during a month where the QBO is negative, on average temperatures in the following 60 days are above average suggesting the positive QBO is more than offsetting any impact from the SSW. When the QBO is between 0 and -8 there is a negative temperature anomaly broadly equal to the average anomaly for all SSW events.
When the QBO is lower than -10 and a SSW occurs the resulting temperature anomaly for the next 60 days is on average about -1C which is very significant indeed.
Again I repeat my earlier comment that there are many influences on temperature so one has to be very careful in drawing any conclusions. As can be seen from the data in my table even in years with a very negative QBO and a strong SSW, the CET in the following 60 days has been warmer than average in some instances (the best example of this being January 1977).
Nevertheless the data would suggest that a major SSW event in early to mid-winter during a strong easterly QBO phase does increase the likelihood of temperatures being below average in the following 60 days.
One final point to highlight is that the temperature figures in my table are averages. I am not saying that a SSW can give 60 days of bitter cold. On the contrary after all such events the temperature patterns are very variable as they always are because of the multitude of any factors that impact on our weather. So even when we have experienced spells of very cold weather following a SSW there are periods of warmer weather as well when temperatures are above average. It is just that overall over a 60 day period temperatures tend to be somewhat below average.
So if we now look at 2014/15 and try to apply some of the thinking above what do we see? Well as we know the QBO is very heavily negative at the moment (below -20 in fact). So that is obviously helpful. The models are showing a potentially significant warming event developing in the stratosphere over the next 10 days.
Major warming over Siberia has already displaced the Polar Vortex somewhat although the vortex itself is still intact and circular in nature at the moment as shown by the 18z GFS run.
http://modeles.meteociel.fr/modeles/gfs/runs/2014122718/gfsnh-10-18.png?18
As we reach the end of the high resolution section of the GFS run we see the Polar Vortex starting to split in two. By T240 the split is complete
http://modeles.meteociel.fr/modeles/gfs/runs/2014122718/gfsnh-10-240.png?18
The 18z GFS(P) run does not quite manage a split vortex (although previous runs have). The vortex is heavily displaced though but does try to reassert itself by T240.
Anything beyond this is way into FI and in low resolution so should be ignored. A splitting vortex has been modelled fairly consistently for a few days now so it is conceivable that a SSW could occur during the first half of January but far too early to be cerain. We will need to see wave breaking as well in order for the cold air from the Pole to be displaced south into the mid latitudes. But the building blocks are there so something to keep a close eye on in the next 10 days.
Finally, on a separate point I have been looking back at some analysis I did in November on CET analogue years. The best fit now we are approaching the end of December is in fact 1984/85. The data is indeed a remarkably good fit since about 8 December.
Now of course January 1985 was very cold indeed and at the moment it seems unlikely that January 2015 will be as cold as this. But I would not rule it out, especially if a SSW does occur. The chart below shows how the daily CET for October to December in 2014 compares to 1984.
Interesting times ahead I think. This winter is just getting started weather wise and I would be surprised if we do not see some breaks in the more zonal flow during the next couple of months.
Edited by user
28 December 2014 07:13:44
|
Reason: Not specified