Winter 16/17


Three consecutive winters with a mild, stormy & snow less start, gives way to something a little different at last.

But just how different?

greeny EG

Long range (LR) forecasting is experimental. No LR forecast can tell you when it will snow, nor can it tell you weeks in advance when a two day northerly cold spell will hit the UK. But assuming the key drivers are correctly identified and enough weighting is given to factors that may well skew prolonged cold chances in Europe, then this forecast will try to pick out periods of winter when prolonged cold is feasible. It won’t be until later down the road, during winter, that we will truly be able to ascertain the extent of any cold air advection to the UK, and of course any subsequent snow risk.

As I’ve done since the first forecast in December 2013, I will lay down the key drivers with a brief explanation. Each section is then scored in an attempt to clearly demonstrate the influence that each driver has on cold winter weather prospects (the scores are more focused on UK). The forecast then follows the conclusions at the end.

As usual, no seasonal model guidance has been factored into this forecast.

If you want to skip the long technical part, please scroll down to the conclusions and then the forecast below that.

1. The stratosphere (polar vortex), QBO, solar activity & ozone.


Before we get into this, I will briefly explain the significance of the relationship between the troposphere and stratosphere and the implications for winter weather.

Each winter the stratosphere cools significantly. The difference in temperatures between the Arctic and warmer latitudes further south results in the formation of a strong area of low pressure, called the stratospheric polar vortex. Below this is the tropospheric polar vortex (the area of low pressure located around the north pole that we see on weather charts). The colder the stratosphere is, the tighter/stronger the stratospheric vortex becomes and consequently, the tropospheric vortex too. The strength/position of the tropospheric vortex influences the AO (Arctic Oscillation), which is a measure of pressure between the north pole and the Azores, as shown in the image below, courtesy of NASA.

ao infographic

A stronger vortex can result in a positive Arctic Oscillation which, for Europe, equates to a less cold winter. The opposite is the case with a weaker vortex.

Sometimes, during winter, a strong vortex is put under pressure, warming it and displacing cold air from the pole down to the mid latitudes. Rossby/planetary waves circumnavigate the globe and during winter, when a powerful large wave encounters a mountain range (eg. Himalayas, Rockies, Andes), if the wave is large enough, some energy is deflected poleward (mountain torque). It needs to be a sizeable deflection to achieve this, but these waves can penetrate into the stratosphere, creating a warming disruption to the otherwise usually cold and stable wintertime stratospheric environment (think of the sea-shore where waves break all the time but only the strongest will push more inland resulting in coastal erosion).

There are two main types of disruption to the vortex via this process; a displaced vortex, where wave breaking and consequent warming moves the core of the vortex away from the pole; or a split vortex where the vortex is put under even more pressure and is split in two. These events are often referred to as sudden stratospheric warmings (SSW) where the zonal winds at 60N/10HPA are reversed from westerly to easterly. In both cases, the warming and movement of the vortex, pushes cold air into the middle latitudes as higher pressure builds over more northern latitudes. Generally speaking the number of the wave (1,2,3) refers to the number of waves at that time. Wave 1 usually displaces, and strong wave 2 can cause splits. Following a displacement, if wave activity subsides, it is common for the vortex to fairly rapidly regroup and cool. Following a split, if wave activity wanes, it can take much longer for the vortex to recover and regroup. So a split vortex is the ideal scenario we are looking for and historically the UK has benefited more in terms of
prolonged cold weather from splits rather than displacements. Split vortexes can also lead to faster response at the surface.

1.1 The QBO (Quasi-biennial oscillation)

The QBO is a measure of wind flow across the equator high up in the stratosphere (measured at 30mb). There are two phases of the QBO, east and west – referring to the direction of those winds. These cycles or phases last roughly 18 months or so. The QBO has a significant effect on the state of the polar stratosphere during wintertime, and is therefore of much interest for the winter forecast.

We began a West QBO phase in the summer of 2015 which continued through the winter. The expectation was that we would then see a transition to an East phase during 2016, but that was not the case. Instead, we have eventually seen a west QBO re-emerge & slowly strengthen during the year. In November 2016, current QBO values are very similar to November 2015, apart from easterly winds still present at the top of the stratosphere. This has not been recorded before, at least going back to 1948. The 1950’s recorded extended periods of east QBO phases, with 39 consecutive months of E QBO between 1948-1951 . In terms of west QBO, the longest duration in the dataset is 16 months. That record has already been beaten, with November 2016 being the 17th consecutive month, and a west QBO is expected to dominate through the coming winter.

wave example

The abundance of high pressure over the poles this Autumn is also very much against the tide of historical west QBO analogues.



Possible implications of West QBO for UK winter 2016/17?

A west QBO gives that bit more credence to the idea of vortex strengthening further into winter vs east phase. (Winter 0 Anti Winter 1)

1.2 Solar Activity


After a recent weak solar maximum, activity is slowly weakening as we head towards our next solar minimum in 2020.

Recently, we saw solar activity drop quite significantly before climbing back towards average/forecasted levels.

Geomagnetic activity usually lags behind, with increased activity in declining phases of (even numbered) solar cycles. It is quite likely that we will see further spikes in activity over the winter.  Indeed, such an occurrence is already looming, with spikes forecast for second week of December, as also tentatively end of week 3 into week 4 December.

Labitzke et al have published several papers alluding to 10.7cm solar flux relationship with the stratosphere and northern hemisphere winters, specifically drawing attention to an increased incidence of SSW occurrence in west QBO winters when the 10.7cm flux is above 110 units.

The image below (Labitzke & Kunze) shows this relationship between solar flux, the qbo and SSW’s clearly. In the sample of 65 years (1942-2007), there was only ONE (two, see text below image) SSW during a west QBO winter when flux was below 110 units (February 2007).

Incidentally, there was another in early February 2009 which is not covered in this graphs dataset, with flux around 75 units. The key point is between 1942-2007 there were 11 SSW’s during west QBO winters with flux above 110 (and 10 of those, the flux was above 150 units).

Flux has been bumping around the 70-100 mark recently and is more than likely set to continue in this fashion for much of winter.


Implications on winter 16/17?

Solar flux & W QBO relationship suggests less chance (but not no chance) of mid winter SSW but that’s for later in winter (see conclusions).

However, until we get closer to that solar minimum in 2020, volatile solar activity has to be treated with due caution, as I’ve done previous 3 winters. (Winter 0 Anti Winter 2)

1.3 Ozone/BDC (Brewer-Dobson Circulation)


Levels of ozone concentrations are also known to have an influence on stratospheric temperatures. The BDC refers to the transport of ozone from tropics to pole during Autumn into the winter. The phase of the QBO is also influential here. During east phases, the tropical stratosphere is cooler than average and the polar stratosphere warmer, because higher levels of ozone release heat into the surrounding air, sharpening the thermal gradient, reducing polar westerlies which in turn can lead to a weakened polar vortex.
During W QBO phases, a generally weaker BDC is more typical, with ozone concentrations tending to linger in the mid latitudes. We have already identified above that this is far from a typical W QBO phase, and this is evidenced further using current tropical stratospheric temperatures (these giving us an indication as to how strong the BDC is). Temperatures have been below average since around September, and closer to the 1979-2015 minimum.

This suggests that despite the west QBO, ozone transport via the BDC is strong this Autumn.


Implications of BDC on winter 16/17?

An active BDC from Autumn into the winter bodes well for northern blocking later in winter.

(Winter 1 Anti Winter 2)

2. ENSO (El Nino Southern Oscillation)

Enso refers to the warm and cold phases of the waters along the equatorial Pacific. Warm being El Nino, cold phases are La Nina.

Climate influences of the warm and cold phases of ENSO have more pronounced implications on weather in the tropical regions, but do still influence weather patterns globally. The effects of weaker or neutral ENSO events in Europe are smaller and can often be overridden by stronger signals elsewhere. In last years winter forecast, i spoke in detail about the evidence for influences in Pacific ENSO region to Europe, via the stratosphere vs more direct tropospheric led weather patterns arising from strong ENSO events. The evidence for that stratospheric pathway during weak/neutral ENSO events is far less clear & isn’t given any weighting in this years forecast.

The tropospheric pathways are, however, crucial this season.

After 2015 – early 2016’s very strong El Nino (first image below), global SST distribution is vastly different this season (second image).




In terms of sea temperatures, we have been at weak La Nina levels for a while now, however, such was the legacy of those exceptionally warm equatorial sea temperatures, that residual atmospheric El Nino tendencies have persisted for much of 2016, with a notable disconnect between ocean and atmosphere developing, as is usually the case after lagged strong ENSO events.  These residual El Nino atmospheric signatures will only slowly rinse out of the atmosphere, whilst at the same time an increasing tendency for La Nina and atmospheric coupling, manifesting in terms of notable pressure patterns, namely North Pacific & mid Atlantic high pressure cells.  The autumn disconnect is clear to see using global wind analysis as shown below (globally averaged angular momentum).


…which is not too dissimilar to the evolution through 1983, as shown below (AAM)


2.1 November Patterns

Sea level pressure during November also highlights that disconnect between ocean and atmosphere



Latest ENSO forecasts tentatively suggest that we may have seen the extent of La Nina for the winter season. ENSO neutral/weak La Nina conditions will persist for winter 16/17.

Combined with the west QBO, the expectation is for a more typical La Nina atmospheric base state to develop over winter, with ridge becoming dominant in the North Pacific.

Analogues based on similar El Nino to La Nina transition yield the following results:


December ENSO


January ENSO


February ENSO

2.3 Azores high pressure (mid latitude ridges)

When La Nina and atmosphere are coupled, we see increase in easterly wind flow which has the effect of agitating mid latitude ridges. We will continue to see the waxing and waning (but increasingly waxing) of this coupled environment through winter. It is therefore likely that we will see high pressure dominating SW Europe and ridging further north and east into Europe at times during periods of low angular momentum (La Nina). It will only be via feedback processes involving stratospheric disruption/forcing in the troposphere that a more amplified Pacific and Atlantic profile may be attained. This having the effect of weakening the influence of Azores HP over SW (west) Europe, tilting on a more NW/SE axis, similar to the January composite above.

Implications on winter 16/17?

Composites show a pretty cold pattern emerge for January, albeit the UK is closest to that mid Atlantic ridge. However this will be reliant of forcing and feedback loops described above.

(Winter 2 Anti winter 2)

2.4 MJO (Madden-Julian Oscillation)

During November we saw a high amplitude convective tropical wave pass through the W Pacific, slowly fading into the Indian Ocean. Previous years with a similar evolution, combined with similar background states, suggest a propensity for tropical convection to be rooted to the West Pacific/East Indian Ocean during December, albeit fairly weak. Rolled forward to January, however, an active, high amplitude wave was present in almost all those years.

So whilst a fairly incoherent MJO signal is likely to persist for the bulk of December, there is an expectation for eventual renewed activity, which will increase pressure on the polar vortex and amplify global circulations, as tropical forcing progresses east.

(Winter 3 Anti winter 2)

2.5 PDO (Pacific Decadal Oscillation)

The PDO is detected as warm or cool surface waters in the Pacific Ocean, north of 20° N.  The phases of the PDO are known to influence global sea level pressure patterns, amongst others.  Historically, the UK’s coldest winters have coincided with low solar activity and cool phase of the PDO.  There are exceptions, but the general theme is consistent.

A cool PDO and low solar output certainly do not guarantee cold winters, but when combined with other variables (eg. east QBO, low sea ice, rapid Oct snow cover/advance etc) they can increase the chances of a negative Arctic Oscillation/NAO and a cold winter for Europe.

The phases last roughly 20-30 years. Despite the main phases of the PDO, there are still
cool PDO years within warm phases and vice versa.  We are currently in a cool phase which we entered in the latter half of 2007, but since January 2014 we have seen +PDO, which slowly built to record values in January 2015 – a notable departure considering we are in a cool phase. +PDO often sees more in way of wave 1 disruption vs wave 2.

Current values are around neutral to slightly negative & the coolest since the end of 2013/early 2014.  PDO during 2016 has been very warm and any cooling off now is unlikely to be of much influence to winter forecast.

(Winter 3 Anti winter 3)


3. Snow cover

Cohen et al’s work on snow cover/extent is well documented.
To summarise, the rate of Eurasian snow cover during October is linked to upper and surface weather patterns during winter, particularly January. During Octobers where snow cover rate is higher than average, distinct feedback mechanisms are observed over the course of the following 3 months.

High snow cover increases diabatic cooling aiding formation of a strong Siberian high pressure cell, which fluxes energy poleward into lower stratosphere. This energy flux (see stratosphere and waves section for more), disrupts the stable stratosphere, causing strong stratospheric warming which then feeds back to the troposphere in the form of high pressure at northern latitudes, pushing the jet stream south. The high pressure/warmer air at higher latitudes effectively squeezes the cold air locked over the pole south towards mid latitudes. Resulting in a typical -AO pattern. The following infographic (Cohen, 2014) explains the processes and feedback mechanisms involved.


Reanalysis of high snow extent October years reveals high pressure at northern latitudes to be a distinct theme and January in particular showing a strong -AO, as shown below:

high snow

October 2016 saw well above average snow advance, the second highest on record.

3.1 Sea Ice

Reduced Arctic sea ice, primarily in August & September, influences tropospheric and stratospheric circulations and has been linked to a propensity for -AO conditions in the following January (Overland and Wang 2010).
Cohen, Jones, Furtado & Tziperman have linked this sea ice loss in late summer to increased snow cover in October.
This years sea ice is well below average.
For the purposes of this forecast, snow cover and sea ice extent are scored together.

(Winter 4 Anti Winter 3 Final score)

Conclusions and summary

  • We have a weak polar vortex, very weak indeed for time of year.
  • Abundance of Northern blocking has been quite remarkable during Autumn.
  • I have identified a likelihood for vortex intensification henceforth but it’s very difficult to ascertain to what extent.
  • snow cover/ice extent, BDC, and analogue based MJO & ENSO analogues point towards cold spells developing during January (Europe in general), but these are offset against the West QBO and volatile solar activity during the declining phase of solar cycle.
  • SSW is a low chance but if it occurs, it would unlikely to be before mid January (analogues)
  • there is no signal for the type of long duration exceptionally mild and stormy winter seen in recent years.
  • Analogues quite bullish in demonstrating Atlantic/Azores ridge close to or over UK at times
  • The coldest winter in some years for eastern and NE Europe seems entirely possible
  • SW Europe is likely to see mild winter overall
  • For the UK, an average winter with below average precipitation and some cold spells at times (most likely in January) seems the most sensible call.

The Forecast

December into January

Confidence is quite low throughout this very tricky seasonal forecast. The December forecast is based on strong positive mountain torques, potential stratospheric disruption and tropical forcing but against the backdrop of strengthening polar westerlies (PV intensification highlighted in write up) and solar spikes.

Frequent cold spells for eastern and NE Europe with stark temperature contrast compared to milder SW Europe. Eventual HP over NW Europe, perhaps migrating north (NW) with time.



From mid January, winds from the North or NE…



Should we see a SSW in mid January, this forecast will need to be changed and trend colder, especially first half of the month.

But as identified in the write up, SSW chances look low this season, and more than likely will come down to what extent the polar vortex has strengthened before the crux of stratospheric disruption occurs mid January.


Please follow on Twitter for continued updates during the winter.