During the winter of 2018/19 ENSO moved from near-neutral to moderately positive, which led to widespread forecasts of a cold February, as easterly and north-easterly types tend to be more common in February during El Nino years. I was one of those who was fairly confident of a cold, blocked February. December 2018 was mild, but a sudden stratospheric warming event started around 22 December, adding to the suspicions that the mid to latter part of the winter would be cold. The first half of January was generally anticyclonic and fairly mild, but the second half had some cold snaps, mainly from north-westerly winds, and there was widespread snow on the last few days.
Towards early February, many of the forecast models were hinting at the onset of an easterly type, but then they spectacularly backtracked and the month ended up dominated by southerly and south-westerly winds. In most areas February was particularly unusual for its combination of exceptionally warm daytime temperatures (including maxima of up to 21C in the London area), bright sunshine, and high diurnal temperature ranges, although the relatively cool nights prevented it from being the warmest February on record. Possible explanations for the failure of the cold easterlies to materialise included an exceptionally persistent polar vortex over Canada, and a tendency for the atmosphere to stay in a relatively La Nina-like state despite the arrival of weak El Nino conditions. The winter illustrated that there is still a lot more to be learnt in the realm of long range forecasting.
Forecasts are mostly going for a weakly positive ENSO state early in the season, followed by a shift towards near-neutral ENSO towards the end of the season. This is reflected by both the NMME ensemble (below) and the Met Office ensemble, although the Met Office ensemble gives a higher probability of the ENSO state remaining weakly positive.
ECMWF ensemble has greater support for ENSO dropping to a neutral state. Uncertainty increases sharply after December 2019, but with high confidence that ENSO will probably be between +1.0 and -0.5 come the end of February.
The nearest analogues for the current situation are 1981, 1991, 2004, 2013 and 2015.
The overall sea level pressure anomaly chart shows an anomalous north to north-westerly flow over the British Isles, indicating a mean westerly flow over the country that is rather veered from normal, i.e. with less of a southerly component than usual. However, anticyclonic west to north-westerly types tend not to be especially cold. The Decembers of these winters had a strong tendency for an anomalous north-westerly flow with a strengthened Azores High and low pressure over Scandinavia. The Januarys of these years had a mean anomalous north-westerly flow over the British Isles with below-average pressure to the north and east, while the Februarys tended to be anticyclonic with a reduced westerly flow relative to normal.
The QBO (Quasi-Biennial Oscillation) is currently in a westerly phase, which means a reduced likelihood of northern blocking and sudden stratospheric warmings, although the extent of the westerly phase is relatively weak and has also been weakening slightly over time since June. This suggests that we are less likely than average to see a significant sudden stratospheric warming event this winter (which means reduced chances of cold snowy weather developing in January and February).
The Madden-Julian Oscillation is currently in a low amplitude phase, but is forecast to head towards phases 3 and 4 towards mid-December. Phase 3 is correlated with anticyclonic weather, and phase 4 with an enhanced westerly flow over the British Isles. Generally the phase area associated with cold snowy outbreaks is 7-8-1, but it is impossible to predict at this range when the MJO will move into that phase area, and whether or not it will be of high amplitude when it gets into that area.
Arctic Sea Ice
The Arctic sea ice extent is relatively close to the long-term normal on "our" side of the Arctic, from Greenland across to the Kara Sea, which means that northerly outbreaks are likely to be relatively potent by recent standards, although probably still less potent than prior to the step-change in the Arctic climate around 2005. In other parts of the Arctic, particularly around the Chuckhi/Bering Sea area, sea ice extent is anomalously low. At this time of the year, the Arctic sea ice extent was very similar in 2010, 2014 and 2017.
The three winters were very different, with 2010/11 having a record-breaking cold December, an average-ish January and a mild February, 2014/15 being mostly westerly/polar maritime-dominated, slightly milder than average but with some cold snaps, and 2017/18 having a fair amount of cold snowy weather in December and February, but a mild January (although January had a snowy "westerly" third week in Scotland and northern England). To varying degrees, though, they all bore out my earlier comment about northerlies being relatively potent when the Arctic sea ice edge in the Greenland/Barents/Kara Sea region ends up close to its long-term average position, and 2014/15 and 2017/18 were also noteworthy for featuring marked examples of "cold zonality" with fairly widespread snowfalls from polar maritime westerlies.
Atmospheric circulation, temperature and rainfall forecasts
The Met Office model ensemble mean is pointing towards a slightly milder than average winter (by 0.5-1.0C), exceptionally mild over inland northern Eurasia, but with parts of the Arctic having relatively close to average temperatures for recent years. Most striking is the mean sea level pressure anomaly map which suggests an enhanced westerly flow over the British Isles and a western European high. To my mind, this is consistent with a rather milder winter for Britain than the ensemble mean temperature anomaly shows, although the enhanced westerlies are consistent with the projection of extensive anomalous warmth over northern Eurasia). Precipitation is forecast to be a little above normal. The long-range projections into spring have the high pressure anomaly heading further north, perhaps suggesting a tendency for more anticyclonic weather during February.
The ECMWF ensemble's projected temperature anomaly patterns are remarkably similar to those of the Met Office ensemble, but its mean sea level pressure anomaly chart shows a more cyclonic tendency over the British Isles with higher than average pressure centred further south, over north Africa and southern Spain and Portugal, which would mean a more unsettled type of westerly-dominated winter than the Met Office ensemble's sea level pressure anomalies would suggest. Again I am surprised, given the ensemble mean pressure forecast, that the winter is only forecast to be 0.5-1.0C warmer than average over most of the country, and less than 0.5C above in the north and west of Scotland, but the sea level pressure forecast is again strongly consistent with exceptional anomalous warmth over much of inland Eurasia.
The long-range CFS forecast from NOAA is also pointing towards a milder and wetter than average winter, but with disagreement between the three sets of ensembles (the coolest have the winter about 0.5C warmer than average, the warmest have it 1 to 2C above average over England and Wales, though still just 0.5-1.0C warmer than average in Scotland and Northern Ireland).
The most up-to-date Netweather.tv atmospheric circulation forecasts go for below average pressure in December, nearer average pressure in January, and a close to average strength of mean westerly flow in both months. For February, in contrast to other models and signals, the model goes for below average pressure and a reduced westerly flow relative to normal, indicating a southerly tracking jet stream and above average precipitation.
The JAMSTEC model, unlike other long range models, is going for a colder than average winter over Britain with above average precipitation over central and southern areas, possibly indicating a continuation of the southerly tracking jet stream that we've seen from late September through to the end of November 2019. CANSIPS is at the other end of the spectrum, forecasting a stronger than average westerly flow in December, January and February, and a cyclonic tendency in January and February. Unsurprisingly, it also forecasts a mild and wet December and January, though with precipitation closer to average in February.
Winter Forecast Conclusions
There is reasonably high confidence that December and January will be dominated mainly by westerly winds, bringing changeable weather and generally mild temperatures especially to the south. There is less confidence in the forecast for January than for December, but the signals point mostly to a continuation of the west to north-westerly type, though January may be a little less unsettled overall than December.
However, there is also a likelihood of short-lived cold snaps, especially in the case of December. These will come mainly from north-westerly and northerly (rather than easterly) winds and will affect northern areas most frequently. With the Arctic sea ice edge looking set to be close to the long-term average in the Barents and Kara seas, any northerlies that we do get will tend to be relatively potent for recent years. As a result, snow events will probably be quite frequent in the north, but lying snow at low levels will tend to be short lived due to the mobile nature of the pattern which will lead to alternating mild and cold spells. There is less confidence in southern areas seeing much snow at low levels, but it cannot be ruled out, as was illustrated by the widespread falling and lying snow from a predominantly west to north-westerly type, even in low-lying parts of the south-west, in late January 2019.
For February there are signs that the weather will turn more anticyclonic, meaning that it will generally be much drier than December or January. There is not much of a signal for extensive blocking to the north of Britain, which implies that February will probably end up as a relatively mild month, though probably nothing like as exceptional as February 2019. With the weather expected to be both dry and fairly mild, it will probably also be less snowy than average. Confidence in the February pattern is relatively low, though, as there are some signals that keep the unsettled weather going through February and have the weather settling down only towards March.
Temperatures for this winter are expected to be above the 1981-2010 long-term average, but not exceptionally so. The most likely outcome is a positive anomaly of around 0.5C over Scotland and Northern Ireland and around 1C in central and southern England.
December looks set to have close to average temperatures across the UK as a whole, with slightly milder than average temperatures (by up to 1C) in the south of Britain, but slightly colder than average temperatures (again by up to 1C) are likely over much of Scotland and Northern Ireland. Confidence in this is relatively high, at around 70%.
In January it looks probable that all parts of the UK will be milder than average, although again with the south being milder relative to normal than the north - positive anomalies of 1 to 2C are likely in the south, less than 1C in the north. Confidence in this is not as high as with December, but the main uncertainty area is the extent of the positive temperature anomalies - there is high confidence that it will be milder than average for most or all of the country.
There is more uncertainty in the temperatures for February but it is most likely to also be milder than average, probably by 1 to 2C in most regions.
Rainfall totals are forecast to be above normal during the winter quarter over most parts of the country, but not exceptionally so.
December and January will probably both be wetter than average for most of the country, although with the forecast being for a mainly westerly type, some sheltered parts of eastern Scotland and north-east England may see near or rather below average rainfall. There is more confidence in the above-average rainfall for December (about 80%) than for January (about 60%).
February will probably be drier than average in most parts of the UK due to the signal for relatively high pressure, but as there is some uncertainty over the transition from a predominantly westerly type to more settled conditions, confidence in the below-average rainfall for February is lower.