London Winter Forecast 2021/22

This winter is most likely to be on the colder side of average with near normal rainfall.

While the modelled prognosis for the first half of December looks unsettled with an Atlantic influence evolving from the current chilly NW’ly to a mild SW’ly, local analogues of the climate of London suggest the season could be a bit of a rollercoaster with spells of wet, windy and mild weather alternating with dry, calm and cold.

A large factor to consider this winter is the presence of a slight La Nina that is forecast to evolve cooler.  

Given the uncertainties involved with the influence of ENSO I’ve decided to stick with analogues found in local data that stretches back to 1797.

Overall then the probabilities for the next 90 days are.

Mean	Probability	Rainfall	Probability
Mild (>5.9C)	14%	Wet (>178mm)	19%
Average (5.1C – 5.9C)	38%	Average (107-177mm)	62%
Cold (<5.1C)	48%	Dry (>107mm)	19%

The above table doesn’t reveal a great deal in that extremes can be hidden in a month or season that finishes broadly average. So I decided to look closer at the winters that were revealed in the analogues.

December

The first month, as already mentioned, looks like it will be on the mild side with possibly a notable storm off the Atlantic before things calm down over Christmas – the period between Boxing Day and New Year’s Eve possibly presenting the best chance of any lying snow in this region.

The below graphs are a smoothed representation of the years revealed in the analogues most similar to this autumn.

What is probably most interesting is that the analogues that shared a similar ENSO / La Nina index to this autumn tended to ‘turbo charge’ any yo-yoing in the weather type, be that super-mild or abnormal cold.

December probabilties for maxima:
Mild: 48%
Average: 19%
Cold: 33%

And precipitation.
Wet: 33%
Average: 29%
Dry: 38%

So, the month overall will be mild and slightly dry

January

The first month of 2022 probably represents this region’s best chance of lying snow this winter. A mild start perhaps with a falling off of temperatures in the final third of the month and a cold spell of a week or so. As with December the influence of La Nina could tend to boost the swings in the pattern.

January probabilties for maxima:
Mild: 33%
Average: 10%
Cold: 57%

And precipitation.
Wet: 19%
Average: 43%
Dry: 38%

Greatest chances for January, then, are cold with average precipitation.

February

The second month may see a slight return of the cold spell in January before temperatures recover for a mild and wet spell in the second half.

February probabilities for maxima.
Mild: 33%
Average: 15%
Cold: 52%

And precipitation.
Wet: 38%
Average: 38%
Dry: 24%

The stats suggest on the cold side overall with average to above average rainfall. Perhaps the depth of the cold skewing any very mild second half of the month?

The extremes that no-one can forecast

As well as the very mild winter of 1989/90 the analogues also revealed the very cold winter of 1978/79. There were others but their occurence makes the probability of a repeat at either extreme at less than 10 per cent.

Thoughts on winter

Lots of hype regarding winter at the moment with contradictory model output being released almost daily.

My early take on this is based on the current state of the El Nino Southern Oscillation (ENSO) index which is currently -0.6 and trending more negative (La Nina).

Using analogues from previous events would suggest that the coming winter in the south-east of England will be broadly average in terms of mean temperature and wetter than average.

Looking more closely at each year suggests the following.

Further scrutiny of the local statistics suggest that the mean temperature for November could be lower than average, some 1.7C below average with rainfall around 20 per cent below average.

I should be noted, however, that in the years that were most similar to this September and October (1958, 1961, 1967, 2001, 2005, 2019) the ENSO index was much closer to neutral than it is now. In other words what happens next month could be vastly different to what happened in those years.

Years prior to those mentioned above add intrigue as ENSO statistics are not available.

1802 (cold January 1803), 1849 (severe January 1850 with ice in the Thames) 1862, 1865 (heavy snow southern England 10/11 January) 1874 (notable December storm)

Obviously a lot can change atmospherically in the next few weeks; this is an early take on my usual winter forecast which will be released on December 1st.

London winter forecast 2020/21

The best chance for lying snow this winter looks like being at the end of the second week of January.

A combination of local analogues and global atmospheric factors including the El Niño-Southern Oscillation suggest that the coming season will be colder than recent years. Though that’s not saying much considering just how mild winters of the past decade have been.

Meteorological autumn was the warmest for 5 years and slightly drier than last year’s wet autumn. Though it shares some similarities with 2015 the external influences are thankfully different to that season which produced the warmest December on record. Considering data back to 1797 I was able to make the following suggestions on how the next 90 days may unfold.

December is most likely to be around average temperature-wise with rainfall also about average. Possibly stormy at the end of the first week. Any snowfall events are likely to be marginal – bad news for anywhere below 70m above sea level. In terms of Christmas a white one in London looks unlikely. There may be interest in the week running up to the big day but I wouldn’t be surprised to see that inexplicable warm up that often happens just as the 25th arrives.
Mean: 5.8°C (5.6°C 1981-2010 average)
Rainfall: 57.1mm (53.2mm 1981-2010 average)

January is the month most likely to see any lying snowfall, particularly during the first half of the month, with the mean temperature about 1C colder than average. Rainfall is likely to be above average.
Mean: 4.1°C (5.2°C 1981-2010 average)
Rainfall: 63.8mm (53.2mm 1981-2010 average)

February looks wet and mild overall.
Mean: 6°C (5.3°C 1981-2010 average)
Rainfall: 55.6mm (39.2mm 1981-2010 average)

Overall the mean for winter: 5.3°C, a little below average.
And rainfall about 120 per cent higher than average.

Looking in closer detail reveals that the coldest period is most likely to be between 13th and 19th January, with anomalies sufficiently low enough for lasting lying snow.

The extremes that no-one can forecast

As well as the extreme December 2015 the analogues also revealed the severe season of 1822-23 which saw ice on the Thames by late December. February 8th saw a great snowstorm in northern England where people had to tunnel through the snow.

Another was 1950-51 which was very snowy at high levels. There were 102 days of lying snow at Dalwhinnie (1000ft), exceeding the 83 days set in 1946-47. December 15th saw 15in of snow in Shanklin, Isle of Wight in 3.5 hours.

London winter forecast 2019/20

For this year’s winter forecast, because of time constraints, I am sticking purely to what the local analogues reveal.

Meteorological autumn was the wettest for 18 years and the coolest for 6 years, revealing some interesting similarities with past climate. Considering data back to 1797 I was able to make the following suggestions on how the next 90 days may unfold.

December is most likely to be on the cold side with rainfall slightly below average.

January is most likely to finish average temperature-wise and wetter than December.

London’s February extremes temperature also look average overall and the precipitation signal also creeps up in comparison to the first two winter months.

Best chance for snow? Impossible to pin down but the coldest weather is likely to be at the beginning of February, the third week of January and just after Christmas.

The warmest period looks to be around January 10th while there’s also a signal for that often phenomena of milder temperatures just before Christmas day.

Most of the probabilities in the above statements are relatively low and are explained in the table below.

Mean temperature 2019/20

December			January			February
1981-2010 average	5.6		1981-2010 average	5.3		1981-2010 average	5.3
Forecast median	4.4		Forecast median	4.7		Forecast median	5
Rather cold	40%		Normal	40%		Normal	44%
Normal	28%		Rather cold	32%		Rather cold	32%
Mild	20%		Very cold	12%		Mild	16%
Cold	4%		Mild	12%		Cold	4%
Very cold	8%		Cold	4%		Very cold	4%

 

Rainfall 2019/20

December			January			February
1981-2010 average	53.2mm		1981-2010 average	55mm		1981-2010 average	40.1mm
Forecast median	48.3mm		Forecast median	55mm		Forecast median	40mm
Wet	32%		Wet	32%		Wet	36%
Normal	20%		Normal	28%		Normal	36%
Dry	48%		Dry	36%		Dry	28%

Winter proper to arrive mid January

Weather models are continuing to struggle in the aftermath of the stratospheric sudden warming on January 1st. The GFS and ECMWF have flip-flopped: on one run decent northern blocking extends southward only for the dreaded European high to appear on the next.

Using a combination of QBO and ENSO data featured in my winter forecast and statistics from previous SSWs (including 2013 and 2018) achieved the following results shown in this graphic.

Although some days in the next week or so will be cold it is not until the 14th that conditions start to bite, the start of a week-long cold spell that will probably be more notable for cold than snowfall.

The rapid recovery in temperature would suggest that the Azores / European high making a return. With the MJO moving back and forth between phase 7 and 8, and looking at the behaviour of previous cold spells, this would make sense.

The GFS has been churning out some very strange charts of late

As for February, unless there are further SSWs to disrupt the polar vortex, and depending on its recovery, it is unlikely we will see a repeat of the winter of 1984/85 that I hinted at last month. The graphic below, however, would suggest another cold spell in the third week of February.

 

 

London winter forecast 2018-19

Long range modelled forecasts have been all over the place of late and, looking at the underlying signals, it is easy to see why.

When I’ve produced these forecast in the past, in terms of QBO and ENSO data, there’s usually a lot of analogues to compare with. This year, however, seems to be an exception.

Considering QBO first I looked back over data to 1950 and found nothing similar for October. However, looking over the whole series the cyclical nature of this circulation may give some clue.

Some 20 months were revealed, ranging from June 1959 to June 2015.  Using NOAA’s  Niño 3.4 region I narrowed this list down to the few that had an ENSO value of around +1 with a rising trend. With NOAA’s forecast of a Modoki El Nino (one that occurs in the central Pacific) this narrowed the list to just 1 period: June 2015. Considering maxima anomalies this would give the following winter.

The above would suggest there being a general cool down through December with a cold spell starting just before Christmas into the new year? And another cold spell end of January into the first week of February?

The above precipitation anomaly chart would suggest a wetter than average December,  January and February, though February by much less so.

It’s been a very busy autumn so I’m keeping this short.

The below figures, particularly January and February, may be different in the event of an SSW occurring. In all then.

The mean:
December: +0.8C
January: -1.2C
February: -1.7C
Overall: -0.3C (broadly average)

Precipitation:
December: 158%
January: 155%
February: 120%
Overall: 134%

 

Winter 2017/18 forecast review

At the beginning of December the opening par for my forecast for winter read:

The mean temperature for winter 2017/18 is looking average overall with average precipitation. While that doesn’t sound exciting for anyone looking for colder weather and snow I think the figures mask frequent 3 or 4 day-long cold snaps from the Arctic interspersed with milder interludes from Atlantic incursions; typical characteristics of a pattern driven by the troposphere. For anything longer term we have to hope for a warming of the stratosphere, a sudden stratospheric warming over the Arctic, that downwells into the troposphere, reversing the general westerly circulation. We are overdue an ‘SSW’ event but, even if one were to happen, its effects wouldn’t be felt until much later in the winter.

Figures generated using QBO statistics in the run up to December were identical to the outcome of 5C. It is figures generated using ENSO statistics, however, that produced the best results considering December, January and February alone.

It would seem that the winter was most influenced by the ENSO variable up until February when the effects of the SSW were enough to overcome it?

 

 

London winter forecast 2017/18

I’ve been doing these written winter forecasts for five years now but this year I’ve never seen such a spread of probabilities as to how winter may pan out. So, if by the end of reading this you are none the wiser, you’re not alone!

The mean temperature for winter 2017/18 is looking average overall with average precipitation. While that doesn’t sound exciting for anyone looking for colder weather and snow I think the figures mask frequent 3 or 4 day-long cold snaps from the Arctic interspersed with milder interludes from Atlantic incursions; typical characteristics of a pattern driven by the troposphere. For anything longer term we have to hope for a warming of the stratosphere, a sudden stratospheric warming over the Arctic, that downwells into the troposphere, reversing the general westerly circulation. We are overdue an ‘SSW’ event but, even if one were to happen, its effects wouldn’t be felt until much later in the winter.

The current cold snap follows the recent disruption of the polar vortex. The ‘displaced polar vortex’ regime seems to be becoming more common as the Arctic gets warmer and maintains more open water during winter. This leads to anomalous ridging over the Arctic, which results in a displaced or bifurcated vortex, which leads to somewhere in midlatitude getting persistent anomalous cold. There was very cold air around last year but the UK missed out; the heart of the cold plunging down far to our east with snow falling on beaches in Crete!

This week’s cold snap looks like it will be short lived though the ‘displaced vortex’ could be an early sign of yet another pattern transition in the troposphere. Looking at the latest model output it is uncertain how long the next polar vortex perturbation will be.

So, whether we see any cold, snow-worthy, air this winter is a lottery – though I think our chances are a lot better than they have been for the past 5 years.

Because October and November have been so dry I’ve had very little to work with in terms of stats to try to find if the current pattern is similar to previous autumns. Instead I’ve looked in depth at ENSO, QBO and local meteorological data.

ENSO

The El Niño–Southern Oscillation (ENSO) is a periodic variation in winds and sea surface temperatures over the tropical eastern Pacific Ocean. It is forecast to be mildly negative (La Nina) over the coming winter. Many doubt whether it has an impact on our winter but there seems to be a connection with strong La Ninas and very mild winters. Because it is forecast to only be mildly negative it’s impact may be very small. I had a look at similar years were ENSO was -0.4 in September and came up with the following.

QBO

I also had a look the quasi-biennial oscillation (QBO), the quasiperiodic oscillation of the equatorial zonal wind between easterlies and westerlies in the tropical stratosphere. This produced the following table of probabilities.

Singularities / patterns

My traditional method of looking back for similar patterns of weather in the autumn produced this table.

 

 So, considering all three tables together, would suggest, probably bizarrely given the cold start to this month, a rather mild December followed by an average January and February.

London winter forecast 2016/17

Frost still lingers on my lawn as I write this winter forecast on the last day of the meteorological autumn. Last night’s minimum was -6.1C, the coldest night this year and, indeed, colder than any night last winter.

It would be very easy to get caught up in the hysteria that happens every November of what the winter will bring. Most of the hysteria comes from ‘coldies’, those folk that hope every winter will be the coldest and snowiest on record. I count myself among them.

The north-west of England looks around 10C above average on December 9

But, despite hints from BBC weather forecasters that something maybe afoot in terms of something colder than usual from mid-December, years of disappointment have taught me to just take winter as it comes. Meteorological models this morning look anomalously warm for the first half of December, the UK being on the warm side of an anticyclone centred over France – the theme of so many winters I can remember.

The current phases of ENSO and QBO suggest that their influence on our winter this year will be far less marked than last year. Looking at teleconnections with a slight La Nina and the current QBO cycle have revealed winters that were fairly uninspiring though by no means write-offs for snow, two examples in the case of QBO being 1985/86 and 1990/91 – both had very cold Februaries.

Before I got too bogged down with trying to find patterns here I decided to move on to my more traditional pattern-matching.

Before I trawled through the figures my initial gut reaction to this winter was that it would be the coldest for at least four years – a not-too-difficult feat giving the mildness of the last three winters which saw very little in the way of snow.

The overall mean temperature for December, January and February came out at as 4.2C – 1.2C colder than the 1981-2010 average, with average precipitation. 

A rather cold winter, over three months doesn’t tell you much. A very mild and stormy December could mask a dry and notably cold February. This makes it worth looking at the individual months of each winter.      

There is a lot of spread in the above tables. Considering them altogether I would guess winter will unfold in the following way:

December to start mild, possibly becoming very mild, before beginning to turn very cold mid-month. This pattern will persist, with varying intensity through into January which could also be very cold. Milder conditions to follow in February though this will probably end a fairly average month and drier after a slightly wetter than average December and average January.

As well as my method of using rainfall and temperature I also considered other methods. One was Russian research that states that the weather pattern in the winter will be the opposite to the weather on September 17th and November 7th. This autumn September 17th was N’ly and November 7th was NNW’ly. So, this would suggest a continental flow from S and SSE.

You can read the method of how I reached my conclusion here.

Late autumn to bring chilly shock

The Met Office’s latest outlook for a relatively high chance of cold weather in the early part of winter has set the internet abuzz with talk that the UK is about to experience its coldest winter in years.

The 3-month outlook, produced by the agency for contingency planners, states:

Lower-than-average temperatures are more probable than higher-than-average values. Overall, the probability that the UK-mean temperature for November-December-January will fall into the coldest of our five categories is 30% and the probability that it will fall into the warmest of our five categories is 10% (the 1981-2010 probability for each of these categories is 20%).

THe ‘30% chance’ for temperatures to be in the coldest of five categories is more significant than it would appear: in this three-year period of mild winters that figure usually hovers far lower.

The continent is not yet cold enough to get excited about this synoptic chart for Monday, November 7th

The forecast has drawn obvious comparisons with their 2009 ‘barbecue summer’ f0recast but the GloSea5 model has made infinite advances on the old one.

The Met Office’s method, which includes a combination of data such as QBO, ENSO and ground based observations also agrees with my own method of finding local October singularities back to 1797. This month I found that an average or cold November is most likely: average and cold are both 30 per cent probability. Rather cold and severe are both 20%! There appears no chance of anything mild.

So, we can look forward to ‘a reduction in the normal westerly flow across the UK from the Atlantic, with a greater frequency of northerly or easterly winds’.

But what does this mean for the London area? If the near continent is not cold enough northerly and easterly winds could just bring a succession of cold rain in the form of showers off the North Sea with a higher than average frequency of frosty nights: nothing out of the ordinary for a slightly cooler than average November.

To try and second guess proceedings into December I decided to plot all year singularities to find when the cold spells would most likely fall and how severe they could be.

The results for November look unexciting: the coldest day is likely to be around the 19th with a maximum of 6.9C. The best chance for something cold looks to be in December, around 6th to the 11th – but there again highs of 3C or 4C would mean that any snow would be short lived. There appears to be a warm up in time to scotch hopes of a white Christmas before another cool down toward new year.

So, in conclusion, late autumn / early winter is likely to be a lot colder than the past three years. It is perhaps this fact, together with frosty mornings being common, that will make the season feel colder than it actually is.

I will be publishing my usual winter forecast on December 1st.

The coldest day in November is likely to be around the 19th