Last November was on the cold side prompting me to investigate whether we were about to record a third month in a row below average. December turned out to be mild and wet, the lack of snow especially stark in Scotland.
The findings of that study showed that any sustained period of colder than average months was more likely to happen during months of March, April and May, nothing unusual there, especially considering H.H Lamb’s weather types.
I decided to scrutinise further all the colder than average months in this area, considering the 1981-2010 average, back to 1981. This gave the below results.
The dataset covers 399 months, of which 200 were colder than average.
The overall picture shows that negative anomalies are becoming more and more rare, though with notable exceptions being March 2013, December 2010 and January 2010.
The only month that has showed any sort of consistent general decrease in negative anomaly is November.
*For good snowfall at this station needs a negative anomaly of 2C during the months of November, December, January and February.
It’s been a while since this region has seen a sustained colder than average period. Though the past few weeks have seen colder than normal weather we have to go back to the beginning of 2015 where at least three consecutive months were colder than the 1981-2010 average.
The spell, which coincided with winter, was unremarkable with just one day of lying snow. The season finished 55th out of 73 of the most recent winters, the anomaly for the 90-day period was -0.3C. For deep, lasting snowfall events a monthly mean temperature must be at least 2C colder than average.
For anything ‘snow-worthy’ you have to go back seven years to a 10-month long colder than average spell that began in September 2012 and coincided with probably the last winter I can remember that had more than one cold spell with snow that lasted longer than a few days. The average mean for this spell was -1.3C.
The most potent cold period occurred during the winter of 1984/85, a winter where snow lay nearly a foot deep by the end of January and where seven ice days were recorded – these days we’re lucky to record just one ice day per winter! Only just behind was the 3-month period January to March 1987.
As with previous blogs I have devised a way of ranking these cold periods, by dividing the mean with the length in months. The first nine months of 1986 achieved the highest ranking, a period that included the 5th coldest February in this area since 1797.
And in chronological order…
It is looking like the mean this November will finish below average, making it the second month in a row, something that hasn’t happened since March last year. Will December make it three in a row and a new addition to the list?
In July 1808 Paris wilted in a heatwave. The average maximum for the 13-day spell that began on the 10th was 31°C, higher than a similar spell last month that saw the all-time record for the French capital broken.
The temperature at the peak of last month’s hot spell reached 42.6°C, some 6.4°C higher than the peak of the 1808 spell but, as the graph below shows, maxima fell back more quickly than 1808.
The average mean and minimum temperature of both spells showed a difference of just 0.2°C.
Since Paris recorded its hottest day ever there have been just two days where the temperature has exceeded 30°C.
The heat in Paris in 1808, like in 2019, was also felt in London. Luke Howard noted the following in The Climate of London.
“Very hot from July 12th to 19th. On the 12th a thermometer in perfect shade in a window in St James’s Park was 81.5 degrees at 3pm, and on the 13th at the same hour, 94 degrees. On the same day four men and seven women were killed by sunstroke in various parts of the Midland counties, and numerous coach and other horses were also killed. On the 15th a very violent and destructive thunderstorm in Gloucestershire, Monmouthshire, and surrounding counties.”
* Values for 2019 were taken from the station Montsouris.
** Though there is no way of knowing how accurate the 1808 values were previous studies have found that historic temperatures can be as much as 3°F too high.
Since February most of this region’s rainfall has been falling during the first half of the month.
May (to 26th) 54%
With only one of these months wetter than average the long-term, rolling trend continues to head downwards.
Considering past rainfall patterns it can only be a matter of time before the graph heads upwards again. Looking further back at rainfall data to 1961 reveals the extremes were higher and that the jump from dry to wet can be sudden.
The wettest 12-month rolling period ended mid way through May 1975 with a total of 873mm. However, just 16 months later, at the end of the hot summer of 1976, the rolling 12-month average was just 302mm, that’s less than half of our average annual rainfall!
Earlier this winter there were many reports on how good snowfall had been in Austria. But on closer inspection it was clear that the weather pattern at that time only favoured certain resorts.
In contrast with last year the totals I’ve used in my cross section of the range don’t look that exciting; Bourg-St-Maurice, the jumping off point for Savoie resorts including Les Arcs and Val d’Isere, has recorded 96mm of precipitation this season, in stark contrast to the 433mm it recorded last season between Christmas Day and February 5th.
In Switzerland Arosa, a resort well placed to pick up snow from any direction, has recorded the same this season as last. Totals in Davos are well down on last season. Similarly San Bernardino has recorded about half the amount of precipitation than it did last season. Its location toward the southern side of the range has been sheltered from the prevailing winds this year. That said it did enjoy a big dump last week.
Going further south and east St Vallentin in Italy has recorded about a third what it did at this stage last season.
Mean temperatures overall are about 2C to 3C lower than they were last season.
It is 21 years ago this month that London and most of the southern half of the UK experienced a remarkably warm spell of weather during what had been a mild winter.
The spell, which saw the record warmest February day on the 13th (19.7C), satisfies the Met Office’s old criteria of a heatwave whereby the maximum temperature is 5C or more above average for five consecutive days.
While these spells are fairly common in summer they are very rare during a meteorological winter. During the last 60 years the only other periods to have experienced a heatwave in winter are December 1966 and 2015.
Weather charts for this week look remarkably similar though, according to the latest forecasts, values will be nothing like they were in 1998.
Since this was published this 2019 spell has satisfied the Met Office criteria. But the anomalies are not yet as impressive as 1998.
More than three weeks have now past since the polar vortex split on January 1st.
London’s weather has turned colder and there’s even been a (small) fall of snow but the negative anomalies recorded so far are less than those experienced during the ‘Beast from the East‘ events last March.
In the graph below I have plotted the mean anomalies since the day of the PV split against the mean anomalies recorded from the day of the PV split in February 2018.
The effect this year is much less pronouced so far: over the last 24 days the average mean anomaly has been -0.7C, compared with -2.1C last February/ March.
The NOAA anomaly map for the week of January 13th-19th shows much of NW Europe has experienced a positive anomaly with any serious cold restricted to Norway and Sweden and Greece.
It has been mentioned that the variables with this SSW are very similar to 2004. Using the anomalies generated after the 2004 event resulted in the graphic below.
Taking it at face value would suggest that two more snow events could be possible at the beginning and middle of February – a cold but not severe pattern currently reflected in the models – but forget any repeat of last year’s Beast from the East.
Rain today (January 16th) is the first meaningful fall since before Christmas, putting an end to the 23-day long dry spell.
This meteorological drought, rare given that it spanned Christmas and New Year Storm singularities, these having 84 and 86 per cent probabilities respectively, was the 3rd equal longest winter drought.
The only other similar droughts in a list that dates back to 1887 were 19/12/2008 – 03/01/2009 and 17/12/1972 – 02/01/1973.
The last precipitation I recorded was from a weak occlusion that followed a cold front on the evening of December 23rd.
This synoptic set-up was followed by a build in air pressure that peaked on the morning of January 3rd; 1043.8mb was the highest reading in this area for at least 10 years and is the highest pressure I have measured.
A fuller version of London droughts in all seasons can be found here.
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.
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.
…And not a flake of snow. There’s been lots of talk about how an SSW will eventually usher in a cold January and February but it is by no means guaranteed.
The average maxima for this December will finish 10.2C, precisely the same as December 1985. Other Decembers that finished close to this average maximum are shown below.
As you can see apart from the middle of the month the temperatures are all over the place. And so are the conditions that followed in January and February.
A closer inspection of every year reveals that just three were very similar to 2018: 1974, 1986 and 1988.
Again, there’s plenty of spread. The 500mb reanalysis charts below show the situation of the northern hemisphere on December 31st of each year, including this year.
December 31st 1974
December 31st 1986
December 31st 1988
December 31st 2018
What followed in January 1975, 1987 and 1989?
January 1975 in this area was the warmest on record, back to 1797, while January 1989 was the 12th warmest. Conversely, January 1987 saw one of the coldest spells on record.
Perhaps it is fair to say that there is a 66 per cent chance of a very mild January, though you cannot discount the 34 per chance of another January 1987…