Category Archives: weather pattern

The disappearing deluges of September

While putting together my September extremes blog I noticed that the month was marked by some big rainfall episodes. On a national scale TORRO statistics show that a south-east climate station holds the daily record for September – unusual in that every other month is dominated by stations in the north and west of Britain.

Further analysis of local data since 1959 shows how September has slowly evolved from being dominated by autumnal to summery weather. The wettest period, from the mid Sixties to the mid Seventies, saw 42 per cent of highest daily rainfall events recorded.

Two of these events, in 1968 and 1973, are well noted and appear in RMET’s Weather magazine.

1968

In Wanstead, between 14th and 15th September, a total of 58.4mm fell, a large total though far less than elsewhere.

The rain became torrential overnight in southeast England and continued through most of the 15th. Rainfall totals ranged from three inches in the London area and nearly four inches in southern Essex to approach the quite abnormal level of nearly seven and a half inches in parts of Kent during the 14th/15th.

The area of greatest precipitation was near the Kent, Surrey and Sussex border where violent downpours in the 12 to 15 hour period from midnight on Saturday to the early afternoon of Sunday 15th led to widespread and disastrous flooding. The heavier rain area moved north and east during the night of 15th/16th and Gorleston recorded.

The highest accepted two-day falls were 201mm at two rain gauges at Tilbury
and Stifford in Essex, and a similar fall north of Petworth in Sussex. The highest ‘rainday’
totals (i.e. nominally 0900–0900 GMT) listed in British Rainfall 1968 were 129 mm at
Bromley and 125 mm at South Godstone sewage works in Kent, both on 15
September.

This exceptional event was described by Bleasdale (1974), Salter and
Richards (1974) and Jackson (1977) and the map below (Fig. 1) is taken from Bleasdale’s
paper in British Rainfall 1968 (p. 231). In all, some 575 km2 received more than 150 mm
in 48 hours.

1968 rainfall

1973

In Wanstead, on 20th, a total of 55.4mm fell, the largest daily total recorded in September. Again it was far less than elsewhere.

5pm gmt 20th 1973

There were notable falls of rain in London, Surrey, West and East Sussex, and particularly Kent. At Manston, near Margate, 172 mm fell in 18 hours 40 minutes commencing 1710 GMT; at nearby West Stourmouth 190.7 mm fell in the rainfall day i.e. the 24 hours commencing 0900 GMT on 20 September (source: Met Office internal list of heavy falls of rainfall in short periods in the United Kingdom during the year 1973; Rainfall/heavy falls section listing, available in manuscript/computer printout form in Met Office archives).

The future

Big September rainfall events seem to be becoming rarer in our part of the UK, the last was in 2014, on the day of the Scottish referendum, though this was convective rather than frontal rainfall.

The outlook for the rest of this month suggests yet another absence of a large rainfall event. The ingredients for large rainfall totals in the south-east – blocking high to the north with slow moving low pressure over Brittany – look unlikely to form during the remainder of the month. We probably will see rain but any fronts are likely to move through quickly, with typical totals being around 5mm.

rain rest.gif

Advertisements

When did September become a summer month?

The answer is around 1993. A look back at mean temperature and rainfall statistics for east London over the past hundred years reveals that the ninth month has, since that date, slowly become warmer and drier.

Putting all the arguments of meteorological and astronomical summer aside, many people of a certain age regard September as an autumnal month, but as recent years have shown it can very often be an extension of summer; September 2016 was the second warmest on record in the local area, warmer than many previous summer months!

Looking back even further, over the past 100 years, the September mean has trended upward, though many peaks and troughs reveal how the month has ebbed and flowed from being summery to autumnal.

The prognosis to the end of this month suggests that air pressure will be anomalously high – so the pattern in the south-east for settled, summery weather in September doesn’t look like ending any time soon.

sept means
The September mean trend has crept generally upward.
sept rainfall
A look at rainfall back over the last 100 years shows that wet Septembers have been on a general decline since 1994.

sept chart.gif

Repeat of March 24th 2013 only 10C colder

The weather this coming weekend reminds me of a March weekend in 2013 when large parts of the south-east woke up to a covering of snow.

Conditions in Wanstead were very knife-edge, the cover being very slushy. I remember going for a bike ride on that Sunday morning. We left Aldersbrook in sleet, the cover grew thicker as we pedalled into Essex. By the time we reached High Beach near Epping there was 8cm of fresh cover. Conditions on the 2-hour plus ride were inclement to say the least – the temperature didn’t rise above 0C!

This weekend we have a similar set-up, only this time the upper air at 5,000ft is a good 10C colder – no knife-edges this time and, if the streamer gets going, many places should see 10cm of cover. Don’t put away that sledge just yet!

 

The most potent cold spells since 1960

With met models now coming into reliable range it now looks odds on that very cold air from the continent will be in place across most of the UK from early next week.

The big question is how long will the cold last and how much snow will fall? While the latter looks likely at some point once the air is in place it is impossible to pinpoint where and how much any given place will receive at this range.

Screen Shot 2018-02-22 at 09.01.04
The 00Z op run of the ECMWF model has the really cold air arriving at 00Z on Monday 26th.

In terms of longevity latest data shows the spell could have real staying power though my experience with models over the years has shown that they can overcook the potential of a cold spell.

I’ve lost count of the number of times when excited enthusiasts proclaim that an incoming cold spell is going to last as least three weeks; the reality being that the intensity of the cold has gone after four or five days. Cold spells since 2008 often arrive as a ‘blob of cold air’ from the continent that eventually gets ‘warmed out’; it’s been a very long time since we had a cold spell that’s been fuelled by a continual feed of air off the continent.

Screen Shot 2018-02-22 at 09.02.23
By Saturday 00Z the really cold air aloft has gone. A typical 4-5 day event?

To illustrate my point I had a look back at every cold spell in this area of suburban east London since 1960. I weeded out the feeble efforts of the last few years by only considering spells where the maximum didn’t exceed 2.8C. The results spanned from the most recent cold spell of March 2013 to the mammoth 31-day Siberian blast that began on Boxing Day 1962.

In another blog I remarked how similar the recent pattern was to February 1962. This cold spell began on the 26th and lasted 9 days. Some 7cm of snow fell, this drifting in the wind, possibly making it seem worse with only 4 hours of sunshine which would have maintained any snow cover.

One of the snowiest cold spells happened in February 2009, eight days after an SSW event that lead to a polar vortex split. This four day spell saw a total of 26cm of snow fall.

February 1991 was even snowier, the intense cold lasting some 11 days.

Looking at other February cold spells a 7-day spell occurred in 1985 about a month after an SSW event.

Overall the median length was 5 days with an average of 6cm of snow and 8hrs of sunshine.

*A survey of winters ranked for temperature and snow can be found here.

spells

rank cold

 

1960

 

 

 

 

 

 

 

 

 

 

 

 

 

The cold spell of February / March 1962

The last week or so has been agony for model watching coldies wishing for a snowy end to winter.

27021962
The Synoptic chart for February 27th which saw 4 inches of snow in Stratford, east London

Solutions have often flip-flopped between a mild and cold outlook. Even this morning at 6 days out the GFS and ECM model temperature outcomes differed by some 20C, a choice between spring and deepest winter!

Sometimes you have to try and second guess what is going to happen by looking at previous patterns from years past.

Yesterday at Philip Eden‘s funeral I spoke to Woodford Green resident Ron Button. He pointed out how much this February reminded him of February 1962, a month that was non-descript for the first three weeks before turning very cold and snowy. Ron, who has kept a meticulous record of the weather ever since his interest was prompted by the severe winter of 1947, produced diaries of 1962 when he was living in Stratford. The entry for 26th / 27th read: “4 inches of snow with drifting”!

The March that followed was the coldest of the 20th century, ranking 10th in my list back to 1797, and 0.3C colder than March 2013. The fact that the monthly sunshine in 1962 was only slightly below average is testament to how cold the source of the air must have been. March 2013 was very dull by comparison!

An entry in London-weather.eu reads that March was colder than any of the previous 3 winter months: “The first three weeks were mostly mild and dry. It was often breezy which resulted in fewer than average night frosts. Frontal systems off the Atlantic passed through the London area, most of them weak though on the 12th, nearly 8mm of rain fell with southwesterly winds gusting to 57 knots. During the last week of the month, it became much colder, and on the 26th snow fell with the temperature not rising above -0.3C all day.”

There was no stratospheric sudden warming that winter and ENSO was neutral.

The winter of 1961/62 ranks only 14th in my list of worst winters, mostly because the core of the cold happened in March which is considered spring in meteorological circles

Screen Shot 2018-02-16 at 13.54.17

march 1962 summary

The Snow Survey of Great Britain also makes interesting reading with these entries for February and March.

feb1962feb19622

march 1962

How does an SSW affect London’s weather

There’s been much anticipation regarding the forthcoming sudden stratospheric warming (SSW) event with many hoping that a resultant split vortex will result in unseasonably cold weather in the UK and… copious snow in the low-lying south-east.

An SSW event, which reverses winds high up in the atmosphere from a westerly to easterly, can downwell into the troposphere, bringing weather from a (usually) cold continent instead of the warm Atlantic.

While a split PV event is usually more conducive for cold weather in the UK, as opposed to a ‘displaced vortex’, which usually favours only the eastern US, it is by no means a guarantee of a cold pattern subsequently evolving.

Screen Shot 2018-02-09 at 18.43.11Using results published in the paper Tropospheric Precursors and Stratospheric Warmings (Judah Cohen and Justin Jones), along with meteorological data for east London, I set out to find what influence past warmings had on the weather in the capital.

Looking 45 days either side of the central date for vortex splits gave the following, chaotic graph.

 

Screen Shot 2018-02-09 at 18.08.00

But every year is different. And there appears to be more likelihood of an SSW making a difference, in terms of prompting a colder pattern, the earlier in the winter it occurs.

The SSW in 1985 was followed by a 45-day mean temperature anomaly of -3.8C! If you look at a shorter timescale, 15 days after a split PV and the anomaly is -10C on January 16th: -4.9C is the second coldest January day in Wanstead of the past 60 years.

At the other end of the scale the SSW event on March 23rd 1965 was followed by a POSITIVE anomaly of 3.4C. Perhaps solar influence this late in the year can override any SSW? Elsewhere, however, according to the website london-weather.eu: “3rd March – A combination of deep snow cover and clear skies allowed minimum temperatures to fall below -21C in northern Scotland.”

During another SSW in 2001 results in London were fairly unremarkable though heavy snow fell in Ireland.

 

Screen Shot 2018-02-09 at 19.33.43
This graph shows a general downward trend in the 15 days following an SSW event

Screen Shot 2018-02-10 at 10.42.39The other result to consider is the influence from ENSO. It seems that when La Nina is ‘too negative’ this can ‘overcook’ proceedings and actually leave our part of the UK with a positive anomaly, as this table shows. It should be noted, however, that thicker Arctic ice in the 1960s would also possibly have had more influence than now.

There hasn’t been a full SSW event for years. The impact this one will have on our weather in London, a tiny part of the globe, is impossible to quantify. Though the latest model output is encouraging for anyone looking for a chilly end to winter.

Screen Shot 2018-02-09 at 19.48.42

And here’s @wxcharts idea on two weeks from now.

 

The ‘nearly cold spell’ of February 2008

Recent model output that has many bullishly predicting a cold spell in February reminds me of the winter of 2008 when the likes of the GFS and ECMWF models kept teasing us that a Scandinavian high pressure was imminent.

This rare pattern brings the south-east its coldest weather in winter, in the past bringing the January 1987 and February 1991 cold spells.

feb16
Midnight on February 16th 2008 just before the high moved south, taking its cold weather into the Balkans

A look back at the pressure patterns in 2008 show that high pressure was in the ‘right place’ between the 11th and 16th but, just as -10C uppers looked poised to flood across from eastern Europe, the high pressure collapsed south, leaving the UK in a SE’ly flow off a mild France.

There was snow in 2008 but it didn’t arrive until the spring, in late March and, the best fall, in early April when I remember waking up on the 6th to a 5cm covering that mostly melted by lunchtime.

I suppose the one silver lining for any coldies reading this is that February 2008 was the sunniest on record.

2017
The mean temperature for December 1st – January 30th is 6C for both 2007/08 and 2017/18

The maps below show the sea surface temperature anomalies. A slight La Nina on both. ENSO, considering my winter forecast, has been the best guide for how this winter has so-far panned out – front loaded.

Screen Shot 2018-01-30 at 11.49.50

Screen Shot 2018-01-30 at 11.50.40

wl
Philip Eden’s column in the journal Weather showed that there was some wintry weather in February 2008 but it was restricted to the north of England northwards

* The focus of these blogs is usually on what the weather was, or may be, like in my backyard – though they usually broadly apply to anywhere in the south-east that is close to sea level.

Best Alpine snow for 30 years?

There’s been countless reports about amazing amounts of snow falling across the French, Swiss and Italian Alps to the point where some agencies have been proclaiming that it has been the best season for the white stuff in 30 years.

Extraordinary totals have fallen in some areas. Bourg-St-Maurice, the jumping off point for Savoie resorts including Les Arcs and Val d’Isere, has recorded over 400mm of precipitation over the past 30 days, equating to around 4m of snow at the resort summits.

In Switzerland, large amounts of snow in a short period caused chaos in Zermatt, stranding tourists after the area’s rail services suffered disruption.

Away from the north and west side of the Alps, however, snowfall, while good, has been less impressive the further south and east you look.

alps arrow
30-day precipitation totals reveal that Bourg St Maurice recorded 422mm while Obertauern in the east recorded 99mm. 

It is a far cry from last year where some resorts on the southern side of the range were particularly dry. San Bernardino, during the last 30-days, has recorded 179mm of precipitation. During the same period last year just 14.6mm fell!

The outlook for the Alps continues to look unsettled with snow forecast to fall at resorts that are in deficit to the Valais and Savoie areas.

snow

 

 

The same old weather every year?

Recurring weather patterns at certain times of the year are well known. The ‘Buchan Cold Spell’, ‘European Monsoon’ and Indian / St Martin’s Summer are all phenomena that have been studied extensively.

An article by the late meteorologist and broadcaster Philip Eden a number of years ago considered many of these patterns and found that, to varying degrees of reliability, they provided a guide to what the weather would be like at any given time of the year.

Considering climate change I wondered how much these patterns could still be relied on. Using my own pressure, rainfall and ‘wind run’ data (the total amount of daily wind) going back to the start of 2013 I had a look at the singularities for January and February

January patterns at the beginning, middle and end of the month appear to be the most reliable. However, it is only the ‘mid-Jan settled’ period that is most reliable.

singulariuty

pressure
The pressure trace most notably shows a general rise from the 17th, the date of last week’s windstorm, consistent with the mid-January settled singularity.
rain
Though this year’s ‘settled’ January spell saw rain the past 5 years have seen mostly dry weather.
windrun
Very little wind has been recorded around January 20th for the past 5 years

The early Feb settled spell occurs with very low probability: just 56 per cent. And this year the pressure, according to the GFS model, plotted below by WXCharts.eu, is predicted to be around 1040mb by February 2!

out to feb 1

new gif
This animation of the GFS model shows the idea of the Iceland low, which drives our SW’ly type weather, ‘taking a holiday’ to southern Iberia, possibly advecting any cold weather in the east to flood the UK

Looking at the results of the past 5 years it could be concluded that the patterns do still occur but because of the nature of the jet stream, which seems to meander far more readily than in the past, these stormy / quiet episodes are becoming shorter than they were in previous studies.

 

A 100% guaranteed snow risk for rest of winter

This weekend marks the halfway point through winter. Though December brought some snow January has been broadly average – really cold air has been absent with only one air frost recorded this month.

The cloudy, anticyclonic type weather is about to be replaced, however, with cold, polar maritime air this week set to flood down from the north-west.

Since December 1st, Wanstead has recorded 12 air frosts – about average. The coldest night was just -3.7C. The current mean temperature this winter to January 14th is 5.7C with rainfall 111mm – statistics that are remarkably similar to the winters of 2012/13 and 1990/91. Both those seasons were followed by cold late winters, February 1991 saw some of the deepest snow that I’ve ever seen in the south-east; the mean temp for that February finished 1.6C, the 14th coldest in the local series going back to 1797.

Using my method for finding patterns stretching back over 50 years to forecast this winter I picked out years that were +/- 10% of the 2017/18 total rainfall. From these I then weeded out the seasons where the average temperature was +/- 10% of the 2017/18 mean.

This gave a list of just two other winters with similar temperature and rainfall. Both winters had above average ‘snow lying’ days, the long terms average for this area being six.

Before any readers accuse me of going all Daily Express with the title of this blog I would emphasise that this piece doesn’t echo the latest long-term output from the models, which are in a state of flux, caused mostly by the evolution of the explosive cyclogenesis expected midweek. It is simply a reflection of what the local data is telling me.

A 100% guarantee of snow isn’t such a fantastical claim as, during the past 10 years, there is only one winter when no snow fell or was lying at 9am!

In terms of the rest of the winter, outside of the models, I would expect a couple more snowfall episodes, similar to the ones we had in January and February 2013. A repeat of February 1991, while not impossible, looks unlikely – there seems to be far too much energy coming from the Atlantic to allow the all important Scandinavian / Russian high to form and exert its influence far enough west for long-lasting cold and snow.