Following on from some posts I made earlier, and given the frankly boring weather outlook, I thought I'd look into the death of my favourite weather phenomenon - the midwinter easterly. We've had easterlies in spring and autumn, but for the purpose of this post I'm concentrating on winter.

Living here in the southeast, near to (but not on) the coast, it's darned hard to get snow these days. I grew up in the 80s and my love of the weather was formed by watching the snow come down, year after year. My gran said to me that it doesn't snow every year, but I'd seen firsthand that it did... that is until winter 1987/8, which was indeed snowless here. After that I learned to treasure snowfall rather than expect it - but I knew that if we didn't get any it'd only be a couple of years until we did.

And so that held through the 90s, until January 1997. That was the last deep cold easterly and was the last time I saw powder snow, icicles, snow sublimating rather than melting, all the traditional bits of winter that I remembered. It was also my last year as a child, for I turned 18 that autumn.

I couldn't have known then that the following 19 years would be icicle-free and largely free of ice days. Little did I realise how frustrating it'd become as a snow fan down here - the former ally of snow (convection over the sea) would turn into the villain, raising temperatures just that bit too much for decent snow.

So, rose-tinted glasses or not? Let's look at some cold, hard data. Here's a chart showing snowfall days at Manston, together with ice days. Note that snowfall can include sleet, or just a flurry of snow - it doesn't mean heavy snow! Data is from Tutiempo, which goes back to January 1973. Note that the chart shows the December of a given year, plus the following January and February.

At first glance it may look like not much has changed. Look more closely though and you'll see a near complete absense of ice days since 1997. Many of the snow days have been sleet since then, whereas before then more of them were snow rather than sleet. (The data may well miss out a couple of days in before 1977, as figures were rounded to the nearest number. There were a few zeroes which may or may not have been ice days).

From this, it's clear that the climate of this part of the world has undergone a marked change in winter. Why is this?

The answer is the lack of midwinter easterlies. You see, down here we rely on easterlies (or NE'lies) for snow. Snow from the north or NW is rare, as we rely on an organised front making it this far. Even then, it's a hell of a job to get ice days from the north or NW.

The other way of getting heavy snow is via a frontal attack from the SW, but that's not only rare but fraught with danger; get it a bit too mild and you just end up with slush.

No, the reliable way is to get a deep cold pool advected from the east. The North Sea causes instability and as the air's really cold although it warms up it remains below freezing - the ideal recipe for heavy convective snow showers. Any troughs/fronts in the flow just add to the mix.

Ideally the cold pool would be modified as little as possible by the sea, which would require a strong easterly flow in order to speed its journey and minimise the modification taking place. A sharply defined cold pool is a bonus too, as the greater contrast seems to aid convective potential. Even without that though snowfall was possible or even likely if an easterly flow was sustained for more than a couple of days.

Here are three examples, excluding the obvious one (1987). In 1986 we had a slack pressure pattern, but that blob to the east contained some seriously cold air (<-20C at 850). 850s fell to -13C over Kent, low enough to trigger slow-moving snow showers.

In 1991 we had an upper cold pool move swifly westwards just to the south. 850s were -17C for a time over Kent, enough to cause persistent heavy, thundery snow showers.

At the end of 1996 a cold pool moved westwards over the southern UK, with -14C 850s accompanying it. At a time of minimum insolation, temperatures plummeted even at the coast and heavy convective snow was the order of the day. This was the last time a high of -1C or lower was recorded here, along with icicles etc.

Some would say these sorts of setups were never common and to be fair they weren't that common - you would expect 2 or 3 a decade in the 70s, 80s and 90s. However, even "near miss" setups would give snow, as conditions weren't as marginal as they are now. Indeed, many of the ice days in the Manston chart were from setups where the core of the cold didn't even reach us!

2005 was the last true easterly spell here, but it simply came too late. In addition, recycled air from the Med got caught up in the flow, diluting the 850s and meaning for most it was just raw drizzle rather than a winter wonderland.

Eventually (after another week of this!) some cold air did arrive - 850s fell to -14C briefly in Kent, but as it was under a ridge and light northerlies there was no convective activity. A pulse of milder air moved down from the north, which brought a return to some wet snow but with the loss of really cold air aloft it was again too little to sustain it - temperatures were above freezing and it just led to a slushfest.

So, with that in mind, there are several factors at work here.

• Synoptically it seems that easterlies of any type are rarer than they were before 1997.
• When we do get easterlies, any cold pools are ill-defined and lack contrast with surrounding air (compare 2005 with the others). This leads to greater stability of the air, reducing the "lake effect" of the North Sea.
• When air does move in from the east, it's not as cold as it was due to greater mixing out of the airmass. Or, to put it another way, something invariably fouls the flow to prevent a direct feed from Russia.

Add to that a general background warming and what was marginal is now rain, what was a safe snow event (-1C) is now marginal.

I've been scouring various archives to try and find what flipped in 1997. It's clear to me that something changed in our atmospheric circulation to prevent easterlies and to warm things generally. I looked at sunspots (not much correlation there, although 2010 did occur during a minimum). I looked at El Nino (inconclusive). I looked at sea temperatures - and found something that fits, namely the AMO.

This is the multidecadal Atlantic oscillation. As can be seen, it flipped from negative (cold) to positive (warm) around 1997 - and it's been warm ever since. As the Atlantic has such a large effect on our weather I wouldn't mind betting this is part of the puzzle - the extra warmth will have added more energy to our weather, promoting a stronger jet (due to greater contrasts between the Atlantic and the Arctic) and thus reducing the frequency of blocking episodes. Furthermore, the extra warmth would help turn what would have been marginal events into rain events, or snow events into sleet.

I don't know why it'd have the effect of reducing the depth of cold pools from the east (the few times they occur), but it could well be tied in to the lack of blocking generally allowing more mixing-out to occur.

This all sounds fine and dandy, now all we need is some weather data before 1963 or so... and, of course, we know all about 1962/3 and 1946/7, both of which were epic cold spells in the middle of a mild AMO. What about the rest of it?

Amazingly, some data from the 30s exists on TuTiempo for Manston. The winter of 36/7, for example, saw one snow day and no ice days. It was 5/0 for 37/8 and 8/5 for 38/9. And lo, 38/9 contained a textbook easterly:

There were others, too, during that period of positive AMO. So it's not the be-all-and-end-all.

However, I feel confident enough to draw some conclusions.

• The AMO has some effect on the UK and surrounding areas, possibly quite a big one. Easterlies and snow seem more common during negative phases.
• Something changed in 1997 to make easterlies of any type much rarer. It also coincided with a marked reduction in ice days.
• Since 1997 snowfall has become much more marginal, with some events which would have been snowfall now coming as sleet/rain instead.
• When the AMO flips back negative again we are likely to see more in the way of cold, wintry weather - however that isn't likely for some time.

This research is, of course, heavily biased to the SE - an area which has seen IMO one of the biggest changes since 1997. Areas which are less dependent on easterlies for snowfall (ie the Midlands north and west) won't have seen anywhere near as much of a change.

Well as Mr.Spock might say: That Was Fascinating. Great analysis without too much technobabble.

Ice days per year since 2000 here:

2015: 0

2014: 0

2013: 7

2012: 2

2011: 0

2010: 14

2009: 1

2008: 2

2007: 0

2006: 1

2005: 1

2004: 0

2003: 1

2002: 3

2001: 1

2000: 0

Yes, I suspect location is important. Since 1996 we've averaged around 1-2 ice days a year. That figure has actually increased post-2010 due to 2010 and 2013. We had a max of -4C on 16 January 2013.

Of course, the overall trend is warmer, but ice days haven't disappeared here. Not yet anyway!

One thing I would say though is that, looking at your Manston graph, although it doesn't go back past the 1970s, the amount of cold weather seems to reduce again back in the 70s. Reading Trevor Harley's summaries backs that up - he comments on how boring the winters of the early to mid 70s were generally.

So a big part of this could be down to the fact that your childhood coincided with a particularly cold and snowy spell, which you regarded as "normal" but was actually rather out of the ordinary. So it's not a case of "what has gone wrong" but more reverting to a more normal pattern (or at least normal for positive AMO eras).

I have the same inbuilt bias, as I am only a couple of years older than you, so I recall plenty of snowy 1980s winters!

Thats a fair shout as I recall late 70's , maybe I was just lucky

We do seem to be getting less in the way of easterly types nowadays, most of the cold spells are assoicated with Arctic airmasses moving Southwards due to a Greenland high.

I do have hope we will see the December 1938 type easterlies again though, February 2012 was so so so close to delivering but the Azores high ridged over us and an exceptionally cold air mass fell towards Spain instead.

We can get severe cold spells if the right synoptics are present, they are still capable of delivering in the winter months.

The key question is what causes Scandinavian Highs to form? Only then can we learn whether they are becoming rarer because of climate change, or some other phenomenon.

Just another post about Feb 2012...

This is what I consider to be the last true winter Scandi high:

http://www.wetter-zentrale.de/pics/archive/2012/noaa/NOAA_1_2012020100_1.gif 

Here are the coldest 850s

http://www.wetter-zentrale.de/pics/archive/2012/noaa/NOAA_1_2012020300_2.gif 

I wouldn't say it's bias - it's a fact that it's 19 years since the last deep cold easterly.

You will struggle to find such a long period without one in the archives!

Even the 20s had some:

http://www.wetterzentrale.de/pics/archive/ra/1926/Rrea00119260114.gif

http://www.wetterzentrale.de/pics/archive/ra/1927/Rrea00119271216.gif

The 30s did, the 40s did, the 60s did and of course the 70s had some too.

http://www.wetterzentrale.de/pics/archive/ra/1970/Rrea00119701225.gif

The last was the textbook white Christmas down here.

The lack of them, 19 years and counting, it's what's remarkable. Of course they've ebbed and flowed over the years, but this long gap is something else entirely.

EDIT: That 2012 one was close, but no cigar. Not only did we not get a strong easterly flow, the upper cold pool stayed well away to the east - the air which did reach us just wasn't that cold.

An interesting analysis Darren.

I think there is good evidence for what the main cause of this change is. But, for many it is a undiscussable and unacceptable answer - so we have to find a answer that we find is both discussable and acceptable.

I'm afraid, if only for the sake of the thread, I can say no more.

Which is why I'm surprised that people are surprised. The decrease in notable cold spells, regardless of their origin, is a worldwide phenomenon and the answer is absolutely obvious. Whether people choose to acknowledge that is another matter entirely.

Speaking from an IMBY point of view, growing up in the 90s and coming of age in the 00s, I didn't notice much if any difference between the two decades in terms of snow. Nearly every year had a decent fall at some point, and nearly every year had an ice day. It's only the past 2 years that have seemed particularly poor in terms of snow for me. 2013/2014 was the first winter here to not have accumulating snow at any point - and only a few days of falling snow. The February 2005 spell seemed snowier than most 90s winters. The only truly great snowfall I recall from the 90s is January 1995.

Darren's analysis is fascinating, and as a Kent resident 1976-2002 and before 1960, I'd generally support his conclusions about the absence of easterlies. Ice days too - growing up in the 50s, it seemed like every other year at least there was ice on ponds strong enough to walk on. When we moved to Chichester in 2002, I was expecting channel lows to combine with a cold pool to our east and give at least an occasional big snow dump over the South Downs, but no such luck.

Dec 2010, justly celebrated in some parts of the country, did produce the only good snowfall in that time around here - but without a good easterly flow following, it was all gone in 3 or 4 days.

A great analysis, and I've enjoyed the discussion, but I have to take issue with the idea that 19 years and counting is an even remotely "remarkable" number. 

The human mind seems to have a blind spot when it comes to accepting how unremarkable a long sequence of positive or negative events really is. Nineteen failed attempts is nowhere near enough to claim that the absence of a phenomenon as fleeting as an icy easterly is evidence that something must have changed.

For the last 24 years I've made my living from gambling (it's the reason for my interest in weather, I need to understand the odds of weather occurring which might affect sporting contests). If those years have taught me anything it's the extraordinary length adverse sequences can continue without any change whatsoever in the underlying principles governing events. You know, think of the worst possible run of ill fortune imaginable, and then accept that it can still just go on and on beyond even that, and with nothing having changed in the general background whatsoever.

You don't even need to be a gambler to recognize that. Human lives are unfortunately short enough for whole lifespans to be nothing more than runs of good or ill fortune, with people being unfairly judged as successes or failures as a consequence.

And I think the same principle applies to weather. We simply don't experience enough years to be able to judge whether our region's climate, with its basis in factors as solid as latitude, altitude and proximity to water, can somehow have shifted.

We've done well enough here despite some rather poor winters (on a nationwide scale), making the best of a bad situation as it were.

It's a fascinating read Darren, what you have put together and there must be some truth in the AMO cycles, even with some notable exceptions. 

We have had easterly flows since such as late November 2010 that gave a dumping of snow. Also there was a Scandi high and easterly during January 2010 but it was a pathetic effort.

An easterly flow or classical one being a Scandi high and Genoa low is in reality a rex block. The question is why a rex block in that location seems to be a struggle to form.

Very interesting thread and analysis.

I think one of the great easterlies of modern times must have been that which occurred throughout Feb '86. I imagine that this must have given some intense cold and snowfall to eastern parts of the UK, but here in Ireland, it proved to be one of the driest and coldest month's on record. It is really hard to imagine a true deep sourced easterly on this scale happening in our 'modern' winters. 

Chart shows mean MSLP and 850 hPa temp anomaly for Feb 1986. Will we ever see the likes again?

This is different I know, but will hopefully it will give some hope... Prior to 2009 I was wondering if I'd ever again see a decent snowfall that lasted a week+ on the ground, with snow-on-snow events topping up what fell. 1991 was the last time it had happened. (December 2000 was the nearest, but even that only last 4 days.) Then in the 4 year period from Feb 2009 to Jan 2013, the 7+ day snow cover, with accompanying snow-on-snow events, that had eluded me since 1991 occurred 6 times (Feb 09, Dec 09, Jan 10, Dec 10, Feb 12, Jan 13)! That came out the blue.... Just as a proper easterly will do some day, perhaps even this coming Feb even!

Actually I don't think Feb 1986 was very snowy. Trevor Harley refers to it as the "forgotten cold month" as it was basically just cold and dull for most. 

Terrific post, Darren.

There was a period during the 1980s when Kent seemed to be the coldest and snowiest part of the country. Now we seem to get less snow than even the south west although that is probably subjective on my part.

We have at least another 10 years of this crud until the AMO turns negative. And, even then, I do wonder whether wider changes in the global climate will scupper things for us for good.

Back when I worked on the Telegraph, I was in regular contact with Philip Eden who told me to make the most of Feb 2005 because he expected it to be one of the last major cold outbreaks for my part of the world. Winter 2009/10 and December 2010 made me scoff at his prediction but I'm now beginning to wonder whether he was right - just out by four or five years.

That's a great post Darren with plenty to think about. We seem to get phases that last longer than they seemed to or at least that is how it looks to me but why the cessation of midwinter easterlies? Maybe the AMO is one of the drivers.

Nick

That is how 90% of easterlies turn out here. Waste of time if you ask me. 

Arctic surface temperature anomalies are certainly changing pace rather rapidly since 1997. How that affects synoptic and the distribution of pressure patterns is beyond my understanding, but possibly one of a number of factors.

http://www.arctic.noaa.gov/reportcard/air_temperature.html

I certainly don't recall a classic snow event and certainly nothing like "near blizzard conditions". More like a few cm that was gone within a day. January 1996 was an absolutely horrible month. Just awful, sunless, grey and depressing with temperatures constantly stuck between 3 and 6C. 10 hours sun total for the month.

I'd love to see the back of easterly and south easterly winds for good but unfortunately they have increased in frequency in recent years. They are totally useless, unpleasant and repellant.

The death of northerlies and north easterlies is, however, very concerning and means we don't get any winter weather any more

Agree with that,the best month for cold I remember,an absolute classic

The cet record backs this up with a mean temp of -1.1c the coldest month since the 1963 winter.

As you rightly mention it was a very dry cold,hardly any snow and with a brutally cold wind chill,somedays as low as -20 parts of Poole harbour froze that month and that's never happened anytime since.

Feb`86 was,imo, the daddy of all easterlies and nothing has come close since, not down here anyway.

Whilst we are on the subject of `The Death Of The Winter Easterly`what about March 2013?? Just because it happened outside the meterlogical winter doesn`t make it any less noteable,it was still winter that year until second half of April!