Impact of a late SSW on spring

[WYorksWeather]

Hi everyone,

As mentioned a few days ago I thought it was worth doing some checking into the impacts of a late-season SSW on the spring season. This will be quite a long post, apologies in advance for that!

Definitions

Here I'm defining a late-season SSW as being any SSW from February 1st onwards that is not a final warming event. 

I'm going to cover all SSWs since 2000 for the purposes of this - I'm relying on the list via Wikipedia here up to 2021. For the years since then, we can add that there was no SSW in 2021/22 (though there was an early final warming in March), then there were late winter SSWs in 2022/23 and 2023/24.

So, we can then divide the winters since 2000 into three categories:

• No SSW
• Late-season SSW
• Other SSW

For the purposes of this, years with multiple SSWs are assigned according to their latest SSW, so a winter with a late-season SSW and an earlier SSW is assigned as late-season.

To be clear I picked the year 2000 as a convenient starting point rather than for any specific reason, and though I have my own preconceived view on this, I'm just going to see what the data shows. Similarly the threshold of February 1st for a late-season SSW is a bit arbitrary, but again no specific reason, just seems like a sensible point to call it a late-season SSW.

Breakdowns

Using those definitions, we get the following breakdowns:

Springs following a winter with no SSW:

2005, 2011, 2012, 2014, 2015, 2016, 2017, 2020, 2022

Springs following a late-season SSW:

2000, 2001, 2002, 2007, 2008, 2010, 2018, 2023, 2024

Springs following any other SSW:

2003, 2004, 2006, 2009, 2013, 2019, 2021

Analysis

Now I'll use the NCEP reanalysis plotter to have a look at composites for these springs. Four charts for each group from the top left - 500mb height anomaly, sea level pressure anomaly, temperature anomaly and precipitation anomaly.

All anomalies are relative to 1991-2020 averages.

Springs following a winter with no SSW

As a reminder, the years in scope are: 2005, 2011, 2012, 2014, 2015, 2016, 2017, 2020, 2022

Overall these springs as a group were generally quite high-pressure dominated. Slightly milder than average, around 0.5C above 1991-2020, and somewhat drier than average.

Springs following a winter with a late-season SSW

As a reminder, the years in scope are: 2000, 2001, 2002, 2007, 2008, 2010, 2018, 2023, 2024

Overall, these springs were much more blocked. A general tendency for heights to the north or north-east, and low pressure through the UK. As a result, temperatures no higher than average. Driest in the north closest to the blocking highs, and wetter further south.

Springs following a winter with any other SSW

As a reminder, the years in scope are: 2003, 2004, 2006, 2009, 2013, 2019, 2021

Overall, these springs exhibit a blocked pattern with low pressure but not quite as pronounced as those following late-season SSWs. Temperatures are somewhat colder than average in the far south, but around average elsewhere. Precipitation is broadly no signal. In broad terms, these springs are best described as mixed, but with a slight cold signal.

Conclusions

Overall I think there are a few broad statements we can make based on these. To get the obvious conclusion out of the way first, springs following winters without an SSW are clearly the warmest and driest.

The other two types are more mixed. Both feature a lot of northern blocking. Those following late-season SSWs tend to feature around average temperatures overall but with a lot of very wet weather - a clear above average precipitation signal. Those following all other SSWs have the coldest signal with below average temperatures in the south, but not quite as unsettled as those following late-season SSWs.

My preference

Of course, spring weather preferences are probably the most varied on here, but for my personal opinion I am looking for dry and warm weather in spring (in that priority order). Based on the charts above I would rank from best to worst:

1. Springs following a winter with no SSW
2. Springs following any other SSW
3. Springs following a late-season SSW

Final points

Just to say that there was a fair amount of manual work involved to assign the years and input them into the plotting tool, so it's definitely possible I've made a mistake somewhere. If you spot one let me know and I'll correct it, whether it's an incorrectly assigned year or the wrong chart.

Hopefully this will provoke some interesting discussion. I do think it makes a decent case for the hatred of late-winter SSWs!

[SunSean]

Nice analysis @WYorksWeather

I get blasted quite a bit by certain members for my hatred of late SSW's but your analysis pretty much backs up my concerns. Id rather not have one of these poxy late SSW's! Sure you could potentially still get a decent Spring out of one like 2007 but it's not a roll of the dice I would rather take.

To get a late one 3 years in a row would be obscene and if we do get one and Spring turns out pants again and we end up with members STILL saying "a late SSW does not mean Spring is in trouble" or whatever then I'm putting any SSW champions on block so I never have to hear their gaslighting again lol.

[richie3846]

 WYorksWeather thanks for the analysis, I wanted to do something similar last year, but didn't have the know-how. I think this helps resolve some of the doubts and question marks many members have had, relating to late SSW events. 

I was one of the members who felt like we needed more information before we could be sure about the claims that a late SSW is a spring disaster. Your post clearly demonstrates that indeed, these events are much more likely to warmth spring, though of course there will always be the exception to this.

I think this post is so informative, that we could add new springs after they happen, and continue this science based approach to help us understand the impact of said events on our springs. To my mind, this is a much better approach than the various disagreements many have had on the subject in recent times. 

[danthetan]

This explains nicely what the chances are of a SSW this year with background drivers . Currently low probability but still possible. 

[WYorksWeather]

 SunSean Yep, there will definitely always be exceptions to any rules - if the weather were as simple as taking a few background factors and making some sort of composite then we'd be able to predict months in advance.

Looking at some of the years in more detail, I think what we can say is that no spring following a winter without an SSW was truly dreadful - 2005 and 2012 were probably the most mixed. That explains why this signal is the most solid.

For the winters with any other type of SSW, the extreme outlier is 2013, probably the main driver of the cold signal overall. One of the coldest springs in many decades. However, most of them were often quite dry or had dry periods even if cold, and of course 2003 bucked the trend in the opposite direction with a very warm and dry spring overall.

For late-winter SSW springs, 2007 was a very good early spring but a dreadful May which of course then setup the pattern for the summer. 2018 was the opposite with a very cold first half of the spring, then a warmer second half. But overall there are some pretty dreadful ones in there.

 

[Maz]

@WYorksWeather Really good analysis, thank-you.  I've now got my fingers firmly crossed that there is no SSW in February!

It would be interesting to see how sunshine levels vary.  I feel that due to the rapidly strenghening sun in Spring, the most important factor in how spring feels, is how much sun there is.  Almost any sunny spring day feels pleasant, even it starts and ends frosty and only gets to the low teens*.  Conversely, due to the lag in sea/air temps catching up with the stenthening sun, cloudy spring days invariably feel unpleasant.

*Unless wind is very strong.

[WYorksWeather]

 Maz Sunshine level is a tricky one - I can't find anything useful to create a composite on cloudiness or sunshine unfortunately.

Agree that it makes an important difference to the feel, for two main reasons. The first is as you say the feel of the sun makes a given temperature feel warmer.

The second is also that those sorts of days tend to have wider diurnal ranges due to the lack of cloud.  For example, a day that started cold and frosty (min of 2C) breaking to a warm and sunny afternoon (max of 16C) would have a mean of 9C. Another day which stayed completely overcast with warm upper air might have a min of 10C and a max of 12C. That would be a mean of 11C, but the first day would probably feel warmer (especially if you weren't heading out early) and would be by most people's standards a subjectively 'better' spring day, despite the lower mean.

 

[feb1991blizzard]

This might help people here, you can get various fields of up to 60 days after an SSW.

NOAA CSL: Chemistry & Climate Processes: SSWC

CSL.NOAA.GOV

---

NOAA CSL: Advancing scientific understanding of the chemical and physical processes that affect Earth's atmospheric composition and climate.

 

Edited January 26 by feb1991blizzard

[WYorksWeather]

Ignore - link added above.

Edited January 26 by WYorksWeather

[WYorksWeather]

 feb1991blizzard Thanks for posting this. The list here of SSWs is easier to read I think than the Wikipedia one:

NOAA CSL: Chemistry & Climate Processes: SSWC

CSL.NOAA.GOV

---

NOAA CSL: Advancing scientific understanding of the chemical and physical processes that affect Earth's atmospheric composition and climate.

Having had a look the plotting tool doesn't really offer an equivalent to the NCAR one to create composites from multiple years, and it also doesn't let me obviously split by SSW. If I wanted to do that I would probably have to download the data and mess about with it, which is an undertaking I don't have time for right now!

[feb1991blizzard]

 WYorksWeather  I must admit my interest in SSW's is sometimes tempered by access to charts, i would like to have re-analysis and real time model output data on 7-10 levels of the atmosphere as postage stamps on one page and just be able to flick through the timeframes.

[WYorksWeather]

 feb1991blizzard Yep, the way I had to do the analysis above was somewhat irritating to say the least. SSW data is annoying as well because there are different re-analyses and different definitions of an SSW.

[Catbrainz]

It would seem that a late SSW does tend to lead to a poor spring. You’ve got quite a few very bad springs like 2000 and 2024 and the rest are mediocre at best. A late SSW tends to create NAO- set ups which is often a spring killer. (Generally for good springs you want NAO+). While if you’ve had a strong PV all winter that can be good for spring as it means more mid latitude blocking over northern blocking. 
 

 

[Summer8906]

What I have noticed is that if late winter (i.e. February) is cold, the following early/mid spring tends to be cold, wet or both.

That pattern holds true in the cold Februaries of 1979, 1981, 1983, 1985, 1986, 1994 (not always cold - rather mixed), 1996, 2010 (to a degree), 2013, and 2018.

I can't think of an example of a cold Feb followed by a warm/dry spring.

It would be interesting to hear if late SSWs were present in any of those 20th-century years.

Edited January 27 by Summer8906

[WYorksWeather]

 Summer8906 There was an SSW in Feb 1979, one in both Feb 1981 and March 1981, no SSW at all in 1982/83, one in Jan 1985, no SSW in 1985/86, 1994, or 1996. Therefore, we end up with a list consisting of late SSW, late SSW, no SSW, other SSW, no SSW, no SSW, no SSW.

It would be interesting to extend this analysis back a bit further and see if the pattern still holds, but will probably have to wait until I have a bit more time. What I would probably do is grab the table and at least automate the process of producing the year classifications, then just enter them into the NCAR/NCEP plotting tool. The main limitation is that without doing some more hacking around the tool will only let me have 20 years per category, which will limit my ability to go all the way back to the 50s - that would have to be done in batches.

A project for when I have time though - probably not in the next few days to be honest.

 

[Summer8906]

 WYorksWeather No SSW in 1983 is an interesting one as that seems like a classic SSW sequence. Dec and Jan are very mild and zonal and then it suddenly becomes much colder and more blocked in February. Then there was the typical dull and wet SSW-style spring. That mild-to-cold Jan to Feb progression is unparalleled in any other year I remember - have there been any investigations on the reasons for the sudden 1983 switch round?

Edited January 27 by Summer8906

[WYorksWeather]

 Summer8906 I'm not sure about that year specifically. There was the El Chichon volcanic eruption in 1982 though, so it's possible I suppose that there was some weirdness that year because of that, it was a reasonably sizeable one.

 

[LetItSnow]

 Summer8906 1983 was a weird one across the pond as well. Winter 1982/1983 was a blowtorch winter across a lot of the northern states and into Canada (unusual after a run of mostly very cold winters. December 1982 was the December 2015 on the Eastern Seaboard) meanwhile down south into Texas etc it was quite "cold", this pattern persisted through March, then there was a very cold April and May for most with blizzards during their spring. I believe the atmosphere was acting very unusually because of one of the strongest ever El Ninos on record to this day.

Interesting how we shared the mild winter, cold spring, hot summer with our American neighbours that year as 1983 was a very hot summer for much of the States too.

[raz.org.rain]

Semi-related question, but what are the current chances of a SSW occurring this season? Last I heard it was unlikely, but I've not heard anything else about it since then.

[LetItSnow]

Also adding (I know this is about SSW's but to further answer @Summer8906's question on 1983) that I've noticed a strong trend for strong Nino years to have wintry and/or poor conditions show up in the spring, especially after a mild winter. I think in 1983 it just happened earlier than usual so it swung back wildy in February. But you can see 1998 and 2016 also do this. Didn't we also have a strong Nino last year too? Wasn't a cold spring but it was dull and unsettled. 

Anyway, don't get hopes up on a marvellous spring if you have a super Nino! 

On 25/01/2025 at 22:16, WYorksWeather said:

2000, 2001, 2002, 2007, 2008, 2010, 2018, 2023, 2024

An extra point on these springs following a late season SSW is a lot of them do feature spells of notably above average temperatures. May 2000 saw a hot first half, May 2001 was very good, April 2002 was warm and fairly sunny, 2007 goes without saying despite going wrong in May. May 2008 was very hot. Spring 2010 was nowhere near as bad as people think, 2018 had an April heatwave and May of couruse - so to add some perspective it's not really a guarantee of a write off. I think there was more at play in why 2023 and 2024 were so grim.

Plus I find that composite charts are often not as useful as we'd like them to be (even though I do find them interesting) as they're not a forecast but just a map of an average of where pressure has been, and a slight deviation in wind direction or pressure etc could completely change the whole surface conditions. So I do think it's interesting but I think some further analysis would be needed.  

[baddie]

 LetItSnow Tbh I find Spring 2023 overhated. The dire weather from the SSW was concentrated in March, and April was changeable but much better than March, and May was good apart from the 2nd week, of course it came after a decent February and proceeded a June which was about as superb as you can get out of an early summer month. Spring 2024 definitely deserves its hate (Didnt even get a full sunny day until.... April 21st)

Back to topic, February SSW's are just simply not worth the risk of dire March (and possibly April). Best to get one in January, where any dire weather from the impact doesnt last deep into the spring

[WYorksWeather]

 LetItSnow I think what I would say though is it's notable that there's not a single one of those springs that would be classified as very good. You're right that some of them had brighter spots within them. I was never expecting that these charts would show a completely terrible signal for late-winter SSW springs, just a below average one. I would also say that I specifically put rainfall in there as well, because some of them were mild and wet. But in spring it is really the dry weather that helps with the feel, especially if the sun comes out from mid-March or so onwards.

I think I'd still stand by the idea that all else equal, if you're hoping for a warm and dry spring, you don't want a late-winter SSW.

[Catbrainz]

I have noticed that Nino springs tend to be quite poor too which makes senses that Nino also often creates NAO- set ups in spring. Here's 1983s set up. Gotta love west based NAO-s eh? (It does look like a very archetypical post late SSW spring). Probably not as wet as say a spring 2000 or 2024 given the low is centred over the North Sea rather than over the Biscay so more Elys and NElys over SWlys (Probably still fairly gloomy though). Reminds me of spring 2008 set up wise actually. 

NAO- is always poor in spring unless you can get a really SE based one like an April 2021 where the Icelandic block extends SE torwards us (Although at that point its barely a NAO- itd be more of a Scandi or north Euro high). 

[Catbrainz]

 baddie I did always find spring 2023 very overhated myself. I would put it above say 2018 or 2021. At least 2023 had the good grace to get all the really bad weather out of the way in March (If a spring month is going to be really bad March is less of a sting than a bad April or May). 

[LetItSnow]

 Catbrainz Spring 2023 reserved its worst conditions for the south-east. April 2023 which IIRC was rather dry in Scotland but very wet down here. Also the easterlies in May 2023 were quite cool in the east. Even early June was cool in the east.

 WYorksWeather By no means do I disagree with your point, it's just warning some to not be too fatalistic. Thankfully there's apparently not any signs for one this February.