Jump to content
Snow?
Local
Radar
Cold?
IGNORED

Winter 2018/19 – cold long range models, SSW events and favourable global drivers… where did it go wrong? A review.


Derecho

Recommended Posts

Posted
  • Location: Camborne
  • Location: Camborne

It has crossed my mind that perhaps not everyone is completely au fait with what the QBO is. I say this because I’m one of them. So this succinct explanation taken from “Stratosphere Troposphere Interactions” by K Mohanakumar may assist  If not, not to worry

Role of Quasi-Biennial Oscillation in Coupling Process

dominant form of variability is a quasi-periodic (2-3-year) wave-driven descending zonal mean wind reversal, called the quasibiennial oscillation (QBO). The period of the QBO at any level varies from 2 to 3 years, and it could probably be predicted for about 1 year. The QBO is observed to affect the global stratospheric circulation. It modulates a variety of tropical and extratropical phenomena including the strength and stability of the wintertime polar vortex, and the distribution of ozone and other gases (Baldwin et al. 2001). 

Figure 9.5 illustrates the schematic picture of the dynamical overview of the QBO during northern winter (Baldwin et al. 2001). In the tropics, the stratospheric QBO is driven by the upward propagating gravity, inertia gravity, and Kelvin and Rossby gravity waves. In the middle and high latitudes, it is maintained by the planetary-scale waves. The contours in the tropics are similar to the observed wind values when the QBO is easterly. It can be seen that the QBO extends to the mesospheric region and even above 80 km.

The QBO is driven by the dissipation of a variety of equatorial waves and gravity waves that are primarily forced by deep cumulus convection in the tropics.  The stratospheric QBO effects extend to Earth’s surface during northern midwinter. There is also observational evidence that the QBO modulates the depth of the troposphere in the tropics and subtropics, affecting convection, monsoon circulations, and hurricanes. Although the amplitude of the QBO decreases rapidly away from the equator, observations and theory show that the QBO affects a much larger region of the atmosphere.

Through wave coupling, the QBO affects the extratropical stratosphere during the winter season, especially in the northern hemisphere where planetary wave amplitudes are large. These effects also appear in constituents such as ozone. In the high-latitude northern winter, the QBOs modulation of the polar vortex may affect the troposphere through downward penetration. Tropical tropospheric observations show intriguing quasi-biennial signals which may be related to the stratospheric QBO. The QBO has been linked to variability in the upper stratosphere, mesosphere, and even ionospheric F layer (see Fig. 9.5).

the paper referenced

https://www.researchgate.net/publication/11709083_Stratospheric_Harbingers_of_Anomalous_Weather_Regimes

qbo.thumb.JPG.95d25b931c44870ae6ea259f02d55d98.JPG

  • Like 2
  • Thanks 3
Link to comment
Share on other sites

Posted
  • Location: st albans
  • Location: st albans

Long range forecasts generally show mean height anomolys as their predictive tool for us who don’t have access to more detail than that

for nw Europe to be cold, that requires either a decent scandi high anom in tandem with a euro trough or a high anom to our nw/n without a w euro high anomoly of any kind 

The size of the anoms is important. Given the usual greeny heights are low, you need a big high anom to show persistent decent blocking up there and the opposite for Europe. 

i just think that we see the Lr charts and the oranges/blues in places we would like them to appear and make dodgy deductions on the back of them. add to that the output isn’t very reliable anyway and you have a recipe for over exuberance on forums like this! 

having said that, we will look at the monthly anoms for winter in a weeks time and perhaps see that we just didn’t get the luck - again. 

Edited by bluearmy
  • Like 1
Link to comment
Share on other sites

4 hours ago, Nick F said:

Also Nina type patterns for Europe means most cold and snow is confined to NE/E/SE Europe.

No statistically significant correlation between the winter multivariate ENSO index and winter CET, however since winter 1950-1, 7/10 coldest CET occurred with negative ENSO (i.e. tending towards La Nina).

  • Like 1
Link to comment
Share on other sites

Posted
  • Location: Hull
  • Weather Preferences: Cold Snowy Winters, Hot Thundery Summers
  • Location: Hull

Seems fitting that this winter ends with such an exceptional mild spell. Again if that cold spell in January came off things would be completely different IMO but the repeated pouring of cold air into the North Atlantic is leading to the mild southerly-westerly winds.

Maxes of upto 17-18C over a few different areas today even in Hull maxima were around 3C higher then forecast. There has got to be a good chance of the winter record going on Tuesday which was 19.7C in February 1998.

We won't be able to beat the 12 month rolling CET record over the next few months though, that is still May 2006-April 2007.

 

Link to comment
Share on other sites

Posted
  • Location: Liphook
  • Location: Liphook

Yeah that 12 months was amazing for warmth. It includes a record breaking July and September, 3rd warmest October, 4th warmest Jan and 2nd warmest April. Several other months such as Dec 06 rank highly as well but not quite as noteworthy as those. I think it's one of the few times the rolling CET has been above 11c as well.

Link to comment
Share on other sites

Posted
  • Location: Worcestershire/Warwickshire/North Oxfordshire
  • Location: Worcestershire/Warwickshire/North Oxfordshire

This Winter did not go wrong.

What did go wrong was people trying to predict the weather more than 7-days in advance, which is always mostly pie in the sky theories.

Too many people hanging on to the long range Metoffice forecast, bellieving it to be gospel.

SSW does not guarantee a cold spell.

I experienced three snow events this winter, but they were only short term one day events, and probably only only because my job involves travel

throughout the UK. But I am happy I have seen snow this winter.

We are constantly reminded about global warming, so we are really lucky to see snow at all in the UK.

 

 

  • Like 1
Link to comment
Share on other sites

Posted
  • Location: Ossett, West Yorkshire
  • Location: Ossett, West Yorkshire

Although SSW increases the chance of a cold spell, it does not guarantee that blocking will set up favourably to influence our small part of the world with cold.  I wonder if anyone has any data to indicate the number or proportion of SSW events that brought notable cold spells to the UK?  

Link to comment
Share on other sites

Posted
  • Location: Ossett, West Yorkshire
  • Location: Ossett, West Yorkshire
On ‎24‎/‎02‎/‎2019 at 17:28, kold weather said:

Yeah that 12 months was amazing for warmth. It includes a record breaking July and September, 3rd warmest October, 4th warmest Jan and 2nd warmest April. Several other months such as Dec 06 rank highly as well but not quite as noteworthy as those. I think it's one of the few times the rolling CET has been above 11c as well.

In that 12 month period, including all the notable months that you mention, every month was much warmer than average, with only one month in that 12 month period being close to average (Aug 2006).  The 12 months from May 2006 to April 2007 were virtually the warm equivalent of 1740.  

The last occasion that the 12 month rolling CET reached 11*C was in the period from Dec 2013 to Nov 2014 (11.01), and in this period every month, with the exception of Aug 2014, was much warmer than average. There is a significant chance that we could surpass this during the next month or two, and see the warmest 12 month period in the CET since 2007.

Edited by North-Easterly Blast
Link to comment
Share on other sites

Posted
  • Location: Ossett, West Yorkshire
  • Location: Ossett, West Yorkshire
33 minutes ago, 'ColdIsBest' said:

This Winter did not go wrong.

What did go wrong was people trying to predict the weather more than 7-days in advance, which is always mostly pie in the sky theories.

Too many people hanging on to the long range Metoffice forecast, bellieving it to be gospel.

SSW does not guarantee a cold spell.

I experienced three snow events this winter, but they were only short term one day events, and probably only only because my job involves travel

throughout the UK. But I am happy I have seen snow this winter.

We are constantly reminded about global warming, so we are really lucky to see snow at all in the UK.

 

 

The Met Office have done well at times in LRFs for cold spells, they were consistent from early Feb 2018 that a cold easterly pattern would take hold.  Thinking back to the Dec 2010 severe spell the UKMO were consistent in their LRF from early Nov 2010 that a severe cold spell was on the way.  Although at times they have not always spotted serious cold, if I remember winter 2009-10 overall turned out to be much colder than the UKMO suggested in their LRF at the time.  I am not sure that the UKMO were consistent in their LRFs in predicting the cold spells in the 2012-13 winter either.

Link to comment
Share on other sites

Posted
  • Location: East Ham, London
  • Location: East Ham, London

Evening all

I thought Singularity made an excellent observation over in the Stratosphere thread. I think the SSW wrecked our winter - normally we got prods of Wave 1 warming which serve to gradually weaken and destabilise the PV. This year we got one intense burst and then nothing - the downwelling failed and indeed served only to strengthen the residual vortex.

Had the Wave 1 occurred later in the winter it might have prevented the PV from re-establishing but it didn't and as it was still midwinter the PV rallied and remains incredibly strong even now.

  • Thanks 1
Link to comment
Share on other sites

Posted
  • Location: Wyke regis overlooking Chesil beach.
  • Weather Preferences: Snowfall
  • Location: Wyke regis overlooking Chesil beach.

As others have stated nothing went wrong in terms of winter 2018/19 from an atmospheric point of view the atmosphere responded exactly as it should to the way the drivers actually played out as opposed to what many ( including myself) thought was the way they should play out.

What did go wrong was the interpretation of the effect the various atmospheric drivers would probably have on our tiny set of islands.

That goes from amateurs like me all the way up to the professionals at the met office. Some one in another post has pointed out that we should not treat the Met office 15-30 day extended as gospel and indeed we should not, but we do have a right to expect it to come somewhere near to fruition when it repeatedly suggests the likelyhood of high pressure to build to the north or over scandi as it pretty much did for six weeks from christmas to mid feb.

I personally feel that what was got wrong was the weighting given to the various drivers. Those drivers need to be weighted in the correct way and in a very narrow of accuracy within a very broad band of context for us to get severe or long lasting cold/snow in the UK.  

In a way I shouldn't complain as I had a very decent 10cm snowfall here in central south Dorset just 5 miles from the channel coast at a hieght of just  132m asl.

The joke of it all is that the snowfall far from being caused by a big scandi high as kept being progged, came instead from a low pressure system running into the base of a big scandi trough. 

  • Like 1
Link to comment
Share on other sites

Posted
  • Location: Ossett, West Yorkshire
  • Location: Ossett, West Yorkshire

At the start of this winter when we saw what was the second warmest December CET in the last 30 years, I did some research on similarly mild Decembers to 2018, and what the rest of each winter was like.  Given that Dec 2018 had a CET of 6.9, and the fact that there has never been a December with a CET of 7.0 or 7.1, I looked at all Decembers that had a CET between 6.6 (the low end) and 7.2 (the high end), which would fit 2018 just in the middle of this range of Decembers with a similar CET to 2018, and looked at what the CETs were for the following January and February.  Here are my results:  (the December CET is the first value, January the second, and February the third value)

1953-54: 6.9  2.9  2.6
1918-19: 6.9  2.9  1.9
1833-34: 6.9  7.1  5.6
1827-28: 6.9  5.1  5.2
1954-55: 6.8  2.6  1.2
1924-25: 6.8  5.3  5.2
1806-07: 6.8  2.8  3.7
1942-43: 6.7  4.9  6.1
1912-13: 6.7  4.5  4.8
1971-72: 6.6  3.9  4.3
1795-96: 6.6  7.3  4.7
1842-43: 7.2  4.0  1.9
1868-69: 7.2  5.6  7.5
1900-01: 7.2  3.5  2.3

Looking at the above list of Decembers that had a similar CET to 2018 (6.9 -0.3 - +0.3), the results are very mixed in relation to the rest of the winter, with some remaining very mild (1868-69, 1795-96, 1833-34), and some that turned much colder (1953-54, 1954-55, 1918-19, 1900-01, also another one with a cold Feb after an average Jan, also some that remained rather mild in Jan and Feb though unexceptionally so (1924-25, 1827-28), and another couple that were close to or slightly above average in Jan / Feb (1912-13, 1971-72).  Putting it all together, the average January CET following a Dec CET from 6.6 to 7.2 is 4.5*C, and the average February CET following a Dec CET in this range is 4.1*C.

When it comes to where the rest of the 2018-19 winter will fit into the above list after a December with a similar CET to 2018, I would say that the closest match will be 1942-43.  This winter had a December just short of but similar to 2018, then a Jan a bit colder but slightly above 2019, then a mild February but not quite as mild as 2019 is likely to come out at.

  • Like 2
Link to comment
Share on other sites

Posted
  • Location: Ossett, West Yorkshire
  • Location: Ossett, West Yorkshire
1 hour ago, Broadmayne blizzard said:

As others have stated nothing went wrong in terms of winter 2018/19 from an atmospheric point of view the atmosphere responded exactly as it should to the way the drivers actually played out as opposed to what many ( including myself) thought was the way they should play out.

What did go wrong was the interpretation of the effect the various atmospheric drivers would probably have on our tiny set of islands.

That goes from amateurs like me all the way up to the professionals at the met office. Some one in another post has pointed out that we should not treat the Met office 15-30 day extended as gospel and indeed we should not, but we do have a right to expect it to come somewhere near to fruition when it repeatedly suggests the likelyhood of high pressure to build to the north or over scandi as it pretty much did for six weeks from christmas to mid feb.

I personally feel that what was got wrong was the weighting given to the various drivers. Those drivers need to be weighted in the correct way and in a very narrow of accuracy within a very broad band of context for us to get severe or long lasting cold/snow in the UK.  

In a way I shouldn't complain as I had a very decent 10cm snowfall here in central south Dorset just 5 miles from the channel coast at a hieght of just  132m asl.

The joke of it all is that the snowfall far from being caused by a big scandi high as kept being progged, came instead from a low pressure system running into the base of a big scandi trough. 

What I interpret from this is that even in a winter where the background signals are good (eQBO, low solar activity, no strong ENSO anomaly either way, a decent SSW), it is of no guarantee that blocking will set up favourably to influence our small part of the world with severe cold spells.  That said, certainly for a winter where the background signals were not poor in the way that they were in 2015-16 or at the time of solar maximum in 2013-14, when a much milder than average winter was favoured; winter 2018-19 has certainly been a very poor and disappointing affair for cold in the UK.

Link to comment
Share on other sites

Posted
  • Location: North Wales Riviera
  • Location: North Wales Riviera
19 hours ago, Broadmayne blizzard said:

That goes from amateurs like me all the way up to the professionals at the met office. Some one in another post has pointed out that we should not treat the Met office 15-30 day extended as gospel and indeed we should not, but we do have a right to expect it to come somewhere near to fruition when it repeatedly suggests the likelyhood of high pressure to build to the north or over scandi as it pretty much did for six weeks from christmas to mid feb. 

Have you heard of the "Gambler's Fallacy"? https://en.wikipedia.org/wiki/Gambler's_fallacy It explains perfectly why because there's 100 times where there's a 20% chance of cold weather why there could be 0 cold spells or 100 cold spells.

  • Like 1
Link to comment
Share on other sites

Posted
  • Location: Crymych, Pembrokeshire. 150m asl
  • Weather Preferences: Extremes of all kinds...
  • Location: Crymych, Pembrokeshire. 150m asl
On 26 February 2019 at 21:56, North-Easterly Blast said:

At the start of this winter when we saw what was the second warmest December CET in the last 30 years, I did some research on similarly mild Decembers to 2018, and what the rest of each winter was like.  Given that Dec 2018 had a CET of 6.9, and the fact that there has never been a December with a CET of 7.0 or 7.1, I looked at all Decembers that had a CET between 6.6 (the low end) and 7.2 (the high end), which would fit 2018 just in the middle of this range of Decembers with a similar CET to 2018, and looked at what the CETs were for the following January and February.  Here are my results:  (the December CET is the first value, January the second, and February the third value)

1953-54: 6.9  2.9  2.6
1918-19: 6.9  2.9  1.9
1833-34: 6.9  7.1  5.6
1827-28: 6.9  5.1  5.2
1954-55: 6.8  2.6  1.2
1924-25: 6.8  5.3  5.2
1806-07: 6.8  2.8  3.7
1942-43: 6.7  4.9  6.1
1912-13: 6.7  4.5  4.8
1971-72: 6.6  3.9  4.3
1795-96: 6.6  7.3  4.7
1842-43: 7.2  4.0  1.9
1868-69: 7.2  5.6  7.5
1900-01: 7.2  3.5  2.3

Looking at the above list of Decembers that had a similar CET to 2018 (6.9 -0.3 - +0.3), the results are very mixed in relation to the rest of the winter, with some remaining very mild (1868-69, 1795-96, 1833-34), and some that turned much colder (1953-54, 1954-55, 1918-19, 1900-01, also another one with a cold Feb after an average Jan, also some that remained rather mild in Jan and Feb though unexceptionally so (1924-25, 1827-28), and another couple that were close to or slightly above average in Jan / Feb (1912-13, 1971-72).  Putting it all together, the average January CET following a Dec CET from 6.6 to 7.2 is 4.5*C, and the average February CET following a Dec CET in this range is 4.1*C.

When it comes to where the rest of the 2018-19 winter will fit into the above list after a December with a similar CET to 2018, I would say that the closest match will be 1942-43.  This winter had a December just short of but similar to 2018, then a Jan a bit colder but slightly above 2019, then a mild February but not quite as mild as 2019 is likely to come out at.

I have a lot of respect for statistics generally and you can often make predictions in many fields using them, but looking at your results I would suggest that you can't use one month's CET's to predict another month, or to forecast the winter ahead, no matter how far back the records go.  The results for January and February just seem completely random, which is really the story of the U.K. weather, month to month.  The best use of the CET records is to track the changes in average temperature over the decades.  No doubt this data is used among other things by the long range climate models to predict where overall temperatures are likely to be in the future.

  • Like 2
Link to comment
Share on other sites

On 06/03/2019 at 15:43, Sky Full said:

I have a lot of respect for statistics generally and you can often make predictions in many fields using them, but looking at your results I would suggest that you can't use one month's CET's to predict another month, or to forecast the winter ahead, no matter how far back the records go.  The results for January and February just seem completely random, which is really the story of the U.K. weather, month to month.

Have to disagree with this, a degree of climatic persistence is evidenced in the data, as should be expected from the duration of patterns of low-frequency intraseasonal variability.

From a quick look at the last 100 years 1920-2019 (arbitrarily chosen, earliest CET values probably poorer quality, and note data should be detrended for a full analysis) there is a weak significant correlation between January and February CET of 0.36 (it was 0.41 for 1919-2018).

If randomly distributed then there would be a 25% chance of both months being below average, 25% above average and 50% of them being different. 26/100 years had both months below average so close to random. However, 37% had both months above average - so a milder January is more likely to be followed by a milder February than would be expected by chance.

Some of these months though may be very close to the average and not register as particularly extreme, and when looking at months that are 1 standard deviation from normal something interesting is apparent. Only once was a very mild January followed by a very mild February (1990) and it is just as uncommon to flip from very mild / very cold (1945) or very cold / very mild (1983). However, a very cold January is seeming quite likely to lead to a very cold February, occurring 6 times in 100 years (1929, 1942, 1947, 1963, 1979 and 1985) and on almost half the occasions (13 v. cold Januaries). This may be because of the stubbornness of blocking patterns perhaps.

Finally, from the linear relationship between the months, using January as a predictor of February has a RMSE of 1.73°C which is marginally better than predicting purely by the February mean, RMSE 1.89°C. The linear relationship may not be stationary though, so its performance may differ over time.

Edited by Interitus
  • Thanks 1
Link to comment
Share on other sites

Posted
  • Location: Crymych, Pembrokeshire. 150m asl
  • Weather Preferences: Extremes of all kinds...
  • Location: Crymych, Pembrokeshire. 150m asl
On 7 March 2019 at 17:30, Interitus said:

Have to disagree with this, a degree of climatic persistence is evidenced in the data, as should be expected from the duration of patterns of low-frequency intraseasonal variability.

From a quick look at the last 100 years 1920-2019 (arbitrarily chosen, earliest CET values probably poorer quality, and note data should be detrended for a full analysis) there is a weak significant correlation between January and February CET of 0.36 (it was 0.41 for 1919-2018).

If randomly distributed then there would be a 25% chance of both months being below average, 25% above average and 50% of them being different. 26/100 years had both months below average so close to random. However, 37% had both months above average - so a milder January is more likely to be followed by a milder February than would be expected by chance.

Some of these months though may be very close to the average and not register as particularly extreme, and when looking at months that are 1 standard deviation from normal something interesting is apparent. Only once was a very mild January followed by a very mild February (1990) and it is just as uncommon to flip from very mild / very cold (1945) or very cold / very mild (1983). However, a very cold January is seeming quite likely to lead to a very cold February, occurring 6 times in 100 years (1929, 1942, 1947, 1963, 1979 and 1985) and on almost half the occasions (13 v. cold Januaries). This may be because of the stubbornness of blocking patterns perhaps.

Finally, from the linear relationship between the months, using January as a predictor of February has a RMSE of 1.73°C which is marginally better than predicting purely by the February mean, RMSE 1.89°C. The linear relationship may not be stationary though, so its performance may differ over time.

A fascinating result which just goes to show how statistics can reveal otherwise hidden information if you are willing to take the time to analyse them correctly, which I didn't!  In my defence I still feel that this particular data set still leaves the theory to be proven especially in respect of a relationship between v. cold Januaries and v. cold Februaries: 50% related and 50% unrelated out of 13 seems closer to random than a trend.  However, I can see that there is an argument for cold Januaries to be followed by cold Februaries (and probably vice versa) due to persistent weather conditions lingering during the climatically coldest months.

  • Like 1
Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...