An Unusual August and September?

[Timmytour]

Starting with a bold statement.......
 
If the mean CET for September turns out to be 15.1°C, it would make this one of only nine years since 1700 to have had an August and September with a difference in mean CET of 0.2°C or less

First of all......is that statement correct?

If so....

Does it feel like we've lived through "history" in the making this past month?

And as five of the last six of those years have gone on to deliver winters that rank in the top 25% of the coldest winters since 1700/01.....have we got something to look forward to?   

(The last three years being 1985,1961 and 1941)

[johnholmes]

Starting with a bold statement.......

 

If the mean CET for September turns out to be 15.1°C, it would make this one of only nine years since 1700 to have had an August and September with a difference in mean CET of 0.2°C or less

First of all......is that statement correct?

If so....

Does it feel like we've lived through "history" in the making this past month?

And as five of the last six of those years have gone on to deliver winters that rank in the top 25% of the coldest winters since 1700/01.....have we got something to look forward to?   

(The last three years being 1985,1961 and 1941)

 

I find it absolutely fascinating when these kind of numbers come up, we will know how high the correlation is 1 March 2015. Fun to wonder about it and perhaps look at similar types of 'patterns'.

Can you post the years both when the correlation occurred please T and when it did not. I am sure more than just me would be very interested to look at them?

[Timmytour]

Hi John

Like you say....just a bit of fun but interesting

The years in question were   1750 1760 1844 1891 1929 1941 1961 1985  and 2014 

I've shown them below as Year (CET diff Sept - Aug)  and ranking as "coldest winters"

In terms of "coldest winters" I've used the months December through to March being those most capable of delivering snow, and taken the sum of the mean CETs for those months ranking them.in a range from warmest to coldest with the warmest ranking as 1,
 
1750 ( 0.2)....ranked  188
1760 (-0.1)....ranked 13
1844 ( -0.2)   ranked  307
1891 ( 0.1)....ranked 264
1929 ( -0.1).. ranked 105
1941 ( -0.2)...ranked 274
1961 ( -0.2)...ranked 242
1985 ( 0.0)....ranked 243
2014 ( 0.2),,,,
Considering 2012/13, which was a winter I really wouldn't mind again, ranked 230 on the same scale, those last three years look mouth-watering!!  :-) 

Correction typo ...230th coldest year is 2012/13 not 2011/2 as previously stated!

[BornFromTheVoid]

I've had a look at similar stats, and there are a number of little statistical quirks like that.

 

The average winter following August - September CET values of between -0.2 and +0.2 is 3.1C.

The average winter following August - September CET values of between +0.3 and +0.5, is 3.0C

 

Growing the rage to include difference of -0.5 to +0.5 (20 years) still keeps the average following winter down at +3.2C, with a max of 5.8C in 1760 and -1.2C in 1683. and 13 out of the 20 lower than the 61-90 average.

[johnholmes]

I've had a look at similar stats, and there are a number of little statistical quirks like that.

 

The average winter following August - September CET values of between -0.2 and +0.2 is 3.1C.

The average winter following August - September CET values of between +0.3 and +0.5, is 3.0C

 

Growing the rage to include difference of -0.5 to +0.5 (20 years) still keeps the average following winter down at +3.2C, with a max of 5.8C in 1760 and -1.2C in 1683. and 13 out of the 20 lower than the 61-90 average.

and another interesting statistic, one I am sure winter cold lovers will relish.

I wonder just why this data suggests cold winters, something to do with blocking, but what has caused that? This might give a better clue to the meteorology/climatology behind what BFTV has shown.

 

Also interesting if you look at the current long range computer output from Met, see link below

http://www.metoffice.gov.uk/research/climate/seasonal-to-decadal/gpc-outlooks/glob-seas-prob

that data, fairly consistent over the past 3 issues I think, does not support the BFTV suggestion. But why, surely Met would have what appears to be anomalous warmth and 500mb height anomalies that support this idea? All quite intriguing.

[BornFromTheVoid]

and another interesting statistic, one I am sure winter cold lovers will relish.

I wonder just why this data suggests cold winters, something to do with blocking, but what has caused that? This might give a better clue to the meteorology/climatology behind what BFTV has shown.

 

Also interesting if you look at the current long range computer output from Met, see link below

http://www.metoffice.gov.uk/research/climate/seasonal-to-decadal/gpc-outlooks/glob-seas-prob

that data, fairly consistent over the past 3 issues I think, does not support the BFTV suggestion. But why, surely Met would have what appears to be anomalous warmth and 500mb height anomalies that support this idea? All quite intriguing.

 

I think it's just a statistical fluke really.

 

However, at a bit of a stretch, it could be related to the AMO. Results from a Met Office conference, which you can read here, suggests that a +ve AMO can promote mild/cool, wet summer. Other studies also link the +ve AMO with more -ve NAO winter values and more extreme winters in western Europe (example here and here.)

 

So the +ve AMO could cause August to be slightly cooler due to increased cloud and rain, while September could be slightly milder due to the increased heat content in the north Atlantic. These combined could mean that a +ve AMO tends to reduce the difference between the September and August CET values, while promoting colder winters. This wouldn't happen every year with a +ve AMO, but it might increase the chances?

 

I know, it's a bit far fetched, but interesting to muse over!

[johnholmes]

 

 

I know, it's a bit far fetched, but interesting to muse over!

as you say-interesting to muse over

[Roger J Smith]

The three largest increases were 1.2 in 1865, 0.9 in 1729 and 0.8 in 1784. The first two were followed by mild winters, much of 1784-85 was quite cold however.

 

Of all years with a >2.0 increase in anomalies Aug to Sep, 1750 seems to have the highest correlation from Jan to July before the anomalous Aug-Sep combination. This was also a year close to the peak of a moderate sunspot cycle (albeit a stronger one than the current declining peak).

 

Looking at all years in play, I can see a general pattern of cool, wet summers in eastern North America, which was the case in 2014 as well. There was a tendency to warmer than average autumns there.

 

Seems that the signal is probably a mixture of two regimes, one being increasing blocking and a good chance for winter cold to follow, the other being a relaxing depressed jet stream that would probably be more often than not followed by a mild winter.

 

I think this case in 2014 shows more signs of falling into the increasing blocking camp.

 

 

 

 

[Nouska]

The three largest increases were 1.2 in 1865, 0.9 in 1729 and 0.8 in 1784. The first two were followed by mild winters, much of 1784-85 was quite cold however.

 

Of all years with a >2.0 increase in anomalies Aug to Sep, 1750 seems to have the highest correlation from Jan to July before the anomalous Aug-Sep combination. This was also a year close to the peak of a moderate sunspot cycle (albeit a stronger one than the current declining peak).

 

Looking at all years in play, I can see a general pattern of cool, wet summers in eastern North America, which was the case in 2014 as well. There was a tendency to warmer than average autumns there.

 

Seems that the signal is probably a mixture of two regimes, one being increasing blocking and a good chance for winter cold to follow, the other being a relaxing depressed jet stream that would probably be more often than not followed by a mild winter.

 

I think this case in 2014 shows more signs of falling into the increasing blocking camp.

 

You should also add 1956 to your list of biggest differential. What stands out is that these summer/autumn change over ambiguities all followed a very cold winter. This year has a very small figure of difference - that seems to be just as likely to follow a mild winter - as has been the case this year.

 

What might be the artefact of a cold winter that would impact August more than September?

[Yarmy]

1784 was post Laki and also a solar minimum. Just to confuse things a little more.

[Timmytour]

Looking at the yeears with the variance of plus or minus -0.2, it seems common to see  a particularly cold February or March in the ensuing winter which has no doubt assisted the end result of a cold Dec - Mar.

All references relate to records for the actual month concerned in the years from 1700....(315 years)  references to winter are as measured Dec - Mar

1750 ( 0.2)....The February of the ensuing winter was within the coldest 25 
1760 (-0.1)....The one complete exception...1760/1 ranks as the 13th warmest winter! (Dec - Mar)
1844 ( -0.2) ...A December, Feb and March that are all ranked within the top 20 coldest for their respective months
1891 ( 0.1).....March 1892 is the 11th coldest
1929 ( -0.1)..  the only winter in the last 250 years not in the coldest 25% of winters in this selection...nonetheless Feb 1930 is only just outside the top 20% of coldest Febs 
1941 ( -0.2)...Feb 1942 is in the top 10 coldest
1961 ( -0.2)...March 1962 is the 11th coldest  (only beaten by March 2013 since)
1985 ( 0.0)....February 1986 is the 5th coldest

2014 ( 0.2),,,, If there is any correlation....and the "if" is obviously enormous, it looks like the kind of winter most of us dream about would later, rather than earlier, in the season 




 

[Timmytour]

Keeping with using August as a pivotal month of the year, and being reasonably confident now that the mean CET for October will be within 3C of the August figure, I've turned my attention to looking at the other years where this was the case.

Since 1700 it's only happened on seven previous occasions, The only year prior to 1900 being 1811, with the other six occasions being 1908, 1921,1954, 1968, 1986 and 2011.

Looking at the winter from the perspective of the sum of the monthly mean CETs from December to March, those seven years have given us five years of relatively cold ensuing winters and one exceptionally warm one in 2011, (principally because of a very warm March).The other tended to the warmer side of winters.

The average mean CET for the months Dec - March since 1700 has been just over 4.23C, though since 1900 this has risen the best part of half a degree to just over 4.67C 

In each of the five cold winters the average CET for the four months is at least one degree under the 1900 to current day average. The super "warm" winter that followed October 2011 was 1.2C above the same average which the other winter, that followed October 1921, was 0.13C over.

All these winters together produce an average monthly CET of just over 4.01C.....over 0.65C below the average of the last century. Remove the exceptional 2011 from the calculations and the average goes down to just under 3.71C....over half a degree colder than the average from 1700 to current day.

Statistical fun perhaps.....or a further pointer towards a colder set up than usual for winter?  






 

 

[cheeky_monkey]

I know I have been out of the country for 6 years but when did March become a winter month?

[Timmytour]

I know I have been out of the country for 6 years but when did March become a winter month?

 

I think it makes sense for the UK to view March as a winter month (for the purposes of stats) since the sea that surrounds us is about at its coldest during the month.

Going on records since 1700, there have been more occasions upon which March has been colder than the following December, then there have been where February has been colder then the following December.

The chances of March being one of the three coldest months of any one year are over 35%, almost double those of November.

In the last six years, the three coldest months have been a December (2010), a January (2010) and a March (2013)

 

I rest my case m'lud

 

[cheeky_monkey]

I think it makes sense for the UK to view March as a winter month (for the purposes of stats) since the sea that surrounds us is about at its coldest during the month.

Going on records since 1700, there have been more occasions upon which March has been colder than the following December, then there have been where February has been colder then the following December.

The chances of March being one of the three coldest months of any one year are over 35%, almost double those of November.

In the last six years, the three coldest months have been a December (2010), a January (2010) and a March (2013)

 

I rest my case m'lud

 

spring is always been March April May you cant move months around to suit stats or because you decide that's the way it should be...in regards to March there is no case to answer m'lud case dismissed

[Timmytour]

I'm not getting excited about the models showing winter being real cold until late on in January.

statistically speaking it's February that delivers the cold after a September and August with such similar CET means

[Lettucing Gutted]

However, at a bit of a stretch, it could be related to the AMO. Results from a Met Office conference, which you can read here, suggests that a +ve AMO can promote mild/cool, wet summer. Other studies also link the +ve AMO with more -ve NAO winter values and more extreme winters in western Europe (example here and here.)

That is very surprising.  I always thought, and many others here and elsewhere have said that the AMO is positively correlated with the NAO, not negatively correlated with it.  More logically any link between the "cool August and mild September" combo and cold winters must surely be due to enhanced and frequent blocking features with the associated highly amplified and unstable main westerly Jetstream?

[Timmytour]

Well it's a pretty chilly start to February anyway.....