^^^^^

Thanks GW! 

Fascinating stuff - I'm shocked that the CET will end up in the 13's and we now go 10 years without a 14C CET finish for September! 

Those analogues do look very interesting. I'll get James (my analogues guru) to have a look and see what we can come up with. 

Cheers - A big thanks Simon there for such an informative post.

Thank you for the info GW.   Might have to look around for next month. Compared to my 5 year average.

Met Office Hadley    15.1c     Anomaly     1.1c. Provisional to 20th.

Metcheck                 15.06c   Anomaly     1.34c

Netweather              15.08c   Anomaly     1.99c

Cheadle Hulme         14.0c     Anomaly     0.4c

Hexam                     13.0c     Anomaly     0.16c

Mansfield Woodhouse 15.0c    Anomaly    1.22c

Peasedown St John 15.44c    Anomaly     0.47c

Treviskey Redruth   15.4c.    Anomaly     0.66c.

1 steady   9 Down  

Mean Of My 10 Watched Stations  15.25c.

The high doesn't quite get far east enough and then retracts west a bit - so near but then so far! Only a week or so ago a finish in the 15s was still plausible - how things have changed... and yet in terms of the broad details, not really. It's the exact way ex-Florence interacts with the polar jet; splitting apart, firing a low across the UK and then kicking off amplification that pumps up a ridge west of us instead of over or just east of us as was otherwise looking to be the case.

This happenstance is well-timed relative to the results of a study I've just completed; the CET returned no significant correlation with any predictors, despite SLP patterns and even rainfall returning a few. This was a disappointing find, as while it's sort of fun seeing whether you happen to land near the mark or not but I was hoping it would have some scientific value too.

So the only real question left is... are you feeling lucky, punk? 

I wouldn't be suprised if we got a notable cold spell in November.

The signals for an average or slightly cooler September were there at the end of August, especially so after the decline towards Autumn started early this year...circa 9th August.

But I have to say, it’s been such a great weather year that it’s made it particularly exciting and interesting on the CET threads!  Nothing at all would surprise me this year now!  

Met Office Hadley    14.9c.    Anomaly    1.0c. Provisional to 21st.

Metcheck                 14.87c   Anomaly    1.14c

Netweather              15.54c   Anomaly    1.85c

Cheadle Hulme         14.6c     Anomaly    0.56c

Darwen                    14.6c    Anomaly     1.44c

Hexam                     13.6c    Anomaly     0.64c

Mansfied                   15.4c   Anomaly     1.62c

Peasedown St John 15.21c   Anomaly   0.24c     

Treviskey Redruth 15.3c   Anomaly    0.56c.

4 Up and 6 Down  Average of my 10 watched stations 15.25c.

Cloud amounts at night will be critical over the next week or so. Meanwhile October looking chilly. 

Referring back to GW's link between late September cool spells and cold winter weather, 1984 could be added to the list. Although the last 10 days were lifted by the last few, 21st to 26th inclusive was a cold spell, below 10C as a mean. Jan and Feb following had notable freezes.

I have had a bit more of a detailed look at cool ends to September and have come up with the following table

TABLE

This shows all years since 1875 where the CET for 21-30 Sept has been 11.2C or less. Currently 2018 looks like coming in around 11C.

I have then looked at the CET for the following 6 months and compared this to the 1971-2000 CET mean for each month. The colour coding is the same as I have used for my CET summary table going back to 1950. Light orange is 0.5-1.4C above the 1971-2000 mean. Dark orange is a 1.5c+ anomaly. Similar for below average with light and dark blue. Anything from -0.4C to +0.4C of the mean is coloured green.

You will notice that for the 34 years in question the average CET is below average for all months. This is not really surprising because many of these years go back 100 years or more when the CET was generally quite a bit lower than it is today. To put that in context the CET mean for 1871-1900 was 0.6C lower than 1971-2000.

Nevertheless even if you assume that the average CET mean over the whole period 1875-2000 was say 0.3C lower than the period 1971-2000 then none of the months in the table below would report an above average anomaly. So we could be in for an extended period of cooler weather, particularly when you factor in the solar minimum.

The anomalies for each month show that in fact October has the biggest negative anomaly followed by November and then March. So a strong signal for a cold October I think and probably quite a cold November as well.

Conversely looking at the winter months the signal is not that strong. Overall a negative anomaly of 0.4C and when you factor in the 0.3C I refer to above then essentially that is suggesting average conditions. However, that does not tell the whole story. The winter CET column shows only 1 very warm winter and 6 reasonably warm winters out of 34. Whereas there are 8 very cold winters and 7 fairly cold winters.

I also looked at averages for a narrower set of years where the late September CET was less than 10.5C and less than 10C. This shows that the colder the end of September is the colder October is as well. There is little effect for November, January or February. December is slightly warmer when the end of September is very cold, as is March. 

I then also factored in solar activity and looked at years with low sun spots. This also had little impact for most months except January and February which actually were somewhat warmer in low solar years, slightly surprisingly.

Finally I looked at the CET for the previous summer. Most of the 34 years had average to poor summers so the cool trend had already started. There was only one summer with a CET anywhere close to this year which was 1899. I took those summers with a CET of 15.5C or more. Again very little difference from the overall average except in December which was colder.

Finally I looked at summers with a CET of 16C or more. There are only 4 of these in the data set so clearly this does not give a very statistically significant result but it did throw up some different numbers for the winter. Both December and February were very cold for these 4 years overall and the winter CET was 1.2C below average. 

So in summary what I take from this is that we could now be in for a spell of 6 months of cooler weather to follow on from the very warm weather between March and August and average overall conditions in September. The solar minimum should help support that position. 

The closest matches to this year in the list are probably 1952, 1887 and 1899. All had low solar conditions and above average summer CETs. However, only 1952 is close to the late Sept CET for this year. The other years are much cooler. But 1952 had a summer only slightly above average and nowhere this year. The other 2 years were warmer. So a blend of these 3 years is probably the best match. All three years had cold winters as well as very cold conditions in March and October. 1899 had a warm November but the other years were very cold.

I am not, based on the above, suggesting we will have a cold winter. Going back to the full list of 34 years the overall CET for winter is very close to average. What stands out more is the likelihood of very cold weather in October and also fairly cool weather in November and March. The winter anomaly may not be that significant, although there are hints that where a cool end to September follows a very warm summer this can also increase the likelihood of winter seeing cold conditions.

Would be interested to see what the synoptic patterns look like overall for the above years. 

By ‘eck GW!    You really have put some thought and work into the above comparison!  I had to read it three times to take it all in and it’s very interesting. Thank you!  

Excellent work there GW. 

Will have a good study of that with my coffee later. 

Latest output suggests a somewhat less cold end to September. In particular Wednesday and Thursday will be fairly pleasant, albeit still only just above average as the nights remain relatively cool. The CET for 21-30 Sept now estimated to be above 11C at 11.21C.

That said the next three days will be chilly. Today and Tuesday will see a CET close to or just under 10C. Monday is the coldest day and could see a CET of only 8.7C. A CET under 9C in late September is not that unusual. We saw 8.3C on 22 Sept 2012, 8.1C on 25 Sept 2010 and 8.3C on 27 Sept 2007.

Current estimate for the final September CET has ticked up to 13.80C but still well under 14C.

Another look at years with very warm Apr - Jul periods. I have taken the top 23 in the CET series which is all years with a CET of 13.5C or more for those four months (remember 2018 is comfortably the warmest for this 4 month period).

What we see from the table below is that the average CET for the following winter is 3.55C which is about 1C below the 1971-2000 mean. Most of the very cold winters were back in the 1700s and early 1800s.

If we also filter by the October and November CET (taking an average CET for these months of less than 8.5C) then the winter CET for the matched years is 3.22C so a bit colder than the average for all years.

Based on the analogues I looked at yesterday there is a suggestion that October and November could be on the cool side this year.

Finally if we add an additional filter of low sunspots then the winter CET is even colder at just 2.67C. The sunspot data only goes back to 1750 so I can't analyse the earlier years in the 1700s.

Of the 5 matching years with all the filters added, if we look at the years post 1900, there are three of them, all three have a CET around 3.2C to 3.5C. 

So overall this data would suggest a reasonable chance of this winter seeing a CET of 1C or more below average.

TABLE

Met Office Hadley      14.7c.    Anomaly    0.8c.  Provisional to 22nd.

Metcheck                   14.63c     Anomaly    0.91c

Netweather                15.33c     Anomaly    1.62c

Darwen                      13.2c      Anomaly    - 1.98c From a 5 year average

Hexam                       13.3c      Anomaly     0.14c   

Mansfield  Weather      15.3c     Anomaly      1.52c

Peasedown St John      14.94c   Anomaly     -0.03c

1 Steady  9 Down today

Mean of my 10 stations   14.87c.                       

More interesting stats!  

For some time I’ve had a feeling that the coming winter is going to be cold, nothing scientific, just a gut feeling, and all the signs and stats seem to point that way too.  However, it’s been such a fantastic year for throwing surprises that I really wouldn’t be surprised if it turns out completely the opposite!  

I think at the end of this month we should archive this thread and revisit it in the Spring for a de-brief!  

[edit] Ahh, I think there might be a rogue zero in there!  It should be 14.7c.   

I guess, we can call that another famous "senior" moment from Arty!!

Blooper .0 now dropped from original post.. 

Met Office Hadley       14.6c.     Anomaly     0.7c.  Provisional to 23rd.

Metcheck                    14.39c    Anomaly     0.67c

Netweather                 15.1c      Anomaly     1.41c

Mansfield                    15.1c      Anomaly     1.32c

Peasedown St John 14.64c    Anomaly   1.32c

Treviskey Redruth 15.0c    Anomaly    0.26c.

1 up and 9 down.

14.72c Average of 10 stations.