1737: 17.4-13.8

2006: 19.7-16.1

1757: 18.4-15.2

1699: 18.2-15.0

So this year’s 19.1-16.6 (or thereabouts) will be quite a sharp drop but nothing unusual.

That 2006 drop really was a bit of a shock to the system, but 1737 must have been utterly depressing.

Hopefully it won't drop quite that far, my guess was 17.0C!

Met Office Hadley             17.2c.      Anomaly     1.4c. Provisional to 29th.

Metcheck                          17.10c     Anomaly     0.88c

Netweather                       17.71c     Anomaly     1.52c

Canvey island                    18.8c       Anomaly     0.6c

Clevedon Weather              18.1c       Anomaly     0.75c

Mansfield Weather              17.4c       Anomaly     0.8c

Peasedown St John             17.0c       Anomaly      -1.32c

Treviskey Redruth               16.7c       Anomaly      0.18c

mean Of My 10 Stations   17.32c.     

Thanks ART!  

We arrested the decline yesterday but I think tomorrow will see it drop to 17.1 as it was quite cold night last.

Another chilly night tonight probably moves us down to 17.0 for 31st and then with the usual .2 or .3 downwards correction we come out close to 16.7 exactly where GW has been pitching! 

I'm impressed that the CET is holding up so well; 17.2 would need to be followed by 2 x 14.2 to drop to 17, rounding excepted. But what will the likely adjustments mean? I remember the disappointment of adjustments in the Julys of 2013 and 2014. Even a 16.9 would just pip 2003 for the summer I think? Be very unlucky now to not beat 2006.

It doesn't always adjust down, so we'll see. Not at all confident that it won't, though, given GW's estimate!

Anyway, to distract myself from the ridiculously large late-stage crash in August CET, we have some entertaining September model output to admire. Next week is generally indicated to be warmer than average, but the midweek period is varying a lot between runs depending on the exact direction of flow and extent of cool air aloft circulating around from the Atlantic (i.e. how much of a gap between ridges we see just E of the UK for 2-3 days). For example, the 00z GFS gave me an estimate of 16.9*C to 7th which is over 1*C above the LTA, but the 12z GFS gave me one of just 15.8*C to that date, barley half a degree above the LTA.

Beyond that, the warm signal is actually more consistent, especially for the 8-14 day range. The 00z GFS produced 'moderate' levels of warmth (widely low-mid 20s) to lift the estimate very slightly; into the low 17s as of 14th September. The 12z GFS, by contrast, delivers some exceptionally high temperatures via an almost optimal positioning of high pressure (dependability alert!), rocketing the estimate into the low 17s by 11th and to very near 18*C as of 14th. That'd be quite some going!

Locally for me, that 12z GFS produces a 7-day (and counting as of run's end) stretch of mid-high 20s maximums following mid-teens minimums, which would leave me local mean very close to 19*C as of 14th. Oh what I'd give to secure that outcome, it would be something to remember for years to come!

When the September CET thread opens I'll move this across - or I might forget, in which case hello mods, please feel free to do that for me... .

I'll do that for you James.

GW

Attempted to send my Sept CET post of 14.7C to your inbox, but it is full.

Grandad

Met Office  Hadley        17.1c.       Anomaly       1.3c.  Provisional to 30th.

Metcheck                      16.99c      Anomaly       0.76c

Netweather                   17.57c      Anomaly       1.38c

Canvey Island               18.7c        Anomaly       0.5c

Mansfield Woodhouse     17.2c       Anomaly       0.6c

Peasedown St John        16.36c      Anomaly       -1.96c

Treviskey Redruth          16.5c.       Anomaly       0.98c.

Mean Of My Watched 10 Stations     16.40c.    

Sorry been rather busy the past couple of days. I have tidied up my inbox so you should now be able to send predictions through. Will start the September thread later this evening.

So in absolute classic fashion, today's runs have taken what looked like a decent strength signal for warmth and absolutely trashed it; GFS sticking to the warm theme but ECM determined to draw cool air in via a low dropping down to our east, and UKMO + GEM moving toward that outcome this evening.

Historical precedent favours GFS but only slightly; it's not enough to make for confident CET estimation. Difference between ECM and GFS in terms of CET to mid-month is in the order of 2*C which is [insert favourite cuss here] massive.

Madness!   

So here are my final figures for August. Final CET mean is 16.69C by my rough calculations based on the publicly available data.

Unadjusted individual stations means were as follows:

Stonyhurst 15.3C

Rothamsted 17.5C

Pershore College 17.3C

Final summer CET mean is 17.32C

Top 7 warmest summers of all time on the CET series

17.79C 1976

17.61C 1826

17.37C 1995

17.35C 2003

17.32C 2018

17.22C 2006

17.11C 1983

ECM Op run seems to have gone off on one this evening. Well out of sync in the latter stages of the run

http://www.wetterzentrale.de/de/show_diagrams.php?model=multi&var=2&geoid=49069&lid=OP&bw=

A great summer but 5th is a real tragedy IMO.

Good month for all the comments on CET. 

Thanks; very useful to have the individual stations to reference too. One thing- you've listed a rm of 16.78 to 30th unadjusted; any idea why  MetO has published 17.1 as its provisional mean to 30th on the HadCET site? Is that not the same as the unadjusted figure? I see that all of their estimated daily values are also higher for the first 10 days. Really helpful if someone can explain.

GW..

Thanks for your excellent analysis above.

we all are extremely grateful.

I will send my Sept CET figure by PM this morning, but I would appreciate it being regarded as being about 16:00 of yesterday, as I was out last evening and unable to send a PM then.

Grandad

Thanks GW.  A tour de force as usual.  All credit to you.

Top 5 given the length of the CET series is pretty good if you ask me

Personally I'd like to see a summer in the bottom 5 occur - now that'd be interesting We'd need daily summer moaning threads if that happened

Just seen the Met Office blog update that shows England coming in 1st or 2nd warmest - goes to show how much warmer it's been in the far south*, plus to a small extent the western bias to the CET region. The way that's behaved this month has indeed been frustrating. Especially for those at the high end of the estimates who had strong model support at all ranges as July finished (remember those sensational warm + dry EPS and CFSv2 means for the third week of the month?) along with backing from the major climate drivers based on typical cycles (it's the unusually large suppression and delay that's caused things to turn out so differently). 

I have long wondered how the competition would evolve if the errors were capped at 1.0*C for each month to make it more a measure of how many months you can do well in, rather than whether you can avoid having any howlers. Given the propensity of the climate to strongly defy most of us at least once a year, I can see that having quite the impact. 

Maybe, if I can find the time, I'll try to interpret this year's CET errors into Excel and have a go at making an experimental alter-table. Just out of interest and if GW doesn't mind me doing so.

* As an example of which - my local August mean finished at 17.93*C with the summer mean winding up at 18.53*C. That's the old leaders 1995 (18.13*C) and 1976 (18.12*C) thoroughly smashed.

August 2018 CET confirmed at 16.6

https://www.metoffice.gov.uk/hadobs/hadcet/cetml1659on.dat

16.64C is the figure to two decimals

https://www.metoffice.gov.uk/hadobs/hadcet/cet_mean_2018

Jerry P is our competition winner this month. His prediction was 16.6C

The figures in my table are adjusted figures using my methodology, not the raw data.

The unadjusted figures are the overall monthly means for the three individual stations that I put at the start of my post.