Jump to content

December 2018 CET forecasts and EWP contest -- start of 2018-19 competitions

Recommended Posts

On behalf of J10 and myself, welcome to another new competition year with the winter season of 2018-19 fast approaching. The contests will include the CET forecast and the (still optional) EWP forecast, with some guidance provided in two sections below ... 


CET averages and extremes for December

--- all CET values 1981 to 2017 are in the table --- they are in bold italic to make them easier to find

--- --- --- and colour coded for warmest, middle and coldest thirds of the 37 most recent years. 


13.1 ... warmest daily mean, 19th 2015

12.9 ... previous warmest daily mean, 12th 1994

 9.7 ... warmest December 2015
 8.1 ... tied second warmest Decembers 1934 and 1974
 7.7 ... fourth warmest December 1852

 7.6 ... fifth warmest December 1733

 7.5 ... 1988


 6.5 ... 2006

 6.4 ... 1994

 6.3 ... 1985, 2013
 6.2 ... 1986

 6.0 ... 2011, 2016

 5.9 ... has never happened in 359 tries

 5.8 ... 1997, 2000

 5.7 ... 2002

 5.6 ... 1983, 1987

 5.5 ... 1993, 1998
 5.4 ... 2004

 5.2 ... 1984, 2014

 5.1 ... average for 1971-2000*

 5.0 ... 1999

 4.9 ... average for 1988-2017, 2001-2017, 1991-2017 (developing 1991-2020) and 1989, 2007

 4.8 ... 2003, 2012, 2017
 4.7 ... average for 1961-90 and 1991
 4.6 ... average for 1981-2010, also 1901-2000
 4.4 ... 1982, 2005

 4.3 ... 1990

 4.1 ... average for 1659-2017 (all 359 years)
 4.0 ... average for 1801-1900
 3.9 ... average for 1701-1800
 3.6 ... 1992, 2001

 3.5 ... 2008

 3.4 ... average for 1659-1700

 3.1 ... 2009

 2.9 ... 1996

 2.3 ... 1995

 0.3 ... 1981

--0.4 ... tied fourth coldest 1788 and 1796
--0.5 ... third coldest 1676
--0.7 ... second coldest 2010
--0.8 ... coldest December 1890

-10.8 ... lowest daily mean 25th 1796


Enter your forecast before midnight on Friday 30 November, or with the increasing late penalties for entries 1-3 Dec.

*1971-2000 is in fact the highest 30-year average for December in the CET series, and the only one that rounds to 5.1, from 5.06. With the exception of June, this is the furthest back one has to go to find the maximum value. For most months it is one of the past three (1987-2016, 1988-2017 or 1989-2018). For June, however, it was 1822-51 (14.72) followed closely by an earlier peak, 14.70 for 1772-1801. Last June was warm enough to boost the running 30-year average to 14.64, its highest value since 1825-54. In between the peaks, the June average got as low as 13.8 (1900-29), rising to 14.57 for 1931-60 and 1932-61, then falling back to 14.07 as recently as 1971-2000. 

As you might expect, the Maunder period produced cold Decembers, the mean was as low as 3.2 (1665 to 1694). It then rose as high as 4.29 by 1709-38. There was a colder period than any in the Maunder; the lowest 30-year average was 3.09 from 1783 to 1812; rounded off 1784 to 1813 was equal.

The running mean climbed as high as 4.61 for 1823-52 before falling back to 3.48 for 1867-96. Decembers in the first third of the 20th century were generally a bit milder than the mid-century period, means as high as 4.73 were reached in 1893-1922 and 1907-36. The lowest since those peaks was 4.33 for 1923-52. The last time that the running mean was 4.5 was 1957 to 1986. The table above tracks the changes since then, but worth noting that the cold December of 2010 dropped the running mean from 4.84 to 4.63 but it then went back up to 4.82 for 1982 to 2011 as that period dropped the cold December of 1981 (mean 0.3). 1981-2010 was the only 30-year mean at 4.6 after 1960-89. 

_____ _ = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = _ ________


Optional Precip Contest -- details

The EWP contest now enters its second year. It remains entirely optional, in the first year about two-thirds of the regular temperature forecasters entered this second contest. (there were no precip-only forecasters). 

As mentioned last year -- scoring will be handled by myself, J10 has lots to do already and so this is more or less a separate contest.  

The contest asks you to predict an amount in mm (not a percentage) for this statistic.

After using the 1910-2017 "NCIC" precip values for the first contest, and scoring Hadley as an alternative, I will just go with the somewhat later appearing Hadley values for this contest year, to reduce the complexity of scoring and reporting. You will have to be patient as Hadley posts their numbers around the 5th of each month (and last February it was more like the 10th, which is what got me over to using NCIC). 

For December, here are the averages and extremes for your guidance. Data (Hadley only from now on) go back to 1766.


Max on record ____193.9 mm (1876)

Max since 1981 ___ 175.2 (2012)

Mean 1988-2017 __ 100.1 mm

Mean 1981-2010 __ 97.4 mm

Min since 1981 ___ 34,1 mm (2010)

Min on record _____ 8.9 mm (1788)


Your forecast should be in mm with one decimal allowed. If you enter a percentage, that will be applied to 1981-2010 mean of 97.3 to derive your contest entry. Any number that appears without a % sign will be taken as whole number of mm. It will be up to you to clarify your contest entry.

Deadlines are the same as the temperature contest (will drop the three hour extension feature this contest year).

Scoring will change slightly from the first contest year.

Top score will be 10.0 regardless of error (we never got into a situation where it was significantly different from existing ranked values). Lowest score will be 0.0, and all other scores will be calculated from step values. This contest year, order of entry will affect your score, but each later (equal ranked error) forecast will drop by 0.1 rather than the full step value (which is often 0.2 or 0.3). The lower scores after these ties will resume at a point determined by full step value drops. Example, three are tied at 7.0 by step values, so they score 7.0, 6.9 and 6.8. Then the next score will be 6.4 (as though three steps had been used).

Late penalties are applied as follows: 0.2 per day late. Example, you score 8.4, but enter on second late day, penalty of 0.4 reduces that score to 8.0. No score will be negative after this rule applies.

The table of ranked monthly amounts (for EWP) can be found here. With a bit of navigating you can also find other products of the precip product. The contest will only use end of month values posted, any later quality control will not be used for recalculation.

I hope we get an even stronger contingent of our temperature regulars involved in this precip forecast. 



If entering both contests, just post once (not a requirement, you can post separate but try to remember to use mm)

A sample "good" entry would be

5.0 C, 82.5 mm

(that's not my entry, I will enter around the 30th). 

Good luck and thanks again on behalf of all of us to J10 for scoring the temperature forecast contest. 

(he may want to let you know about any scoring plans for 2018-19, I have assumed we are just going forward status quo). 

Edited by Roger J Smith
  • Like 1
  • Thanks 5
Link to post
Share on other sites
  • Replies 287
  • Created
  • Last Reply

Top Posters In This Topic

Top Posters In This Topic

Popular Posts

On behalf of J10 and myself, welcome to another new competition year with the winter season of 2018-19 fast approaching. The contests will include the CET forecast and the (still optional) EWP forecas

Another mild washout whimper to end another naff year and my worst year ever.  Rest in pieces 2018!  12C    Last sub 0C, last sub 1C, last sub 2C, last sub 3C and last sub 4C December all in 2010.

Hadley is confirmed as 6.9C. Joint 15th warmest. The annual CET was 10.69C, 4th warmest.

Posted Images

Another mild washout whimper to end another naff year and my worst year ever.  Rest in pieces 2018!  12C  :oldsad:  Last sub 0C, last sub 1C, last sub 2C, last sub 3C and last sub 4C December all in 2010.

200mm of Atlantic garbage

Edited by Lettucing Gutted
  • Like 2
  • Thanks 4
Link to post
Share on other sites

Stick me at an average 4.6C. 

My analogues (2002, 2004, 2008, 2009) basically split into two camps however. One goes about a degree warmer and the other colder. 

2002 got unlucky (blocking to our north and east but like the coming week, it looks likely we had an insufficient undercut). 

2009 saw a monster -AO deliver.

2008 although it records a +AO actually saw strong blocking over and north/east of the UK so the AO value does not tell all. 

2004 actually saw a few topplers.


So i will stick with average but expect we might end below. 

Link to post
Share on other sites

Using nonsense logic for my guess. I'm looking at the difference between July and August CETs  (this year 2.5C ) and looking at Decembers in the years where the difference has been greater than 2.2C (26 years)

In all those years the corresponding December has never been brutally cold, and only two of them have been below 2.7C  (2C in 1677 and 1.8C in  1830). In the last 100 years the corrsponding December has only once been below 5.2C

So, expecting a relatively mild December as a result, i am going for the average of the 7 corresponding Decembers over the last 100 years, despite the fact it has never come in as the December CET!!! 

My guess  is 5.9C    (77mm) 


  • Like 1
Link to post
Share on other sites

Again very early for me to be calling a month, I will go with a below average 3.1 degrees - being cautious, good chance we could see something in the low 2s. A mild start will be deceptive, and suspect the models will underestimate the strength of the block to the north and east.

  • Like 1
  • Thanks 1
Link to post
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...