January 2021 C.E.T. and EWP forecast contests

[Timmytour]

Good weird stuff WH! 

[North-Easterly Blast]

2 hours ago, Weather-history said:

Here's one peculiar quirk, the last notably above average January for an odd number year was in 2007 with 7.0C

2009: 3.0C

2011: 3.7C

2013: 3.2C

2015: 4.4C

2017: 4.0C

2019: 4.0C

2021: 3.1C

The even year Januarys  since 2010

2012: 5.4C

2014: 5.7C

2016: 5.4C

2018: 5.3C

2020: 6.4C

 

 

Prior to 2007, January 2005 was also well above average at 6.0 CET, and at 7.0 CET, January 2007 was actually the fifth warmest in the entire CET series; but as you rightly say, since 2010, Januarys appear to have followed a checkerboard, alternating between the milder even years and closer to average, or occasionally colder (like 2013 and 2021) odd numbered years.

[North-Easterly Blast]

6 hours ago, BornFromTheVoid said:

The final CET came in at 3.1C so, 1.3C below the 81-10 average, 0.7C below the 20th century average and 0.1C above the 19th century average.

Compared with the 81-10 values, there were 9 days above and 22 days below average. One day reached the top 10 warmest (20th), but no days were in the bottom 10 coldest.

January 2021 confirmed as 3.14*C (3.1 rounded down) gives us the coldest January since 2010, and the first sub 3.5 January CET also since 2010.  11 years is a pretty remarkable timespan without a sub 3.5 January CET.  It was at least a cold month by today's standards, although pre 1988 it would only have been slightly below average.  The only significantly cold January in recent times was 2010.  Last month at least felt a decent winter month with generally colder than average conditions even though there was no big freeze.

[Weather-history]

3 minutes ago, North-Easterly Blast said:

January 2021 confirmed as 3.14*C (3.1 rounded down) gives us the coldest January since 2010, and the first sub 3.5 January CET also since 2010.  11 years is a pretty remarkable timespan without a sub 3.5 January CET.  It was at least a cold month by today's standards, although pre 1988 it would only have been slightly below average.  The only significantly cold January in recent times was 2010.  Last month at least felt a decent winter month with generally colder than average conditions even though there was no big freeze.

And colder than 8 Januarys of the 1960s....

[North-Easterly Blast]

13 minutes ago, Weather-history said:

And colder than 8 Januarys of the 1960s....

Yes you are right, apart from 1962-63 there were not any other notably cold winters overall in the 1960s.  The only other standout of the 1960s is that most of that decade saw Decembers on the cold side, whereas the Januarys and Februarys of that decade were not that particularly cold overall (although 1963 apart, February 1969 was also very cold, and Feb 1968 was pretty cold too).  So apart from the 1962-63 winter the 1960s only stand out as having mostly cold Decembers and a cold February in 1969. 

It was really the 1977-87 period that stands out as having cold Januarys and Februarys overall, as the whole of that period generally saw cold Januarys and Februarys with only occasional exceptions. 

Edited February 1, 2021 by North-Easterly Blast

[North-Easterly Blast]

3 minutes ago, North-Easterly Blast said:

 

 

Edited February 1, 2021 by North-Easterly Blast

[Relativistic]

27 minutes ago, North-Easterly Blast said:

January 2021 confirmed as 3.14*C (3.1 rounded down) gives us the coldest January since 2010, and the first sub 3.5 January CET also since 2010.  11 years is a pretty remarkable timespan without a sub 3.5 January CET.  It was at least a cold month by today's standards, although pre 1988 it would only have been slightly below average.  The only significantly cold January in recent times was 2010.  Last month at least felt a decent winter month with generally colder than average conditions even though there was no big freeze.

Not entirely true, as it would have been cold by early 20th-century standards. The 30-year rolling mean from 1894-1923 through to 1916-1945 remained at or above 4.1C, so 2021 was at least 1C cooler than the averages throughout this period. The same is true for the 1919-1948, 1920-1949, and 1921-1950 averages. The peak 30-year means during this period were 1898-1927, 1909-1938, 1910-1939, and 1911-1940, which were all 4.4C.

[North-Easterly Blast]

18 minutes ago, Relativistic said:

Not entirely true, as it would have been cold by early 20th-century standards. The 30-year rolling mean from 1894-1923 through to 1916-1945 remained at or above 4.1C, so 2021 was at least 1C cooler than the averages throughout this period. The same is true for the 1919-1948, 1920-1949, and 1921-1950 averages. The peak 30-year means during this period were 1898-1927, 1909-1938, 1910-1939, and 1911-1940, which were all 4.4C.

The period from 1940 to 1988 generally saw frequent cold winters, much more so than the 1895-1940 period (apart from 1916-17 and 1928-29), so what I was implying is that at least for the 1940-1988 period, a January CET of 3.1 like this year would not be much colder than average.

[Optimus Prime]

From memory, January 2009 felt a lot colder than the one just gone. But the greatest departure from average was measured in the SE where it was >1.5 °C below average and some unusually low overnight temperatures despite the lack of snow cover. This year there were some very mild CET days recorded there. 

Will be interesting to see how cold it was for Scotland. Perhaps 2nd coldest since the 90's after 2010.

[Walsall Wood Snow]

1 hour ago, North-Easterly Blast said:

 

It was really the 1977-87 period that stands out as having cold Januarys and Februarys overall, as the whole of that period generally saw cold Januarys and Februarys with only occasional exceptions. 

Although that period covers 11 January's, 2021 was still colder than 5 of them including 1978, 81, 83, 84 and 86. That's not too bad. It'd be interesting to know though how snowy Jan 21 would be ranked (especially from the Midlands north) in that period, or more especially compared to the 5 it was colder than.

Jan 84 had a CET 3.8c, so 0.7c milder than Jan 21. But that was apparently quite a snowy January in the north. I wonder if it was snowier than the one just gone. 

Jan 86 CET 3.5c had some snow near the beginning I believe but don't know more than that.

Jan 83 CET 6.7c, so I doubt there was much snow around that month

As for January's 81 CET 4.9c and 78 3.4c I've no idea how snowy or not they were, but I've heard there was a snowy period at some point in early 78, but that could have been in Feb for all I know. 

Speaking of Feb though, here's hoping Feb 21 turns out to be one that had it occurred in the 1977 to 87 period, would have been considered a good one for cold and snow.

 

[damianslaw]

Suspect Jan 2021 in Scotland was colder than 2009 by some margin.

For the Lake District it will go down as average on the snowfall front - in terms of number of days of snow falling, but in terms of depth below par..

Temperature wise despite no ice day, we came close on a number of days, and many days brought maxima below 2 degrees. The low maxima rather than particularly low minima has been the most notable feature of the month.

Looking back at past decades, the Januaries of the 1940s, 1950s were often cold or very cold. Januaries of the 1970s often mild, apart from 1978 and most notably 1979.

Januaries of the 1980s very mixed, 1982, 1985 and 1987 very cold, but some very mild ones 1983, 1988 and 1989.

Januaries of the 1990s often mostly mild, none were notably cold, coldest were 1991, 1997.

 

[Don]

1 hour ago, damianslaw said:

Suspect Jan 2021 in Scotland was colder than 2009 by some margin.

I think northern Scotland was milder than average in 2009?  I seem to recall this in the late Philip Eden's monthly review.

[phil b]

2.7C for my location with 15 frosts 109.2mm of rain and 7 snow events! interesting months weather that's for sure.

[Weather-history]

26th December 2020-25th January 2021 about 2.5C?  Is that right?

 

 

[Roger J Smith]

EWP Tracker has finished on 143 mm. Will post some scoring updates when the excel file is updated later today. 

The monthly winner was John88b, followed by Polar Gael second and moffat in third. 

The annual top five total scores are now shillitocettwo, snowray, JeffC, Feb1991blizzard and virtualsphere. 

More details in the attached excel file. 

(will also hunt down the lowest 31d average from Dec-Jan, separate post for that info)

EWP2020_21.xlsx

Edited February 2, 2021 by Roger J Smith

[Roger J Smith]

3 hours ago, Weather-history said:

26th December 2020-25th January 2021 about 2.5C?  Is that right?

 

 

2.5 is correct for any of 26 Dec - 25 Jan, to 28 Dec - 27 Jan. (2.549, 2.516, 2.542).

The middle one (27 Dec to 26 Jan) is coldest by second decimal (2.52). 

[North-Easterly Blast]

17 hours ago, Walsall Wood Snow said:

Although that period covers 11 January's, 2021 was still colder than 5 of them including 1978, 81, 83, 84 and 86. That's not too bad. It'd be interesting to know though how snowy Jan 21 would be ranked (especially from the Midlands north) in that period, or more especially compared to the 5 it was colder than.

Jan 84 had a CET 3.8c, so 0.7c milder than Jan 21. But that was apparently quite a snowy January in the north. I wonder if it was snowier than the one just gone. 

Jan 86 CET 3.5c had some snow near the beginning I believe but don't know more than that.

Jan 83 CET 6.7c, so I doubt there was much snow around that month

As for January's 81 CET 4.9c and 78 3.4c I've no idea how snowy or not they were, but I've heard there was a snowy period at some point in early 78, but that could have been in Feb for all I know. 

Speaking of Feb though, here's hoping Feb 21 turns out to be one that had it occurred in the 1977 to 87 period, would have been considered a good one for cold and snow.

 

I think that last month, a good number of parts of the country have seen some snowfall at times, although it rarely fell in large amounts and was never significantly widespread or lay on low ground at least, for any length of time.  January 2021 felt a decent although not a classic cold winter month, in that it saw generally colder than average conditions for the most part but there was not a big freeze.

January 1984 was fairly average CET wise, but saw frequent good snowfalls for northern britain and these spread further south during the second half of that month, so I would say that taking the UK as a whole Jan 1984 was snowier than Jan 2021.  January 1986 I believe saw snow for a good part of the country in the first third, and also in the final week I believe some areas saw snowfall, so possibly fairly similar snow-wise to Jan 2021.  Jan 1978 (only marginally less cold than Jan 2021) I believe saw some good snowfalls in the north at times, but I am not sure about how much further south.  Jan 1981 although milder than average I believe saw some short lived northerly outbreaks which did give some snow for the north but not in the south.  Jan 1983 despite being one of the mildest on record I believe did see some short lived snowfall in northern britain in the closing days of the month but that was that.

[Timmytour]

and still we can still look back over a period of ten years where the coldest month in that time was a March......I'm beginning to think the next time we get a month colder than March 2013, it will be another March!

[J10]

Spreadsheet January 21 CET.xlsx

Summary January 21 Summary.pdf

Monthly Competition

15 players were within 0.2c and two got it spot on.

Shillitocettwo and General Cluster, with the latter making a very wise decision to make a late change to 3.1c.

Seasonal Competition

All change at the top 3 General Cluster, Summer18 and nonameweather. 

Overall Competition

The same top 3 as the seasonal Competition.

Edited February 2, 2021 by J10

[Roger J Smith]

The 3.1 result ends the discussion about which anomaly is stranger, the missing 5.9 from December, or the formerly lone wolf 3.1 at almost the top of the bell curve of January CET values (in statistical terms, the 1832 lone wolf was probably less likely at random). The fact that the first quarter of the record was confined mainly to multiples of 0.5 would have to be considered in any analysis of that. Still, it's sort of the equivalent of flipping ten heads in a row in heads or tails. The missing 5.9, while remarkable perhaps, is after all towards the 25th percentile area of the December spread, albeit closing in on the median rather steadily in recent decades. I just had a look and found that the new lone wolf closest to the median is the coldest December (1855) at 2.4 which is almost exactly a mirror image of the missing 5.9 (if that happened this year it would rank 56th warmest). There are two other lone wolves that rank in the 40s.

The entire list of lone wolf CET values closest to the median values can be seen below (in my historical CET thread there's a list of the closest missing values, a lone wolf value is one sole outcome at the level shown -- it has to be kept in mind that a warming climate means that the colder lone wolves are probably further from the median than the warmer ones in terms of their chances of being joined at any time). 

There are other cases where a missing value appears before a lone wolf beside the missing 5.9 in December.  The lone wolf coldest March (2.3) is colder than missing 2.4 or 2.6. The lone wolf warmest March is warmer than the first missing value also. See notes below table for details and some other cases. 

 

<<< CET LONE WOLVES CLOSEST TO THE MEDIAN VALUES >>>

 

Month ____ Cold Lone Wolf _ rank ______ Warm Lone Wolf _ rank 

JAN _______ 0.2 1895 _______ 26th ________ 6.1 1989 _______ 20th

FEB _______ 1.6 1844 _______ 43rd ________ 6.6 1790 _______ 19th

MAR ______ 2.3* 1837 ______ 9th _________ 8.1^ 1834 ______ 10th

APR _______ 6.0 1908 _______22nd _______ 10.5 1943 _______ 4th

MAY _______ 9.8 1968 ______ 37th ________ 13.7 1808 ______ 5th

JUN _______ 12.6 1927 ______20th ________ 15.8 1857 ______ 31st

JUL ________14.4 1767 ______ 26th ________18.1 1794 ______ 21st

AUG ______ 13.8 1737 ______ 18th ________ 17.5 1705 _____ 20th

SEP _______ 14.8 1784 ______ 32nd ________10.7**1952 ____ 6th 

OCT _______ 7.6^^ 1782 _____ 25th  _______11.2 1706 _____ 42nd

NOV _______ 3.7a 1739 ______19th ________ 8.3 2018 ______ 22nd

DEC ________ 2.4 1855 ______ 58th _________ 7.6 1733 ______ 5th (see last note)

_______________________________________________________

* lone wolf coldest March (2.3, 1837, 9th coldest) is colder than two missing values, 2.4 which if it happened this year would be 10th coldest, or 2.6 which if it happened this year would be 12th coldest. There were two March CETs at 2.5. 

^ lone wolf warmest March (8.1, 1834, 10th warmest) is warmer than the missing value 8.0 which would rank 11th if it happened this year. 

** lone wolf coldest September (10.7, 1952, 6th coldest) is colder than missing values 10.8 and 10.9. Either of those would rank 7th coldest if they happened this year (which they won't). 

^^ lone wolf coldest October (7.6, 1782, 25th coldest) is colder than missing value 7.7 which would be 26th coldest if it ever happens (before any colder value). Seems very unlikely in the christmas pudding but you never know. 

a _ lone wolf coldest November (3.7, 1739, 19th coldest) is colder than missing value 3.8 which would be 20th coldest if it ever happens (before any colder value). Once again, seems unlikely.

Finally, 5th warmest Dec 1733 (7.6) is the warm lone wolf but 7.1, 7.0 and 5.9 are all missing values; the next time 7.1 or 7.0 are hit without an intervening warmer value they would rank 15th, the first time 5.9 hits will rank (56+n)th warmest where n is the number of 6.0+ Decembers before that happens. If it happens soon, its 56th to 60th rank won't be any different from current cold lone wolf December (58th coldest) or that much more unusual than current lone wolf outliers 42nd warmest October or 43rd coldest February. 

As an indicator of how unusual lone wolf 3.1 1832 was, until 2021 joined it, it ranked 149th coldest (now tied 149th with 2021).  

[Roger J Smith]

The final value for the EWP was posted this morning at 144.2 mm. That landed in 13th place overall for the 256 years 1766 to 2021. A table of the top 20 was posted back on page 12 or on 24 Jan if that makes it easier to find. 

This slight upward revision will not change the points listed in the earlier excel file but it will make minor differences to the average errors and change a few ranks there. If I get to making those edits to the table today, I will post the file. But if you wanted to check your points total for the month or two-month contest year to date, then scroll back a few posts and find my post on 2 Feb with the provisional value of 142 mm used at that time. 

I had a look at EWP and CET rankings and the forecasters with the best combined ranks after two months are as follows: 

FORECASTER ________ CET __ EWP ____ combined / avg

Summer18 ___________ 2 _____ T14 ____ 16 / 8.0

JeffC __________________15 _____ 3 ______ 18 / 9.0

Feb1991Blizzard _____ 19 _____ 4 ______ 23 /11.5

Reef __________________17 _____ 9 ______ 26 /13.0

2010cold ______________ 7 ____ T28 ____ 35 /17.5

Don __________________T23 ___ T14 ____ 37 /18.5

DR(S)NO ______________20 ____ 20 _____ 40 /20.0

Federico ______________ 6 _____ 34 _____ 40 /20.0

snowray ______________40 _____ 2 ______ 42 /21.0

General Cluster _______ 1 _____ 42 _____ 43 /21.5

Born from the Void ___22 ____ 22 _____ 44 /22.0

nonameweather ______ 3 ____ 41 _____ 44 /22.0

stewfox _______________ 8 ____ 36 _____ 44 /22.0

Mulzy _________________24 ____ 23 _____ 47 /23.5

bobd29 _______________31 ____ 16 _____ 47 /23.5

davehsug _____________29 ____ 18 _____ 47 /23.5

moffat ________________47 ____ 11 _____ 58 /29.0

 

Edited February 5, 2021 by Roger J Smith

[Roger J Smith]

This is the updated EWP scoring file. There were just a few minor changes to ranks of average error, points unchanged in general. 

EWP2020_21.xlsx